US Army Corps
       of Engineers
       Construction Engineering
       Research Laboratory
                      Sponsored By
                     University of Pittsburgh
                    In Cooperation With
U.S. Environmental

Protection Agency
 U.S. National
Science Foundation
Proceedings:
•••••••

FIRST INTERNATIONAL CONFERENCE
ON FIXED-FILM BIOLOGICAL PROCESSES
April 20-23,1982
Kings Island, Ohio
Edited by Y.C. Wu, Ed D. Smith,

     R.D. Miller, and EJ. Opatken
                       Vol.   I

-------

-------
Members Of Organizing Committee:


Yeun C. Wu (Chiarman)
Department of Civil Engineering
University of Pittsburgh
Pittsburgh, Pa,


James V. Basilico
Office of Research and Development
U. S. Environmental Protection Agency
Washington, D.C.

Ed. J. Opatken
Wastewater Research Division
U.S. Environmental Protection Agency
Cincinnati, Ohio


Ed. D. Smith
Environmental Division
U.S. Army Construction Engineering
Research Laboratory
Champaign, Illinois

Ed. H. Bryan
Civil and Environmental Engineering
National Science Foundation
Washington, D. C.

Roy D. Miller
Environmental Health Engineering Branch
U.S. Army Environmental Hygiene Agency
Fort Meade, Maryland

Richard Dick
Department of Civil Engineering
Cornell University
Ithaca, New York

-------
                                 DISCLAIMER
     These proceedings have been reviewed by the US Army Construction
Engineering Research Laboratory, the University of Pittsburgh and the
US Environmental Protection Agency, and approved for publication.  Approval
does not signify that the contents necessarily reflect the views and policies
of the US Army Construction Engineering Research Laboratory, the University
of Pittsburgh or the US Environmental Agency, nor does mention of trade
names or commercial products constitute endorsement or recommendation for
use.
                                      ii

-------
                         FOREWORD
     Biological wastewater treatment has been practiced in many
forms since the early part of this century, but fixed-film bio-
logical processes have only recently been more intensively stud-
ied and applied by water pollution control researchers and engi-
neers,  Because of some inherent advantages over suspended growth
processes, today there is a greater interest in fixed-film bio-
logical treatment processes than ever before.  This Conference
was designed to provide a forum for that interest and to help
accelerate further development of this technology.
     The objective of this Conference was to assess the State of
Knowledge and identify the research needs regarding the full
spectrum of fixed-film biological processes.  The Conference
addressed many new approaches to anaerobic as well as aerobic treat-
ment.  Many practical applications and new research findings
were presented and many of the speakers expressed optimism for
significant progress in the future.  Because of their keen interest
and the dedication of those who attended, this Conference was
truly a professionally stimulating experience.  There was much
interaction and exchange between all participants.
     The  Conference consisted of 13 technical sessions with a
total 80 presentations, one workshop on research needs for fixed-
film biological wastewater treatment processes, and a field tour
to the LeSourdsville Regional Rotating Bioloigcal Contactor Plant.
The Conference Proceedings consisted of 77 of those papers present-
ed by the authors.   More than 300 participants representing a wide
spectrum of researchers and practitioners attended the Conference,
Worldwide interest was also evident from the 31 foreign participants
who traveled from Canada, India, Saudi Arabia, Yugoslavia, Japan,
Norway, Switzerland, Republic of China, Italy, South Korea, France,
Belgium, England, West Germany, and Scotland,
     The material presented herein is published as submitted by the
authors.  No attempt was made by the Conference Co-sponsors to edit,
reformat or alter the material provided except where necessary for
production requirements or where obvious errors were detected.  Any
statements or views here presented are totally those of the authors,
and are neither condoned nor disputed by the Conference Co-sponsors,
Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
May 24, 1982                          James V, Basilico
                                      Yeun C, Wu

-------
                          ABSTRACT
     The First International Conference on Fixed-Film Biological
Processes was held at the Kings Island Resort, Kings Island, Ohio
on April 20-23, 1982,  This Conference serves as an opportunity
to assess the applicability of this advanced technology for the
treatment of municipal and industrial wastewaters.
     The proceedings are essentially the papers and discussion
given by authors and participants.  The papers are divided into
13 major topic areas;


                 1, Current Status and Future Trends
                 2, Biofiltn and Btoktnetics
                 3, Concepts and Models
                 4, Small Scale/On Site Systems
                 5. Municipal Wastewater Treatment-
                    Case Histories
                 6, Nitrification and Denitrification
                 7, Industrial Wastewater Treatment
                    Part I-* Rotating Biological Contactors
                 8, Industrial Wastewater Treatment
                    Part II- Biofiltration, Packed Bed
                    Reactors
                 9, Innovative Research
                10. Aerobic and Anaerobic Treatment-
                    Submerged Media Reactors
                11, Industrial Wastewater Treatment
                    Part III- Submerged, Anaerobic
                    Fixed-Film Reactors
                12, process Evaluation and Design
                13, Experiences With Fixed-Film
                    Treatment Facilities


     The discussion occurred during the Research Needs Workshop
was taped and printed as an appendix.  This document was submitted
in fulfillment of Research Grant No, DACW88-81-R-005 by the Uni-
versity of Pittsburgh under the sponsorship of the U.S. Environ-
mental Protection Agency, the U, S, Army Construction Engineering
Research Laboratory, and the U, S. National Science Foundation.

-------
                      ACKNOWLEDGEMENT
     The Conference Organizing Committee would like to
acknowledge the invaluable contributions of the individuals
listed below.

     We thank all  keynote speakers Dr. Ed D. Smith of the U.S.
Army Construction  Engineering Research Laboratory, Dr. William
Jewell of Cornell  University, Dr. Ed D. Schroeder of the Uni-
versity of California at Davis, and Dr. Mark Williams of the
University of Pittsburgh.
     We thank Dr.  Joel I. Abrams of the University of Pittsburgh,
Dr. Roy D. Miller  of the U.S. Army Environmental Hygiene Agency,
Dr. John A. Roth of Vanderbilt University, Mr. Marvin E. Lambert
of Columbus City Utility in Ohio, Mr. Dick Brenner of the U.S.
Environmental Protection Agency, Dr. Michael Saunders of Georgia
Institute of Technology, Dr. Ed D. Smith of the U.S. Army
Construction Engineering Research Laboratory, Dr. A. A. Friedman
of Syracuse University, Mr. Michael Sweet of Engineering Science
Ltd in Ohio, Mr. James V. Basilico of the U.S. Environmental
Protection Agency, Mr. Ed D. Opatken of the U.S. Environmental
Protection Agency, Dr. Richard Dick of Cornell University,
Dr. Hallvard Odegaard of the University of Trondeheim, and
Dr. John Bandy of  the U.S. Army Construction Engineering Research
Laboratory for presiding all technical sessions.  The sincere
appreciation of the Organizing Committee to Dr. A. F. Gaudy, Jr.
of the University  of Delaware, Professor Wesley W. Eckenfelder of
Vanderbilt University, Dr. C. P. Leslie Grady, Jr. of Clemson
University, and Dr. A. A. Freidman of Syracuse University for
chairing the workshop on research needs for fixed-film biological
processes.
     The Organizing Committee is indebted to the following manu-
facturers for their support in equipment exhibit.  These companys
are:  A.O.S. Smith Company, B.F. Goodrich, Crane Company, CMS equip-
ment Limited, Mass Transfer, Inc. , Mid-south Distributor, Munters
Corporation, and Neptune Microfloc.
     Assistances from Mrs. Joyce Wingham, Mrs. Diana Casteel of the
Kings Island Resort, Mrs. Reita Bender of the U.S. Environmental
Protection Agency, and Mrs. D. Dixion of the Cincinnati Convention
Bureau are greatly appreciated.
     The co-editors thank Ms. Debra Moore and Ms. Lynn Smith for
their design and artwork for the proceedings cover.
                             v

-------
Keynote Speakers;


Mark Williams
Dean, School of Engineering
University of Pittsburgh
Pittsburgh, Pa,

Ed. D. Smith
Environmental Division
U. S. Army Construction Engineering
  Research Laboratory
Champaign, Illinois

William Jewell
Department of Agricultural Engineering
Cornell University
Ithaca, New York

Ed, D. Schoreder
Department of Civil Engineering
Univesrity of California
Davis, California
Conference Assistants

John C. Kennedy
Chung C. Chen
Shen Y. Lien
Sin N. Hsieh
Jeff Greenfield
Li L. Lin
              Department of Civil Engineering
              University of Pittsburgh
              Pittsburgh, Pa,

T. Casteel                  J. Wingham

              Kings Island Resort
              Kings Island, Ohio

Session Chairman;

Joel I, Abrams
Department of Civil Engineering
Univesrity of Pittsburgh
Pittsburgh, Pa.
                            VI

-------
Roy D. Miller
Environmental Health Engineering Branch
U.S. Army Environmental Hygiene Agency
Fort Meade, Maryland

John A. Roth
Center for Environmental Quality Management
Vanderbilt University
Nashville, TN

Richard Dick
Department of Civil Engineering
Cornell University
Ithaca, New York

Hallvard Odegaard
Department of Sanitary Engineering
University of Trondeheim
Trondheim-NTH, Norway

Marvin E. Lambert
Columbus Gas utility
Columbus, Ohio

Dick Brenner
Municipal Environmental Research Lab.
U.S. Environmental Protection Agency
Cincinnati, Ohio

Michael Saunders
School of Civil Engineering
Georgia  Institute of Technology
Atlanta, 6A

Ed. D. Smith
Environmental Division
U.S. Army Construction Engineering Research Lab,
Champaign, IL

A. A. Friedman
Department of Civil Engineering
Syracuse University
Syracuse, New York

Michael Sweet
Engineering Science Ltd.
Cleveland, Ohio

-------
James V. Basilico
Office of Research and and Development
U.S. Environmental Protection Agency
Washington, D.C.

Ed. J. Opatken
Municipal Environmental Research Lab.
U.S. Environmental Protection Agency
Cincinnati, Ohio

John Bandy
Envrionmental Division
U.S. Army Construction En gineering Research Lab,
Champaign, IL.


Workshop Organizers:


A. F. Gaudy, Jr.(Chairman)
Department of Civil Engineering
University of Delaware
Newark, Delaware

Ed. D. Opatken
Municipal Environmenatl Research Lab.
U. S. Environmental Protection Agency
Cincinnati, Ohio

A. A. Friedman
Department of Civil Engineering
Syracuse University
Syracuse, New York

W. W. Eckenfelder, Jr.
Department of Environmental Engineering
Vanderbilt University
Nashville, TN

C. P. Leslie Grady, Jr.
Department of Environmental Engineering
Clemson University
Clemson, South Carlonia

Yeun C. Wu
Department of Civil Engineering
University of Pittsburgh
Pittsburgh,  Pa,
                    viii

-------
                 Table of Contents

   CONFERENCE ORGANIZING COMMITTEE „ o o . o o o . . . o o » o o . o » o o o o     i
   DISCLAIMER. . . „ <, ... o .... » . ° . „ ... o .,. » . „ . . . o o . „ , „ . . o o „ .    ^
   FOREWORD o o . . . o . o o o . o o . o . o . . o ..... o . o o o . o . o . o o . . o .....   ii:L
   ACKNOWLEDGEMENTS 0 . o o . o . . . o . » „ . . » . » » » . o » . » 0 o o <, » » <, , o . o .     v
   KEYNOTE SPEAKERS ....................... ..... .........    vi
PART I:  GENERAL SESSION

   Keynote Address
   "State of Knowledge for Rotating Biological Contactor
   Technology" 0 . . o . . » 0 o » ..... , . . . , 0 . • .............. o . . . . o     1
     Ed D. Smith

   "Anaerobic Attached Film Expanded Bed Fundamentals". „    17
     William J. Jewell

   "Trickling Filters:  Reliability, Stability and
   Potential Perf ormance". „ 0 0 . o . o <> 0 .<>... 0 o ... .o o o. 0 .° o 0 .    ^3
     E. D. Schroeder
PART II:  CURRENT STATUS AND FUTURE TRENDS

   "The History of Fixed-Film Wastewater Treatment
   Systems" . . 0 . » ...... o0°o.. .0.0.000.. .00 ........ ooo.oo.    60
     Robert W. Peters and James E. Alleman

   "Development of Synthetic Media For Biological
   Treatment of Municipal and Industrial Wastewaters", „ .    89
     Edward H0 Bryan

   "Current Status and Future Trends of Rotating
   Biological Contactor in Japan". 0 ° o . 0 ». o . o „ .. o o o 0 . o . o .
     Masayoshi Ishiguro

   "RBC Unit:  Best in Sewage Treatment for
   Saudi Arabia" ..... „ «, . „ „ . . . „ . „ . „ ..... , . . . „ . . . . 0 » . o . . . .    132
     Sharaf Eldin I. Banna ga

   "The Future of Biological Fixed-Film Processes and
   Their Application  to Environmental Problems". ». „<>... o
     Stanley L0 Klemetson and Gary L0 Rogers
                         ix

-------
PART III:  BIOFILM AND BIOKINETICS

   "Processes Involved in Early Biofilm Formation"..0..o    155
     James D. Bryers

   "The Microbiology of Rotating Biological Contactor
   Films".	    184
     Nancy E0 Kinner, David L. Balkwill and
     Paul L. Bishop

   "Rotating Biological Contactors - Second Order
   Kinetics"	„......... „	    210
     Edward J0 Opatken

   "Assessments of the Kinetic Performance of a
   Rotating Biological Contactor System"...........„o..o    233
     Ta-Shon Yu and Randolph G, Denny

   "The Kinetics of Rotating Biological Contactors at
   Temperatures:  5°C,, 15°C, and 20°C"	    261
     Abraham Pano and E. Joe Middlebrooks

   "Kinetics and Simulation of Nitrification in a
   Rotating Biological Contactor".	„...    309
     Yoshimasa Watanabe, Kiyoshi Nishidome,
     Chalermraj Thanantaseth, and Masayoshi Ishiguro

PART IV:  CONCEPTS AND MODELS

   "Selection and Optimization Protocols For Attached
   Growth Biological Packed Columns"	.........    331
     Sheldon F« Roe, Jr0, and Edward B. Hanf

   "Modeling of Biological Fixed Films - A State-of-the-
   Ar t Review"...........................................    344
     C. P. Leslie Grady, Jr.

   "Investigation of Some Parameters in RBC Modeling"...   405
     Khalil Z. Atasi and Jack A. Borchardt

   "Analysis of Steady State Substrate Removal Models
   For the RBC"	   438
     David E. Schafer, James C0 O'Shaughnessy, and
     Frederic C. Blanc

-------
   "Mathematical Modeling For Assessing Development
   and Sloughing of Fixed Films and Their Effects
   on Waste Stabilization"	
     Ju-Chang Huang,  Shoou-Yuh Chang,  and Yow-Chyun Liu

   "Evaluation of RBC Scale-Up"	   474
     Yeun C. Wu, Ed D. Smith, Chiu Y.  Chen,  and'
     Roy Miller

PART V:  SMALL-SCALE/ON-SITE SYSTEMS

   "Small Wastewater Treatment Systems Using Soil
   Purification Method"	   487
     Masaaki Niimi

   "A New Fixed-Film System Covered by Surface Soils"...   516
     Tsutomu Arimizu

   "Study of Fixed-Film Biological Contactors For
   Recreational Area Wastewater Treatment Application"..   524
     Calvin P. C. Poon, Edgar D. Smith, and
     Vicki A. Strickler

   "Start-Up and Shock Loading Characteristics of a
   Rotating Biological Contactor Package Plant"	   542
     Farley F. Fry, Tom G. Smith, and Joseph H. Sherrard

   "Upgrading With Submerged Biological Filters"	   570
     Orval Q. Matteson            •<

PART VI:  MUNICIPAL WASTEWATER TREATMENT - CASE HISTORIES

   "RBC For BOD and Ammonia Nitrogen Removals at
   Princeton Wastewater Treatment Plant"	   590
     Shundar Lin, Ralph L. Evans, and Warren Dawson

   "Upgrading Activated Sludge Process With Rotating
   Biological Contactors"	   617
     Roger C. Ward and James F. Coble

   "Use of Supplemental Aeration and pH Adjustment to
   Improve Nitrification in a Full-Scale Rotating
   Biological Contactor System"	   "33
     James L. Albert
                          XI

-------
   "Application of Rotary Screens, Biological
   Contactors, and Gravity Plate Settlers to Treat
   Wastewaters in Hoboken and North Bergen,
   New Jersey"	   658
     Joseph M. Lynch, Jiunn Min Huang, and
     C. H. Joseph Yang

   "An In-Depth Compliance and Performance Analysis
   of the REG Process at Municipal Sewage Treatment
   Plants in the United States"..	,.   697
     Robert J. Hynek and Richard A. Sullivan

   "The Use of Plastic Media Trickling Filters - Two
   Case Histories"	   708
     Felix F. Sampayo

PART VII:  NITRIFICATION AND DENITR1F1CATION

   "Nitrification of a Municipal Trickling Filter
   Effluent Using Rotating Biological Contactors"	
     Frederic .C. Blanc, James C. O'Shaughnessy,
     Charles H. Miller, and John E. O'Connell

   "Improvement of Nitrification in Rotating Biological
   Contactors by Means of Alkaline Chemical Addition"...
     James M. Stratta, David A. Long, and
     Michael C. Doherty

   "Simultaneous Nitrification and Denitrification
   in a Rotating Biological Contactor"	   802
     Sumio Masuda, Yoshimasa Wantanabe,  and
     Masayoshi Ishiguro

   "Denitrification in a Submerged Bio-Disc System
   With Raw Sewage as Carbon Source"	   823
     Bjorn Rusten and Hallvard Odegaard

   "Operation of a Retained Biomass Nitrification
   System For Treating Aquaculture Water For Reuse"	   845
     D. E. Brune and R. Piedrahita

   "Nitrified Secondary Treatment Effluent by
   Plastic-Media Trickling Filter"	   87°
     Jiumm Min Huang, Yeun C. Wu and Alan Molof
                          xix

-------
PART VIII:  INDUSTRIAL WASTEWATER TREATMENT

   "Upgrading Slaughterhouse Effluent With Rotating
   Biological Contactors"'	    892
     Torleiv Bilstad

   "Evaluation of an Anaerobic Rotating Biological
   Contactor System For Treatment of a Munition
   Wastewater Containing Organic and Inorganic
   Nitrates"	    913.
     Leonard L. Smith

   "Application of Rotating Biological Contactor (RBC)
   Process For Treatment%of Wastewater Containing a
   Firefighting Agent  (AFFF)"	•..    927
     Susan Landon-Arnold and Deh Bin Chan

   "Operation of an RBC Facility For the Treatment of
   Munition Manufacturing plant Wastewater"	    944
   '  Leonard L. Smith  and Wayne G. Greene

   "Treatment of Starch Industrial Waste by RBCs"	    960
     Chun Teh Li, Huo'o Tein Chen, and Yeun C. Wu

   "Inhibition of Nitrification by Chromium in a
   Biodisc System"	    990
     Shin Joh Rang and Jack A.  Borchardt

PART IX:  INDUSTRIAL WASTEWATER TREATMENT '

   "Scale-Up and Process Analysis Techniques For Plastic
   Media Trickling Filtration"	   100?
     Thomas P. Quirk and W. Wesley Eckenfelder, Jr.

   "Treatment of Coke  Plan? Wastewaters in Packed Bed
   Reactors".	   1042
     Meint.Olthof, Jan'Oleszkiewi.cz, and
     William R. O'Donnell

   "Trickling Filter Expansion  of POTW by Snack Food
   Manufacturer"	;	   106°
     Michael R. Morlino,  Sajiel M. Frenkil,  and
     Paul Trahan

   "The Evaluation of.a  Biological Tower For Treating
   Aquaculture Wastewater For Reuse"	,	   1071
     Gary L. Rogers and  Stanley L. Klemetson
                          xin

-------
   "Bio filtration of Tannery Wastewater" ............ ....   1093
     Ahmed A. Hamza, Fahray M. El-Sharkawi, and
     Mohamed A. Younis

PART X:  INNOVATIVE RESEARCH

   "Effect of Periodic Flow Reversal Upon RBC
   Performance". , .......... ...... .................... ...   1113
     John T. Bandy and Manette C, Messenger

   "An Assessment of Dissolved Oxygen Limitations and
   Interstage Design in Rotating Biological Contactor
   (RBC) Systems" .......................................   1121
     Warren H. Chesner, John J. lannone, and
     Jeremiah J. McCarthy

   "Combined Biological/Chemical Treatment in RBC
   Plants"... .................. .............. ...........   H39
     Hallvard Odegaard

   "Treatment of Domestic Sewage by Aquatic Ribbon
   System" ................................... . ........ , .
     Chun-Teh Li, James S. Whang, and T. N. Chiang
   "Activated Fixed Film Biosys terns in Wastewater
   Treatment" ................ . ................ . .........  1 1 75
     John W. Smith and Hraj A. Khararjian

   "Comparison of Fixed-Film Reactors With a Modified
   Sludge Blanket Reactor" ......... ... ............ . .....  1 192
     Andre* Bachmann, Virginia L. Beard, and
     Perry L. McCarty

PART XI:  AEROBIC AND ANAEROBIC TREATMENT - SUBMERGED   .
          MEDIA REACTORS

   "Treatment of High-Strength Organic Wastes- by
   Submerged Media Anaerobic Reactors State-of-the-Art
   Review" ................... ... ........ . . ..............  12 12
     Yeun C. Wu, John C. Kennedy, A. F. Gaudy, Jr., and
     Ed D. Smith

   "Alcohol Production With the Bacterium Zymomonas" . . . .  1239
     Robert A. Clyde

   "Dynamics and Simulation of a Biological Fluidized
   Bed Reactor" .........................................   1247
     David K. Stevens, P. M. Berthouex, and
     Thomas W, Chapman
                        xiv

-------
  "Hydrodynamics of Fluidized Bed Reactors For
  Wastewater Treatment"...	  1288
    Boris M. Khudenko and Rocco M. Palazzolo

  "Retention and Distribution of Biological Solids in
  Fixed-Bed Anaerobic Filters"	  1337
    Mohamed F.  Dahab and James C. Young

  "Application  of Standard Rate and High Rate
  Anaerobic Treatment Processes"	  1352
    William F.  Owen

  "Application  of Packed-Bed Upflow Towers in Two-
  Phase Anaerobic Digestion"......	  1392
    Sambhunath  Ghosh and Michael P. Henry

ART XIII:  INDUSTRIAL WASTEWATER TREATMENT

  "Performance  Characteristics of Anaerobic Downflow
  Stationary Fixed Film Reactors".	  1414
    L. van den  Berg and K. J. Kennedy

  "Tannery Effluent:  A Challenge Met by Anaerobic
  Fixed Film Treatment"	  143?
    A. A. Friedman, D. P. Dowalski, and D. G. Bailey

  "Anaerobic Fluidized Bed Treatment of Whey:  Effect
  of Organic Loading Rate, Temperature and Substrate
  Concentration"	  1456
    Robert F. Hickey

  "Treatment of Phenol With an Innovative Fluidized
  Bed Activated Carbon Anaerobic Filter".	  1476
    Sheng S. Cheng and Edward S, K. Chian

  "Anaerobic Treatment of Landfill Leachate by an
  Upflow Two-Stage Biological Filter"....	  !495
    Yeun C. Wu, John C. Kennedy, and Ed D. Smith

  "Energy Recovery From Pretreatment of Industrial
  Wastes in the Anaerobic Fluidized Bed Process"	  1521
    Alan Li, Paul M. Button, and Joseph J. Corrado
                       xv

-------
PART XIV:  PROCESS EVALUATION AND DESIGN

   "The Hydrodynamic Evaluation of a Fixed Media
   Biological Process"	  1542
     Euiso Choi and Carl E. Burkhead

   "The Effects of Hydraulic Variation on Fixed Film
   Reactor Performance"	„	  1566
     Roy 0. Ball

   "Importance of Ecological Considerations on Design
   and Operation of Trickling Filters"	  1599
     Peter A. Wilderer, Ludwig Hartmann, and
     Thomas Nahrgang

   "Evaluation of Biological Tower Design Methods"	  1623
     Don F. Kincannon

   "Anaerobic Biofiltration - Process Modification and
   System Design"	  1644
     Jan A. Oleszkiewicz and Meint Olthof

   "Rotating Biological Contactor Scale-Up and Design"..  1667
     Enos L. Stover and Don F. Kincannon

PART XV:  EXPERIENCES WITH FIXED-FILM TREATMENT FACILITIES

   "RBC Supplemental Air:  Continuous or Intermittent?
   Youghiogheny Wastewater Treatment Plant,
   North Huntingdon Township, Pennsylvania"	  1688
     Jeffrey W. Hartung

   "The Operator's Viewpoint of Wastewater Treatment
   Using Rotating Biological Contactors"	  1695
     Mary A. Bergs

   "Troubleshooting an Existing RBC Facility". „	  1710
     B. W. Newbry, M. N. Macaulay, J. L. Musterman,  and
     W. E. Davison, Jr.

   "Structural Engineering of Plastics Media For Waste-
   water Treatment by Fixed Film Reactors"	  1731
     Jean W. Mabbott
                        xvi

-------
   "Criteria For Fatigue Design as Applicable to
   Rotating Biological Contactors".	   1756
     Sib S, Banerjee

   "The Air Force Experience in Fixed-Film Biological
   Processes"	   1777
     Ching—San Huang

WORKSHOP ON RESEARCH NEEDS FOR FIXED-FILM BIOLOGICAL
WASTEWATER TREATMENT	   1806

LIST OF ATTENDEES	   1845
                      xvii

-------

-------
                  PART I:   GENERAL SESSION
                       KEYNOTE ADDRESS
               STATE  OF  KNOWLEDGE  FOR  ROTATING
               BIOLOGICAL  CONTACTOR TECHNOLOGY
     E. D. Smith.  Environmental Engineer and Leader of the
     Environmental Water Quality Management Team, U.S. Army
     Construction Engineering Research Laboratory, Cham-
     paign, IL  61820
     J. T. Bandy.  Environmental Engineer, U.S. Army Con-
     struction Engineering Research Laboratory, Champaign,
     IL  61820
INTRODUCTION

     It is a real pleasure for me to be here this morning to
discuss the state-of-knowledge on Rotating Biological Con-
tactors (RJBC's).  When I made the Keynote address at the
1980 First National Symposium/Workshop on RBC's, held at
Champion,  PA in 1980, I had hoped that this type of confer-
ence might become a tradition.  I believe that the 1980
conference was beneficial to the RBC industry.  I expect
this conference to be equally useful.  Recently it has
becomfe more difficult to obtain funding for this type of
symposium.  I believe that the benefits of these meetings
far outweigh their costs.  I hope that the success of our
conference will encourage the sponsoring of similar gather-
ings in the future.
     I am happy to see the excellent turnout for this sympo-
sium.  Most of the experience and competency in the field of

-------
RBC technology are represented here today.  The agenda indi-
cates you will be quite busy in the next three days.  I am
confident that it will be a productive and pleasant experi-
ence for you.  I am confident that the proceedings, which
will be published from this meeting (and for which I am
responsible), will provide very excellent technical guidance
to those who could not attend.
     Today,  I plan to provide a state-of-knowledge defini-
tion of RBC  technology.  I plan to do this by discussing how
the RBC scenario has changed from 1980 - the year that a
state—of—knowledge definition was given at the 1st National
Symposium/Workshop on RBC Technology.
PROGRESS AND PROBLEMS

     It was reported at the 1980 conference that, in compar-
ison with many other wastewater treatment technologies
(e.g., activated sludge), few dollars and man-years of
research had been devoted to RBC technology.  The many
excellent papers presented at the 1980 symposium signifi-
cantly narrowed that disparity.  Numerous additional
research studies and field evaluations have been described
in the literature since that time.  Many new RBC installa-
tions have come on line since the First National Symposium.
However, despite RBC technology's continued spread and
despite the incorporation of field experience and research
findings into RBC process and equipment design recommenda-
tions; one unfortunate characteristic of the technology has
remained unchanged.  Those who design and operate RBC facil-
ities must largely rely upon the design recommendations and
operation guidance of the vendors of RBC equipment.
     The present state—of-knowledge is such that there is no
single best design procedure or set of relationships that
are universally applicable.  No well-defined theory of RBC
design and operation is accepted by all RBC manufacturers.
Activated sludge, trickling filter and most other wastewater
treatment processes may be designed and constructed without
significant dependence upon equipment proprietors.  This is
not the case with RBC technology.  Design engineers who have
selected RBC technology are extremely dependent upon
proprietors' design curves.  The situation is compounded by
the fact that each manufacturer has a different approach to
media fabrication, configuration, and shaft attachment and
shaft design, and there exist many conflicting stories and

-------
opinions as to the suitability of the alternative equipment
for even conventional "applications.
     Although millions of dollars have been spent by Ameri-
can industries and municipalities for RBC process equipment,
the latest wastewater treatment guidance documents still
reveal a conspicuous lack of information regarding the RBC
unit process.  For instance, .many excellent .documents which
provide design and operation and maintenance
criteria/guidelines•are readily available for traditional
technologies such as the activated sludge and trickling
filter process.  'An example of such a publication is the
excellent EPA report - Process Control Manual for Aerobic
Wastewater Treatment Facilities(1).  The purpose of the pub-
lication is.to provide guidance to optimize the performance
and to help establish process control techniques for trick-
ling filter and activated sludge systems.  There is no com-
parable manual for RBC technology.  Other examples which
demonstrate the novel nature of RBC technology in the United
States are two excellent EPA documents - (1) Upgrading
Trickling Fi 11 e r s (2) and(2) Proce ss Des_ign Manua 1 for
Upgrading Existing Wastewater TreatmentPlants(3).They do
not mention RBC technology.  In addition, commonly used
"state-of-the-knowledge" documents which are designed as
guidance for the selection of wastewater treatment systems
based upon economic consideration .either do not have RBC
cost curves (capital, O&M, energy, etc.) or the curves are
dated.  Guidance remains scarce with regard to RBC applica-
bility, design, O&M and economic considerations.
     To make matters worse, the RBC industry has suffered a
public relations problem because of numerous equipment
failures.  Premature shaft failures, stub end failures and
media separation/degradation have been experienced at exist-
ing installations.  The durability of the polyethylene is
still uncertain because of the relatively short service
record (most facilities with RBC systems were built during
the past five years).  Industry is attempting to rectify
these problems.
     To be fair to the RBC equipment vendors it must be
noted that the RBC process has unique characteristics which
almost guarantee that problems would occur during the early
development of the technology.  It would be very difficult
to destroy or damage wastewater treatment technologies such
as activated sludge or trickling-fliters through improper
design or operations.  Improper design or operation of RBC
units potentially could result in structural failure

-------
problems.  Even with proprietary assurances that current
designs are much improved, the life expectancy of major com-
ponents is not fully known.  Consequently, choice of the RBC
alternative should be accompanied by a negotiated
performance/equipment warranty.  This consideration is
important when a pollution abatement engineer wants to be
confident of the reliability of any wastewater treatment
technology.  However, one should keep in mind that if the
manufacturers' assurances are accurate, current designs are
much improved.  Then RBC technology should be the technology
of choice wherever it is applicable.  It is significant that
hundreds of RBC plants have been in operation for several
years without experiencing media/shaft failure problems.
     All manufacturers offer a warranty against defects in
materials and workmanship after delivery or after plant
start-up.  The warranty period and conditions vary depending
on system components and the manufacturer, and are often
negotiable.  For example, the Plainville Plant in Connecti-
cut was given a warranty period of 30 years for the shafts,
10 years for the surface media, and 5 years for mechanical
equipment.
     Many RBC manufacturers offer performance guarantees
that generally provide a specified effluent with the equip-
ment installed and operating at design conditions.  The
guarantee usually obligates the manufacturer to provide new
equipment or a partial refund if the design effluent stan-
dards are not met.  This guarantee is predicated on the fact
that influent characteristics are within the specified
limit.  Generally, the manufacturers are willing to nego-
tiate a guarantee as long as they agree with the treatment
design.  During the maturational period of the RBC process,
these guarantees and warranties will be especially important
to the RBC user community.
STATUS OF RBC TECHNOLOGY

     Even with these problems, the extent and magnitude of
interest regarding RBC technology continues to increase.
The participation at the conference session dedicated to
RBC's is evidence of the interest of various sectors
(private, academic, research, government agency,  regulatory,
A/E, professional organization, design engineer,  industrial,
and plant operators).  All of the above and other profes—

-------
sionals involved in wastewater treatment and mangement are
represented at this symposium.
     Two years ago it was reported that RBC technology was
popular in Europe for both municipal and industrial applica-
tions and that it was being utilized ever more frequently in
the U.S.  Today, it can safely be reported that RBC technol-
ogy has truly made the transition into a truly cosmopolitan
treatment technology.  There are more than 30 RBC manufac-
turers in Japan alone.
     In the U.S., RBC's have been in operation treating mun-
icipal wastewater for more than 10 years.  Over 250 instal-
lations are presently in operation with design flow rates
ranging from less than 0.01 mgd to 54 mgd.  Approximately 25
percent of existing RBC municipal facilities in the U.S. are
package plants.  RBC's are currently being evaluated for
potential application to a 200 mgd plant which would have
several hundred shafts.  Approximately 70 percent of the RBC
systems operating in the U.S.A and Canada are designed for
organic carbon removal only, 25 percent for combined organic
removal and notification and 5 percent for nitrification of
secondary effluent.
     Several significant developments in RBC technology are
occurring.  Some of these directly address the problems to
which I alluded earlier.  All are potentially important.
     a.  The U.S.E.P.A. has chosen RBC's as the topic of
their first publication of a Design Information Series (DIS)
document, the purpose of which is to provide selected design
information.  The document is currently under review.  The
DIS is not a manual specifying design criteria.  It supple-
ments commonly accepted RBC design procedures or approaches
by providing appropriate qualifiers and/or information not
readily available to the design community.  The document
seeks to address important design parameters and relation-
ships (or lack of them) in order to provide a more rational
RBC design approach.  Topics considered are design loadings
for carbonaceous removal, nitrification and denitrifiction,
equipment reliability and service life, power requirements
for air and mechanically driven units and structural design
considerations such as flexibiity and hydraulics.  The docu-
ment attempts to provide practical usable design information
rather than to emphasize theoretical considerations.  The
information in the document is intended to assist the design
engineer by providing a more in-depth perspective on some of
the key design considerations than is normally available in
other design manuals.

-------
     b.  The American Society of Civil Engineers  (ASCE) has
formed a "Rotating Biological Contactor Task  Subcommittee."
The subcommittee has prepared a report entitled "RBC for
Secondary Treatment" which should, be published in  1982.
     c.  The US Army Construction Engineering Research
Laboratory (CERL) has prepared a report which provides
assistance in determining when trickling filter plants can
be effectively and economically upgraded using RBCs and
which provides guidance in designing the RBCs.
     d.  USACERL will publish a lessons learned document
based upon Dept. of Army and Corps of Engineers RBC applica-
tions at Fort Bragg, Fort Ritchie, Fort Knox, Jwalein
Island, Korea, and Saudi Arabia.
     e.  The Corps of Engineers has sponsored a study dedi-
cated to evaluating the potential of RBC's  in recreational
area applications.
     f.  Finally, this conference session devoted  to RBC's
is taking place.  All the various sectors of  the RBC commun-
ity are represented in these meetings (academic/researchers,
A/Es, manufacturers, government representatives, municipal
and industrial engineers and operating personnel).  My work.
in compiling these proceedings has convinced me that our
meeting has already been productive.  Much  more will come of
our personal interactions this week.
     g.  Rotating Biological Contactor related research
reported in the technical literature is much more  common
during the last few years.
LITERATURE REVIEW

     A literature search for 1980-1981 was performed which
identified 126 studies.  The following review provides
information concerning various aspects of theory, design and
operating experience associated with RBC systems.
     The First National Symposium/Workshop on Rotating Bio-
logical Contractors held at Champion, PA, on February 4—6,
1980 more than doubled the literature of the technology(4).
The proceedings are a compilation of the 68 papers delivered
during the meeting and a transcription of the associated
workshop.  Eleven major topics are covered in the papers:
perspective, overview, history, process variables and
biofilm properties, municipal wastewater treatment, biok-
inetic studies, air drive and supplemental aeration, indus-
trial wastewater treatment, concepts and models, upgrading

-------
waste treatment systems with RBC's, design and operation,
nitrification and dentrification, and selection and econom-
ics.  Requirements for further research were identified at
the workshop.
     In September of 1980, Kneel and Godfrey announced in
Civil Engineering a cooperative effort 'of the U.S. EPA and
the ASCE to produce a new series of design books which would
meet the twin goals of reducing the time required f.or new
knowledge to be reflected in design manuals and of securing
profession—wide peer review of design manuals as they are
produced(5).  One of the first manuals to be produced will '.
cover rotating biological contactors.
     Numerous papers have appeared since the First National
Symposium in early 1980.  Hitdlebaugh and Miller(6) dis-
cussed the .operational problems of RBC's.  Dehkordi(7) and
Keihan:(8)  described the effects of heavy metals upon RBC
performance.  Trinh(9), Allen(lO) and Bauer et al.(ll)
assessed the applicability of RBC's for remote or on—site
applications.  Mueller et al.(12) discussed the impact of
mass transfer considerations upon RBC and trickling filter
design.  Factors to be considered in scaling up were-identi-
fied by Wilson et al.(13).  Kinetics for domestic wastewater
treatment were explored by Pano(14).
     Reports of RBC applications to the secondary treatment
of domestic wastewater continued to appear.  Regent(15)
reported several years of successful RBC operation in Yugos-
lavia.  Interestingly, he described no mechanical failures.
Spink(16) described the role of RBC's in the Province of
Alberta, Canada.  Rushbrook and Wilke(17) described an inno-^
vative treatment facility in Hillsborough, NH which will
include RBC's, solar-heated anaerobic digestion and methane
recovery.  Shifts in sewage solids distribution across RBC
installations were studied by Nunch et al.(18).  Sapin-
sky(19) emphasized the importance of energy conservation in
wastewater treatment and cited RBC plants at Hillsborough
NH, Minneapolis and Chicago for their efficient use of
energy.
     Some interesting process modifications were explored as
were some unusual applications of RBC technology at conven-
tional wastewater treatment plants.  Given(20) reported on
the RBC treatment of dilute wastewater.  Huang and Bates(21)
compared RBC treatment of a synthetic milk waste using air
and pure oxygen.  RBC's were used in an innovative anaerobic
treatment system for high strength carbonaceous wastes by
Tait and Friedman(22).  Cheung and Krauth(23) investigated

-------
the feasibility of replacing conventional sedimentation by
microstrainers in the RBC system.  Hong(24) evaluated the
use of RBC's in treating aluminum sulfate coagulated septage
supernatant.
     The use of RBCs for tertiary treateraent continued to
develop.  Noss and Miller(25) described the use of an RBC
for secondary treatment and recarbonation following low-
level lime addition for phosphorus removal.  The effects fo
nitrate concentration and retention period upon RBC denit—
rification were investigated by Cheung and Krauth(26).
Stephenson and Murphy(27) characterized the kinetics of den-
itrification in a biological fluidized bed.  Buckingham(28)
performed an engineering and marketing analysis of the
rotating disk evaporator, a device physically similar to RBC
which is designed to evaporate wastewater rather than bio-
logically treat it.
     Numerous nitrification studies have appeared since
early 1980.  Wu et al.(29) used data from many previous stu-
dies to derive and validate a model for the prediction of
RBC nitrification performance.  Mueller et al.(30) developed
and verified a steady state model of nitrification and
organic carbon oxidation in the RBC.  Smith et al.(31)
evaluated RBC's as an upgrading-retrofit process for BOD
reduction and nitrification.  Bridle(32) discussed RBC's in
the context of biological processes for nitrogen conversion
along with other processes capable of achieving the same
ends.  The kinetics of the nitrification process were
modeled by Watanabe et al.(33) and by Margaritas et al.(34).
Stratta(35) investigated the feasibility of enhancing nit-
rification by controlling the pH in RBC's,  Marsh et al.
described a coupled trickling filter - RBC nitrification
process(36).
     Additional nutrient removal work included the investi-
gation by Knoetze et al.(37) into chemical inhibition of
biological removal processes.  Murphy and Wilson(38) per-
formed pilot plant studies of BOD removal, nitrification and
phosphorus removal.  Singhal(39) described RBC nitrification
at an advanced wastewater treatment plant in Cadillac,
Michigan.  An energy efficient extension to the Guelph,
Ontario wastewater treatment plant was described(40).  This
plant effectively removes BOD, ammonium-nitrogen and phos-
phorus with RBC's followed by filtration.
     The feasibility of using RBC's to upgrade existing
plants was explored in several studies.  Gutierrez et
al.(41) evaluated upgrading primary tanks with RBC's.  Smith

-------
et al.(42) considered RBC's as an upgrade for existing
trickling filter plants.  Poon et al.(43,44) evaluated the
effectiveness of RBC's in supplementing the BOD and ammonium
nitrogen removals achieved at trickling filter plants.  The
Surfact process' developed by the Philadelphia Water Depart-
ment was described by Guarino et al.(45).  The Surfact pro-
cess which physically merges an RBC with a diffused aeration
tank provides an inexpensive upgrade.  Very little construc-
tion is required.
     A final area of activity has involved industrial or
primarily industrial wastewaters and rotating biological
contactors.  Chesler and Eskelund(46) evaluated RBC's for
the treatment of explosives manufacturing wastes.  Acid mine
wastes were treated in pilot scale and prototype studies
conducted by Olem and Unz(47) at Hollywood, PA.  Dairy
wastes were treated in an innovative process involving an
aerated equalization tank and RBC's by Waggener et al.(48).
Suria Pandian and Agarwal(49) also described RBC treatment
of dairy wastes.  The city of Monett, Missouri, overcame
problems posed by industrial discharges to its sewage plant
equivalent to 7 times its population by using RBC's(SO).
O'Shaughnessy et al.(51) applied RBC's to oil shale retort
wastewater.  Blanc et al.(52) evaluated RBC's for the treat-
ment of beef slaughtering and processing wastewaters.  The
influence of the rotational speed of RBC's on the reaction
rates observed was investigated by Odai et al.(53).
Borghei(54) described treatment of the effluent of a
glucose-production plant using a rotating biological packed
bed,
     The emphasis of a report by Chesner and Bender (55) was
to review and compare current design procedures and perfor-
mance capabilities of the RBC process.  This was accom-
plished by a review of the literature, an evaluation of the
process, O&M, equipment and power performance at RBC plants
approaching design flow conditions and a comparison of
current design guide information.

          Sixteen domestic RBC facilities providing
     carbonaceous BOD removal and approaching design
     flow conditions, supplied monitoring data that
     were used to evaluate process performance.  The
     reliability of these systems in meeting effluent
     concentration and removal efficiency criteria,
     defined by NPDES as 30 mg/L BOD effluent concen-
     tration and 85 percent BOD removal efficiency,

-------
respectively, was evaluated.  The results indi-
cated that the plants exceeded effluent criteria
12 percent of the time and failed to meet percent
BOD removal 67 percent of the time.  An analysis
of performance data demonstrated that average
values for both mass BOD removal rates (Ibs BOD
removed/day/I,000 sf of media) and BOD removal
efficiencies increased with increasing influent
waste strength.  For the range of conditions found
at the plants surveyed, RBC process performance
followed design predictions for mass BOD removal
rates and percent BOD removals for high wastewater
influent strength (175 to 350 mg/L BOD), and pro-
gressively lagged behind those predictions as
waste strength decreased below 175 mg/L.
     Low labor requirements to operate and main-
tain an RBC secondary treatment unit are attrac-
tive features of an RBC system.  Hourly labor
requirements were reported in the range of 1 to 7
hours per week, averaging 2.6 hours per week for
23 plants, with an average design flow of 1.4 mgd.
Power measurements were performed during the
course of this investigatin to identify RBC energy
consumption.  The results established power con-
sumption to rotate 100,000 sf of standard density
media to be 3.46 kw for mechanical drive (1.6 rpm)
and 2.93 kw for air drive (1.2 rpm).  To rotate
150,000 sf/shaft of high density media at 1.6 rpm
mechanical units used 3.77 kw of power.
     Equipment performance is a severe problem in
existing RBC systems.  The nature of the problem
centers on shaft failures and media degradation.
Of the plants surveyed there were 12 shaft
failures reported and the media in three plants
had become brittle or failed due to shifting.  As
a result of this poor operating history it was
concluded that design engineers should seek an -RBC
eqiupraent warranty sufficient to protect the owner
against equipment failures (55).
                         10

-------
                         REFERENCES
1.  U.S. EPA, "Process Control Manual for Aerobic Wastewater
   'Treatment Facilities," EPA-4Q3/9-77-006,  (1977).

2.  U.S. EPA, Office of Water Program Operations  (WH-547),
    Washington, DC, 20460, EPA-430/9-78-004,  MCD-42,  (1978).

3.  U.S. EPA Technology Transfer Publication,  "Process
    Design Manual for Upgrading Existing Wastewater Treat-
    ment Plants," (1974).

4.  Wu, Y. C.; Smith, E. D.; Opatken, E. J.;  Miller,  R.  D.;
    Borchardt, J. A.; Proceedings;  First National
    Symposium/Workshop on Rotating Biological  Contactor
    Technology.  Champion, PA; FEbruary 4-6,  1980.  Spon-
    sored by University of Pittsburgh, Municipal Environmen-
    tal Research Laboratory, Cincinnati, Ohio  and U.S. Army
    Construction Engineering Research Laboratory, Champaign,
    Illinois.  June 1980.  Volumes I and II.

5.  Godfrey, Kneeland A. Jr., "New Developments in Wastewa-
    ter Treatment — EPA-ASCE Design Books  Planned",  Civil
    Engineering (NY) 50(9) 96-101 (1980).

6.  Hitdlebaugh,  J. A. and R. D. Miller, "Operational Prob-
    lems with Rotating Biological Contactors", Journal Water
    Pollution Control Federation 53(8) 1283-1293 (1981).

7.  Dehkordi, F.  G., "The Effect of Heavy Metal on the Per-
    formance of Rotating Biological Contactors (RBC)",
    Master's thesis, Oklahoma State University, OWRT-A-087-
    OKLA(3), (1980).

8.  Keihani, A.,  "Long Term Effects of Chromium and Copper
    on the Rotating Biological Contactor",  Master's thesis,
    Oklahoma State University, OWRT-A-087-OKLA(2), (1980).

9.  Trinh, D. T., "Exploration Camp Wastewater Characteriza-
    tion and Treatment Plant Assessment," Technol Dev. Rep.
    EPS No. 4 (1981).
                               11

-------
10.  Allen, G. A., "Methods  Used  to  Provide  Services to Out-
     lying Mine Townsites",  Water and  Pollution Control
     118(2) 18-19, (1980).

11.  Bauer, D. H.; Conrad, E.  T.  and D.  G.  Sherman,  "Evalua-
     tion of On-site Wastewater Treatment and  Disposal
     Options", SCS ENGINEERS,  EPA-600/2-81-178,  (1981).

12.  Mueller, J.  A., Paquin,  P. and  J.  Famularo,  "Mass
     Transfer Impact on RBC  and Trickling Filter  Design",
     AIChE Symposium Series  No. 197, Vol. 76,  (1980).

13.  Wilson, R. W.; Murphy,  K. L., and  J. P.  Stephenson,
     "Scale Up In Rotating Biological  Contactor Design",
     Journal Water Pollution Control Federation 52(3)  610-
     621, (1980).

14.  Pano, Abraham, "The  Kinetics of Rotating  Biological
     Contactors Treating  Domestic Wastewater",  doctoral
     dissertation, Utah State  University, (1981).

15.  Regent, Aleksander,  "Small RBC's  Logging  Hours  in
     Yugoslavia", Water and  Sewage Works 127(8) 42,  (1980).

16.  Spink, D., "Getting  the Treatment", Env.  Views  3(1)7,
     (1980).

17.  Rushbrook, E. L. and D.  A. Wilke,  "Energy Conservation  '
     and Alternative Energy  Sources  in  Wastewater  Treat-
     ment", Journal Water Pollution  Control  Federation
     52(10) 2477-2483, (1980).

18.  Munch, R.; Hwang, C. P. and  T.  H.  Lackie,  "Wastewater
     Fractions Add to Total  Treatment  Picture",  Water  and
     Sewage Works 127(12) 49-54,  (1980).

19.  Sapinsky, C. P., "Energy  Conservation is  a Dire Neces-
     sity", Water and Wastes Engineering 17(8)  28-32,
     (1980).

20.  Given, P. W., "RBC Treatment of Dilute Wastewater",
     Annual Conf. of the Water Pollution Control  Federation,
     53rd, Proceedings of the  Research  Symposium,  Las  Vegas,
     Nevada, WPCF, (1980).
                                12

-------
21.  Huang, J. C., and V. T. Bates, "Comparative  Performance
     of Rotating Biological Contactors Using Mr  and  Pure
     Oxygen",  Journal Water Pollution Control Federation
     52(11) 2686-2703, (1980).

22.  Tait, S.  J. and A. A. Friedman, "Anaerobic Rotating
     Biological Contactor for Carbonaceous Wastewaters",
     Journal Water Pollution Control Federation 52(8)  2257-
     2269, (1980).

23.  Cheung, P. S. and K. Krauth, "Investigation  to Replace
     the Conventional Sedimentation Tauk by a Microstrainer
     in the Rotating Disk System," Water Res 14(1) 67-75,
     (1980).

24.  Hoag, G.  E., "Rotating Biological Contactor  Treatment
     of Aluminum Sulfate Coagulated Septage Supernatant",
     Master's thesis, University of Lowell, (1980).

25.  Moss, C.  I. and R. D. Miller, "Rotating Biological Con-
     tactor Process for Secondary Treatment and Recarbona-
     tion Following Low-Level Lime Addition for Phosphorus
     Removal," Final Report, Army Medical Bioengineering
     Research and Development Laboratory, Fort Deitrick, MD,
     USAMBRDL-TR-8007, (1980).

26.  Cheung, P. S. and K. Krauth, "Effects of Nitrate  Con-
     centration and Roetention Period on Biological Denit-
     rification in the Rotating-Disc System," Water Polution
     Control (London) 79(1) 99-105, (1980).

27.  Stephenson, J. P. and K. L. Murphy, "Kinetics of  Bio-
     logical Fluidized Bed Wastewater Denitrification,"
     Tenth Int. Conference International Association  for
     Water Pollution Research, Toronto, Ontario,  June  23-27,
     1980 Prog. Water Tech 12(6).

28.  Buckingham, P. L., "Production Engineering and Market-
     ing Analysis of the Rotating Disk Evaporator", Munici-
     pal Environmental Research Laboratory, Cincinnati,
     Ohio, EPA-600/2-81-179, (1981).

29.  Wu, Y. C.; Smith, E. D. and John Gratz, "Prediction of
     RBC Performance for Nitrification", Journal  Environmen-
     tal Engineering Division, ASCE 107(4) 635-652, (1981).
                                 13

-------
30.  Mueller, J. A., Paquin, P., and  J.  Famularo,  "Nitrifi-
     cation in Rotating Biological  Contactor",  Journal  Water
     Pollution Control Federation 52(4)  688-710,  (1980).

31.  Smith, E. D.; Poon, C. P. C.;  Mikucki, W.  and J.  T.
     Bandy, "Tertiary Treatment of  Wastewater  Using A  Rotat-
     ing Biological Contactor System", U.S. Army  Construc-
     tion Engineering Research Laboratory,  Champaign,  IL,
     CERL Tech. Report N-85, (1980).

32.  Bridle, T. R., "Fundamentals of  Biological Processes
     for Nitrogen Conversion," Presented at Env.  Canada et
     al Nutrient Control Technology Conference, Calgary,
     February 7-8, 1980, P7-A(42).

33.  Watanabe, Y., Ishiguro, M., and  K.  Nishidorae,  "Nitrifi-
     cation Kinetics in a Rotating  Biological  Disk Reactor,"
     Tenth Int. Conference  International Association for
     Water Pollution Research, Toronto,  Ontario,  June  23-27,
     1980, Prog. Water Tech. 12(6).

34.  Margaritas, A., Watanabe, Y.,  Ishiguro, M.,  Nishidome,
     K. and P. Harremoes, "Nitrification Kinetics  in a
     Rotating Biological Disk Reactor",  Water  Science and
     Technology 13(4-5) 1219-1225,  (1981).

35.  Stratta, J. M., "Nitrification Enhancement Through PH
     Control With Rotating  Biological  Contactors",  doctoral
     dissertation, Pennsylvania State  University,  (1981).

36.  Marsh, D., Benefield,  L., Bennett,  E., Lindstedt,  D.
     and R. Hartman, "Coupled Trickling  Filter-Rotating Bio-
     logical Contactor Nitrification  Process,"  Journal  Water
     Pollution Control Federation 53(10) 1469-1480,  (1981).

37.  Knoetze, C., Davies, T. R., and  S.  G.  Weichers, "Chemi-
     cal Inhibition of Biological Nutrient  Removal
     Processes," Water SA 6(4) 171, (1980).

38.  Murphy, K. L., and R.  W. Wilson,  "Pilot Plant  Studies
     of Rotating Biological Contactors Treating Municipal
     Wastewater", Env. Canada Env.  Protection  Service Report
     SCAT-2, (1980).
                               14

-------
39.  Singhal, A. K., "Phosphorus and  Nitrogen  Removal  at
     Cadillac, Michigan," Journal Water  Pollution  Control
     Federation 52(11) 2761, (1980).

40.  Anonymous, "Innovative Nutrient  Removal Process
     Attracts International Attention,"  Water  and  Pollution
     Control 118(10) 14-15, (1980).

41.  Gutierrez, A.,-Bogart, I. L«,  Scheibe,.D.  K.,  and T.  J.
     Mulligan, "Upgrading Primary Tanks  With Rotating  Bio-
     logical Contactors", Municipal Environmental  Research
     Laboratory, Cincinnati, Ohio,  EPA-600/2-80-003,  (1980).

42.  Smith, E. D., Poon, C. P., and R. D.  Miller,  "Upgrading
     DA Trickling-Filter Sewage Treatment  Plants,"  U.S. Army
     Construction Engineering Research Laboratory,  Cham-
     paign, IL, CERL Tech. Report N-102,  (1981).

43.  Poon, C. P. C., Chin, H. K., Smith,  E. D., and W. J.
     Mikucki, "Upgrading With Rotating Biological  Contactors
     for BOD Removals," Journal Water Pollution Control
     Federation 53(4) 474-481, (1981).

44.  Poon, C. P. C., Chin, H. K., Smith,  E. D., and W. J.
     Mikucki, "Upgrading With Rotating Biological  Contactors
     For Ammonia Nitrogen Removal," Journal Water  Pollution
     Control Federation 53(7) 1158, (1981).

45.  Guarino, C. F,, Nelson, M. D., and  T. E.  Wilson,
     "Uprating Activated-Sludge Plants Using Rotary Biologi-
     cal Contactors," Water Pollution Control  79(2) 255,
     (1980).

46.  Chesler, P. G. and G. R. Eskelund,  "Rotating  Biological
     Contactors for Munitions Wastewater Treatment,"  Final
     Technical Report, Army Mobility  Equipment Research and
     Development Command, Fort Belvoir,  VA, MERADCOM-2319,
     (1981).

47.  Olem, H. and R. F. Unz, "Rotating Disc Biological
     Treatment of Acid Mine Drainage," Final Report,  Indus-
     trial Environmental Research Laboratory,  Cincinnati,
     Ohio, EPA-600/7-80-006, (1980).
                                 15

-------
48.  Waggener, J. E., Fitzhugh, M. L«, and  G.  E.  Flann,
     "Innovative Approach to the Treatment  of  Dairy  Wastewa-
     ter with Rotating Biological Contactors,"  53rd  Annual
     Conference of the Water Pollution Control  Federation,
     Proceedings of the Industrial Wastes Symposium,  Las
     Vegas, Nevada, Sep 28-Oct 3, 1980, Published by WPCF,
     (1980).

49.  Suria Pandian, P. S. and I. C. Agarwal, "Removal of
     Dairy Waste Organics by Rotating Biological  Contac-
     tors," Indian Journal of Environmental Health 23(1)  27,
     (1981).

50.  Riddle, W. G., "Small City Requires Large  WWTP," Water
     and Wastes Engineering 17(7) 42-44, (1980).

51.  O'Shaughnessy, J., Blanc, F. C., Wei,  I. W.,  and J.
     Patinskas, "Biological Treatment of Oil-Shale Retort
     Wastewater Using Rotating Biological Contactors,"
     Abstracts of Papers of the American Chemical Society,
     Vol. 181, p. 76, (1981).

52.  Blanc, F. C., O'Shaughnessy, J. and S. H.  Corr,  "Treat-
     ment of Beef Slaughtering and Processing  Wastewaters
     Using Rotating Biological Contactors," Abstracts of
     Papers of the American Chemical Society, Vol. 181, p.
     74, (1981).

53.  Odai, S., Fujie, K., and H. Kubota, "Effect  of  Rotation
     Speed on Reaction—Rate on a Rotating Biological Contac-
     tor," Journal of Fermentation Technology  59(3)  227-234,
     (1981).

54.  Borghei, S. M., "Treatment of the Effluent of a
     Glucose-Production Plant Using a Rotating  Biological
     Packed Bed," Process Biochemistry 16(2) 29,  (1981).

55.  Chesner, W. H. lannone, J. and J. Bender,  "Review of
     Current RBC Performance and Design Procedures," Report
     prepared for th Municipal Environmental Research
     Laboratory Office of Research and Development USEPA  22
     W. St. Clair Street, Cincinnati, OH  45268 (contract
     No. 68-02-2775), March 1981.
                               16

-------
             ANAEROBIC ATTACHED FILM EXPANDED BED
                         FU1DAMENTALS
             William J.  Jewell,  Department  of
             Agricultural Engineering, Cornell
             University, Ithaca, Hew York
ABSTRACT

     The anaerobic attached film expanded bed process (AAFEB)
has been shown to be capable of treating low strength waste-
waters  at  low temperatures and at relatively short retention
times.  Such capability leads to the unexpected conclusion
that the AAFEB is a municipal wastewater treatment alternative
capable of meeting secondary effluent quality without producing
a secondary waste sludge.  In order to understand reasons for
this extraordinary capability, recent research has focussed on
reproducing the phenomena, evaluating the kinetics with soluble
and insoluble substrate at varying temperatures, testing of
the system under shock loads, and evaluation of the potential
applications (algae harvested or operation under thermophilic
temperatures, 50°C).  This paper will relate the research data
to process fundamentals (active biomass, solids retention time,
substrate kinetics) and design requirements.

INTRODUCTION TO THE ANAEROBIC          BED

     An anaerobic biological treatment process capable of
treating dilute domestic sewage to secondary quality without
the production of waste secondary sludges, and processing
                             17

-------
capability superior to aerobic biological systems are among
the characteristics suggested by the results of studies on
the anaerobic attached microbial film expanded bed reactor
(AAFEB) (1, 2).  These surprising results have been obtained
from nearly a decade of research and development efforts on
a new process that has attempted to optimize the capability
of the anaerobic fermentation process for wastewater treatment
(3, U, 5, 6).
     The anaerobic methane fermentation process has been
applied for many decades to waste management, as has been the
expanded bed process.  The definition of each unit process is
well-known, but the combination of these two unit operations
into one process has only recently been achieved.  The applica-
tion of the expanded bed physical process as a biological con-
verter appears to be an optimum method of achieving fine solids
separation and microbial conversions,

Goals and Objectives

     The main goal of this paper is to synthesize the basic
fundamentals of anaerobic biological processes and the char-
acteristics of the physical expanded bed to illustrate the
basis for the AAFEB process.  The specific objectives are to
review the physical considerations required in operating the
expanded bed, to summarize the biological capability, to
compare the resulting process to other attached film processes,
and finally to briefly consider future research and develop-
ment efforts required to clarify the process capabilities.

Background

     The conceptual diagram of the expanded bed process is
shown in Figure 1.  It is a fine particle upflow filter in
which the particles are slightly expanded but remain in close
proximity to other particles.  The basic mode of operation
is similar to packed filters and fluidized processes.  These
similarities have led to some confusion between the processes
and the terminology used to describe each.
     A comparison of these various processes that use inert
packing material is shown in Figure 2.  The static filters,
or packed filters, were developed in 1963 by Young and McCarty
(T).  By decreasing particle size and increasing upflow veloc-
ities , there will be a point at which the particles begin to
be lifted in a slightly expanded form.  The relationship
between the porosity of the bed (the ratio of the void space
to the total volume of the bed) and the head loss that occurs
                               18

-------
INFLUENT
                                     BIOGAS
                                         EFFLUENT
                                     RECYCLE
Figxire 1.  Schematic dxagraia of the attached laicrobial fiJLm expanded
         bed Drocess.
                           19

-------
      FLUIDIZEO BED
  t
STATIC
                        FLOWING
                                                        EXPANDED  BED
                                                  STATIC
t
"•^r
    FLOWING I
Figure 2.  Qualitative comparison of reactor volumes occupied by media in typical fluidized and

         expanded beds.

-------
In these units has been well defined for process applications
such as water filter backwashing (8, 9)-
     The general relationships between the upflow rate and the
bed characteristics are summarized in Figure 3.  At low fluid
velocities5 the particles remain in contact or in packed form,
As the velocity of the liquid is increased, the particles'
resistance causes them to be slightly expanded or "fluidized."
The operation of the expanded bed is most effective when the
expansion is limited to a small fraction of the packed bed
volume.  This requirement" enables the bed to inhibit the flow
of fine solids through the filter but to avoid clogging, which
occurs in a packed bed.  Further increases in velocity cause
further separation of the bed.  As the velocity is further
increased, the individual particles separate, and true fluid-
ization begins.  All particles are in motion, and the bed con-
tinues to expand; and particles move in more rapid and more
independent motion.  The bed continues to expand as the veloc-
ity is increased and maintains a uniform character.  Particles
move in random directions through all parts of the liquid at
this state. . Strong transient currents with many particles
temporarily traveling in the same direction can be observed,
but in general, particles move randomly as individuals.  This
phenomenon is known as particulate fluidization.  It is the
common state for fluidized processes in which the bed is
fluidized up to 300 percent or more of its static, packed bed
form.
     Eventually, as 'the upflow of the velocities increases,
the superficial velocity approaches the terminal settling
velocities of particles, and the particles become'.entrained
in the liquid and are carried out of the reactor.  Thus, in
relation to the diagram shown in Figure 3, the expanded bed
operates as close to the fixed bed characteristics as possible,
whereas the fluidized bed often operates at much higher super-
ficial velocities further to the right of the abscissa on the
diagram.
     A comparison of anaerobic and aerobic microbial processes
can be made if the capability of microorganisms.is known under
ideal conditions and these characteristics can be adjusted for
the application to specific processes.  Attached film processes
complicate the comparison because of the increased influence of
mass transfer limitations.  The main parameters would be the
temperature of operation, the concentration of substrate
required to achieve a given removal rate, and microbial yield.
Microbial yield is one of the most important criteria for
comparison since it also is related to the minimum sludge
retention time (SRTmj_n) that can be achieved in a biological
                              21

-------
     FIXED   EXPAND.

       BED     BED
FLUID1ZED

  BED
TRANSPORT
CO
o
tr
o
a.


(9
O
a.
o
a:
o

UJ
a:
iy
tr
a.
               LOG  SUPERFICIAL VELOCITY
    Figure 3.  Effect of increasing upflow velocities with a filter

              of small particles on the friction losses' and the

              void space or  filter porosity.
                             22

-------
process.  The microbial solids retention time ultimately
defines both the stability of the process as well as the
safety factor under which the process is designed or operating.
A summary of example values for anaerobic and aerobic treat-
ment systems  is  given in Table 1.  These data emphasize
several well known process differences.  The high reproduction
rates of aerobic organisms lead  one to conclude that they
have significantly more capability for substrate removal and
require a much smaller reactor volume than anaerobic processes.
Minimum solids retention times in treating soluble substrates
significantly greater than 10 days at an operating temperature
of 25°C would not be unusual for a conventional anaerobic
treatment process.  The relationship of 'the solids retention
time to the microbial mass in the system is given by the
following equation:
                             '  X0 •  V
                         SRT = -~ - -
     where  .    SRT =-the solids retention time
                  V = reactor volume
                 XQ = bacterial concentration in the reactor
                 Xe = biomass lost from the reactor in the
                      effluent or intentionally wasted each
                      day

     Whenever a process operates at a sludge retention time
less than the microorganism reproduction time, it is, in the
process of failing and/or going through a change to a situa-
tion where the process efficiency is changing as the micro-
bial mass adjusts its concentration.  A simple method of
reviewing the process capability under a given set of operat-
ing conditions is to determine the solids retention time that
can be achieved by various processes and compare it to the
minimum  acceptable  with the system.  This will be done in an
example later.

EXPANDED BED PROCESS DEFINITION

     The development of the expanded bed process is based on
efforts to optimize the conditions required to achieve maximum
microbial concentrations while good control over the microbial
biomass in the fluid media is maintained.  The goal of a
biological process is to minimize the cost of the system that
is  designed  to achieve a specific purpose.  The maximum con-
version  rate  per unit volume of reactor will lead to lower

-------
                       Table I.   Kinetic  Coefficients  for Anaerobic and Aerobic  Treatment  Systems
ro
Biological
Systems

Anaerobic
Systems
1. Acetic
Acid
2. Acetic
Acid
3. Milk
Waste
Aerobic
Systems
1. Domestic
Waste
2. Skim
Milk
Temp.
°C


25
35
25

20
20
Ks Microbial Minimum
Half Rate Yield, Cell Coefficient
Coefficient mg VSS Residence Basis
mg/£ mg Substrate Time, Days

869 0.051* H.2 Acetic Acid
159 0.0l*l* 3.1 Acetic Acid
2k 0.370 	 COD

22 0.670 0.27 COD
100 O.H8 O.H2 BOD..
Reference


15
15
16

17
18

-------
costs, and this should minimize the total treatment system
cost.  It follows that the achievement of maximum removal
rates can be obtained by either using superior microorganisms
or higher microbial mass- concentrations.  Since we have few
opportunities to select microorganisms in waste management,
the emphasis has focused on maximizing microbial concentra-
tions.  However, it is essential that the microbial mass.be
"active"; that is, that it be exposed to an available sub-
strate in the bulk liquid.  Therefore ..--the first part of the
definition of the optimization of a process is that it must
achieve a maximum active "biomass- concentration.
     Of course, achievement of maximum biomass is only the
first step in designing an optimized biological process.  The
second requirement  for the system is to be able to achieve
efficient, reliable management of microorganisms.  A process
that clogs or accumulates thick films through which substrate
cannot penetrate is -not acceptable.  The approach that was
taken with the' development of the expanded bed was to first
define a process that would achieve a maximum biomass per unit
volume and  then superimpose these requirements on-the physical
requirements that are necessary to operate the hydraulic flow
regime.   .                                              .

Design Requirements for Maximum "Active" Microbial Concentra-
tions

     There are a number of factors to be considered in defining
the maximum biomass accumulation potential of the expanded
bed—mass diffusion characteristics of soluble and particulate
organics, microbial growth rates, substrate requirements, and
process kinetics.  The growth kinetics and process require-
ments can be assumed to be similar to those shown in Table I
in the absence of mass diffusion limiting processes.  These
emphasize the .problem of low substrate removal rates whenever
the substrate concentration is low and the requirement for
long solids retention times,                        •  •
     It is well known that methane-producing bacteria are
amongst the slower growing bacteria and at 35°C, under optimum
conditions, have a maximum reproduction time of three to four
days, as shown in "Table I.  Of course, temperatures less than
20°C are often experienced with domestic sewage, and the
microorganism reproduction time may have to be significantly
greater than 10 to 30 days under the cooler winter tempera-
tures ,
     Two questions represent the challenge in the understand-
ing of the process biomass requirements:  (l) what are the
                               25

-------
conditions required to accumulate a maximum active biomass in
the system? and (2) how would we manage those bacteria such
that they would be exposed to the substrate at maximum flow
rates and still maintain control over the bacteria?  Without
the addition of inert particles and the growth of attached
microbial film, it is obvious that there .are a limited number
of ways to try to increase the biomass concentration in the
reactors .  It would appear that reactors without inert media
are limited to somewhere around 5 gm VS/2. unless we go to
highly elaborate methods of maintaining bacteria within the
system.
     The depth of substrate penetration to a microbial floe or
an attached film is well known for the lower substrate concen-
trations that are common in sewage (10, 11, 12).  Substrate
diffusion depths exceeding 60 microns occur at relatively low
substrate concentrations.  In an aerobic film LaMotta (10)
showed that 5.2 mg/£ of glucose penetrated to greater than
10 microns.  These diffusion-limited depths indicate that
attached films thicker than 0.05 mm would result in some sub-
strate-limited biomass, especially where the substrate in
solution is low.
     The substrate diffusion depth limitations for optimum
microbial particle dimensions can be estimated as follows:
     where        = maximum total particle diameter of
                     the inert particle and the attached
                     microbial film

                S  = substrate diffusion depth
                4>  = inert particle diameter

     The background development work at Cornell University has
developed two surprising results in relation to methane-
forming bacterial films ,  The thickness of the film is thin
and usually around 0.020 mm.  This unexpectedly thin film
has a limited impact on particle management and indicates that
all of the attached film will be active since the substrate
diffusion depths, even at low bulk solution concentrations,
will be no greater than 0.05 mm.  Thus, the optimum inert
particle diameter is exceptionally small, being around 0,02 mm.
However, as will be seen, this particle is so small that it is
difficult to 'manage with the practical hydraulic retention
times .
                               26

-------
     The second major observation has been that the bulk
density  of  these thin films is much higher than expected,
being greater than 200 gm/£ of film in some cases.  This
compares to aerobic films that have densitit.es of about
3^ gm VS/&.   A comparison of the particle size and the active
biomass achieved with the particles, assuming that they all
achieve a-thin microbial film and bulk density as indicated
above, is given in Table II.  Maximum biomass concentrations
that have been observed in expanded beds have exceeded kO gm/£.
Data in Table II show that the goal for the expanded bed should
be the achievement of as much as 100 gm/£ of active biomass.

          Table II.  Effect of Inert Particle Size
                     On .Maximum Microbial Mass Goal.
Particle
Description
None (activated
sludge)
Large rocks
Plastic media
Coarse sand
Fine sand
Size
Particle
mm

50 to 75
25
0.2 to 2 :
0.02 to 0.2
Area per
Volume of
Reactor
cm2/cm3

1.0
5-0
21.0 •
210.0
Active
Biomass As
Volatile
Solids
gm/£
2
2
7
16
150
Microbial Management Requirements

     The velocities in the expanded bed and at the top of the
bed should be 'less than or equal to those required for micro-
bial management if we want to achieve maximum solids capture
and separation with the process.  Typical clarifier overflow
rates are approximately 0.7 gal/min/ft2.  This is equivalent
to an upflow velocity of approximately 6 ft/hr.  Flocculated
microbial particles or films that have been scraped off the
inert particles would settle at velocities higher than these
clarifier overflow rates.  Thus, we would expect if the pro-
cess could be designed at these lower velocities, any solids
passing through the bed or escaping from the films would
                             27

-------
collect at the top of the expanded bed.  These solids could be
recycled or removed at this point.

Bed Expansion Requirements

     The liquid velocity required for bed expansion is a func-
tion of its viscosity and density and the particle size, shape,
and specific gravity.  Numerous attempts have been made to
estimate expansion velocity and requirements.  Most of these
efforts have been directed at backwash requirements for various
physical filters.  The author is unaware of any specific work
that has been completed with inert  particles coated with mature
microbial films.  In the case of the expanded bed, it is pos-
sible to use the theory as developed for bed expansion for
backwashing since the microbial films are thin and insignifi-
cant in most cases.  Figure k summarizes example interactions
between the physical factors controlling bed expansion at 25 °C
and the particle diameter.  This figure also contrasts the
unhindered terminal settling velocity to the superficial up-
flow velocity required for expansion for the specific gravity
particles of 2.65.  These data indicate that the lower den-
sities (l.2 specific gravity) have  upflow expansion velocities
in the region where solids management is compatible with
requirements.  This is achieved with particle diameters between
0.1 and O.i* mm.  The operating zones for fluidized beds and
expanded "beds are qualitatively indicated on this diagram,
indicating that the higher velocity requirements for the fluid-
ized bed achieve a shorter hydraulic retention time.  These
reactors also require higher specific gravity particles.  For
example, an  expansion velocity of  60 ft/hr is required for a
particle diameter of 0.2 mm with sand.  A 1 mm size sand par-
ticle requires velocities of 300 ft/hr or greater.

Head Loss Considerations

     Head loss through an expanded  bed or a fluidized bed is
given by the general relationship equation as follows:

                         L (0  - 0)               '   -
                    AP o   -^L_ • (1 . e)


     where       P = head loss through the filter
                 L = bed height
                 e = porosity of the expanded bed
                as = specific gravity of the particle
                 o = specific gravity of the fluid
                            28

-------
lOOOnr
    ooi   5o2    oJoe
        0,2.    .04  ,06   1.0
PARTICLE  DIAMETER, mm
10
              Relationship between spherical particle diameter» superficial xipflov velocity
              required for bed expansion, particle specific gravity at. 25°C.
     V
                                           29

-------
     In general, the friction pressure loss through media with
a specific gravity similar to sand (2,65) results in head
losses of 1 ft per ft of bed depth or greater, whereas the
lower specific gravity particles required in an expanded bed
results in much lower head losses, usually on the order of
1 in per ft of bed or less.

Expanded Bed Dimensions

     Since no full scale expanded beds have been built to date,
the information on the size of the system can be discussed in
general terms only.  Once the capability of the biological
process  is  established and the particle management has been
defined, the  remaining concerns relate to volumetric require-
ments  for  flow and the dimensions of the unit.  The relation-
ship between substrate concentration, depth of the columns,-
the loading rate, and hydraulic retention time,are illustrated
in Figures 5 and 6.  Since these are arithmetic relationships,
they are only presented here as design guides for considera-'
tion of various processes. • The height of the process is
intimately involved in the particle selection and the bed
management requirements.  The shorter reactors would result
in lower retention times at velocities that are acceptable
for solids management.  The typical range of upflow velocities
in the fluidized bed tend to favor deeper beds.  Of course,
the relationship between velocity and depth can be changed by
adding recycle to the system.  The recycle requirements should
be minimized and only utilized for bed management purposes.'
Note that at minimum flow requirements with the expanded .bed ••
it may be necessary to include a pumped recycle.

STATUS OF PROCESS DEVELOPMENT

     Previous studies have focused on defining the character-
istics of  the attached film, in relation to synthetic sub-
strate concentration in sewage (l) and the effects of tempera-
ture (U), shock loadings (5), and particulates on -the inter-
action (6).   Ongoing studies are evaluating the thermophilic
kinetics on both soluble and particulate substrates (13).
Although a complete review of the kinetics of the process is
beyond "the scope of this paper, a review of selected data is
included here to indicate the process capability.  The rela-
tionship between temperature, substrate concentration, and
process loading rate is summarized in Figure 7-  These data
show a wide scatter but indicate,  as  reported earlier by
Switzenbaum (U), that the temperature effects are not as
                               30

-------
       HIGH  EFFICIENCY UPTAKE RATES
                                               THEORETICAL

                                               MAXIMUM
                                                         I

                                                         c

                                                         C
                                                         o


                                                         33
                                                         m
                                                         H
                                                         m
                                                         z


                                                         o
                                                         o
       Q2    0.4   1.0    2    4      10   20  40


       ORGANIC LOADING  RATE, KgCOD/m3-d
Figure 5- Relationship between process volumetric organic loading rate,

        reactor volume requirements, and hydraulic retention titne

        for varying substrate concentrations.
                              31

-------
  100

   60

   40
c EO-
E

t '1
o  8
Q  c
TYPICAL RANGE
FLUIDIZED  BEDS
IU
O

fc  a
   IX)
  as
  0.6

  0.4
  0.2
  0,1
                         TYPICAL
                         RANGE
                         EXPANDED
                         BED
          0.2    0,4  0.6 OS 10
               6  8 10
20
40
                           H R T, hours
            6. Relationship between reactor height, hydraulic retention time,
               and upflow velocity (superficial or empty feed basis),
                                      32

-------
U)
                                                                  OHQANIC LOADING  RATE,  Kfl  COD/nT-d
                                                                                                                   100
                                                Figure  7,
Organic  loading rate effects  on  total effluent soluble COD for a wide
range of steady state operating  conditions  (HRT values from 0.3 to 6 hr,
50 to 600 mg/« influent TCOD, temperature 10°C to 30°C) from reference  (1|).

-------
significant as one might expect with anaerobic processes
applied to dilute wastewaters.  Figure 8 contrasts the effluent
when primary settled sewage was treated with the expanded bed
to that reported in a pure oxygen fluidized bed study (lU).
     A study "by Morris (6) focused on the interaction of
particles  in  the expanded bed process.  It was found that at
35°C pure cellulose particles loaded at a reactor loading rate
of less than 7 kg/m3/day resulted in a total effluent COD con-
centration of less than 60 mg/£.  Thus the loading rate and
effluent quality relationship shown in Figure 7 for soluble
organics also appears to hold true for particulate matter.
     Although much additional work is required to define the
detailed kinetics of the expanded bed process, Switzenbaum (U)
showed that a highly simplified equation could describe the
biological reaction rates.  At low influent concentrations, the
following equation was found to correlate the effects of sub-
strate concentration on the process efficiency in the expanded
bed:

                          S,
                          5* - K2 . A
                           o

     where      S^ = effluent COD concentration
                So = influent COD concentration
                K2 = removal rate coefficient
                 A = specific film•substrate utilization
                     rat e, day~1

     The coefficient, K£» was closely correlated in a tempera-
ture relationship.  This temperature relationship was used to
extrapolate the reaction kinetics to 55°C, with the relation-
ship between substrate removal efficiency, temperature, and
removal rates and reactor volume shown in Figure 9-  It is
interesting to note that the early data developed by Schraa
(13) on the thermophilic films' interaction with soluble
substrates indicates that the rates achieved in Figure 9 will
be supported.

DISCUSSION

     The previous review of principles involve'd in controlling
and defining the expanded bed process can be used to illustrate
the AAFEB potential.  As was indicated in the Introduction,
the process appears to be capable of producing a secondary
effluent quality without production of significant secondary
microbial sludge.  If the reactor upflow velocities are
                              34

-------
oo
en
  140



  130



  120




6

 " 10$*-


o  on
m  w



I80
o

"  70


£  60



   50



§  40



i  30



   20



   10
                                                                        TOTAL COD
                                                                                            BOD
                                              J	I  I I I Illl
                                I,   I  I  I  I Illl
I    II
                                         0.1  0,2    0.4   0,81     t     4       10    20    40       100


                                                     ORGANIC UOAOING RATE, kg BOD or COD/m3-d


                                         Figure  8,  Comparison of process efficiency of the AAFEB  (the COD data, reference [!»]!

                                                   and  a pure oxygen fluidized bed (the BOD data, reference [ill]) when applied

                                                   to primary settled sewage.

-------
   1.0

   0.9


   0.8


   0.7
«f 0.6
 09
CO
-L 0.5
o
z 0.4
iii ^*^
5     .01
10° 20°   30°
0.04
0.1
                          0.4
                                                   1.0
B  1.0

I  Oi9
nj
""as
<
K 0.7
CO
CD
w 0.6
   0.5
   0.4
       I          4        IO            40        IOO         400

       REACTOR VOLUMETRIC  REMOVAL  RATE, gm COD/l-d
      Figure  9.  Comparison of AAFEB removal kinetics  at varying
                temperatures.and process  efficiencies.  The specific
                removal rate was calculated i'rom the  relationship
                3»/30 = KoA where Kg = 1.T7, 1.21, 0.75, and 0.25 l/d
                for temperatures of 10, 20, 30, and 55°C, respectively.
                Volumetric removal rates  were calculated assuming that
                the AAFEB aiicrobial mass  was 50 gm VS/l.  Values for A
                from reference ( U ) for 10, 20, and 30°C and the values
                for 55°C were estimated using temperature effect on Kg
                frora reference (U, 13).
                                   36

-------
compared to the clarifier velocities, it is clear that one can
expect effluent suspended solids to be low.  If the process is
operating at a design organic loading rate of 7 kg/m3/day, the
reactor will have a hydraulic retention time of approximately
one hour.  Tests with primary settled sewage indicate that the
effluent quality at this loading rate should be high and that
one could expect much of the BOD to be converted.  If 200 mg/£
of BOD is converted, this should result in a net yield of
approximately•0.75 gm/£/day.  If the effluent suspended solids
are lost equivalent to 30 mg/£, the net change in volatile
solids in the system is zero.
     The long sludge retention times required to achieve an
efficient anaerobic reaction and the high substrate concen-
trations required to drive the reaction combine to make the•
task of treating dilute, low temperature wastewaters amongst
the most difficult challenges for anaerobic processes.  It is
essential that the solids management as well as the biological
process be carefully controlled.
     A comparison of .various particle sizes and SET values
illustrates the problems that will occur if the fluidized bed
process is used for sewage treatment, as compared to the
expanded bed process.  If it is assumed that sewage has an
organic content of 230 mg/£ of BOD and a temperature of 20°C
and effluent solids from the reactors are limited to 15 mg/Jl,
the following compares the solids retention time and therefore
the capability of the processes to produce the" required
effluent.  It is assumed that both reactors have equal effi-
ciencies, even though this will probably not be the case.  It
will also be assumed that 200 mg/£ BOD is removed in each.
Both units will be 20 ft deep.  The expanded bed will have a
50 minute hydraulic retention time, whereas the fluidized bed'
will have 6.5 minutes.  Due to the expansion and the small
particle use in the expanded bed, the operating mass is
estimated to be approximately kO gm/£ of reactor.  The fluid-
ized bed will have an operating mass of approximately 8 gm/£
if 300 percent expansion is used.  Based on the above assump-
tions, the net yield in the expanded bed is O.U gm/£/day,
whereas in the fluidized bed it would be 3.31 gm/£/day because
of the increased reaction rate per unit volume that is
required.  The resulting solids retention time is nearly 100
days in the expanded bed, as compared to 2.U days in the
fluidized bed.  Clearly, the' velocities and the solids manage-
ment in the expanded bed result in the requirements for both
the biological and physical processes to treat low strength
wastewaters, whereas the fluidized bed is operating at very
short solids retention times and will achieve a low quality
                               37

-------
of effluent.
     Finally, it is possible to make some gross comparisons
between aerobic and anaerobic processes based on the data that
are available.  Figure 10 illustrates the relationship between
the substrate removal rates and the resulting solids retention
time in various aerobic and anaerobic processes that are able
to achieve different reactor concentrations of biological
solids.  There are numerous assumptions included in this
figure.  For example, it is assumed that the efficiencies and
the removal rates of the processes are compatible.  The sur-
prising results that this overview emphasizes is that the
anaerobic process capability exceeds all aerobic processes.
The high concentration of microorganisms in the AAFEB result
in much longer solids retention times than the aerobic pro-
cesses under comparable loading rates.  This indicates that
high organic loadings that are achieved either at high flow
rates or organic concentrations with the aerobic processes
can easily lead to unstable situations; whereas the anaerobic
processes can still have an acceptably long solids retention
time so that they can continue to operate successfully.
     The AAPEB studies show significant promise for the
application to a wide variety of wastewater purification
problems.  Areas that require further research and development
are as follows:
     - process scale-up to large pilot or full scale;
     - impact of toxic substances j
     — fundamental study of the biological reaction kinetics
       as affected by film thickness, substrate characteristics,
       and temperature;
     - definition of the physical filtering capabilities of an
       expanded bed;
     - definition and application of the thermophilic expanded
       bed;
     - definition, of impact on major practical problems such
       as:  algae and eutrophication management, waste acti-
       vated sludge treatment, and retrofit to existing pro-
       cesses ;
     - definition of physical process requirements for inex-
       pensive inert particles coated with anaerobic microbial
       films, i.e., upflow velocity, bed management needs,
       recycle, solids, wasting;
     - development of the specific application of series
       anaerobic-aerobic treatment to achieve high efficiency
       of carbon and nitrogen removal without any chemical
       additives.
                              38

-------
100
       I   I   I  -I	I—I—1—I
                                                       WIN. YIELD
                                                       lOOKg/m3
                                            >ANAEROB1C
                                                    \ MAX. YIELD
                                                       3OKg/m3
                                                    / MIN. YIELD
                                            1        /  30Kg/m5
                                 f          /AEROBIC
                                 (SOLIDS     I        \
                                 v ccccn-r J        \ MAX. YIELD
                                                       30Kg/m3
/SOLIDS
\EFFECT
  EFFECT
     0           50           100
       SLUDGE  RETENTION  TIME, days
                                  ACT. SLUDGE 5Kg/nrr
                               j	i	l	
         10. Comparison of resulting solids retention time and varying
            substrate removal rates for anaerobic and aerobic processes.
                                39

-------
     The title of this paper indicated that the topic was to
be the fundamentals of a new process referred to as the
expanded bed process.  It is clear that the fundamentals that
apply still remain to be defined in relation to many applica-
tions of the anaerobic expanded bed process.  The work to date
has focused on attempting to define the limits of the biolog-
ical capability of a high biomass anaerobic system.  Ongoing
work indicates exciting possibilities for the applications of
high temperature films, especially to concentrated waste
stream management, excess waste activated sludge and substrates
such as algae and weeds for energy production and pollution
control.

SUMMARY AND CONCLUSIONS

     The combination of process characteristics of the expanded
bed filter with anaerobic microbial films has resulted in a
process that provides the opportunity for maximum biomass con-
centration development while good control over the fluid forces
required to retain solids is achieved.  This enables the pro-
cess to produce such surprising results as secondary treatment
quality effluents from dilute wastes even at low temperatures,
and substrate removal capability greater than any biological
process, including all aerobic alternatives.
     Two major unexpected results account for the capabilities
of the AAFEB process.  The anaerobic microbial films are ex-
ceptionally thin (around 0,020 mm) at low substrate concentra-
tions, thus preventing mass diffusion limitations, and high
bulk densities (calculated values in excess of 200 gm VS/£ for
anaerobic films contrasted to 3^ gm VS /£ for the thick aerobic
films). Due to the combined characteristics of the expanded
bed to achieve maximum biomass and suspended solids control at
relatively high processing rates, it should be a desirable
process for all microbial conversions.  Additionally, the
small particulate filtering capacities of the expanded bed
appear  to  be significant but undefined.
     Studies in progress on thermophilic films show promise
of  developing  high rate processes for concentrated waste
streams.  Further research and development efforts should
focus on both the fundamentals of the process and scale-up
applications.

ACKNOWLEDGEMENTS

     The AAFEB research has been conducted with the support of
Cornell University and many dedicated .individuals.  To date
                             40

-------
no major grant has been available to support the development
of this technology.  Major contributions to this technology
have been made by Michael S. .Switzenbaum, James W. Morris,
Robert J. Cummings, and A. P. Leuschner.  Other individuals
who contributed to one or more studies include: W. W. Clarkson,
R. Lobdell., S. Morris, J. Neander, J. Nolfi, R. Orenstein,
J. Simpson, and S. H. Zinder.  Mr. Gosse Schraa is presently
conducting experiments with high temperature attached films.
     The author is grateful to the following for financial
support of short-term feasibility efforts and for graduate
student assistantships:  the U.S. Department of Energy
(Contract EY-S-02-298l), the Solar Energy Research Institute
(Contracts XB-9-8263-1 and XB-0-9038-1), and Canadian Liquid
Air.

REFERENCES

1.  Jewell, W. J., Switzenbaum, M. S., and Morris, J. W.,
    "Municipal Wastewater Treatment with the Anaerobic
    Attached Film Expanded Bed Process."  Journal Water
    Pollution Control Federation, 53(U):H82-U90.  1981.
2.  Jewell, W. J., "Development of the Attached Microbial
    Film Expanded Bed Process for Aerobic and Anaerobic
    Waste Treatment."  Paper presented at the Biological
    Fluidised Bed Treatment of Water and Wastewater Con-
    ference, University of Manchester Institute of Science  .  .
    and Technology, England, April lU-17, 1980.
3-  Jewell, W. J., "Future Trends in Digester Design." In:
    Anaerobic Digestion, Proceedings of the First Inter-
    national Symposium on Anaerobic Digestion, University
    College, Cardiff, Wales, September 17-21, 1979-  1980.
    pp. U67-U91.
U.  Switzenbaum, M. S. and Jewell, W. J., "Anaerobic
    Attached Film Expanded Bed Reactor Treatment.  Journal
    Water Pollution Control Federation, 52(7):1953-1965.
    1980.    .                                    .
5.  Jewell, W. J. and Morris, J. W., "Influence of Varying
    Temperature, Flow Rate and Substrate Concentration on
    the Anaerobic Attached Film Expanded Bed Process."  In:
    Proceedings of the 36th Industrial Waste Conference,
    Purdue University, May 12-lU, 1981.  pp. 655-66H.  1982.
•6.  Morris, J. W. and Jewell, W. J. , "Organic Particulate
    Removal with the Anaerobic Attached-Film Expanded-Bed
    Process."  In:  Proceedings of the 36th Industrial Waste
    Conference, Purdue University, May 12-lU, 1981.  pp. 621-
    630. 1982.
                             41

-------
 7.  Young, J. C. and McCarty, P. L., "The Anaerobic Filter for
     Waste Treatment."  Journal Water Pollution Control Federa*-
     tion3 1*1 5(2)Rl60.  1969.
 8.  Rich, L. G. , Unit Operations of Sanitary Engineering, John
     Wiley and Sons, pp. 11*6-151.  1961.
 9-  Cleasby, J. L. and Fan, K., "Predicting Fluidization and
     Expansion of Filter Media."  Journal of the Env. Eng. Div.t
     June, 1981, 107 (No. EE3), pp. 1*55-^71.  Paper No. 16321
     in:  Proceedings of the American Society of Civil Engineers.
10.  LaMotta, F. J. , "Internal Diffusion and Reaction in Bio-
     logical Films."  Journal Environmental Science and Tech-
     nology, 10  (no. 8).  August, 1976, pp. 765-769.
11.  Kornegay, B. H. and Andres, J. F., "Characteristics and
     Kinetics of Biological Film Reactors."  Federal Water
     Pollution Control Administration, Final Report, Research
     Grant ¥P-01l8ll.  Dept. of Environmental Systems Engineer-
     ing. Clemson University, Clemson, SC.  1969-
12.  Rittmann, B. E. and McCarty, P. L, "Substrate Flux Into
     Biofilms of Any Thickness."  Journal of the Env. Eng. Div.3
     August, 1981, 107 (No. EEl*), pp. 831-849 in: Proceedings
     of the American Society of Civil Engineers.
13»  Schraa, G., "Thermophilic Anaerobic Attached Film Expanded
     Bed Treatment of Soluble Organics."  Ph.D. Dissertation,
     Dept. of Agricultural Engineering, Cornell University,
     Ithaca, New York.  In preparation.
lU.  Nutt, S. G. , Stephenson, J.  P. and Pries,  "Aerobic
     Fluidized Bed Treatment of Municipal Wastewater for Organic
     Carbon Removal."  Presented at the Water Pollution Control
     Federation Conference, Houston, Texas.  1979-
15.  Lawrence, A. W. and McCarty, P. L, "Kinetics of Methane
     Fermentation in Anaerobic Treatment,"  Journal Water Pol-
     lution Control Federation, 1*1 (no. 2), part 2.  1969•
16,  Gates, W. E., et_ al., "A Rational Model for the Anaerobic
     Contact Process."  Journal Water Pollution Control Federa-
     tion, 39  (no. 12).  1967.
17.  Beneder, P. and Horvath, I., "A Practical Approach to
     Activated Sludge Kinetics."  Water Research, 1  (no. 10),
     1967.
18.  Gram, A. L., "Reaction Kinetics of Aerobic Biological
     Processes."  I.E.R. Series 90, Report 2, Sanitary Engi-
     neering Research Laboratory, University of California,
     Berkeley.   1956.
                              42

-------
        TRICKLING  FILTERS:  RELIABILITY,  STABILITY
                AND POTENTIAL PERFORMANCE
           E D Schroeder. Department of Civil
           Engineering,  University of  California,
           Davis, California
INTRODUCTION

     Trickling filtration is probably the oldest and least understood
of the  modern systems for wastewater treatment.   The  process
was  developed shortly before the turn of  the  century  and in its
original form was  an intermittant or periodic  treatment  system.
Development  resulted in  two  general types  of operation;  the
standard or low  rate process which  is basically the original  one,
and the high rate process which  incorporates effluent recirculation
and  higher  hydraulic and  organic  loading  rates.   Some  design
parameters  and  operating characteristics  of  the  two  types  of
process are given in Table 1.
     The basic operations difference  between  the  standard  and
high rate operation  is effluent  recirculation, and as  can be  seen
in Table 1  the loading characteristics are  considerably  different.
Perhaps  more  interesting  are  the  differences  in   operating
characteristics. Effluent BOD5 and suspended so Eds from standard
rate trickling  filters  are  usually  comparable to  activated sludge
processes,   while   effluent  from  high   rate   systems   is   less
satisfactory.  In standard rate filters the biological film builds up
for long periods of time.  Large scale sloughing occurs periodically,
most notably  in the  spring.  In  high rate  filters  sloughing  occurs
continuously.
                            43

-------
                                  TABLE 1
                 DESIGN AND OPERATING CHARACTERISTICS
                           OF TRICKLING FILTERS
CHARACTERISTIC
Depth, m ? 3
Specific Surface m /m
Porosity
Media size, mm
Hydraulic Loading*
3 2
Rate, m /m -d
Organic Loading*
Rate, kg BOD5/m3-d
Recirculation Ratio
Sloughing
Nitrification
3
Effluent BOD5 g/m
Effluent SS, g/rn3
STANDARD RATE
1.8 -3
M -65
0.45 -0.55

25 - 75
0.9 -2.8

0.11 - 0.37
0
intermittant
yes
<25
<25
HIGH RATE
ROCK
1 -2.5
M -65
0.45 -0.55

25 - 75
9 -28

0.37 - 1.8
1 - 4
continuous
at lower
loading rates
>30
>30
PLASTIC
4 -10
80 -100
0.90 -0.97
Dependent upon
configuration
20 - 75

up to 15
1 - 4
continuous
not in economic
range of operatioi
>30
>30
* Calculated  using influent flow rate and Bod concentration

-------
      Because the loading conditions are quite different the  actual
effect of  recirculation  is difficult  to  determine.  Obviously the
actual hydraulic loading  rates are increased over the nominal value
found  by dividing influent flow rate by cross-sectional area. Flow
variation is damped because of,-the steady recycle component, and
presumably distribution  over  the  media  is  more  uniform  and
complete.  Larger  organisms, such  as fly larvae,  that feed  on the
slime  are  washed  out.  Thus the microbial community should be
different   in  high  and   standard  rate   systems.   Because  the
predator/ grazing  organism  population  is lower  more  overgrowth
and  plugging  problems  might be expected  in high rate  trickling
filters.  The  opposite is the case, however.  Two factors  appear
to be involved. First, the higher flow rates result in more complete
distribution of the nutrients through the  volume and the  result  is
more uniform growth. More important  are the  higher  shear rates
associated  with the larger flow  rates.  Bruce (1)  reported that at
higher  hydraulic  loading  rates   shear  was  the   principle control
mechanism  and that at  lower rates slime accumulation the most
important control mechanism was grazing by invertebrates.   Solbe
and  Roberts (1) performed an inventory of invertebrate organisms,
in an  experimental standard  rate  unit  over a three  year  period
and  found  both  the total  slime mass  and the populations to be
highly variable.  The  spring sloughing resulted  in large decreases
in invertebrates  as well as  bacterial slime.  It  is assumed  that
the  large  accumulation  of film  during  the  winter months  is the
result of decreased  invertebrate activity at lower  temperatures
and  the spring sloughing is  caused by their renewed  activity.
Recycle
      The  effect  of recycle on  process  performance  has  always
been controversial.   Many  workers have considered  the recycle
stream to  provide additional  passes through the reactor (3), and
therefore improving process performance.  This would  actually be
true  only  if  the  recycle stream  remained segregated  from the
influent, a situation that is difficult to  conceive.  A similar result
would be obtained  if the higher flow  rates of a recycle system
caused a more complete wetting of the trickling filter surfaces.
Over  designed  units  and  systems   with  high influent  BOD
concentrations where  the organic loading rate controls  the process
design would be examples that might appear  to follow  the multiple
pass concept.
      Recycling  the  process effluent  should have three  physical
consequences; 1) diluting the influent  stream, 2) increasing the
liquid  film  depth  and 3) incorporating  sloughed  microbial culture
into the liquid film 4, 5,  6.   The  first  two factors will  decrease
process performance because liquid phase transport will be  slower
                              45

-------
at lower concentrations and greater distances.  Recycling sloughed
cells could  result in significant quasi-homogeneous  reaction rates
in the liquid  film.   Oxygen transfer would  be less of a problem
because  the portion of the  reactions  taking place  in  liquid film
would  be closer to the air-liquid-interface.   The summary effects
of these  three factors is not clear.  High  rate trickling filters
remove  considerably  more organic  material  than  standard rate
units  but effluent quality  is considerably lower also.
Media
     Highly porous plastic media has been  increasingly  used  in
recent years.   The  greater porosity and  regular  shapes can  be
expected to result in more uniform flow  distribution and improved
oxygen transfer.   The  advantages of  plastic  media are realized
only at  high   loading  rates  where  conventional media would  be
quickly plugged (7).  At lower  rates  rock media systems perform
as well or better than plastic media units.  This is not particularly
surprising because the available surface area per  unit volume  is
not greatly different.
Biofilm
     The attached biological  slime  in trickling filters  is  highly
variable.   Mass distribution varies  with time,  season and  flow
(1,8,9).
Bacteria  are the dominant types of organisms  although fungi such
35 Geotrichium are often present in significant amounts.   There
has been very little  study of  the  structure of trickling  filter
biofilms.  They are obviously not uniform and vary greatly in depth.
Pivetti (8) observed the accumulation of biomass as a  function of
depth  in a pilot  scale  (0.25  m diameter,  2.4  m  media  depth)
trickling  filter using  5 cm  plastic  pall  rings (Nortoa Actifil   ).
The hydraulic loading  rate was held constant  at 9.9 m /m -d (10.6
mgad or  4.13  d~ ) but three organic loading rates of  approximately
0.31, 0.81 and 1.05 kg BODjm-d. Both  the  hydraulic and organic
loading rates  fall  in the lower range  for high rate trickling filters.
Slime  mass varied with depth for all three loading  rates as  shown
in Figure 1.
     Make  up of the biofilm includes many filamentous.
Sloughing
     Little information is available on the mass  rate of sloughing
because  most  workers report secondary clarifier effluent suspended
solids  concentrations  rather than  trickling filter effluent values.
Eden  et  al (7) s"bJdied plastic media (Surffpac) at high loading
rates (6.7-21.5 m /m  d  and  0.9-2.4 kg  BOD5 /m   d)  over  a one
year period and reported  trickling filter  effkient suspended  solids
concentrations ranging from 119 to 198 g/m3.  Influent suspended
solids  concentrations  were  approximately 25 percent higher  and
                             46

-------
g
ca
    500 _
    WO |_
    300 |_
                            INFLUENT COD,  g/m'
o
3
w   200 U
y
     100 (_
  FIGURE 1     VARIATION IN ATTACHED SOLIDS WITH
               DEPTH A.T HYDRAULIC LOADING RATE
               OF 10 in /nrT-d 8
                               47

-------
thus  a large  fraction  of  the  effluent  solids could  well have
originated  in the influent.  Pivetti worked  with a soluable fed and
reported average effluent suspended  soiids values  (unsettled)  of
10, W and 72 g/m   for  the three organic loading rates  he used.
      Aitken (10) used a  hydraulic  loading  rate  15 m /m  *d and
an  organic loading  rate  of 2 kg/m 'd  in  studies with a 0.15  m
diameter,  1.1  m deep model plastic media trickling filter.  Like
Pivetti ,the feed was  soluble.  Effluent  suspended solids averaged
23 g/m  over a  103  day  period,  with  a standard deviation  of 7.7
g/m .   The media  used  by Aitken (10) was  1.3  cm  plastic pall
rings.   Quite possibly the small media size was  a factor  in the
low effluent suspended solids values.    Periodically  Aitken  pulsed
the hydraulic  loading rate to determine the  effect  on  effluent
suspended   solids.     Pulses  consisted  of  increasing  the   liquid
application rate by  a factor of  ^  or  8  for  a 30  or  60  minute
duration period.   Large  increases  in sloughed solids  during the
pulse resulted.  For  a period of  approximately one  day after the
pulse effluent suspended solids were less than  normal steady state
value  (Figure)  2, but after this  relatively brief  period  effluent
suspended  solids approached the  steady state  values.
Nitrification
      Control of nitrification in  biological wastewater treatment
is still a  somewhat  elusive objective.   The  fact  that extensive
nitrification is normal in slow rate  trickling filters  and relatively
insignificant  in  high rate  systems is  a consistant observation.
There  is no reason to suspect that nitrifying organisms are washed
out of the  high rate systems and another cause must  be considered.
A reasonable  conjecture  is that competition for  oxygen  is  too
great  in high  rate systems for  nitrification to occur.  If  this is
correct nitrifiers activity should  be limited to the  lower  depth  of
low  rate  units  but  this has not  been  demonstrated.   The  actual
mechanisms of  nitrification have  not  been  established.   Quite
possibly  adsorption  of NH   on  slime surfaces is the ammonia
removal mechanism, with oxidation of the adsorbed NH  + following.
An interesting set of experiments could be  developed^to  test this
hypothesis.
CONCEPTUALIZED  AND REAL TRICKLING  FILTERS
      A considerable  amount of mathematical modeling of trickling
filters has been done.  Examples include the early work of Velz
(11),  Howland (12) and Eckenfelder (13) that utilized  first order
reaction models, the work of   the  mid-196Q's characterized  by
Swilley and Atkinson  (4), Kehrberger and Bush (5),  Meir et  al (14)
and  Kornegay  and   Andrews    and the more recent  studies  by
Atkinson and his co  workers (16,17,18),  Williamson and  McCarty
(19),  Rittman and McCarty (20) and Harromoes (21). All  of these
                            48

-------
workers  were  forced to  make idealizing  assumptions about  the
fluid  flow conditions.   Generally  these  assumptions are  uniform
steady flow over the entire surface, a smooth microbial slime of
uniform depth  throughout the  trickling filter  and all reactions in
the slime.  Such  assumptions  are  very unrealistic.  All trickling
filters are loaded periodically  if one considers  a particular point
or flow path.   Thus  flow would be  expected to occur in a rippling
pattern with  mixing occuring  at  points  were sections of media
intersect. The microbial slime can be seen to vary in characteristics
throughout trickling filters and sloughing would  result in  a patchy
surface of variable friction characteristics.  Under such conditions
it  would  be  surprising  to  find the wastewater  running  smoothly
over  the  entire media surface.   Quite likely   the  flow runs in
riveluts for short distances before joining  with other streams which
are then separated out  into smaller  flows  at  media  interfaces or
junctions.  As growth  and sloughing  occurs the pathways of the
flow  would  be  expected  to   change  resulting  in  an extremely
dynamic system.
      The  third assumption, all reactions  occur in the microbial
film,  has  varying validity.   Sloughed  film would be  biologically
active. Recycling from  a  point prior to  the  secondary  clarifies
would enrich  the liquid  film  with  micro  organisms  and  result in
what  Swilley (22) termed a pseudo-homogeneous reaction system.
Swilley    concluded     from     theoretical     studies    that
pseudo-homogeneous systems would have better performance  than
heterogeneous  systems  and that   recycle  would  improve process
performance  if  suspended  cells   were   included  and  decrease
efficiency if suspended cells were excluded. Kehrberger and Busch
experimentally  validated Swilleys conclusion for  an inclined plate
system.  Unfortunately the idealized flow conditions of an inclined
plate  are  quite different from real trickling filters and  there  is
substantial evidence that removal rates are increased when recycle
is used, regardless  of  the configureation  (ie  before or after the
secondary  clarifes).
      Most of the recent attached growth models (14-22) are based
on mass transport concepts.  None  consider the possibility of more
than  one  limiting nutrient.   In their most easily  applied forms
there is an assumption  that the same mechanism is rate limiting
throughout the depth,  but  models such  as Atkinsons  (16,17) are
based on  spatially varying  conditions.  These models are useful  in
delineating the  interaction of  system variables and parameters but
are  far too  simplified  to  predict process  performance without
extensive, system  specific calibration. This was demonstrated by
Atkinson  and  All (28)   and  Pivetti  (5)   in their studies  with
simplified systems.
                              49

-------
ACTUAL PROCESS PERFORMANCE
      A number of reports on actual performance of trickling filter
systems  over extended  periods  of time are  in  the  literature.
Particularly notable in  the historical sense is  the NRC report (23)
the  used annual average values to develop  loading/performance
relationships.   Although  the results were at  best shakey  (24) the
NRC formula  is  still  used in  design.   Gallen and  Gotaas  (25)
developed a performance relationship based on the best fit of data
from trickling filters and Fairall (26) and Rankin (27)  worked with
average  data  from a number of systems also.
      More recently Haugh  et  al (28) and Niku et al (29) reported
the results of the analysis of one years daily  composite data  from
11 high rate trickling filters systems located through the midwest.
Average daily flows ranged from  900 m /d (0.5 mgd) to 130,000
m /d (34 mgd). In these studies summary statistics (mean, standard
deviation, show  etc) were examined to determine general process
characteristics.   Five common probability density  functions  were
tested with the  effluent BOD
-------
          TABLE 2
EFFLUENT BOD5 DATA FROM 11
MIDWESTERN TRICKLING FILTERS
Plant
1
2
3
^
5
6
7
8
9
10
11
Daily
Average
g/m3
33.3
10.7
10.1
5S.f
29.2
27.0
23.2
43.1
51.1
21.0
18.3.
7 day
Running
Average
g/m3
33.1
10.6
10.0
58.3
29.2
27.0
22.9
4-3.0
51.5
21.0
18.0
30 day
• Running
Average
g/m3
32.1
10.0
9.8
57.7
29.0
' 26.8
22.6
^2.8
52.9
,21.1
17.1
            51

-------
               TABLE 3
   EFFLUENT SUSPENDED SOLIDS DATA
FROM 11 MIDWESTERN TRICKLING FILTERS
Plant
1
2
3
4
5
6
7
8
9
10
11
Daily
Average
g/m3
52.5
21.5
21.0
54.9
18.3
15.1
24.1
34.0
41.1
23.6
16.2
7 day
Running
Average
g/m3
52.5
21.2
21.1
54.8
18.4
15.1
23.6
34.0
41.2
23.9
16.1
30 day
Running
Average
g/m3
51.6
20.9
21.1
54.3
18.6
14.9
23.2
34.0
41.8
24.0
15.8
                  52

-------
      Process  reliability has been defined by Niku, Sarnaniego and
Schroeder  (29) as the  probability  that a given system will meet
a chosen standard.  The proposed a coefficient  of  reliability (COR)
based  on the log normal distribution that  uses  and a processes
coefficient  of variation,  V
      COR = (Vx2  +  I)* exp  j- z{_a  [ln(V
where  z ,    =' standard normal variate  for  the  distribution.
     Process  reliability can  be plotted as a function  of  V  and
the nomalized mean, m /X (m  - actual or design value ana X   =
standard value) as shown in  Figure 2.   Based  on the  data from
the  11  plants  V   values  for  effluent BOD^  and  SS  should be
approximately 0.50 and 0.55,  respectively.  Using Figure 2 and X
=  30 g/m  we can conclude  that 95 percent reliability would
require  an  average effluent  BOD^  concentration  of slightly less
than 15 g/m .
     Evaluating process stability is a  more qualitative procedure
than evaluating reliability.   Defining  stability  is  somewhat of  a
problem  in  itself.   Normalized  parameters  generally  do not
differentiate between systems with  good and poor effluents.  For
example of the ratio of the standard deviation  to the mean were
used a  plant  with a  ratio of one  could have  standard deviation
and mean values of 5 and 5 or  100 and  100.  The standard deviation
is  a useful value  in  estimating process  stability,  but does not
provide  information  on  what caused  a  particular  value.   For
example a given standard deviation might be the result of a number
of  small  deviations or  one or  two colosal failures.   The  range
provides information on  maximum  values experienced but not on
their frequency.  A plant  with one  failure  per  year  would not be
called  unstable by most  people, but might be  identified  as such
if  range were  the sole criteria. For  these reasons an  ideal stability
measure does  not  exist, but a somewhat qualitative estimate can
be  developed  by plotting range vs  standard  deviation  (Figures  3
and  4).  As can be seen  the range  tends to increase  with  standard
deviation    for   both  effluent   BOD,-   and   suspended   solids
concentration.   A standard deviation value of  10 g/m   was  taken
as the  stability  cut off point for both  variables. The decision was
based to a  large extent on a similar analysis for activated sludge
processes where the differences are considerably more clearcut.
SUMMARY AND CONCLUSIONS
     Trickling filters can be best designed using pilot scale studies.
Estimates of process performance can be made using  simple models
incorporating loading rate  parameters and the reliability of process
can  be  estimated  for a given effluent standard using  Figure  2.
                             53

-------
SJ
oi
       0.2  -
       0.1
         O.I  0.2  O.3 0.4 0.5  O.S  0.7 OS 0.9  1.0  1.1  1.2 1.3  1.4  1.5


                     Normalized mean mx/X


    FIGURE 2      RELIABILITY  AS A  FUNCTION  OF
                   COEFFICIENT OF VARIATION AND

                   NORMALIZED MEAN
                                 54

-------
    ISO
    140
    120
•*">   100
     so
     60
1 ' I . I 	 I "
Stable


n Mean
fartf Deviat
	 T~
Unsta&e
"on
u
~ Plant Number —r
-
-
c
i i
$3
£
1 P
<*i
i
•j


%
•)








t
	 T 	 1 — 	
0
•» fs







1
r~
i
"«l




	 1 	
r- _
-
-
,
-
                   Standard Deviation g/m

   FIGURE 3     VARIABILITY OF EFFLUENT BOD
                 AS A FUNCTION OF STANDARD
                 DEVIATION AND RANGE
                                     16      2O     24
                       55

-------
CJ

BO
c
a

-------
Process  stability  cannot  be predicted,  but  it  is  clear  that the
lower  the  effluent  BOD and  suspended  solids concentrations are
the more stable a plant will  be.
      High rate trickling filters can produce  good quality  effluent
as shown in Tables  2 and  3.   In general the  lower the organic
and  hydraulic  loading  rates the  better  the  effluent quality will
be.  The performance history  of  low rate systems is quite good,
but capital and land requirements are  high and this  will probably
restrict their use to smaller communities even under current energy
restrictions.  High  rate  trickling  filters can  be competitive with
activated sludge processes in terms  of land and cost, but  effluent
quality  is  definitly  lower   than   activated  sludge processes.
Application of 30-30 standards  makes  selection  of high rate
trickling  filters risky  unless  further  treatment is  included.   If
standards  are  allowed  to reflect a  particular  set  of discharge
conditions trickling filters would be a more widely used alternative.
                             57

-------
REFERENCES

1.  Bruce, A.M.  "Percolating Filters," Process  Biochemistry,  4,
    April,  1969.
2.  Solbe,  3.F. and H. Roberts  "The colinization  of a Percolating
    Filter  by  Invertebrates  and  Their  Effect on  Settlement  of
    Humus Solids", Water Pollution.
3.  Fair,  G.M.  and  3.C.  Geyer   Water Supply  and  Wastewater
    Disposal,  3ohn Wiley and  Sons,  New York, 1956.
4.  Swilley,  E.L.  and B.  Atkinson    "A Mathematical  Model For
    The  Trickling  Filter"   Proceedings   18th  Industrial  Waste
    Conference, Purdue University,  1963.
5.  Kehrberger, G.3.  and A.W. Bush  "Mass  Transfer  Effects  in
    Maintaining Aerobic Conditions in Film Flow Reactors" 3ournal
    Water  Pollution Control  Federation, 43,  1514,  1971.
6.  Schroeder, E.D.   Water  and  Wastewater Treatment,  McGraw
    Hall Book  Company, New York,  1977.
7.  Eden,  G.E.,  G.A.  Truesdale and  H.T.  Mann    "Biological
    Filtration using Plastic Filter Medium" 3. Institute  of Sewage
    Purification, 65,  562, 1966.
8.  Pivetti, D.A.  "Influent Concentration,  Slime Mass Distribution
    and  Substate   Removal  In  A   Trickling  Filter"  MS Thesis,
    University of  California, Davis,   1976.
9.  Howell, 3.A. and B.  Atkinson   "Sloughing of Microbial Film  In
    Trickling  Filters", Water Research,  10,  307,  1976.
10. Aitken, M.D.    "Hydraulic Modeling of  Suspended Solids  In
    Trickling Filter Effluents" MS Thesis, University of  California,
    Davis,  1980.
11. Velz,  C.3.  "A Basic Law For The Performance of Biological
    Beds",  Sewage Works  3ournal, 20, No.  4,  1948.
12. Howland,  W.E.   "Flow Over  Porous Media as  In A  Trickling
    Filter", Proceedings 12th Industrial Wastes Conference, Purdue
    University,  1958.
13. Eckenfeder, W.W.  "Trickling Filter Design and Performance",
    3. Sanitary Engineering Division, ASCE,  87,  SA3,  33, 1968.
14. Maier, W.3.,  V.C.  Behn  and  G.D.  Gates   "Simulation of the
    Trickling  Filter  Process",  3. Sanitary  Engineering  Division,
    ASCE, 93, SA4,  91,  1967.
15. Kornegay,  B.H. and 3.F. Andrews  "Kinetics  of  Fixed Film
    Biological   Reactors"    3ournal   Water    Pollution   Control
    Federation, 40, 460, 1968.
16. Atkinson, B.  Biochemical Reactors, Pios Press,  London,  1974.
17. Atkinson B. and I.S.  Daovd  "Diffusion Effects Within Microbial
    Films" Trans. Inst.  Chemical Engineers, 48, 245, 1970.
                               58

-------
18.  Atkinson,  B. and  M.E. Abdel Rahman  All   "The Effectiveness
    of Biomass Hold-Up and Packing Surface in Trickling  Filters"
    Water Research,  12,  147,  1978.
19.  Williamson,  K.  and  P.  McCarty    "A  Model of  Substrate
    Utilization By  Bacterial  Films"  3.  Water  Pollution  Control
    Federation, 48, 9,  1976.
20.  Rittmann, B.E. and  P.L. McCarty "Substrate Flux Into Biofilms
    Of  Any  Thickness"  3.  Environmental  Engineering Division,
    ASCE, 107,  831,  1981.
21.  Harromoes,  P.   "Significance  of  Pore Diffusion  To Filter
    Denitrification"  Half  Order Reactions  in  Biofilm  and Filter
    Kinetics", VATTEN, 2, 1977.
22,  Swilley, E.L.   Ph.D.  Thesis, Rice University,  1964.
23.  National Research Council  Subcommittee on Sewage Treatment
    "Sewage  Treatment  at  Military Installations,  Sewage Works
    Journal, 18, May 1946.
24.  Schroeder, E.D. and G. Tchobanoglous   "Another Look at the
    NRC  Formula" Water and Sewage  Works, 122, 58,  1978.
25.  Gallen, W.S.  and H.B.  Gotaas  "Analysis of Biological Filter
    Variables" 3. Sanitary  Engineering  Division, ASCE, 90,  SA6,
    59, 1964.
26.  Fairall,  3.M.   "Correlation of Trickling  Filter Data"  Sewage
    Works Journal, 28,  1069, 1956.
27.  Rankin, R.S.  "Evaluation of The Performance of Biofiltration
    Plants" Transactions  ASCE,  120, 1955.
28.  Haugh,  R. ,  S.   Niku,  E.D.  Schroeder and  G. Tchobanoglous
    Performance of Trickling Filter Plants, PB 82-108 143, National
    Technical Information Service,  1981.
29.  Niku,  S., E.D. Schroeder  and F. Samaniego   "Performance of
    Activated Sludge Processes and Reliability Based  Design" CL
    Water Pollution  Control Federation, 51,  284,  1979.
                                59

-------
          PART II:  CURRENT STATUS AND FUTURE TRENDS
   THE HISTORY OF FIXED-FILM WASTEWATER TREATMENT SYSTEMS
     Robert W. Peters.  Department of Civil Engineering,
     Purdue University, West Lafayette, Indiana.

     James E. Alleman.  Department of Civil Engineering,
     University of Maryland, College Park, Maryland.
INTRODUCTION

     The science and 'art1 of wastewater engineering stretches
only slightly beyond one hundred years.  Within this period,
the applied technology has certainly made significant strides
in promoting disease control and environmental protection.
Fixed-film treatment unquestionably plays an important role in
this history, particularly since it represented the original
biological mechanism.   Beginning with options like the
trickling filter, intermittent filter and contact  bed,
fixed—film systems dominated the technology of wastewater
treatment for several decades.  And although this status has
subsequently been assumed by suspended growth process, there
is unquestionably a resurgence of interest in fixed-film
applications.
                              60

-------
     Given the relative historical significance, and projected
future of fixed—film systems, a chronological review of the
associated progressive developments should be both interesting
and informative.  This paper will therefore explore the gene-
alogy behind our current fixed—film technology, condensing the
relevant yesteryear literature into twenty-five year incre-
ments.  While attempting to limit this synopsis to a reasonable
length, every effort has been made to facilitate a thorough
documentation of the associated literature.

1850 - 1875

     As described by the classic Dickens tale in 1859, "It was •
the best of times, it was the worst of times..." (1)  This
literary image poignantly portrays a mid-nineteenth century era
freshly endowed with the blessings of an Industrial Revolution,
yet virtually helpless in the face of rampant, epidemic dis-
ease.  Cholera, alone, flared through the British Isles in four
deadly outbreaks within one terrifying ten year period. (2)
     Without question, these problems with communicable disease
provide a sad reflection on the existing deficiencies in en-
vironmental sanitation.  However, the concurrent infancy of
bacteriology yielded only vague clues regarding the dangerous
correlation between fecal contamination and disease trans-
mission.  Existing efforts towards sewage disposal, let alone
treatment, were virtually non-existent. (3)  Certainly it was
fortuitous,  then, that legislation (i.e. the Nuisance Removal
Act) was enacted in 1858 to control sewage discharge, albeit
more so a function of safeguarding asthetics rather than a
perceived health hazard.   (4)  This emphasis quickly shifted
towards disease control, though, following Dr. John Snow's
monumental publication on epidemiology within the same year,
(2,4,5)
     England shortly organized a series of Royal Committees
(6,7,8) charged with the study.of problems relating to sewage
disposal and treatment.  Their initial findings categorized
the existing state-of-the-art according to chemical precipi-
tation, filtration and irrigation, with the latter two pro-
cedures generally associated with land treatment.  While land
systems carried a traditional background extending several
centuries ,(4,9) some of the other available options were
rather curious.  One such precipitation procedure, the ABC
process, employed a bizarre mixture of alum, blood and clay.
(4,10)
     None of the available treatment mechanisms were, however,
                               61

-------
recognized as biologically-related systems.  Hence, Dr. Alex-
ander Mueller's demonstration in 1865 that sewage could be
purified by living organisms in a filtration column provided a
major revelation. (11)  Dr. Mueller, a prominent City Chemist
of Berlin, subsequently patented his biological purification
process several years later.  Unquestionably avant-garde,
neither the patent nor the fundamental concept attracted much
attention, though.
     In 1868, one of the Commission members, Sir Edward Frank-
land, began an epic study of filtration performance on raw
London sewage in laboratory columns packed with media ranging
from coarse gravel to peaty soil.  Using a twice daily dosing
pattern, Sir Frankland maintained successful filtration per-
formance for over four months. (11,12,13)  Although the
filter's treatment capability was solely credited to physical-
chemical means, the associated establishment of the inter-
mittent filtration concept had notably introduced a necessity
for resting or aeration periods between sewage applications.
     Based on these results, the Royal Commision began to
place considerable emphasis on the use of intermittent land
filtration. (14)  In 1871, J. Bailey-Denton initiated the first
full-scale operation at Merthyr Tydvil, Wales.,(14)  Success at
this facility, and others subsequently developed by Bailey-
Denton, soon promoted several engineers to apply Frankland's
concept.  (4,11,14)   Unfortunately, these engineers oftentimes
neglected critical factors such as soil permeability and/or the
necessity for intermittent dosing, such that failures became
commonplace.  And with subsequent documentation of 38 such
failures, (4,11,14)  technical interest in the intermittent
concept quickly faded.

1875 - 1900

     Following upon the singular work by Mueller over a decade
earlier, several researchers successively explored the microbi-
al aspect of sewage purification,  Schloesing and Miintz (15)
first demonstrated soil nitrification in 1877.  Five years
later, Warrington (16) confirmed that sterilized solutions lost
their nitrifying ability until inoculated by fresh soil.  And
in 1890, Winogradsky (17) succeeded in identifying Nitrosomonas
bacteria.  These pioneers were, however, still uncertain as to
the pragmatic application of these bacterial mechanisms to
effective treatment.
     Up to this point, Europe had dominated the developments in
wastewater treatment technology.  Within the United States,
                                62

-------
though, comparable concern for pollution control resulted in
the establishment of the Lawrence Experimental Station by the
Massachusetts State Board of Health. (4,18)  Organization of
the Lawrence facility was handled by Hiram F. Mills, a dis-
tinguished hydrologist, and Professors Sedwick and Drown from
the Massachusetts Institute of Technology. (4,19)  Under the
direction of Allen Hazen, the Lawrence group began a series of
filtration experiments in 1887 which were comparable to the
prior Frankland tests on intermittent dosing.  In this case,
however, the filters were significantly larger, at 1/200th
acre per unit.  While their results subsequently verified the
treatment potential afforded by an intermittent filtration
mechanism, the Lawrence group's first publication in 1890 pro-
vided a monumental analysis of the associated microbial ac-
tivity. (18)  Indeed, their findings truly furnished the hall-
mark demonstration that microorganisms carried within the
filter media could degrade sewage in an aerobic environment
facilitated by intermittent dosing.
     Given the success of the Lawrence experiments, biological
treatment systems rapidly expanded in terms of application and
sophistication.  Considerable controversy had arisen in the
1890's over patent rights obtained by Donald Cameron for septic
tanks, (4) such that most municipalities were anxious to find
suitable treatment alteratives.  Several full-scale inter-
mittent filtration systems were therefore constructed in the
New England area, most of which were successfully maintained
for several decades.
     In Europe, though, sanitary engineers were still hesitant
to accept the intermittent filtration concept.  This opinion
likely stemmed either from a lingering dissatisfaction with
the Frankland-era facilities, or because of the widespread
unsuitability of European soil. (4)  Instead, they chose to
intensify filtration rates using coarser media such as coke
breeze, gravel, burnt clay and coarse chalk.  Scott-Moncrief
(9) probably began the first such tests, using sewage perco-
lation through sequential trays of 1 inch diameter coke media.
In 1893, J. Corbett (20) also employed a serial filter scheme,
with an additional wooden trough to continuously distribute in-
fluent sewage across the bed.  And in the same year, F. Wallis
Stoddart (21) reported on the use of a coarse media filter
receiving a continuous, trickling flow.  Of these two latter
researchers, Corbett acknowledged the impetus and direction
provided by the previous Lawrence findings.  Stoddart, however,
insisted that his work stemmed from Frankland's principles and
that his continuously percolated units  were the first of their
                                63

-------
kind.  In either case, the trickling filter had been conceived.
     Another classic European option which developed at much
the same time was the contact bed.  Acting along the lines of
the Lawrence experiments, W. Santo Crimp and W. J. Dibdin
decided in 1891 to experiment with a dosing pattern which
flooded a coarse media filtration bed for 8 hours, followed by
16 hours in a drained state. (4,9)  Of the coarse media materi-
als tested on chemically-treated London sewage, Dibdin found
that the coke breeze provided satisfactory treatment, while
sand clogged extensively.  In subsequent tests, Dibdin experi-
mented with a double—contact approach, using primary and
secondary beds respectively containing successively smaller
media. (4)  The success of this operation quickly led to
several full-scale installations, all of which maintained the
cyclic fill, drain and react periods. And in their fifth report
(6), the Royal Commission provided extensive technical support
for the installation and operation of such contact beds.
     Dosing strategies for both the trickling filters and con-
tact bed systems received intensive study in the years immedi-
ately following their development.  For uniform loading of
intermittent filter units, Waring and Lowcock devised a sim-
plistic technique in 1892 based on an overlying fine gravel
layer to promote equivalent flow distribution. (4,14,23)  How-
ever, this procedure retarded desired bed aeration.  Perhaps as
a consequence, Waring also devised and patented a trickling
filter system which employed forced aeration. (4,14)
     Stoddart's (4,21) approach to flow distribution was that
of corrugated sheet-metal plates with symmetrical discharge
ports.  Although considered satisfactory, leveling of these
horizontal plates required tedious adjustment.  Corbett (4,20)
initially used slotted wooden troughs and then switched to a
variety of fixed-spray jets.  In 1896, Carfield (4,14) im-
proved the fixed distributor concept by adding a siphoned dos-
ing tank.  The siphon action insured an intermittent dosing
procedure which prevented localized flooding at the media.
     Rotary flow distributors were originally tested in 1889,
with additional refinement by Corbett in 1894. (20)  Two years
later, Whittaker and Bryant (4) introduced a rotary sprinkler
equipped with a pulsometer.  This latter addition not only
produced a pulsed, intermittent flow, but also warmed the in-
fluent sewage.  However, their model employed perforated pipe
distribution arms prone to clogging.  Rotary wooden troughs
were then introduced by Mather and Platt to avoid this plugging
problem. (4,14)
                               64

-------
1900 - 1925

     Given the classic technical advancements made by Hazen,
Stoddart, Corbett and Dibdin in the past quarter century, the
next twenty-five years could be viewed as an era of practical
application and refinement.  Of the available biological treat-
ment systems (i.e. intermittent filtration, trickling filters
and contact beds), it is interesting to note that each com-
prised a fixed-film process.  Rudimental experiments in sewage
aeration were underway at the time, but suspended growth
systems did not originate for several years. (4)
     Trickling filters were first introduced to the U.S. in
1901 at Madison, Wisconsin.. (4)  By 1910, several additions in
mid-west and eastern cities brought the total to ten. (9)
Monumental in size alone, the 31 acre Baltimore trickling
filter system is remarkably still in operation some seventy-
five years after its initial development. (24)
     Amongst these early U.S. trickling filter units, and for
several decades, fixed spray jets served as the norm for flow
distribution. (4)  Contemporary sewage treatment texts typical-
ly carried several pages devoted to spray jet design and in-
stallation.  (4,9,14)  In most cases, these distributions were
also equipped with siphon dosing tanks.  While rotating dis-
tributors were only randomly tested in the United States (i.e.
Springfield, MO in 1912 and Pontiac, MI in 1920),(4) European
trickling filter designs favored the rotary or travelling
sprinkler approach. (11)
     With, the advent of trickling filter applications, interest
in intermittent-filtration began to fade.  Experimentation
continued on both options at Lawrence, (19) demonstrating that
the higher loading rates provided by coarse media design could
significantly reduce the requisite land area.  Mathematical
modeling of these biological filters was also initiated in 1916
by Tatham. (25)  In using a mass-balance derivation based on
first-order kinetics, this study classically sought to define
the purification process according to precise chemical engi-
neering principles.
     As for contact bed design, several full—scale applications
were recorded. (4,26)  Although a few large scale units were
built in the United States, (4) contact beds did not receive
much interest outside Europe.  Because of the involved flooding
routine, anaerobic conditions tended to lower final effluent
quality. (26)  This circumstance, combined with frequent clog-
ging of the bed media by entrained sludge, (4,26) certainly
began to cast doubts on the usefulness of contact bed treat-
                               65

-------
meat.
     Recognizing the desirability of an aerobic biofilm, Dibdin
decided in 1904 to experiment with forced bed aeration. (27)
And to facilitate flushing solid matter from the bed, the
coarse media was replaced with slate slabs packed in horizon-
tal layers.  Operation of the modified unit still followed the
phased fill-and-draw routine. (4,28)  After 12 months of labor-
atory study, Dibdin successfully progressed to a full-scale
demonstration of his slate bed design at Devizes in 1905. (29)
However, in their fifth report, the Royal Commission indicated
that the slate bed approach should only be considered as a
primary sedimentation mechanism.  (6)
     Within  the U.S., Dibdin's slate bed technique drew
immediate interest.  Experimental testing was initiated in
Plainfield, New Jersey in 1905. (30)  Historically important
experimentation on slate bed treatment was also begun at
Lawrence under the direction of H. W. Clark and S, Gage. (19,
31)  In comparing aerated slate bed units and aerated bottles
containing algal suspensions, these investigators reported in
1913 that the bottles provided better treatment efficiency.
(31)  This variance was attributed to a failure by the pre-
viously scrapped slate plates to accumulate a suitable biofilm
during the short period of study.
     Shortly thereafter, Gilbert John Fowler, a British Pro-
fessor of Chemistry at Victoria University, visited the
Lawrence labs and witnessed these same experiments. (31)  Upon
returning to England, Dr. Fowler's students Edward Ardern and
William Lockett began the historic study of suspended growth
treatment.  In 1914, these two students, then published the
first account of an activated sludge process; sticking with the
accepted intermittent (i.e. fill-and-draw) pattern, but dis-
tinctively switching to a suspended biomass. (32)  Speaking on
behalf of his students. Fowler did acknowledge the contributing
and inspiration provided by Clark and Gage, referring to
Lawrence as "the Mecca of sewage purification..." (32)
     In much the same vein as Dibdin's slate bed, Dr. William
Owen Travis also sought to improve upon the contact bed pro-
cedure. (22)  As the local health officer in charge of a con-
tact filter at Hampton, England, Dr. Travis was quite familiar
with the problem of bed clogging. (4)  His solution, introduced
in 1904 as the Travis Hydrolytic or Colloider Tank, was es-
sentially configured as a multi-stage septic tank.  Successive-
ly divided into detritus, hydrolytic and finishing tanks, the
latter two zones contained wooden colloider baffles or laths
placed in a parallel array.  These baffles were intended to
                                 66

-------
attract fine particulates for subsequent degradation.  Only one
such plant was ever built, at Norwich, England in 1909, (4) The
construction at another Travis facility by the Emsher Drainage
District Board was discontinued after the death of the in-
volved design engineer, Wattenberg. (4)  His replacement, Karl
Imhoff, subsequently convinced the Board to switch to his
personal design, known thereafter as the Emscher or Imhoff,
Tank.  (9,14)
     As a footnote to this era, mention should also be made of
two unique patents obtained for rotating support media. (33,34)
The first, conceived by Weigand in 1900, (33)comprised a moving
cylinder with wooden slats.  Poujoulat's patent in 1916 (34)
employed agglomerated slag or porous brick fashioned as a
hollow cylinder and rotated about its horizontal axis.  Flow
distribution was provided using a perforated pipe placed over
the cylinder.  Although neither option attracted much attention
at the time, these designs could well be considered vintage
predecessors to rotating biological contactor technology.

1925 - 1950.

     Over the next twenty-five years, intermittent filtration
and contact bed systems were effectively discarded in favor of
trickling filter design.  Within the U.S., extensive efforts
were made to improve and upgrade trickling filter performance,
including the development and adoption of technical standards
for design loading, bed construction and system operation, (35)
High-rate designs, developed to increase hydraulic capacity,
were marketed by several companies,- including:  Lakeside
Engineering (Aero-filter), D0rr/Link-BeIt Comp. (Bio-filter)
and Infilco (Accelo-filter).  (35)  In most cases, fixed-spray
jets were also discarded in favor of rotating distributor
systems.
     Much of the popularity of these trickling filter units
could certainly be attributed to their relative simplicity,
ease of operation and cost-effective performance capabilities.
Activated sludge was still a somewhat innovative process, and
one which prompted considerable concern regarding its intensive
energy demand for aeration. (31,36,37)  Legal problems also
plagued the activated sludge process, with costly patent in-
fringement suits filed against several.major cities by Aeti-
.vated Sludge, Ltd. (38)  Many municipalities consequently turn-
ed away from suspended growth systems in favor of the more
conservative trickling filter option.
     There were, however, several tangential developments in
                                67

-------
fixed-film technology which deserve considerations.  The appli-
cation of one such option, the Hays process, actually rivaled
the installation of trickling filters for the period of 1930 to
1940.  (39)  Developed in 1930 by Clifford Hays, a chemist from
Waco, Texas, this procedure employed large asbestos-concrete
sheets vertically stacked with a 1" to 2" spacing.  (39)  This
design approach was physically analogous to the Dibdin slate
bed (although vertically arrayed, rather than horizontal) or
the Travis colloider system (with the added feature of a dif-
fused aeration system).  By 1942, there were 63 such units in
operation throughout the U.S., many of which were located at
military installations. (40)  However, the limited availability
of corrugated asbestos-concrete sheets during wartime con-
ditions necessitated the use of flat sheets.  (41)  Lacking
surface rigidity, these latter sheets frequently buckled and
collapsed, resulting in process failures which doomed its
future consideration.
     Another such resurrected concept was that of the Nidus
Rack. (42)  Developed by A.M. Buswell in 1929, the Nidus Rack
was intended to advance the Travis Colloider principle by
significantly increasing the surface area for colloid/parti-
culate attraction.  Numerous woven lattice units constructed of
veneer or basket wood were placed into a contact tank and
mechanically agitated to promote deposition into an underlying
settling compartment.  Buswell's article also mentions a number
of related studies incorporating straw and corncob filter
arrays. (42)
     Following along the research line established by Weigand
and Poujoulat,  a number of investigators independently studied
the use of rotating support media.  J. Doman (43) reported in
1929 on the development of a contact filter using partially
submerged rotating plates constructed from galvanized steel.
The schematic overview provided with.this report (43) bears
a striking resemblance to modern RBC designs.
     One further option on rotating media, the Biological
Wheel, was patented by A. T. Maltby shortly before 1930. (44)
The unit consisted of a series of paddle wheels partially sub-
merged in, and rotated by, sewage flowing through a surrounding
channel.  Biofilm attached to these wheels consequently rotated
in alternating fashion through the sewage and into the atmos-
phere .

1950 -PRESENT

     Mohlman's Sewage Works Journal (45) editorial entitled,
                               68

-------
"Revival of the Trickling Filter," provides an excellent
commentary on the mid-twentieth century state-of—the—art for
fixed-film systems.  Despite referencing the relative ad-
vantages of system reliability and economy, this editorial
acknowledged that trickling filters, "were almost relegated to
limbo." (45)  Indeed, over the next few years, conventional
trickling filter construction using rock media was unquestion-
ably surpassed by activated sludge,  Mohlman also provided a
timely reference to the related technologies recently developed
by Buswe11, Maltby, Doman and others.  In essence, he collect-
ively defended fixed-film treatment as a worthy alternative to
the rapidly advancing suspended-growth concept.
     At much the same time, significant developments were oc-
curing with the incorporation of plastic-based support media
Into various fixed-film treatment systems.  These synthesized
media forms offered several advantages over naturally available
materials particularly in terms of surface  contact area, void-
age fraction, packing density, and construction flexibility.
     Research and development on plastic media proceeded along
two distinct lines during the early 1950's.  In America,
bundled plastic units were being proposed and tested as inno-
vative packing for stationary filter applications.  (46)  In-
vestigators in Europe, though, began testing rotating plastic
discs in much the same manner as Doinan's rotating cast iron
system.  (47)  These latter researchers at Stuggart University,
West Germany, conducted extensive testing on wooden and plastic
discs, 1 meter in diameter. (47)  Further improvement by Popel
and Harttnan (48,49) led to the use of expanded polystyrene
media which then opened ChtSt-door for commercial application.
     By 1957, the J. Conrad Stengelin Company in Tuttlingen,
West Germany had begun manufacturing expanded polystyrene discs
2 and 3 meters in diameter for use in wastewater treatment
plants.  The first commercial installation went into operation
in 1960, (44,45) and soon thereafter the process began to
attract considerable interest through Europe.
     During the early 1960's, the research division of Allis
Chalmers Corporation also investigated the use of rotating
discs in various chemical processing applications.  Their disc
was called a two-phase contactor (TPC), and was tested for
applications of gas absorption and stripping,/ liquid-liquid
extraction, liquid-liquid heat transfer, and other mass and
energy transfer applications.  Eventually, the device was con-
sidered for oxygen transfer.  In the summer of 1965, three-
foot diameter metal discs were evaluated at the Jones Island
treatment plant in Milwaukee, Wisconsin.  These units were
                              69

-------
initially' employed for oxygen transfer in an extended aeration
process, and then tested without sludge recycle and with an
attached bioraass (i.e. as a biological contactor), [in retro-
spect, the Jones Island site was an ironic location, as it re-
presents the original application of activated sludge on a
large commercial basis], To confirm the favorable results of
these initial tests and to learn more about the treatment pro-
cess , laboratory tests were subsequently conducted using a
synthetic dairy waste and 3-foot diameter aluminum discs. (49)
     After learning of the European activities s Allis-Chalmers
reached a licensing agreement in 1968 _with the German manufact-
urer for production and sales distribution in the U.S.  The
treatment process was marketed under the trade name Bio-Disc.
The first commercial installation in the U.S. went into oper-
ation at a small cheese factory in 1969. (50)
     In 1970, Allis-Chalmers sold its rotating biological con-
tactor technology to Autotrol Corporation.  At that time, poly-
styrene discs were still not competitive with the activated
sludge process, primarily due to the high capital cost of the
polystyrene discs.  However, in 1972, Autotrol announced the
development of new rotating contactor media constructed from
corrugated sheets of polyethylene.  Until then, (51) the RBC
unit consisted of a series of parallel, flat 0.5 inch thick
expanded polystyrene sheets, each separated by a 0.75 inch
space.  The new arrangement used 1/16 inch thick polyethylene
sheets with a 1.2 inch space.
     Numerous terms are used throughout the wastewater treat-
ment literature to describe RBC's.  Among the terms in current
use are the following:  rotating biological contactors, rotat-
ing biological discs, rotating biological surfaces, RBS,. bio-
disks, bio-discs, biological rotating discs, rotating filters,
rotating biological filters, as well as trade names such as
Bio-Surf, Aero-Surf, Surfact, and BioSpiral.
     Several proprietary RBC options are currently available,
including the following variations on media construction:
parallel disc media attached perpendicular to the rotational
shift, media sheets spiral wound about the shaft, and segmented
media bundles placed as pie-shaped wedges about the shaft cir-
cumference.  Another recent development amongst the field of
rotating media units is that of providing supplemental aeration,
either for enhanced oxygen transport and/or to provide for
shaft rotation.  In one instance, a full-scale system employing
mechanical shaft rotation will shortly be retrofitted with such
aeration capabilities in an effort to enhance system perfor-
mance. (52) Numerous additional research, pilot-scale and full-
                               70

-------
scale (including commercial and industrial) investigations have
also been reported in the literature, notably including the
Proceedings of the First National Symposium/Workshop on Rotat-
ing Biological Contactor Technology, held at Champion, Penn-
sylvania, February 4-6, 1980,
     RBC's have a number of characteristics which make them an
attractive process for the design engineer.  They can provide
a high degree of treatment and, like trickling filters, have
lower energy and maintenance requirements than activated sludge
units.  RBC's require less land area than most other comparable
processes.  A large microbial population in the form of mix-
tures of filamentous and non-filamentous bacteria and fungi
grow on the contactor surface. (53)  A large active surface
area is obtained by the filamentous character of the growth.
RBC's can provide a highly nitrified effluent, since different
biological communities can be developed and maintained in
separate stages.  Because the biofilm is exposed to air rough-
ly 50% of the time, concentrated industrial wastes can be
treated without becoming anaerobic.  RBC's systems can be
designed to handle a wide range of flows, from less than 1 MGD
to over 100 MGD. (54)  No recycle is required.  The .sloughed
biomass generally has good settling characteristics and can
easily be separated from waste streams.
     Rotating biological contactors show high efficiency in
oxygen transfer.  Organic overloads are handled well due to the
large biomass on the discs. (51)  Since they involve attached
growth, they are less likely to fail through washout when con-
ditions adverse to biological growth occur.  No bulking, foam-
ing, or floating of sludge occurs to interfere with a plant's
overall efficiency.  Short circuiting in the RBC cannot occur,
due to the effect of staging in this plug flow system.  Shock
loads are dampened.  (55)
     In designing a plant, RBC's have advantages beyond their
low area requirements.  Most RBC's" .operate with nominal hy-
draulic head, so that pumping which otherwise might be required
may be avoided.  The change in head through the disc sections
is less than 1.0 ft.  Less excavation.is required for RBC's
than for activated sludge aeration tanks.  RBC's are versatile
both in the functions they perform and in the flexibility with
which they can be configured.  The discs can either be rotated
by mechanical drive (such as the Bio-Surf process) or use an
air drive mechanism (such as the Aero-Surf process) which has
fewer moving parts and uses less energy. (56)  For the mechani-
cal drive systems, a 25 ft by 12 ft diameter module which con-
tains 104,000 ft2 of total surface area, can be driven by a
                                71

-------
5-hp motor. (54)  The Bio-Surf process can be designed to pro-
duce an effluent BODs of 10 rag/1.  The composition of effluents
between 10 and 20 rag/1 BODs generally consists of approximately
1/3 soluble and 2/3 insoluble BOD5. (54)  However  the discs
are rotated, RBC technology use up to 50% less energy than
activated sludge units.  The low speed of the mechanical drive
units reduces maintenance requirements and prolongs their
lives.  Air driven RBC's allow the rotational speed to be ad-
justed by turning a few valves.
     The power requirements are low because the buoyancy of the
plastic discs offsets their weight, the weight of the biomass,
and the weight of their support structure so that the shaft
structure half submerged, has almost no resultant downward
force. (53)  The process is virtually absent of nuisances: no
clogging of the disc surface, no flies present, and no object-
ionable odors or noise.  A high treatment capacity exists be-
cause of the large microbial population which is contacted
xdLth wastewater and aerated.  BOD removal of 90% or more are
obtained on domestic or industrial wastewaters for retention
times of 60 minutes or less.  Toxic shock loads affect only the
more completely exposed organisms so recovery is rapid and
complete.  Cyclical fluctuations in wastewater flowrate are
absorbed with no loss in overall treatment efficiency.  The
time required in introducing waste to the discs to steady state
operation is usually one week.
     RBC's are simple to operate.  Nominal skill is required in
plant operation.  Since the sloughed biomass settles well and
can be removed more reliably than solids from activated sludge
tanks, clarifier design and operation is far less critical in
1BC installations.
     The RBC process also lends itself well to upgrading exist-
ing treatment facilities.  Because of its modular construction,
low head loss, and shallow excavation, it can be installed to
follow existing primary treatment plants.  Reliable winter per-
formance is obtained when the discs are sheltered by a modest
enclosure.
     RBC technology is not without its share of problems, how-
ever, the structural integrity of RBC units is untested by
time.  Plastic media has torn loose from its drive shaft in one
instance. (51)  Tie rods can loosen and cause uneven rotation
and need for realignment.  Oil leaks from drive units are
common.  Friedman (57) has discussed some of the failure modes
for RBC's.  Failure can be defined as any situation where the
process does not effluent goals, or does so in an objectionable
manner.  Situations such as process instability to meet
                                72

-------
effluent BOD and/or ammonia standards, or production of solids
that won't settle or cannot be separated readily from the
carrier stream, or production of objectionable odors are ex-
amples of process failure modes.  Media separation, shaft,
bearing and mechanical drive train problems are precursors of
process failure.  The authors of this paper know of at least
15 process failures. (58)  The reasons for failure were: shaft
failure, bearing failures, plastic weld failure, structural
support failure, steel shaft failure, and failure of the media.
Smith and Bandy (51) point out that although maintenance costs
are cited as an advantage, the costs are proportional to plant
capacity, exhibiting none of the economies of scale observed
with other non-modular technologies.  Area requirements are
also proportional to plant capacity.  Mechanically driven RBC's
are not able to vary the rotational speed easily; each drive
unit must be modified.
     Enclosures are necessary where low air and wastewater
temperatures occur to achieve acceptable performance.  RBC
systems must ordinarily be protected by a roof since heavy
rains may strip off the slime growth and hail may damage the
plastic discs. (59)  In northern climates, an enclosed heated
building may be necessary to prevent freezing during the
winter.  Provision for enclosures increases an RBC instal-
lation's initial cost, which is a disadvantage.  However, pro-
tected RBC's probably operate more stably in winter.
     With inadequate grit and primary solids removal, suspended
solids may accumulate in RBC reactors, resulting in lower
process efficiency and possible foul odors.  This can be avoid-
ed by providing adequate primary treatment.  The RBC operation
is subject to influent fluctuations which upset other
processes; although RBC's handle organic and hydraulic shock
loadings comparatively well, but with some loss in process
efficiency.  Toxic substances can cause a catastrophic loss of
biomass from the discs, although recovery is more rapid than
that of trickling filters under similar toxic loadings.  Ex-
tremes of pH have an adverse effect on RBC performance.  Over-
loadings on the first stage of RBC'c can cause an odor problem
and loss of efficiency.
     The conclusions on the advantages and disadvantages of the
RBC process are varied.  Antonie and Hynek (60) concluded the
RBC processes are stable, versatile, and competitive with
activated sludge.  Their studies included a wide variety of
municipal and industrial wastewaters.  Thomas and Koehrsen (61)
worked with distillery wastewaters, concluding that the acti-
vated sludge process was more stable when subjected to shock
                               73

-------
loads, provided better removals, and was less expensive on both
capital outlay and annual cost basis.  Some disadvantages
charged to the RBC provess will probably, disappear as the
technology matures.  Controversy exists regarding design
criteria, matrix design, surface-to-volume ratio for the re-
action chambers, optimum rotational speeds, appropriate scale
up procedures, recirculation requirements, and media config-
uration.  Antonie (48) further compares the rotating biological
contactor with the trickling filter process.  Further opera-
tional experience, additional research, and symposia such as
this one can be expected to remedy these shortcomings.
     At much the same time (i.e. early 1950's) that the West
German researchers began exploring plastic RBC's, American
investigators at Dow Chemical Company were initiating their
experiments with the production and use at plastic packing
media. (46)  Two initial plastic units were devised at Dow
including a modified  'berl-saddle1 (tradetnarked as Dowpac FN-
90) and bundled arrays of nested, corrugated sheets (trade-
marked as Dowpac HCS). (46)  Dow subsequently reassigned the
Dowpac term, substituting it with 'Surfpac.' Further detailed
review of the genealogy for these synthetic media is provided
in the following paper by Bryan. (62)
     Pilot-scale tests were conducted on both Dow packing
materials using various types of industrial wastes.  Both per-
formed acceptably well, but future emphasis was given to the
bundled form (i.e. Dowpac HCS) because of its perceived cost-
effectiveness and operational flexibility.  This material was
designed to distribute falling liquid wastes in thin films over
large surface areas so that maximum efficiency of contact with
aerobic micro-organisms was attained.  It provided a high per-
centage of void space for unimpeded draft circulation and
waste flow.  It provided large surface area adherence of bio-
logical slimes.  The material produced by Dow Chemical Company
consisted of individual sheets of polystyrene or Saran plastic
material,  (63,64) corrugated in two directions, having di-
mensions of 3 ft by 1.75 ft.  The individual sheets were typi-
cally shipped stacked in bundles, and then assembled into
structurally self-supporting modules at the point of use.  In
assembly, the sheets provided approximately 1 inch of free
space.  These modules were laid in the filter structure in a
layered grid pattern to provide good distribution of flow of
liquid, and to assist in structural stability.  Void space
within the assembled filter bed was about 94 percent.  As-
sembled weight of the individual modules was 4 to 6 lb/ft3.
This enables the modules to be stacked to depths of 30 to 40
                                74

-------
feet, conserving the use of land space.
     Because of the variable character of.different wastes, Dow
developed two types of plastic, each suitable for certain waste
streams:
     1*  Dowpac 10, which has good resistance to alkalies,
         salts, dilute mineral acids, and water, and is stated
         not to be suitable for some hydrocarbons, ketones,
         oxidizing acids, vegetable fats, and oils.
     2.  Saran, originally known as Dowpac 20, which is ex-
         tremely chemically resistant to all common acids and
         alkalies, with the exception of strong ammonium
         hydroxide.  It is suitable for most alcohols, esters,
         ketones, nitroparaffins, benzene, xylene, and toluene
         which diffuse slowly through the interstices between
         the modules, and have little effect on the material
         itself.
The sheets of Dowpac 10 are assembled with a solvent adhesive
supplied by the manufacturer.  Dowpac 20 is heat welded by
special assembly machines supplied by Dow.  In estimating the
cost of plastic media for trickling filters, the cost of
assembly must be included.  The use of heat welding caused some
modules to literally go up in smoke, which was a common failure
(64).
     Because of the light weight of this new material and its
available void space, the development of small diameter towers
with great height has occurred.  This has incorporated im-
portant savings in the use of the filter since it materially
reduced the amount of underdrain required.  The enclosing
structure for the trickling filter may be made of aluminum or
other light metal or wood, since no structural containment
walls are necessary.  In place of the vitrified underdrain tile
used in ordinary trickling filters, these under drains may be
made of pressure-treated lumber concrete partition blocks, sub-
way grating, etc. (63)  Since the assembled modules are rec-
tangular in shape, to avoid expensive cutting and shaping of
the material, the tower structures are  usually rectangular or
hexagonal in shape.
     The advantages of this lightweight and resistance sub-
stance generated the interest of other manufacturers.  Since
that time, similar plastic materials have been developed.  ICI
offers a polyvinyl chloride (PVC) packing named Flocor, which
was formerly available from Ethyl Corporation.  This was de-
veloped in England by the Imperial Chemical Industries, Ltd,
and consists of flat and corrugated sheets bonded into a module
2 feet in width and depth, and 4 feet in length.  The configu-

-------
ration of the sheets is a patented feature.  It offers such
advantages as large practical surface area with the lowest
bulk density.
     Another development in the field is the use of a poly-
vinyl chloride plastic called Koroseal developed in 1963, pro-
duced by B.F. Goodrich Industrial Products Company. (63,64)
The material is shipped, packaged, and assembled into modules
in the field.   The most outstanding demonstration of the use
of this material was at the Rome, Georgia mill of a manu-
facturer of kraft paper. (63)  This filter handles a flow of
16 MGD and is 80 feet in diameter, with a medium depth of 20
feet, and a total medium volume of 100,000 ft3.  The material
is supported on epoxy-coated steel gratings in the tower, which
has concrete block walls, with a total height of 30 ft.  To fit
the rectangular module shape, the tower is octagonal in shape.
B. F. Goodrich next changed to a lower density medium (4) but
contained less surface area (27 ft2/ft3).  This material,
derived from polyvinylidine chloride, required thicker sheets.
The sine wave corrugations had a wave length of 4 inches and
an amplitude of 2 inches.  This compares with the Koroseal,
having 37 ft2/ft3, of a sine wave corrugation with wavelength
3 inches and 1.5 inch amplitude.  A 1.5 inch amplitude is
generally the smallest amplitude put into use commercially,
otherwise bridging and plugging problems occur,especially for
high BOD wastewaters.  B. F. Goodrich (4) has developed a cool-
ing tower media, which can be used for nitrification-denitri-
fication operations.  This material has a surface area to
volume ratio of 44 ft2/ft3.  The corrugations are of wavelength
1.5 inches and amplitude 1 inch.  Another recent development
was Vinyl Core.(65)
     An additional development in the plastic line was provided
by American-Standard, New York.  A cellulose-fiber sheet im-
pregnated with plastic resin was made in a honeycomb design and
was suitable for stacking in a column.  Other varieties of
plastic material for trickling filters are offered by Tex-Vit
Company at Texas and Norton Chemical Process Products Division.
(66)  The structural engineering aspects of the plastic media
has been addressed by Mabbott.(67)
                                 76

-------
               Table I.  Comparison of Plastic
           and other Trickling Filter Media (63)
  Source             Brand     Density  Surface area,    Void
                      Name     lb/ft3      ft2/ft3     Space
Dow Chemical Co,   Surfpac       3.6         25          94

B. F. Goodrich     Koroseal    2.7-3.5       40          94

ICI                Flocor        4,06        —          95
                                                         74 9
Raschig Rings        —         30.3         22.7

Blast furnace        —         68.0         20          49
   slag

Stone, granite       —         90.5         30          45
                                 77

-------
         Table II.  Available Synthetic Media (48)
Supplier Trade Name
Envirotech Corp.
Brisbane, CA
B.F. Goodrich
Marietta, OH
1CI
Great Britain
Neptune-Micro floe
Corvallis, OR
Koch Eng. Co.
New York, NY
Norton Chemical Co.
Akron, OH
Institute de
Reserche Chimique
Applique, France
Surfpac
Koroseal
Vinyl Core
Flocor

Del-PakC
Flexirings
Actifil
Cloisonyle
Specific
Construction Surface Area
ft2/ft3
Flat and Corru- 27
gated PVC sheets
Flat and Corru- 30,5
gated PVC sheets
Flat and Corru- 29
gated PVC sheets
Horizontal wood- 14
en slats
Plastic pall 28
rings
Plastic pall 29
rings
PVC tubes 68.5
 formerly available from Dow Chemical Co.,  Midland,  MI


 Formerly available from Ethyl Corp., Baton Rouge, LA

c
 Formerly available from Del-Pak Corp.,  Corvallis, OR
                                78

-------
     Table I illustrates the weight and surface area advantages
of these synthetic materials.  Diversity within the current
market of proprietery plastic packing media is demonstrated by
Table II.
     The primary merits associated with trickling filters stem
from their simplicity low operating cost and ease of operation,
which makes them ideal for remote sites or small communities
(68).  Because large masses of organisms .must be present to
achieve high quality effluents, they possess substantial
reserve capacities making them robust and tolerant to changes
in the influent.  The dense nature of the raicrobial film which
slough from the media produces sludges of relatively constant
character which can be removed by sedimentation.  Trickling
filters have an ability to survive shock loads of toxic wastes
(69) due to the relatively short retention time of the waste-
water in the reactor (70) and/or because only organisms on the
surface may be killed.  If a shock load of long duration is
applied or of a type which will be adsorbed onto the biofilm,
then the trickling filter can be severaly affected (71,72).
     Problems of clogging by excess biomass have been experi-
enced when using a trickling filter, due to having too small an
interstitial volume within the stones.  The clogged areas be-
come anaerobic, generating objectionable odors, and are diffi-
cult to clear once clogged.  Filter flies often breed in a
trickling filter to cause a further nuisance. The major opera-
tional problems of trickling filters are associated with cold
weather operation, producing excessive cooling of the waste-
water and ice formation on the surface of the stones.  Ef-
ficiency in high rate filters is reduced with decreased
temperature by approximately 30 percent per 10°C.  Freezing
may cause partial plugging of the filter medium and resulting
over load in the open area.  In northern climates, fiberglass
covers or windbreaks have been employed to prevent ice for-
mation.  Covers also help contain odors which may be produced
in the filter.
     The main reason for the gradual loss of popularity of the
trickling filter is the limited degree of treatment achieve-
able.  Some of the largest plants have been built in recent
years, but the use of the trickling filter is steadily de-
creasing, due to its inability to consistently achieve high
degrees of soluble BOD removal.  The short wastewater retention
time limits the soluble BOD removal to the extent that it can-
not meet the levels of treatment possible in an activated
sludge system with a much longer retention time.  With effluent
discharge  requirements becoming more stringent, the trickling
                                79

-------
filter could no longer compete economically with the activated
sludge process.  The popularity of the trickling filter has
also lost some of its popularity in favor of the rotating bio-
logical contactor,  (49,73)
     Generally operated as aerobic systems, these latter packed
bed units typically receive a trickling flow which facilitates
desired tower ventilation.  Submerged contact has been recently
tested, though, both for aerobic and anaerobic treatment,
Tunick et. al (74) and Mines and Weeter (75) have accordingly
.reported on the behaviour of upflow anaerobic contact systems
packed with selected media materials.  A down—flow submerged
contact process has also been marketed by Cytox (76), incor-
porating a parallel array of vertically stacked plastic sheets.
Continuous fluid recycle within the vessel is directed towards
a splash pad above the tank which then promotes oxygen trans-
port.  Aside from this latter aeration mechanism, the Cytox
system could well be considered a resurrected Hays process.
Another option for submerged media will be presented by a
subsequent author, Li and Whang (77),  This unique approach
employs a synthetic ribbon media design which is then unfurled
and weighted to maintain extension.
                               80

-------
SUMMARY

     This paper has described the important historical develop-
ments of fixed-film wastewater treatment systems.  Beginning
in the 1860's x^ith filtration columns, various methodologies
have been developed for wastewater treatment.  This paper
addressed the development of such fixed-film systems like
trickling filters, intermittent filtration, contact beds,
hydrolytic tanks, and rotating biological contactors.  This
paper can not possibly include all the relevant references on
fixed-film processes.  Rather, the goal of this paper is to
highlight the technological advances which have occurred within
the field,  Fluidized bed systems have not been included in
this discussion.  They were intentionally omitted since they
are semi-suspended growth cum fixed growth systems.  Figure 1
highlights the important chronological developments of fixed-
film wastewater treatment systems.  This figure provides a
quick synopsis of the involved genealogy described in this
paper.
     With the resurgence of interest in fixed-film applications,
these processes are indeed consistent with the current federal
policy regarding "trickle down theory."  (78)
                               81

-------
                                   CHRONOLOGICAL PROFILE OF FIXED-FILM
   1860
1870
1880
1890     1900     1910     1920
Figure 1.  Chronological Development
of Fixed-Film Wastewater Treatment
Systems.
                             82

-------
WASTEWATER TREATMENT SYSTEMS
    19*30
1940
1950
1960
1970     1980
           	•— Acltwatwd Sfadgo I
                              83

-------
REFERENCES

 1.  Dickens, C., A Tale of Two Cities, Oxford University
     Press, London (1949).
 2.  Charnock, A., "When Beer Was Safer Than'Water."  New
     Civil Engineer, 18-19, 21/28 December (1978),
 3.  ASCE Committee on History and Heritage, Pure and Whole-
     some, Amer, Soc. Civil Engin. (1982).
 4.  Metcalf, L, and Eddy, H, P.,•American Sewerage Practice,
     Vol.Ill - Disposal of Sewage, McGraw-Hill Book Company,
     Inc., New York (1916).
 5,  Chanlett, E. T., Environmental Protection.  McGraw-Hill
     Book Company, New York, NY (1979).
 6.  Thompson, J. T., "The Fifth Report of the Royal Commission
     on Sewage Disposal."  Jour. Soc. Chemical Industry,  _33 (2)
     72-76 (1909) .
 7.  Calvert, H, T., "The Eighth Report of the Royal Commission
     on Sewage Disposal."  Jour. Soc. Chemical Industry,  3_2 (6),
     265-274 (1913).
 8.  Johnston, J. H., "Final Report of Royal Commission on
     Treating and Disposal of Sewage."  Jour.   Soc. Chemical
     Industry, 34, 507-508 (1915) .
 9.  Kinnicut, L. P., Winslow, C. E. A., and Pratt, R. W.,
     Sewage Disposal, John Wiley and Sons, Inc., London (1919).
10.  Anonymous, "Sewage Treatment and Disposal of Town Refuse."
     Engineering, 30, 519 (1884).
11.  Pearse, L., Modern Sewage Disposal, Lancaster Press, Inc.
     Lancaster, PA (1938).
12.  Kiersted, ¥., Sewage Disposal, John Wiley and Sons,  New
     York, NY (1894)
13.  Hutchinson, C. C., "On the Disposal of Sewage Sludge."
     Jour. Soc. Chemical Industry, _3, 41-51 (1889).
14,  Buswell, A. M., "The Chemistry ofWater and SewageTreat-
     ment, Chemical Catalog Company, Inc., New York, NY (1928).
15.  Schloesing, T. and Muntz, A., "Sur la Nitrification  Par
     les Ferments Organises,"  C. R. L'Acadamie Des Sciences,
     84, 301 (1977).
16.  Warrington, R., "On Nitrification,"  Jour. Chemical  So-
     ciety, 59_, 484 (1891) .
17.  Winogradsky, S., "Sur les Organismes de la Nitrification."
     C. R. L'AcadamieDes Science, 110, 1013 (1890).
18.  Hazen, A., "(Abstract)  The Lawrence Experiments on the
     Purification of Sewage in 1890 and 1891."  Engineering
     News-Record, 28, 559-560  (1892).
                              84

-------
19.  Clark, H. W., "Past and Present Developments in Sewage
     Disposal and Purification."  Sewage Works Journal,  2_,  561—
     571 (1930),
20.,  Corbett, J., "A Dozen Years of Sewage Purification
     Experiments  on a Large Scale at Salford, England."   Engi-
     neering News-Record, 49, 191-192 (1903).
21.  Stoddart, F. W., "Nitrification and the  Absorption
     Theory."  Jour. Soc.Chemical Industry,  30.  236-237,
     (1911).
22.  Folwell, A.  P., Sewerage, John Wiley and Sons, Inc., New
     York, NY  (1918).  .     •
23,  Lowcock, S.  R., "Experiments on the Filtration of Sewage."
     Jour. Inst.  Civi1 Engin., 115, 297 (1893).
24.  Kratz, H,, "Operation of Trickling Filters at Baltimore."
     Sewage Works Journal. _4, 135-141, (1932).
25.  Tatham, G. T. P., "Some Theoretical Considerations  Bearing
    • on the Performance of Biological Sewage  Purification
     Plant."  Jour. Soc. Chemical Industry, 35, 711-715, (1916).
26.  Hurley, J.,  "An English View on Contact  Aeration."   Engi-
     neering News-Record, 130, 895 (1943).
27.  Reefer, C. E., Sewage-Treatment Works, McGraw-Hill  Book
     Company, Inc., New York, NY (1940).
28.  Babbitt, H.  E., Sewerage andSewage Treatment, 3rd
     edition, John Wiley and Sons, Inc., New  York, NY (1928).
29.  Dibdin, W. J., "The Slate Bed Treatment  of Sewage." Jour.
     Soc. Chemical Industry, 32_, 55-61 (1913).
30.  Lanphear, R. S.s "Sewage Purification Results at Plain-
     field During 1910-1911."  Engineering Record, 64_, 29
     (1911).
31,  Martin, A. J., The Activated Sludge Process, MacDonald
     and Evans, London (1927).
32,  Ardern, E. and Lockett, W. T., "Experiments  on the  Oxid-
     ation of Sewage Without the Aid of Filters,"  Jour. Soc.
     Chemical Industry, 33, 523-539 (1914).
33.  Weigand, Ph., "Varfahren zur Biologischen Reinigung von
     Abwassern,"  German Patent No. 135755, 1900.
34.  Poujoulat, G., "Apparatus for the Purification of Sewage."
     French Patent No. 478,669;  11 May 1915, Jour. Soc. Chem-
     ical Industry, 3_5, 1129 (1916).
35.  Hardenbergh, W. A., Sewerage and Sewage  Treatment,  Inter-
     national High Rate Filters Textbook Company, Scranton,
     PA (1936) .
36.  Ardern, E.,  "A Resume of the Present Position of the Acti-
     vated Sludge Process of Sewage Purification."  Jour. Soc.
     Chemical Industry, 36, 822-830  (1917).
                                85

-------
37.  Hammond, G. T., "Some Remarks on the Biological Treat-
     ment of Sewage, With Especial Reference to the Activated
     Sludge Method."  Jour. Ind. Engin. Chemistry.,  9_, 399-403,
     (1917).
38.  Anonymous, "The Activated Sludge Patent."  Engineering
     News-Record, 113, 507 (1934).
39.  Griffith, L, B., "Contact Aeration for Sewage"Treatment."
     Engineering News-Record, 130, 138-142 (1943),
40.  Kessler, L. H, and Norgaard, J.  T., "Operation of Army
     Sewage Treatment Plants."  Sewage Works Journal, JL4,  1228-
     1261 (1942).
41.  Otts, L. S»,Personal Communication, University of Mary-
     land, College Park, MD (1982).
42.  Buswell, A. M. and Pearson, E. L., "The Nidus  (Nest)  Rack,
     A Modern Development of the Travis Colloider."  .Sewage
     Works Journal, _!, 187-195 (1929).
43.  Doraan, J», "Results of Operation of Experimental Contact
     Filter with Partially Submerged Rotating Plates."  Sewage
     Works Journal,  I, 555-560 (1929).
44.  Allen, K., "The Biological Wheel."  SewageWorks Journal,
     I, 560 (1929).
45.  Mohlmann, P. W., "Editorial: Revival of the Trickling
     Filter."  Sewage Works Journal,  8, 647-648 (1936).
46.  Bryan, E. H., "Molded Polystyrene Media for Trickling
     Filters."  Proceedings of the 10th Purdue Industrial
     Waste Conference, Purdue University, Engineering Bulletin,
     164-172 (1955).
47.  Daillaire, G., "Behind the Rapid Rise of the Rotating
     Biological Contactor."  Civil Englneerlng-ASCE, 49_,  70-73
     (1979).
48.  Antonie, R. L., Fixed Biological Surfaces -_ Wastewater
     Treatment, Chemical Rubber Company Press, Cleveland,  OH
     1976.
49.  Antonie, R. L., Private Communication, Autotrol Corpora-
     tion, Milwaukee, WI, 1982.
50.  Birks, C. W., and R. J. Hynek, "Treatment of Cheese Pro-
     cessing Wastes by the Bio-Disc Process," Proceedings  of
     the 26th Purdue Industrial Waste Conference, 26:89-105,
     Purdue University, West Lafayette, IN (1971).
51.  Smith, E. D., and J. T. Bandy, "A History of the Rotat-
     ing Contactor Process,"  Proceedings of the First Nat-
     ional Symposium/Workshop on Rotating Biological Contactor
     Technology, Vol. I, pp. 11-23, 1980.
52.  Canaday, J., Personal Communication, University of Mary-
     land, College Park, MD (1982).
                                86

-------
53.  Antonie, R.  and K.  Van Aacken,  "Rotating Discs  Fulfill
     Dual Wastewater Role,"  Water and Wastes Engineering,. 8
     (1): 37-38,  1971.
54.  RenefieId, L.  D,,  and C. W.  Randall,  Biological Process
     Design forWastewater Treatment,  Prentice Hall, Englewood
     Cliffs, NJ,  1980.
55.  Metcalf and  Eddy,  Inc., WastewaterEngineering:  Treat-
     ment , Disposal, Reuse, 2nd edition, McGraw-Hill, New  York,
     NY,  1979.
56.  Autotrol Wastewater Treatment Systems:   Design  Manual,
     Autotrol Corp., Bio-Systems  Div,, -Miluwakee,  WI (1978),
57.  Friedman, A. A., "RBC Design Considerations  for Industrial
     Wastewaters,"  Paper presented at the 74th Annual Meeting
     of the American Institute of Chemical Engineers, New
     Orleans, Louisiana, November 8-12, 1981.
58.  Anderson, B. L., Private Communication,  B. L. Anderson
     Co.. West Lafayette, IN, 1982.
59.  Steel,E. W., and T. J. McGhee,  Water  Supply  and Sewerage,
     Fifth edition, McGraw-Hill,  New York, NY, 1979.
60.  Antonie, R.  L., and R. J. Hynek,  "Operating  Experience
     with Bio Surf Process Treatment o.f Food  Processing
     Wastes,"  Proceedings of the 28th Purdue Industrial
     Waste Conference,  28: 849-860,  1973.
61.  Thomas, J. L., and L. G. .Koehrsen, "Activated Sludge-Bio
     Disc Treatment of Distillery Wastewater," EPA-6601 2-74-
     014, Superintendent of Documents, Washington, DC, 1974.
62.  Bryan, E. H.,  "Development of.Synthetic'Media for Bio-
     logical Treatment of Municipal  and Industrial Wastewaters."
     Proc. 1st Internat. Conf. Fixed-Film  Biol. Processes,
     Kings Island,  OH,  22 April (1982) [in press].
63.  Besselievre, E. B,, and M. Schwartz,  TheTreatment of
     Industrial Wastes,  2nd edition, McGraw-Hill,  New York,
     NY,  1976.
64.  Grieves, C.  R., "Dynamic and Steady State Models for  the
     Rotating Biological Disc Reactor," Ph.D. Dissertation,
     Clemson University, Clemson, SC,  1972.
65.  B. F. Goodrich, "Vinyl Core  and Koro-Z:  PVC  Filter Media
     for Wastewater Treatment," Bulletin FPD-81-EP-5027,
     B. F. Goodrich, Akron, OH, 1981.
66.  Norton Chemical Process Products  Division, "Actifil
     Media for Packed Biological  Reactors,"   Bulletin ACT-30,
     Norton Chemical Process Products  Division, Akron, OH.
67.  Mabbott, J.  W. , "Structural  Engineering  of Plastic Media
     for Wastewater Treatment by  Fixed-Film Reactors." Paper
     presented at the First International  Conference on Fixed-
                              87

-------
      Film Biological Processes, King's Island,  OH,  April  20-
      23, 1982.
68.   Chipperfield, P.N.J.,  M.  W.  Askew,  and J.  H.  Benton,
      "Multistage, Plastic-Media Treatment Plants,"  JWPCF, 44,
      1955-1967, 1972.
69.   Dow Chemical Co., Functional  Products and  Systems  Depart-
      ment, "A Literature Search and Critical Analysis of
      Biological Trickling Filter Studies," Vol.  I and II,
      Environmental Protection Agency Water Pollution Control
      Research Series, Report No. 17050 DDY 12/71, December
      1971.
70.   Cook, E. E., and L. P.  Herning, "Shock Load Attenuation,
      Trickling Filter,"  J.  Environ,. Engr. Div., ASCE,  104:
      461-469, 1978.
71.   R. E. McKinney, "Biological Treatment Systems  for  Refin-
      ery Wastes,"  JWPCF, 39: 346-359, 1967.
72.   Gillie, G. G., "The Effect of Toxic  Loads  on Effluent
      Purification Systems,"  WaterSA, 3_:  83-89,  1977.
73.   Gray, K. L., Private Communication,  B.F. Goodrich  Co.,
      Engineered Products Division, Akron, OH, 1982,.
74.   Tunick, M. H., Friedman, A. A., and  Bailey, D.G,,  "Treat-
      ment of Tannery Beamhouse  Waste with a Bench-Scale
      Anaerobic Reactor."  Proc. 13th Mid-Atlantic Ind.  Waste
      Conf., Ann Arbor Science Publishers, Inc.,  197-207 (1981),
75.   Hines, D. and Weeter, D. W.,  "Production of Ammonia  in a
      Packed Bed Anaerobic (ANFLOW) Bioreactor,"  Presented at
      the 1980 Annual AIChE Meeting, Chicago, IL, 19  November
      (1980) .
76.   CTX™ BIOX Treatment System,  Cytox Equipment Corp.,
      Dayton, OH, U.S. Patent No. 5.  3,966,608;  3,972,965;
      3,966,599.
77.   Li, C. T. and Whang, J. S., "Treatment of Domestic Sewage
      by Aquatic Ribbon System," Proc. 1st Internat. Conf.
      Fixed-Film Biol. Processes, King's Island,  OH,  22  April,
      1982 [in press].
78.   Hao, Oliver.  Personal  Communication, University of
      California-Berkeley, Berkeley, CA, 1982.

-------
   DEVELOPMENT OF SYNTHETIC MEDIA FOR BIOLOGICAL TREATMENT
           OF MUNICIPAL AND INDUSTRIAL WASTEWATERS
   Edward H. Bryan.  Division of Civil and Environmental
   Engineering, National Science Foundation, Washington, D.C.
ABSTRACT

     In Midland, Michigan, during June of 1954, a pilot-scale
experimental trickling filter ten-feet in diameter and ten-
feet deep began receiving the unsettled effluent from The Dow
Chemical Company's four conventional trickling filters, the
first of three stages for biological treatment of its strong
phenolic wastewater.  Half of the experimental unit was filled
with crushed blast furnace slag identical to that used in the
four large filters.  The other half was packed with a fabri-
cated plastic medium trademarked Dowpac HCS (since re-named
Surfpac).   With biological activity evident after eleven days,
the feed was changed to a synthetic wastewater containing pure
phenol and ammonium phosphate dissolved in Midland tapwater.
     A paper presenting results of the direct comparison
between performance of the two media in the experimental unit
was presented in May of 1955 at the Tenth Purdue Industrial
Wastes Conference and subsequently published in its Proceed-
ings.  From 1954 through 1960, an extensive research and
development program was conducted by The Dow Chemical Company
with cooperation of potential industrial users, municipalities,
consulting engineers, educators and government personnel at
local, state and federal levels.  During this period, results
from design, construction and/or operation of approximately
35 units provided guidance for decisions made during the
development period.
     This paper presents aspects of the critical early stages
in the development of plastic media, experiences with relevance
and potential applicability to current implementation of
                              89

-------
innovative and alternative solutions to problems of wastewater
treatment and management.  Previously unpublished results from
operation of several experimental units during the period from
1954 through 1958 are presented.  Included are data and results
from units that were packed to depths of 42 feet and which were
constructed to make intermediate depth-sampling possible.  One
unit, which was constructed to permit measurement of air-flow
through the packing, provided data confirming the previously
known but sparsely documented potential for stagnation in
trickling filters, a factor potentially affecting performance.

INTRODUCTION

     During 1953, The Dow Chemical Company's effort to produce
a tower packing for its own internal needs resulted in the
successful development of two types of media that could be
produced from synthetic plastics.  Then 'trademarked "Dowpac
FN-90" and "Dox-jpac HCS"*, efforts were initiated early in 1954
to investigate broadening their potential application to
cooling of water and biological treatment of wastewaters.
     The earliest public disclosure of Dow's pioneering work
in development of synthetic media for biological treatment of
municipal and industrial wastewaters was by Griess in a paper
presented at a meeting of the American Chemical Society in
1954 (1).  This paper contained initial, preliminary data from
operation of a pilot—scale experimental trickling filter, ten-
feet in diameter and ten-feet deep, half-filled with crushed
blast—furnace slag and the other half containing Dowpac HCS
(Figure 1).
     In contrast to Dowpac FN-90, a unique modification of the
conventional "berl-saddle" type of packing, which was
injection-molded; Dowpac HCS was vacuum-formed from flat
sheets of plastic.  The forming process produced corrugations
at right-angles to each other, and ribs that served to stiffen
the individual sheets, produce an average spacing of one-inch
between sheets when assembled into packs, and as positions of
additional contact for joining sheets into packs.

*The Dow trademark "Dowpac" was reassigned to other products
and replaced by "Surfpac" after the period of time during
which the author of this paper was responsible for conduct of
the research and development program described in this paper.
To avoid any misunderstanding, the trademark designation used
in this paper coincides with that in use when the work was
conducted that led to results cited.
                                90

-------
Figure -1.  Pilot-scale experimental trickling filter, 10-feet
           in diameter, 10-feet deep filled with'conventional
           crushed stone and Dowpac HCS media at The Dow
           Chemical Company in Midland, Michigan (1954).

     The unique design of Dowpac HCS permitted individual
sheets to "nest" in one position, but when alternate sheets
were rotated 180° in the plane of each sheet, the pack
expanded to produce a structure with the appearance of a
"honeycomb" when viewed from either end.  The combination of
edge—loading, rib—stiffening and composite-sheet action
produced modules of remarkable strength when subjected to ,-
compressive loading (Figure 2).         .   •
    'Experimental operation of the original pilot-scale unit
which began in June of 1954 continued until September of 1955.
With the exception of the initial eleven days during which
the unit was inoculated by passing through it the effluent
(unsettled) from the full-scale, conventionally packed Dow
phenolic wastewater treatment plant trickling filters, the
unit was operated until June of 1955 using a synthetic waste-
                            91

-------
water consisting of pure phenol and a proportional amount of
ammonium phosphate dissolved in City of Midland .tapwater
(treated Lake Huron water).  In June of 1955, the unit was
put on—line in parallel operation with the full-scale Dow
trickling filters and was used to evaluate other potential
packing shapes, materials and configurations as part of the
materials/fabrication component of the development program.
The range of phenol concentrations to which the pilot unit was
subjected during the initial phase of its operation (on pure
phenol) was from 10 to 536 mg/1.  Results of this pilot plant
study were presented by Bryan at the Tenth Purdue Industrial
Waste Conference in,May of 1955 (2) and were subsequently
also published in Industrial Wastes magazine (3).
Figure 2.  Stanley Mogelnicki, Supervisor of Waste Treatment
           Operations, The Dow Chemical Company standing on a
           module of Dowpac HCS illustrating its ability to
           support, weight of treatment plant personnel.
                          92;

-------
     A unit of identical dimensions to the initial pilot plant
was constructed and packed with Dowpac FN-90.  It was operated
in series with the original unit.  Despite favorable results,
it was concluded that Dowpac FN-90 was likely to be more
expensive and less likely to provide the flexibility in design
of full-scale units for biological treatment of wastewaters
when compared with Dowpac HCS.
     During the preparations for operating the initial pilot-
scale unit, it became evident that while polystyrene resin
used in fabrication of the media would be satisfactory for
process evaluation, it would not be satisfactory for the wide
range of conditions to .which full-scale units would be
subjected.  Test coupons of alternative materials were placed
on and buried within the full-scale Dow trickling filters.
While there was some variance in the length of time it took to
form the initial films, all plastics tested responded
favorably.  Process studies to assist in identifying and
characterizing the potential market were given priority over
further research on materials of fabrication.  The Dow Plastics
Technical Service Bulletin issued in October of 1955 (4)
announced availability of the two packings, suggested some
potential applications, listed their physical properties,
contained results of research to date, and contained a note of
caution regarding limitations of polystyrene with regard to
its chemical resistance.

TECHNICAL PROCESS EVALUATION PROGRAM

     During 1954, it became evident that patent protection
was likely for the unique designs of both packings but process
patent protection in conventional applications for biological
treatment of wastewaters was not.  Accordingly, the decision
was reached to utilize the technique of full public disclosure
and offers of cooperative assistance to industries, municip-
alities and consulting engineers who expressed interest in
assessing the potential applicability of the packings to meet
their needs as the principal component of the Dow development
strategy.
     The previously cited paper presented at the Purdue
Industrial Waste Conference was followed by technical•papers
which were essentially reports of progress on the Dow research
and development program at the Michigan (June 1955), West
Virginia (October 1955), Kansas (April 1956), Central States
(June 1956), and Pennsylvania (August 1956) Sewage and
Industrial Waste Association annual meetings, and at the Texas
A & M Short Schools in March of 1956 and 1957 by Bryan (5)(6).
                            93

-------
Kountz of the Pennsylvania State University referred to the
results of his Dow-funded studies using catalyzed sodium
sulfite to measure capacity for oxygen-transfer in a
"philosophical" paper on "total oxidation treatment" at the
Purdue (7) and Honey Harbour, Ontario Industrial Waste
Conferences in May and June of 1956, respectively.  In May of
1956, Towne and Becher of the U. S. Public Health Service's
Robert A. Taft Sanitary Engineering Center presented a brief
report on a Dowpac HCS research project that was in progress
at the Battle Creek, Michigan wastewater treatment plant to
the annual meeting of the Michigan Sewage and Industrial
Wastes Association meeting in Benton Harbor.
     All personnel who were cooperating with Dow in this
development program were encouraged to present their findings
in technical papers at conferences and meeting that were
appropriate to their content.  Stack presented results of a
pilot plant study conducted at the Union Carbide Chemicals
Company's South Charlestown, West Virginia plant at a
meeting of the Manufacturing Chemists Association's Air and
Water Pollution Abatement Committee's Joint Conference in
Washington, D. C. on April 4, 1957.  Trepanier (8) presented
results of his research that was conducted at the Ford Motor
Company's coke production plant in Dearborn, Michigan at a
conference in Pittsburgh, Pennsylvania on April 8, 1957.
Mills of the Naugatuck Chemicals Company in Elmira, Ontario
discussed his research at the Ontario Industrial Waste
Conference in Honey Harbour, Ontario on June 10, 1957.
     Results from this expanding external evaluation program
continued to be encouraging, equalling or exceeding the
original process-related expectations and confirming results
of a continued, parallel internal research and development
program.   While providing gradually increasing encouragement
for its process-potential, the program was equally effective
in disclosing weaknesses that would need to be addressed
before marketing Dowpac HCS.  Problems disclosed included
confirmation of the already well-documented property of
polystyrene to sustain combustion, its already well-estab-
lished solubility in gasoline, and its tendency to absorb some
organic compounds from wastewater which weakened its
structural integrity to the point where it would no longer
support the combined dead and live loads imposed on it in
packed towers.
     Increasing confidence in its technical promise was
instrumental in increasing attention to alternative plastics
for fabrication of Dowpac HCS early in 1956.  A number of
                             94

-------
approaches were tried centering around potential use of
polyethylenes and polyvinyl chloride resins.  By May of 1957,
two test packs fabricated from Saran were sent to the Great
Northern Oil Company's petroleum refinery in Pine Bend,
Minnesota for preliminary testing in their trickling filter
that had been originally packed with Dowpac HCS produced from
polystyrene.  In December of that same year, the unit was
completely re-packed with 13,300 cubic feet of Dowpace HCS
fabricated from Saran.  The design and preliminary operation
of this first, full-scale installation of a plastic-media
packed trickling filter was described by Anderegg (9) in 1959
and by Bryan (10) in 1962.
     During the initial, critical years while the product was
in Dow's "development stage", it was necessary to simultan-
eously excite the interest of potential users, establish and
maintain credibility regarding the relationship between its
promise and proven performance, and maintain Dow internal
interest to sustain.the research and development program.
Efforts to utilize technical forums in pursuit of public
disclosure sometimes led to misunderstandings of intent.  This
is evident from the following abstract of a letter received
from a consulting engineer in October of 1956:

  "Is Dowpac HCS available for purchase by my clients?  In the
   plant designs I am not commiting your company as to its
   effectiveness nor as to claims for its use...and...if your
   answers are negative then I am confused.  You never should
   have disclosed your information in technical society
   meetings and their journals if you did not want the
   engineering profession to be interested and to help you
   develop the ideas applications-.  It would seem that Dow
   takes the attitude of giving supreme and final approval to
   the engineering profession when Dow is ready.  This is
   neither a scientific approach nor enticing to engineers
   interested in process development"

In response, the engineer was informed that:

  "...Dowpac HCS is not presently available for purchase
   except for experimental use.  The magnitude of our existing
   program precludes duplication of experimental installations.
   All installations at the present time would be regarded by
   us as experimental...we have been pleased to participate
   in many technical programs by presenting 'progress reports'
   dealing with our work in this field.  We have never
                             95

-------
   expressed nor attempted to imply that Dowpac HCS is
   currently available as a product at such meetings.  Our
   desire to proceed to full-scale usage through the
   experimental pilot plant stage can hardly be considered
   by the profession as a 'neither scientific approach nor
   (one) enticing to engineers interested in product
   development'.  We are sorry that your impression of our
   effort to provide the profession with a new and perhaps
   better tool for the solution of waste treatment problems
   is summarized by the preceding extract from your letter."

     In response to another consulting engineer in October of
1957, it was necessary to emphasize again that:

  "Dowpac HCS is still considered by us to be a
   developmental product.  We feel that Dowpac HCS offers to
   the potential user a number of unique properties which
   will result under many circumstances in performance and
   economic advantages over conventional technology.  We have
   strongly urged the prospective user to recognize the
   essential uniqueness of his particular waste treatment
   problem attacking it through pilot plant experimentation."

     Even development of a product with such limited public
appeal as a packing for wastewater treatment processes had its
moments of difficulty with "the press".  The March 1957 issue
of Chemical Engineering contained a statement that:

  "Entry of a big chemical company like Dow, with its
   technical and promotional skills, should produce results
   in a field long dominated by sanitary engineers."

In the conventional wisdom of public relations that it doesn't
matter what is said about one in the media just as long as
one's name is spelled right, a decision was made to not
request a printed correction of this misinterpretation of the
"Dow" approach which was to work through rather than around
the traditional methods of obtaining product acceptability.
     A somewhat more intriguing error occurred in the article
by Egan and Sandlin in the August 1960 issue of Industrial
Wastes (11).  In their article, while correctly identifying
Mead-Core, a plastic packing being developed by the Mead
Corporation, Dowpac HCS and Polygrid (a plastic packing being
developed by the Fluor Corporation) were reversed as to their
identity in a set of four pictures and their captions.
                            96

-------
     Both Mead-Core and Polygrid bore historical ties to
Dowpac HCS and The Dow Chemical Company's development
strategy.  The Fluor Corporation has the unique distinction of
being the first customer for a full-scale installation of
plastic media in their design of the previously cited Great
Northern Oil Company refinery in Pine Bend, Minnesota.
Recognizing promise of the basic concept inherent in its
design, Fluor, in cooperation with Dow and Great Northern Oil
Company personnel worked together to resolve the technical
issues associated with that initial, full-scale installation
while simultaneously beginning its development of the Polygrid
packing, primarily for application in cooling of water.
Almost forgotten "heroes" in the risk that was inherent in
that initial installation were the personnel of the Minnesota
Department of Health who approved the initial plan and who
were patient during subsequent efforts to functionally
integrate the Dowpac HCS unit into routine operation.
     The Mead Corporation's interest which led to development
of Mead-Core was directly related to its comparative studies
of Dowpac HCS and Polygrid packings at pilot-scale (11).
Cawley, who reported subsequently on full-scale use of Mead-
Core at the Rome Kraft Company (12) himself conducted a
Dowpac HCS pilot-scale study while with the Rayonier Corp-
oration in Jessup, Georgia during the late 1950's.
     Entry of other potentially competitive plastic media into
the "arena" was an important factor in maintaining internal
interest within The Dow Chemical Company, where assessment of
its continued development program seemed to be subject to
re-evaluation every other week.  Equally important to the
emergence of competitive packings was the continued evidence
of technical superiority that Dowpac HCS was exhibiting over
conventional media,- emerging competitive shapes, and
alternative processes.
     During the period from May of 1955 to January of 1957,
an average of one pilot plant study was initiated each month
over a wide spectrum of potential applications, as summarized
in Table I.  The general arrangement was that The Dow Chemical
Company would .-provide the packing and technical assistance in
planning, conduct of the study, and evaluation of the results
in return for a technical report of performance.  While
emphasis was on the external effort during this period, a
complementary internal program was maintained and modestly
expanded.  By August of 1957, 28 pilot-plant studies had been
conducted or were underway and an additional 6 were at an
advanced stage in planning, design or construction.
                             97

-------
        Table I.  Dowpac Development  Program Summary
                     from  1954  -  1958
Date  of
Internal
Dow-
  Report

May 1955

November
  1955
May 1956
January
  1957
Number of Pilot Studies
Ongoing and/or Complete
      (Summation)
 Internal     External
    2

    4
                20
August
  1957
                21 plus
                plans
                for 6
Types  of Wastewater and/or
        Application
    (Items are additive)

Synthetic phenol, Cooling
water, Brine settling,
Construction prototype,
Semi-chemical boxboard,
Kraft  pulping

Domestic wastewater, Coke
oven,  De-inking, Glycol,
hydrolyzer, Sulfite oxi-
dation, Dehumidification
Ammonia removal, Corn steep-
Vegetable oil refinery, Oil
gas processing, Chlorinated
phenols, Water treatment to
remove carbon dioxide and
hydrogen sulfide, Contact
aeration, Milk waste, Sour-
water scrubber, Solids
flotation

Alternative materials for
media fabrication
     Shortly after initiating its initial Dowpac HCS and
conventional stone-packed unit, a construction prototype was
designed and constructed at the Dow plant in Midland, Michigan
(Figure 3).  Another pilot plant containing a packed depth of
42 feet was constructed and operated at the City of Midland,
Michigan Sewage Treatment Plant (Figure 4), initial results
of which were presented by Bryan at the Michigan Sewage and
Industrial Wastes Association meeting in June of 1955 and at
the Texas Water and Sewage Works Short School in March of
1956 (6),  Both units provided breadth to the development pro-
gram not possible by response to external interests.
                              98

-------
          S^S^S*^'-;'\V'••-,;'.•!; -^ji.rC1' r
          ">tj%St- ^^^ ••' *   ' ^ \     t  ' V  ,
Figure 3.  Dowpac HCS Construction Prototvpe, The Dow Chemical
           Company, Midland, Michigan  (1955).

     With the Dowpac HCS Construction Prototype, Handt (13)
observed an efficiency of phenol removal.of 96%.for the 20-
foot packed depth in comparison with 82% for the original unit
containing a packed depth of 10 feet at the same hydraulic and
organic loading rates.  Brelsford (14) continued to operate
this unit with the objective of determining the "protein-
value" of harvested slimes, concluding their protein-equiv-
alent based upon their organic nitrogen content was between
31.6 and 34.4 percent.  In a subsequent study, Froman (15)
found the unit to remove between 86.4 and 88.4 percent of the
acrylonitrile in a synthetic wastewater using ammonium
                              99

-------
Figure 4.  Dowpac HCS pilot plant at the City of Midland,
           Michigan Sewage Treatment Plant, containing a
           packed depth of 42 feet (1954).

phosphate as a supplemental source of nutrients, at loading
rates of 83 to 162 pounds of oxygen-equivalent per 1000 cubic
feet per day.  The data from this study was used by Roy F.
Weston, Inc. in the design of the full-scale plant for the
treatment of wastewater at The Dow Chemical Company's acrylic
fiber production facility near Williamsburg, Virginia which
was placed into operation during 1958 (16).
     The experimental operation of the pilot plant at the
City of Midland Sewage Treatment Plant (Figure 4) included
observations of air-flow by Heckeroth (17) and Greene (18).
Their data (Figure 5) provided clear evidence of the potential
                              100

-------
for reversal of air-flow through trickling filters and
therefore the potential for stagnation and its consequences
in affecting the performance characteristics as suggested by
Bryan (19.
   p
   D
   o
  •H
  J3
   3
  O
   C
  •H

   3
   O
  i-H
50
_40
                            Variation in Air Flow with
                              Temperature Difference
                           (Infl Air) - (Avg Water)°C.
                               Midland Pilot Plant
      Data from 6/10 -9/13
       1955 at Irregular
             Intervals
    -30
        -2
                   0  +1
       Temperature Difference in Degrees Centigrade

Figure 5.  Relationship between temperature difference (Air -
           Wastewater) and air-flow through the 42-foot
           Dowpac HCS experimental unit at the City of Midland,
           Michigan Sewage Treatment Plant.

     Greene conducted a four-week study in which solids from
the settling tank at the City of Midland, Michigan experimental
unit were returned to the Dowpac HCS tower as a "test" of the
"total oxidation" concept of Kountz (7).  Greene found that the
loss of solids over the settling—tank weir was approximately
equal to solids produced which were, in turn, produced in
direct proportion to the reduction in chemical oxygen demand of
the wastewater treated.  His brief, preliminary study of the
relationship between air-flow and performance suggested that
theories of trickling filter performance and consequent
"formulations" that ignore the effect that potential stagna-
tion may have on availability of oxygen to the biologically
active films may poorly represent the performance of actual
trickling filters.  These observations clearly indicated the
                             101

-------
 superiority of Dowpac HCS over conventional media in the
 freedom with which air could move under natural conditions
 and the relative ease with which forced-ventilation could be
 implemented in design,  construction and operation of full-
 scale units.
      Within the range of  organic and hydraulic  loadings  used
 in studies  with the City  of Midland unit,  its performance was
 found to be dependent only on the hydraulic rate of applic-
 ation.   Results of the  two rates most comparable to those
 used  in trickling filters studied by the National Research
 Council were compared with the empirical formula resulting
 from  those  studies and  found to be in essential agreement
 with  those  findings (Figure 6).
                                       50  100
                                  !  I  I  I I M
       Application Rate of BOD5_da -Lbs/1000ft3/Day
Figure 6.  Performance of the Dowpac HCS unit at the City of
           Midland Municipal Sewage Treatment Plant which
           contained a packed depth of 42 feet.  Hydraulic
           application rates for results compared to those of
           the National Research Council (1946) were 18 and
           36 million gallons per acre per day.
                            102

-------
     The concept of returning solids to the influent with the
packing "functioning as an aeration device for mixed liquors
in addition to supporting bacterial slimes" as suggested by
Bryan (5) was tested initially by Bauer (20) who found that
the strength of the City of Midland wastewater was insuffic-
ient to build—up enough activated sludge for a good test of
this concept.  In a subsequent study, Ellis (21) used acti-
vated sludge from the Dow general wastewater treatment plant
(10), whey from a local dairy and ammonium phosphate as a
source of supplemental nutrients in tests ranging from two to
nine hours in duration.  He found the oxidation rate to be
in a range of from 2.4 to 6.2 pounds per cubic foot per day
(Chemical Oxygen Demand).
     Since slimes had been chemically cleaned from'the
packing prior to his tests, the reduction in Chemical Oxygen
Demand was solely attributed to the packing acting as an
aerator.  However, Ellis felt those rates were "exaggerated"
by his assumptions in sampling, but after accounting for
potential error, he concluded that:

  "...removal rates of greater than 1,000 pounds of Chemical
   Oxygen Demand per day per 1000 cubic feet were obtained."

This rate, was in the-mid-range of those plotted by Bryan
(Figure 7) from data obtained by Kountz (22) using the
cobalt-catalyzed, sodium sulfite technique in studies he
conducted at the Pennsylvania State University.
     Late in 1955, while studies were in progress at the City
of Midland Sewage Treatment Plant, an opportunity arose to
conduct a similar study in Battle Creek, Michigan,  Following
some preliminary discussions between personnel of The Dow
Chemical Company, City of Battle Creek, and the firm of
Jones, Henry and Williams (consultants to Battle Creek), a
meeting was held in Battle Creek on January 4, 1956.  The
eleven persons present included personnel from the State of
Michigan Department of Health and the Water Resources Comm-  '
ission and the U. S. Public Health Service's Robert A. Taft
Sanitary Engineering Center in Cincinnati.  A decision was
reached to conduct a pilot-scale evaluation of Dowpac HCS at
Battle Creek in a unit analogous to the unit in operation at
the City of Midland.  Financial support, estimated at $10,000,
was agreed would be equally shared by the City of Battle
Creek, The Dow Chemical Company, General Foods and Kellogg
Corporations.  A Steering Committee was appointed to include
representation from all participants in the proposed study.
                             103

-------
                         Oxygen Transfer as Related
                          to Hydraulic Application
                         (using the cobalt-catalyzed
                           sodium—sulfite technique)
      Hydraulic Application Rate - Gallons/cubic ft/Day

Figure  7.  Relationship between oxygen  transfer rate and
           hydraulic application rate for Dowpac HCS using
           cobalt-catalyzed sodium sulfite  (Kountz data).

     On January 13, 1956 - only nine days after the initial
meeting at which the Battle Creek Study was formulated, the
pilot plant was placed into operation.  Activities during the
intervening nine days between the initial meeting and the
start of operation included construction of the pilot plant
(Figure 8), fabrication of a settling tank, construction of
a large BOD-incubator, augmentation of the City of Battle
Creek Treatment Plant's laboratory for conduct of Chemical
Oxygen Demand, and correlation of the 5-day Biochemical Oxygen
Demand and the Chemical Oxygen Demand for the City's primary
effluent.  This pilot unit was operated continuously through
April 28, 1956 while it was intensively studied.  It was on
"stand-by" operation until June 11, 1956 when it was oper-
ated at a low dosing rate to provide data for extending the
range of operation to include the the highest hydraulic
dosing rate then in general use for design of conventionally
packed trickling filters.   During the entire period of oper-
ation, the Steering Committee provided guidance to the study.
                             104

-------
 Figure 8.  Dowpac HCS pilot plant constructed at the City of
            Battle Creek.  The unit contained a packed-depth of
            42 feet with provision for intermediate sampling»

     Details regarding the "construction and operation of the
Battle Creek-pilot plant, guidance provided by the Steering
Committee, results of operation and their analysis were con-
tained in a Report of the Steering Committee authored by Becher
and Bryan (23) with a statistical .analysis by Busch.  Table II .
contains a summary of results.  An Appendix to the Report (23)
contains all observed data obtained during the reported study.
Stack (24) commended the Stee.ring Committee for; "accomplishing
an excellent study...the most,thorough study of trickling
filtration treatment of sewage that I have seen."
                              105

-------
               Table II.   Summary of Results From the Dowpac HCS  Pilot  Plant at Battle Creek,  Michigan (1956)
o
cr>
                                                                   Chemical Oxygen Demand
                                                             Temperature-
 Time  Hydraulic 5-Day Biochemical Oxygen Demand
Period   Rate     Infl C  lb/1000-  Efficiency-%  Inf1 C  lb/1000- Efficiency-%  (Degrees F)
(1956) (gpm/ft2)  (mg/1)  ft3/day   Tank    Lab*  (mg/1)  ft3/day  Tank    Lab*  Air  Water**
2/3-17   3.39      249      241     34.0    48.1   453      440    29.2    44.9  26.2  57.9
              2/17-
                3/14
                       1.63
222
                             83.2   56.0    71.9   389
156
                                                                   52.4
              3/20-    1.6 +
                       1,6(R)*** 24i
                                 227
         101     62.2

          53.0   76.0
                                            66.7   420

                                            80.3   397
176    49.3

 92.9  63.5
60.3  29.0  53.0


51.1  31.8  51.0

69.2  45.6  54.2
4/10-28  0.820

   ,/,,  Standby operation of unit while data were being analyzed - No data were obtained

6/11-27  0.318     218       19.9    -      86.7   393       35.8   -      72.2  75.2  66.5**
              Notes:  *Tower effluent  was given 60-minutes  of  quiescent  sedimentation in  a  laboratory
                      graduated cylinder.
                    **Average of influent  and  effluent.  Passage  through  unit  reduced temperature by 3.4°F.
                   ***1.6 gpm direct  (primary  effluent)  plus  1.6  gpm  recirculation  (unsettled  filter eff1).
                    **Influent temperature only,  effluent  temperature was not  obtained in  this period.
                      Tower depth was 42 feet.   In calculation  of loadings,  it was  assumed sewage applied
                      by a 3-foot diameter rotary distributor contacted all  packing in the 37-l/2"square
                      section.  No adjustment  was made  for packing-equivalent  of  sidewalls which provided
                      a maximum of an additional  5% surface area  in the packed tower.

-------
     As likewise determined from operation of the pilot .plant
in Midland, Michigan, the efficiency of operation at Battle
Creek was linearly and inversely proportional to the hydraulic
application rate (Figure 9).  However, with respect to removal
of oxygen demand, within the limits of hydraulic application
rates studied, removal of both biochemical and chemical
oxygen demand was linear (in two "regimes") proportional to
the hydraulic application rate.  The particular advantage of
Dowpac HCS as a "roughing" unit was obvious from this study as
it was from all other prior studies.
                    1.0
2.0
3.0
            Hydraulic Application Rate - gal/min/ft


Figure 9.  Efficiency and total removal of biochemical and
           chemical oxygen demand as affected by the hydraulic
           application rate of the Dowpac HCS pilot unit at
           Battle Creek, Michigan.
                              107

-------
     It was concluded that the Battle Creek pilot plant data
were best represented by the following empirical equation:
      R
                                      2 /•}
              0.0148 D1'748 10-°'127 Q
 where: R is the "fraction" of 5-day BOD remaining at depth D
        D is the depth of the Dowpac unit in feet
        Q is the hydraulic application rate in gallons/minute
          (Note - the unit was 9.77 square feet in area)

The exponent of "Q" in the above equation was noted to be in
accord with Rowland's theoretical development (25) and with the
results of studies of laboratory trickling filters conducted by
Bloodgood, Teletzke and Pohland (26).

SUMMARY

     Although the general principle of trickling filtration had
been, previously well established and prior attempts had been made
with little success to introduce synthetic media, the process by
which synthetic media fabricated from plastic resins were
developed was without precedent.  In 1960, Zwick and Benstock
(27), in a draft of their "Study Group Report on Water Pollution,"
attributed the origin of plastic media to an undocumented source -
a person who had suggested replacing conventional media with
wooden planks mounted in a box.  Correspondence in which they were
provided with a copy of the Battle Creek Report (28) resulted in
some modification of the draft to provide a more balanced and
accurate description of the origin of the concept and the role
of personnel from The Dow Chemical Company, the U.S. Public Health
Service, and others in its development.
     The period during which The Dow Chemical Company's effort
took place was one in which plastics were emerging to take the
place of other materials in applications that went beyond the
production of toys and novelties.   Its own internal needs were
the initiating cause for action taken by Dow in development of
plastic media.  The initial step is most accurately attributed
to R. S. Chamberlin, D, E. Lake and F. E. Dulmage of The Dow
Chemical Company who conceived the basic design of Dowpac FN-90
and related shapes.  Dowpac HCS was a product of the joint
efforts of D. E. Lake and Thomas J. Powers, Sr.   The distinction
of recognizing their potential for treatment of wastewaters
belongs to Powers who provided the initial context within which
                            108

-------
the development effort was initiated and nurtured.  His
seemingly unlimited capacity for seeking simpler and more
direct ways of solving problems was coupled with a gift of
almost infinite patience up to a point where action was both
necessary and wise...attributes which, in the complex process
of product innovation and development, are essential if not
indispensible to balance potential risk with reward.

ACKNOWLEDGMENTS

     Initial stages in the development of plastic media that
are described in this paper were under the joint responsibility
of James A. Struthers, Plastics Technical Service and Edward H.
Bryan, Waste Disposal, The Dow Chemical Company, Midland, Mich.
Other Dow personnel not specifically noted in the text or in
the References section who made significant contributions
during the early stages of development described in this paper
included:  E. E. Chamberlin, G. F. Dressell, John Hoy, William
C. Goggin, A. A. Gunkler, Frank H. Justin, Earl Kropscott, Paul
H. Lipke, Frank J. MacRae, Del H. Moeller, Stanley Mogelnicki,
W. L. Nelson and Gordon B. Thayer.
     Dr. Edward H. Bryan is currently Program Director, Water
Resources and Environmental Engineering in the Division of
Civil and Environmental Engineering, National Science Found-
ation, Washington, D.  C.  He was responsible for the technical
and process related research and development activity, subject
of this paper from 1954-58 while employed by The Dow Chemical
Company.  Responsibility for development of alternative
plastics and methods  of fabrication was initially that of
James A. Struthers who was ably succeeded in this portion of
the development program by Del H. Moeller, who assumed full
responsibility for the program in 1958.
     The content of this paper and any opinions expressed are
solely those of the author and do not reflect a position by
either The Dow Chemical Company or The National Science
Foundation.  The author expresses his appreciation to The Dow
Chemical Company for  permission to use photographs contained
in this paper.

REFERENCES

1. Griess, G. A., "Plastics in Plants Manufacturing Heavy
   Chemicals", Industrial and Engineering Chemistry, Vol 47,
   pp 1343-1349 (July 1955).
                            109

-------
 2. Bryan, Edward H., "Molded Polystyrene Media for Trickling
    Filters," Proceedings of the Tenth Purdue Industrial Waste
    Conference, pp 164-172  (May 1955).
 3. Bryan, Edward H., "Molded Polystyrene Media for Trickling
    Filters," Industrial Wastes, pp 80-84 (November-December
    1955).
 4. Dowpac FN-90 and Dowpac HCS, Bulletin of The Dow Chemical
    Company, Plastics Technical Service, 16 pp, (October 1955).
 5. Bryan, Edward H., "Dowpac Tower Packing for Gas-Liquid
    Contact Systems," Proceedings of the 38th Texas Water and
    Sewage Works Short School, pp 121-122 (March 1956).
 6. Bryan, Edward H., "The Role of Oxygen in Sewage Treatment,"
    Proceedings of the 39th Texas Water and Sewage Works Short
    School, pp 88-89 (March 1957).
 7. Kountz, R. Rupert, "Total Oxidation Treatment," Proceedings
    of the Eleventh Purdue Industrial Waste Conference, pp 157-
    159 (May 1956).
 8. Trepanier, Norman W., "Biological Treatment of By-product
    Coke Plant Phenolic Wastes," Blast Furnace, Coke Oven, and
    Raw Materials Conference, American Institute of Mining,
    Metallurgical and Petroleum Engineers, Discussion by:
    Edward H. Bryan, Charles Drake and Hayes H. Black, pp 204-
    210 (1957).
 9. Anderegg, Fred C. "Biological Disposal of Refinery Wastes,"
    Proceedings of the 14th Purdue Industrial Waste Conference,
    (May 1959).
10. Bryan, Edward H., "Two-Stage Biological Treatment - Indust-
    rial Experience," Proceedings of the Eleventh Southern
    Municipal and Industrial Waste Conference, pp 136-153
    (April 1962).
11. Egan,  John T. and McDewain Sandlin, "Evaluation of Plastic
    Trickling Filter Media," Industrial Wastes,. Vol 5, No 4,
    pp 71-77 (August 1960).
12. Cawley, William A., "Polyvinyl Chloride for Trickling
    Filter Media," Industrial Water & Wastes, pp 111-  (July-
    August 1962).
13. Handt, Paul R., "Progress Report on Evaluation of Dowpac
    HCS as used in Trickling Filters," Student Trainee Report,
    The Dow Chemical Company (1956).
14. Brelsford, Donald L., "Harvesting Protein-Rich Bacterial
    Material From a Dowpac HCS Type Biological Trickling
    Filter," Summer Technical Employee Report, The Dow Chemical
    Company (1956).
15. Froman, Charles 0., "Biological Oxidation of Acrylonitrile
    on Dowpac HCS Pilot Plant," Internal Report, The Dow
    Chemical Company (March 1957).
                              110

-------
16. Sadow, Ronald D., "Acrylonitrile and Zinc Wastes Treatment,"
    Parts 1 and 2, Industrial Water & Wastes, pp 42-45 and 66-70,
    (March-April 1961) and (May-June 1961), respectively.
17. Heckeroth, Earl T., "Progress Report on the Evaluation of
    Dowpac HCS as a Trickling Filter Medium," Student Trainee
    Report, The Dow Chemical Company (September 1955).
18. Greene, Robert E., "Progress Report on the Evaluation of
    Dowpac HCS Trickling Filter at the Midland Municipal Sewage
    Treatment Plant," Chemical Engineer Trainee Report, The Dow
    Chemical Company (September 1955).
19. Bryan, Edward H. and D. H. Moeller, "Aerobic Biological
    Oxidation Using Dowpac," in Advances in Biological Treat-
    ment , Proceedings of the Third Conference on Biological
    Waste Treatment, Manhattan College, edited by: W. W.
    Eckenfelder and Joseph McCabe, pp 341-346, Pergamon Press
    (1963).
20. Bauer, K. C., "Progress Report on Dowpac HCS'as a Trickling
    Filter Medium - Midland Sewage Plant," Summer Technical
    Employee Report, The Dow Chemical Company (1956).
21. Ellis, William J., "Progress Report on the Evaluation of
    Dowpac HCS as a Trickling Filter Medium," Student Trainee.
    Report, The Dow Chemical Company (1956).
22. Kountz, R. Rupert, "Oxygen Solution" Capacity of Wetted
    Dowpac HCS Towers," Report to The Dow Chemical Company
    (1956).
23. Becher, A. E. and Edward H. Bryan, "A Study of the Perform-
    ance of Dowpac HCS When Applied to the Treatment of Settled
    Sewage from the City of Battle Creek, Michigan," Report of
    the Project Steering Committee, Statistical Analysis by
    K. A. Busch, 149 pp, The Dow Chemical Company (June 1958).
24. Stack, Vernon T., Personal Communication (February 1959).
25. Howland, W. E., "Flow Over Porous Media as in a Trickling
    Filter," Proceedings of the 12th Purdue Industrial Wastes
    Conference, pp 435-465 (May 1958).
26. Bloodgood, Don E., G. H, Teletzke and F. G. Pohland,
    .Fundamental Hydraulic Principles of Trickling Filters,"
    Sewage and Industrial Wastes, Vol 31, pp 243-253 (1959).
27. Water Wasteland, Nader Task Force Report on Water Pollution,
    Edited by David R. Zwick and Marcy Benstock, Vol 2, pp XIX-
    28 and 29, Preliminary Draft, Center for Study of Respon-
    sive Law, Washington, D. C. (1971).
28. Bryan, Edward H., Personal Letter to David Zwick (July 1971).
29- Water Wasteland, Edited by David R. Zwick and Marcy Benstock,
    pp 381-382, Grossman Publishers (1971).
                             Ill

-------
     CURRENT STATUS AMD FUTURE TRENDS OF ROTATING BIOLOGICAL
                      CONTACTOR IN JAPAN
     Masayoshi Ishiguro.  Professor of Civil Engineering,
     Miyazaki University, Kirishima 1-1-1, Miyazaki, Japan.
INTRODUCTION

     This paper presents the current status and future trends
for the use of Rotating Biological Contactors in Japan.  It
includes a historical survey. Since 1966, the number of waste-
water treatment plants using RBCs has risen to over 1,323.
The total flow was 4^3,000 m3/day in June 1981.  Over 300 add-
itional plants are now under construction.  Most of these are
utilized for secondary waste-water treatment, but U2 plants
have been installed for nitrification and 17 other plants for
BOD and nitrogen removal.  The first denitrification RBC plant
has been in operation since 1976.  In 198l the first objective
RBC nitrification plant was built for treating surface water
prior to water purification.  Another special nitrification
plant is under construction for nitrification of rice field
irrigation water and has a design flow of 8U,000 m3/day.  At
the present time there are 22 RBC manufacturers in Japan.
Seven of them have technical tie-ups with foreign enterprises,
the other 15 manufacturers have developed their own technology.
Investigation of the RBC is very active.  About 100 papers
were presented at Annual conference and published in the Jour-
nal of the Japan Society of Civil Engineers, and the Journal
                             112

-------
of the Japan Sewage Works Association, etc. last year.

1.  HISTORICAL REVIEW OF THE RBC PROCESS IN JAPAI

     K. Kohyama, Department of Sanitary Engineering,' Hokkaido
University conducted the first experimental research of the
RBC in I960, for the treatment of Potato starch wastewater (l).
In 196U, M.Ishiguro,Department of Civil Engineering, Miyazaki
University, began studies'on the BBC for treatment of Sweet
Potato starch wastewater.  As a result of this research, the
first full scale RBC process in Japan for the treatment of
Sweet Potato starch wastewater was installed at the end of
1966 in Miyazaki prefecture.  This plant was constructed with
five stages^ a disk diameter of 2.0 m, and a surface area of
1,500 m2^ Polystyrene was used for the discs.  The concentra-
tion of influent BOD5 is 10,000 mg/1, and the flow rate is 600
m3/day.  The design for BOD loading is 900 gBOD/m2day which
achieves a 70% BOD reduction (2).  Hot many more RBC process
plants were installed, until 1971, "but investigation of the RBC
continued'Steadily'at both the above mentioned Universities
and at other places (3, ^+, 55 6).
     Table - 1 summerizes the number of operating•RBC plants
from 1972 up to June 30, 1981.

           Table 1. Number of.RBC Plants in Japan
       Year        1972   1973   197^  .1975   -1976-  1977
  No. of Plants       h     25    -60     96  .  252   ,.k69
       Year        1978   1979   1980   198r ( June. 30
  No. of Plants     701    9^8   1206        1323
In 1972, there were only h wastewater treatment facilities
utilizing the RBC.  Since 1973, the number has increased from
year to year to more than 1,323, with another 300 now under
construction (7).
                               113

-------
2.  CURRENT STATUS

2-1 Ixsisting RBC Plants
     Table 2. summerizes the number of locations and the quan-
tity of flow for the six major wastewater categories -identi-
fied by source; pre-purification surface water (tap water sou-
rces), domestic, food processing, industrial (e.g. the pulp
and chemical industries), waste treatment and disposal (e.g.
landfills and wastematerial treatment plants), and animal
breeding (?)•

   Table 2. Summary of RBC Plants in Japan (June 30, 198l)


 Wastewater                   Flow (m3/d),Flow(f3)5,Site5Site(^)
Tapwater sources
Domestic
Pood processing
Industrial
Waste treatment and disposal
Animal breeding
Total
lU,200
22U,321
35,81* U
121, it?1*
23,675
3,381
It22,875
3
53
8
29
6
1
100
1
65U
2U3
261
135
29
1323
0.1
50
18
20
10
2
100
     There were over 1,323 RBC plants with a total flow of
1|J|2,875 m^/day by June 1981.  The tap water sources in Table 2
reflects the fact that in Japan the largest volume of water
for municipal use is taken from surface water.  The water
sources have become polluted with organic wastes and nitrogen.
Therefore, an RBC nitrification process has been installed for
surface water prior to the water's treatment in the water pur-
ification plant.  Further details of the plant are given in
section 3.
     There are approximately  651* RBC plants currently treat-
ing municipal wastwater.  The. largest operating RBC facility
in Tokushima City has 32 shafts, and a flow of- 31,600 m3/day
(Design flow: 63,200 m3/day)(8).  Two hundred forty three
instalations treat food processing wastewater, two hundred
sixty one installations treat industrial wastewater.  The
largest operating RBC facility has Uo shafts, a flow of 12,000
                               114

-------
m3/day for the treatment of water from the manufacture of pulp.
One hundred thirty five installations treat landfills -(garbage
dump) for BOD removal, nitrification and denitrification.  The
first such RBC plant has "been in operation in Miyazaki City
since 1976 (9', 10).  There are twenty-nine plants treating
wastewater from.animal breeding (7,ll).
     Table 3. lists the distribution of Table 2. summary trea-
tment facilities by flow range.   Approximately 53% of the ex-
isting facilities are package plants treating a discharge
flow below 100 m3/day (0.03 MOD.).

            Table 3- Total Number of Operating RBC
                     Installations (June 30, 198l)


 Flow range (m3/day)   Total No.    Sub.total        %
0 -
100 -
300 -
500 -
1,000 -
3,000 -
5,000 -
10,000 -
20,000 -
30,000 -
99
-299
U99-
999
2,999
M99
9,999
19,999
29,999
39,999
702
U05
101
57-
37
8
5
6
1
1
1,170
1,208
1,265
1,302
1,310
1,315 •
1,321
1,322
1,323
53
Qk
91
96
98
99 -
99. U
99.8
99-9
100.0
     Approval and financing by the Japanese Ministry of Con-
struction for the RBC process for municipal wastewater is
about ten years behind Europe and the U.S.A.  The first RBC
plant for public sewerage treatment was constructed in 1978.
For that reason, in the early years after the RBC process was
introduced into Japan, almost none were installed for munici-
pal sewerage; therefore Japanese RBC engineers concentrated
their efforts on the most difficult wastewater treatment for
industrial etc.  They have achieved to success with that
wastewater treatment.
     Table U. summarizes the design criteria for surface load-
ing of BOD in order to achieve the concentration effluents of
BOD below 20 mg/1 except for domestic wastewater (ll).
                               115

-------
  Table k.  Design surface loading rates for all types of
            wastewater  excluding domestic
No . Wast ewat er
1 Marine product process
2 Fish meat process
3 Pish market place
k Meat process
5 Eatable bird process
6 Bean paste (Mi so) Soy manf. process
7 Eatable food oil manf. process
8 Pickles manf.
9 Sake brewing (brewery)
10 Dairy
11 Fruit Canning
12 Orange Canning
13 A taro Canning
Ik Center of feeding
15 Silk yarn manf.
l6 Dyeing manf.
17 Paint material product
18 Woolmil manf.
19 Wood pulp manf.
20 Refinery bleaching
21 Old paper reproduct
22 Bleaching paper manf.
23 Petrochemistry manf.
2h Cleaning (wet)
25 Cleaning (dry)
26 Medicine manf.
27 Hospital wastewater
28 Slaughter-house
29 Hog yard
30 Diluted night soil
31 Waste material treatment plant
32 Garbage dump
Influent
BOD5 (mg/1)
koo -
150 -
100 -
100 -
300 -
150 -
1*00 -
500. -
700 -
300 -
1000 -
200 -
100 -
200 -
lUoo -
120 -
70 -
150 -
1000 -
800 -
300 -
50 -
100 -
80 -
300 -
600 -
120 -
750 -
200 -
1500 -
300 -
10 -
1000
1*50
600
1500
1500
600
600
1500
2000
Uoo
1600
ikOQ •
200
500*
6000
200
lUO
200
2300
1000
800
100
800
lUO
500
1000
l»50
2500
1300
2000
1000
200
BOD loading
(g/m2d)
30
25
15
10
15
5
20
30
15
30
20
30
15
10
10
20
5
20
10
50
10
10
5
8
10
5
10
80
5
5
10
2
- 90
- 60
- 20
- 20
- 20
- 25
- 25
- 50
- 20
- 60
- 60
- Iio
- 25
- 30
- 20
- ^0
- 10
- 2?.
- 80
- 65
- 20
- 15
- 80
- 10
- 20
- 25
- 15
-100
- 50
- 30
- 20
- 20
2-2 RBC Manufacture
     At the present time there are 22 RBC manufactures in
Japan.  Seven of them have technical tie-ups with foreign en-
terprises: Autotrol (U.S.A.) with Nippon Autotrol (1972),
Schuler-Stengelin (West Germany) with Pacific Engineering and
also with Mitsuitoatsu (1973), Ames Croster (U.K.) with
                              116

-------
Niigata Tekko (1973), Mecana (Switzerland) with Takuma (197M,
Clow Envirodisc (U.S.A.) with Sinko-Pfaudra (1978) and Bio-
Shaft (U.S.A.) with Maezawa Kogyo (1980).  The other 15 manu-
factures have developed their own technology:  Kurita-Kogyo,
Shinmeiwa Kogyo, Dengyosha, Tore-Engineering,  Sekisui Kagaku
Kogyo, Asahi Engineering, Unichica, Matsushita Seiko, Showa
Koji, Sanki-Kogyo, Meiden-sha, Kyushu Denko, Organo, 'Sekine
Sangyo and Tsutsunaka Plastic (ll).

2-3 Existing Facilities
     Table 5- shows the nominal parameters associated with the
media and mechanical components for the 22 RBC manufactures in
Japan.  Each equipment manufacturer offers variations of the
media and drive components.  The media material, support,
shaft strength, tank shape, and clearance are some of the
items which have affected RBC'performance.  The maximum values
of disc diameter, surface area, and shaft length are 5-0 m,
19,170 m^, and 8.8 m, respectively.  There are many shapes
for the-disc surface, e.g. flat, combined flat and corrugated,
waved, double-waved, two flat plates combined, flat-netted,
etc.

             Table 5- RBC Equipment Dimentions
  Media  :   Disc
         Shape
         Material
         Diameter
         Thickness
         Surface area
         Spacing
         Construction
  Mechanical  :
         Shape
Shaft
         Material
         Thickness
         Length
Circular, Octagonal
High density Polystyrene, Polyethylene,
Hard Polyvinyl Chloride, FRP (Fiber
glass Reinforced Plastic)
Standard: 3.6 m, Range: 1.0 - 5-0 m
0.7 - 7.0 mm
300 - 19,170 m2/shaft
1.0 - 3.2 cm
Segmented (12, 8 or 5 pieces) :  Steel
supported.  Unitized :  Heat welded
self-supported.

Cross-section  : Circular, Round Square
Octagonal
Steel
1.90 - 3.80 cm
Standard  : 7-5 m, Range  : 1.0 - 8.8 m"
                               117

-------
              :   Motors;
         Horsepower   :  0.5 - 15-0
                 Drive Units              .....
                        Multi-V Belts, Chain and sprocket,
                        Enclosed cartridge, Air Driven,
                        Water Driven.
     Recentry, a new drive unit process has been developed by
Kurita—Kogyo : Water is introduced to aid in rotating the di-
scs. Plastic water cups are welded onto the periphery of the
media over the entire length of the contactor.  The waste
water is dropped from a height of about 1.0 m above the top
periphery of the media and is captured "by the plastic cups.
The falling wastewater causes the 3-6 m-diameter RBC disc to
rotate.  The process could be combined with Activated Sludge
process, e.g.  Northeast sewage treatment plant in Philadel-
phia, U.S.A. (12, 13, l»6).
     Another more highly technique, developed by Meiden-sha,
is the automated control of the rotational speed of the discs
depending on the quality of the influent water.  It is well
recognized that when the BOD concentration in the influent in-
creases, the additional BOD removal can be achieved by incr-
ease the rotational speed of the disc.  Self-variation of ro-
tational speed by the newly developed equipment could match
the variation of influent flow rate, temperature, and con-
centration of organics.  In the operation of the RBC process
for variable loading such as industrial wastewater, this new
technique and equipments will improve the maintenance and
treatment efficiency (lU, 15).

2-k RBC System Study Mission to Foreign Countries
     RBC system study mission have been organised six times
since 1975 and have visited foreign RBC manufacturers and
RBC plants under construction or operating :  (l) August 1975
(U.S.A.), (2) November 1975 (U.S.A.), (3) September 1977 (50-
th annual conference of the Water Pollution Control Federation
in Philadelphia), (U) June 1978 (Europe-Denmark, Sweden, West
Germany, Austria, France, Switzerland, and the U.K. including
the 9th International Conference of the International Associ-•
ation on Water Pollution Research in Stockholm, Sweden), (5)
June 1980 (Canada and the U.S.A., the 10th International Con-
ference of the IAWPR, Toronto, Canada), (6) April 1982 (U.S.A.
attended the 1st International Conference on Fixed - Film Bio-
logical Processes, Kings Island, Ohio, U.S.A.) (11, 16, 17).
                               118

-------
 3.   SPECIAL APPLICATION  OF  THE RBC

      As  already  mentioned,  most of  the  RBG  plants are utilized
 for  secondary  waste-water treatment.  About  5% of the total
 number are  utilized for  nitrification and denitrification.
 Tth  following  discussion will  be concerned  with two specisl
 applications of  RBC plants  for the  removal  of low concentre^
 tions of ammonia-nitrogen.

 3-1  Nitrification  Prior  to  Water Purification
      In  Japan, the largest  volume of water  for municipal use
 has  been taken from surface water.  The surface water has be-
'came polluted  with organics and nitrogen, so that the cost for
 prechlorination  (addition of chlorine at the mixing basin) and
 for  other chemicals have greatly increased  at water purifica-
 tion plants.  Therefore, it has caused  a rise in the cost of
 water supply and plant maintenance.  Trihalomethane (THM: a
 cancerous growth matter) is produced in the reaction between
 organics and chlorine, is becoming  a world-wide problem.  In
 addition, the  rejection  of  high amounts of  ammonia-nitrogen
 in raw water requires a  large  amount of chlorine, which might
 cause the production of  THM.
      Several studies have been made to  find a process which
 could be installed prior to the water purification process in
 order to solve the problem.  The unit processes evaluated were
 activated sludge,  trickling filters, submerged biological
 filters, stripping, and  the RBC process.  The RBC process was
 selected because of its  simplicity  of construction, operation,
 maintenance, and low energy requirements.
      Field  tests using a RBC pilot  plant were carried out from
 April 19T6  to  October 1980  in  order to  examine the effect of
 the  reduction  of organics and  the oxidation of ammonia-nitro-
 gen  in low  concentrations in river  surface  water.  Based on
 the  results of the field test, the  first RBC nitrification
 plant for use  prior to water purification plant was construct-
 ed in apri1-1981 with the approval  by the Ministry of Health
 and  Welfare (l8).   The plant is installed in'Nakama-City,
 Fukuoka-Prefecture, in the  island of Kyushu and treats most
 of the downstream  water  of  the Onga-River,  running into the
 Genkai Sea.
      Figure 1. is  a diagramatic sketch  of a rapid sand filter.
 It shows the path  of the water through  the  various units.
 Conventional types of water purification plants 'are shown by
 the  broken  line  and RBC  nitrification unit  by solid lines.
                               119

-------
    from surface
    water supply
                  |Pre-         |  (Mixing,
                  'chlorinationi~^'tank  i
                  i	i  i	j
(Floeeulation!
[SettlingJ_. [Rapid      S
'basin   i  'Sand filter1
i ______ i  i _______ i
                                            i_\
                                'chlorinatiom   pump
                                               r
                                               to service
          Fig 1.  Schematic flow diagram of an upgraded
                  Water purification plant with RBC
                  nitrification

     Table 6. shows the characteristics of the Onga River
water and the percent.reductions in the listed items by the
RBC pilot plant. It indicates that the concentration of PIH3-N
is higher in winter than in summer due to the small discharge
rate of the river in winter.

    Table 6. Characteristics of River water and RBC test
Concentration ___ , ,
RBC test

Water temperature (C)
DO (mg/1)
PH
COD (mg/1)
BOD (mg/1)
SS (mg/1)
NH3-M (mg/1)
Degree of turbidity (mg/1)
Color (mg/l)
Threshold odor number (TO)
Chlorine requirement (mg/l)
Total iron (mg/l)
Manganese (mg/l)
Max.
30. i*
10. T
8.3
13.6
5-3
1*7.2
3.0
11.0
1*6.0
50.0
15.5
0.31*
0.22
0
Min. Mean '
u.o 15.9
I*. 5 7.7
7.3 7-8
9.8 11.6
1.2 3-3
7.6 15.2
0.02 0.67
U.5 8.2
2k 36
8.0 25.0
8.8 11.5
0.18 0.29
0.09 0.13
5 reduction



32
70
32
90
58
30
60
59
70
80
     Design criteria for the flow rate, loading rate and equ-
ipment are summarized in Table 7- for the total installation.
                              120

-------
              Table T-  Nitrification Criteria
 A.  Design Flow Rate (Water Consumption)
    Maximum - Summer
    Yearly Average
    Maximum - Winter

 B.  Loading Rates
    Hydraulics
    Hydraulics
    Hydraulics
    NH3-N
lU,200 m3/D (high water temperature)
12,000 m3/D (mid-term water temp.)
11,000 m3/D (low water temp.)
200 i/m2D (5 gpd/ft2D)  (winter)
259 l/m^D               (summer)
219 l/m2D               (mid-term)
0.256 g/m2D             (winter)
 C. RBC Equipment Dimensions

    Total Surface area
    One shaft surface area
    No. of shafts
    No. of trains
    Length of shafts
    Diameter of discs
    Material of discs
    Type of disc surface

    Peripheral' rotational speed
    Electric power consumption
    Detention time
    Detention time
    Detention time
    RBC manufacturer
         55,000 m2 (592,020 ft2)
         9,150 m2/shaft
         6
         3
         l.k m
         3.6m
         high-density Polyethylene
         composed of flat and corru-
         guted sheet
         18 m/min. (1.6 rpm)
         5.5 KW/shaft
         U3.^ min. (winter)
         UO.O min. (average)
         33.6 min. (summer)
         Nippon Autotrol
     In winter, the average values of the NH3-N concentration
in the river water and the required dosage of chlorine are
1.28 mg/1 and 13.5 mg/1, respectively.   However, the new pur-
ification plant with RBC nitrification unit has achieved the
effluent NH3-N concentration of O.l8 mg/1 (86% reduction) and
the chlorine dosage of 3.3 mg/1 (l6% reduction).  Yearly ave-
rages of the feeding ratios  for prechlorination have decre-
ased from 11.3 mg/1 to ^.1 mg/1 (a 58% reduction).   Moreover,
the yearly average feeding ratios  of activated carbon for the
elimination of odor, etc. of 11.2 mg/1 has decreased to ^.7
mg/1 (a 57% decrease).  The reduction in expense for chemicals
is about ¥ 10 million (U.S. $ 50,000) a year.
                              121

-------
     The plant was started-up of April 1, 198l at a flow rate
of 55220 m3/day (maximum) and It,219 m3/day (average), which
corresponded to 37 and 35$ of the design flows, respectively.
The performance of the RBC treatment has almost coincided with
the design criteria from start-up to the present day.
     A similar BBC nitrification plant is also under conside-
ration  in Nakama-City and will have a design flow of 19,700
mS/day.

3-2 Nitrification of River Water for Rice Field Irrigation
     Rice is the staple food for the Japanese.  The 3,081,000
hectares (7,700,000 acres) of rice field comprise $6% of the
total farm land in Japan.  The largest volume of water for
rice field irrigation has been taken from the surface water
of natural rivers  and from irrigation reservoirs.  Poor rice
yields have been traced to high NHg-N which causes excessive
stalk growth compared to desired kernel growth.
     The RBC process was selected to solve this problem.  The
field test of an RBC pilot plant with disc diameters 2.0 m
and a flat and waved media surface was carried out from Sep-
tember 1977 to October 1979-  The tested discs peripheral
rotational speeds were 10.0, 13-5, 18.0, 2k.0, 27.0, 30.0,
and 36.0 m/min.  The hydraulic loadings were 200, 300, 1*00,
1*50, 600, and 800 l/m2day.
     Table 8. summarizes the water quality of the river water
and the performance of the RBC pilot plant.  Hydraulic load-
ing is 600 l/m2day and peripheral rotation speed is 27 m/min.
which are the optimum conditions for the removal of ammonia-
nitrogen (19, 20).

    Table 8. Characteristics of river water and RBC test
      Items
Concentration  Concentration    Percent of
of influent    of RBC effluent  reduction
Water temp. (°C)
DO (mg/1)
COD (mg/1)
BOD (mg/1)
SS (mg/1)
N%-I (mg/1)
N03-N (mg/1)
Org-N (mg/l)
Kej-N (mg/1)
19.2 -
5.2 -
7.8 -
3.7 -
17.7 -
0.9 -
0.6 -
0.5 -
1.1* -
22.5
6.5
lU.O
18.2
U8.3
l.U
0.7
1.0
2.2
18.8 -
7.7 -
U.6 -
1.5 -
7.0 -
0.0 -
1.2 -
0.3 -
o.u -
20.0
8.7
5-9
5-5
15-9
0.3
1.9
0.8
0.8


21*
1*6
1*8
79
200
20
61*


- 1*3
- 70
- 67
-100
-283
- 50
- 80
                              122

-------
     The first RBC nitrification plant for rice field irriga-
tion water with a design flow of 70,000 m3/day (l8.5 MOD) is
•under construction with the approval of the Ministry of the
Agriculture and Forestry.  The RBC nitrification plant is be-
ing installed in Ibaragi-City, Osaka Prefecture in Central
Japan on the Yodo River, which into the Seto  Inland Sea (Seto-
naikai).
     Irrigation water  for rice fields is required from June to.
September, therefore,  the RBC plant is operated only four
months  a year.  Water  quality standards for rice field irriga-
tion water is as follows: pH (6,0 - 7.5), COD (  6 mg/l), SS
(  100  mg/l), DO (   5.0 mg/l), and TN:Kej-N (  l.Omg/l).
     The characteristics of the RBC influents are summarized
in Table 9.

         Table 9- Characteristics of RBC influents

    Items   ' HH-NNO-NOrg-NKej-N  T-N   DO   COD   BOD
Concentration Ii2               2^              Q        g'
     (mg/l)

     Final effluent values from the RBC plant of Kej-N, COD
and  DO were defined as   1.0,   6.0, and    5.0 mg/l respect-
ively.  Design  flow rate, loading rate, and equipment are
summarized in Table 10.

         Table  10. Nitrification criteria for rice
                   field irrigation water
 A. Design  flow rate
       Average flow                  : 70,000 m3/D  (l8.5 MOD)

 B. Loading rate
       Hydraulics                    : 600 l/m2D
       NHs-N                         : 0.507 g/nA)
 C. Design  RBC equipment dimensions
       Total  surface area            : l6U,l60 m2
       One  shaft surface area        : 13,680 m2/shaft
       Number of shafts              : 12
       Number of trains              : 6
                               123

-------
       Length of shafts
       Diameter of discs
       Spacing of discs
       Thickness of discs
       Material of discs
       Type of disc surface
       Peripheral rotational speed
       Electric power consumption
       Detention time
       RBC manufacturer
8.85 m
5.0 m
17.5 iran
1. 3 mm
high-density Polyethlene
flat
27 m/min
9.2 KW/shaft
20 min.
Dengyo — sha Machine
Works
     This RBC nitrification plant has six trains mechanically
driven.  There are two shafts per train.  All twelve shafts
have "been installed in one building with 91-2 m in length,
15-6 m in width, and 5-9 m in height.
     An additional RBC nitrification plant for field irriga-
tion water with a design flow of 1^,000 m3/day (3-7 MOD) will
"be eonstraeted within a few years in the same area.

1*.          STATUS OF

     Research of the RBC is very active.  Last year 100 papers
were presented at Annual Conference and published in the Jou-
rnal of the Japan Society of Civil Engineers (JSCE), the Jou-
rnal of the Japan Sewage Works Association (JSWA), the Nation-
al Symposium on RBC Technology of the Environmental Conserva-
tion Engineering Association (ECEA), and other journal of
wastewater treatment, etc, (7).  The first special edition on
the RBC process appeared in Journal of the Environmental Con-
servation Engineering (ECl), Vol.U, No. 7, July, 1975 C1*, 5,
6).  Another Journal of Engineering has edited a special issue
on RBC every year.
     The first Seminar ori the RBC was held in September 1975
and sponsored by the ECEA.  Since then, the Seminar was held
annualy until 1979.  In November, 1977, the RBC Wastewater
Treatment Div. of the Association was established.  As .a res-
ult of the first National Symposium on RBC Technology with
the ECEA, November 13 - 15, 1979 (2l), an RBC Symposium has
been held every year.  At the end of the Symposium, field
tours observe operating RBC plants.  In October, 1979, the
first Annual Conference of the Fixed-Film Biological process
Research Association was held in Tokyo (22).   The conference
presents research in every year that is on the Rotating Bio-
                              124

-------
logical Contactor, Trickling filter, Submerged Biofliter,etc.
     To date, research on the RBC has been conducted at the
following institutions :  Hokkaido University, Kitami Institute
of Technology, Tohoku University (23), Tokyo University (2U),
Tokyo Institute of Technology (25,26), Tokyo Metropolitan
University, Miyazaki University (27 - 30), Kagoshima Technical
College, the National Institute for Environmental Studies (31,
32), the Japan Sewerage Works Agency (12), the Consulting En-
gineers Co. (33, 3^), and among RBC manufacturers.
     The most foundamental research on the RBC was conducted
by: M. Ishigurp, Y. Watanabe, S. Masuda, K. Yamaguchi, and
H. Uchida, "Advanced Wastewater Treatment by RBC Unit (l-IV)"
and published in the Journal of the Japan Sewage Works Asso-
ciation, Vol. 12-16, from 1975 to 1979-  These papers were
awarded the 1980 thesis prize of the JSWA (35).  At the 9th,
10th, and llth International Conferences of the IAWPR in 1978,
'80, and  '82, Y. Watanabe, M. Ishiguro, and K. Nishidome pre-
sented papers on Denitrification Kinetics, Nitrification Kine-
tics and Simulation of Nitrification in an RBC Unit (36, 37,
38).  At the 1st National Symposium on RBC Technology in Feb.
1980, Pa. U.S.A., K. Ito and T. Matsuo presented a paper on
"The Effect of Organic loading on Nitrification in RBC Waste-
water Treatment Processes" (2M, and H.Iemura and R.J.Hynek
presented a paper on "Nitrogen and Phosphorus Removal with
RBC" (10).
     Many books have been published concerning RBCs: l) The
Newest Technique of wastewater treatment by Biochemical Pro-
cesses (39),, 2) Guide book for night soil treatment (Uo), 3)
Wastewater treatment "by the RBC (Ul), h) Compilation "book of
Wastewater treatment technique by RBC (1*2), 5) Guide book for
sewage, industrial wastewater, and sludge treatment (Us),
6) The Fixed-Film Biological Process (ll), 7) Guide book for
Domestic wastewater treatment (UU), etc.  In particular the
literature 6) includes the newest theory and the design pro-
cedures for Trickling filter, Rotating Biological Contactor,
and Submerged Biofilter systems.

5.  FUTURE TRENDS

     To date, the pace of development of sewerage facilities
in Japan has been slow.  There are many reasons for this.
Until recently, night soil was plowed back into farmland.
Because water pollution problems were rare in the past, the
importance of sewerage tended to be minimized.  After World
War II, farmers began to use chemical fertilizers instead of
                              125

-------
night soil (septage).  Night soil was disposed of in other
ways, primarily being discharged into rivers and other bodies
of water, eventually polluting them (^5).
     The J.k% of the population was served by sewerage in 19-
63.  The spread of sewer system in Japan lags far behind that
of other developed countries.  The systematic construction of
sewerage facilities began with the First Five-year plan of
Sewerage Construction (1963 - 67) in 1963.  Although the pop-
ulation served by sewerage has increased along with increased
Five-year plan,  it was still only 30% at the end of fiscal
year 1980 ( the end of the Fourth Five-year plan).  Coverage
was 70$ in 10 major cities having populations of more than 1
million, whereas that in other cities was under 20%, showing
that wastewater treatment works in smaller cities lags far
behind that in large cities.  By about the year 2,000, B0%
of the population should be served by sewerage.  According to
the Fifth Five-year plan, coverage should increase to about
kk% by about 1985-  Because the RBC process is characterized
"by low maintenance costs and low energy consumption, Sewerage
facilities using the RBC process would be constructed espee-
eally in smaller cities in Japan (i*6).  Over twenty cities
are constructing or planning sewage treatment plants with the
RBC process.
     In Japan, approximately 75% of the night soil collected
by vaccum trucks.  The collected night soil is treated at pub-
lic wastewater treatment plants, home night soil purification
tanks or collected night soil treatment plants.  There are '
1,186 night soil treatment plants, with a total planned pro-
cessing capacity of 9^,126 kl/day in 1980.  These plants are
viewed as a transitional measure until public sewerage could
be established.  Thus, Japan will continue to depend heavily
on the collected night soil treatment plants for some time.
RBC design criteria for night soil purification tanks (inclu-
ding domestic wastewater) have been authorized by the Minist-
ry of Health and Welfare since July, 1980 (^7).  It seems
likely that the number of RBC plants should increase in small-
er cities, towns, rural communities (farm, fishing and moun-
tain villages) etc.
     Finally, the RBC process should increase rapidly in the
following field in Japan :'
       a)  public sewage treatment plants
       b)  community plants
       c)  home night soil purification tanks
       d)  treatment plants for
                i)  food processing
                               126

-------
               ii)  waste material treatment and disposal,
              iii)  animal breading, and other industrial
                    wastewater.
       e)  special applications
                i)  pre-nitrification of surface water for
                    water purification plant,
               ii)  nitrification plants treating river water
                    for rice field irrigation.
     In addition, research on EEC will continue actively in
Universities, Government Institute, Consulting Engineering
Companies, and RBC manufacturers in Japan.

REFERENCES

 1,   Kohyama  K., Inoue I., and Takayasu M. ,"Studies on the
      Biochemical Treatment for Wastewater in the Manufacture
      of Potato Starch", Journal of Water and Wastewater, Vol.
      3, No.12, Dec, 196l, pp.1-10.  (in Japanese)
 2.   Ishiguro M., Takahata S., and Wakamura I.,"Studies on
      the Sweet Potato Wastewater Treatment by the Rotating
      Biological Contactor", Proc.  21st Annual Conference of
      Japan Society of Civil Engineers, May, 1966,•pp.1^6-
      1^9- (in Japanese)
 3.   Ishiguro M.,"Wastewater Treatment by the Rotating Bio-
      logical Contactor", Journal of Japan Sewage Works Asso-
      ciation, Vol.10, No.Ill,  Aug.1972, pp. 1.8-29. (in Japan-
      ese)
 k.   Ishiguro M.  ,"The Secondary and Tertiary Treatment of
      Municipal and Industrial Wastewater by the RBC", Envi- -
      ronmental Conservation Engineering {ECE),Vol.U, Mo.7,
      July,19755 pp.1-21. (in Japanese)
 5.   Ishiguro M.  ,"Current Status of RBC in Abroad", ECE.,
      Vol. It i  No.7, July.1975, pp.^2-52. (in Japanese)
 6.   Kohyama K, and Kato Y. ,"Purification Mechanism of the
      RBC", ECE.,  Vol.it, No. 7,  July. 1975, pp.Sl-1*!.  (in Japan
      ese)
 7.   Ishiguro M. /'Current Status and Future Trends of RBC",
      Pr.oc.,  The Third National Symposium on RBC Technology.
      Oct.1981, pp.61-71. and ECE.,  Vol.10,No.12, Dec.l98l,
      pp.37-*t6. (in Japanese)
 8.   Yamashiro Y. ,"RBC Process in Tokushima City Central
      Sewage Treatment Plant",  Froc.  The Fifth Seminar of the
      RBC., May.1979, pp.57-70. (in Japanese)
 9-   Ishiguro M., Watanabe Y., and Masuda S. ,"Treatment of
      Leachate from  Sanitary  Landfill"  ECE., Vol.7,
                              127

-------
      No.6, June.1978j pp.3-11- (in Japanese)
10.   Hynek R.J., and lemura H. ."Nitrification and Phosphorus
      Removal with RBC", First National Symposium on RBC Tech-
      nology. Proc., pp.295-32it. Pa. U.S.A., Feb.1980.
11.   Iwai S., Kusumoto M., and Ishiguro M. et.al,, "Fixed-
      Film Biological Process", Chapter U, RBC.process. San~
      gyo-yosui-Chosakai. Co., Nov.1980. (in Japanese)
12.   Qkuno N.,"Upgrading of the Existing Activated Sludge
      Process "by the RBC", Proc., l8th Annual Conference of
      JSWA. May.1981 pp.198-200. (in Japanese)
13.   Kurita Kogyo »"Pilot Scale Studies on the Combination of
      a new developed RBC process and the existing Activated
      Sludge process", SK NEWS. No.36, March.1981. (in Japan-
      ese)
lU.   Meiden-sha ,"RBC Process with Auto Speed Regulator",
      Technical Data. Io.GP-2028. Sept.1980. (in Japanese)
15.   Meidensha/'Meiden Wastewater Recycling System for Buil-
      ding", Catalog.Wo.BB52-2038. Sept.l98l. "Bio-rotacon:
      Rotating Disc Type Biological Contactor for Water Treat-
      ment", MP-9118. March.1980.
16.   Mori T.»"The latest dates Status of RBC in U.S.A.", ECE.
      Vol.lt, No.12, Dec.1975, pp.tit-it?, (in Japanese)
17.   Ishiguro M.."Current Status of RBC in U.S.A.", ECE.,
      Vol.7, No.it, April.1978, pp.UO-U6, ."Current Status of
      RBC in Europe", ECE., Vol.7, No.11, Sept.1978, pp.9-23,
      "Current Status of RBC in North America", Proc., The
      2nd National Simposium on RBC Technology, Oct.1980,
      pp.13-22 . (in Japanese)
l8.   Water Supply Department of Nakama-City /'Experimental Re-
      port on Pre-Treatment by the RBC for Polluted Water
      Sources to use of Water Supply", June.1978. The Task
      Committee Report on Biological Treatment Process for
      Polluted Water Sources to use Water Supply. Jan.1980.
      (in Japanese)
19-   Shinan Tochi Kairyouku and Ninon Meintenace Engineering
      Co., "Experimental Study Report on the Purification
      Effect of Rice Field Irrigation Water by RBC Pilot Plant"
      Dec.1978.  (in Japanese)
20.   Sasaki A.,"Treatment of Agricultural Water containing
      low level of Contaminated by RBC Process", ICE., Vol.8,
      No.10, Oct.1979, pp.22-29. Proc., The First National
      Symposium on RBC Technology, ECE Association, Nov.1979
      pp.55-58.  (in Japanese)
21.   Proc.,"1st. National Symposium on RBC Technology" Envi-
      ronmental Conservation Engineering Association. Nov.
                             128

-------
      1979- (in Japanese)
22.   Proc. ,"lst Annual Conference on Fixed-Film Biological
      Processes", FFBP. Research Association. Oct.1979. (in
      Japanese)
23.   Nakamura K. , Matsumoto J. , and. Noike T. ,"With a Iron
      Bacteria RBC Treatment of Strength Accid Mine Drainage"
      Proc., The 35th Annual Conference, JSCE.  Sept.I960, pp.
      670-671. (in Japanese)
2k.   Ito K. and Matsuo T. ,"The Effect of Organic Loading on
      Nitrification in RBC Wastewater Treatment Processes"
      Proc., First National Symposium on RBC Technology, Pa.
      U.S.A., Feb.1980, pp.1165-1176.
25.   Kubota H., and Takahashi M. ,"Effect of Daily-Fluctuation
      on Rotating Biological Wastewater Treatment Contactor",
      Japan Journal of Water Pollution Research, Vol.1, No.3
      Dec.1978, pp.175-182. (in Japanese)
26.   Kubota H. ,"Operation and Design Proceedures of RBC",
      ECE., Vol.9, No.6, June.1980, pp.50-56. (in Japanese)
27.   Ishiguro M., Watanabe Y., and Masuda S. ,"Advanced Waste-
      water Treatment by RBC", Proc., llth Symposium of
      Sanitary Engineering, JSCE., Jan.1975, pp.!09-llU. (in
      Japanese)
28.   Watanabe Y., Ishiguro M., and Nishidome K. ,"Kinetic
      Analysis of Denitrification by RBC Unit", Proc, of JSCE
      No.276, Aug.1978, pp.35-^3. (in Japanese)  and Trans-
      actions of JSCE., Vol.10, Aug.1978, pp.166-169.
29.   Watanabe Y., Ishiguro M., and Nishidome K.,"A Mechanism
      of Substrate Removal in RBC Reactor (l),  (ll)", Journal
      of JSWA., Vol.15, No.172, Sept.1978, pp.2i-3U. Vol.17,:
      No.195, Aug.1980, pp.lU-23. (in Japanese)
30.   Masuda S. , Ishiguro M. , and Watanabe Y. .."Nitrogen Remo-
      val in RBC (l).  Simultaneous Denitrification with
      Nitrification in Biofilm", Journal of JSWA., Vol.l6,
      Ho.187, Dec.1979, pp.2^-32. (in Japanese)
31.   Sudo R., Okada M., and Mori C.,"An Aspect of Micro-Orga-
      nism in the Bio-film for RBC"5 Journal of Japan Society
      of Fermentation Engineering. Vol.56, Feb.1978, p.580.
      (in Japanese)
32.   Sudo R., Okada M., and Mori C. ,"The Microorganisms
      Control in a RBC units", Water and Wastewater, Vol.19
      No.7, July.1977, pp.6l-70.  (in Japanese)
33-   Takahata S., and Fujishima M. ,"The Rational Design Pro-
      ceedure of RBC", Proc., The First National Symposium on
      RBC Technology, Nov.1979, pp.27-3^. (in Japanese)
3^-.   Kato Y. ,"Energy Saving in Wastewater Treatment Plants
                              129

-------
      and RBC", ECE.,Vol.9, No.8, Dec.1980, pp.69-77. (in Ja-
      panese)
35-   Ishiguro M., Watanabe Y., Masuda S. , Yamaguehi K., and
      Uchida H.,"Advanced Wastewater Treatment "by RBC Unit (I
      -IV)", Journal of JSWA. , Vol.12, No'. 129, Feb.1975, pp.
      h6~5h, Vol.lU, No.152, Jan.1977, pp.32-Ul, Vol.lU, No.
      l6l, Oct.1977, pp.53-59  Vol.l6, Mo.185, Oct.1979, pp.
      1*0-^8. (in Japanese)
36.   Watanabe Y., and Ishiguro M.,"Benitrifieation Kinetics
      in a Submerged Biological Disk Unit", Progress Water
      Technology, Vol.10, Nos.5/6, pp.187-195- IAWPR.,9th In-
      ternational Conference, Stockholm, Sweden, June.1978.
37.   Watanabe Y., Ishiguro M. , and Nishidome K. .."Nitrifica-
      tion Kinetics in a Rotating Biological Disk Reactor",
      Progress Water Technology, Vol.12, pp.233-251. IAWPR.,
      10th International Conference, Toronto, Canada, June.
      1980.
38.   Watanabe Y., Bravo H.E., and Nishidome K. /'Simulation
      of Nitrification and its Dynamics in a Rotating Biolo-
      gical Contactor", IAWPR., llth International Conference,
      Kapetown, South Africa, March.1982.
39-   Kojima S., and Ishiguro M. et.al., "Newest Technique of
      Wastewater Treatment by Biochemical Process" Management
      Co., July.1975. (in Japanese)
kO.   Iwai S. ,"Guide book of Night Soil Treatment", Kankyo-
      gijutsu Co., May.1978.  (in Japanese)
Ul.   Research Groupe of RBC ,"Wastewater Treatment by RBC Pro-
      cess", Sankaido Co., Sep.1978, (in Japanese)
1*2.   Ishiguro M. et.al. ,"Compillation book of Wastewater
      Treatment Technique by RBC", IPC Co., March.1979- (in
      Japanese)
U3-   Iwai S. ,"Guide book of Sewage, Industrial Wastewater
      and Sludge Treatment", Kankyo Gijutsu Co., Aug.1979-
      (in Japanese)
UU.   Iwai S., Kato Y. et.al.,"Guide Book of Domestic Waste-
      water Treatment", Kankyo Gijutsu Co., Aug.l98l. (in
      Japanese)
^5.   Special Edition ,"Sewerage in Japan",Journal of ¥PCF.,
      Vol.52, No. 5, May.1980, pp. 8U1-1051*.
1*6.   Nakamoto I. ,"Current Status of Sewage Treatment Techno-
      logy and RBC Process In Japan",  The Third National
      Symposium of RBC Technology, Oct.1981, Proc. pp.72-78.
      and ECE., Vol.10, No.12, Dec,198l, pp.U7-51- (in Japan-
      ese)
kj.   Supervisions of Ministry of Construction, Ministry of
                             130

-------
Health & Welfare and Environmental Protection Agency
"Design Criteria of Night Soil Purification Tank and
its Explanation", Japan Architecture Center, Aug.1980.
(in Japanese)
                         131

-------
       RBC UNIT: BEST  IN SEWAGE TREATMENT FOR SAUDI ARABIA
       Sharaf Eldin  I. Bannaga. Directorate for Housing and
       Military Cfties, Saudi Arabian National Guard, Riyadh,
       Saudi Arabia.
  INTRODUCTION: THE NEED FOR ADEQUATE WASTEWATER TREATMENT
               FOR SAUDI ARABIA.
      Jhe Kingdom of Saudi Arabia which boundaries are the Red
sea from the West, the Arabian Gulf from the East, Latitude
17  from the South and Latitude 30 from the North, covers an
area of more than two million Sq. km and is inhabited by seven
million people approximately.  The region occupies a leading
place in the Islamic World and sustains its heritage and
culture.
     Being a mdjor exporter of petroleum, the Kingdom has
been spending vast sums of money on ambitious development
programmes, aimed at every sector of the economy.  This very
fact provides a link with the question of water supply and
waste water purification,^for just as all vital processes
depend functionally on water as the medium, so do almost all
Industrial  production processes.
     Due to the scarcity of water, which imposes a major
problem for the Kingdom of Saudi Arabia, waste water treat-
ment should not be aimed only at disposal for public health
considerations,but also at recirculation of treated effluent
for special use.  Use of treated effluent is at the moment
                               132

-------
                              o
                              o
                              o
133

-------
.limited  to  Irrigation and  landscaping, but further develop-
ments could  Include use  in  industries and drinking water
supplies.   It  is encouraging to note the Fatwa(legal opinion)
announced by the Board of  Scientifie(Religious)Research,  Ifta
(delivery of legal opinion), Dawa  (invitation  to  islam) and
Guidance regarding the possible use of adequate,  safe treated
effluent for religious purposes.
     It  is  therefore imperative that great attention be given
to development of an adequate waste water treatment processes
that will optimise in the  Kingdom capabi1 Ity,manpower and
materials. This need can best be demonstrated  by  some schemes
executed in similar countries which are not  in-country compa-
tible.
     To pick a process that will discharge total ability and
satisfy the Kingdom requirements for waste water  treatment
the writer has to recommend the RBC(Rotating Biological
Contactors)process.
     The purpose of this report is to present a short account
of literature about the RBC process which may be beneficial
to engineers, consultants and research workers working in
Saudi Arabia and elsewhere. However, the writer wishes to
state that the opinions expressed in this report are entirely
his own and do not necessary reflect the views of the govern-
mental  organization by whom.he is employed.

     RBC UNIT:  PRACTICAL APPLICATION AND TECHNICAL PARTICULARS

     The Rotating Biological Contactors (RBC) unit is a
secondary biological  treatment unit for waste water. The
system consists of a number of large diameter plastic or
expanded metal  discs mounted on a horizontal  shaft and placed
in a reaction vessel  which is often of a semi-circular cross-
section. Numerous terms are used throughout  the wastewater
treatment literature to designate RBC's.  Among the trade
terms in current use are the Bio-Disc,  Bio-Surf, Aero-Surf,
Surfact, Bio-Sperial,  Rotating Disc, etc.
     The RBC process was developed initially in Europe in the
1950's.  Further development of the process began In the
United States In 1965 and has. continued to the present time.
However, active commercial  use of RBC plants  had not started
until  only in the early 70's in the U.S.
     In  Saudi Arabia practical  application of the RBC
process  started a few years ago in the form  of package plants
serving  small communities,  such as university campuses,
Hospitals,  army bases  etc.  Initial  emphasis  was directed to
                           134

-------
secondary treatment of municipal wastewater and may continue
years ahead.
     There are now more than 50 RBC plants operating in Saudi
Arabia treating of over 20 MGD of mainly domestic wastewater
These plants range in size from one to twenty four RBC
assemblies and treat wastewater flows up to 4.8 MGD plant.
King Abdul Aziz International Airport at Jeddah, is the
largest Aero-Surf facility in Saudi Arabia.The RBC process
is now gaining wide acceptance in the Kingdom for a large
number of plants are under construction. These include the
7-5 MGD Yanbu RBC plant near the Red Sea and those awaiting
construction for use of the Saudi Arabian National  Guard
Housing Project-Phase one at five different localities and
capacity of which exceeds 8 MGD.
     It should be emphasised that the growth of RBC process
with regard to product development and commercial utiliza-
tion seems promising and could be rapid when its applicabili-
ty is recognised by the local authorities.

     The RBC unit consists of;

     a) Large plastic discs mounted on a horizontal metallic
        shaft (refer to figure 1). The discs are so mounted
        that slightly less than half of their surface area
        is immersed In waste water.
     b) The discs and shaft assembly is placed In a tank
        which has a rectangular surface area.  The tank is
        usually constructed of reinforced concrete.
     c) A driving system is incorporated with the disc and
        shaft assembly. The mechanical driving system
        incorporates an electric motor that rotates the disc
        and shaft assembly.In Autotrol Aero-Surf units the
        disc and shaft assembly is rotated by buoyant force
        exerted by captured air{refer to figure 1). Aero-
        Surf assembly consists of corrugated media  with
        plastic cups attached around the outer perimeter and
        over the entire length of the contactor.  The media
        assembly Is installed in a tank in the same manner
        as a conventional  unit with the addition of an air
        header at a low pressure Into the attached  cups.
                       The captured air exerts a buoyant
        force, which in turn exerts a torque on the shaft
        sufficient for rotation.
     d) The tank is divided by a number of baffles  for flow
        direction as well  as creation of stages to  the
                           135

-------
 Eigure    1-
Figure     2
     136

-------
       process.
       Statically the disc and shaft assembly support the
  biological growth. The micro-organisms which are present
  naturally in wastewater adhere  to the disc surface and mul-
  tiply quickly within a few days of start-up and they cover
  the entire surface area of the discs.
        B/ rotating the disc and shaft assembly two functions
  are fulfi1 led:

       a) Increasing the dissolved oxygen content of the mixed
          1iquor.
       b) Provide contact between the biological  growth and
          the waste water. The rotation  alternately submerges
          the attached biological growth and then exposes them
          to air. The mixing and agitation a/iables food and
          oxygen to penetrate further Into the biomass.

       The process is continuous as the  biological slime is
  alternately exposed to waste water and then to air. This
  provides a means of exposing the biological growth to the
  organic polluting load and of aerating the waste water.
  Replacing the mechanical,drive system  with an air drive
  system serves a number of purposes:

       a) Mixed liquid dissolved oxygen  concentration is
          increased through supplemental aeration.
       b) Thinner biomass Is achieved as a result of increased
          shearing action as air bubbles rise through the
          radial passages and corrugations in the media.
       c) Power consumption is optimized through variable
          speed control and reduction of biomass.

  RBC PROCESS: COMPARISON OF OPERATIONAL CHARACTERISTICS WITH
               CONVENTIONAL PROCESSES

       The RBC process may be defined as the biological de-
  composition of organic waste materials In aerobic condition
 : and without offensive odours as opposed to the anaerobic
 »' process of putrefaction with which small nuisances are
I inveriably associated. In this respect the RBC process is
- comparable to the conventional aerobic processes of the
  percolating filter and the activated sludge.
                           137

-------
    Comparison of  the RBC Process with  the Percolating
    Filter Process:

    Both  the RBC process and the percolating filter process
.are fixed film biological reactors. The growth  in both units
 is supported by fixed solid surfaces, the discs  in the RBC
unit and  the media  in the percolating filter. The difference
between the two biological processes is that the microbial
mass in the RBC system  is passed through the waste water,
while  in  the?percolating filter the waste water  is passed
over the microbial mass.
    A  factor contributing to the advantage of the RBC proce-
ss is  that the rotating discs provide an intimate contact
between the biological slime and the waste water. The rota-
ting discs also increase the degree of mixing, agitation
and turbulence in  the reaction vessel and in doing so, the
organic pollutants  in the waste water will  stand better
chances of diffusing into the biological film. Efficiency
of the percolating filter process is usually impeded by
unevenness of the settled sewage over the whole surface of
the filter and of bad circulation of air through the bed
which must reach the surface of each piece of the filter
material to keep the right kind of bacteria and other orga-
nisms  fltlive and active.
    The clogging that occurs in the percolating filter sys-
tem is prevented with the RBC unit by the sloughing action
of the excess biomass from the discs caused by the shearing
forces developed as the discs rotate. Percolating filters
are susceptible to clogging by grit settlements, moss,and
weeds.
    The dissolved oxygen content of the waste water Is
Increased by the rotating discs In the RBC units and the
supplemental  air used for driving the system of the Aero-
Surf unit. This may prevent .the development of anaerobic
conditions and hence avoiding foul  smells and objectionable
sights both on and,off the sewerage works.
    Ronald Antonie  and associates, reported that there were
no nuisances,  no objectionable odours and no files at the
village of Millwankee,  Wis., USA and so did Simpson2^.  This
is becuase the development of flies, which are often associa-
ted with the percolating filter operation,  is prevented in_
the RBC unit operation by the continuous wetting of their
biological growth.
    A substantial  amount of research work has been carried
out by a number of research workers to specify the BOD
                             138

-------
Loading that would be acceptable to the RBC process. The
Water Pollution Research Laboratory, •*  at Stevenage, UK,
suggested a BOD Loading of 5~6 g/m  of disc median, Ellis
and Bannaga'5 reported 20 g/m  while Autotrol Corporation
used 12 g/m^. For comparison the appropriate BOD Loading on
a low-rate single pass percolating filter containing a med-
ium of 50mm nominal size would only be 1-2 g/m  according
to the Water Pollution Research Laboratory. This Indicates
that the area required by a percolating filter to purify the
same amount of settled sewage is much greater. Borchardt
reported that the actual area occupied by the RBC unit was
about t/IOth of that required by a percolating filter.
     It is unnecessary to recycle the effluent to achieve
maximum wetting, dilution and flushing action in the RBC
process which is required for the percolating filter. The
report produced by the British Ministry of Housing and
Local Government'2 recommended the use of recirculation for
strong waste that makes the sewage more difficult to treat
using percolating filters.
     Both the RBC and the percolating filter systems are
simple to maintain and have a relatively low cost.  However
the percolating filter requires more labourfor the sparge
holes of its distribution arms and the arms themselves need
to be regularly cleaned and brushed out and its dosing cham-
ber and air pipes need to be maintained properly. The RBC
unit requires minimal attention from operators but its belt
and chains require checking for alignment.

     Comparison of the RBC Process with the Activated
     Sludge Process.

     The RBC process  is somewhat similar to the activated
sludge process in that it has a suspended culture of bio-
mass in its mixed liquor and both processes possess aeration
devices. However,  the part of the bio-mass  that is  in sus-
pension in the mixed  liquor is too small  to compare with
the total  amount of the biological  growth supported by the
surfaced of the discs and would therefore contribute only
marginally to the treatment.
     The RBC process  retains a large fixed  biological  film
and a great micro-organism population and because of this
the RBC process is less upset by the variation in hydraulic
loading than the activated sludge process.  Activated sludge
process is easily upset by industrial  wastes and  is incapa-
ble of handling shock loads.
                               139

-------
    The RBC unit  is more efficient per unit volume than
.activated sludge  unit. Ainsworth  reported that a settled
sewage BOD Loading of about  (0.48-1.28 ) kg/m-' of tank capa-
city  is suitable  for fairly  good purification by an activa-
ted sludge process. For comparison the appropriate BOD loadt-
ing on a RBC unit used by Ellis and Bannaga was 3 kg/mr of
tank  capacity. This indicates that the treatment capabili-
ties  of the RBC process are  much greater than those of the
activated sludge process.
    Unlike the activated sludge process oxidation of ammonia
can be attained in the RBC process within normal retention
periods.
    The sludge solids from khe RBC process have favourable
concentration characteristics thus eleminating the need for
special thickening. De-watering of sludge generated by a
RBC unit through vacuum fiIteration was satisfactorily
accomplished according to Ellis and Bannaga,  Sludge genera-
ted by an activated sludge unit was not amenable to de-water
-ing  by vacuum filters according to Quirk'°.
    The only disadvantage to RBC process is the need for
covering the unit to protect the discs from wind, sand
storms and rains.
    The RBC unit requires little maintenance and minimal
operator's attention when compared with the activated sludge
unit  for the RBC unit Is mechanically simple. The activated
sludge unit requires careful supervision. The British
Ministry of Housing and Local Government recommends that
specialist advice from the manufacturer should be obtained
because the great complexity of plant piping arrangement
and multiplicity of aeration devices etc and because the
effectiveness of the plant is dependent upon the human
element.
    The power requirements for the RBC system are considera-
bly less than an activated sludge system, because power is
only  required to rotate the discs.

    RBC: ECONOMIC  FEASIBILITY AND SUITABILITY FOR SAUDI
         ARABIA.

    In order to examine the economic feasibility of the
RBC unit,  the unit has to be matched with the conventional
ones  in regard to  capital expenditure and running costs.
It is very difficult to compare the capital costs of waste
water treatment units  in Saudi  Arabia since these units
are mostly parts of large projects  usually awarded on
                       140

-------
4umpsum basis. The capital cost required for supplying»cons-
tructing and  installing a RBC unit may be comparative to
that of a percolating filter of same capacity for the cost
of land is becoming very expensive in urban areas of Saudi
Arabia. The RBC is now recognized as a cost-effective and
cost-competitive since the annual operation and maintenance
costs play an important factor in determining the selective.
Lundberg and Pierce   present a summary of the results of
cost-effectiveness analyses which compared the aii—drive and
mechanical RBC processes with air and pure oxygen activated =
sludge processes over a range of design flow capacities. The
results of their studies indicated that RBC process,through-
out the range of design flow capacities they used in analysis
were less costly than activated sludge processes in supply
and construction as well as operation and maintenance. The
comparison of an extended aeration plant, which operates on
activated sludge principles, and RBC plant for the Makkah, 14
Saudi Arabia, municipality showed that the extended aeration
plant was about 70% more expensive to operate and maintain
than Aero-Surf.
     The particular problems of waste water treatment plants
for Saudi Arabia are related to such factors as lack of
skilled supervision, high operating temperatures, high rate
of expansion due to intensive development and urbanization,
recognition of scattered small communities such as the
presence of military cantonments, isolated camps, special
settlements etc, scarcity of water,  dry weather, supply of
local materials, availability of funds etc. All these
factors may be adequately dealt with by the application of
RBC system for the following considerations.

     a. The system claims the benefit of reliability without
        frequent supervision.
     b. The system does not depend substantially on oxygen
        dissolved in water which saturation concentration
        decreases as temperature rises.
     c. The system is susceptible to upgrading or extension.
     d. The system is well  established for small communities
        applicat ion.
     e. Due to scarcity of water,  recirculation of waste
        water may be required for supplementing drinking
        water supplies.  The system is capable of removing
        objectionable ammonia and  producing nitrate required
        for potable water.
     f. The system is capable of producing sludge of adequ-
                          141

-------
         ate quality liable  for  disposal  on  drying  beds which
         effectively operate in  dry weather  conditions.
      g.  The discs  which  cost makes a  large  proportion of  the
         system equipment costs  can be manufactured  locally
         because they ace of plastic material  which  is a petro-
         chemical product. Firms  such  as  SABIC(Saudi Arabian
         Basic  Industries Co), SAPPCO  (Saudi Arabian Plastic
         Products Co)  etc are wel1  established for manufacture
         of such product.
      h.  Funds  for  installation of  a technically  sound system
         are easily allocated and therefore  there is no neces-
         sity for application of  systems  operating on  interme-
         diate  technology principles or committing nuisance
         and obnoxious  to the community or restricting the
         freedom of expansion and advancement.

           DESIGN CRITER10R

     .One would  ask,  why  the RBC process  is not widely used in
•Saudi Arabia if it is  adequately acceptable for waste water
 treatment  and particularly  satisfies  the Kingdom requirements.
      Being comparatively young  in  the market,  in contrast to
 the conventional treatment  processes, use of  RBC process  in
 Saudi Arabia is hampered partly by  lack of adequate litera-
 ture  needed for engineers and consultants who handle waste
water treatment in the Kingdom and  partly by  manufacturers
who make little effort to pass on  knowledge and convey correct
 information. Absence of  such valuable  information may subject
 the RBC  process to reservation within the engineering commu-
 nity. Since waste  water  treatment  is a supporting facility in
most  large scale projects,  which are usually  awarded on lump-
 sum contract basis to  general contractors, its emphasis is
 not greatly aknowledged.
      The ambitious development programmes launched in Saudi
Arabia are so great  that justified  employment of multi-natio-
nals  who possess different  technical background and approach
and that coordination  between different organizations could h-
ardly be secured.Some  programmes have been rushed and their
periods  squeezed for time saving.  In absence of such a govern-
mental   body whose main  task would  be to furnish engineering
departments  of  all  organizations with sufficient and correct
data and monitor their performance accordingly, the performan-
ce of such departments will   depend  totally on the type,qua-
 lity and talent of personnel employed. In these circumstances'
 if personnel employed  are European  for example they will
                              142

-------
obviously choose European products since they are familiar
with them, the America! will choose American products etc.
     At present designs of waste water collection and dis-
posal schemes are generally based either on the existing
facilities with annual percentage growth factor, or on cui—
rent design criterior  In industrial countries.Neither of
these bases is satisfactory. Since literature and studies
leading to identification of such design criterior are
limited, it is insufficient for the engineer practising in
Saudi Arabia just to exercise his technical competence and
skill. His task has to be extended, within his own initia-
tives, to include gathering of information to arrive at
reasonable design criterior applicable to Saudi Arabia. His
success depends on the ability to take a positive interest
in this direction.

     CONCLUSIONS

     The RBC units have become widely accepted In recent
years, notably in the United States,  which is evidenced
by the great number of plants operating or under construc-
tion and by the influx of several new RBC manufacturers
competing into the market.  The unit suits best the require-
ments of Saudi Arabia and its manufacture can be carried out
locally.
     Engineers operating in Saudi Arabia should be encourag-
ed to keep their designs as simple as possible and to avoid
complicated features and sophisticated equipment for the
completed plants cannot operate satisfactorily without
skilled, talented operators particularly if they depend
substantially on mechanical and electrical  equipments.
     There is a need to establish an  organization whose
prime tasks would be to identify the  objectives of waste
water treatment, review waste water treatment processes
in application and recommend their use for varied purposes
in Saudi Arabia, Lay design criterior, conduct research
leading to development of suCto-ble processes as wel 1  as
therr local  manufacture etc.
                          143

-------
REFERENCES

1. Ainsworth, G. "The Activated Sludge Process"
   Water Pollution Control Engineering, 6, pp 60-74
   (H.M.S.O, 1970)
2. A.C.M., U.K. "Bio-DIs.c Process for small Community
   Sewage Purification". Patents granted or pending
   throughout the world.
3. Antonie, R.L. and Hynek, R.J. "Operating Experience
   with Bio-Surf Process Treatment of Food-Processing
   Wastes", Paper presented at the 28th Annual Purdue
   Industrial Waste Conference,  West Lafayette,Indiana
   May 1973-
4. Antonie, R.L. and Welch, F.M. " Preliminary Results
   of a Novel Biological Process for Dairy Wastes".
   Presented at the 24th Purdue  Industrial Waste
   Conference,  May 6th-8th, 1969.
5. Antonie, R.L."Applicatlon of  the Bio-Disc Process
   to Treatment of Domestic Waste Water"
   Federal  Water Quality Administration of the U.S.
   Department of Interior, Contract No.14.12.810.
6. Antonie R.L. "Response of the Bio-Disc Process to
   Fluctuating Waste Water Flows".
   Presented at the 25th Purdue  Industrial Waste
   Conference,  May 5th-7th, 1970.
7- Antonie, R.L."Rotating Disc Dual  Waste Water Role".
   Wat. Wasts Engng.  1971, 8NO.  1, pp 37~38.
8. Autotrol Corporation, U.S.A."!ndustrial Waste Water
   Treatment".  Bio-Disc Information Bulletin.
9. Autotrol Corporation, "Waste  Water Treatment Systems
   Design Manual".  Bio-Systems Division,  MilwagJ
-------
15. Ellis, K.V.and Bannaga,  S.E.I."A Study of Rotating
    -Disc Treatment Unit Operating at Different Tem-
    peratures". Journal of the Institution of Water
    Pollution Control, London, January 1976-
16. Hartmann, H. " A Dipping Percolating Filter Plant"
    G.P. 1,275, 967, Korresp.Abwarss, 1969, No.5, 102.
17. Hartmann, H. " The Dipping Contact Filter".
    Ost. Wasserw, 1965, 17,  264-269.
18. Lundberg, L.A. and Pierce, J.L.  "Comparative Cost-
    Effectiveness Analysis of RBC and Activated
    Sludge Process for Carbon Oxidation".
    Schneider Consulting Engineers,  Bridgville,
    Pennsylvania, U.S.A.
19- Quirk, T.P. and Hellman, J.  "Activated Sludge and
    Trickling Filter Treatment of Whey Effluents".
    Wat. Poll. Cont. Fed.  1972,  kk,  2277.
20. Simpson,  J.R. "Waste Water Treatment for Small
    Communities". Process  Biochem.,  1972, Vol. 7 18-
     21.
21. Simpson,  J.R. "Technical Basis for Assessing the
    Strength  Charges for Treatment and Treatability
    of Trade  Wastes".  Wa.  Poll,  Control,  1967, Vol
    66, pp 165-181.
22. Tropey, W.H., Hewkelekian, H., Kaplovsky, A. and
    Epstein,  L. "Effect of Exposing  Slimes on Rotat-
    ing Discs, to Atmosphere Enriched with Oxygen".
     Presented at the 6th  International  Water
    Pollution Research, June l8th-23rd,  1972.
23. Water Pollution Research Laboratory,  Stevenage,
    U.K. " Rotary Biological Contactors".
    J. Inst.  of Pub. Health  Eng., May, 1973(3)pp 116-
    130.
                    145

-------
      THE FUTURE OF BIOLOGICAL FIXED-FILM PROCESSES AND
         THEIR APPLICATION TO ENVIRONMENTAL PROBLEMS
    Stanley L. Klemetson.  Department of Civil Engineering,
    Brigham Young University, Provo, Utah.

    Gary L. Rogers.  Department of Civil Engineering,
    Brigham Young University, Provo, Utah.
INTRODUCTION
    The needs and design of municipal and industrial
wastewater treatment facilities will be much different in the
late 1980's than in the past.  Cutbacks in federal funding
and strained financial condition of local taxpayers will
require that less expensive and more reliable wastewater
treatment system*be built. While it was once desirable to
build significant excess capacity in new plants to improve
the economies of scale at lower current costs, the goal now
is to build smaller plants in hopes of technological advances
in the future. In many cases older plants are being upgraded
for both treatment efficiency and flow capacity.  Operational
costs are being carefully reviewed by both municipalities and
industries. The systems with the lowest energy, maintenance,
and labor costs will be chosen when possible.
    To appreciate the changes that will occur it is necessary
to review the past.  In both the United States and Britain
the development of biological treatment progressed from
sewage farms and discharge into waterways, through
intermittent sand filters and contact beds, to trickling
                            146

-------
filters, activated sludge basins, rotating biological
contactors, and land applications.  The terra Biological
Fixed-Film Process is really a new name for an old process.
As will soon be shown we have gone full circle and are
returning the semi-passive fixed-film systems.
EARLY DEVELOPMENTS
    The disposal of raw and partially treated sewage onto the
land was a natural consequence of good farming practices.
The sewage placed or poured on the fields from the urban
drainage ditches could irrigate the land and provide needed
nutrients to the soils and the crops.  Since disease control
was not a great concern in the nineteenth century, it was not
until land became scarce and the potential for profit from
sewage-irrigated lands was diminished that that sewage
farming was abandoned.
    In areas where sandy soils existed, the practice of land
application continued in the form of intermittent sand
filtration.  Dosages were continually intensified, and in
many cases, wastewaters were pretreated by use of settling
tanks or other biological treatment units. This practice
continued until more advanced treatment units were developed.
    In areas where tight clay soils existed, it was necessary
to build contact beds, which were relatively shallow tanks
containing many layers of slate supported on a layer of
bricks or filled with broken stone or slag. The original beds
used fill-and-draw and resting eyelet  The beds provided an
excellent site for large populations of microorganisms and
removed dissolved as well as suspended solids from the
wastewaters.  The efficiency of the beds were soon increased
by the addition of sprays which permitted application of the
wastewaters to the beds on a more continuous basis.  The
discharged wastewater was also saturated with oxygen.   This
modification no longer required that the beds be flooded, but
rather permitted oxygen to pass through the bed continuously
to keep it aerobic.   These improved beds were first called
bacteria beds, but this was later changed to trickling
filters.
    Much later the activated sludge process was developed,
and promoted as the ultimate in biological treatment.  Even
this method used a biological film provided by the raicrobial
floe, however, it is not considered a fixed-film process.
                            147

-------
    With the advent of PL 92-500, a variety of new processes
were developed to advance the wastewater treatment technology
available.  Innovative new processes were developed, not all
of which worked well.  The Environmental Protection Agency
provided funds to encourage the application of these
processes in new plants. Even industries were willing to try
some of these processes that would reduce their total costs.
    Among the developments were the high rate aerated contact
bed and the high rate anaerobic system  .  Both of these
systems are improvements of original contact bed. Both are
suitable for municipal wastes (1), but  the latter will see
its greatest use in treating high strength industrial wastes.
Specialized bacteria are also being developed to degrade
non-conventional 'organic industrial waste compounds in the
fixed film system (2).
    While each of the processes have gone through a variety
of revisions and updatings, about the last major different
type of treatment unit is the rotating  biological contactor.
In this system, the raicrobial fixed-film alternatively is
rotated into the wastewater and into the air.  While this is
based on the same principles as the trickling filter, it
should provide a higher level, of treatment in a smaller area.
PROCESS APPLICATIONS
     The fixed—film treatment systems have been applied
successfully to a variety of applications.  Its role in
wastewater treatment has been long anisuccessful.  Industrial
wastewater pretreatraent is more recent.  Anaerobic have
frequently been used for strong organic wastewaters.  The use
of anaerobic fixed-film filters is much more recent.
    With the tightening of effluent standards, the concern
for nitrification and denitrification became extremely
important.  Again, fixed-film systems proved very adequate.
    Water conservation efforts have required the evaluation
of many treatment methods.  In the power industry, fixed-film
systemshave been used to remove organic contaminates before
reuse within the plant.
    Another use that will receive increased attention is
aquaculture.  The wastewaters used produced by fish and prawn
farming must be treated before discharge. Also the internal
recycle in the ponds requires that harmful biproducts be
removed from the water on a continuous basis to maintain the
aquatic population.
                            148

-------
CURRENT APPLICATIONS
    The future of biological fixed-film system will be
discussed in the following sections in the context of their
advantages and disadvantages.
Introductory Comments
    The activated sludge process has, during the past fifteen
years, jreceived significant development to meet the needs of
the wastewater treatment industry.  If it were not for its
high power requirements and the current high costs of buying
this power, it is likely that this process would continue to
be highly favored.  However, the need for economies of
operation for treatment plants requires that alternative
treatment methods be considered in the design of new
treatment plants and the up grading of old plants. The U.S.
Environmental Protection Agency has also required this plan.
Land Applications
    Land applications of wastewaters, while not considered by
most of the industry to be fixed-film process, is a return to
the concept of the sewage farm.  However, the concern about
disease is very important now.  While raw sewage is no 'longer
applied directly to the land, treated effluents and partially
treated sludges are being applied.  The three methods of
application: Spray irrigation, overland treatment, and rapid
infiltration, are modifications of the fixed-film process.
In this case the biological growth occurs on the plant
structure or within the soil. This system has limited usage
in some regions of the country because of land costs, cold
weather operational requirements, and the loss of water
discharges to subsequent users.
                              149

-------
Coupled Trickling Filter/Activated Sludge Systems
    Efforts to upgrade existing wastewater treatment plants
have led the development of coupled trickling
filter/activated sludge treatment systems.  These systems
have the advantage of reducing future operating costs while
meeting required effluent limitations.  In some cases the
trickling filter is only for roughing to reduce the load on
the activated sludge.
    This system will continue to be built for upgrading of
existing plants, the multiplication of equipment for this
dual system does not recommend it for new small and medium
size wastewater treatment plants.  In large treatment
systems, it is quite possible that the operational advantages
and treatment efficiencies of activated sludge systems will
be combined with the economies of trickling filters (or other
fixed film system) to provide the least cost alternative
treatment system.
Trickling Filters
    Trickling filters have been subjected to a variety of
modifications to improve their operation.  They have
relatively low operating costs but suffer from Inadequate
removal efficiencies.  Probably the most significant recent
modification has been the introduction of plastic media.
While rock media was significantly affected by changes in
flow, the plastic media only requires a minimum quantity of
wastewater for wetting and nutrient source.  Beyond that
flowrate the variation in flow does not significantly affect
treatment efficiency. Trickling filters have, in the past,
been considered inadequate to meet effluent standards without
additional treatment.  Therefore some additional treatment
has been required.
    Trickling filters will increase in their importance in
the design of new wastewater treatment plants.  However, in
some applications, other fixed film biological treatment
system will be more applicable.  Among the limiting factors
will be loading rate and area requirements.
                             150

-------
Biological Towers
    A modification of Che trickling filter is the biological
tower.  Either plastic media or redwood can be used in the
system.  Either natural or forced aeration can be used,
depending upon the design of the unit.  Depending upon the
aeration requirements, the operating costs are low.
Additionally, the manpower requirements are low..   Very high
loadings can be applied to the filter and nitrification can
be achieved in some of the filters.  It is therefore possible
to achieve quite adequate treatment efficiencies.
    Biological towers will enjoy a strong role in the future
of wastewater treatment systems.  Industrial applications are
expected to be significant.  A number modifications will be
developed in the next few years that will make the system
more reliable for specific applications.
Biological Aerated Filter
    Variations of the biological contact basin have been
developed and will continue, to receive development.  These
systems have low area requirements and low capital costs.
About 10 years ago a fluidized activated carbon system was
developed by Weber (3).  There were other modifications,
including the addition of air and pure oxygen.  Each of these
systems had a variety of advantages and disadvantages.  A
more recent development was the Biological Aerated Filter
which uses a fixed bed of granular materials (2).  The basic
difference in the systems are structural and operational
modifications.
    The current systems will continue to experience
popularity in the future, however, wide—spread full-scale
applications will be slow in coming.  Special purpose and
industrial applications are and will continue to be the most
likely use.
                             151

-------
Anaerobic Filters
    Anaerobic systems have often been used for strong organic
wastewaters.  The high rate anaerobic filter, a modification
of the old contact bed system, effectively treats high
strength organic wastes.  It has a low operating cost, has
high removal efficiencies, but does require some post-
treatment system prior to discharge of effluent.  Among the
wastes that have been treated are food processing,
pharmaceutical, sugar, potato, and beet sugar.
    These systems will experience an increasing demand from
industry and little demand from municipal waste treatment
sys terns.
Rotating Biological Contactors
    The development of the rotating biological contactors has
opened up a new process of treating wastewaters in small to
medium size systems with a minimum of equipment or manpower.
Power costs are low for each shaft, and both rotation and
aeration can be achieved by using the air drives.  The
systems are quite suitable for upgrading existing plants by
adding the units to existing aeration tanks.
    The systems have moderate land requirements, but have
high capital costs.  While the concept is good, not all of
the manufacturers have produced good equipment.  There have
been a higher than expected number of equipment failures,
including shaft failures and media failures.  In addition,
the published design curves are unrealisticly low and promote
under design of treatment systems.  Once these difficulties
have been cleared up the systems have a good potential for
the future.
Summary Comments
    All of the systems can be compared on the basis of
loading rates and capital costs.  Starting with rock
trickling filters as having the lowest loading rate.
Improvements can be obtained by using plastic media.  At the
top of the loading scale, the anaerobic filters can receive
                            152

-------
the highest loading rates.  The other units fall in between
these limits.  The required areas are inversely proportional
to the loading rates.
    The biological anaerobic and aerated filters have the
lowest costs with rotating biological contactors being the
highest.  The other units fall in between these limits.
    The overall comparisonsare more clearcut than they should
be.  Each of the fixed film units have an optimum application
and optimum size of operation.  Each application will have to
be analyzed for specific needs and locations.
FUTURE DEVELOPMENTS AND NEEDS
    Some of the fixed-film biological treatment systems have
been unable to meet effluent standards.  The current push to
relax those standards to about 50 mg/1 BOD and no limit on
suspended solids on selected waterbodies has made the return
to trickling filters for the complete secondary treatment,
much more realistic.
    As the cost of energy increases, more treatment plants
will be equipped': with energy conserving equipment. This
requirement will mean that more fixed—film systems will be
constructed.  The trickling filter, with its low energy
requirements will experience continued improvements in design
and media to meet effluent requirements.  Combined Activated
Sludge/Trickling Filter systems will be built to reduce the
costs of operation.  Improvements in media design to improve
efficiency and to reduce cost will continue to be made.
Alternative media will have to be developed to sever the ties
to petroleum products.
    Rotating Biological Contactors, which still hold a future
promise of success, have several problems to overcome.  The
design curves need to made realistic so more success plants
can be designed and built.  The great, uneven, weight of the
rotating bioraass will require that design changes be made to
prevent failures of the shafts.  The media does not have the
lifetime necessary for economical operation at all plants.
These are challenges that need to be met since the system can
provide a low cost operation for both power and manpower.
    Biological towers, basicly a tall trickling filter, can
provide an economical operating unit for selected
applications.  The biological aerated filter will continue to
                             153

-------
be developed for specific applications.  It is unlikely that
it will be used as the sole biological treatment process for
mauy plants.
SUMMARY
    Biological fixed-film processes have been      around for
a long time.  While many of them have been placed on the
shelf for many years, their usefulness is being re-
established, and they are being used again as a viable method
of wastewater treatment.
REFERENCES

1.  Switzenbaum, M.S. and Jewell, W.J., "The Anaerobic
    Attached Film Expanded Bed Reactor for the Treatment of
    Dilute Organic Wastes."  TID-29398, National Technical
    Information Service, Department of Commerce, Springfield,
    Virginia, August 1978.
2.  Stensel, H.D., "Biological Treatment Systems for the
    1980fs."  Proceedings Utah Water Pollution Control
    Association Annual Meeting, April 16-17, 1982,  Salt Lake
    City, Utah, Ed. Stanley L. Klemetson, Brigham Young
    University, Provo, Utah, 1982, pp. 22-26.
3.  Weber, W. J., Jr., Physicochemical Process for Water
    Quality Control.  John Wiley & Sons, Inc., p 164,  1972.
                            154

-------
         PART III:  BIOFILM AND BIOKINETICS
     PROCESSES INVOLVED IN EARLY BIOFILM FORMATION
     James D.  Bryers.  Engineering Science, Swiss Federal
     Institute for Water Resources and Water Pollution
     Control (EAWAG),  Diibendorf  CH-8600  Switzerland.
INTRODUCTION
     Fundamental and applied research in fixed-film biologi-
cal processes has steadily progressed in the past ten years.
Atkinson and Fowler (l) review the significance of microbial
films in the fermentation industry while Cooper and Atkinson
(2) and Smith et.al., (3) provide state-of-the-art symposia
on fixed-film bioreactors In wastewater treatment.
     A large portion of this research has focused on the ma-
thematical description of substrate depletion within a bio-
film - i.e., "biofilm kinetics". Typically, such kinetic mo-
dels describe one dimensional mass transfer of substrate with
simultaneous biological reaction; the resulting differential
equation is

                  D d2S/dx2  =  -r.                   (l)

where S = substrate concentration in biofilm (ML  ), D = ef-
fective substrate diffusivity {L^t~l), rj. = intrinsic sub-
strate depletion rate {ML~3t~l), x = direction of substrate
                         155

-------
flux (L.). Solutions to Equation 1 depend upon (a) prevailing
boundary conditions and (b) the dependency of rj_ on substrate
concentration. Harremoes (U) and Riemer (5) provide excellent
reviews of the extensive literature on various solutions to
Equation 1. Those intrinsic kinetic forms assumed for ri most
relevant to sanitary engineers are the following:

     First order (ref. 6)   :  r. = k S                (2)

     Zero order  (ref, 7)   :  r. = k                  (3)
                               i    o
                                       1/2
     Half order  (ref. 8)   :  r± = ky2                ^)

     Saturation  (ref  1      r_ = k S/K   s          (5)
                  8, 9, 10,    i        s
                  11, 12)   :
     Unfortunately, these models only deal with substrate
removal kinetics and ignore biofilm development. Equation 1
tacitly requires that r. be either zero order in biofilm
concentration or, if first order, that biofilm mass remain
constant. Otherwise, an additional equation.describing bio-
film accumulation is required. Past works have either simply
ignored biofilm production (8) or assumed zero biofilm accu-
mulation by equating biofilm production to endogeneous decay
processes (11, 12). In most cases, processes governing bio-
film formation and, thus eventual fixed-film reactor per-
formance are neglected; consequently, important information
about reactor design, start-up procedures, and control of
biofilm thickness remains unknown.
Contributing Processes
     Biofilm net accumulation within a turbulent flow field
proceeds as shown in Figure 1. Five stages are evident: (l)
induction or lag, (2) exponential accumulation, (3) decrea-
sing rate, (k) plateau, and (5) sloughing. Processes involved
in this net accumulation can include (Figure 2):
                       156

-------
BIOFILM
                                                  sloughing
                exponential
                 growth
                         TIME
         FIGURE 1.  FIVE STAGES OF BIOFILM DEVELOPMENT

-------
en
Co
              FLOW
                             O
                                           "O"
       O
                  OCELL
                              T
 O sr&Tfrf\\~~P~ 'Pr-Jf



Pfoyuyimftfti^ytfttt^
wtyXJpyHp^j&fiqjtyiupM.
                       INERT SURFACE
             1.ADSORPTION

             2.TRANSPORT

             3.ATTACHMENT
        4.GROWTH

        5.REENTRAINMENT
            FIGURE 2. PROCESSES INVOLVED IN BIOFILM ACCUMULATION.

-------
     1. adsorption of organic molecules to the wetted surface.

     2. deposition of bacterial cells to the organic-treated
        surface. Deposition rate can be considered the sum
        of bacterial cell transport and cellular attachment
        rates.

     3. cellular growth, reproduction, and extracellular
        polyme r format ion.

     h, detachment of biofilm and entrainment of debris into
        the fluid.
Trulear and Characklis (13), Bryers (lk), and Characklis (15)
provide extensive reviews on these processes and their in-
volvement in fouling biofilm development. This paper will pre-
sent methodology used to quantify the physical transport and
microbiological processes involved in early biofilm formation.
EXPERIMENTAL PROTOCOL
     Individual processes, and thus, net biofilm developement
 are  considered in this study to be functions of the follo-
wing:

     1. prevailing hydrodynamic conditions - i.e., linear
        velocity, shear stress at the wetted surf ace,or Reynolds
        number, Re.

     2. concentration of bacteria suspended in the ambient
        fluid, X.

     3- metabolic activity of suspended bacteria as indicated
        by their growth rate, y.

     U. biofilm concentration as COD mass per area, B.

     Consequently, the laboratory reactor system shown in
 Figure 3 was employed for it allowed continuous surveillance
 of biofilm development under defined conditions of Re, y,
 and X. The system  is operated as two completely stirred tank
                         159

-------
cr>
o

                                                                 •«	 NUTRIENTS


                                                                 «	 DILUTION WATER
                                                                     THE CHEMOSTAT
                                              -NUTRIENTS



                                              	DILUTION WATER
                                           F,X°,8«

c
r


2,Sj


/


'I 'T 	
! i
By Pass ' 	
1
1 SAMPLER S-1
-n — L—H 	
SAMPLER S-3
	 ^
1
	 AP 	 i

SAMPLER S-2
71 	 1 — 	 	 	 	 j 	 J
                                                       THE BIOFILM REACTOR
                           FIGURE 3.   REACTOR SYSTEM DIAGRAM. CSTR 1 OPERATED AS A CHEMOSTAT WJILE

                                      CSTR 2  WAS THE BIOFILM REACTOR. OPERATING CHARACTERISTICS

                                      GIVEN IN TABLE I.

-------
                                         Table I.
                      Pertinent  Characteristics of CSTR 1 and CSTR 2.
                                       CSTR l:
                                       The Chemostat
                                                                   CSTR 2:
                                                                   The Biofllni Reactor
System_S£e£lfi£s

Reaction Volume (cm3)
Total Wetted Surface Area Ccm2>
Surface Area:  volume (cm~l)
Dilution Rate (h~l)
Mean Residence Time (h)
Dilution Water Flow (cm3h-l)
Effluent Flow Rate (cm^-1)

£ro_-«th_S£ec_l_fi£s

Inlet Substrate
  TS3:  Glucose (wt; wt)

  Combined Concentration
    (rug H)
    (mgCOO 1~1)

Mlcrooial Feed
Temperature  (°C)
                                         3000
                                         1070
                                         0.36
                                         0.33
                                         3.0
                                         998
                                         1000
                                         9:1
        1000
         850

Initial inoculation
with heterogeneous
population

        31
pH
                                          8.1
       , L_oop_S£e£l£l£s_(CSTR_2_onl^)
                                     4750
                                     5934
                                     1.23
                                     4.0
                                     0.25
                                     13000
                                     1SGDO
                                                                      1:1
                                                                      20.0
                                                                      23.0

                                                                   CSTR 1
                                                                   effluent
                                     31

                                      7.3
Recycle Loop Tube length (cm)
Inside Tube Diameter (cm)
Recycle Reynolds Number
Recycle Flow Rate (cm3 - s~l)
Recycle Velocity (cm - s~l)
Test Sections
Length (cm)
Inside Diameter (cm)


13000
104
82
SI
91.4
1.27
1219.0
1.27
26300
203
164
32,3
ie-4
1.27
 Sample  Tubes  (52,3)
   Number
   Length/Tube (cm)
   Inner Surf5ce Area/Tube
   Total Sampling  Surface Area  (cm2)
   X Sampling  Area of Total  System  Area
                                                                         40
                                                                         5.2
                                                                        20.7
                                                                       330.0
                                                                        14.0
                                   161

-------
reactors (CSTR), in series, such that operation of the first
reactor  is independent of the second. The first reactor
(CSTR l)  is a conventional chemostat. The second reactor
(CSTR 2)  is a tubular reactor with internal recycle flow.
Recycle flow rate in the CSTR 2 tube  is far greater than
the volumetric flow rate of influent to CSTR 2. The tubular
geometry of CSTR 2  is choosen to simulate biofilm develop-
ment under known hydrodynamic conditions.
     6STR 1 was operated at a residence time of 3 h and
serves only as the source of suspended biomass for CSTR 2.
CSTR 2  is operated at a residence time of 0.25 h» consequent-
ly, the majority of biological activity in CSTR 2  is due to
biofilm development.
     CSTR 2 contains  two sampling sections which allows  for
periodic determination of biofilm accumulation as COD mass
per area. A third section of the tubular reactor serves to
monitor the increase in frictional resistance due to biofilm
development (15). For this study, the early biofj.iLm, formation
period is_defined as that amount of biofilm accumulated prior
to any increase in frictional resistance.
     Table I summarizes pertinent operating characteristics
of both reactors. Details of the reactor system, start-up
procedures, sampling and analytical methods are provided else-
where (lU, 16).
     Experimentation  is divided into two parts:
     Series I.  development of an empirical rate expression
                describing net biofilm development as a
                function of Re, x, and p.
     Series II. Estimation of the relative magnitudes of in-
                dividual  processes contributing to net bio-
                film development and the effect of Re and X
                on. those magnitudes.
RESULTS

Series I

     Conditions for Series I experiments are given in
Table II.
                         162

-------
     Table  II.   Series  I   Experimental  Conditions,
Experiment X
Number Group (mg-TSST1)
1 Biomass
2
3
4
5
6
7
8
9 Reynolds
10 number
11
12
13
14 Suspended
15 growth rate
16
17
18
4.4
1 2.0
2.8
13.0
23.0
4.0
10.1
2.5
12.0
12.0
12.0
12.0
12.0
18.0
18.0
18.0
18.0
18.0
ft
(h-1)
0.28
0.28
0.28
0.28
0.28
0.28
0.28
0.28
0.28
0.28
0.28
0.28
0.28
1.0
0.28
0.16
0.13
0.13
Re
17.200
17,200
17.200
17.200
17.200
17,200
17,200
17,200
17,200
10.600
19,300
23.900
28,800
17.200
17.200
17.200
17.200
17,200
Biofilm development, as COD mass per area, is shown in
Figures Ua-c as a function of either Re, X, or y. Results
in Figures Ua-c suggest a first order rate expression of the
form:
                      as/at  =
                               (6)
The numerical value of the rate constant for biofilm net
accumulation, k ,   is determined from statistical regression
of B vs time according to the integrated form of Equation 6.
These values of kjj are illustrated for each experimental
group in Figures 5a-c.
     The net accumulation rate constant, kN is actually a
function of the three parameters  considered:
=  k.
                                Re
(7)
                                                    -1,
where kM = biofilm net accumulation rate constant (t  ), k.=
intrinsic biofilm accumulation rate constant, and (a, b, c) =
empirical constants. Linear regression of data in Figures
5a-c provides the following estimates of the empirical con-
                         163

-------
                        FIGURE 4A. BIOFELM ACCUI-IJLATION
                        AS A FUNCTION OF SUSPENDED
                        BIOMASS. X = 23.0 mg/1 (  • ),
                        12.0 rag/1  (<\7), and 2.4  mg/1
                        ( • ) . Re = 17,200 and p =
                        0.28 h"1.
                        (CURVES       ON EQ. 8 )
20    «o    eo    BO
   TlME(h)
20    4O    60
    TIME Ch!
                   too
                        FIGURE  4B.  BIOFUM AOCUMULAIION
                        AS A FUNCTION OF REYNOLDS
                        NUMBER.   Re = 10,000" ( Q ) ,
                        17,200  (*  ), and 23,900 (^).
                        X = 12.0  mg/1 and p = 0.28 h"1.

                         (CURVES BASED ON EQ. 8)
4O    ao    120
    TIME {h}
               160   200
                        FIGURE 4C.  BIOFIM ACOJMULATIOii'
                        AS A FUNCTION OF SUSPENDED
                        BIOMASS GROWTH RATE. Ji = 1.0
"1
                                    0.28 h
                                          "1
                         ,and
                               -1
                         0.13 h - ( * ). X = 18.0 rog/1
                         and Re = 17,200.

                         (CURVES BASED ON EQ. 8)
                     164

-------
ooe -
006 -
OO4 -
002
  0       6      12      18    24
  DISPERSED BIOMASS CONCENTRATION x
             (mg TSS I'1)
FIGURE  5A.  ACCUMULATION
RATE CONSTANT, kn,  AS A
FUNCTION OF SUSPENDED BIO-
MASS CONCENTRATION.

Re= 17,200  and y =  0.28 h"1
005
   10
         IS
               20
                            30
    Reynolds number x 10~3
                                  FIGURE 5B. ACCUMULATION
                                  RATE CONSTANT,  k , AS A
                                  FUNCTION OF  REYN8LDS NUM-
                                  BER.

                                  X=.18 mg/1 and  y = 0.28  h
                           -1
 010
 005
  O           OS           10
    DISPERSED BIOMASS GROWTH RATE./i
FIGURE  5C. ACCUMULATION
RATE CONSTANT, k  ,  AS A
FUNCTION OF SUSpiNDED BIO-
MASS GROWTH RATE.

X=  18 mg/1 and Re=  17,200
                           165

-------
stants: a = 1,0, b = -1.0, and c = 1.0. Once the empirical
constants a, b, c are known, the intrinsic rate constant, kj_,
can be calculated from any set of known experimental condi-
tions (l6). The resultant integrated form of Equation 6 can
be written as follows:
     B(t)  =  B0exp [ (kXM/Re) • t ]                (8)
•where k£ = 125.0 * 25 mg TSS-1!*1 and Bo = biofilm COD per
area at time zero (ML~2) . (Range of Bo observed was 0.5-1.0 ug
COD cm"2).
Series II

     Details of CSTR 1 operation are given elsewhere (lU)
and are only summarized here in Table III.
Table III. Operational Results of CSTR 1, Series II Experi-
           ments.

Duration of CSTR 1 continuous operation =  kh days

Dilution rate                           -  0.33 h~
                                                           — "1
Inlet soluble COD concentration         =  6UO-850 mg COD 1

Effluent total COD concentration        =  390-^00 mg COD 1~

Effluent soluble COD concentration
prior to dilution water to CSTR 2       =   Uo-50  mg COD 1~
                                                mm "i
Dilution rate at culture "wash-out"     =  2.2 h

Biomass yield (g biomass/g-COD)         =  O.U2-0.56

Microorganisms present : Klebsiella oxytora, Klebsiella
pneumoniae, Enterobacter cloace.
     Operating conditions for CSTR 2 are  given in Table IV.
Inlet flow to CSTR 2 consists  of dilution water, fresh ste-
rile substrate, and CSTR 1 effluent. Primary substrate
fed to CSTR 2 in all experiments  is 10 mg 1~1 trypticase
soy broth and 10 mg 1~1 glucose, after dilution.
                         166

-------
 Table IV.  Operating Conditions of  CSTR 2.  The  Biofilm Reactor,
                                EXPERIMENT

                                         2.            3
  CSTR 1 Effluent
  Suspended BioiMss
  Concentration
   (mgCOO l-l)               370.            35?.   .        368.         380.

  CSTR 1 Effluent
  Delivered to CSTR 2
   (cm3 h-1)        '        1000.           1000.          1000.         200.

  Measured Freshl
  Inlet Suostrate
  Concentration
   (mgCOO 1-1)                22.7   ..       22.9   .       33.0         23.8

  Measured Inlet
  Suspended Bicmass
  Concentration
(mgTSS l-l)
(mgCOO 1-1)
Reynolds Number
19.5
22.2
13000.
18.9
21.6
13000.
19.4
22.1
26000.
4.0
4.6
13000.
  Mean Residence Time
     (h)       .              0.25           0-25          0.25         0.25


  1 Fresh  substrate delivered to CSTR 2 consisted of 10 mg 1-1 TSB and 10 mg I"1
  glucose.  Reported concentration Is after dilution with CSTR 1 effluent and fresh
  dilution
 Materi-al Balances                          •  '


      Presentations  of Series II  experiments  is facilitated by
 material balances for substrate  and suspended biomass as  well
'as a constitutive equation for "biofilm accumulation:


 Substrate:             V dS/dt  =   F(S.-S) -   yVX/Y-R A/I    (9)
                                          1                 B,

 Suspended Biomass:    V dX/dt  =   F(X.-X) +   uXV+R  A - R A (10)
                                          i              r     d

 Biofilm:               [ dB/dt   =   R  +  R  - R ] A            (ll)
                                        Q     *•*     •*
                              167

-------
vhere S = substrate concentration measured as COD (ML  ),

X = suspended biomass concentration measured as COD (ML  ),

                                        —2
B = attached biomass measured as COD (ML  ), t = time (t),

S, = inlet substrate concentration measured as COD (ML  ),

                     3                              2
V = reactor volume (L ), A = reactor surface area (L ), F =

volumetric flow rate (L t  ), y = specific growth rate of

suspended biomass (t  ), Y = biomass yield measured as COD

(MM  ), R  = net biofilm production rate due to metabolic
         6
                             _2  — i
processes measured as COD (ML   t  ), R, = deposition rate of •
                                       d
                                     _2 —j_
suspended biomass measured as COD (ML  t  )» R  = detachment
                                   _P 	i
rate of biofilm measured as COD (ML  t  ).

These material balances can be simplified with the following
assumptions:

1.    Rates of accumulation of S and X (i.e., dS/dt and dX/dt)
      are negligible and the system can be considered at steady
      state.
2.    Although an increase in suspended cell numbers is un-
      likely at residence times of 0.25 h, increases in sus-
      pended biomass concentration may be significant. Con-
      sequently, suspended biomass growth in CSTR 2  is not
      ignored.

3.    Substrate depletion rate by suspended biomass  is also
      considered significant in CSTR 2 (see Assumption 2).

k.    Net biofilm production rate  is assumed the sum of bio-
      film production processes (i.e., growth of organisms
      and product formation) and maintenance energy require-
      ments, i.e.,
                          168

-------
                    R   =  (y  - k ) B                ,(12)
                     g       p    e

where  y  = specific biofilm production rate (t  ), k  = decay
        p     _T                                     e
       rate (t x).

Equation 12 tacitly assumes specific biofilm production and
decay rates are first order in biofilm accumulation. Conse-
quently, Equations 9,10, 11, and 12 reduce to the following:

                 F(S.-S)  =  (u BA + pXV)/Y           (13)
                    i          p
                 F(X.-X) + (uXV) = R A - R A          (lU)
                    i               d     r

                 dB/dt = (p -k ) B + R,- R            (15)
                           p  e       d   r
Determining Individual Process J?ates

     Equation 15 describes biofilm accumulation throughout an
experiment as the sum of four processes: biofilm production
(UpB), biofilm mainenance decay (keB), suspended biomass depo-
sition (Rfj), and biofilm removal  (Rr). However, analytical
methods provide for direct measurement of only biofilm accu-
mulation - e.g., B and dB/dt, and the decay rate, k  .
                                          *        S
     Consequently, changes in CSTR 2 experimental conditions
 are made'"periodically to simplify Equation 15. These pertur-
bations consist   of depriving CSTR 2 of inlet substrate and/
or  inlet suspended biomass during four two—hour periods in
each experiment. Figure 6 details these perturbations and
their intended purpose. This technique allows estimation of
the following:

1.    Suspended biomass  .deposition rate (R^) on the "clean"
      surface at time equal zero. In further calculations,
      _Rd_ i_s_ jtssumed constant and independent of biofilm
      aecumulat ion.

2.    Equations 17 and 18  (Figure 6) can be used to  calcu-
      late  the biofilm removal rate, Rr, at each perturbation
      period knowing values of k  , R  and the slope  of the
      biofilm accumulation curve  (i.e., dB/dt).
                         169

-------
FIGURE 6.  DEFINITION  AND SIGNIFICANCE OF PERTURB-
             ATIONS TO CSTR 2.
                                     NORMAL  OPERATION
                                     FRCSH SUBSTRATE, SUSPENDED
                                     BIOMASS, AND DILUTION WATER
                                     SUPPLIED TO CSTR 2. BIQFILM
                                     NET ACCUMULATION DESCRIBED
                                     BY:
                                      dB/dt =   R, + (u -k )B  - R   (is)
                                               d    ' p  e     r
	 inflow 2. U *
cell
» 0 "0
1 R 1 -S>X, fc
I d I organics
.O^. .* ~. ^.O~~- ~T~

0
v
o

V_-O,^

«.
	 .-flow t O
U
M °^
o
V
_Rr
i
O^Nr-7^0^1- V'-^T1 — ' ••v"OC'/*.


	 »- flow O^
»\ o*- ^ o
1R4 ° ^
3^^>P7rC-"^o'
» -c O
Rr,
^-^-^-

                                     PERIOD  1
                                     ELAPSED TIME = 0-2 hours.
                                     NO SUBSTRATE TO CSTR 2, ONLY
                                     DILUTION WATER AND SUSPENDED
                                     BIOMASS, EQUATION (15 ) REDUCES
                                     TO:
 dB/dt =   R,
           d
PERIOD  2
                                                                 (16)
                                     ELAPSED TIME = 18-20 hours.
                                     NO SUBSTRATE TO CSTR 2, ONLY
                                     DILUTION WATER AND SUSPENDED
                                     BIOMASS. EQUATION ( 15 ) REDUCES
                                     TO:
                                      dB/dt = R. - k B - R        (17)
                                               a    e     r

                                     PERIOD  3
                                     ELAPSED TIME = 40-42 hours.
                                     SAME CONDITIONS AS PERIOD  2
                                     EXCEPT BIOFILM ACCUMULATION
                                     IS GREATER.
                                                       - R        (17)
      flow
                                     PERIOD  4
                                     ELAPSED TIME = 50-52 hours.
                                     NEITHER SUBSTRATE NOR SUSPENDED
                                     BIOMASS TO CSTR 2, OJJLY DILUTION
                                     WATER,  EQUATION (15 )  BECOMES;
                                      dB/dt «  -R  -k B
                                                r   e
                             (18)
                                   170

-------
     Figure 7 indicates biofilm. COD accumulation including
perturbed and non-perturbed intervals for a typical Series II
experiment. Biofilm accumulation, dB/dt, during the perturba-
tion periods only are presented, for all experiments, in
Table V. Values of the biofilm decay rate, kg, determined via
respirometer measurements (1^), are also given in Table V.
     Calculations of Rr in Periods 2-U, from data .in Table V
and Equations  17  and 18, are summarized in Table VI.
Figure 8 illustrates resultant Rr values versus the average
biofilm COD present during the perturbation. Data from Trulear
and Characklis (13), obtained from an annular rotating reac-
tor, are also included and indicate a similar magnitude of
biofilm removal rates.
     Specific biofilm production rates, \i , throughout the un-
perturbed portion of each experiment ,canbg determined using
Equation 15 and the values R^, ke, and Rr above. Resultant Mp
values as a function of biofilm COD are shown in Figure 9.
 DISCUSSION

 Deposition

      Rate of deposition,  R(j, was  considered  constant  through-
 out  any experiment  at  the value of  dB/dt  determined during
 Period 1 (ref.,  eq..  16,  Figure 6).  This  assumption provides  an
 estimate of deposition  rate at "clean"surface  conditions and
 most likely underestimates the enhanced effect a fouled sur-
 face would  have on  particle deposition at later stages  of
 biofilm development.
      Adsorption of  organic molecules  (e.g.,  polysaceharides
 and/or glycoproteins)  can contribute to the  total  amount de-
 posited (and rate of deposition)  as detected by COD analysis.
 However, this adsorption  occurs within minutes of  exposure
 (17) and the maximum amount of adherent material due  to orga-
 nic  adsorption  in this  system is  estimated - 0.01 ugCOD cm~l,
 Consequently, rates of  organic adsorption are  assumed  in-
 stantaneous and independent of Reynolds number (Re) and sus-
 pended biomass  concentration (X).
      Mass flux  of particles, suspended in a  turbulent flow
 field, across a boundary  layer is directly proportional to
 the  bulk fluid  concentration of particles (l8, 19, 20).
                          171

-------
     0
10
20
30    40
TIME  (h)
FIGURE 7.  BIOFILM NET COD MX1MJLATION DURING EXPERI-
           4 INDICATING BOTH NORMM, GKMTH AND FOUR
           PERTURBATION PERIODS.
                       172

-------
                         Table V.  Biofilm Accumulation during CSTR 2 Perturbations.
CO

Experiment One
Period 1
2
3
1*
Experiment Two
Period 1
2
3
1*
Experiment Three
Period 1
2
3
1+
Experiment Four
Period 1
2
3
!*
(l) average biofilm
. only.
(ygCOD cm )

1.2
ND
63.5
1*8.3

1.1
7.3
38.8
57.5

0.8 -
2.5
38.5
61.0

0.3
2.8
28.0
33.0
during perturbation
dB/dt(2l k B
(pgCOD cm h ) (pgCOD cm h )

1.2
ND
8.0
-5.3

1.1
-2.1*
-2.9
-2.5

0.8
-o.i*
0
-10.8

0.3
-2.2
-3.0
-6.0
(2) accumulation

ND
ID
0.38
ND

NKD
ND
0.23
ND

ND
ND
0,22
ND

ND
ND
0.17
ND
during perturbation period

-------
              fable VI.  Summary of Biofilm Removal Rate Calculations,
B _2
(ygCOD cm

Experiment 1
Period 1
2
3
If
Experiment 2
Period 1
2
3
1*
Experiment 3
Period 1
2
3
It
Experiment It
Period 1
2
3
it


1.8
ND
63.5
tt8.3

1.1
7.3
38.8
57.5

0.8
2.5
38.5
61.0

0.3
2.8
28.0
33.0
V
) as/at


+1.2
ND
+8.0
-5.3

+1.1
-2.it
-2.9
-1.lt

0.8
-0.lt
0
-10.8

0.3
-2.2
-3.0
-6.0
Rd
(a)

1.2
1.2
1.2
(e)

1.1
1.1
1.1
(e)

0.8
0.8
0.8
(e)

0.3
0.3
0.3
(e)
(VigCOD em"2!!""1)
k B R
e r
(*)

(d)
ND
0.1*
0.3

(d)
O.OU
0.2
0.3

(d)
0.2
0.2
O.lt

(d)
0.02
0.2
0.2
(c)

0
ND
-7.2
5.0

0
3.5
3.8
2.2

0
1.2
0.6
10.it

0
2.5
3.1
5.8
ND = not determined; (a) = deposition rate assumed constant at value of dB/dt determined
in Period 1; (b) = biofilm specific decay constant ke= 0.006 h~^-for all experiments;
(c) calculated from Equations 17 or l8(see Figure 6); (d) assumed zero during Period 1;
(e) assumed zero during Period it.

-------
    12.
'E
o
Q
O
O
CD
=L


c:"

UJ


a:
 o
 2
 LLI
 O
 m
    10.
     8.
     6.
     2-
         EXPERIMENT   SYMBOL
    1

    2

    3

    4

DATA FROM
TRULEAR AND
CHARACKLIS
O
o
a
A
                  13
                                               -T-"
              10.     20.     30.     40.     50.

                       BIOFILM. B (jagCOD cm"z)
                                        60.
                                70.
FIGURE 8.   BIOFUJV1 REMOVAL RATE, R  , AS A FUNCTION

             OF BIOFILM COD.
                       175

-------
     0.8
     0.7
~   0.6
m
j-
2
g

o
3
Q
O
CE
Q.
LL
g
m
     05
0.4
0.3
02
     0.1  .
                                     EXPERIMENT   SYMBOL
                                     1

                                     2

                                     3
                                     4
D

A
              10.      20.    30.    40.    50.    GO.


                               BIOFILM. B (jigCOD cm"2 )
                                                   70.
         80
FIGURE 9.  BIOFUM SPECIFIC PRODUCTION RATE, p. ,  AS A


            FUNCTION OF BIOFILM COD.
                            176

-------
A reduction in suspended biomass concentration from 19-5 "to
U.O mgTSS-l"1(Experiment 1 and U, respectively) does  result
in a proportional decrease of  1.2 to  0.3  UgCOD cnf"2h~lin de-
position rate, R^ref.Table" VI,'  Period  1  data).
     The effect of changing fluid flow regime on particle
transport is a complicated function of fluid velocity, fluid
properties, particle size and particle physical properties.
Increasing fluid velocity can have the following two effects:

     1. Increased turbulence may increase or decrease the mass
        transfer coefficient depending on characteristics of
        the suspended particle (19).
     2. Increased turbulence may decrease the boundary layer
        thickness and, thus, increase transport to the sur-
        face.

Suspended biomass generally has  specific  gravity less than
1.1 and suspended biomass aggregates  in these experiments
measured 3.0 - 5.0 um in equivalent diameter. Consequently,
the suspended biomass particles  are.  assumed uniformily distri-
buted and concentration gradients did not exist in the bulk
fluid. Table VI, Period 1 data shows  that deposition rate R^
decreases only slightly with a doubling in recycle Reynolds
number. In experiments at Re = 13000  (0.8 m/sec), deposition
rate was 1.1 - 1.2 ygCOD cm~2h~l while at Re = 26000  (1.6
m/sec), Rd was 0.8 pgCOD cm~2h~l, yet   Beal  (19) pre-
dicts  an increase in tranport  rate. This  discrepancy arises
since deposition rate is the sum of both  the particle trans-
port rate and microbial cell adhesion rate. Therefore, while
the particle transport rate may  be increasing with Re, the
deposition rate (the measured  parameter)  may not; suggesting
that cell "sticking efficiency"  is changing with Reynolds
number.
Biofilm production

     Figure 9 gives values of  up throughout each experiment as
determined from Equation  15. In all  cases,  up asymptotically
decreases with increasing biofilm  COD  to the  same value,
0.1-0.2 h~l. Wide variations in u  initially  may result from
errors in biofilm measurements at  very low levels or changes
                          177

-------
in. cellular metabolism upon attachment. Decreases in up with
increasing biofilm COD could result from changes in cell me-
tabolism or increasing internal resistance to substrate mass
transfer within biofilm.
     Instantaneous yield values in all experiments were cal-
culated from the substrate material balance, Equation 15 or
upon rearrangement:

               Y  =  (y BA +  yXV)/F(S.-S)             (19)
                       P              i
which tacitly assumes  that yield coefficients for attached
and suspended growth are the same. A summary of yield calcu-
lations is given in Table VII and suggests an average yield
of approximately 0.5 nig COD biomass per mg COD removed. This
value compares favorably with those obtained by Stathopoulos
(21)for similar experimental systems.
Biofilm Decay

     The spec
corresponds to values reported by Lawrence
The specific biofilm decay'rate is  0.006 h   and
and MeGary (22) for suspended biomass (0.0019-0.22 h  ) and
Stathopoulos (21) for biofilm experiments at temperatures
ranging from 15-60°C (O.OU-0.22 h"1).
Biofilm Removal

     Biofilm removal rates, due to existing shear stresses,
are shown in Figure 8 as a function of biofilm COD. Over the
range of biofilm COD observed, biofilm removal rates, Rr, were
less than 5 ygCOD cm~2h-1 and appeared independent of both
Reynolds number and suspended biomass concentration. This is
true except for the R  value determined in Period U of Expe-
riment 3; at Re = 26000 the removal rate suddenly increases
from 1.0 ygCOD cm~2h~l at a biofilm COD = U5 ygCOD cm"2 to
10.8 ygCOD cm-2 h"1 at a biofiljn COD = 6l ygCOD cm~l.
     This increase in removal rate can be explained by con-
sidering the changes in hydrodynamic  conditions that occur-
between these two levels of biofilm COD. Biofilm levels of
U5 and 6l ygCOD cm~2 correspond approximately to biofilm
                          178

-------
                           Table VII.  Summary of Calculations for Yield Coefficients.
U3
time F(Si-S)
(h) (mgCOD h"1]
Vp(1>
\ (mgCOD cm2) (h"1)
\i Bh X UXV<2>
P _1 _T
(mgCOD h ) (mgCOD 1 } (mgCOD h
Y(3)
-1) (mgCOD/mgCOD)
Experiment 1
0
10
34
50
Experiment 2
0
2
6
20
46
Experiment 3
6
18
2*
30
50
Experiment 4
6
20
44
50
51
52.7
79.6
150.6
149.6

0
96.9
101.1
103,4
208.4

81.6
81.6
84.0
96.2
106.4

14.3
31.4
110.5
106.8
89.9
.001
.005 •
.019
.035

.001
.003
.005
.007
,030

.005
.005
,006
.011
.055

.001
,002
.020
.035
.033
.70
.40
.20
.13

.10
.25
.47
.45
.18

.35
.35
,32
.25
•16

.75
.70
.24
.24
.24
0.5
11.8
22.5
27,0

0.5
4.4
13.9
18.7
32.0

10.4
10.4
11.6
16.3
52.2

4.5
8.3
28.5
49.8
46.9
18.7
24.0
28.9
26.5

13,9
25.5
24.
22.5
25,0

22.7
22.7
22.1
22.0
22.5

3.0
4.2
12.0
12.0
7.0
29.3
37.6
45.3
41.5

29.6
39.9
37.6
35.3
39.2

34.6
34.6
34.6
34,5
35.3

4.7
6.6
18.8
18.0
11.0
.57
.62
.45
.46

—
.46
.51
.52
.34

.55
.55
.55
.53
.42

.64
.47
.43
.64
.64
                     (l) biofilm production rate constant  taken from Figure 9 at specific biofilm COD.
                     (2) Suspended biomass growth rate, y, assumed  value =0.33 h~^. (3) From Eqn. 19.

-------
thickness of 39-5 and 53.0 ym, respectively (where 1 mg bio-
film = 1.7U mg biofilm COD and biofilm density = 10.0 mg bio-
film cm~3) (lU). A viscous sublayer thickness, 6, of UU ym
can be calculated as follows (15):

                      6 = 25 d (Rep875              (20)

with d = pipe diameter (1.27 x ICr ym) and Re = 26000. This
calculation indicates the biofilm thickness, just prior to
Period U of Experiment 3,exceeded the viscous sublayer, there-
fore, increasing the system friction factor and, consequently,
the shear stresses at the biofilm-fluid interface. This in-
crease in shear stress could result in the dramatic increase
in biofilm removal rate. Viscous sublayer thickness at
Re = 13000 is 80 ym and  biofilm thicknesses  in the three  expi-
riments at Re = 13000 never exceeded 55 Mm; consequently, no
radical increase in biofilm removal rate  is  expected and  none
are  observed.
     During Period U in all cases, -suspended biomass from
CSTR 1 is  not supplied to CSTR 2. Therfore, after this
two hour period (eight CSTR,2 reactor residence  times) any
suspended biomass leaving the system must originate as
biofilm.  Consequently, the rate of any suspended biomass  lea-
ving CSTR 2 after this perturbation  is  considered equal  to
the rate of biofilm removal - i.e.,

                       R  = FX/A                      (21)
                        r

Table ¥111 indicates the biofilm removal rates determined from
biofilm COD (Equation 18) are somewhat less, but of the same
order, as values determined from the suspended biomass mate-
rial balance, Equation 21.
                            180

-------
Table VIII. Comparison of Biofilm Removal Rate Estimates


               CSTR 2 Effluent     R  (Eq.2l)    R (Eq.18)
                 Biomass (a)
Exp . No . Re
3 26000
U 13000
( ugCOD 1 -1)
5358
1938
(ygCOD
16.3
5.9
cm"2 h"1)
10.8
5.8
(a) determined after eight residence times from start of
    Period U.
                               —1                2
(b) evaluated using F = 18 1-hr   and A = 593^ cm
SUMMARY
Series I Experiments.

     Results in Series I experiments provide the following
in format ion:
     1. the rate of biofilm COD accumulation during its ear-
        ly formation stages was described mathematically
        using a first—order rate expression. The resultant
        first—order rate constant was a linear function of
        suspended biomass concentration and growth rate, and
        Reynolds number
Series II Experiments.

     Results of Series II work, for the thin aerobic biofilms
investigated, show the following:
     1. Although particle deposition contributes significantly
        to the initiation of biofilm development, its relative
        role in biofilm net accumulation decreases with time.
                        181

-------
     2. Biofilm decay due to mainentance requirement is  in-
        significant for the thin biofilms considered.

     3. Biofilm production and shear removal processes con-
        tribute significantly to early biofilm accumulation.
        Shear removal rates drastically increase as the film
        exceeds the viscous sublayer.
ACKNOWLEDGMENTS

     This work was carried out at the Department of Environ-
mental Science and Engineering, Rice University, Houston,
Texas. Preparation of the manusript was supported by the
EAWAG and the typing quality is due solely to Fr. Frieda
Schlumpf.
REFERENCES

1.   Atkinson, B. and Fowler, H.W. "The Significance of
     Microbial Films in Fermenters" Chap 6. Advances in Bio-
     chemical Engineering^ Vol 3. Ghose, T.K. et.al. (Eds).
     Springer-Verlag. NY., 19TU, pp 221-277.
2.   Cooper, P.F. and Atkinson, B. (Eds), Biological Flui-
     dised Bed Treatment ofWater and Wastewater Ellis Hor-
     wood Limited, Chichester, UK. 1981.
3.   Smith, E.D. et.al., (Ed). Proceedings First National
     Symposium on Rotating Biological Contactor Technology.
     Champion, Pennsylvania, February U-6, 1980.
1*.   Harremoes, P. "Biofilm Kinetics". Chap. U. Water Pollu-
     tion Microbiology, Vol. 2., Wiley Interscience (Ralph
     Mitchell, Ed.), 1976.
5.   Riemer, M.W. "Kinetics of Denitrification in Submerged
     Filters. Part I" Ph.D. Dissertation. Technical Univer-
     sity of Denmark, 1978.
6.   Levenspiel, 0. Chemical Reaction Engineering, Wiley and
     Sons, New York, 2nd Ed. 1972.
T.   La Motta, E. "Evaluation of Diffusional Resistances in
     Substrate Utilization by Biological Films" Ph.D.  Disser-
     tation, Univ.of North Carolina, Chapel Hill, 197U.
                         182

-------
 8.   Harremoes, P. "Half order Reactions in Biofilm and Fil-
      ter Kinetics" Vatten 2 1977, p 22.
 9.   Williamson, K.J. and McCarty, P,L. "A Model of Substrate
      Utilization by Bacterial Films", J.W.P.C.F. 48 (l)t 1976,
      pp. 9-2U.
10.   Willimason, K.J. and McCarty, P.L. "Verification Studies
      of the Biofilm Model for Bacterial Substrate Utilization",
      J.W.P.C.F. 48 (2) 1976, pp 281-296/
11.   Rittmann, B.E. and McCarty, P.L. "Evaluation of Steady-
      State-Biofilm Kinetics", Bioteahn. Bioengr. 22. 1980,
      pp 2359-2373.
13.   Trulear, M.G. and Characklis, W.G. "Dynamics of Biofilm
      Processes" in 34th Annual Purdue Industrial Waste Con-
      ference 3 West Lafayette3 Indiana, May 8-103 1979 (Ann
      Arbor Publishers, Ann Arbor3 MI3 1979).
lU.   Bryers, J.D. "Dynamics of Early Biofilm Formation",
      Ph.D. Dissertation, Rice University, Houston, Tx. 1980.
15.   Characklis, W.G. "Bioengineering Report: Fouling Bio-
      film Developement - A Process Analysis" Bioteahn, Bio-
      engr. 23 1981, pp. 1923-1960.
l6.   Bryers, J.D. and Characklis, W.G. "Early Fouling Biofilm
      Formation in a Turbulent Flow System: Overall Kinetics."
      Vat.Res.., 15 1981, pp. U83-U91.
17.   Baier, R.E. "Influence of the Initial Surface Conditions
      of Materials on Broadhesion" Proa. 3rd Initial Congr.
      Marine Corrosion and Fouling. Nat'l Bureau of Standards,
      Gaithersburg, Maryland, Oct. 2-6, 1972.
18.   Freidlander, S.K. and Johnstone, H.F. "Deposition of
      suspended particles from turbulent gas streams". Ind.
      & Engr. Chem. 49  (?), 1970, pp. 1-11.
19.   Beal, S.K. "Deposition of particles in turbulent flow
      on channel and pipe walls". Nuol.Soi. and Eng. 40 1970
      pp. 1-11.
20.   Browne, L.W.B. "Deposition of particles on rough surfa-
   \   ces during turbulent gas flow in a pipe?. Atm.Envir. 83
      1971*, pp. 801-815.
21.   Stathopoulous, N. "Influence of Temperature on Biofilm
      Processes", M.S. Thesis, Rice University, Houston, Tx
      1981.
22.   Lawrence, A.W. and McCarty, P.L, "A unified basis for
      biological treatment design and operation". J.Sanit Eng.
      Div.3 ASCE, 96, 1970, SA3.
                           183

-------
                THE MICROBIOLOGY OF ROTATING
                 BIOLOGICAL CONTACTOR FILMS
     Nancy E. Kinner.  Department of Civil Engineering,
     University of New Hampshire, Durham, New Hampshire.

     David L. Balkwill.  Department of Microbiology,
     University of New Hampshire, Durham, New Hampshire.

     Paul L. Bishop.  Department of Civil Engineering,
     University of New Hampshire, Durham, New Hampshire.

INTRODUCTION

     The treatment of municipal and industrial wastewater
generated by modern society is rapidly becoming an intract-
able problem.  The continuing demand for a pollutant-free
environment (1) is exceeding the ability of traditional waste
treatment processes to produce high quality effluents at
reasonable costs.  A recent GAO study reported that, of the
242 wastewater treatment facilities examined, 87 percent were
in violation of their NPDES permits at least one month of the
year, with 56 percent being in violation more than half of
the year (2).  Many of the violations, resulted from the
municipality's inability to afford the high operation and
maintenance costs (3).  Consequently, economical and innova-
tive wastewater treatment techniques are needed immediately
to meet the legal and public demand for water pollution
control.  The rotating biological contactor (RBC), a rela-
tively new technique for aerobic biological wastewater
treatment, offers a cost effective solution to this demand
with the advantages of low energy and maintenance require-
ments, high organic removal efficiencies at short retention
                             184

-------
times, modular flexibility in design, and adaptability to a
wide range of wastewater types and flows.
     Until recently, most of the research on RBCs has been
conducted using traditional engineering methods in an effort
to determine their overall organic removal efficiency and
design parameters during the treatment of various kinds of
wastes (4,5,6,7,8,9,10,11).  Design equations, based on
            32'
hydraulic [ra  applied/m *d] (12,13) and organic [gms organic
                              2
matter as BOD or COD applied/m "d] (14,15) loading rates,
have employed general empirical relationships and large,
conservative safety factors.  With the increasing demand for
cost effective designs, optimization of RBCs has become
important.  To optimize organic removal, one must understand
the interactions between RBC microorganisms and their physico-
chemical environment.  This results from the fact that the
RBC process is a product of the microbial ecosystem which
operates within its confines.  As a first step towards under-
standing the microbial interactions which occur in the RBC it
is necessary to have a general knowledge of 1) the micro-
organisms present, 2) their relative abundance, and 3)
their ultrastructural characteristics which may be indicative
of their physiological state.
     The bacteria inhabiting RBCs during secondary wastewater
treatment have not been thoroughly examined.  Most published
research which contains a description of the biofilm con-
stituents provides it as supplementary information.  No
examination of the microflora of the suspended floes has been
conducted, though Kincannon and Groves (16) assert that they
can play a major role in organic removal.  Information on the
biofilm has been collected by observing its gross morphology
and by examining wet mount slides.
     The RBC biofilm is usually characterized as shaggy and
filamentous (5,17).  The effects of compartmentalization,
however, are apparent; the biofilm's color and density varies
along the length of the unit.  When treating municipal or
artificial sewage the first compartment usually contains a
thick, white to gray growth (7,12,18) which grades to a dark,
brown—black and thinner biofilm in the final compartments
(7,18,19).  The sparse growth is attributed to protozoan
predation and low organic concentrations.  These character-
istics may differ- when industrial wastes are being treated
(6).
     The first attempt at a complete categorization of the
biofilm constituents was made by Antonie and Welch (20) as
                             185

-------
part of a study of RBCs during dairy waste  treatment.  Sever-
al microorganisms were identified  (Table I).  They concluded
that the most important species were the filaments Geotpichium
candidum and Bacillus cereus, and  the nonfilamentous bacteria
Zooffloea filipendula, Pseudomonas denit^ifiaans^ AeTobaotev
aspogenes3 and Esaherichia coli.  Unfortunately, the authors
did not discuss the techniques used to isolate and identify
the bacteria to the species  level nor did they mention their
relative numbers and distribution within the RBC.

                           Table I
         Organisms Identified in an RBC Biomass  (20)

Predominant Organisms	:	Non-Predominant Organisms

Zooffloea filipendula               Pseudomonas fluoresaens
Pseudomonas denitrificans          Pseudomonas aermginosa
Aepobactep aepogenes               Neissevia catavrhalis
Escherichia coli                   Geotyichium candidum
Escherichia freundii Type I        Torula spp,
Bsoherichia spp.                   Rhodotorula spp.
Bacillus ceyeus var, mycoides
Bacillus oereus
Micrococcus aonglomeratus
Micrococaus luteus
     Several authors have described the indigenous biofilm
populations inhabiting properly loaded RBCs treating munici-
pal or artificial sewage.  They have identified the bacteria
present by examining wet mount slides of the biofilm.  In the
first compartments, the most commonly observed filamentous
bacterium is Sphaerotilus (21,22,23,24,25,26).  Beggiatoa
(22,23,27), Fusarium (26), Nocardia (25), Cladothrisc (23),
and Oscillatopia (26) are found less frequently.  Nonfila-
mentous forms observed in the first compartments are Zoogloea
and zoogloeal masses (21,23); unicellular algae (26); and
unicellular rods, spirilla, and spirochaetes (27).  The final
compartments contain most of the same forms as well as
Stz»eptamyces (27) and Athx>o"botrys (22).  Protozoan popula-
tions have been characterized microscopically, but will not
be discussed in this paper.
     In this research traditional light microscopy, inter-
fence optics, and transmission electron microscopy were used
                             186

-------
to examine the biofilm constituents of the first compartments
of RBCs treating domestic wastewater.  Two different types of
RBC pilot plants were studied.  The smaller unit consisted of
a single compartment which had 18 cm diameter, polyurethane
               •n
coated Masonite  disks.  The larger unit had four equally
sized compartments each of which contained a section of 0.5 m
diameter corrugated plastic disk media.  Both RBCs were
                3  2
loaded at 0.04 m /m *d; typically hydraulic loading rates for
                             3  2
RBCs vary from 0.04 to 0.08 m /m «d (17).  Biofilm from the
first compartment of each unit was examined after steady
state operation was achieved.  Some staining was done for
light and transmission electron microscopy.  Filaments were
isolated on special microbiological media.  Particular atten-
tion was directed to determining 1) the identity of the
predominant filaments, 2) the morphological characteristics
of single-celled bacteria present, and 3) the ultra-structural
characteristics of the bacteria as a possible indicator of
their physiological and ecological conditions.

MATERIALS AND METHODS

RBC Pilot Plant Descriptions

     One laboratory-scale RBC unit was operated under a fume
hood in an environmental engineering laboratory.  It had one
compartment constructed from an acrylic half cylinder 30 cm
long and 20 cm in diameter.  A horizontal stainless steel
shaft supported 16 disks, each with an 18 cm diameter, for 'a
                                   2
total wetted surface area of 0.78 m .  The equally spaced
                           •n
disks were made of Masonite  sealed with polyurethane.
Effluent flowed from the RBC to an ajoining basin through
each of four 1.3 cm diameter ports located at the base of the
end wall and over a notched weir located at the top of the
end wall.  A peripheral disk velocity of 0.31 m/s was main-
tained by a mechanical drive.  The unit was exposed to a low
                                             2
level fluorescent light of less than 100 Im/m  for a maximum
of 12 hours per day.  Ambient air and wastewater temperatures
were 20°C.  All disks were approximately 40 percent submerged
in wastewater at any given time.
     A second RBC pilot plant was housed in a laboratory
trailer located at the Durham, New Hampshire wastewater
treatment plant.  It was a 0,5 m diameter, 4 compartment Bio-
                           187

-------
Surf  unit with corrugated polyethylene disk media.  Influent
was delivered to a wet well and was distributed to the first
compartment by four rotating scoops.  Wastewater flowed from
one compartment to the next through each of two 2.6 cm dia-
meter ports located in the baffle walls.  Effluent passed out
of the fourth compartment via an overflow pipe.  The peri-
pheral velocity was 0.31 m/s (mechanical drive) and the
submergence level was 40 percent.  The unit was exposed to no
longer than 10 hours of natural light per day.  Ambient air
temperature was maintained at 20°C; wastewater temperature
was no less than 17°C.
     Raw sewage, the influent for the 18 cm RBC, was obtained
in 20 1 carboys from the Durham sewage pumping station.  The
carboys were stored at 4°C until used (a maximum holding time
of 3 days).  During this period solids settling occurred.
The settled sewage was transported from the carboys to the
small RBC via a peristaltic pump set to deliver 67 1/d.  This
flow was sufficient to operate the unit at an hydraulic load-
                  3  2
ing rate of 0.04 m /m *d and an organic loading rate averaging
3.2g TOC/m2-d.
     The 0.5 m diameter RBC received 0.95 m  of fresh primary
effluent from the Durham treatment plant per day.  This was
pumped continuously to the wet well of the RBC achieving an
                                        3  2
overall hydraulic loading rate of 0.04 m /m *d and an organic
                                 2
loading rate averaging 3.2g TOC/m -d.  The hydraulic loading
                                   3  2
to the first compartment was 0.16 m /m *d and the organic
                                 2
loading rate averaged 12.8g TOC/m «d.
     After a three week start-up period both RBCs had achieved
steady state operation as determined by obtaining similar
effluent total organic carbon (TOC) concentrations on three
consecutive days.  TOC measurements were performed on the RBC
influent and effluent samples after filtration through
Whatman #40 paper, according to the ampule method outlined
                                  2
for the Oceanography International  Model 526.  The RBC influent
wastewater:  settled raw sewage and the fresh primary effluent,
had average TOC's of 80 rag C/l.  The effluent concentration
from the 18 cm diameter and 0.5 m diameter RBCs were 17.5
 Autotrol Corporation; Milwaukee, Wisconsin.
2
 Oceanography International Corporation; College Station, Texas.
                               188

-------
and 23.0 mg C/l, respectively.  Microbial samples of the 18
cm diameter RBC biofilm for both light and electron microscopy
were randomly scraped from the surface of the first disk
after steady state was achieved.  Biofilm from the 0.5 m
diameter unit was randomly scraped from the front, middle and
end surfaces of the disk media in the first compartment.

Light Microscopy

     The biofilm removed from the RBC disks was too dense to
examine directly.  To prepare samples for light microscopy
the biofilm was rinsed in several petri dishes containing
         R
Nannopure  water and then repeatedly drawn up into a Pasteur
capillary pipette to separate the densely tangled mass.
Several wet mount slides of each washed sample were examined
under a Nikon Biophot Research Microscope equipped with
Nomarski differential interference bright field optics.  A
photographic record of observations was made using Panatomic
X (ASA 32) film.
     Pieces of each sample of rinsed biofilm were then run
                                                  P
through another series of four rinses in Nannopure  water.
These were further separated by the Pasteur pipette technique
and by micromanipulation.  Most of the constituents were
removed from the sample by these procedures except for the
filaments and zoogloeal masses.  Staining methods were em-
ployed to determine the presence of poly-B-hydroxybutyrate
                         +3
(PHB) and ferric iron (Fe  ) in these samples.
     Burden's method, as outlined in the Manual of Microbio-
logical Methods (?8), was used to determine if the micro-
organisms contained PHB.  After staining with 0.3% alcoholic
Sudan Black B and counterstaining with 0.5% aqueous .S.afranin,
PHB appeared blue-black while the rest of the cell was pink.
     Ferric iron on the filaments was reacted with 0.1%
aqueous potassium ferrocyanide, under acidic conditions, to
produce the Prussian blue reaction (29) .  Special care was
taken to insure that soluble ferric iron in the biofilm was
                                                   P
removed by washing these samples in extra Nannopure  water
rinses.              •
     Color photographs were taken of the samples after stain-
ing procedures were performed.  An Olympus BHA microscope was
used with Kodachrome ASA/25 and ASA/64 film.
                            189

-------
Isolation Experiments

     The isolation techniques developed by Dondero, Phillips,
and Heukelekian  (30) for Sphaerotilus were followed.  Biofilm
                                  P
washed in four rinses of Nannopure  water and teased apart
using the Pasteur pipette technique was placed in a blender,
                             P
containing 50 ml of Nannopure  water, for 30 seconds.  The
homogenate was streaked on petri dishes of CGY and CG agar
media and incubated for 48 hours at 28°C.  After incubation
the plates were observed under a dissecting microscope.
Tangled, curled, filamentous growth suspected of being Sphaero-
titus, was reisolated on fresh plates of the media and in-
cubated for 48 hours at 28°C.  The filamentous growth formed
after reisolation was observed using the Olympus microscope
and the PHB test was performed according to the procedures
described above.  Color photographs were taken of these
samples.

Phototatic Experiments

     To test the response of the biofilm constituents to
light, a series of phototactic experiments were conducted.
These procedures were recommended by Dr. Jane Gibson of
Cornell University (31) .  Extract agar plates were prepared
by adding 2 gm of agar to 1 liter of filtered (Whatman #40
paper) mixed liquor from the first compartments of the RBCs.
Plates were poured after the medium was sterilized for 20
minutes at 15 psi.
     Six plates of the RBC extract media were streaked with
rinsed biofilm samples.  Three plates were incubated in the
dark; three plates were incubated in continuous light which
was provided by two fluorescent lights.  All incubations were
at 25°C for one week.
     Biofilm samples removed directly from the RBC were
teased apart.and placed on one side of each of nine plates
containing RBC extract media.  Six of the plates were then
covered with aluminum foil.   Three of these plates had a 1 mm
diameter hole placed in the foil on the side opposite the
sample.   The pinhole provided a fixed light source to which
photosynthetic organisms would migrate.  All of the plates
were placed in a 25°C incubator, which was continuously
illuminated by two fluorescent lights, for one week.
                            190

-------
Electron Microscopy

     All biofilm specimens were prepared for electron micro-
scopy by the thin sectioning technique.  Two fixation pro-
cedures were used to prepare each sample for thin section-
ing.  For the Kellenberger fixation, pieces of biofilm mater-
ial were suspended in Kellenberger buffer (32) and sufficient
1% OsO. (in Kellenberger buffer) was added to bring the final

concentration of OsO, to 0.1%.  The samples were prefixed in

this suspension for 30 minutes at room temperature, after
which they were concentrated and washed by centrifugation in
Kellenberger buffer.  The resulting pellet was resuspended in
2-3 drops tryptone-salt solution (1% tryptone, 0.5% NaCl) and
mixed with approximately 0.5 ml molten 2% Difco Noble Agar
at 50°C.  The agar-specimen mixture was then transferred to
a glass slide, allowed to solidify, and cut into small blocks
(less than 1mm on a side).  These blocks were postfixed 12-18
h at room temperature in 1% OsO, (in Kellenberger buffer) and

prestained 2 h at room temperature in 0.5% uranyl acetate (in
Kellenberger buffer).  For the glutaraldehyde-osmium tetroxide
fixation, pieces of biofilm material were suspended in 0.1 M
sodium cacodylate buffer (pH 7.5) and sufficient 12.5% glutar—
aldehyde (in 0.1 M sodium cacodylate buffer) was added to
bring the final concentration of glutaraldehyde to 3%,
Following prefixation in this suspension for 2 h at room
temperature, the specimens were concentrated and washed twice
by centrifugation in sodium cacodylate buffer.  The final pel-
let was resuspended in tryptone—salt solution and embedded in
agar as described above.  The resulting blocks of agar were
then postfixed 12-18 h at room temperature in 1% OsO, (in 0.1
M sodium cacodylate buffer).
     Samples from both fixations were dehydrated through a
graded ethanol series and then embedded in Spurr's low-visco-
sity epoxy resin (33).  Thin sections were cut on an LKB
Ultratome III ultramicrotome, using glass knives or a Diatome
diamond knife.  The sections were retrieved on uncoated, 400-
mesh copper specimen grids, after which they were poststained
15 minutes with 0.5% uranyl acetate (in 50% methanol) and 2
minutes with 0.4% lead citrate (34),
     Thin sections were viewed with a JEOL JEM-100S transmis-
sion electron microscope at an accelerating potential of 80 kV.
The specimens were examined and photographed extensively in


 Difco Laboratories; Detroit, Michigan.
                              191

-------
order to ensure that a representative sampling of microbial
cells was obtained.  Comparisons were also made with light
microscopical observations (above) for this purpose.  Both
fixation procedures used for thin sectioning gave equivalent
results.  Micrographs of samples prepared with the glutaralde-
hyde-osmium tetroxide fixation were chosen for purposes of
illustration in this study.

RESULTS

Gross Morphology

     The biofilm on the first disk in the 18 cm diameter RBC
and in the first compartment of the 0.5 m diameter RBC was
gray-brown and filamentous with a subsurface black layer.
Growth was fairly uniform; maximum film thickness was 1 mm.•
Sloughing occurred randomly and recolonization was immediate.
On the terminal disks in the 18 cm RBC and in the last com-
partments of the 0.5 m RBC the biofilm was thinner and dark
brown appearing mottled because recolonization occurred more
slowly.  These RBC biofilm characteristics were similar to
those observed previously during domestic wastewater treat-
ment (7,12,18,19).

Light Microscopy

     Biofilm from both of the RBC pilot plants was extremely
dense forming an interwoven mat.  It was composed of two
principal constituents:  filaments and single-celled bacteria
grouped together in amorphous clumps.  The biofilm constitu-
ents were similar to ones previously observed in RBC biofilms
in this laboratory (35).  The filaments consisted of a series
of sausage—shaped cells, approximately 1-2 ym in diameter and
2—5 ym long, which were tightly encased in an outer sheath.
The sheath was most visible at the ends of the filaments
where the cell chain terminated leaving only the empty cas-
ing.  No flagellated cells were observed exiting from the
broken ends of the filaments.  Concurrently, no holdfasts
were seen, though these may have been lost when the sample
was scraped from the RBC.  The sheaths were very flexible and
were often bent to severe angles without rupturing.  It was
impossible to determine the overall length of the filaments
because they were too intertwined with one another.  False
branching was rarely observed.  The filaments did not move or
oscillate during examination.
                              192

-------
     Most of the cells within the filaments contained blue-
black inclusion bodies after staining with Sudan Black B
indicating that PHB was stored,  A small number of filaments
did not contain PHB or contained it only in a localized
region.  Cells with PHB usually contained at least three of
the blue-black inclusion bodies; in some filaments the PHB
storage appeared to involve as much as 3/4 of the cell.  The
zoogloeal masses also contained these inclusion bodies,
     The sheaths of the filaments stained a dark Prussian
blue after exposure to potassium ferrocyanide under acidic
conditions.  As great care was exercised to insure that no
soluble ferric iron was in solution, it appears that the iron
precipitated out onto the sheaths of the filaments during
wastewater treatment.

Isolation Experiments

     Tangled and curled filamentous growth appeared on all of
the initial isolation plates of CGY and CG media.  After
reisolation the filaments were examined using the light
microscope.  They were similar to those observed in the
biofilm samples:  sausage-shaped cells within a sheath ex-
hibiting the same morphological characteristics.  Most of the
individual cells contained PHB.

Phototactic Experiments

     There was no visible difference in the amount of growth
on the RBC extract media plates after the one week incuba-
tion.  Both the plates exposed to continuous light and
complete darkness contained a substantial number of bacterial
colonies.  No microorganisms moved toward the light on those
plates with the pinhole in the aluminum foil covering, except
for nematodes which burrowed throughout the media.

Ultrastructure of the Non-filamentous Population

     Transmission electron microscopy of thin-sectioned sam-
ples confirmed the presence of the non-filamentous bacterial
cells seen by light microscopy.  From low-magnification
micrographs, it was evident that both a large number and a
wide variety of these organisms were present (Fig. 1).  Cell
diameters varied considerably, ranging, from 0.25 to 1.5 urn.
     The ultrastructural characteristics of sever-al repre-
sentative types of the non-filamentous bacteria are shown in
Fig. 2.  A number of these organisms regularly contained one
                             193

-------
Fig. 1.  Low-magnification electron micrograph of the non-
filameatous bacterial forms present in the RBC biofilni.
Bar - 1.0 vim.  Note variety of morphological types and the
tendency for similar types to be grouped in a relatively
confined area.
                              194

-------
            ^"Ws£a«**      ^  * *J» *•»%!* "taSK^'4 *   V* •J**V A?PF3^1' *

                                                 '•«=%. "Pii <*«?* rt
Fig. 2.  Electron micrographs of representative non-filamen-
tous forms,*showing  typical ultrastructural features.  Bars
= 0,5 pm.  a. Cell with PHB Inclusion bodies,  b. Cell with  .
electron-dense inclusions,  possibly polyglucoside granules.
c. Cell with prominent  mesosome.  d. Cell with unidentified,
inclusions of medium electron density.

Fig. 3.  Electron micrograph of amoeboid cell containing .two
bacterial cells  (arrows)  within a vacuole.  Bar = 1.0 ym._
                           195

-------
or more inclusion bodies, and most of them possessed mesosome-
like structures.  In several instances, cells were seen that
appeared to be infected with bacteriophage.
     The non-filamentous bacteria often appeared as groups of
cells possessing identical morphological and ultrastructural
characteristics.  These groups, which included 3 to 25
cells, apparently represented microcolonies or colonies.  One
or more of the cells in such groups were sometimes seen to be
undergoing cell division.
     The only eukaryotes detected with any regularity in the
electron microscopical investigations were amoeboid organisms
because the larger protozoa and metazoa were lost in the
fixation process (Fig. 3).  Interestingly, these organisms
always appeared to have several vacuole—like structures which
contained one or more intact bacterial cells.

Ultrastructure of the Predominant Filaments

     Transmission electron microscopy also confirmed that the
predominant filaments seen by light microscopy (above) con-
sisted of independent bacterial cells surrounded by a common
sheath (Fig. 4),  The filaments ranged from 1.35 to 1.55 pm
in diameter, including the sheath.  Individual cells within
the sheaths ranged from 1.0 to 1.2 pm in diameter and from
1.9 to 4.5 um in length.
     All filaments possessed a relatively dense layer of
sheath material that was situated quite close to the surface
of the underlying cell walls.  Many filaments possessed an
additional layer of sheath material which was external to the
dense layer.
     The cells within the filaments were independent of one
another.  The cells possessed a typical Gram-negative cell
envelope (36).  The cells in some, but not all, filaments
contained as many as 15 electron-transparent inclusions
surrounded by a. single electron-dense bounding layer (Fig.
5).  They corresponded in size and location to the Sudan
Black B-staining granules seen by light microscopy, and in
their ultrastructural characteristics to PHB bodies (37,38).
in one instance, a filamentous organism appeared to be in-
fected with bacteriophage.
     Most cells in the filaments contained prominant mesosome-
like structures.  These were peripherally located and some-
times appeared to be associated with the polar walls.
                              196

-------
                 S-.^'.rfjAjV/^t^-.V'-^s-K'-^'j-'' .,v.v>"tr»*->,-t,: --g   •*
                         1 *•• v  >'?  ••  , ;H, % •fc*.
Fig. 4.  Electron micrograph of predominant Sphae^otitus—
like filaments,  showing arrangement of cells within the
common sheath  (S).   Note the prominent mesosome-like
structures  (M) .  Bar =, 1.0 wm.

Fig. 5.  Electron micrograph of 'cell within a Sphaevo'tilus-
like filament  containing many PHB inclusion bodies  (P).
Note sheath (S)  struucure, rnesosome-like structure  (M), and
typical Gram-negative cell wall (W).  Bar = 0.5 urn.
                              197

-------
DISCUSSION

     This study has served to document for the first time the
morphological and ultrastructural characteristics of micro-
organisms living within the biofilm formed on a rotating
biological contactor.  It makes a significant contribution to
the RBC literature because the information available on the
morphology, physiology, and ecology of RBC biofilm micro-
organisms is extremely limited.  Ultimately a greater under-
standing of the biofilm's function can help the engineer
optimize RBC design and performance.  In addition, this study
provides information on microorganisms operating in their
natural environment; making the conclusions drawn from this
information available for practical application.
     The predominant organism in the biofilms examined here
was a filamentous bacterium consisting of rod-shaped cells
enclosed by a common sheath.  The data presented suggest
strongly that this bacterium is a Sphaerotilus species accord-
ing to the taxonomic structure of the Sphaepotilus-Lept.othr'ix
group recently established by van Veen et al (39) .  It is
important to note, however, that the species-level taxonomy
of this group has been controversial (39,40,41).  The princi-
pal methods of identification for Sphaerotilus'are based on
microscopic examination, plating and isolation,  and on dis-
counting the possibility of the specimen being another type
of filamentous form (40).  The light microscopical morphology
of the organism was identical to that described for Sphaero-
tilus by several authors (39,40,42,43,44).  The RBC filaments
                                    +3
had:  1) smooth, thin, colorless, Fe   encrusted sheaths
which tightly encased the cells and were often partially
evacuated on one end, and 2) Gram-negative cells within the
size range 1.-3.5 pm wide and 2.5-16 pm long arranged in a
single row.  Other filamentous forms were ruled out because
the RBC species lacked endospores and crosswalls, were non-
motile, and did not demonstrate a positive phototactic
response.  The filaments isolated on both the CGY and CG
media were similar to those described by Dondero et. al. (30)
further indicating that they were Sphaevotilus.   The ultra-
structural characteristics of the filaments were also similar
to those of Sphaerotilus species noted in other studies (39,
43,44,45), especially the strain examined by Petitprez et.
al. (46).  The principal similarities included sheath morphology,
wall structures, presence of PHB granules, and presence of
prominent mesosomes.
                                 198

-------
     Most microscopic studies of the biofilm have not in-
volved a thorough examination of the filaments present.  As a
result they have been identified as various types of bacter-
ia, algae, and fungi. . This study has confirmed that the pre- .
dominant filament growing in healthy RBC biofilms during
domestic wastewater treatment is the bacterium, Sphaerotilus.
Initially the iron encrustation on its sheath -led invest!-.
gators to believe Sphaerotilus was an autotroph (47,48); how-
ever, it is now considered an aerobic heterotroph (39,49,50).
The role of iron deposition in Sphaerotilus remains unknown,
though the mechanism may be associated with a moiety of the
organisms sheath which catalyzes the reaction -(51,52)'.
Sphaerotilus-based films growing in laboratory and natural
                                             2
environments can remove 0.5-7.4 g organic C/m *d (53,54);
suggesting that this bacterium may contribute significantly
to the organic uptake capacity of the RBC biofilm.  Though
Sphaerotilus requires oxygen as a terminal electron acceptor
it can function in microaerophillic conditions (39,55).  This
is particularly significant because the filaments may con-
tinue to assimilate organic matter in spite of the rapid
decrease in oxygen concentration with depth in the RBC bio-
film.  Sphaepotilus can exist as a filament or free-swimming
flagellated cell.  This flexible morphology is also uniquely
suited to the RBC process.  Swarmers can rapidly recolonize  "
disk surfaces after sloughing and filaments can attach to the
disks and/or serve as a stabilizing force within the biofilm
in a manner similar to that of reinforcing rods in concrete.
The maximum growth of Sphaevotilus occurs when the fluid
velocity is between 0.18 and 0,45 m/s (56) which coincides
with the peripheral velocities used to optimize effluent
quality in RBCs (17,18).
     The fixation procedure used here to prepare the biofilm
for electron microscopy did not specifically preserve eukary-
otic cells, especially large protozoa and metazoa. . Amoebae,
however, were regularly observed indicating that they may
play a significant role in the trophic structure of the
biofilm.  Most previous studies have determined that ciliates
are the major protozoa present in wastewater treatment systems
(57,58,59).  A few researchers (60,61,62) found that amoebae
were often overlooked or identified as detritus.  Sydenham
(61) concluded that they may be ecologically as important as
ciliates in improving the efficiency of-wastewater treatment
systems.  The amoebae in this study contained single-celled
bacteria in individual vacuoles.  Ciliates in activated
sludge systems have been observed to prey upon single-celled
                               199

-------
bacteria; predominantly on enteric species from the raw
sewage (63,64,65).  The ciliate's predatory activity results
in lower organic and suspended solids concentrations in the
effluent.  The amoebae may play a similar role in the RBC
biofilm.  It is likely that they live on or near the bio-
film's surface where oxygen and influent bacteria are more
abundant.
     A cell's ultrastructural characteristics can indicate
the microorganism's physiological condition.  The data pre-
sented in this study confirm the presence of a metabolically
active population in the biofilm.  It was apparent that the
population was quite active from:  the numbers of cells seen,
the variation in cell size, the presence of microcolonies,
and the presence of dividing cells within these microcolonies.
The presence of mesosomes in both filaments and non-fila-
mentous cells may also be evidence for active metabolism and
growth.  Although mesosomes are currently somewhat controver-
sial in terms of their true ultrastructure and their func-
tions (if any) in the bacterial cell, they are often seen in
dividing cells or metabolically active cells (66).
     Both the Sphaerotilus filaments and many of the non-
filamentous bacteria contained PHB granules.  PHB is stored
by bacterial cells when carbon concentrations available in
the environment are not limiting (44,67).  The large number
of PHB granules found in the RBC bacteria indicates that
excess carbon was present and had been metabolized.  Organic
carbon assimilated by bacteria may be used in 1) respiration,
2) cell growth and division, 3) PHB production, or 4)
extracellular matrix and sheath production.  The storage of
PHB by biofilm bacteria may serve as an important intraeellu—
lar sink for organic carbon in RBCs.  PHB can account for 11-
22.5% of the dry weight of Sphaeratilus (68) and 12.0-50.5%
of the dry weight of loogloea (69) .  The variation in the
percent of cell volume involved in PHB storage may be a
function of the amount of organic matter available to the
cells.  Therefore, as the organic loading in the RBC in-
creases the bacteria may store more carbon as PHB until some
critical amount of the cell's volume is occupied by this
substance.  PHB storage, however, cannot be considered
exclusively of the other cellular metabolic processes because
it acts concommitantly with them in determining the fate of
assimilated carbon in the biofilm.  PHB also serves as a
carbon and energy source for the cells during low nutrient
concentrations (70,71) and in this capacity it may mitigate
against the effect of fluctuating hydraulic and organic
loadings in the RBC.
                                 200

-------
     In Sphaerotilus the thickness of the sheath and the
formation of an additional layer of sheath-like material has
been observed in cells exposed to high organic loadings (44,
72).   These external cell structures may function in a manner
similar to PHB and/or may additionally function like the
extracellular polysaccharide matrices described for zoogloeal
bacteria.  In aerobic waste treatment systems zoogloeal
matrices are important as 1) a storehouse of carbon and
energy, 2) an effective adsorbent of metals and organic
compounds, 3) an adhesive mechanism, and 4) a buffer during
high carbon and nitrogen growth conditions (73).
     •Some understanding of the ecological conditions in the
biofilm may also be drawn from examining the biofilm micro-
organisms.  Both light and transmission electron microscopy
revealed the presence of many different types of bacterial
cells.  Eukaryotic organisms were seen as well.  This work
supports the contention of other researchers who found
various types of bacteria present in wastewater treatment
systems (74,75,76,77).  The greater the diversity of the
biofilm community, the greater its stability, which increases
its ability to efficiently degrade wastes and withstand
fluctuations in the environment.  The presence of different
types.of bacteria, protozoa, and metozoa indicate that a
complex trophic structure may be operating in the biofilm
which helps it to continue functioning in spite of external
perturbations.  The appearance of groups of cells, either as
filaments or microcolonies, suggests that these forms are
favored over single cells.  The presence of phage within some
of the bacteria may be indicative of deteriorating conditions
in the biofilm.  Whatever the cause, bacteriophage may act as
natural enemies of biofilm bacteria by reducing their ability
to assimilate organic matter from the wastewater.
     While this study has shown that the RBC biofilm contains
a large and diverse population of microorganisms which form a
metabolically active ecosystem it leaves many questions about
the microbial ecology of the film unanswered.  Its greatest
significance may be that it prompts more research aimed at
optimizing RBC design and evaluation by increasing the engi-
neer's understanding of the biofilm's mode of operation.
Initially additional studies must be performed on the biofilm
in the other RBC compartments and as a function of radial
distance from the center of the disks.  Similarly, the
profile of microorganisms must be examined as a function of
time and depth within the biofilm.  The role of PHB in the
physiology of the cells and as a function of organic loading
                                201

-------
must be understood to determine the limits of its ability as
an intracellular carbon sink in the RBC.  The presence or
absence of extracellular polysaccharide matrices in the
biofilm should be determined because the role of these struc-
tures in the metabolism of organic carbon is suspected and
deserves further examination.  Finally the role of the proto-
zoa and the overall predator-prey relations of the RBC
biofilm must be determined to give a clearer picture of its
tropic structure.  Research of this kind is continuing in our
laboratories in an effort to answer some of these questions
and to ultimately optimize RBC design and evaluation through
an increased understanding of microbial interactions and
processes.
                              202

-------
REFERENCES

 1.	,  1980, "Environmental Support Still
     Strong," EPA J.  6(10); 16-18.  -

 2.	__,  1980, "Costly Wastewater Treatment
     Plants Fail to Perform as Expected," Govt. Acct. Office
     CED-81-9.

 3.  Hill, W.R., Regan, T.M., and Zickefoose, C.S., 1979,
     "Operation and Maintenance of Water Pollution Control
     Facilities: a WPCF White Paper," JWPCF 51: 899-906.

 4.  Birks, C.W., and Hynek, R.J., 1971, "Treatment of Cheese
     Processing Wastes by the Bio-Disc Process," Proc.
     Purdue Ind, Waste Conf. 26: 89-105.

 5,  Corneille, R.W., and OfShaughnessy, J.C., 1975, "Treat-
     ment of Apple Wastes Using Rotating Biological Con-
     tactors," Proc.  Purdue Ind. Waste Conf. 30: 675-688.

 6.  Hudson, J.W., Smith, J.P., and Pohland, F.G., 1976,
     "Rotating Biological Contactor Treatment of Shellfish
     Processing Wastewaters," Proc. Purdue Ind. Waste Conf.
     31: 193-205.

 7.  Griffith, G.T.,  Young, R.H.F., and Chun, M.J., 1978,
     "Rotating Disc Sewage Treatment Systems for Suburban
     Developments and High-Density Resorts of Hawaii,"
     Technical Rept.  116, Water Resource Center, University
     of Hawaii, Honolulu.

 8.  Mikula, W.J., et. al., 1980, "A Kinetic Model for Treat-
     ment of Cheese Processing Wastewater with a Rotating
     Biological Contactor," First Nat. Symp. on Rotating
     Biological Technology 1: 491-514.

 9.  Huang, J.C., 1980, "Operational Experience of Oxygen
     Enriched Rotating Biological Contactors.'" First Nat.
     Symp. on Rotating Biological Contactor Technology 1:
     637-659.

10.  Tanacredi, J.T., 1980, "Removal of Waste Petroleum
     Derived Polynuclear Aromatic Hydrocarbons by Rotating
     Biological Discs," First Nat. Symp. on Rotating Bio-
     logical Contactor Technology 1:725-732.
                               203

-------
11.  Pano, A., Reynolds, J.H. and Middlebrooks, E.J., 1980,
     "The Kinetics of a Rotating Biological Contactor Treat-
     ing Domestic Sewage," First Nat. Symp. on Rotating Bio-
     logical Contactor Technology 1: 449-467.

12.  Antonie, R.L., Kluge, D.L., and Mielke, J.H., 1974,
     "Evaluation of a Rotating Disk Wastewater Treatment
     Plant," JWPCF 46: 498-511.

13.  Barnes, D., and Wilson, F., 1976, The Design and Opera-
     tion of Small Sewage Works, E. and F.N. Spon, Ltd.,
     London.

14.  Stover, E.L., and Kincannon, D.F., 1976, "Evaluating Ro-
     tating Biological Contactor Performance," Wat. .Sew.
     Works 123: 88-91.

15.  Poon, C.P.C., Chao, Y.L., and Mikucki, W.J., 1979,
     "Factors Controlling Rotating Biological Contactor Per-
     formance," JWPCF 51: 601-611.

16.  Kincannon, D.F., and Groves, S., 1980, "Role of Suspended
     Solids in the Kinetics of RBC Systems," First Nat. Symp.
     on Rotating Biological Contactor Technology 1: 433-488.

17.  Antonie, R.L., 1976, Fixed Biological Surfaces Waste-
     water Treatment - The Rotating Biological Contactor, CRC
     Press Inc., Cleveland.

18.  Friedman, A.A., Robbins, L.E., and Woods, R.C., 1979,
     "Effects of Disk Rotational Speed on Biological Con-
     tactor Efficiency," JWPCF 51: 2678-2691.

19.  Ellis, K.V., and Banaga, S.E.I., 1976, "A Study of
     Rotating Disc Treatment Units Operating at Different
     Temperatures," Wat, Pollut. Control 75: 73-91.

20.  Antonie, R.L., and Welch, F.M., 1969, "Preliminary
     Results of a Novel Biological Process for Treating Dairy
     Wastes," Proc. Purdue Ind. Waste Conf. 24: 115-125.

21.  Torpey, W.N., et. al., 1971, "Rotating Disks with Bio-
     logical Growths Prepare Wastewater for Disposal or
     Reuse," JWPCF 43: 2181-2188,
                              204

-------
22.  Pretorius, W.A., 1971, "Some-Operating Characteristics
     of a Bacterial Disc Unit," Wat. Res. 5: 1141-1146.

23.  Pescod, M.B., and Nair, J.V., 1972,  "Biological Disc
     Filtration for Tropical Waste Treatment.  Experimental
     Studies," Wat. Res. 6: 1509-1523.

24.  Sudo, R. , Okada, M., and Mori, T.,  1977, "Rotating Bio-
     logical Contactor — Microbe Control in RBC, "J. Wat.
     Waste 19: 1.

25.  Hoag, G.E., Widmer, W.J., and Hovey, W.H., 1980, "Micro-
     fauna and RBC Performance: Laboratory and Full-Scale
     Systems.  First Nat. Symp. on Rotating Biological Con-
     tactor Technology 1: 167-187.

26.  Khan, A.N., and Raman, ?., 1980, "Rotating Biological
     Contactor for the Treatment of Wastewater in India,"
     First Nat. Symp. on Rotating Biological Contactor Tech-
     nology 1: 235-258.

27.  Clark, J.H.,  Moseng, E.M., and Asano, T., 1978, "Per-
     formance of a Rotating Biological  Contactor Under
     Varying Wastewater Flow," JWPCF, 50: 896-911.

28.  Pelczar, M.J., Manual of Microbiological Methods,
     McGraw-Hill,  New York, 1957.

29.  Sherman, V.B.D., A Guide to the Identification of the
     Genera of Bacteria, Williams and Wilkins, Baltimore,
     1967.

30.  Dondero, N.C., Phillips, R.A., and  Heukelekian, 1961,
     "Isolation and Preservation of Cultures of Sphaepo-
     tilus," Appl. Microbiol. 9: 219-227. -

31.  Gibson, Jane, 1980, Personal Communication.

32.  Kellenberger, E., Ryter, A., and Sechaud, J., 1958,
     "Electron Microscope Study of DNA-Containing Plasms.
     II. Vegetative and Mature Phage DNA as Compared with
     Normal Bacterial Nucleoids in Different Physiological
     Studies", J.  Biophys. Biochem. Cytol. 4: 671-687.
                              205

-------
33.  Spurr, A.R., 1969, "A Low-Viscosity Epoxy Resin Embedding
     Medium for Electron Microscopy," J, Ultrastruet. Res,
     26: 31-43.

34.  Reynolds, E.S., 1963, "The Use of Lead Citrate at High
     pH as an Electron-Opaque Stain in Electron Microscopy,"
     J. Cell. Biol. 17: 208-212.

35.  Kinner, N.E., and Bishop, P.L., 1981, "Rotating Bio-
     logical Contactor Microbiology During Saline and Non-
     Saline Wastewater Treatment," Submitted for Publication
     in JWPCF.

36.  Costerton, J.W., Ingram, J.M.'V and Cheng, K.J., 1974,
     "Structure and Function of the Cell Envelope of Gram-
     Negative Bacteria," Bacteriol. Rev. 38: 87-110.

37.  Dunlop, W.F., and Robards, A.W., 1973, "Ultrastructural
     Study of Poly-g-Hydroxybutyrate Granules from Bacillus
     aereus, J. Bacteriol. 114: 1271-1280.

38.  Snively, J.M., 1974, "Inclusion Bodies of Prokaryotes,"
     Annu. lev. Microbiol. 28: 167-187.

39.  van Veen, W.L., Mulder, E.G., and Deinema, M.H., 1978,
     "The Sphaerotilus-Lepthothyix Group of Bacteria," Micro-
     biol. Rev. 42: 329-356.

40.  Dondero, N.C., 1975, "The Sphaerotilus-Leptothrix Group,"
     Annu. Rev. Microbiol., 29: '407-428.

41.  Pringsheim, E., 1949, "Iron Bacteria," Biol. Rev. Cam-
     bridge Phil. Soc. 24: 200-245.

42.  Mulder, E.G., and van Veen, W.L., 1963, "Investigations
     on the SphaeTotilus-Leptothrix Group." Ant. van Leeuwenhoek
     29: 121-153.

43.  Stokes, J.L., 1954, "Studies on the Filamentous Sheathed
     Iron Bacterium Sphaerotilus natans," J. Bacteriol. 67:
     278-291.

44.  Hoeniger, J.F.M., and Tauschel, H.D., 1973, "The Fine
     Structure of Sphaerot'i'lus natans," Can. J. Microbiol.
     19: 309-313.
                               206

-------
45.  Doetsch, R.N., 1966, "Notes on Some Structural Features
     of Sphaerotilus natans," Arch, Mikrobiol. 54; 46-55.

46.  Petitprez, M., et, al., 1969, "Quelques Aspects Struc-
     turaux de Sphaerotilus natans," Ann. Inst. Pasteur-Lille
     20: 103-114.

47.  Winogradsky, S., 1888, "Ueber Eisenbakterien," Botan.
     Ztg. 46: 262.

48.  Winogradsky, S., 1922, "Eisenbakterien als Anorgoxydan-
     ten," Zbl. Bakteriol. 57: 1.

49.  Petitprez, M., and LeClerc, H., 1969, "Les Bacteries du
     Groupe Sphaerotitus-Leptothpix," Ann. Inst. Pasteur-
     Lille 20: 115-140.

50.  Dondero, N.C., 1961, "Sphaerotitus, Its Nature and
     Economic Significance," Adv. Appl. Microbiol. 3: 77-107.

51.  Rogers, S.R., and Anderson, J.J., 1976, "Measurement of
     Growth and Iron Deposition in Sphaerotilus disQOphorus3"'
     J. Bacteriol. 126: 257-263.

52.  Rogers, S.R., and Anderson, J.J., 1976, "Role of Iron
     Deposition in Sphaevotilus discophoi>us," J. Bacteriol.
     126: 264-271.

53.  Curtis, E.J.C., Delves-Broughton, J., and Harrington,
     D.W., 1971, "Sewage Fungus;  Studies of Sphaepotilus
     Slimes Using Laboratory Recirculating Channels," Wat.
     Res. 5: 267-269.

54.  Capblancq, J., and Cassan, M., 1979, "Etude du Periphy-
     ton d'une Riviere Polluee (L'Agout).  I. Structure et
     Developpement des Communautes sur Substrats Artificiels,"
     Annls. Limnol. 15: 193-210.

55.  Dias, F.F., Dondero, N.C., and Finstein, M.S., 1968,
     "Attacked Growth of Sphaepot-ilus and Mixed Populations
     in a Continuous-Flow Apparatus," Appl. Microbiol, 16:
     1191-1199.

56.  Phaup, J.D., and Gannon, J., 1967, "Ecology of Sphaero-
     ti-tus in an Experimental Outdoor Channel," Wat. Res. 1:
     523-541.
                                   207

-------
57.  Curds, C.R., and Cockburn, A., 1970, "Protozoa in Bio-
     logical Sewage-Treatment Processes, I. A Survey of the
     Protozoan Fauna of British Percolating Filters and
     Activated Sludge Plants, "Wat. Res. 4: 225-236.

58.  Pike, E.B., and Curds, C.R., 1971, "The Microbial
     Ecology of the Activated Sludge Process."  In Microbial
     Aspects of Pollution, Ed. Sykes and Skinner, Academic
     Press, pp. 123-147.

59.  Curds, C.R., 1975, "Protozoa," In Ecological Aspects of
     Used-Water Treatment, Ed. Curds, C.R., and Hawkes, H.A.,
     Academic Press, London, Vol. I: 203-268.

60.  Brown, T.J., 1965, "A Study of the Protozoa in a Diffused-
     Air Activated Sludge Plant," Wat. Pollut. Control 64:
     375.

61.  Sydenham, D.H.J., 1971, "A Re-Assessment of the Relative
     Importance of Ciliates, Rhizopods, and Rotatorians in
     the Ecology of Activated Sludge," Hydrobiol. 38:  553-
     563.

62.  Schofield, T., 1971, "Some Biological Aspects of the
     Activated Sludge Plant at Leicester," Wat. Pollut.
     Control 70: 32.

63.  Curds, C.R., Cockburn, A., and Vandyke, J.M., 1968,  "An
     Experimental Study of the Role of the Ciliated Protozoa
     in the Activated Sludge Process," Wat. Pollut. Control
     67: 312-324.

64.  Curds, C.R., and Fey, G.J., 1969, "The Effect of Ciliated
     Protozoa in the Fate of Escherichia coli in the Activated
     Sludge Process," Wat. Res. 3: 853-867.

65.  Curds, C.R., 1971, "Computer Simulations of Microbial
     Population Dynamics in the Activated Sludge Process,"
     Wat. Res. 5: 1049-1066.

66.  Greenawalt, J.W., and Whiteside, T.L., 1975, "Mesosomes:
     Membranous Bacterial Organelles," Bacteriol. Rev.  39:
     405-463.

67.  Fukui, T. et. al., 1976, "Enzymatic Synthesis of Poly-B-
     Hydroxybutyrate in Zoogloea ramigera," Arch. Microbiol.
     110: 149-156.
                                 208

-------
68.   Rouf, M.A., and Stokes, J.L., 1962, "Isolation and
     Identification of the Sudlanophilic Granules of Sphaero-
     t-ilus natans," J. Bacteriol. 83:  343-347.

69.   Crabtree, K.,  McCoy, E., Boyle, W.C., and Rohlich, G.A.,
     1965, "Isolation, Identification, and Metabolic Role of
     the Sudanophilic Granules of Zoogloea ramigera," Appl.
     Microbiol. 13: 218-226.

70.   Stokes, J.L.,  and Parson, W.L., 1968, "Role of Poly-g-   '
     Hydroxybutyrate in Survival of Sphaeroti-lus discophorus
     during Starvation," Can. J. Microbiol. 14:  785-789.

71.   Parsons, A.B., and Dugan, P.R., 1971, "Production of
     Extracellular  Polysaccharide Matrix by Zoogloea ramigera,"
     Appl. Microbiol. 21: 657661.

72.   Phaup, J.D., 1968, "The Biology of Sphaerotilus Species,"
     Wat. Res. 2: 597-614.

73.   Joyce, G.H., and Dugan, P.R., 1970, "The Role of Floe-
     Forming Bacteria in BOD Removal from Wastewater," Dev.
     Ind. Microbiol. 11: 377-386.

74.   James, A., 1964, "The Bacteriology of Trickling Filters,"
     J. Appl. Bacteriol. 27: 197-207.

75.   Taber, W.A., 1976, "Wastewater Microbiology," Annu. Rev.
     Microbiol. 30: 263-277.

76.   Lighthart, B., and Loew, G.A., 1972, "Identification Key
     for Bacteria Clusters from an Activated Sludge Plant,"
     JWPCF 44: 2078-2085.

77.   Unz, R.F., and Dondero, N.C., 1970, "Nonzoogloeal Bac-
     teria in Wastewater Zoogloeas," Wat. Res. 4: 575-579.
                               209

-------
          ROTATING BIOLOGICAL CONTACTORS - SECOND ORDER KINETICS
                                    by
                            Edward J. Opatken
                   U.S.  Environmental  Protection Agency
                       Wastewater Research Division
               Municipal Environmental Research  Laboratory
                         Cincinnati, Ohio  45268
     The R8C process is uniquely adaptable for kinetic studies on secondary
treatment of wastewater".  Secondary treatment, for this specific kinetic
study, is defined as the removal or reduction of soluble substrate with
time. The substrate is identified in the reaction phase as soluble chemical
oxygen demand (sCOD) and/or soluble biochemical oxygen demand (sBOD).  The
reduction of insoluble oxygen demanding material is not applicable since:
     1.  It is the function of the reactor (RBC) to convert soluble
         organic matter into carbon dioxide and insoluble matter for
         later removal by the secondary clarifier.
     2.  The use of unfiltered oxygen demand would require the kinetic
         study to treat the RBC process as a heterogeneous reaction
         instead of a homogeneous reaction.
                                  210

-------
     The applicability of the RBC process to study reaction kinetics is
attributed to its process configuration, and operation mechanism.  The RBC
process usually consists of modular units (shafts) that are normally installed
in series. Each RBC shaft contains either 100,000 sq ft (9,300 m2) or 150,000
sq ft (14,000 m2) of surface area.  The volume of the trough to surface
area of the discs (V/SA) ratio is fixed by the manufacturer at 0.12 gal/sq
ft (4.9 L/m2). These basic geometric standards enable the reaction time to
be determined at each stage.  Each RBC shaft rotates at approximately 1.6
rev/min or at a peripheral speed of 60 ft/min (18.3 m/min).  The rotation
of the RBC mixes the wastewater, and thus simulates a stirred tank reactor.
The RBC is divided into independent stages by a baffle which enables the
disappearance of soluble oxygen demand to be quantified at each stage for
specific time intervals.
Second order kinetics.
     The published data by A. A. Friedman 0) on the disappearance of sCOO
was incorporated into the rate expression for a second order equation.
            where   r  =  AC
                    f  ~  -r   =  rate °f disappearance  of  sCOO, ing  sCOO/L
                    k  =  reaction rate constant, L/mg-h
                    ^  =  the  square of the concentration  of sCOD in
                          the  nth stage,
                                   211

-------
                   AC  =   (Cn_i-Cn)  =  the difference  in concentration of
                                        the influent into a stage from the
                                        concentration within that stage
                                        in mg sCOO/L
                   At  =   reaction time in h
     A plot of AC/At  vs  C^ is  shown  in  Figure  1.   The  slope  of the  line  is
the reaction rate constant, k, which has a value of 0.0062 L/mg-h.  The
intercept on this curve should theoretically go through zero;  however,
there is a fraction of sCOD that can be assumed as refractory.  This fraction
will not undergo biochemical conversion and is represented by  the "x" intercept.
This fraction is 33 mg sCOD/L for the synthetic influent used  by A. A.
Friedman in his RBC pilot  plant study.
      Another published paper by R. 0. Hynek(2) was used to obtain .interstage
data on the disappearance  of sCOD.  The hydraulic  loading rate was used to
calculate the residence time in each stage and again the data  was incorporated
into a second order rate expression.  The data consisted of results from
both a mechanical drive and an air drive RBC shaft.
     Data for the first five runs on both air and mechanical  drive systems
were plotted showing the concentration of sCOD at specific time intervals
                                                      *•
based on the retention time within a stage.  The curves for only three of
these runs are shown in Figures 2 and 3 to improve the clarity of the plot.
The plot indicated that the removal rate of sCOD decreased as  the reaction
time increased, which indicated second order rates of reaction.  The data
                                   212

-------
were then used to plot AC/At vs C^ to  obtain  the  reaction  rate  constant
from the slope of the plot.  For the five runs the mean reaction rate constant
was 0.015 l/mg-h for the mechanical drive system and the mean refractory
concentration for these five runs was 25 mg sCOD/L.  The air drive system
had a mean reaction rate constant of 0.025 L/mg-h and the mean  refractory
concentration was 21 mg sCOD/L.  These data are shown in Table  1.
      Table 1.  Reaction Rate ConstantsDerived fromR.J. Hynek Data

Run Number
VA-M
VA-A
VB-M
VB-A
V1A-M
V1A-A
V1B-M
V1B-A
V1C-M
V1C-A
mean value (M):
mean value (A):
k
(l/mg-h)
0.013
0.015
0.018
0.024
0.019
0.022
0.013
0.056
0.014
0.010
0.015
0.025
Refractory sCOD
Correlation Coefficient
0.999
0.999
0.997
0.995
0.998
0.996
0.999
0.987
0.997
0.999
0.998
0.995
• (mg/L)
32
30
27
26
27
24
14
0
26
25
25
21
 M = Mechanical drive
 A = Air drive
 Field verification of second order reaction.
      Three RBC facilities within a 80 km radius of the Andrew W. Breidenbach
 Environmental Research Center were sampled to obtain interstage data on the
 disappearance of sCOD.  The three facilities were LeSourdsville, Ohio;
 Indian Creek (Cleves, Ohio); and Brookville, Indiana.  Table 2 summarizes
 the characteristics of these facilities.
                                    213

-------
                Table 2.  Characteristics of RBC Facilities

                                                  Indian
                              LeSourdsville       CreekBrookville
No. of trains                      4                2             3
No. of shafts                     20                6             3
No. of stages per train            53             4
Diameter of disc, ft (m)      12(3.7)          12(3.7)       12(3.7)
Stages per shaft                   1.1             4
Total surface area, ft?        2.6xl06          4.8xl05        S.OxlO5
                   (tn2)      (2.4xl05)        (4.5xl04)      (2.8xl04)
 Surface area, per stage «2    IxlflS*           8xl04        2.5xl04
                               l.SxlO5**
                     (m2)      (9,300)*         (7}500)       (2,300)
                              (14,000)**
Design flow, mgd                 4.0              0.5            0.6
          (m3/d)               (15x103)         (1.9x103)     (2.3xl03)
Design hydraulic load,  gpd/sq ft 1.5              1.0            2.0
          (tn3/m2-d)           (0.062)            (.042)        (0.081)

 *Stages 1&2
**Stages 3,4,5
                                   214

-------
     On each sampling date, the following were obtained:
       1.  Influent, effluent, and stage samples
       2.  Influent and effluent temperature
       3.  Plant flow rate during the sampling period
     The samples were filtered and stabilized with acid before submittal to
the MERL Waste Identification and Analysis Section for sCOD analysis.
     The data were then incorporated into.a second order reaction rate
equation to determine the rate constant for-these systems.
     The interstage data on sCOD obtained for LeSourdsville were treated in
the following two modes.  The first mode consisted of plotting sCOD against
time and a curve was drawn to represent an approximate fit.
     The data from this curve were used to determine the reaction rate
constant by determining the slope when AC/At  was  plotted against C^.   The
C^ intercept was used to predict the refractory portion of the sCOD.  The
results  are shown in Table 3.
            Table 3.  Reaction Rate Constants for LeSourdsville
Run Number
L0815
L0825
L0903
L0909
L0925
L1002
L1008
LI1017
LII1017
LI1024
LII1024
LI1031
LII1031
LI1105
mean value:
o :
mg
k (L/mg-h)
0.016
0.028
0.032
0.018
0.023
0.015
0.022
0.015
0.026
0.019
0.026
0.026
0.021
0.009
TTUFT
0.0062
Refractory sCOD,
(mg/L)
, 37
41
6
21
36
28
43
47
26
30
36
27
16
49
mean value: ""32
a: 12
                                    215

-------
     A second approach was to combine the data from all the runs to obtain a
mean value for the influent, the four intermediate stages, and the effluent.
A plot of the disappearance of sCOD with time is shown in Figure 4.  The
reaction rate constant, k, was determined from the slope of the line shown in
Figure 5, where AC/At  was  plotted  against  C^.  The  reaction  rate  is  0.024
L/mg*h and the refractory portion  is 40 mg/L sCOD.  The k value of 0.024
L/mg*h is similar to the k value of 0.021 L/mg-h obtained by determining  the
mean of the 14 individual runs at  LeSourdsville.
     The k value at LeSourdsville  also is similar to the k value obtained
from the Hynek data at the South Shore plant, which is 0.015 for mechanical
drive and 0.025 L/mg-h for air drive RBC.
     There are wide variations in  the analytical data from Indian Creek and
these may be attributed to the low level of sCOD in the influent and the
long residence time, which at times were over 6 hours.  The maximum sCOD
obtained at Indian Creek was 105 mg/L and only one sample out of 36 was above
100 mg/L.  Another factor that impaired the analyses at Indian Creek was  the
physical layout which consisted of only three stages.  This limited the number
of sample points and reduced the probability of determining a curve for repre-
senting the disappearance of sCOD.
     There were nine sampling dates at Indian Creek.  Of the nine dates,  four
were discarded because a curve could not be drawn that would adequately repre-
sent the data to describe the disappearance of sCOD.  Figure 6 is an example
of a wide scatter analytical result that could not be used in the data reduction.
Figure 7 is an example of the disappearance of sCOD with time that could  be
represented by drawing a curve to  represent the selected data.  For the five
dates that could be described by drawing the best curve for the disappearance
                                      216

-------
of sCOD with time, the reaction rate constant was 0.018 L/mg-h and the mean
sCOD refractory was 27 mg/L.  Again, the data from these'five runs showed a
                                                      *•
k value similar to the value at LeSourdsvil le and byHynek, even though the
level of influent sCOD was significantly below the sCOD levels at the other
locations.
     The result at Indian Creek behaves as though it were biochemical reaction
rate limited and the kinetics obey a second order rate'expression.  The
Indian Creek results show that the low level of sCOD  in the influent does
not alter its kinetic behavior, and obeys a second order rate expression, whose
reaction rate constant is similar to the values obtained at LeSourdsville and
with Hynek data.
Oxygen transfer limitation.
     During Hynek's test, four runs were operated at  a significantly higher
hydraulic loading rate, ranging between 2.1 and 2.9 gpd/sq ft  (86 to 120
L/d-m2).  This, in effect, reduced the reaction time  by approximately 50%.
To accomplish the same sCOD reduction at the high hydraulic loading, as was
obtained at. the low hydraulic  loading, would require  doubling  the oxygen
transfer rate and an adequate  level of biomass to handle the additional
sCOD removal requirements resulting from the increase  in the hydraulic
loading rate.
     The plot of  sCOO with time is represented by Run  VIII for the air
drive system and  is shown in Figure 8.  The data show a  linear relationship
for the disappearance of sCOD  with time.  This relationship indicates zero
order kinetics.  A possible explanation is; as the hydraulic  loading increased
there was insufficient time to transfer the oxygen required for  converting
the sCOD; thus changing the system from a biochemical  reaction limiting
process into an oxygen transfer limiting process; and  the kinetic rate
changed from a second order expression into a zero order expression.
                                   217

-------
     The Brookville, Indiana facility was sampled on  10 dates.  The data
reduction for the ten sampling dates resulted in an apparent oxygen limiting
operation.
     A plot of sCOD against time for the Brookville data showed that seven
of the ten dates could best be represented by a zero  order rate equation.
The data were combined to obtain a mean value of sCOD at each of the four
stages and the average retention time at each stage.  These values are
plotted in Figure 9 and show an excellent correlation for a zero order rate
equation.
     A comparison was made of the loading levels at LeSourdsville, Indian
Creek, and Brookville.  A significantly higher loading  is evident at Brookville
when compared with LeSourdsville or Indian Creek.
     The pseudo oxygen mass transfers were calculated for the first stage
of the RBC at LeSourdsville, Brookville, and  Indian Creek.  A sample
calculation for the pseudo oxygen transfer at LeSourdsville follows.
     The hydraulic loading at LeSourdsville averaged  0.82 gal/d-sq ft
       0.82  gal    x   d   x  [2 x 100,000 + 3 x 150,000] sq ft  x 3.8 L =
            d-sq ft    "Z4TT                                          "gal
       0.82  gal    x   d  x  650,000 sq ft x  3.8  L  =  85,000 L/h
            d:sq ft    24h                         gal
     The oxygen required to satisfy the disappearance of 54 mg/L of sCOD  in
the first stage is determined by:
     85,000  r   x  54 ma  x     1      =   37 mg_ 02  (4°0 "ig 02)
             h         r    100,000 sq ft  h-sq ft            ~
                                    218

-------
              Table 4.  Loading Levels  at  the  RBC  Facilities

Hydraulic load

ing, gpd/sq ft
(L/d-m2)
Influent, mg sCOD/L
Retention time
Pseudo oxygen
(first stage)
, h
transfer, mg02/h-sq

LeSourdsvil le
0.88
(36)
118
3.5
ft, 37
(400)

Bropkvil le
1.5
(62)
288
2.0
48
(520)
Indian
Creek
.5
(20)
65
6.0
9
(100)
     It is evident from this comparison that Brookville has 170% greater
hydraulic loading than LeSourdsvilie and the influent concentration in sCOD
is 240% greater at Brookville resulting in an exceptionally high organic
loading.  The oxygen mass transfer is assumed to be at a maximum, and the
limiting factor at Brookville appears to be oxygen transfer rate limited.
     If it is assumed that Brookville were limited by a second order rate
equation with a rate constant equal to the rate constant obtained at LeSourdsville,
0.021 L./mg/-h, then the disappearance of sCOD would follow the curve as
shown in Figure 10.  The concentration leaving the first stage, 0.5h reaction
time, is 123 mg/L.  The pseudo oxygen transfer rate to achieve this level
of reduction is 158 mg/h-sq ft (1700 mg/h-m2).  This is more than 4 times
the pseudo oxygen transfer rate calculated for LeSourdsville and 3 times
the actual rate calculated for Brookville.  It is for these reasons that
the oxygen transfer rate is believed controlling the reaction mechanism at
Brookvilie.
     The selection of the three facilities, LeSourdsville, Brookville and
Indian Creek, was based on the proximity of these sites to MERL.  Yet these
three facilities provide a good mix for this evaluation because of the wide
                                      219

-------
variation in their loadings.
     1.  LeSourdsville operates at an organic loading that appears to be
         within 20% of the upper limit for oxygen mass transfer rates and
         behaves as though it were biochemical reaction rate  limited.
     2.  Brookville operates at an organic loading that appears to be limited
         by the oxygen mass transfer rate.
     3.  Indian Creek operates at an organic  loading considerably below
         LeSourdsville and appears to follow  a biochemical reaction rate
         limting process, whose rate constant is similar  in value to the
         rate constant obtained at LeSourdsville and from Hynek.
     When the hydraulic load increased, as Hynek did in his evaluation,
then the process appears to change from a kinetically limited system to an
oxygen limited process.
     These results present a new approach in  the analyses of  RBC performance.
The applicability of a second order reaction  rate expression  to follow the
disappearance of the soluble organic fraction was demonstrated with Friedman's
pilot plant data, Hynek data, LeSourdsville,  and Indian Creek.
     The second order expression failed to follow the disappearance of sCOD
with time at Brookville, and with Hynek results when the hydraulic loading
was doubled.  These two operations obeyed zero order kinetics and were
assumed to be oxygen mass transfer limited.
                                  220

-------
     The similar k values obtained from RBC's at different field sites
treating municipal wastewaters indicates that the reaction rate constant,
k, can be used to predict the performance for RBC's when they are employed
for secondary treatment.  For a series of stirred tank reactors or RBC
stages, the' concentration of soluble organics can be determined at any
stage in the process by use of Levenspiel's equation(^) if the following
parameters are known:
     1.  Reaction rate constant based on sCOD or sBOD, L/mg-h
     2.  Residence time, h
     3.  Influent organic concentration, mg/L
     Levenspiel's equation for staged reactors that follow second order
kinetics is mathematically derived from a mass balance, and is applicable
for calculating the soluble organic concentration at any stage.  The equation
is:
        Cn  =  -1 +  1 + 4
                      2T¥t)
 where  Cn  =  concentration of soluble organics in n-stage, mg/L
        k   =  second order reaction rate constant, L/mg-h
        t   =  residence time, h
       Cn-l -  influent soluble organic concentration to stage n, mg/L
     This equation can then be programmed into a computer and by inserting
the number of stages, n, the initial concentration, Cn_i, the residence
time within each stage, t, and the reaction rate constant, k; the concentra-
tion, Cn, in terms of soluble organics can be readily obtained at any stage
in the process train.
                                     221

-------
     To test the applicability of second order kinetics to predict the
concentration of soluble organics in any stage of a RBC train, interstage
data was obtained from lanone(^) on the disappearance'of sBOD at 9 plants
using air drive RBC.  The results obtained by Hynek(2) and analyzed earlier
in this paper to obtain k values based on sCOD for both, air and mechanical
drive RBC also included interstage data on the disappearance of sBOD.  Hynek's
data with sBOD using air drive shafts were incorporated into the second
order rate expression to obtain a reaction rate constant for sBOD of 0.083
L/mg-h.  The k value was incorporated into Levenspiel's equation to predict
the sBOD concentration at any stage for each of the 9 air drive RBC plants.
These results are shown in Table 5, and are displayed with the actual results
for comparative purposes.
                                    222

-------
  Table 5.  Comparison of  the Predicted and Actual Disappearance of sBOD
         Cleves
   Shafts/Stage = 1-1-1
t(h) = 2.5, 2.5, 2.5
            Predicted   Actual
 Cin  =                   40
   Shafts/Stage = 1-1-1-1.5**
t(h) = 1.4, 1.4, 1.4, 2.2
            Predicted   Actual
 Cin  =                  218
                          78
                          22
                          14
                           8
Cl =
C2 =
C3 -
C4 -
39
15
8
4
78*
22
10
5
       Enumclaw
  Shafts/Stage = 3-1-1-1
t(h) = 1.4, .46, .46, .46
            Predicted   Actual
 Cin  =                  168
Ci
C2
C3


12
5
= 3

Lancaster
8
5
3


Cl
C2
C3
C4

34
20
. 13
10
Lower East Fork
14
9
7
6

                                                Shafts/Stage = 1.5-1-1-1
                                         t(h) = .97, .64, .64, .64
                                                     Predicted   Actual
                                          Cin =                     20
                                          Ci  =         11
                                          C2  =
                                          C3  =
                                          C4  =
                6
                5
  *Assume overloaded first stage and determine concentrations
    of sBOD in succeeding stages.
 **High density media
   C   = influent
11
 6
 5
                                     223

-------
Table 5.  Comparisonof the Predicted and Actual Disappearance of sBOD  (cont'd)
              Woodburn
    Shafts/Stage = 4-2-1.5-1.5
  t(h) = 1.69, .84, .63, .63
   Greenwood-Springs
 Shafts/Stage = 1-1-1-1.5
t(h) = .56, .56*, .56, .84
Predicted Actual
Cjn =
Cl =
C2 =
C3 =
C4 =

Predicted
226 Cin =
37
17
11
8
Dodgeville
28
12
7
7

CT
C2
C3
C4

22
13
9
6
West Dundee
Actual
43
20
14
4
5

    Shafts/Stage = 2-1-1
  t(h) = 2.6, 1.3, 1.3
           Predicted   Actual
  r.   -                 07
   i n  ~~
  C]          11          9
  C2   =       7          7
  Ca   =       4          4
 Shafts/Stage = 1-1-1,5
t(h) = .76, .76, 1.2
         Predicted   Actual
Cin  =                 101
Cl   =       33         33
C2   =       16         15
C3   =        9          8
                             Hartford
                          Shafts/Stage =  1-1-1-1
                         t(h) = .25,  .25,  .25
                                 Predicted   Actual
                        Cin  =                  17
                        Ci   =      13          13
                        C2   =      11          12
                        C3   =        9          9
                        C4   =        8          8
                                     224

-------
     There is good agreement between the predicted and actual sBOD at seven
of the nine plants.  There was a difference at Enumclaw and Lancaster.  The
calculation for Lancaster was modified by assuming an inadequate oxygen
transfer rate in the first stage, due to the high organic  loading, and then
applying second order kinetics to the following stages.  By using the actual
value of 78 mg/L sBOD, that was obtained in the second stage, as the  initial
concentration, and then calculating the sBOD in the ensuing stages, good
agreement was then obtained for Lancaster between the'predicted and actual
sBOD concentrations.  There is no explanation that can be  theorized at this
time for the discrepancy at Enumclaw.  An analysis similar to Lancaster is
not valid because the actual concentration of sBOD in the  first stage was
considerably below the predicted value, and therefore oxygen transfer require-
ments were satisfied at Enumclaw.
     These results provide added evidence that RBC obey second order  kinetics
and when the reaction rate constant is known, can be used  to predict  perfor-
mance, design optimum train configurations, and can be used to reduce capital
costs.
                                    225

-------
                                 References
1.  Friedman, A. A. "Kinetic Response of Rotating Biological Contactors,"
    31st Annual Purdue Industrial Waste Conference, 1976.

2.  Hynek, R. J., and Chou, C.S. "Development and Performance of Air Driven
    Rotating Biological Contactors," 31st Annual Purdue  Industrial Waste
    Conference, 1976.

3.  Levenspiel, 0. Chemical Reaction Engineering, John Wiley & Sons, 124-149,
    1972.

4.  lannone, J., Personal communication (Roy F. Weston), December 7, 1981.
                                    226

-------
   140
   120
   TOO
ra
E
   80
u
h-
<
CHL



1
J—
U
<
LU
   40
   20
                                              I         I          I
                        10000              20000


            CONCENTRATION SQUARED C*  (mg/lf
30000
        Figure 1.Reaction rate at various concentrations squared.

-------
                                                                0)
                                                                s:
                                                               Q
                                                               o
                                                               u
                                                                01
                                                                u
                                                                C
                                                                a
                                                                ^.
                                                                c
                                                                0)
                                                                a
                                                                a
                                                                o
                                                                in

                                                               a
                                                               >
                                                               w
                                                              ~a

                                                               o
                                                               u


                                                               a
                                                              _c
                                                               w
                                                               3
                                                               D5
                                                  3
                                                  O
                                                 _c
                                                                      39
                                                                      E
                                                                     a
                                                                     O
                                                                     u
                                                                      Ul

                                                                      O

                                                                      OJ
                                                                      a
                                                                      4)
                                                                      a
                                                                      a
                                                                      o
                                                                     0)
                                                                     >
228
                                                                     0)
                                                                     l~
                                                                     3
                                                                     CS

-------
          120
ro
ro
                        1          2         3
                            TIME, hours
                       Figure 4. LeSourdsville —
                    Disappearance  of sCOD  with time.
 120
                                                              100
                                                            J,  80
                                                            \
                                                            O)
                                                            E
                                                            if  60
                                                            Z
                                                            2
                                                            t—
                                                            u
  40
                                                               20
                                                                                       I
                                  1
            2000     4000
    CONCENTRATION SQUARED, (C f~ (mg/Lp
          Figure 5. LeSourdsville.
Reaction rafe al  various concenlralions squared.

-------
                                                                                o>
                                                                                £
                                                                 3
                                                                 O
              1/Bui  'QODS
                                                                                o
                                                                                O
                                                                                u
                                    u
                                    c
                                    D
                                    U.
                                    D
                                    en
                                    a.
                                    a,
                                    o
                                    tn

                                    Q
                 cc
                                                                               u

                                                                                M
                                                                                5
                                                                               N.

                                                                                (U
                                                                                i.
                                                                                3
                                                                                CS
                                                                 D
                                                                 O
                                   Q
                                   O
                                   U
                                                                               0)
                                                                               u
                                                                               C
                                                                               o
                                                                               w
                                                                               D
                                                                               CD
                                                                               a
                                                                               a
                                                                               o
                                                                              c
                                                                              o
o
CO
O
cs
a
u
                         230
                                                                             D
                                                                             09

-------
                           V*-
120
 100
                                                                      1.5
         Figure 8. Hynek — Disappearance of  sCOO  with time.
                                 231

-------
                                                            D
                                                            O
                                                                      0!

                                                                      E
Q
O
u
 m
*^»
 o

 CD
                                                                      O
                                                                      0)
                                                                      D.
                                                                      CL
                                                                      O
                                                                      v>

                                                                     5
                                                                      3
                                                                      CD
                                                                      c
                                                                      c
                                                                      c
                                                                      c
                                                                      u
                                                                     a
                                                                     o-

                                                                      Q
232

-------
          ASSESSMENTS OF THE KINETIC PERFORMANCE OF
            A ROTATING BIOLOGICAL CONTACTOR SYSTEM


          Ta-Shon Yu, Ph.D., P.E.
          Office of Environmental Programs
          State of Maryland


          Randolph G. Denny
          Office of Environmental Programs
          State of Maryland
INTRODUCTION

     The employment of a rotating biological contactor (RBC)
for wastewater treatment was pioneered by Hans Hartmann
and Franz Popel of Germany in 1955 on a scale of technolo-
gical research basis.  It had not been developed into the
extent of commercial applications until early 1970's when
the technological practice became economically competitive
with the activated sludge process.  From 1974 to 1980,
escalation of energy costs and abundance of federal funds
for construction in the United States prompted this waste-
water treatment technique into its prospective market place
within a short time frame during which the sales represen-
tatives of the biological contactor manufacturers were the
only authorities in the structural design as well as the
functional forecast.  As a result, the owners and operation
personnel associated with the biological contactors would
either take in a pride of prudent decisions in selection of
this specific  treatment process, or be dismayed by the out-
come of functional performance for the entire life span.
     Whether the rotating biological contactor process can
live up with the expectations of cost-effectiveness and
                             233

-------
and functional reliability should not be assessed solely
based on the numerous incidents of the structural failure
or simply based on the reports of successful performance
with short-time experience.  The manufacturers have been
pressed to improve the structural integrity for the reason
of business survival.  Only the time can tell if improve-
ments have been made to a satisfactory manner that requires
a functional life of at least 25 years to justify its cost-
effective claim.  The structural set-back could be viewed
as a typical problem of any technological transition.
However, it should be born in mind what damage can be done
with the business once the reputation is ruined.  Should the
biological contactor industry strive to stay in business, it
would be a viable wastewater treatment technique which
deserves a fair consideration.
     The first installation of the rotating biological con-
tactors in the State of Maryland is at the St. Michael
Wastewater Treatment Plant.  The design capacity is 0.5 mgd
to accommodate the projected needs for the year of 1990's.
This is a tertiary plant which consists of primary sedimen-
tation, biological treatment by rotating contactors,
secondary sedimentation followed by filtration, chlorination,
dechlorination and post-aeration.  It is designed primarily
to treat domestic wastewaters containing 240 mg/1 of BOD_
and suspended solids respectively to meet 20 mg/1 of BOD_
and 10 mg/1 of suspended solids as monthly average effluent
quality limitations set forth by the National Pollution
Discharge Elimination System  (NPDES) permit.  The design
criteria are shown in Table I.  It should be noted that the .
design of primary and secondary clarifiers is not intended
to be conservative, but to satisfy the performance relia-
bility which requires at least two units for each sedimen-
tation process.
     The plant operation was initiated in late 1979.  The
current flows approximate 0.25 mgd with 210 mg/1 of BOD,.
and 120 mg/1 of suspended solids on a yearly basis.  Because
of the current low flow conditions, one primary clarifier
and one secondary clarifier are in line with the remaining
treatment processes.  The biological contactors are 'driven
by 5-hp gearmotors with a rotating speed of 1.6 rpm.

PERFORMANCE STUDY

     Attempts were made to assess microbial behaviors of
the rotating biological contactors on performance of
                              234

-------
         Table I - Criteria Used for Design of the
                   St. Michael Wastewater Treatment Plant
Average Daily Flow
     Initial (1979)
     Design (1990)
Influent Characteristics

     BOD
     Suspended Solids

Primary Clarifier  (2 Units)
     Dimensions
     Surface Overflow Rate
     Detention Time
Biological Contactor (3 Units)
     Operation Mode
     Shaft Dimensions
     Surface Area - Each
     Nominal Volume - Each
     Nominal Detention Time - Each
Secondary Clarifier (2 Units)
     Dimensions
     Surface Overflow Rate
     Detention Time

Filtration
     Operation Mode
     Surface Area
     Filtration Rate
Chlorination
     Detention Time

Dechlofination/Post-aeration
     Detention Time
0.25 mgd
0.50 mgd
240 mg/1
240 mg/1
30' dia. x 10' SWD
350 gpd/sq.. ft.
5 hrs.
in series
25' x ll'-6"
100,000 sq. ft.
10,500 gal.
0.5 hr.
30' dia. x 8' SWD
350 gpd/sq. ft.
4 hrs.
continuous backwash
180 sq. ft.
2 gpm/sq. ft.
60 rain.
15 min.
                             235

-------
carbonaceous removal and nitrification so as to acquire
relevant information for optimal design.  Evaluations were
conducted under both normal and abnormal operating conditions
by analyzing samples taken from each stage of the biological
treatment.  The area of relative microbial activity at each
stage of the contactor was also exploited.
     The magnitude of pollutants permissible for discharge,
except for the toxic substances, is indicative of the assim-
ilative capacity of the receiving water through the natural
purification process to satisfy oxygen demands exerted by
the carbonaceous and nitrogenous compounds.  No matter these
bio-degradable compounds are soluble or insoluble, the
receiving water is obligated to replenish the total amount
of oxygen required until the assimilative capacity is
exhausted.  The current approach in evaluation of the perfor-
mance efficiency of the rotating biological contactor appar-
ently tends to place its importance upon removal of soluble
and readily oxidizable constituents.  This study is intended
to reiterate the significance of the fundamental principle
of pollution abatement related to the capability of the
biological contactor in removing insoluble bio-degradable
organic substances.
     The primary effluent was introduced into the biological
contactor in a direction perpendicular to the shaft.  The
compartment of each stage was so confined that the mixed
liquor in a practical sense represented a completely mixed
system.  Samples taken from the contactor compartments had
been allowed to settle for 30 minutes before the supernatants
were drained for laboratory analyses conducted by the
Laboratories Administration of the Maryland State Department
of Health and Mental Hygiene.  The analytical results of the
supernatants would provide the accessory information of
relative settleability of the mixed liquor suspended solids
in each stage of contactor in comparison with that of the
secondary effluent.
     The rotating biological contactor system installed at
the St. Michael Wastewater Treatment Plant was manufactured
by George A. Hormel & Co., EPCO - Hormel RBS Bio-Shaft,
Model M3707, Serial No. 179.  In two years operation, the
structural failures were experienced.  As a State regulatory
agency in approving construction contract plans and specifi-
cations and in implementation of plant performance, such
unwanted problems must be resolved.  In order to live up
with the expectation that the rotating biological contactor
                             236

-------
is a viable and dependable technique in wastewater treatment,
recommendations are made to control structural integrity
in the process of the construction contract procurement.

RESULTS AND DISCUSSIONS

              Under Normal Operating Conditions

     The current domestic flows at the St. Michael
Wastewater Treatment Plant average 0.25 mgd.  Figure 1
represents the typical pattern of the rotating biological
contactor performance in removal of carbonaceous compounds
and achievement of nitrification under normal operating
conditions.  It is interesting to note that the first stage
contactor is capable of performing two distinctly different
metabolic functions simultaneously.  The result indicates
that the heterotrophic micro-organisms responsible for BOD
removal and the autotrophic micro-organisms responsible for
oxidation of ammonia nitrogen co-exist on the same environ-
ment favorable for their growth and propagation.
     The bio-mass attached to the contactors is roughly
equivalent to 10,000 mg/1 of mixed liquor suspended solids
in the first stage compartment, 7,500 mg/1 in the second
stage compartment, and 3,750 mg/1 in the third stage
compartment.  The BOD  applied to the first stage contactor
approximates 100 mg/1 that is 200 pounds of BOD  at the flow
of 0.25 mgd.  The corresponding organic loading lies in the
neighborhood of 2 Ibs. BOD  / day / 1000 sq. ft. or 0.2 Ib.
of BOD  per pound of bio-mass.  The organic loading of this
magnitude is comparable to the operation mode of the extended
aeration process.
     In the presence of high concentrations of alkalinity
(200 mg/1 to 300 mg/lj and slightly alkaline pH conditions
(7.5 to 8.0), a complete nitrification can be expected at
temperatures above 10°C.  As the nitrification takes place,
it consumes approximately 8 mg/1 of alkalinity for 1 mg/1
of ammonia nitrogen oxidized.  Other than the favorable
environmental factors with respect to alkalinity, pH and
temperatures, the successful nitrification may have, been
attributed to the low BOD loading which refrains the
heterotrophic micro-organisms from rapid growth to the
extent that permits Nitrosomonas and Nitrobacters to
reproduce themselves.
     As shown in Figure 1, the second and the third stages
of contactors contribute little wastewater treatment under
                             237

-------
   300 <*€}
   250
   200   ->
M
2?  150
8
CQ
   100
                                                    30
    50
                                                    20
                                                    10
1
z
Figure  1  -  Metabolic  Responses  on Oxidation of Carbonaceous
            and  Nitrogenous  Compounds Under Normal Operating
            Conditions
                           238

-------
the normal operating conditions.  It is conceivable that the
first stage contactor alone will be able to treat the same
characteristics of sewage at the design flow of 0.5 mgd under
the organic loading condition of 4 Ibs. BOD  /day/ 1000
sq. ft.  The question arises as to whether this single-train
system with three stages in series was over-designed or was
intended to provide the necessary redundant capability for
operations.  It is of the opinion that if the structural
reliability is sound, the second stage contactor should be
incorporated into the single-train system with a reserved
capacity to treat the unexpected peak or concentrated waste-
waters; however, if the structural reliability becomes
questionable, there is no room for,criticisms against a
single-train system equipped with three stages of contactors.

              Under Abnormal Operating Conditions

     The rotating biological contactor system at the
St. Michael Wastewater Treatment Plant has experienced both
mechanical problem and structural failure.
     Approximately one year after the system was installed,
the first stage contactor's shaft bearings had to be replaced.
The suspected cause of the bearing failure was thought to be
due to the drainage of the mixed liquor down on the shaft
and into the bearings.  The problem was corrected by putting
a bead of silicone rubber around the shaft to divert the
mixed liquor from entering the bearings.
     The system had been operated in a satisfactory manner
for two years until a severe structural failure developed
in the early winter of 1981.  The tie rods holding the
individual polyethylene discs of the first stage contactor
began to shear and dismember the disc assembly.  This
problem caused noise and shaft vibration and the unit was
taken out of service as the result.  Nevertheless, in an
attempt to alleviate the possible development of a differ-
ential torque applied to the shaft caused by non-uniform
microbial growth, it was managed to operate the first
stage contactor for 10 minutes twice daily under the
stressed crippling conditions.
     During the down-time, the primary effluent continued
passing through the first stage compartment.  The principal
responsibility of wastewater treatment depended to a great
extent upon the second stage contactor.  The metabolic
responses to the abnormal operating conditions shortly after
shut-down of the first stage contactor are shown in Figure 2
                              239

-------
   300
   250
   200
 a
 a
 o
 CD
   100
    50
                           STAGES



Figure 2 - Metabolic Responses on Oxidation of Carbonaceous

           and Nitrogenous Compounds Shortly After Shut-down

           of the First Stage Contactor
                           240

-------
which indicates that the carbonaceous and nitrogenous
removal rates were considerably low at the immediate juncture
of the transition state.  It is also noted in Figure 3 that
the rates of BOD removal and nitrification achieved by the
second stage contactor were lower than the corresponding
rates accomplished by the first stage contactor under normal
operating conditions.  Notwithstanding the disruption of the
first stage contactor operations, the over-all efficiency
of the system performance in every respect remained excep-
tionally high.  Such an accomplishment of a high degree  !
treatment should be credited to the second stage contactor
and the third stage contactor as well.
     Since the shut-down of the first stage contactor, it was
found that the bio-mass on both second stage and the third
stage contactors was gradually developed.  This natural
phenomenon reflected higher organic loadings being applied
to them.
     In order to .prevent an anaerobic environment from
development and to prevent sedimentation from taking place in
the first stage compartment, the operation personnel decided
to remove the partition between the first stage and the
second stage compartments two weeks after shut-down of the
first stage contactor.  This arrangement would permit
fluxing the wastewaters in a common compartment in which
oxygen was supplied and sedimentation was prevented as a
result of the second stage contactor operations.
     The metabolic responses to the operation improvement,
as presented in Figure 4, illustrate that the metabolizable
components of the carbonaceous and nitrogenous compounds
were readily removed in the common compartment of the first
stage and the second stage contactors.  However, the low
temperature at 9°C either curtailed the capability or
diminished the population of the autotrophic micro-organisms
to achieve nitrification.  A slight reduction of ammonia
nitrogen was reasoned on the grounds for supporting microbial
growth in the processes of catalSolism and bio-synthesis.
     Several weeks later, the second stage contactor
experienced the same problem.  It was decided that the entire
system should be taken out of service and repaired.

                 •Modes of Substrate Removal
     The primary effluent contains approximately  100  mg/1 of
BOD  in which  25 mg/1 to 35 mg/1 are  soluble and  65 mg/1  to
                             241

-------
        150
        100
 i
 Q
 O
 B»
 B
 z
         50
         40
         30
         20
         10
                              I        I        I
                             O   Normal Operation
                              I
                                 Abnormal Operation
                                      i        I
                                Normal Operation
Abnormal Operation
                             STAGES
Figure 3 - Comparison of Metabolic Rates:  First Stage
           Performance Under Normal Operating Conditions
           versus Second Stage Performance Under Abnormal
           Operating Conditions
                          242

-------
     300
 x.
 CF>
 E-
 M

 2
 Q
 O
 CQ
     250
    200
    150
100  J
                                                 20
                                                     10
                                                      z
                                                      M
                                                      H

                                                      s
                                                      O
                                                      M
                                                      Z
                                                      o

                                                      co

                                                      n
                                                      o
                                                           PI
                                                           z
                                                           H
                                                           CO
Figure 4 - Metabolic Characteristics After Removal of

           Partition Between  First  Stage and Second Stage

           Compartments
                            243

-------
75 mg/1 are insoluble.  It also contains approximately 30
mg/1 of Total Kjedahl Nitrogen  (TKN) in which 20 mg/1 to 25
mg/1 are ammonia nitrogen and 5 mg/1 to 10 mg/1 are organic
nitrogen.  Since the heterotrophic and the autotrophic
micro-organisms contained in the bio-mass are not distin-
guishable, the loadings cannot be meaningfully expressed
on the mass.ratio basis. The term expressed as "Ib./day/
1000 sq. ft." for the various loading conditions applied to
the first stage contactor are given in Table II.

     Table II - Various Loading Conditions Applied To
                The First Stage Contactor At 0.25 MGD

  Constituents               Loadings (Ib./day/10  sq. ft.)

  Soluble BOD                          0.5 to 0.7

  Insoluble BOD                        1.3 to 1.5

  NH -N                                0.4 to 0.5

  Organic - N                          0.1 to 0.2

     Nitrosomonas and Nitrobacter are chemosynthetic nitri-
fiers, a kind of autotrophic micro-organisms.  The biosyn-
thesis is undertaken through utilization of energy supplied
by oxidation of ammonia.  On the contrary, the heterotrophic
micro-organisirs metabolize organic carbon as well as nitrogen
and release nitrogen as ammonia which can be further oxidized
by nitrifiers.  The degree of nitrification of a heterogeneous
microbial system is the measurement of the nitrifiers' capa-
bility to convert TKN into nitrate,
     Under the normal operating conditions, as shown in
Figure 5, the heterotrophic micro-organisms on the first
stage contactor swiftly remove carbonaceous compounds of the
constituents in forms of soluble BOD or insoluble BOD, while,
organic nitrogen remained essentially untouched.  At the same
time, nitrifiers readily oxidized ammonia nitrogen.  With
respect to carbonaceous metabolism, the result indicates
that the carbonaceous compounds required for the hetero-
trophic micro-organisms exceeded the amount of soluble BOD
available as the low concentrations of soluble BOD failed
to exert inhibitory effects on microbial utilization of
insoluble BOD. Consequently, soluble BOD and insoluble BOD
were removed concurrently. On the other hand, the pattern of
nitrogen metabolism displayed a sequential mode.  This
phenomenon can be deduced as the result that the amount of
                            244

-------
        150
  EH
  2
  UJ
  2
  O
  U
  O
  H
  I
  m
  u
  O
  Or

  8
  O
  a
100
 50
         10
Figure 5 -1  Metabolic Responses on Oxidation of Carbonaceous
            and Nitrogenous Compounds Under Normal Operating
            Conditions
                             245

-------
ammonia nitrogen in excess of what was required for metabolic
needs inhibited the heterotrophic micro-organisms from
further degradation of organic nitrogen.
     Under the abnormal operating conditions, before the
bio-mass on the second and the third stages of contactors
was fully developed, the microbial activities decreased
significantly.  The slow metabolic rates provided an avenue
to gain insight into the microbial behavior on the mode of
substrate removal.  As shown in Figure 6, it is evident that
removal of insoluble BOD took place immediately after soluble
BOD had been utilized.  This mode of sequential substrate
removal reflected the metabolic responses from the samples
taken when the first stage contactor was not in service. The
submerged heterotrophic micro-organisms utilized soluble BOD
and by-passed insoluble BOD to the second stage contactor,
where soluble BOD was not available and the heterotrophic
micro-organisms must metabolize insoluble BOD for survival.
     Figure 7 portrays a similar sequential mode of nitrogen
metabolism as that illustrated in Figure 5.  It is reasonable
to conclude that only an inappreciable amount of organic
nitrogen removal can be expected by the rotating biological
contactor process when the wastewater contains an excessive
amount of ammonia nitrogen.  The inherent nature of a short
detention time provided for biological treatment of waste-
water also plays an important role in limiting microbial
degradation of organic nitrogen.  The combined effect of
metabolic inhibition and short reaction time causes removal
of organic nitrogen ineffective.  Even if the environmental
factors favor nitrification, achievement of nitrification in
a large measure depends upon the amount of organic nitrogen
contained in the wastewaters.  In order to assure a greater
degree of nitrification, organic nitrogen should be removed
by the sedimentation process which proves to be the most
effective and simplest means of treatment.

                  Evaluation of Kinetics

     There are two unique features imbeded in the rotating
biological contactor treatment process: (1) the predominating
micro-organisms differ from one stage to another due to
substrate gradient distribution, and (2) the mixed liquor in
each stage of compartment displays a complete mix system due
to a through agitation in s confined reactor.  With these
two inherent features coupled with a continuous flow pattern,
assessment of kinetic performance within a specific stage
of contactor beconres a matter of art of which beauty is in
                           246

-------
      200
 en
 to
 H
 &J
 o
 u

 to
 o
 o
 w
 CJ
 oa
 IX
 <
 0
      150
      100
Figure 6 - Metabolic Responses on Oxidation  of Carbonaceous

           Compounds by Heterotrophic  Micro-organisms Under

           Abnormal Operating Conditions
                           247

-------
        50
 o
 Q.
 O
 o
 o
 cc
        40
       30
       20
       10
                               STAGES
Figure  7- Metabolic Responses on Oxidation of Nitrogenous

           Compounds by Autotrophic Micro-organisms  Under

           Abnormal Operating Conditions
                           248

-------
the eyes of the beholder.  The metabolic rates among stages
portray responses of various predominating groups of microbial
population to specific loading and environmental conditions.
The environmental factors and the characteristics of the
wastewaters which vary from time to time determine selection
of certain predominating microbial species to grow on various
stages of the contactors.  Unless those influencing elements
can be properly controlled, the kinetic order merely reflects
the shape of a specific metabolic rate curve and the kinetic
value simply stands for a numerical figure.  No meaningful
engineering application in the process design for wastewater
treatment is expected.
     The curves plotted in Figure 1 through Figure 7 are
illustrations of the concentration changes in wastewater
constituents from stage to stage.  A line between two points
where a slope exists, should not be construed as an implica-
tion of a gradual decrease or increase in the concentration
of a specific constituent, because each compartment is a
completely mixed reactor in which the concentration gradient
does not exist.  In order to convey this concept, all data
points shown in Figure 1 and Figure 2 are respectively plotted
in Figure 8 and Figure 9.  The sampling points on the desig-
nated line number are explained below:

   Line Number         Location of Samples Taken

        L               Primary Influent
        L               Primary Effluent
        L               First Stage Compartment
        L               Second Stage Compartment

        L               Third Stage Compartment
        L               Secondary Effluent
         o
     The primary and the secondary clarifiers are designed
on the plug flow pattern.  The changes in the concentration
gradient are best represented by the lines connecting data
points on L  and L  or L_ and L .  Nevertheless, the represen-
tative lines for tne, rotating biological contactors' perform-
ance should be drawn' horizontally from points on L  to L ,
L  to L  and L  to L , and then vertically connecting points
on L , L  and L  to where the horizontal lines intersect. In
agreement with this concept, the metabolic responses should
reflect zero order kinetics.
                                249

-------
                                                         H
                                                         Z
                                                         z
                                                         o
                                                    10
                           STAGES
Figure 8 - Kinetic Performance at Various Stages Under
           Normal Operating Conditions
                            250

-------
      300
      250
      200
 Z
 H
      150
 Q
 O
      100
                                                      60
                              STAGES


Figure 9 - Kinetic Performance at Various Stages Under

           Abnormal Operating Conditions
                              251

-------
No matter how the rotating biological contactor process is
designed, the engineers can control neither waste character-
istics nor environmental factors.  The design should there-
fore be based on the operation experience as well as the
emperical equation to size the unit.
     Wu and Smith (1) developed an emperical model based on
full scale operations to predict the over-all system perform-
ance and assist engineers in the process design.  The Wu's
model as shown below describes the relationship between
percent BOD removal and percent BOD remaining as a function
of process variables including surface hydraulic loading,
influent soluble BOD concentration, number of stages, and
temperature.
  Wu's Model
                   14.2
        F = -     - T  0.6837 - 67247^ --- Equation 1
                   x L         x T
                      o
  where,

        F = fraction of influent soluble BOD remaining in
            the effluent, %

        q = surface hydraulic loading, gpd/ft

        N = number of stages

        L = influent soluble BOD concentration, mg/1
                         o
        T = temperature,  C


     If the Wu's model represents a general characteristic
of the system performance, the relationship among variables
should be independent on the number of stages. Therefore,
the model can be generalized as Equation 2.

                     . .  _    0.5579
                     14.2 x q
         F  = 	 — Equation 2
          n     0.32     0.6837   ,,,0.2477
               e     x L        x T
                        n
                             252

-------
   where"; n = footnote referring to stage number

     The number of stages can be determined by repeating
calculations of Equation 2 until L    meets the discharge
quality limitations.

   For example:

        1st stage -  Use L  to find F

                     L2 = L1X>1

        2nd Stage -  Use L  to find F

                     L3 ' L2 X F2
        N stage   -  Use L  to find F
                          n          n .
                     L  , = L  x F  = discharge quality
                      n+1    n    n    .  .    :
                                      limitation

     When Equation 1 is rearranged to solve N, Equation 3
is obtained.
   N = 3,125 x log
                        14.2 x q
                                0.5575
                             0,6837    0.2477
                      F x L         x T
                           o
— Equation 3
     A paradoxical relationship between N and L  is found in
Equation 3, i.e., the number of stages required decreases
as the concentration of soluble BOD increases, when variables
q, F, and T are constant.  This relationship can be explained
by the fact that the higher concentration of influent BOD
stimulates higher microbial activities and the percent of
BOD remaining can be easily maintained.  As a result, it
requires fewer contactors for treating wastewaters with
higher concentrations of BOD than the number of contactors
needed for treating wastewaters with lower concentrations of
BOD in order to achieve the same degree of percent BOD
reductions.
     Clark, Moseng and Anaso (2) developed a complete -
mix model and claimed that the principle of Monod's Equation
should also apply to each stage of the contactor at the
steady state.

  Clark's Model
                           253

-------
                         K       1      1
        AW = F(So~ Sl) (~PJ~ X - S~  + ~P~]~~ E<3uation 4
        P  = (u    / Y )  X  ----------------- Equation 5
              'max    a   a

   where, A   = wetted area of bio-disc, m
          F   = wastewater flow rate, 1/s
          S   = influent substrate concentration, mg/1
          S   = effluent substrate concentration, mg/1
          K   = the Monod half-velocity coefficient, mg/1
           o
          P   = area capacity constant, the amount of
                substrate removed per day per unit surface
                area of disc
          u   = maximum specific growth rate for the attached
           max  i_-       / j
                bio-mass / day
          Y   = apparent yield of suspended oraganisms
           a
          X   = concentration of suspended organisms,
     When Equation 4 is rearranged to solve S ,  Equation 6
is obtained.
        Sl -
                               2      2      $
              [AWP + FK  - FS )   + 4 F S K P]
                               2 F
                       (V + FV
                                           	 Equation 6
                               2 F
     Equation 6 can be generalized and expressed as
Equation 7.
                             254

-------
    nSl ~
                                  22
            (A P  +F  K  -F  S)+4F   S   K  P ]
              wn     n s     no        nonsn
                               2F
                   (AP+FK-FS)
                     w  n     n s     no
                              2F
                                            —  Equation 7
   Where, n = footnote referring to stage number.

     The number of stages can be determined by repeating calcu-
lation of Equation 7 until  S  meets the discharge quality
limitation.
     Application of these models is restricted to the soluble
BOD system.  Such a restriction brings about a serious question
as to their validity for design purposes, when insoluble BOD
must be removed and the ratio of soluble BOD to insoluble BOD
is not available.  The complications are further extended to
the system where sequential substrate  removal occurs.
     In application of the Clark's model, the fundamental
problem lies in the fact that the rotating biological contac-
tor process has never been operated under a steady state.
Consequently, u   , K , X , and Y  cannot be easily determined
  . , .           max   s   a'      a         ,.      :  ,  _
within a reasonable range of accuracy.   In fact, the informa-
tion relevant to u   , K , X  and Y  may not be available at
                 'max   s   a      a
the design stage.
     The manufacturers (3) published various charts which
correlate mass .loading with hydraulic  loading to predict
effluent quality under specific influent wastewater character-
istics and temperature conditions.  These charts have been
widely acceptable because of their simplicity in usage.  The
charts were developed on the assumptions that insoluble BOD
and soluble BOD would be removed concurrently at the same rate,
and the ratio of these two components  was 1. These assumptions
may not present a problem for domestic wastewater treatment
design because' of low substrate concentrations in both
insoluble BOD and soluble BOD.  However, it is hard to
comprehend that the same charts can also be applicable to the
                               255

-------
design of an industrial wastewater treatment process without
a pilot plant study.
     In accordance with the design procedures, the total
surface area required is calculated by dividing the design
hydraulic loading  (gpd/ft ) into the average design flow  (gpd)
The hydraulic loading is in turn figured from a chart showing
hydraulic loading rate  (gpd/ft )' vs. effluent BOD concentra-
tion (mg/1).  There are two linear relationships existing
between hydraulic loading rate and effluent concentration:
one above 15 mg/1 of soluble BOD and the other below 15 mg/1
of soluble BOD.  The design manual did not explain why the
number of 15 mg/1 was so magic as to render the microbial
population to behave differently in the process of metabolism.
The existence of linear relationship claimed by the manufac-
turer is principally in contradition to the Wu's and the
Clark's models.
     If the required total surface area is proportional to
the hydraulic loading rate, it implies that the microbial
population will uniformly grow on the surface'of the contac-
tors and the metabolic rates will be identical among the
stages.  Of course, the manual for design purposes may not
be intended to address the kinetic matter.  Nevertheless, it
may consequently overload the up-stream stages and underload
the down-stream  stages of the contactors.  In order to
achieve the most cost-effective design and the most efficient
operation, the flow distributions into parallel trains must
be carefully arranged.  For example, the treatment capacity
of a system consisting of 4 stages in 2 trains is not as
great as a system consisting of 2 stages in 4 trains.  The
latter arrangement not only distributes a great magnitude of
the organic loading into the four first stage contactors of
which the operation reliability can be backed up by the four
second stage contactors.  In addition, it may also avoid the
overloading condition imposed upon the four first stage
contactors.
     All models were developed under different theoritical
assumptions.  It is impossible to correlate and express them
in an explicit mathematical language.  However, the model
makers confidently insist that the rotating biological contac-
tor wastewater treatment plants can be easily and precisely
designed and performed in accordance with the models.  This
conclusion may be statistically correct without guaranty,
because there are numerous factors uncontrollable. The impor-
tance of the water pollution abatement program is what quality
of the plant effluent discharges, not what model is based for
                             256

-------
the plant design.
     Historically, many models have been developed for the
design of the activated sludge process.  With the valuable
information given, engineers still felt uncomfortable to use
a specific model because they must consider all factual
conditions and include built-in redundancy as required by the
governmental guidelines or regulations.  As a result,•almost
all designs on the activated sludge process followed the
established criteria of the organic and hydraulic loadings.
If the history repeats itself, engineers are bound to adopt
the same kind of criteria published by the manufacturers for
the future design of the rotating biological contactor process,

              Recommendations .forStructure Design

     In the name of cost-effectiveness, the public has been
led to believe that the rotating biological contactor
technique would be a dependable wastewater treatment process.
In fact, the application history has been too short to assess
its success or to condemn its failure, especially many systems
in operations have not reached the design loading conditions.
At the early stage of the market promotion, few consulting
engineers undertook stress range analysis of the rotating
biological contactor structures.  This caused general concern
as well as disappointment of the technique dependability.
     The reliable plant performance lies in the structural
dependability of which the importance cannot be over-
emphasized.  Historically, the shaft has been the main issue
of the problem.  In order to meet the quality of the struct-
ural design, it is strongly recommended that the maximum
stress range for the main central shaft, stub shafts and all
weldments to the shaft shall not exceed the allowable values
defined under American Welding Society Inc.'s Structural
Welding Code - Steel, AWS D 1,1 - 81 for a minimum fatigue
life of 25 years.  The stress range is defined as the peak-to-
trough magnitude of stress fluctuations.  In the case of
stress reversal where the rotating biological contactor shaft
applies, the stress range shall be computed as the numerical
sum (algebraic difference) of maximum repeated tensile and
compressive stresses, or the sum of shearing stresses of
opposite direction -at a given point, resulting from changing
conditions of load.  The stress range shall be determined
using calculated dead loads, torsion loads, and live loads
corrected for buoyancy using actual media percent submergences
and the appropriate AWS projected curve category for the
                              257

-------
tubular  structures,  as outlined  in  AWS  D  1.1  -  81,  Chapter  10,
Section  10.7.  The  live  loads  shall be  based  on a bio-mass
thickness of 0.125"  for  the  standard density  contactors  and
0.075" for  the high  density  contactors.   The  most important
of all is that the manufacturer  shall submit  the design  calcu-
lations  to  the consulting engineers at  the  time of  the shop
drawing  approval to  substantiate compliance.
     Failures associated with  the media have  also been report-
ed. An equal distribution of flows  to various trains of  the
rotating biological  contactor  system should help alleviation
of developing a thick layer of bio-mass on  the  contactor media
and help preservation of the media  stiffness.   The  manufac-
turers for  the sake  of business  survival  should improve  the
media durability and resistance  to  temperature  as well.

CONCLUSIONS

     The rotating biological contactor process has demonstrated
its capability in removal of soluble BOD and oxidation of
ammonia.  When the metabolic rates are high, soluble BOD and
insoluble BOD can be removed concurrently.  However,  when the
metabolic rates are low,  soluble BOD becomes a preferred
carbonaceous component for metabolisms.
     With respect to the microbial  responses to the nitrogenous
compounds, ammonia is readily oxidized or utilized by the micro-
organisms. While, organic nitrogen cannot be catabolized to an
appreciable extent in the presence of ammonia in excess of  the
amount required for the metabolic needs. As nitrification takes
place, oxidation of 1 mg/1 of ammonia nitrogen  consumes about
8 mg/1 of alkalinity. When temperatures stay above  10  C and
other favorable environmental factors prevail,' a complete
oxidation of ammonia can be expected. On the other hand, when
temperatures fall below 10 C regardless of  other environmental
conditions, a complete oxidation of ammonia cannot be achieved.
     In  cognizance of the process limitations,  cautions must
be exercised in evaluations of its  treatability toward removal
of insoluble BOD and oxidation of organic nitrogen.  It deems
necessary to conduct a pilot plant  study and determine if the
rotating biological contactor is an applicable process for  the
treatment of industrial wastewaters.
     The primary treatment is not a prerequisite in conjunction
with the rotating biological contactor process, but the capa-
bility of primary clarifiers in  removal of  insoluble BOD and
organic  nitrogen is  too great to be ignored. The rotating bio-
logical  contactor process in line with the  primary  treatment
                             258

-------
definitely improves the efficiency of the over-all plant per-
formance .
     The mixed liquor suspended solids generated from the
rotating biological contactor process settle rapidly. A deten-
time of 30 minites proves to be adequate for the sedimentation
purpose. In considerations of flow fluctuations, it is recom-
mended that a detention time of 30 minutes be provided to
accommodate the peak flow rate entering the secondary sedimen-
tation process. This unique settling characteristic will result
in cost savings for the construction of secondary clarifiers.
     As wastewaters enter the rotating biological contactor
process in a direction perpendicular to the shaft, the mixed
liquor in each compartment represents a completely mixed system.
The metabolic response to a certain substrate component in each
compartment should follow zero order kinetics under a continuous
flow condition. A great effort has been made to develop models
for the design and operation guidance. Before a specific model
is used for engineering applications, the model's practical
implications and built-in limitations must be fully understood.
     It is known to all that the metabolic activities are sub-
stantially high at the upstream stages, while, substantially
low at the downstream stages. A good engineering practice
requires the following considerations: (1) how to maximize the
over-all performance efficiency,  (2)  how to minimize the unex-
pected organic over-loading condition, (3) how to prevent
occurrence of the oxygen deficit condition, and (4) how to
increase an additional redundant capability at a minimum cost.
These ideal goals can be accomplished by promoting parallel
treatment schemes through flow distributions to as many trains
of contactors as possible, and by planning future expansions
in phases as the need arises.
     There is no doubt that the rotating biological contactor
is one of the viable alternatives for the treatment of waste-
waters. The past history in many instances has not proved its
structural dependability. Manufacturers are urged to make all
necessary improvements so that the technological reputation can
be built on an unshakable foundation.
     Engineers are indebted to their clients for the fiduciary
reward in expectations of the service being rendered with the
highest degree of professionalism. Responsibilities and obli-
gations must be fulfilled at both the design and construction
stages. The shaft design should meet the minimum requirements
as outlined in AWS D 1.1 - 81, Chapter 10, Section 10.7. The
live load should be calculated on the basis of a bio-mass
thickness of 0.125" for the standard density unit and 0.075"
for the high density one.
                              259

-------
REFERENCES

1. Wu,Y.C.,Smith,E.D.,and Hung,Y,T.," Modeling of Rotating
   Biological Contactor Systems ", Biotechnology and
   Bioengineering, Vol. 12, pp. 2055-2064, 1980

2. Clark,J.H.,Moseng, E.M., and Asano,T.," Performance of a
   Rotating Biological Contactor Under Varying Wastewater
   Flow ", Journal Water Pollution Control, Vol. 50, pp.
   896- 911, 1978

3. Autotrol Wastewater Treatment Systems - Design Manual,
   Autotrol Corporation, 1979
                              260

-------
       THE KINETICS OF ROTATING BIOLOGICAL CONTACTORS
            AT TEMPERATURES: 5°C, 15°C, AND 20° C
     Abraham  Pano.  Culp-Wesner-Gulp  Consulting  Engineers,
     Denver, Colorado.

     E. Joe  Middlebrooks.  Newman Chair Professor, Department
     of Agricultural  Engineering,  Clemson  University,  Clem-
     son,  South Carolina.
INTRODUCTION

     Rotating biological  contactors  (RBC) treating municipal
wastewater  have  been  shown  to  be  efficient in  carbon and
ammonia' nitrogen  removal  (1,2,3).  In  recent  years  in the
U.S., the use of the RBC  process has increased mainly because
of the simplicity of operation and the low power consumption.
     The  design  of RBC  systems  has been  based  primarily on
empirical  relationships  between the  pollutant  removal  effi-
ciency  and the  hydraulic  loading  rates based on  the  total
surface  area of the  RBC.  Presently,  the  design  hydraulic
loading rates are  adjusted  by a  safety factor for wastewater
temperatures  below 12.8°C  (55°F)   (4).  The  employed  safety
factor  generally varies  according  to the  RBC  manufacturer
recommendations,  because  of lack  of established kinetic con-
stants  associated  with  RBC  substrate removal  at  different
temperatures.  Also  there  is  little  information  available
concerning  the  effects of  staging on the kinetic constants
associated with RBC substrate removal.
     The  existing  data  from  RBC  studies  generally  indicate
that  the  kinetics  for   carbonaceous  substrate removal and
                               261

-------
 ammonia  nitrogen  removal  are  first  order,  with  substrate
 limiting phenomenon  (1,5,6,7,8,9).
     Several  studies  (10,11,12)  employed Mbnod  kinetics  to
 describe  carbonaceous substrate removal  in  fixed film  reac-
 tors.  Kornegay and  Andrews based their  model  on a  constant
 amount  of  active  attached  biomass  (10), Clark,  Moseng  and
 Asano  (11)  used 70  percent  of  the total attached biomass  to
 determine  the  kinetic  constants  for  Monod  growth  kinetics.
 Mikula  (12)  based  his kinetic  model on  the  total  attached
 biomass  and  the  biomass  in  suspension.  Other investigators
 developed  conceptual models  (13,14,15)  incorporating funda-
 mentals  of  substrate  and   oxygen  diffusion  and biological
 reaction.  Friedman  and  his  co-workers  (16,17)   used  a mass
 transport  model to  determine the kinetic constants of sub-
 strate removal  in  an RBC unit.  Also ammonia nitrogen removal
 in  RBC units  was  described either by  Monod growth  kinetics
 (18),  or  by mass transport  models  (19),  Some  of  the studies
 mentioned   above  were  conducted  with   synthetic   substrate
 (10,16,17,18)  and  others at fluctuating wastewater  tempera-
 ture (11,12).
     The general objective  of this study was to determine  the
 kinetics of carbon and ammonia nitrogen removal as a  function
 of  temperature  in an RBC  system  treating domestic wastewater.
     The specific objectives were:
     1. To  develop  kinetic  models  for  different  processes
 associated  with carbonaceous and ammonia nitrogen removal  in
 the first and  following stages of  an RBC system.
     2, To  determine the kinetic  constants  for each process
 at  each stage and each temperature.
     3. To determine the  effect  of temperature on the kinetic
 constants.

 MATERIALS AND METHODS

     Four  experimental rotating biological  contactor   (RBC)
 units were  operated  from  late October, 1979, until mid-July,
 1980,  in  the  laboratories  of Utah  State University, Logan,
 Utah  (20) .  The  study  was  conducted  in three  consecutive
 phases  at  three  different  temperatures  of  5°C,  15°C,  and
 20° C. Bach  phase was started with "clean" RBC units  (without
 bioraass).   Table I contains a summary of the detailed dimen-
 sions  of  the RBC units  employed during the three  phases  of
 the study.
     Comminuted wastewater  was  collected at  the Hyrum,  Utah,
wastewater  treatment plant,  and  hauled to the  laboratory  for
                                262

-------
        TABLE I. SUMMARY OF THE DIMENSIONS OF THE RBC
                 EXPERIMENTAL UNITS
Phase
Parameter
Number of stages
Number of discs/stage
Discs diameter, cm
Inflation factor
Side discs diameter, cm
Total surface area/stage, m^
Water volume/stage, liter
Submergence, %
Rotational speed, rpm
1

4
4
37.
1.
22.
1.
6
33.
16




5
37
9
375

3

2,3

4
4
39.0
1.37
22.9
1.474
7
35.5
16
use  as the  influent to  the RBC  units.  The  wastewater was
stored in a refrigerated tank with the temperature controlled
at 2°C.
     The  experimental  units were  operated  continuously  at
constant  influent  flow  rates,  constant wastewater percentage
and  constant  temperature.  The .influent  wastewater was main-
tained  at a  constant  temperature,  and  the experimental RBC
units were located in a constant temperature room to maintain
the  desired water  temperature  through  the four stages .of the
RBC  units. A  schematic  diagram of the experimental apparatus
is shown  in Figure 1.
     Table II  contains  a summary of the operating conditions
used during  the study.  Table  III  contains a  summary  of the
mean  liquid  temperatures  in the various stages  of  the four
experimental units. There was a gradual decline in the liquid
temperature due  to evaporation heat losses as the wastewater
flowed through the RBC units.
     Table IV  contains  a  summary  of  the mean  pH  values and
dissolved  oxygen  concentrations  measured  in  the  various
stages  of the  four  experimental units.  An  examination  of
Table  IV  shows  the  units  were operating as  an aerobic bio-
logical system,
     The  influent  to the  system and the  effluent  from each
stage  was  monitored  by collecting  24-hour  composite samples
at  20-minute  intervals  during  the  period of  steady-state
operation. Temperature,  dissolved  oxygen and  pH  values were
measured on grab samples.
     The ampule  technique (21)  was used to measure both total
and  filtered  COD.  Nitrogen  compounds  (Kjeldahl,  nitrate and
                               263

-------
                                                                                         TAP  WATER
cn
                                                                                 WATER
                                                                                 HEATING
                                                                                 CHAMBER
                                                                                         RBC
                                                                                         UNIT A
                              DRAINAGE
                                Figure 1.   Schematic diagram of  experimental apparatus.

-------
                               TABLE  II.   SUMMARY  OF       STEADY-STATE OPERATING CONDITIONS
cr>
en
Temperature,
°C Parameter
3 2
Hydraulic loading rate, m /m /day
(gpd/sq ft) 2
5 Organic loading rate, gCQD/ra /day
Influent COD concentration, mg/L
Influent NH.-N concentration, mg/L
Hydraulic loading rate, m-Vm^/day
(gpd/sq ft) 2
15 Organic loading rate, gCOD/m /day
Influent COD concentration, mg/L
Influent NH.-N concentration, mg/L
Hydraulic loading rate, rnVm-^/day
(gpd/sq ft)
20 Organic loading rate, gCOD/m /day
Influent COD concentration, mg/L
Influent NH.-N concentration, mg/L

Unit A
0,049
(1.2)
5.76
118.4
13.34
0.050
(1.2)
3.98
79.3
7.70
0.048
(1.2)
6.92
145.5
10.00

Unit B
0.048
(1.2)
4.13
85.6
9.69
0.052
(1.3)
7.50
144.5
14.76
0.048
(1.2)
9.73
202.3
13.00

Unit C
0.050
(1.2)
7.08
142.0
15.98
0.051
(1.3)
9.88
192.6
22.30
0.049
(1.2)
12.51
256.7
17.50

Unit D
0.051
(1.3)
8.90
173.3
20.27
0.053
(1.3)
13.92
265.2
29.79
0.050
(1.2)
13.97
281.9
22.30

-------
               TABLE III.  SUMMARY OF THE OBSERVED MEAN AND RANGE OF VALUES FOR THE LIQUID TEMPERATURE

                           (°C) IN THE VARIOUS STAGES OF THE FOUR EXPERIMENTAL RBC UNITS
ro
en
CTl
Phase
S tage Mean
First 16.3
Second 15.4
Third 14.7
Fourth 14. 4
Overall 15.2
TABLE IV. SUMMARY OF
AND FOURTH
Temperature, °C 5
Unit First
Stage
pH DO
A 7.97 7.6 8
B 8.03 8.8 8
C 8.00 7.9 8
D 8.03 7.4 8
I II
Range Mean
16.0-16.7 20.8
15. OrlS. 6 20.3
14.4-15.1 19.7
14.1-14.8 19.3
20.0
MEAN PH VALUES AND DISSOLVED
STAGES OF THE RBC UNITS
15
Fourth First
Stage Stage
pH DO pH DO
.22 9.4 7.80 5.0 8
.23 9.6 7.70 3.9 7
.25 8.8 7.70 3.6 7
.27 8.3 7.73 2.7 7
Range
20.5-21.2
20.0-20.7
19.1-20.1
18.6-19.8
III
Mean
5.9
5.1
4.5
4.1
4.9
OXYGEN CONCENTRATIONS (MG/L) IN

Fourth
Stage
pH DO
.00 7.8
.90 7.5
.73 7.9
.68 7.1
20
First
Stage
pH DO p
7.95 3.6 8.
7.98 2.9 8.
7.95 2.4 8.
7.80 1.9 7.
Range
5.4-6.1
4.7-5.3
4.2-4.8
3.8-4.5
FIRST

Fourth
Stage
H DO
13 6.9
08 6.5
05 6.5
98 5.8

-------
 nitrite)  were  measured  with a  Technicon Auto  Analyzer  II
 (22,23,24).  Other analytical  methods  employed in  this  study
 were  conducted according  to  Standard  Methods (25). Four  to
 five  samples were  collected  for  each  stage   effluent  during
 the  steady-state  period. The  influent  was generally sampled
 ten times during  a steady—state period.
     At  the  end  of  each phase,  the  total amount of biomass
 attached to  the discs  in each stage was measured by weighing
 the  discs  and  biomass  after drying  at  105°C  and weighing
 the  clean  dried  discs.  Several samples were  taken from  the
 dried  biomass  to determine  the VSS  fraction  as  outlined  in
 Standard Methods.

 PROCESS PERFORMANCE

 Attached Biomass

     In each phase of  the study  after a week  of  operation,  a
 thin  layer of  growth covered  the  discs  in the first stages.
 Generally  in  the  second week  some  biomass  sloughing  was
 observed in  the  first  stages, and  within a  few days  a  new
 biofilm was  built-up.  After  3 to  4 weeks of  operation,  the
 discs  in  the  first  stages  were covered with  a  thick, dark
 brown  or grey  biofilm,  and  further detectable changes in  its
 appearance were not observed.  The  structure of the biofilm in
 the  first  stages  seemed  to  be spongy,   rather than a smooth
 structure.  A filamentous growth in these stages may have been
 the reason for this type of structure.
     In the  successive  stages, the discs were covered with  a
 thinner biofilm  layer  and were  relatively smooth in appear-
 ance.  In  the experiments  at  temperatures of  15° C  and   20° C,
 the color of the  biomass was tan-brown.  In the experiment  at
 5°C, the biomass  in  the  second through  the fourth stages  had
 a black-brown appearance. The tan  color observed  at 15°C  and
 20"C  was  probably   due   to  growth  of  nitrifiers   in   these
 stages. Figures 2, 3 and 4 show the variation  in  the quantity
 of  attached  biomass  in  the  four  stages of the RBC units  at
 5°C, 15°C,  and 20"C, respectively.  In all  three phases,  there
was a successive decrease in the quantity  of biomass attached
 to  the  discs from the  first  to the  fourth  stages.  At   lower
 organic  loading  rates  and  higher  temperatures,  there  was  a
 sharp  decline  in  the quantity of  attached biomass following
 the  first  or second stages.  At  lower  organic loading  rates
 and higher temperatures, less  substrate and less  unstabilized
                               267

-------
 en
 1
 o
CD
   40
   35 -
   30 -
   25 -
   20 -
   15 -
   10 -
    o
      0
               -e- —
               -x—-
               -a—
  UNIT-A
-  UNIT-B
••  UNIT-C
-  UNIT-D
1          2
    Stage Number
            —i
             4
   Figure 2.  Attached biomass in the four  stages of
             the RBC units operating at 5°C.
                          268

-------
   45 i
   40 -
   35 -
   30 -
 •  25
 E
 o
CD
   20 -
   15 -
   I 0 -
    5 -
—-- UNIT-A

-— UNIT-B

— UNIT-C

    UNIT-D
                1          23

                    Stage Number
  Figure 3.  Attached biomass  in  the four stages of

            the  RBC units operating at 15°C.
                         269

-------
 o
OQ
   45 -
   40 -
   35 -
   30 i
   25 -
   20 -
 o
 a
•"  15 H
   10 -
    5 -
    0 -1
•*»-•  UNIT-A
•x-	  UNIT-B
  -  UNIT-C
  -  UNIT-D
                          2         3
                    Stage Number
  Figure 4.  Attached biomass in the four  stages of
            the RBC units operating at  20°C.
                         270

-------
sloughed  biomass  were available to establish attached  growth
in these  later stages.

Carbon Removal

     Figures  5,   6,  and  7  show the  mean steady-state mixed
liquor  filtered   COD  concentrations  when operating  the RBC
units at  5, 15 and 20°C,  respectively. An analysis of Figures
5, 6  and  7 show   that for the first stages  of the units, the
removal of filtered  COD  (influent filtered  COD minus stage
filtered  COD)  increased  when the  influent  filtered  COD was
increased. This  observation supports  the contention that the
removal of filtered COD  can be described by  a substrate  lim-
iting reaction. As the temperature was increased, the removal
of  the  filtered  COD  increased,  even beyond 15°C,  which  is
contrary  to the results reported by others (1).  In stages two
through four,  there  was  further removal of  filtered  COD  in
the higher loaded units,  but  the  removal rate  per stage was
much less  than in the first stage.  There was an  inconsistent
pattern of filtered COD  removal  in stages  two   through four,
probably  attributable to cell  lysing. In the last stages  of
the RBC units,  the differences  in  the  effluent  filtered COD
for  each   of  the  four  units  were much  smaller  than  those
observed  in the first stages.
     Figures 8,  9, and 10  show  the mixed liquor particulate
(total-filter)   COD,   when  operating  the RBC  units  at  5°C,
15°C, and  20°C,  respectively. As shown  in Figures  8,  9, and
10, particulate COD removal occurred as the wastewater  flowed
through the stages, but with  an irregular pattern of decline,
due to  the instability of  the attached biomass  of  the  last
stages.
     The removal of the influent particulate COD  in the first
stage,  although  the mixed liquor contained  sloughed biomass,
implies that  the influent  particulate COD  is available  sub-
strate,  as well  as the  soluble COD.  Table  V shows  the  sub-
strate removal efficiencies when considering total COD  as the
available  influent substrate and  the   filtered COD  as  the
remaining  substrate in the effluent from the RBC units.
     A linear relationship exists between the overall removal
of COD in  terms of grams  of COD removal per unit area and the
influent  substrate  loading  rate.  The  slopes  of  the relation-
ships increased  as the temperature increased:  0.811,  0.897,
0.976 for  5°C,  15°C, and  20°C, respectively  (all are signifi-
cant at  a level  of  0.01).  The  increase in  slope  shows  the
temperature dependency  of the substrate  removal performance.
                              271

-------
   150 -i
   125 -
_  100 -
 I
Q
O
O
 o
 c_
    75 -
    50 -
   — UNIT-A
   	 UNIT-B
   — UNIT-C
   — UNIT-D
                                               -o
                                               -Q
                                               -X
     0
       0
—[—
 2
T~
 3
                     Stage  Number
  Figure 5.  Mean steady-state mixed liquor filtered
            COD concentrations in  the four stages of
            the RBC units operating at 5°C.
                        272

-------
   150 n
   125 -
_  100 -
 I
Q
O
O

-o
 
-------
   150 i
   125 -
   1 00 -
— UNIT-A

-- UN1T-B

— UNIT-C

   UNIT-D
 e
 i
O
o
o
 o
 l_
 o
    25 H
                     Stage  Number
    Figure 7.  Mean steady-state mixed liquor  filtered

              COD concentrations in the four  stages of

              the RBC units operating at 20°C.
                         274

-------
   250 -i
   200 -
 1  1 50
o
o
o
 "100-1
    50 -
— UNIT-A
	 UNIT-B
— UNIT-C
— UNIT-D
               --.  *•'
                 ''X'"
                            —r~
                             2
                      Stage  Number
   Figure 8.   Mean  steady-stage mixed liquor particulate
              COD concentrations in the four stages of
              the RBC  units operating at 5°C.
                          275

-------
   250 i
   200 -
                           - UNIT-A

                           - UNIT-B

                           - UNIT-C

                           — UNIT-D
 CD
 E

O
O
O
 3
 o
 L.
 to
Q_
150 -
   100 i
    50 -
     0 -1
       0
                     S t a. g e  Number
  Figure 9.  Mean steady-state mixed  liquor particulate
            COD concentrations in the four stages of
            the RBC units operating  at 15°C.
                         276

-------
   250 i
   200 -
   150 -
o
o
   1 00
        — UNIT-A
        -	 UNIT-B
        — UNIT-C
        	 UNIT-D
o_
    50 -
                                    '"••x
                                               —!
                                                4
      2          3
Stage  Number
   Figure  10.  Mean steady-state mixed liquor particulate
              COD concentrations in the four stages of
              the RBC units  operating at 20°C.
                         277

-------
TABLE V.  SUMMARY OF  SUBSTRATE  (COD) REMOVAL EFFICIENCIES  (%)
Temperature,
°C

Unit
A
B
C
D

First
Stage
78.3
74.4
78.3
77.2
5

Overall
79.8
76.9
80.2
79.2

First
Stage
73.5
80.4
80.1
81.3
15

Overall
73.1
85.1
86.8
85.0

First
Stage
82.7
86.8
85.2
85.7
20

Overall
82,5
86.3
89.1
90.0
A  similar relationship  was obtained  with a  full-scale  RBC
plant  treating municipal wastewater  at  Kirksville,  Missouri
(26).  The substrate concentration was measured  as BOD^  ancj
the  slope  of  the  relationship  was  0.893  based upon  data
collected over a period of  two years.
     The  mixed liquor VSS  production  in terms of mg  per mg
COD removed was 0.50, 0.38, and 0.38 for 5°C, 15°C, and 20°C,
respectively.  The  increase in  sludge  production at  lower
temperatures  was  probably   due   to  lower  decay   rates.  The
increase  of  sludge  production  at   lower temperatures  was
observed also in other studies (1).

Ammonia Nitrogen Removal

     Figures 11,  12,  and  13 show  the  mean steady-state mixed
liquor ammonia nitrogen  concentration when operating the RBC
units  at  5°C,  15°C, and 20° C  (first  period).* At  5°C, there
was no ammonia removal in  the  system.  Analyses of Figures 12
and  13  show  that,  generally,  in  the  first stages  there  was
limited ammonia nitrogen removal,  except in Unit A, which was
receiving the  lowest  organic loading  rate."Significant ammo-
nia nitrogen removal  occurred  in  the  second  stages,  and pro-
ceeded  in the following  stages  in the  units  receiving  high
organic loading rates. The declining ammonia nitrogen removal
rates  were  observed in the stages containing  low concentra-
tions  of  ammonia  nitrogen  and  indicate  substrate  limiting
conditions.   In the region where  substrate was not limiting,
the decline  in ammonia nitrogen  removal  followed  a  straight
line,  and the lines  for  the different  units  were generally
parallel.  At a temperature of 20°C, the slopes  of these lines

* During  the  experiments  conducted   at  20°C,   significant
  changes in the influent ammonia concentrations  necessitated
  dividing this period for ammonia removal analysis.
                             278

-------
    5 -
    o 4
   40 !
   35 -
                            *•--- UNIT-A
                            -x-	 UNIT-B
   30 ^                     -B	UNIT-C

                            -*	 UNIT-D

-  25 H
01


T  20

o
e     i.	a	o--
   I 5 -



   1 o 4.      	*'
                                     -X--
                 1          2          3
                    Stage Number
Figure 11.  Mean steady-state mixed  liquor ammonia nitro-
            gen concentrations in the  four stages of the
            RBC units operating at 5°C.
                           279

-------
    40 -
                            ••-•-• UNIT-A
                            -x-	 UNIT-B
                            *— UNIT-C
                            	 UNIT-D
     0
Figure 12.  Mean steady-state ammonia nitrogen concen-
           trations in the four stages of the RBC units
           operating at 15°C.
                         280

-------
    40 i
                             *-•— UNIT-A
                             •x-	 UNIT-B

                             -a	UNIT-C

                             •*	 UNIT-D
       0
Figure 13,   Mean steady-state ammonia nitrogen concen-
            trations  in  the four stages of the RBC units
            operating at 20°C (first period).
                           281

-------
were greater  than  the  slopes of the lines at 15°C, emphasiz-
ing  the effect  of temperature  on ammonia  nitrogen  removal
rates  beyond  15°C.  Based  upon  the  above  observations,  it
appears  that  ammonia nitrogen  removal  could be described by
Michaelis—Menten enzyme kinetics.
     Table  VI  presents  a  summary  of  the  overall  ammonia
nitrogen  removal  efficiencies  at  15°C  and 20°C. Only  the
second  sampling period  data were  considered  from  the 20° C
experiments,  because  an  adequate  number  of  data were  not
available  during the  first  sampling  period. The  results in
Table VI show  that 98  to  99 percent ammonia nitrogen removal
was  obtained  at organic  loading  rates up  to  10 to  12.5  g
COD/m2/day. The removal efficiency decreased by approximately
10 percent  at  organic  loading rates of  14 g COD/m2/ciay.  The
percentage  removal of  ammonia  nitrogen  was higher  at 20° C
(Table VI).
     Nitrate and nitrite  nitrogen data revealed that  90 per-
cent of ammonia removal in the system occurred through nitri-
fication.  The  remaining portion  of  ammonia removal  probably
occurred because of stripping and assimilation.

  TABLE VI. SUMMARY OF ORGANIC AND AMMONIA NITROGEN LOADING
            RATES AND THE AMMONIA-N REMOVAL EFFICIENCY
Temperature,
°c
15
20
Unit
A
B
C
D
A
B
C
D
Organic Load
g COD/m2/d
3.984
7.496
9.875
13.916
6.915
9.734
12.513
13.971
Ammonia-N Load
g N/m2/d
0.387
0.766
1.143
1.563
0.362
0.520
0.663
0.786
Removal
Efficiency
Percent
94.8
97.6
98.1
86.9
99.0
98.0
99.4
90.5
KINETIC MODEL DETERMINATION

Carbon Removal

General

     The  first  stages  of  the RBC units performed differently
from the  other  stages  as  shown  earlier,  and the first stages
                              282

-------
were considered  separately  in the analyses.  The distinguish-
ing features of  the  first  stages  were that they were receiv-
ing raw  wastewater,  while  the influent to  the other stages
contained  sloughed  biomass  from  the  preceding stages  and
unconsumed  substrate.  This  difference in  substrate  affects
the processes taking place  in the  RBC stages.  The major pro-
cesses that can be related to the biomass in the first stages
are the  carbonaceous  substrate removal and. endogenous respi-
ration of  the  attached growth. In the following stages,  the
attached biomass  is  associated with  several processes,  i.e.,
stabilization of reattached biomass,  nitrification,  and exog-
enous  substrate consumption.
     The determinations  of  the kinetic  constants  were based
upon the mean  steady-state  values  of the parameters measured
for each  unit.   The  concentrations  of the  pollutants  in  the
effluent were independent of the fluctuations in the influent
concentrations;  therefore,  the mean  concentrations  from each
unit were utilized in the calculations.
     The following  assumptions were made  in the development
of the kinetic  model:
     1. The available  substrate  in the influent to  the first
stage  is the total COD.
     2. The  particulate  material  in  the  mixed   liquor  is
sloughed biomass.  Consequently the  available  exogenous sub-
strate in the mixed liquor is the filtered COD.
     3. The substrate  consumption reaction  takes  place only
in the attached growth.
     4. The  kinetics  of  substrate  removal  in the  second,
third,  and fourth stages can be expressed by a common model.

First Stage Substrate Removal Kinetics

     A mass balance of  the  biomass in the  first stage yields
the following equation under steady-state conditions:

     biomass produced - sloughed biomass - decay =0     (1)

     In  mathematical  terms  the  equation  can  be written  as
follows:
               Y Q(S0-Si-) - Q Xi - kdAiXi  = 0            (2)

     where

         Y       = yield coefficient, g VSS produced per
                   g COD consumed
                               283

-------
         Q       =  influent  flow  rate,  m3/day
         ^0»  Si  =  influent  and first  stage  effluent  sub-
                    strate  concentration,  mg/L-COD
         Xl      =  first stage effluent VSS,  mg/L
         Xl      *  first stage attached biomass g VS/m^
         AI      =  first stage discs area, m^
         kd      =  decay coefficient,  day ~1

A mass balance of the  substrate in  the first stage yields  the
following equation  under steady-state  conditions:

         substrate  consumed  - reaction  =0               (3)

In mathematical terms  Equation 3  can be written as follows:

         Q(So-Sl) - Air =0                              (4)

     where

         r       =  reaction  rate,  g COD/m2/d

     The reaction rate r in  Equation 4  can be expressed using
several kinetic  models.  The three  models  used  in this study
are summarized below:

1)   Monod growth kinetics,  incorporating the total  attached
     b iomas s .
                           K 4-S-
                            S  I                         (5)
                   k is defined as u/Y
     where

         k       = maximum reaction rate, day ~*
         p       ™ maximum specific growth rate, day~l
         Ks      =* half saturation constant, mg/L COD

2)   Monod  growth  kinetics,  incorporating  a  constant amount
     of active biomass.
                              284

-------
                          kaSl

                          V*T
     where

         ka      = maximum reaction rate,  g COD/m^/d

3)   Mass transport model (15,16,17)
                      r =
                           <+Sl
                                                         (7)
     where

         km      = maximum reaction rate,  g COD/m2/ mg/L-
                   COD/d

         KM      = constant,   mg/L COD

     The  reaction  rate expressions  (Equations  5, 6,  and  7)
were substituted into Equation 4, and the resulting equations
were rearranged  in the  following format,  to  carry out linear
regression analyses.
                                   f^i   ,1             do)
                                   k  S.   k
                                    ml    m
                              285

-------
      To carry out  linear  regression analyses for determining
 yield  and  decay   constants,  Equation  2  was  rearranged  as
 follows:
                  QXi
                  Aixi
      Table  VII  summarizes the  results  obtained  from  linear
regression  of  Equation  11,   and  Table VIII summarizes  the
results  obtained  from  linear  regression  of Equations  8,  9,
and  10.

 TABLE VII.   SUMMARY OF THE RESULTS OF THE LINEAR REGRESSION
              ANALYSES OF THE DATA USED TO  CALCULATE YIELD
              COEFFICIENTS AND DECAY RATES
Parameter
Yield coefficient, mg VS/mg COD
Decay rate, day~l
Regression coefficient
Significance level
5°C
0.66
0.07
0.998*
0.05
15°C
0.80
0.22
0.934
0.10
20°C
0.63
0.26
0.950
0.05
*Based  on  three  units;  B,  C,  D.
     The data  from Unit A  at  5°C  was  excluded from the analy-
sis  because the flow  was  changed during  the  experiment,  and
the  unit  did not approach steady-state  conditions.  Table  VII
emphasizes  that the  optimum growth and  yield  occurs at 15°C.
Muck  and  Grady  (27),  using  activated  sludge mixed culture,
observed an  optimum  in yield  coefficient at  20°C.  The differ-
ence in optimum temperature might be  because of  the  different
types of cultures growing  in  these systems.
     Consistent  results  were obtained with  Equation 8,  which
was  derived from Equation 5,  yielding  reasonable values  for
the  kinetic  constants  for all the  temperatures  (Table VIII).
The  mass   transport  model  (Equation 7)  produced reasonable
values only  with the data obtained at 5°C and 15°C. At these
temperatures,  the values  for  T/^ were  20.8 mg/L  and 42.5
rag/1,  which are  close to those  obtained by  Friedman  et  al
(16,17).  At  20°C the mass  transport  model resulted  in  a
                            286

-------
   TABLE VIII.  SUMMARY OF THE KINETIC CONSTANTS FOR CARBONACEOUS SUBSTRATE REMOVAL  IN
                THE FIRST STAGES3
Tempera ture?
Equation
Q(SO-SI) - AI
n/c o \ A
Q(so si} Ai
n/c • o \ A
Q(S0-S1) - Ax
no a o
S2 For 5 C and
°C

k¥i
Ks+Sl
Vi2
kaSl
K +S,
s 1
20°C only
5 15
k k
k K k
, m --S. . m
R a *Si R a
0.965 2.85 61.6 0.950 7.76
0.893 1.12 20.8 0.886 1.80

the data from units B, C, and D were

K
Ks
*M R
262.2 0.999
42.5 -0.808
-81.6 0.983
used, for 15 C
20
k
k K
9.44 276.4
0.98 -5.8
174 111.5
the data from
 units A, B, C, and D were used.
R = Correlation coefficient.

-------
negative  value  for Km> as  shown  in Table VIII. The negative
value may have occurred,  among other reasons, because  at high
temperatures  the kinetics are described by  substrate limiting
conditions  and  not diffusion.  Applying fixed biomass  quanti-
ties  with Monod growth kinetics as expressed  in Equation 6
resulted  in  negative  values  for  the  reaction rates  ka at
temperatures  of  5°C and 15°C. At 20°C  the reaction rate was
determined  to be  174 grams/m2/d  and the half saturation con-
stant was 111.5 mg/L.                         ^
     Clark,  et  al   (11),  reported values  of  U  ,  Kg  and Y of
4.4,  431  and 0,96, respectively. These values  were based on
soluble  BOD,  and  obtained  from  experiments   conducted  at
uncontrolled  temperature conditions. These values were calcu-
lated  from an  equation similar  to Equation 5,  except that
only  70 percent of the total attached  biomass was applied as
active  biomass.  Considering that assumption,  the   y  values
from  their  studies and this study  are  comparable.  The Y and
KS  values  differ  significantly  from the  values obtained in
this  study,  probably  because of  the differences in substrate
and  the fact  that Clark  et al  (11)  did  not  incorporate a
decay factor  in their equations.
     The  temperature  relationship  for k^  and k was obtained
by using Equations  12 and 13:
                       - (kd>2o
                  (k)T .
     where
  (k.)  ,  (k)      — decay rate and reaction rate at tempera-
    a L     T      ture T (°C), day -1
  ^d^ZO*  ^^20   = ^ecay rate a°d reaction rate at tempera-
                   ture 20°C, day"1

     Table  IX  summarizes  the  values  obtained  from  linear
regression analyses.
     The  temperature  factor  of 1.09  obtained  with the Equa-
tions 12  and  13 is similar  to  the  typical  value  of 1.08 for
the trickling filter process  (28).
     The  experimental  data,  as discussed  previously,  showed
that  the mass  of  attached   growth was  dependent  upon  the
organic  loading rate  and  could  be defined by a saturation
                               288

-------
 TABLE IX.  TEMPERATURE DEPENDENCY OF MAXIMUM REACTION  RATE
            AND DECAY RATE
Parameter
Correlation coefficient
Significance level
k20, day""1
•e-s
(kd)20» day'1
%
Equation 12
0.989
0.1

0.27
1.09
Equation 13
0.990
0.1
9.5
1.09


  function.  A  saturation type  relationship was  developed for
  the  first  stages that can be  used  for a given temperature to
  predict  the  quantity of attached bioaass.

                              k M.
                         I  =  K l                         (14)
                          1    K +M.
                               x  1
       where

          Xi       - the quantity of attached biomass in the
                     first stage per unit surface area, g VS/m^
          Mi       - organic load per first stage surface  area,
                     g COD /m2/ day
          kx       = constant, g VS/wr-
          Kx       = constant, g COD/m2/day

  A regression  analysis  of  Equation 14 in  its  linear (Eq. 15)
  form resulted in values as summarized in Table X.
                            kx Mi   \

    TABLE X.  SUMMARY OF THE  FIRST  STAGE  ATTACHED BIOMASS
_ CONSTANTS
Temperature,

  Constant              5.9             16.3              20.8
k
X
K
X
46.15

31.07

52.54

23.77

58.50

23.77

                                289

-------
     The  values of kx were  related to the temperature using
the  relationship shown  in Equation 16. The correlation coef-
ficient   obtained   from   a   linear  regression   analysis  was
0.986.

                  (kx)T =  56.9(1.015)T-20            (16)
                      (kx)T = g VS/m2

Carbon Removal Kinetics in Stages 2-4

     As  discussed  previously,  the  last  stages  of some units
revealed instability. To compensate for this instability, all
three  stages  were considered  as  one  reactor  where common
reactions were taking place. Equation 17 was used to  describe
substrate  removal  as a function  of temperature and  influent
substrate concentration to the second stages.



                    -S) =   I  A.(kL)20^T-2°S.n
                           .__  i  L 20  L     1
     where

         Q       = influent flow rate, m3/d
         £l      = first stage substrate concentration, mg/L
         S       = the mean substrate concentration in the
                   second through the fourth stages, mg/L
                 - total available surface area/stage, m^
              0  = reaction rate at 20° C, g COD/tn^/d
           ,      = temperature factor
         T       * temperature, "C
         n       = apparent reaction order

     Multiple  regression  analysis   with seven  steady-state
values  (where  substrate  removal  occurred)   resulted  in  a
regression coefficient of  0.986,  which  is  significant at the
0.01 level. The values obtained were:

         n       = reaction order = 0.763
        (kL)2Q   = reaction rate at 20°C - 0.0444 g COD/m2/d
         %,      = temperature factor =1.11
                                290

-------
     The  apparent  reaction order 0.763 obtained  for  the  sec-
ond  through  fourth stages  is  in agreement with  the  apparent
reaction  order  of  0.5-1.0  resulting  from  mass  transport
models  for  attached  growth (14).  The  temperature  factor  of
1.11  is  approximately the  same  as  the temperature  factor  in
the  first stages.

Ammonia Nitrogen Removal Kinetics

     A mass balance of ammonia nitrogen at  stage  i yields  the
following equation, at steady-state conditions.

            Q Cj..! -Q Ci =  Air                            (18)

     where

         Q       = flow rate, m-Vday
         C       = ammonia  nitrogen concentration, mg/L
         A       = surface  area  of discs, m
         r       = reaction rate, grams/m^/day

     The  reaction  rate,  r, can be expressed by the following
kinetic models:

     (a)     Monod growth  kinetics


                      -
     where

         kN      = maximum reaction  rate,  grams/m^/day
         Kjq      = half saturation constant, mg/L

     (b)     Caperon and Meyer kinetics  (29)


                          MC.-C  .  )
                           N  i  mm
                   .  r *  K  + (C -C . )
                           N     i   rain
                           291

-------
     where
                  .   =  minimum  ammonia  nitrogen  concentra-
                        tion  below  which  ammonia  nitrogen
                        removal  does   not  occur   (it  is
                        related  to  minimum  intercellular
                        stored  nutrient necessary  to sustain
                        growth) .

     To  carry out  linear repression analyses.  Equations 19
     and 20 were rearranged as  follows:
                 A.
                   1
                           -1
                                             'N
                                                         (21)
              QCC  ,.,-c.)
                 A.
-1
           (C.-C ,  )
             i  mm
(22)
     To  avoid  large errors  with  the  independent  variable
      in  Equations  21   and  22,  where  possible,  data  from
stages with nitrogen concentrations less than 1 mg/L were not
used in the linear regression analyses.
     Figure 14  shows the measured  concentrations  of ammonia
nitrogen   in  the   RBC   units   operating  at  15°C  and  the
regression  line   calculated   using  the  kinetic  constants
obtained  from the  linear regression analyses.  The lower part
of  the  prediction  curve  does  not pass  through  the  measured
data, indicating that there may be a threshold concentration
of  approximately  0.4  mg/L-N  below  which  ammonia  nitrogen
removal  does  not  occur.  Using  Equation  22  with  a  Cmin
of  0.4  mg/L,  better correlation was  obtained  as  shown  in
Figure 15.
     Figure 16 shows a plot of the data collected at  20°C and
the curve plotted using the kinetic constants obtained from a
linear  regression  of Equation 21. The  plot of  Equation  21
deviates  from measured  data points at  the higher concentra-
tions of  ammonia nitrogen.  Regression analyses of the  Monod
growth equation  in the   linearized  form  does  not  necessarily
                             292

-------
                         4.0 n
ro
10
oo
                          0.0
Me a sur e d


Z=2.439«C/(0.76+C)
                                                              R-0.945
                            0.0    2.0    4.0   6.0    8.0    10.0   12.0   14.0

                                           C-Ammon I a~N Concentration mg/l
             16.0
18.0
                        Figure  14.  Relationship between the ammonia nitrogen removal  rate and

                                  • the ammonia nitrogen concentration at 15°C.

-------
    "O
       4.0 i
       3.5 •
     0 3.0 -I
       2.5-
       2.0
       1 .5 -
     - 1.0 •
     c
     o
     6
     E
     -C
     I
0.5
       0.0
                               (C-0.4)/(0.45+ (C-0.4))
          0.0   2.0    4.0    6.0   8.0   !0.0   12.0   14.0
                         C~AmmonlB~N  Concentration *g/l
                                                      6.0   18.0
Figure 15.   Comparison of  the two predictive  equations  showing  the  relationship
            between ammonia  nitrogen  removal  rate  and the ammonia nitrogen
            concentration  at 15°C.

-------
ro
VD
en
                            0.0
Ma a s y r e d

ZM.624»C/(4.68
                                                                R-0.966
                              0.0    2.0    4.0   6.0    8.0   10-0   12.0   14.0

                                             C~Ammonla~N Concentration mg/I
             16.0   18.0
                    Figure 16.   Relationship  between ammonia nitrogen removal rate and the ammonia

                                nitrogen  concentration at 20°C,

-------
provide  the best fit for the Monod growth equation,  although
it  is the  best  fit of  the linearized- form.  The reason  for
deviation is that the  low and medium  concentrations have more
impact  than the high  concentrations  on the determination  of
the intercept and the  slope.
     An  attempt  was  made to improve the fit of the theoreti-
cal expression and  the measured data by choosing the pair  of
kinetic  constants  which yield  the  minimum sum  of  squares
(SSQ) between the predicted and observed values.
     Values  of  kN in  the  range 3.00  to  4.60  g N/m2/day  and
Kjj values  from  1.0  to 4.6  mg/LN  were evaluated.  The minimum
SSQ was  obtained using the values  of kN =  3.74 g/tn^/day  and
% = 2.8 mg/L.
     Figure 17 shows the curve plotted using Equation 19 with
the values  obtained  from linear regression  and with the val-
ues obtained from non-linear fit analysis.
     Table  XI  summarizes  the  kinetic  constants  for ammonia
nitrogen removal.

   Table XI. SUMMARY OF THE KINETIC CONSTANTS FOR AMMONIA
             NITROGEN  REMOVAL


Temperature
15
20


N
5
8


R
0.97
0.97
k
N
g/m2/d
2.334
3.74
V
N
mg/L
0.45
2.80
C .
mm
mg/L
0.40
0.00
N — Number of observations
R - Correlation coefficient

     The  results  obtained from  the  experiments conducted at
15"C show that a minimum concentration of 0.4 mg/L was neces-
sary to  maintain growth, while  at  20° C  this minimum concen-
tration  was  not  required. A  possible  explanation  is that at
higher temperatures  the mass  transport  resistance decreases
and,  as   a  result,   the  requirement  for stored material is
less.
     The kinetic constants obtained in this study for ammonia
nitrogen  removal are  comparable with values  obtained  with
synthetic substrate.  Saunders et al  (18)  reported K^ values
of 0.18  to  1  mg/L,  and  Ito  and  Matsuo (8)  reported kjg value
of 4 g/m2/day.
                                296

-------

TO
X
CM
E
X
Z
' ot
V
•*••
cc
ID
0
E
a:
z
i
n
c
o
E
E
-C
I
4.0 -i

3.5 -



3.0 -

2.5 -

2.0 -

1.5-


1 .0 -


0.5 -

                               Me a s ured
                               ZM.624»C/(4.68 +
                               Z«3.74»C/(2.8+C)
        0.0 -4
           0.0   2.0    4.0    6.0   8-0   10.0   12.0  14.0
                          C~Ammonla~N  Concentration mg/l
16.0   18.0
Figure 17.   Comparison of  the  two  predictive equations  showing  the  relationship
            between ammonia  nitrogen removal rate and the  ammonia nitrogen
            concentration  at 20°C.

-------
     The  effect of temperature on  the  maximum  reaction rate,
    can be  expressed  as  follows:
                                    T— 20
                                                          (23)
     where

          (^N)T (kN)2Q = maximum  reaction  rate  at  T  and
                        20°C,  g/m2/d
         •®f^           = temperature factor
         T            = temperature,  °G

     The  temperature factor,  -eN) derived  using Equation  23
is   1.1,   is  in  agreement   with  temperature   relationship
developed for nitrification  (30).
     The  inhibition  of  nitrification in  the first  stages was
related  to  organic loadings  and resulted in an equation with
a  correlation  coefficient -of  0.971  (significance  level   =
0.01):
          fj.      = 1.43 - 0.1M;  4.3
-------
     where
                 =  first stage ammonia nitrogen concentra-
     •               tion at simulated maximum nitrification

         CQ>C2jC3,C4 = ammonia nitrogen concentration in
                    influent, stage 2, 3, and 4,
                    respect ive ly

ENGINEERING SIGNIFICANCE

     The steady-state  kinetic  models developed in this study
for  the RBC  process  treating  domestic  wastewater  and  the
kinetic  constants  determined  as  a~ function  of  temperature
provide  a  rational design approach  for  the RBC process.  The
mathematical  expressions  presented  provide a basis, for  the
calculation  of  the  required RBC .surface  area to  meet pre-
scribed  effluent  standards  for  carbonaceous  subs'trate,  .and
ammonia  nitrogen  concentrations  at  temperatures ranging from
5°C to 20"C.
     Design curves  developed in this  study for carbonaceous
substrate  removal   in  a  four-stage  RBC  process at  20° C  are
presented in Figure 18. The corresponding temperature correc-
tion curves are presented in Figure  19.
     To  estimate  the  ammonia  nitrogen concentration  in  the
effluent,  design  curves  based on  the results  of  this  study
are . presented in  Figures  20 and 21  for  an  influent COD con-
centration of 300  mg/L and for  temperatures of  15  and 20° C,
respectively.  For  other  influent COD concentrations, similar
design curves can  be  developed using the equations presented
herein.
     When  the  RBC  system  is  designed  primarily   to  remove
carbonaceous  substrate,   a  different  configuration  of  RBC
staging can treat significantly higher loading rates than the
conventional  design,   without  bringing  the  first   stage  to
anoxic-anaerobic conditions. The configuration can incorpor-
ate four shafts in three stages,  with two of the shafts serv-
ing as  the  first  stage,  i.e.,  removal of  the  baffle between
the  first  and 'second  stages in the conventional  configura-
tion.  A design example is presented below:
     Assume that a  design  flow rate of 3800 m3/d  (]_  mgd)  of
domestic wastewater with a primary effluent COD concentration
of  300  mg/L COD  and  ammonia nitrogen of  30 mg/L,   is  to  be
treated  with  a  RBC  system  to a degree  that will  produce  a
final effluent of 45 mg/L COD,  or 85  percent removal.
                                299

-------
co
o
o
yo -I
90-
i
<: ss -
LU
n
S Rn -
o ou
£
o
a:
75-
7n

Infiu











ent COD (mg/U
300-400 -
200










100




























\
\
s










S
\











S
\










^
\











\
\










\
s]
X
x










X
x^



i







X
^










s^
X











\




























                         0.0     0.5     1.0     I .5    2-0     2.5     3.0     3.5

                                 Hydraulic Loading g pd/s qf i (• 04m3 /m2 /d)
               Figure 18.  Design  chart for COD removal in domestic wastewater treatment at 20°C,

-------
 o
-f
 o
 a
Ll_
 L.

 V

 CL
o • o
2.5 -
2n .
• u
1 -5 -
1 n
I > U
0.5 -
n n -

Treatment Efficiency
9O% X
X
85% \^













8O% ^






















"V
^^
^^













\

^\














\
N,

















X^






V







l^v.1









^_








\










K






\








































       0
   10       15       20       25

Temperature  °C
   Figure 19.   Temperature correction for COD removal,
                          301

-------
       100
   UJ
   OS
            Influent COO 3OO mg/1
         0-0      0.5     1-0      1.5     2-0     2.5
                  Q/A  gpd/sqf t (.04m3 /m2 /d)
Figure 20.  Design chart  for ammonia nitrogen removal at 15°C.
                              302

-------
   100
    90 -
LU
 o
 e
CC
    60
    50 -
    40
    30 -
    20
                                               \
      0-0     0-5     1-0      1.5     2.0     2-5

              Q/A  gpd/sqf t (.04m3 /m2 /d)
   Figure 21.  Design chart  for ammonia nitrogen removal
              at 20°C.
                            303

-------
     Design winter  temperature  is  5°C,  and summer temperature
is 15°C.
     1. Conventional  Design 4-stage RBC:  from Figure  18,  the
hydraulic  load  is found to  be  0.07 w?/nfl/d (1.75  gpd/sq  ft)
at 20°C.  At  5°C, the  temperature  factor is 2.7  (Figure  19).
To meet  the  required effluent quality  at 5°C,   the  designed
hydraulic  loading rate  will be 1.75/2.7;  i.e., 0.65  gpd/sq
ft.  The  required  total  effective  contactor   area  will  be
1.5  x  10°  sq  ft.  The  ammonia nitrogen  removal  efficiency
during the summer will be  about 98  percent (Figure  20),  i.e.,
the effluent will contain  about 0.6 mg/L NH^-N.
     2. Three-stage RBC,  first stage contactor  area,  40  per-
cent of the total RBC  surface  area.*
     Based on the equations  and  kinetic  constants presented
in this  study for  first  stage and  the  later stages  of  RBC,
the  total  surface area required will  be about  1 x 1Q6 sq  ft
to meet the effluent  requirements  at  5°C.  The  organic'loading
rate to the  first stage will be about 30  g/COD/m^/day, which
will assure aerobic conditions  at  summer temperatures.
     The  ammonia nitrogen  concentrations  in the effluent  at
summer  conditions  will   be  about  4.7  mg/L   in   this   RBC
configuration.

CONCLUSIONS

Carbonaceous Substrate Removal

     1. Carbon removal  in  RBC  units was  influenced  by temper-
ature  and organic  loading  rate.  The  overall   removal  effi-
ciencies  in  this study  were 80 percent,  85 percent,  and  90
percent for 5°C,  15°C, and 20°C, respectively.
     2. Majority  of  carbon  removal   occurred   in  the  first
stages. The COD removals in the  first  stages were 77 percent,
80   percent,   and  85  percent  for  5°C,   15°C,   and 20°C,
respectively.
     3. The -kinetics  for  carbon  removal in the  first  stages
can be described  by Monod  growth kinetics.
     4. The  temperature factor  for  the carbon  removal  reac-
tion rate and the decay  rate is  1.09.
     5. The  kinetics   for  carbon  removal  in the last  stages
can  be  described by  variable  order kinetics (in this  study,
0.763), and a temperature  factor of 1.11.
* The  current  common design  employs  shafts of 100,000  sq  ft
  in the  first stages,  and  150,000  sq  ft  in the last  stages.
                              304

-------
      6. The kinetic  constants determined  in  this  study  can be
used  to design RBC systems  (minimum DO of 2  mg/L  in the  first
stages)  for  carbon  removal in  a  temperature range of 5°C to
20° C.
      7. For  low  temperature  design,  providing  more  surface
area  in  the first stages  can  reduce  significantly the  total
RBC area required,

Ammonia Nitrogen Removal

      1, Ammonia nitrogen  removal  in RBC units was influenced
by temperature and  organic loading rate.  The overall ammonia
removal  ranged  from  87  percent  to  98  percent at  15°C, and
from  91  percent  to  99 percent  at 20°C.  At  5°C, there was no
ammonia  removal.  As  the  influent  organic   loading   rates
increased, the overall ammonia  removal decreased.
      2. The  kinetics  for  ammonia  nitrogen  removal  can  be
described by Monod growth kinetics. At 15°C, the model incor-
porated  a 'minimum  concentration  of  0.4  mg/L,  below  which
ammonia removal did not occur,
      3. The  temperature  factor for ammonia  removal  reaction
rate  was 1,10.
      4. The inhibition of ammonia removal in the first stages
was proportional to the organic loading rates.
      5, The resulted  kinetic  constants in this  study  can be
used  to  predict  ammonia   nitrogen  removal  in RBC  systems
within a temperature range of 5" to 20°C.
                                305

-------
REFERENCES

 1»  Antonie,  R.L. ,  1976,  Fixed Biological Surfaces - Waste-
     water Treatment, CRC Press, Inc., West Palm Beach, Flor-
     ida, 200  p.
 2.  Banerji,  S.K.,   1980,  ASCE  Water  Pollution  Management
     Task  Committee  ReportonRotating  Biological Contactor
     for  Secondary Treatment,  Proc.  1st National Symposium/
     Workshop  on  Rotating  Biological  Contactor  Technology,
     Champion, Pennsylvania,  1:31-52.
 3.  Smith, E.D.5  and J.T.  Bandy,  1980,  AHistory of the RBC
     Process, Proc. 1st National Symposium/Workshop on Rotat-
     ing Biological  Contactor Technology,  Champion,  Pennsyl-
     vania, 1:11-26.
 4.  Chesner,  W.H.  and  T.T.I.  lonnone,  1980,  Current Status
     o£ Municipal Wastewater  Treatment with RBC Technology in
     the  U.S.,  Proc.  1st   National  Symposium/Workshop  on
     Rotating   Biological   Contactor  Technology,   Champion,
     Pennsylvania,  1:53-70.
 5.  Torpey, W. ,  et  al,  1971, Rotating Discs with Biological
     Growth Prepare  Wastewater for  Disposal or  Reuse, JWPCF
     43 (11): 2181-2188.
 6.  Pescod, N.B., and  J.V.  Nair,  1972,  Biological Disc Fil-
     tration for Tropical Waste Treatment, Water Res.  (G.B.),
     6 (12): 1509-1523."~
 7.  Malhotra,  S.K.,  T.C.  Williams,  and W.L.  Morley,  1975,
     Performance of  a Bio—DiscPlant  in a Northern Michigan
     Community,  Presented  at the  48th  Annual Conf. ,  Water
     Poll.  Contr.  Fed. ,  Miami  Beach,  Florida,   Oct.  5-10,
     1975, 29 p.
 8.  Ito,  K. ,   and  T.  Matsuo,  1980,  The  Effect   of  Organic
     Loading  on  Nitrification  in  RBC  Wastewater  Treatment
     Processes',Proc.  1st  Nat ionalSymposium/Workshop  on
     Rotating   Biological   Contactor  Technology,   Champion,
     Pennsylvania, 2:1165-1175.
 9.  Zenz,, D.R.,  et  al,   1980, Pilot-Scale  Studies  on  the
     Nitrification ofPrimary and  Secondary  Effluents  Using
     Rotating  Biological  Discs at  the  Metropolitan  Sanitary
     District   of Greater   Chicago,  Proc.  1st  National
     Symposium/Work shop   on   Rotating  Biological   Contactor
     Technology, Champion, Pennsylvania,  2:1221-1246.
10.  Kornegay,  B.H.,  and  J.F.  Andrews,  1968,  Kinetics  of
     Fixed Film Biological Reactors, JWPCF 40 (11):R460-R468.
                                306

-------
11.  Clark,  J.A.,  'E.M. Moseng,- and T.  Asano,  1978, Perfor-
     mance of  a Rotating  Biological  Contactor Under Varying
     Wastewater Flow, JWPCF 50  (5): 896-911.
12.  Mikula,   W.J.,  1979,  Performance  Characteristics  and
     Kinetics  of  Substrate  Removal  in the  Treatment  of  a
     Cheese  Processing  Wastewater with a Rotating  Biological
     Contactor,  M.S.  Thesis,  Utah State  University, Logan,
     Utah, 195 p.
13,  Hansford,  G.S.,  et  al,  1978, A  S teady-State Mode1 for
     the Rotating Biological Disc  Re'actor, Water Res.(G. B.),
     12:855-868.
14.  Rittman,  B.E.,  and  P.L.  McCarty,  1978,   Variable-Order
     Model of  Bacterial  Film  Kinetics, J.  Env.  Eng.  Div. ,
     ASCE 104 (EE5):889-900.
15.  Schroeder,  E.D, ,  1977, Water  and Wastewater  Treatment,
     McGraw-Hill Book Co., Inc., New York,  pp.  288-312.
16.  Friedman,   A.A.,  R.C.  Woods,  and R.C.   Wilkey,  1976,
     Kinetic   Response  of  Rotating  Biological  Contactors,
     Proc. 31st  Ind..  Waste Conf.,  Purdue  Univ.,  Ann  Arbor
     Science Publishers,  Inc. ,  Ann Arbor,  Michigan,  pp. 420-
     423.                                                 .
17.  Friedman,   A.A.,  L.E.  Robbins,  and  R.C. Woods,  1979,
     Effect  of  Disc  Rotational  Speed onBiological Contactor
     Efficiency, JWPCF 51 (11):2678-2680.
18.  Saunders,  P.M., R.L.  Pope,  and M.A.  Cruz, 1980, Effects
     of  Organic  Loading  and  Mean  SolidsRetention  Time  on
     Nitrification   in   RBC  Systems,  Proc.   1st   National
     Symposium/Workshop   on   Rotating  Biological  Contactor
     Technology, Champion, Pennsylvania, 1:409-432.
19.  Watanabe,   Y. ,  M.  Tshiguro,   and K.  Nishidome,  1980,
     Nitrification Kinetics  in a  Rotating  Biological  Disc
     Reactor. Prog. Water Tech (G.B.), 12:233-251.
20.  Pano, A.,  1981,  The  Kinetics  of Rotating   Biological
   1  Contactors, Treating  Domestic  Wastewater, Ph.D.  Disser-
     tation,  Utah State University, Logan,  Utah.
21.  Oceanography International, 1978, Chemical Oxygen Demand
     (Standard Ampule  Method),  Federal Register Vol.  43, No.
     45,  Tuesday,  March  7,  1978,  Oceanography International
     Corporation, College Station, Texas.
22.  Technicon   Industrial  Systems,   1977,   Total  Kjeldahl
     Nitrogen,  Industrial  Method   No.  376-75W/B,  Technicon
     Industrial  Systems,  Division  of  Technicon  Instruments
     Corp., Terrytown,  New York.
                                307

-------
23.  Technicon  Industrial Systems,  1973a,  Nitrite  in Water
     and  Seawater,  Industrial Method.  No.  161—71W, Technicon
     Industrial  Systems,   Division of  Technicon  Instruments
     Corp., Terrytown, New York.
24.  Technicon Industrial  Systems, 1973b, Nitrate  and Nitrite
     in Water  and  Wastewater, Industrial Method No. 100-70W,
     Technicon  Industrial  Systems,  Division,   of  Technicon
     Instruments Corp., Terrytown, New York,
25.  American Public Health Association, 1976, Standard Meth-
     ods  for  the Examination  of  Water  and  Wastewater, 14th
     Ed.,  Washington, B.C.
26.  Dupont, R.R. ,  and R.E.  McKinney,  1980,  Data Evaluation
     of a  Municipal  Installation, Kirkville, Missouri, Proc.
     1st  National  Symposium/Workshop on  Rotating Biological
     Contactor Technology, Champion, Pennsylvania, 1:205-234.
27.  Muck,  R.E,,  and  C.P.L.  Grady, 1974, Temperature Effects
     on Microbial  Growth  in  CSTR's, J.  Env.  Eng.  Div., ASCE
     100 (EE5):1147-1163.
28.  Metcalf and  Eddy, Inc.,  1979, Wastewater  Engineering,
     Treatment  Disposal andReuse, 2nd Ed.,  McGraw-Hill Book
     Company,  New York, 920 p.
29.  Caperon,  J.  and  J. Meyer,  1972, Nitrogen-Limited Growth
     ofMarine Phytoplankton  I  and II"Deep-Sea Res.(G.B.),
     19:601-632.
30.  Parker, D.S.,   R.W.   Stone,   and  R.J.   Stenquist,  1975,
     ProcessDesign  Manualfor  Nitrogen  Control,  USEPA,
     Technology Transfer,  October, 1975.
                              308

-------
      KINETICS AND SIMULATION OF NITRIFICATION IN A ROTATING
      BIOLOGICAL CONTACTOR
      Yoshitnasa Watanabe. Department of Civil Engineering,
      Miyazaki University, Miyazaki 880, Japan

      Kiyoshi Nishidome. Department of Civil Engineering,
      Kagoshima Technical College, Hayato 899-51, Japan

      Chalermraj Thanantaseth. Department of Chemical
      Engineering, King Mongkut Institute of Technology,
      Thonbuli Campus, Bangkok, Thailand

      Masayoshi Ishlguro. Department of Civil Engineering,
      Miyazaki University, Miyazaki 880, Japan
INTRODUCTION
     A steady-state kinetics for fixed-biofilm reaction has
been developed and applied to the denitrification and the
nitrification processes in rotating biological contactors (1,2,
3),  The proposed kinetics can be described as a process of
molecular diffusion with a simultaneous zero-order biochemical
reaction. The proposed kinetics has adequately explained all
experimental results of nitrification in a partially submerged
rotating biological contactor (RBC),  but it cannot be strictly
applicable to a partially submerged RBC process in which the
biofilm alternately rotates into water and air. The partially
submerged RBC has no steady-state substrate concentration pro-
                                 309

-------
 file wichin  the  biofilm,  even  though  the  concentration of  the
 substrate  in  che  bulk water  is  the steady  state.
     In  this  paper,  the authors  report  the results of a  com-
 puter simulation  of  nitrification in  a  partially  submerged RBC
 process  to find out  the reasoning behind  the  application of the
 steady—state  kinetics. An analysis of the  experimental data on
 combined carbon oxidation-nitrification in the  same process is
 also included. All fluxes in this paper are expressed on the
 basis of the  submerged disk surface area.

 APPLICATION OF STEADY-STATE KINETICS TO A  PARTIALLY SUBMERGED
 RBC PROCESS

 Modification  of Steady-State Kinetics
     The proposed kinetics can be applied  to  nitrification in
 a fully submerged biofilm process, summarized below. At  steady
 state, the transfer  rate of ammonia to the biofilm surface
 through the diffusion layer is equal to that  at the biofilm
 surface. Thus, the ammonia flux  to the biofilm  surface can be
 expressed by  Eq. 1,  if the amount of ammonia  used for cell syn-
 thesis of the nitrifying bacteria is negligibly small compared
 to that nitrified by the same bacteria.


     TT (CbA - CsA>=FA                                    (1)
       d

Therefore, the relationship between bulk and  surface ammonia
concentrations is,

                              F
                        C .  + rF-                          (2)
                         sA   K,,
                               dA

Ammonia flux at the biofilm surface (FA) is represented  by Eq.
 3 for partial ammonia penetration and jby  Eq.  4  for complete
 ammonia penetration.

      F = 1/2D.R C .        C . < C *                        (3)
       A  V  A n sA        sA =  sA
      FA= F.    = ,/2D.R C* = RL    C   > C *             (4)
       A   A,max  V  A n sA    n n    sA =  sA

     However, the proposed steady-state kinetics would not be
completely applied to a partially submerged RBC process for
the following reasons. A steady-state substrate concentration
profile within the biofilm cannot be assumed, even though the
bulk substrate concentration is the steady state, since the
biofilm alternately rotates into the air and
                                310

-------
          c

          n

          o
          H-
          Hi
          H.
          O
          w
          rt
          H-
          O
          a
                                                    Nitrifying
                                                    Bacteria  Layer
CR
                                     Rotating Disk
W
O
MI
Cu
to
   cr


o  o
H- O
rt  £3
i-t  cr
M- H-
t-rt P
H- (B
O  CL

rt  O

O  f<
         O
         X
         H.
         CL
         (a
         rt
         H-
         O
                                    Layer
                                                 Attached
                                                 Water-Layer
                                                      Heterotrophlc

                                                      Bacteria  Layer
                                                    Nitrifying
                                                    Bacteria  Layer
                                              Anaerobic

                                              Layer
                                      locating Disk
                          311

-------
 E
 OC
  0.4

  0.3
 <.
 u.
  0.2
  0.1
                                          '-BA—%—*—e—
                                                           o   o
                             CiA=2Smg/l     O CiA=200mg/l

                             CiA=50tng/l     Q Calculated Value

                             CiA=100mg/l
      L.l-1.
                J	IN 1   I  I  I I I
                          5       10
                    Bulk ^iponia Cone., C.
                                                   50
                                                          100
                                      .
 Fig. 2  Relationship  between bulk ammonia concentration
         and ammonia flux
         (Disk diameter=30cm, Disk rotating velocity=7.5rpm
          Water temp.=23.5°C)
 ~  0.4-

M|  
-------
the water. The authors (7) have developed a hypohesis about
oxygen transfer which would be applicable to a partially sub-
merged RBC. The hypothesis states that the oxygen transfer to
the biofilm mainly occurs through the attached water-layer,
during the time the biofilm rotates in the air. The nitrifica-
tion biofilm model for a partially submerged RBC is shown in
Fig. 1 (a). The penetration depth of oxygen (Ln) can be ex-
pressed as follows I
           D0CC0* - C  )
      Fo- - I—55-                           •        (5)
                  w
           ¥      F
           _
       n   R    4.33R
            o        n

Oxygen  consumption  for biological nitrification is 4.33
g NH^-N (4). Bintanja et al (5) proposed the following equa-
tion for the estimation of the thickness of the attached wa-
ter-layer on the disk surface!

                                                           (7)

Employing Eq. 2 to 7, we can find the relationship between the
bulk ammonia concentration and the ammonia flux.

Experimental Verification of the Modified Kinetics
     The data obtained in a partially submerged RBC have pre-
viously been presented (2). Figs. 2 and 3 are examples of some
of the data. The theoretical data calculated from Eq. 2 to 7
have also been plotted in Figs. 2 and 3. The experimental unit
had a disk diameter of 30 cm, L^ equalled 42 jjira at a disk ro-
tating velocity of 7.5 rpm and water temperature of 23 °C (Eq.
7) . These corresponded to the experimental conditions for the
data shown in Fig. 2. Hartman (6) measured the attached water-
layer thickness at about 40 ym in an actual RBC plant. As
shown in the next section, the oxygen concentration at the
biofilm surface (CSQ) was estimated at about 2 mg/1. Therefore,
the oxygen flux to the biofilm was calculated at 1.36 g 02/m^h
from Eq. 5. Eq. 6 gave 60 ym as the value of Ln. Fig. 2 shows
the FA>max was 0.27 g NH^-N/m2!!. Eq. 4 then gave 52 ym as the
Ln. The intrinsic nitrification rate (Rn) was determined as
5200 g/m2h in the previous experiment (7), The penetration
thickness of oxygen calculated from Eqs. 4 and 6 were almost
the same. Therefore, the proposed hypothesis on oxygen  transfer
was confirmed.

COMPUTER SIMULATION OF NITRIFICATION IN A PARTIALLY SUBMERGED
                              313

-------
RBC PROCESS

Model Development
     The biofilm attached to a partially submerged RBC rotates
alternately into the air and water. In the air phase, oxygen
is supplied to the biofilm from the air, but there is no ammo-
nia transport to the biofilm. In the water phase, ammonia dif-
fuses into the biofilm from the bulk water.A computer simula-
tion to identify the change in the ammonia and oxygen profiles
in the system was carried out based on the assumptions which
had been made for the development of our steady-state biofilm
kinetics, namely!
1. The bulk water is completely mixed,
2. Only molecular diffusion occurs through the diffusion layer,
3. Molecular diffusion with a simultaneous zero-order biochemi-
   cal reaction occurs within the active biofilm.
Fig. 4 illustrates the biofilm system divisions consisting of
the attached water-layer, the diffusion layer, and the biofilm.
The disk surface was divided into n small sectors each with an
area equal to AA. The biofilm,the attached water-layer, and the
diffusion layer were divided into sub-layers, each of them AZ
thick.
     The basic equation of the simulation was Pick's Second Law
of Diffusion

      8 CA      8 2C


Eq. 8 was directly applied to the attached water-layer and the
diffusion layer,  but a biochemical reaction term had to be
added to take into account the substrate uptake within the bio-
film as follows!

      8C.      92C.
        A _ .,     A
      dt     Aaz2
The difference form of Eq. 9 is

      CA(n+l,i)= K(CA(n,i-l)~2CA(n,i)+CA(n,i+l))
               + C.,   .v-R At                              (10)
                  A(n,x)  n

                                                           (11)

where the subscript n refers to the number of At time and the
                            314

-------
          Lw
Cone.
     Fig. 4  Divisions of the biofilm  system
     Table  1  Simulation conditions
Parameter
Biofilra Thickness (L
-------
 subscript  i refers to the concentration reference plane. The
 ammonia  flux to the elemental biofilm at any time can be ob-
 tained as  follows!

                                                           (12)
      "Asn  AZ  wA(n,l) ~A(n,2)'
The average flux for all elements in the water phase at any
time is shown in Eq. 13

           . n=n          D.  n=n
      F. = i  EAt. F.  = -Ar- I  (C.,  ,v-CA,  „,)       (13)
       At,     A,n  n4 Z   ,   A(n,l)  A(n,2)'       v  '

Results and Discussion
     The thickness of the diffusion layer, the intrinsic ni-
trification rate, and the relationship between the bulk con-
centrations of ammonia and dissolved oxygen were obtained in
the previous experiment (2). The thickness of the attached wa-
ter-layer was changed to match the simulated results with the
experimental results. The thickness of SQ^im gave the best fit
in both cases (Table 1). Fig. 5 shows the changes in the simu-
lated concentration of ammonia and oxygen in the elemental
biofilm with varying detention times for air and water phases.
As shown in Fig. 5, the profile changes depended on the deten-
tion time in each phase, even for the steady-state conditions
of bulk ammonia and dissolved oxygen concentrations, The dot-
ted line represents the steady-state ammonia concentration
profile predicted by the modified kinetics. Fig. 6 shows the
average ammonia flux as a function of detention time in the
bulk water. Fig. 7 shows the comparison between the simulated
average flux and the flux obtained in the experiment. The
simulation results based on the conditions shown in Table 1
compared favorably with the experimental data. Fig. 8 shows
the effect of the attached water-layer thickness on the aver-
age flux. The thinner thickness gave a higher flux because of
the high oxygen flux. However, the attached water-layer thick-
ness was naturally determined as formulated in Eq. 7, Fig. 9
shows the effects of the dissolved oxygen concentration on the
average flux. DO concentration was also naturally set at a
level depending on the operational condition. Therefore, Figs,
8 and 9 show how to change the average flux, if the attached
water—layer thickness and DO concentration are artificially
controlled. Fig. 10 shows the average flux change with the
disk rotating velocity at a fixed disk peripheral velocity of
7 m/min.  With the disk at a fixed peripheral velocity, the
average flux increased with the increase of the disk rotation-
                             316

-------
Fig.  5   Ammonia and DO profiles in air  and
         water phases
    1.2
                      C>3.5.mg/l
                      CbA=3.0 mg/1 (Av.FA=0.299)

                      CbA=2.0 mg/1 (Av.FA=«.2l2)
         Detention Time in Water Phase  t (sec)
Fig.  6   Relationship between detention time
         in water  and ammonia flux
                     317

-------
   0.4
 .c
•si
 E
 U>
   0.3 —
   0.1 _
                    Experiment (Run 1)
  	 Experiment  (Run 2)

   O  Simulation  (Run 1)

   •  Simulation  (Run 2)
                         I
                            _L
                   I
J_
                                           -o-
2      4       6
  Bulk Ammonia Cone.
                                    8       10
                                    (wg/1)
 Fig, 7  Comparison of simulated flux with
          experimental flux
     0.5 r-
        Q       20       40       60      80
           Attached Water-Layer Thick.,  Lw(un)

   Fig. 8   Effect of  attached water-layer
            thickness  on ammonia flux
                318

-------
 E

 oo
 X
 D
t— (
U,

 tfl
TH
 c
 O
0.42


 .40


  38


 .36


 .34
              T=28.5 C

              Disk Rotatim
                Velocity=7..
                 2         4
               Bulk  DO Cone,
Fig.  9  Effect  of DO concentration on
         ammonia flux
 JZ
>J
 E
   0.40
   0.38
   0.36
 3 0.34
 c
 1 0.
 Fig.
  32
           Disk peripheral
            velocity = 7 m/rain
                               10
           Disk Rotating Velocity
                                (min"1)
                                            15
   10   Effect of disk rotating velocity
        on ammonia  flux
               319

-------
al velocity under oxygen limiting, but the increment of the
flux was very small compared with that of the disk rotational
velocity.
     The simulation study had clearly shown that the nitrifi-
cation rate predicted by the modified kinetics would be equal
to the average nitrification flux of each elemental bio^ilm.
This fact provided the reasoning behind the application of the
modified steady-state kinetics to the nitrification process in
a partially submerged EEC. We concluded that the amount of
ammonia nitrified within a biofilm rotating in the air phase
can diffuse into a biofilm rotating in the water phase.

COMBINED CARBON OXIDATION-NITRIFICATION IN A PARTIALLY
SUBMERGED RBC
Experimental Procedure
     Two units with disk diameter of 30 cm were used for
the experiment. Unit 1 consisted of 13 polywood disks mounted 2
cm apart on a horizontal shaft and a trough with a volume of 15
liters. Unit 2 consisted of 15 polywood disks, 2 cm apart on a
horizontal shaft and a trough with a volume of 18 liters. The
direction of the flow in both reactors was perpendicular to the
rotating shaft. The residence time distribution of the water in
the two reactors without biofilm perfectly coincided with that
of a single completely mixed—flow reactor. The experimental
variables were (a) hydraulic loading, (b) influent BOD5 concen-
tration, and (c) the type of organic matter (glucose and tapio-
ca starch). Water temperature and pH were in the range of 23 °C
to 27 °C and 7.8 to 8.2, respectively. The disk rotating veloc-
ity and the influent ammonia concentration were fixed at 8.5
rpm and 45 mg/1, respectively. Experimental conditions are sum-
marized in Table 2. In each Run, samples were collected two or
three times a day until the system reached a steady state,
Results and Discussion
(a) Effect of Organic Oxidation on Nitrification
     The reduction of ammonia in combined carbon oxidation-ni-
trification process consists of a reduction due to nitrifica-
tion and one due to cell synthesis, since the specific growth
rate of heterotrophic bacteria is normally much higher than
that of autotrophic nitrifying bacteria. Therefore, the ammonia
utilized for the cell synthesis of heterotrophic bacteria can-
not be neglected. The authors (7) have already evaluated the
ammount of ammonia which would be utilized due to cell synthe-
sis of heterotrophic bacteria at about 10 % of the 8005 reduc-
tion. Therefore, the total ammonia flux to the biofilm can be
expressed by Eq. 14.
                             320

-------
      Table  2   Experimental  conditions
Type of organic
carbonaceous substrate







Glucose












Starch





Run
number
1-1
1-2
1-3
1-4
1-5
2-1
2-2
2-3
2-4
2-5
3-1
3-2
3-3
3-4
3-5
4-1
4-2
4-3
4-4
4-5
4-6
5-1
5-2
5-3
5-4
5-5
B005/N
0.6
0,6
0.6
0.6
0.6
1.2
1.2
1.2
1.2
1.2
2.8
2.8
2.8
2.8
2.8
2.7
2.7
2.7
2.7
2.7
2.7
0.7
1.4
2.7
4.0
5.8
Hydraulic .
loading 1/m h
5.1
7.6
10.2
14.2
15.3
5.1
7.6
10.2
15.3
21.2
3.9
5.1
7.6
10.2
14.2
3.4
5.1
6.8
9.1
12.7
15.3
5.1
5.1
5.1
5.1
5.1
Unit
2
1
2
2
1
2
1
2
1
1
1
2
1
2
2
2
2
1
2
1
1
2
2
2
1
1
SJater
Temp . C


27




25




25




25





23


                                     0> B005/N-2.68 (Starch)
                                                   2.
   0.3
            A  BOD5/N=0.56 (Glucose)
            O  BODs/N-1.20 (Glucose)    »  Q/Aw=S,l 1/i/h
            *.  BOD5/N=2.80 (Glucose)   	 NitrtEtcation at 8005=0
               .	  ^N.max
e
CO
c
o
" 0.1 -
                                      25 °C
                                     -o~ _ .	^__
                                                   -O-.
                              _l_
Fig.  11
                      10       15      20       25
                   Bulk Ammonia Concentration,  C|,A  (mg/1)
                                                       30
          Relationship between bulk ammonia concentration
          and nitrification flux
                         321

-------
      FA= FN + °'1FB

     The relationship between the nitrification flux and the
bulk ammonia concentration was  obtained as shown in Fig. 11
by using measured values of F^ and Fg, and Eq. 14. Fig. 11
clearly shows the effects of organic carbon oxidation on ni-
trification. Carbon oxidation was also influenced by nitrifi-
cation as shown in Fig. 12. The data represented by triangles
were calculated by the kinetics for starch oxidation without
nitrification. The data represented by circles were obtained
in the experiment. In Fig. 12, it can be seen that the reduc-
tion of starch flux due to simultaneous nitrification was neg-
ligibly small until the bulk starch concentration reached
about 40 mg/1 (the corresponding BODs was about 20 mg/1). When
the bulk starch concentration increased beyond about 40 mg/1,
the reduction of starch flux became remarkable, because the
inner part of the aerobic biofilm mainly consisted of nitrie
fying bacteria. The results shown in Pigs, 11 and 12 would come
from the following hypothesis!
1) Most of the heterotrophic bacteria exist in the outer part
of the aerobic biofilm while most of the nitrifying bacteria
grow in the inner part of the aerobic biofilm as shown sche-
matically in Fig. 1 (b) . However, this would only be true when
the specific growth rate of heterotrophic bacteria is much
higher than that of nitrifying bacteria.
2) Both heterotrophic and nitrifying bacteria are aerobic and
the amount of oxygen supplied to the biofilm would be almost
the same under fixed operating conditions, independent of the
aerobic bacteria composition under oxygen limiting.
     Based on the above hypotheses, the following relation-
ships applicable to a combined carbon oxidation-nitrification
process in a partially submerged RBC have been developed.

      F0= 4.33 v     =4.33 F.T + Fn                       (15)
       0        N,max        N    B
Oxygen flux (Fo) is shown in Eq. 5. In addition, BOD5 flux can
be used to express the oxygen flux caused by the heterotrophic
bacteria, if organic matter produces a straight increase of
BOD against incubation time,  i.e., the oxygen uptake rate is
assumed to be constant independent of the residual organic
concentration. This was almost true in the case of glucose and
starch used in our experiment. Their BOD^ per unit mass were
0.71 g 02/g glucose and 0.55 g 02/g starch. Both values were
experimentally obtained (7).  In dimensionless  form,  Eq.  15
becomes,
                              322

-------
    10 c:
 A
N

 6


 00
 gO.5
 f.
 o
• Fs vs. C




O Fs vs. C
                          O .>
                             (Experiment)




                             (Experiment)
                                                  Fs,raax''2.1 g/m h
                                      or Csg (g/m )
     Fig.  12  Logarithm  plots of  starch  flux
     i.o
                    Di«ensionless BOI>5 Flux,  F /F
                                          N.max
      Fig. 13   Plots of  BODs  flux and nitrification  flux

                (straight line shows Eq.16)
                          323

-------
                      1     FB   - -, _ A 23   FB
                    4 33  F      ~~     u.4.j p
                           N,max             N,max

 Introducing  Eq.  14  into Eq. 16 gives the relationship between
 the ammonia  flux and  the  BODg flux. In dimensionless form, it
 is expressed by  Eq. 17.
         F                 F
      ,  A    = 1 -  0.13 -_!—                            (17)
       N.max             N,max

The maximum 8005  flux  (Fg max)  *s obtained when FN  is  equal  to
zero,  i.e., F^ is equal  to 0.1  FB*

       F_    =4.33 FM                                      (18)
       Btmax        N,max
Figs. 13 and 14 show the  experimental verification of Eqs.  16
and 17, respectively.  In Runs 1 to 4 where the hydraulic load-
ing increased at a fixed influent WD^ to NH4-N ratio, the fil-
amentous bacteria grew on the biofilm surface with the increase
of hydraulic loading.  Organic oxidation by the filamentous bac-
teria was not considered in the biofilm kinetics, therefore,
the obtained 8005 and ammonia flux were higher than the pre-
dicted flux.  As a result, most of the data in Runs 1 to 4 were
slightly higher than the predictions.

(b) Comparison of Predicted Values with Existing Data
     Fig. 15 shows the relationship between  the bulk BODs  and
the corresponding BOD^ flux obtained in Run  5. Circles in  Fig.
15 show BOD5 flux without simultaneous nitrification as calcu-
lated  by our kinetic model, explained below. Under  operating
conditions in  Run 5,

       Fn    =  4.33 FM    = 1.33 g/m2h  ,
       B,max        N,max        6
The molecular  diffusion  coefficient of starch was estimated  by
Fig. 16, because  our kinetics states that the overall  rate
constant (K*)  is  equal to the mass transfer  coefficient  (K
-------
The diffusion  layer thickness was estimated at  75 pm on a  disk
of 30  cm diameter, rotating at  7.5 rpm,  in water at a. tempera-
                    a  Run I

                    A  Run 2

                    O  Run 3
      Q  Run 4

      *  Run 5

      •  Tertiary Treaciaenc
                    1.0         2,0          3.0
                         Dimensionless BOftFlux, F_/F.,
                                      -*      B N>rcax

       Fig.  14  Plots of  BOD5 flux and ammonia flux
                 (straight line shows  Eq.17)
                                                                  570
           1.2F
           1.0
         ^ °'8
         •i
         e
         "oo 0.6
         x 0.4
         3
         o 0.2
F0    -4.33F.,    =1.
 B.max     N.max
                                  • Experiment  (F HO)
                I   I   I   i  i   I   I  I   I   i   i  i   i   i   i  I
             0    20    40    60    80    100    120   140   160
                             Bulk BODj Cone., CjjB  
-------
            o.lrr
          °-001l              5     10             50    100
                    Bulk Substrate Cone. C,- Starch as mg/1 COD
                                          Ammonia as mg/1 N

      Fig.  16  Overall  rate constant vs.  bulk substrate
                concentration
100 r
                        40

                       (mg/1)
                                       Disk RPM  Temp. C

                                     O  4,3-5     12-19

                                     O  3,2       14-20

                                     A  2.0       17-19

                                     ®  5.0       4-10-J

                                     »  8.5       23
                                                     (EPA Experiment)



                                                     (Calculation)
0          20
  Effluent BODjConc.

    Fig. 17   Comparison of  calculated value
              with USEPA data
                          326

-------
ture of 23.5 °C. The thickness would be inversely proportional
Co the root of the disk peripheral velocity  (3), therefore, in
Run 5, it was estimated at about 70 ym. The molecular diffu-
sion coefficient of starch was calculated as follows!
      D = K* L    ( 4 x 10-2 m/h)(7Q x 1Q-6 m) = 2.8 x 10~6 m2/h
       s   so

Ds equals 2.4xlO~6 m2/h, using Wilk and Chang's Equation (9).
We employed 2.5xlO~" m^/h as the molecular diffusion coeffi-
cient of starch in the calculation. Then, the  intristic starch
oxidation rate was determined as 3x10^ g/m^h,  using the half-
order plots shown in Fig. 12.
      The ammonia flux at 23°C for any bulk BODt- can be predi-
cted by using the curve in Fig.15 and Fig.lU. The caluculated
relationship between the effluent BOD,, and the percent ammonia
removal for a multi-stage completely mixed-flow RBC with the
same total disk surface area is shown in Fig.17 along with US
EPA data (1-0).  The caluculation was made for an influent ammonia
concentration of 30 mg/1 and an influent BOD- concentration of
150 mg/1.  USEPA data were collected in a two-stage RBC in which
the direction of flow was parallel to the rotating shaft.  The
average influent Kj eldanl nitrogen and BODc concentrations were
28.9 ffig/1  and 3.1*7 mg/1,respectively.
SUMMARY AND CONCLUSIONS

    A modified biofilm kinetics for a partially submerged RBC
process was 'developed. The proposed kinetics was applied to
the nitrification process with and without simultaneous carbon
oxidation. A computer simulation of nitrification based on the
assumptions for the model development produced an average am-
monia flux of the elemental biofilm rotating in the water
phase that was almost equal to the flux calculated by the
modified kinetics. This provided the reasoning behind the "ap-
plication of the modified steady-state kinetics to a partially
submerged RBC in which the biofilm alternately rotates into
the water and air.
     In a partially submerged RBC, the oxygen  transfer to the
biofilm mainiy occurs when the biofilm rotates in the air
phase, while the biofilm rotating in the water phase adsorbs
the substrates. The oxygen transfer rate through the attached
water-layer or the oxygen flux to the biofilm  rotating in the
air can be shown by the following equationl

          D (C* - C .)
      _    oo    sCr
      V	L	
               w
                             327

-------
The maximum nitrification flux without simultaneous carbon ox-
idation is represented by the following equationl
       N.max  4.33 R

The authors considered that the amount of oxygen transported
to the biofilm or consumed within the biofilm would be almost
the same under fixed operating conditions, independent of the
aerobic bacteria composition under oxygen limiting. Based on
the above hypothesis, we proposed the following equations for
nitrification flux and for ammonia flux in combined carbon
oxidation-nitrification in a partially submerged RBC.
Nitrification flux (FN) I
             = 1 - 0.23;
       N,raax

Ammonia flux
        F,
                N.max
       N,max
               1 - 0-.13:
               "N.max
The relationship between the effluent BOD5 concentration and
the percent ammonia removal calculated by the above equation
and the experimental data almost coincided with that obtained
in USEPA experiments.
Acknowledgment
     The authors would like to express their appreciation and
thanks to Japan International Cooperation Agency (JICA) for
financial assistance.
NOMENCLATURE
Symbol
  JbA
  "sA
  -. *
  -sA
  CbS

  CsS

  CiA
Dimension
 g/m3

 g/m3
 g/m3

 g/m3


 g/m3

 g/m3

 g/m3
Description
Ammonia concentration within biofilm

Bulk ammonia concentration

Ammonia concentration at biofilm surface

Critical ammonia concentration at biofilm
surface

Bulk starch concentration

Starch concentration at biofilm surface

Influent ammonia concentration
                            328

-------
  cj     g/m       Saturation concentration of oxygen
            *3
  CSQ    g/m       Oxygen concentration at biofilm surface
  DA     tn^/h      Molecular diffusion coefficient of ammonia
          •y
  D0     m /h      Molecular diffusion coefficient of oxygen
          o
  Ds     mz/h      Molecular diffusion coefficient of starch
  F£     g/m h     Ammonia flux
  FB     g/m2h     BOD5 flux
  FJJ     g/m^h     Nitrification flux or ammonia flux due to
                   nitrification
  Fo     g/m h     Oxygen flux
  K
-------
   Treatment Plant',1 Water  Research, Vol.9,  1975,  pp.1147-1153
 6. Hartman H.  "Untersuchung uber die  Biologische  Reinigung  von
   Abwasser nit Hife von Tauchtropfkrorpern,  Kommissionsverlag
   R. Oldenbourg Munchen,  1960
 7. Watanabe Y. and Thanantaseth C. "A Study on  Purification
   Mechanism of Rotating Biological Contactor (IE)','  submitted
   to Journal  of Japan  Sewage Works Association
 8. Ishiguro M. and Watanabe Y. "A Study on  Tertiary  Treatment
   of Municipal Sewage  by  Rotating Biological Contactor  (IT) ,
   Journal of  Japan Sewage Works Association, Vol.14,  No.152.
   1977, pp.32-41
 9. Welty J.R.  et.al. "Fundamentals of Momentum, Heat and Mass
   Transfer, John Wiley and Sons Inc., 1969,  pp.463-465
10. USEPA "Application of Rotating Disc Process  to Municipal
   Wastewater  Treatment',' Water Pollution  Control  Research Set*
   ries, 17050 DAM 11/71
                          330

-------
             PART  IV:   CONCEPTS AND MODELS
    SELECTION AND OPTIMIZATION PROTOCOLS FOR ATTACHED

           GROWTH BIOLOGICAL PACKED COLUMNS
    Sheldon F. Roe, Jr., P.E., Manager, Technical Market
    Research, The Hunters Corporation, Fort Myers, Florida

    Edward B. Hanf, Vice President, Director Sales and
    Marketing, The Munters Corporation, Fort Myers, Florida
1.0  INTRODUCTION

     Attached growth has long been used for water treatment
in packed columns.  Now the idea of producing fuels, foods,
or chemicals by similar techniques (anaerobic digesters for
methane or columns for ethanol) promises an exciting future
in this field.
     Our reference point is cooling towers, S0? scrubbers,
chemical absorption-desorption and tube settlers, in addition
to various attached growth mechanisms.  Many of these columns
share common problems-, but they also share common advantages
for optimization and for adaptation to new processes and
systems.
     Too often we hear the comment "I tried your fixed film
system and it didn't make any difference."  Most likely the
system was not operating at capacity and, indeed, the at-
tached growth didn't make any difference.  It is the purpose
of this paper to provide guidelines for process selection and
optimization.
     Topics of discussion include:  trade-offs, column char-
acteristics, operating characteristics, solids handling,
process selection, operating analysis, and recommendations.
The objective is to design the process to fit the biocolumn
rather than to adapt the biocolumn to the process.
                              331

-------
2.0 DISCUSSION

2.1  Efficiency, Through-put, Pressure Drop Trade-off
    One common characteristic of columns is the trade-off
between efficiency, through-put and pressure drop as
illustrated in Figure 1.  Traditionally, high efficiency,
high through-put, and low pressure drop are the ideals we
are looking for, while, on the other hand, low efficiency,
low through-put and high pressure drop are what we most
seek to avoid.
    Between these extremes there exist twelve combinations
of high and low efficiency, through-put and pressure drop
which we must consider.  Let us start with the most for-
giving combinations.
    Traditionally, high efficiency and low pressure drop
(sometimes called HELPD Packings) are the ideal combination,
allowing for some trade-off as far as through-put is con-
cerned.  The combination of high efficiency and high
through—put, on the other hand, allows for flexibility
regarding pressure drop.  While the combination of high
through-put and low pressure drop permits a tolerable
trade-off on the efficiency.
    At the other extreme, low through-put and high pressure
drop represents the least forgiving combination and can only
be compensated for by a high efficiency.  Similarly, low
efficiency and low through-put don't offer much opportunity
for compensation.  Listed below are combinations which do,
on occasion, offer viable trade-offs.  In each case, the
third component of the trade-off or some other un—named
factor must compensate for the undesirable aspect as listed.
    1.  High efficiency and high pressure drop
    2.  Low efficiency and high pressure drop
    3.  Low efficiency and low pressure drop
    4.  Low efficiency and high through-put
    5.  High efficiency and low through-put
    6.  High through-put and high pressure drop
    7.  Low through-put and low pressure drop
    For the purpose of this discussion, pressure drop and
pumping head are used interchangeably since both represent
resistance to flow and operating costs.
    The above is a somewhat over simplified view of complex
real-life situations.  But let us throw in another factor
here which comes with attached growth.  This is fouling.  We
will use the same approach as above, but now with the extra
factor added,  (see Figure 2)
                             332

-------
            Figurt-  1
Figure 2.  Track-off Tt-rrLtory
              333

-------
2.2  Definition of Column Characteristics
    An arbitrary list of 8 parameters which we think are
important for biocolumns is given in Figure 3 (a matrix of
these parameters with themselves).  The shaded areas define
the subject of principle interest in this paper.  As the
light areas indicate we will talk very little about time,
biochemistry, energy, or economic analysis.  The areas are
subdivided as shown in Figure 4.       This results in a
far more complex matrix which demonstrates not only combin-
ations of parameters but combinations within parameters,
And it will, in fact, be desirable to discuss some of these
internal combinations (for example 2.3 with 2,1 - the
combination of an entrained bed with a fixed bed).  The
purpose of the matrix is to ensure that all possible alt-
ernatives are evaluated in die selection of a given process.
    After considering the above, an intensive look at
the operating characteristics of an individual column is
in order.  The morphology (shape) of the column and packing
material are important.   These are considered in the next
two sections.
2.3  Operating Characteristics of an Individual Column—
     Column Morphology
    The interaction of flow and velocity is illustrated in
Figure 5.  For a given through-put, a counterflow column
can have either of two extreme shapes.  In Case No. 1 the
column is long and slender, operating at high velocity.
The pressure drop of the fixed bed per A£ of the column
length must be low or the pressure drop of the total column
will be prohibitively expensive.  The efficiency per length
of column may be low, but the column can be extended in
length to compensate for this low efficiency per unit of
length.
    The shape of the packing material in a fixed bed in
such a column of course influences the length required for
a suitable efficiency.  Here, large edge effects may be
expected because the substance flowing through the column
would rather go to the walls than through the center of the
column.  Redistribution may be necessary to counteract this.
Opportunities for channeling in this column of Case No. 1
are minimal when compared to the column of Case No. 2.  Flow
conditions can be important; i.e., under conditions of a low
Reynolds Number separation can take place.  But with a high
Reynolds Number the column can actually be a mixer.
                            334

-------
CO
to
en
                                            1.0  2.0  3.0  4.0  5.0  6.0  7.0  8.0
                       1.0  Biochemistry


                       2.0  Kind  of  Bed
                       3.0
                            Packing
     Characteristics


4.0  Flow



5.0  Phase-State



6.0  Time



7,0  Energy


     Economic
                       S.O
                            Analysis
                                       Figure  3.   Biocolumn  Matrix

-------
         2.0   KIND OF BED
              2.1   FIXED
              2.2   FLUIDI7.ED
              2.3   ENTRAINED
              2.4   MOVING-MECHANICAL
              2.5   DEGRADABLE

         3.0   PACKING CHARACTERISTICS
              3.1   ORDERED SHAPES
                   3.1.1  TUBES
                   3.1.2  SHEET
                   3.1.3  OTHER-DIRECTION
                              ORIENTATION
              3.2   RANDOM SHAPES
                   3.2.1  SPHERICAL
                   3.2.2  FIBERS
                   3.2.3  OTHER-BLOCKS
              3.3   SURFACE AREA
                   3.3.1  MACRO
                   3.3.2  MICRO
                   3.3.3  MOLECULAR
              3.4   STEADY STATE-CHANGES
                   3.4.1  DEGRADABLE

         4.0   FLOW
              4.1   FLOW DIRECTION
                   4.1.1  CROSSFLOW
                   4.1.2  COUNTERFLOW
                   4.1.3  COCURRENTFLOW
              4.2   FLOW CHARACTERISTICS
                   4.2.1  VELOCITY
                   4.2.2  GRAVITY
              4.3   INTERMEDIATE FEEDS,
                   RECYCLE,  STAGING

         5.0   PHASE-STATE
              5.1
SOLIDS
5.1.1
5.1.2
5.1.3
5.1.4
5.1.5
GAS
5.2.1
5.2.2
LIQUID
5.3.1
5.3.2
SAND
COAL
BIOSOLIDS
FIXED FILM
PLUGGING
BUBBLES -FOAMING
CONTINUOUS PHASE
LIQUID FILM
LIQUID DROPS
              5.2


              5.3


              5.4  REACTANT VS.  CARRIER

         6.0  TIME
              6.1  SOLIDS RETENTION TIME
              6.2  LIQUID RETENTION TIME
              6.3  GAS RETENTION TIME

         7.0  ENERGY
              7.1  MATERIAL-ENERGY BALANCE
              7.2  HEAT
              7.3  KINETIC
              7.4  CHEMICAL REACTION
Figure 4.   Matrix Detail
                        336

-------
    CounterI 1ow
Case 1
   High velocity
   Low AP/A£
   Largo edge offsets
   Lower
                                         Cross!' 1 ow
                                             t
                   Cast- 3
                                             I
                       Low velocity
                       Higher AP/Jl
                       Flow distribution
                          problems
                       Higher effidency/A£
   Case 2 Counter flow
   Low velocity
   Higher AP/£
   Flow distribution
      problems
   Higher efficiency/A£
                       Case 4 Crossflow

                            T   fStaging
                                      •[
                                  ]•
                     High velocity
                     Low AP/A2.
                     Large edge effects
                     Lower e f f iciency /j\£ '
     Figure
(Iross Morphology of Blocolumns
                          337

-------
     In case No. 2 we have a column of large flow area and
low velocity.  Rather than edge effect, here the problem
is one of obtaining equal flow distribution over the 'large
area.  Since this column is operating at a lower Reynolds
Number, it may be a better separator than mixer.  A plug
or  piston-type flow may be more difficult to obtain in a
large column and may present problems in scaling up into
larger diameters.  For example:  how does one get plug flow
in a fixed bed 100 feet in diameter and 6 feet high?
     The crossflow versions of this problem are illustrated
in Cases 3 and 4 of Figure 5.  In Case 4, staging offers a
means of successively making counterflow in a horizontal
direction.
2.4  Handling Solids—Packing Morphology
     Handling solids in a fixed bed with a continuous liquid
phase presents an important problem in controlling fouling.
An analysis of the surface area yields the conclusion that
the shape of the surface is as important as the amount of
area.  This is illustrated in Figure 6 where the distri-
bution of area with the column axis is presented.  The
right-hand side of the horizontal axis in Figure 6 repre-
sents area perpendicular to the flow.  This area must be
kept to a minimum to prevent solids build up.  Because this
area is generally small for all cases described (in some
cases only 10 or 20% of the column cross section), differ-
ences in solids build-up can be experienced which are by
orders of magnitude.
     The left-hand of the horizontal axis represents sur-
faces parallel to flow.  From the biofilm standpoint, these
surfaces would more closely approach laminar flow while the
surfaces toward the right-hand of the axis represent pro-
gressively higher shear rates.  However, this is not abso-
lutely true because the distribution along the column axis
must also be considered.
     It is also interesting to define the parameters in
Figure 6 in terms of the phase, with the following possi-
bilities of contacting the shape of the surface:
     1.  Continuous gas phase contacting a liquid film.
     2.  Continuous gas phase contacting liquid drops to a
         liquid film.
     3.  Continuous gas phase contacting solid'particles on
         a liquid film.
                            338

-------
                 000
                 X X X
                Sheet fill No. 1
                Sheet fill No. 2
                Dump or random fill
                60° tube settler
                Splash fill or trays
                                          x
                                          x
                                          X
                                             X  X
                                          X
                                          x
                                          X
                                          X
   XOXOXOXOXOXOXO &OXOXOXOXOXOXOXOXQX 0  O '
                                             .To
0°
I'5°
       30°     45°    60°     75°
Angle with column axis, degrees
 Structural
                                Fouling^
                  High Spreading
                         High Pressure Drop
     Separation
  Figure 6.  Area Orientation  of  Fill  in  a Tower—
             Packing Morphology
                           339

-------
    4.  Continuous liquid phase contacting gas bubbles.
    5.  Continuous liquid phase contacting solid particles.
    6.  Continuous liquid phase contacting a biofiltn
        combined with gas bubbles or solid particles as
        in 4 or 5 above.
    After considering the morphology of columns (Figure 5)
and the morphology of packings (Figure 6), the protocol
proceeds to a selection decision tree and an intensive
analysis of column operating parameters in the next two
sections.
2.5  Selecting a Process
    Each of the expanded categories in Figure 4 have been
arranged by coded number in a decision analysis tree.
This, in combination with a matrix as in Figure 3, yields
one means for selecting a process.  The interdependence of
alternates is not accounted for in this system.  The
selection decision in Figure 7 is a fixed bed 2.1
(Figure 4), sheet packing (3.1.2), crossflow (4.1.1), and
fixed film (5.1.4).  Admittedly, this is arbitrary and the
selection sequence can be changed to emphasize important
criteria first.  Figure 4 also straddles the fence between
pure logic and existing columns such as:  fluidized beds
with sand or coal, woven fabric, glass fiber discs, glass
beads, Raschig rings, sheet packings, bricks, corn stalks,
or chunks of foam.
2.6  Analyzing Operation of a Column
    Another step to be considered is given in Figure 8.'
This is a more intensive analysis of the given column.
Indeed, this is the subject of the usual operational study
of a column.  Each column has its own operating character-
istics or range of values for acceptable operation depending
on the value analysis in the trade-off of efficiency,
through-put, and pressure drop.
    In addition to the parameters named in Figure 8,
several other common sense limitations should be consid-
ered including:  dissolved gas limitation, film transfer
limitation, substrate limitation, and fixed film solids vs.
suspended solids.  Processes such as gas bubble release,
settling, coagulation, or liquid drop mechanics must be
analyzed on a niicrostructure basis.
                            340

-------
yyyyy    yy    y
                                                    6 o
                                                      "•
                                                    5 
-------
                      Select crowsflow VH. count IT flow
                           Surface contact media
    Too small—
    FouI ing
    Low through—put
    High pressure drop
Optimize media flute slzv
  For attached growth
     952 void area
    30 to 1202ft/ft3
Too small—
Unter evaporates and
Solids remain in media—
High pressure drop	
     Too large—
     Reduced efficiency—
     Large vessels—deeper bed
     Low pressure drop—
     Optimize liquid
        Flow rate
Too small
Large liquid handling
Facilities
Too large—
Flooding or tow solids retention
Large water treatment facl lities
.High pressure drop	
  Optimize solids-liquid
      Concent rat ion
   Too small—
   Reduced efficiency
   Larger tower
    Too large—
    Solids build-up
 Optimize gas flow rate
Too large —
High pressure
drop —
                      Select material of construction
                        Stainless,  fibers or  plastic
                                Select staging
                                  or  recycle
             Figure 8.  Optimisation of  Surface  Contact Media
                                     342

-------
3.0  RECOMMENDATIONS

     Selection and optimization of biological columns is
complex but may contribute as much to the success of an
operation as the biochemical process itself.  Fouling,
channeling, and low efficiency are to be avoided.  However,
with proper selection through optimization, the contri-
bution of a column can be as important as the biochemical
process in the success of the many new processes on the
horizon.
     Several alternatives to conventional processes are
suggested below:
     1.  The combination of entrained or fluidized beds
         with fixed beds where particles in the contin-
         uous liquid phase contribute surface area and
         the fixed bed contributes operating stability
         of attached growth, flow modification, or a
         separation characteristic.
     2.  Trickle beds with gases other than air - e.g.
         nitrogen or carbon dioxide.
     3.  Crossflow trickle beds.
     4.  Crossflow columns with fixed beds and liquid
         continuous phase for release of carbon dioxide
         along the length of the column.
     5.  Staging in both crossflow and counterflow where
         higher surface area fixed bed is utilized at the
         latter stages.
     6.  Microstructures  such as glass or cellulose fibers
         utilized in fixed bed for immobilization while
         maintaining adequate passages to prevent fouling.
     7.  Moving mechanical beds which operate at non-
         steady conditions where the fixed film is
         subjected to alternating conditions as catalysts
         and regeneration in conventional chemical
         processing.
     8.  The stability (resistance to operating upsets)
         of attached growth beds should be utilized more
         fully:  a. to treat toxic chemicals, b. for pure
         cultures (not biological soups) to produce
         chemicals.  This stability, known for years in
         water treatment, might be compared to immobil-
         ization in biotechnology.  For monocultures
         sterilization could be a solvable problem.
         Sterilization could be considered the antonym
         of operating stability.
                           343

-------
            MODELING OF BIOLOGICAL  FIXED  FILMS  —
                  A STATE-OF-THE-ART  REVIEW
C. P« Leslie Grady, Jr. Department  of  Environmental Systems
Engineering, Clemson Univeristy, Clemson,  South  Carolina.
INTRODOCTION
     Although fixed-film biological  processes  found early
application to wastewater  treatment  their  use  declined with
the development and wide—scale  adoption  of the activated
sludge process.  There were many  reasons for  this,  ranging
from trivial to well-founded, but  the  net  result  was that for
many years fixed-film processes  (most  notably  the trickling
filter) were relegated primarily  to  the  treatment of low flow
domestic wastewater or to  the pretreatment of  industrial
wastewater.  During the early 1960's however,  with the advent
of plastic media for trickling  filters,  there  was a resurgance
of interest in fixed—film  reactors.  This  interest was
stimulated in the late 60's and early  70 *s by  the development
and commercialization of rotating  disk reactors which provided
many of the benefits of trickling  filters  without some of the
disadvantages.  Finally, in the late 1970"s and early 1980's
fluidized bed biological reactors  moved  from  the  laboratory to
the field, thereby opening up a whole  new  area for application
of biological fixed films.  Because  of these  advances in
                          344

-------
process development and  because  fixed-film reactors are
generally less energy  intensive  than  activated sludge,  engi-
neers are now employing  fixed-film biological processes in a
host of new applications with a  great deal of success.
     Concurrently with the  new developments in fixed-film
reactors has come a renewed  interest  in  their modeling.  There
are at least two reasons for this.  One  is that models  are the
basic tools of engineering which facilitate the design
process.  The other is that  models  help  us achieve a better
understanding of something  by guiding our  analysis of it.
Modeling and experimentation are interdependent, with each
providing input to and taking information  from the other.
Consequently, as we have learned more about fixed-film
processes we have been able  to develop better models which
have helped us to see  new applications and to develop better
methods for design.
     Mathematical models may be  divided  into two categories:
empirical and mechanistic.   Empirical models simply relate
operating input and output  variables  to  each other and  make
little pretense of representing  individual phenomena.  Such
"black box" descriptions are quite  useful  for design from
pilot plant data and have found  wide  use in environmental
engineering.  Many of  the models for  biological fixed-film
processes fall into this category.  Mechanistic models, on the
other hand, express the  influences  and interrelationships of
individual mechanistic phenomena in a manner which allows  the
investigator to discover how the system  might respond under
unexplored conditions.   Thus one might argue that the primary
purpose of a mechanistic model is to  further understanding.
This additional understanding will  be of direct benefit to the
practitioner, however, because it is  the nature of practice to
apply knowledge to areas in  which no  prior experience exists.
Mechanistic models have  broader  utility  than empirical  ones.
Consequently this review will be limited to models of that
type.
     Mechanistic models  of  biochemical processes generally are
developed by application of  reactor engineering principles,
i.e., they combine expressions representing the intrinsic
kinetic and transport  events with mass balance equations de-
scribing the characteristics of  the particular physical system
under consideration.   Consequently, simulation with such
models give insight into the basic  events  occurring within a
process as well as the influence that the  system configuration
has upon the outcome of  those events. When we examine  fixed-
film biological processes we see that in spite of the
                          345

-------
I
                             I  INPUT
                         Waste Characteristics
                         Media Characteristics
                         Flow Rate  etc.
                             [ OUTPUT]
                         Btomass Concentration
                           Biofilm Thickness
                        Degree of Bed Expansion
      BIOFILM MODEL
]
I
REACTOR FLOW MODEL
J
     Effectiveness Factor
 Reactor Reaction Rate
              Boundary Condition at the Biofilm- Liquid Interface
                       Substrate Conversion Rate
                             [ OUTPUT|
                     Effluent Substrate Concentration
Figure  1.   Flow diagram of  model for  a fluidized bed
             biological  reactor illustrating interfacing of
             biofilm model with model of the physical
             characteristics  of the reactor.  (From Shieh  and
             Mulcahy (I).).
                              346

-------
diversity in physical  configuration,  all contain biofilms.
Consequently, we might model  these  processes by developing a
mechanistic model for  the  biofilm and then interfacing it with
appropriate models for each of  the  physical process
configurations.  Figure  1  adapted from Shieh and Mulcahy (1)
illustrates the application of  this approach to a fluidized
bed biological reactor.  Similar  flow diagrams could be
developed for other fixed—film  reactors but they would differ
in the way in which the  biofilm submodel is interfaced with
the other system submodels.   It therefore follows that a
prerequisite to successful modeling of fixed-film biological
processes is a realistic model  for  the biofilm.  Do we have
one?  How have researchers sought to  develop one?  Is there
concensus in the approach  that  is being taken?  The purpose of
this paper is to address questions  such as those.
THE BIOFILM

     The first step  in  the  development of a mechanistic model
for a system is its  reduction  to  its  essential components.
Figure 2 is a schematic of  the essential components required
to model a biofilm.  As  shown  there,  organisms in a biofilm
with density or concentration  X^  grow attached to a solid
support.  In a trickling filter that  support is either rock or
plastic, in a rotating  disk reactor  it is plastic, and in a
•n
Of
S

o
a
a
3
CO
T>
I
\
\
L. ' ' Biof Mm ".•'''
'C».Density= x'f." '.'
- ' ' '•' '
_ - ,-
3. -o
tj a <
c ~ ~
a>3. o
c  -*>
— i- Q.
S o o
(Q C O
•« O o
% "V «!
2 = c
2il
3 O O
~ 3, 2
_j a o
                                      DISTANCE, x
Figure 2.  Schematic diagram  of  biofilm.
                         347

-------
fluidized bed biological  reactor it is sand, coal, or some
other granular material.   In  the first two reactor types the
radius of curvature  of  the support is large with respect to
the biofilin thickness so  that  the support may be considered to
be flat.  In the  third  this may  not be true, so spherical
particles are generally assumed,  although it has been shown
that the solutions for  particles are similar to those for
slabs when a suitable characteristic dimension is chosen (2).
Growth continues  until  some thickness Lf is attained, with
the method of control of  that  thickness  being a function of
the type of process  being considered.  Adjacent to and
permeating the biofilm  is  a liquid layer whose total thickness
depends upon the  type of  process, and in some cases, upon the
time within the process.   Growth of the  organisms is dependent
upon the transport of an  electron donor, an electron acceptor
and nutrients through the  liquid layer and into the film.
Generally, nutrients are  provided in excess, so the electron
acceptor and donor are  the only  constituents considered.
Since it is possible for  transport of either the donor or
acceptor to limit the rate of  growth of  the organisms in the
biofilm, .knowledge of their concentrations in the bulk liquid
(C  or C7), at the interface  (C   or C ), and in the biofilm (C
or C ), is quite  important.   The relationship between C  and
C  is influenced  by  the nature of the process and represents
a place where the biofilm model  must be  interfaced with the
process model.  Furthermore,  the change of C^ or CA with
depth in the biofilm is influenced by the relative
concentrations of the electron donor and acceptor at the
interface, the thickness  and  physical properties of the
biofilm, and the  kinetics  and  stoichiometry of the biochemical
reactions.  Ths first two  groups  of characteristics are
process dependent, and  thus represent additional connections
with the process  model.   The  last group  is an intrinsic
characteristic of the transformations being carried out in the
reactor.
     From consideration of the essential characteristics
illustrated in Figure 2 it can be deduced that development of
a biofilm model requires  knowledge in the following areas:
(a) transport of  materials in  the liquid phase; (b)
characteristics of the  biofilm,  including its thickness,
density, and composition;  (c)  transport  and reaction within
the biofilm; and  (d) techniques  for solving the resultant
equations.  In the following  sections each of these areas will
be reviewed in depth.
                           348

-------
TRANSPORT OF MATERIALS  IN  THE  LIQUID PHASE
     There is abundant  evidence  that the rate of transport of
materials from  the  bulk liquid to the biofilm:liquid interface
can be an important  determinant  of  the performance of a fixed-
film process.   The  references  cited below to demonstrate this
phenomenon should be viewed  as representative of the broader
body of literature  rather  than as all-inclusive.  This same
caveat should also  be applied  to the remainder of this article
because inclusion of all literature dealing with the mechanis-
tic modeling of  fixed-films  was  beyond the scope of this
endeavor.
     The clearest evidence for external mass transport and the
necessity for its inclusion  comes from microprobe measure-
ments of the dissolved oxygen  profile up to and through a bio-
film.  Bungay and his coworkers  have been the primary utili-
zers of this technique and Figure 3 from their most recent
work (3) clearly demonstrates  that  the oxygen concentration at
the biofilm:liquid  interface can be appreciably less than that
   o>
   E
      8
O  6
O
z
UJ
o
   O
   UJ
   (0
   (0
               Outside Film
 Inside Film
                     I
       200  150  100   50
50   100  150  200
                          DISTANCE (pm)

Figure  3.   Oxygen concentration profile up to and through  the
            biofilm on a rock from a trickling filter.  Data
            from Chen and Bungay (3).
                           349

-------
in the bulk  liquid.   This  particular study was conducted with
slime-covered rock media  removed from, a trickling filter and
placed into  an experimental  apparatus which allowed the
maintenance  of liquid velocities similar to those found in
field installations.   The  key  point  to note is that reaction
rates calculated by using  the  bulk  liquid concentration in the
intrinsic rate equation would  be in  error because the
concentration at the  biofilm:liquid  interface was lower than
the bulk concentration.   Similar findings have been reported
by others (4).
     Indirect evidence for the importance of external mass
transport has been obtained  by observing changes in the reac-
tion rate when the fluid  velocity past a biofilm is changed.
The effects  of only external mass transport may be isolated by
using an extremely thin biofilm so  that mass transfer effects
in its interior are minimized.  This was done by LaMotta (5)
using a rotating annular  reactor. He found that the reaction
rate increased until  the  velocity past the biofilm was around
0.8 m.s   but that thereafter  it was constant.  This
suggests that at higher velocities  the possible transport rate
of reactants to the biofilm  exceeded the maximum reaction
rate whereas at lower velocities transport limited the reac-
tion rate.  With thicker  films the  transport of reactants into
the biofilm makes calculation  of the rate constant for exter-
nal transport more difficult,  but from a qualitative point of
view it is still possible  to show that the rate will increase
with increasing velocity  until some  limiting point is reached.
This has been done by Trulear  and Characklis (6) in a reactor
similar to that employed  by  LaMotta  (5) and by Castaldi and
Malina (7) in a rotating  tube.  Although Trulear and
Characklis (6) found  that  the  reaction rate became independent
of velocity  at a value of  0.93 m.s   , a value remarkably
close to that of LaMotta  (5),  it should not be concluded that
velocities in this range  will  always make external mass trans-
fer effects unimportant.   Rather, it will depend upon both the
concentrations of reactants  in the  bulk liquid and the poten-
tial reaction rates in the biofilm.   Consequently, values well
above or below that value  may  be required.
     Even though external  mass transfer effects can be
important, a number of workers have  concluded that they were
not significant in their  systems and therefore have excluded
them from their biofilm models (B-13).  Howell and Atkinson
(8) modeled sloughing in a trickling filter and stated "It is
reasonable to assume  that  the  liquid phase diffusional resis-
tances in the packing units  are negligible..." and reference
                          350

-------
El Amin  (14) for  evidence regarding that assertion.   Shieh and
coworkers  (9,10)  based thetr decision to ignore  external mass
transfer effects  upon calculations of expected  reaction rates
in the presence and  absence of them.  Since  the  maximum
difference was no more than 7 percent for  the expected
reaction conditions  they concluded that the  error  was not
large enough to justify the additional mathematical  complexity
which the  inclusion  of external mass transfer would  introduce.
Jansen and Kristensen (11) used a rotating annular reactor
similar  to that employed by LaMotta (5) and  thus were able to
adjust the rotational speed to make external mass  transport
limitations insignificant.  Andrews and Tien (12)  and Wang
(13) simply assumed  that external mass transport limitations
were insignificant without giving their reasoning.  It should
be noted,  however, that the physical system  they were using
was similar to that  of Shieh and coworkers (9,10)  and thus the
reasoning  of the  latter workers may be valid in  this case as
well.
     Consideration of all available evidence suggests that
external mass  transport limitations should be considered when
developing biofilm models unless it can be specifically shown
that the effects  are negligible for the entire  range of condi-
tions under considerations.  This will generally require some
way of predicting these effects and thus the ability to model
external mass  transport effects is important regardless of
whether  those  effects are ultimately incorporated  into the
biofilm  model.
  Support
  surface"
          Support
          surface ~
                                                  Microorganisms
      (a) Pseudo homogeneous
           Model
(b) Heterogeneous
   Model
                                           (c) Hybrid
Figure 4.  Characterization of the biofilm:liquid interface.
           (From Atkinson and Howell  (17)).
                           351

-------
Modeling Techniques
     Before external mass  transport  can  be  modeled  it  is
necessary to consider the  nature of  the  biofilm:liquid  inter-
face.  Figure 4, taken from  the work of  Atkinson and Howell
(17) shows three ways in which that  interface  might be
viewed.  In the pseudohomogeneous view (4a)  the  liquid  in  film
flow is considered to move through the biofilm so that  no
clear liquid film exists.  While this concept  might charac-
terize trickling filters it would not accurately  depict other
fixed—film processes.  The heterogeneous  view, on the  other
hand, depicts a clear interface between  the  liquid  and  the
biofilm, so external mass  transport  occurs  in  a  totally liquid
layer.  With regard to the true conformation,  Atkinson  and
Howell (17) state:  "While the true  description  of  the
                                          Bulk  Liquid
   cc
   H
   Z
   u
   o
   z
   o
   o
                        DISTANCE,  x
Figure 5.  Idealized biofilm and stagnent  liquid layer
           illustrating concentration profiles  of  the electron
           donor (D) and the electron acceptor  (A).
                          352

-------
geometry of the slime surface  is  probably a hybrid between the
pseudohomogeneous and heterogeneous  systems,  visual
observations of slime layers and  the experiments of Atkinson
et al. (18), suggest that  the  heterogeneous system is to be
preferred."  Consequently, most researchers have adopted that
view and their models treat the biofilm:liquid interface as if
it were analogous to the interface  between a flowing fluid and
a solid support.  In contrast, Williamson and McCarty (19,20)
have depicted the liquid side  of  the interface as containing
two components:  one .adjacent  to  the biofilm which cannot be
removed by mixing and one  whose thickness depends upon the
turbulence in the liquid phase and  approaches zero at very
high velocity.  The former, which they considered to be
approximately 60pm thick,  would always present a resistance to
external mass transport.   They do state,  however (19):
"Whether such a layer exists in all  biofilms  is currently
unknown."  The implications of the  existance of such a layer
are quite important to modeling,  however, and will be
discussed more below.
     The usual way to depict the  necessity for external mass
transport is to imagine a  hypothetical stagnant liquid film or
boundary layer of thickness Lw between the biofilm liquid
interface and the bulk liquid  phase  as shown in Figure 5.  All
resistance to the transport of materials from the bulk liquid
to the biofilm is then assumed to occur in that layer.  Two
methods of modeling that transport  are commonly used.
     One method assumes that transport across the liquid layer
is by molecular diffusion, with diffusivity Dw.
Consequently, the flux, or mass of  substrate transported per
unit area per unit time, is given by


     N = ^ (Cb - C*)                                    (1)
          w

Because the diffusivity is an  intrinsic characteristic of the
mateial being transported  (the fluid is assumed to be water)
the thickness of the liquid layer,  1^, becomes the parameter
which must be evaluated before Eq.  1 can be used to depict the
rate of transport of the reactants  up to the biofilm.  On the
other hand, the value of LW will  depend upon the physical
configuration of the particular reactor being employed and the
fluid velocity within it.  Consequently,  equations relating
LW to those features represent one  place in which the
biofilm model interfaces with  the process model.
                          353

-------
     As discussed  above,  Williamson and McCarty (19,20) found
that the stagnant  liquid  layer  consisted of two layers which
they termed L^ and L£:

     Lw = L! + L2                                     (2)

The thickness of  the  outer  layer,  L^,  is dependent upon the
level of turbulence and can be  reduced to zero with adequate
mixing.  The thickness of the  inner layer, L£, is dependent
upon, the physical  characteristics  of the biofilm and was con-
sidered to be constant.   They  used the correlations of Welty
et al. (21) to relate L^  to the fluid  and physical charac-
teristics of flow  inside  pipes,  over flat plates, and through
packed solids.  A  similar approach was used by Famularo et al.
(22) and subsequently by  Mueller et al. (23) to model rotating
disk reactors.  In this case L£ was estimated from consider-
ation of the depth of surface  irregularities in the biofilm
and L]^ was calculated from  the  relationship of Levich (24)
for the thickness  of  liquid entrained  on a flat plate.
     Other investigators  have  also used Eq. 1 to estimate ex-
ternal mass transport but they  did not incorporate a fixed
layer, L£.  Even  though Williamson (19,20) originally pro-
posed the incorporation of  L£  into Lw, in a later investi-
gation he and Meunier (25)  stated  "From the review of various
formulas for this  parameter (Lw),  it appears that the ex-
pression proposed  by  Snowdon and Turner (26) for flow past
particles most closely models  conditions in packed and ex-
panded bed biofilm reactors."   The nature of this expression
is such tht Lw will appraoch zero  as the velocity of flow
gets large.  Since no mention  was  made of L£ one is uncer-
tain as to why it  wasn't  considered and whether the authors
have concludedd from  further work  that it is unimportant.
Mulcahy etal. (11) also  used  the  equation of Snowdon and
Turner (26) when  they determined that  external mass transport
resistances were unimportant for dentrification in fluidized
beds.  Since that  determination was based on computations of
utilization rates  with and  without a stagnant layer one must
wonder if the same conclusion  would have been reached if a
permanent layer had been  included.  To add further uncertainty
to the importance  of  the  layer  L£, KcCarty, who advised
Williamson's original work  (9,20), has since coauthored papers
with Rittiaann (27,28) which did not include such a layer.  In
one of these papers (27)  they  state "Many empirical formulas
for evaluating L in porous  media were  reviewed, and the one
presented by Jennings (29)  was  felt most appropriate...".
                          354

-------
However, in the other paper  they  appear to have used the
relationship of Welty et al.  (21)  as  used  originally by
Williamson (19).  Whether L£  was  included  was not stated.
     The other approach  to modeling external mass transport is
to use a mass transfer coefficient, k,  such that

     N = k(Cb - C*)                                     (3)

Comparison of this equation  to  Eq. 1  reveals that k is equiva-
lent to DW/LW and consequently  the comments made in the
preceding paragraphs are equally  applicable here.
Investigators using this approach  have  tended to use empirical
equations for the mass transfer coefficient taken from the
chemical engineering literature.   For example, Grady and Liia
(30,31) used the correlations of Mixon  and Garberry (32) (for
flow over flat plates),  Wilson  aad Geankoplis (33) (packed
beds) and von Karraan as  given by  Levich (24) (rotating disk)
in their work.  Dahodwala etal.  (34) used the relationship of
Brian and Hales (35) to  estimate k for  gently stirred
suspended particles.  Finally,  Mueller  et  al. (36) used the
relationship of Charpaatier  (37)  for  mass  transfer for clean
packed beds for their trickling filter  model.  The fact that
different models must be used for  different types of reactors
demonstrates agait how the biofilm model may be interfaced
with the process model.  The  significant fact about all of
these relationships is that  they were determined for clean,
nonporous material.  Given the  nature of the biofilm,  however,
one must wonder how applicable  they really are for situations
with high turbulence.  While  the mass transfer coefficient Cor
transfer to a nonporous  material  might  become quite large in
turbulent flow, is it possible  that the coefficient for
biofilms will approach some  maximum value  due to a permanent
stagnant layer like L.2?  This is  an area that needs further
investigation because the available evidence is not clear, yet
the implications of decisions made from the models are quite
important.
CHARACTERISTICS OF BIOFILM
     Substrate removal  in  any  heterogeneous  environment is the
result of interaction between  the  rates  of  transport  and the
intrinsic rates of  reaction.   Intrinsic  rates  of  biological
reactions are generally expressed  on  the basis  of a unit of
                           355

-------
biomass, e.g.,  the  Intrinsic rate of substrate removal is ex-
pressed as the  mass  of  substrate removed per unit time per
unit of biomass  and  the intrinsic rate of bioraass growth is
expressed as the amount of  biomass formed pec unit time per
unit of biomass  present.  Before considering the form of the
intrinsic rate  expressions,  consideration should be given to
the mass of organisms  likely to  be found in the bio film.
Since the surface area  available for biofilm colonization and
development is  generally a  physical characteristic of the type
of reactor being modeled, the  mass of biofilm in the reactor
becomes a function  of  the thickness and density of the bio-
film.  Furthermore,  it  is possible that not all of the organ-
isms present in  a biofilm wil  be capable of utilizing all sub-
strates entering it| e.g.,  if  organic matter and ammonia
nitrogen were both  present  only  th-s heterotrophic organisms
would be capable of  oxidizing  the organic matter whereas only
the autotrophs  could oxidize the ammonia nitrogen.  Conse-
quently, the composition of  the  blofilta :nay also be an impor-
tant determinant of  the potential reaction rates.
Biofilm Thickness
     When considering  biofilm thickness it is important that a
distinction be made  between  the  total film thickness and the
active film thickness.   In a. review of 10 papers in which bio-
Eiln thicknesses were  measured,  Atkinson and Fowler (2) found
that the total film  thickness was  between 0.07 and 4.0 mm.
They divided the biofilms into two groups, however, to better
reflect the growth conditions.  When the films vrere subjected
to mechanical or hydrodynamic control, the thickness was
generally less than  0.2  mm.   When  the films were uncontrolled,
however, they were as  thick  as 4.0 mm, though it has been
asserted that in turbulent flow systems, biofilm thickness
seldom exceeds 1 mm  (38).  Thicker,  uncontrolled films are not
likely to have greater substrate removal rates than thin
films because diflusional resistances within the film limit
the amount that is actually  contributing to substrate removal.
This amount is termed  the active layer and two types of evi-
dence for its existence  have been  gathered.
     The earliest evidence was based upon observations of
changes in the rate  of substrate removal as the depth of bio-
film increased in a  reactor  with a fixed surface area.  Those
observations indicated that  the  rate of substrate consumption
                           356

-------
increased as the biofilm  depth  increased up to a limiting
depth of 70—100  pm; after  that  the  removal rate was independ-
ent of depth (39, 40),  The  depth  at which substrate consump-
tion reached a maximum value was defined as the active depth.
Trulear and Characklis (6)  observed  a similar phenomenon,
although they also found  that the  active depth increased as
the substrate concentration  in  the liquid phase increased.
     Shortly after the observation that  the substrate removal
rate increased with depth up to  a  maximum,  Bungay et al» (41)
used a microprobe technique  to  determine oxygen profiles with-
in a film.  Their results indicated  that respiration ceased at
depths of 50-150um, depending  upon  the  substrate concentra-
tion is the medium.  This is consistent  with the interpreta-
tion that only the organisms in  the  active  layer are contri-
buting to substrate removal.  Similar observations have been
made by Hoehn and Ray (4) and by Chen and Bungay (3) as shown
in Figure 3.  The latter  workers also found, however, that at
low bulk substrate concentration,  the oxygen concentration in
the biofilm reached a constant  value at  some depth, thereby
demonstrating that the active layer  may  be  defined by deple-
tion of either the electron  donor  or the electron acceptor.
     It is now generally  accepted  that the  active thickness is
a result of transport limitations  within the biofilm.  Only
when the film is very thin, when the electron donor and accep-
tor concentrations are very  high,  or when the rates of trans-
port are large in relation  to the  reaction  rates will the
active film thickness approach  the total film thickness.  For
many biofilm reactors these  circumstances will not exist, with
the result that the total film  thickness (and by extension the
total amount of biomass)  has no  impact upon reactor perform-
ance.  If one knew in advance that the total film thickness
was in excess of the active  film thickness  then the system
could be accurately modeled  with any arbitrarily assumed
thickness because the differential equations depicting trans-
port and reaction within  the biofilm would  automatically show
a cessation of substrate  removal when either the electron
donor or acceptor was exhausted..  The need  to know the film
thickness arises, however, when  the  potential active film
thickness exceeds the thickness  that could  actually exist
under the given physical  circumstances because then the extent
of reaction will be limited  by  the actual film thickness.
     Prediction of the biofilm  thickness within a fixed—film
reactor is the least developed  of  all of techniques needed for
adequate modeling, primarily because relatively little funda-
                           357

-------
mental study has been devoted to the factors governing biofilra
development.  Some of the best experimental work on biofilm
development has been done by researchers interested in bio-
fouling of heat exchangers,  pipes, etc., and the reader is
referred to the review by Characklis (38) in this regard.
Basically a biofilm will continue to increase in thickness as
long as the rate at which the microorganisms are growing
exceeds their rate of loss by decay and by attrition.  In a
highly turbulent regime, attrition will be relatively constant
and appreciable so that, as  mentioned earlier,  biofilm
thicknesses seldom exceed 1000 pm.  Even in less turbulent
regimes, however,  steady state biofilms can develop when the
available substrate concentration is low because then cell
decay will balance the growth.  Unfortunately,  the general
         1.5
  ll

1.0
  a
  o
  m
         0.0
                           T
                               T
                           !
                               I
            100        150        2OO        250

                ROTATIONAL SPEED (rev/min)
 Figure 6.  Effect of rotational speed in an annular reactor on
           the detachment rate of a biofilm with a mass of
           150-160 mg.  (From Trulear and Characklis (6)).
                          358

-------
situation for most biofilm reactors in wastewater treatment is
one in which continual attrition is not sufficient to balance
the net growth, with the result that the film grows until
conditions develop near the support:biofilm interface which
cause adhesion to be lost and the film to slough away.  This
results in a continually dynamic state for the reactor, which
makes analysis particularly difficult.  Consequently, Atkinson
and Fowler (2) have suggested the application of positive
control over biofilm thickness and fluidized bed biofilm
reactors represent one reactor configuration within which such
control can be practiced.  In that type of reactor the height
of the bed is functionally related to the thickness of the
biofilm on the particles so that maintenance of a constant 'bed
height by the removal and cleaning of particles results in a
maximum known film thickness (11,42,43,44).
     The majority of the biofilm reactors used in wastewater
treatment are of a configurtion which prevents positive con-
trol of biofilm thickness.  This means that the film will
either reach a natural steady state in which growth is just
balanced by decay and attrition losses or it will increase
tu

i-s

P
w  «
Q  x
  IL.
  o
  m
                   r        i         i
             D RL  = 37.2 mg/sq m-min
             O RL  = 4.2 mg/sq m-min
                                         T
                 200
                       400
600
800
1000
                      BIOFILM MASS (mg)
Figure 7.
         Effect of biofilm mass (proportional to thickness)
         in an annular reactor rotating at constant speed on
         the detachment rate of biofilm.  RL refers to the
         organic loading applied.  (From Trulear and
         Characklis (6)).
                           359

-------
 continually until sloughing occurs.  Knowledge of the
 coadLtions  controlling which of those conditions exist is
 required for accurate modeling.  Unfortunately, relatively few
 studies  have been done on factors affecting attrition and thus
 the  data are limited.  The most complete study is that
 reported by Trulear  and Characklis (6) who grew fixed-films in
 an annular  reactor" consisting of two concentric cylinders, one
 stationary  and  the othec rotating.  The rotational speed
 determined  the  shear stress developing at the biofilm:liquid
 interface and the biofilm detachment rate increased as the
 rotational  speed increased, as shown in Figure 6 (6).  This
 suggests that the rate of biofilm detachment increases as the
 shear  stress at the  interface increases.  Furthermore, as
 shown  in Figure 7,  the detachment rate also increases as the
 biofilm  mass increases (6).  This suggests a mechanism whereby
 films  of different  thickness  can be attained in a reactor with
 a fixed  shear stcess as substrate is applied at various rates.

1
a.
at
W
tu
z
id
O
X
S
—
2
£0
*
X

S


1 WVW


800



600

400


200




n
1 K ' ' ' ' '
§
\
\ RL = 18 - 30 mg/sq m-min
\
it
\
\
\
X.
% ">%,
X "*"^.
\ "**^*»,
\^RC = 2,6-7 mg/sq m-min **'""""^^ -
<*»^
n ^--0-_ _ _ °


i i i i i i
                 2     4      6      8     10     12
                 INITIAL SHEAR STRESS (N/sq m)
14
Figure 8.  Effect of fluid shear stress and organic  loading
           (RL) on the maximum biofilm thickness attained in
           an annular reactor.  (Data from Zelvar  (45) as
           presented by Characklis (38)).
                          360

-------
At  low  substrate application rates (R^) the rate of new
biofilm growth would be low and would be balanced by  the
attrition  rate when a relatively low mass (or thickness) was
present.   At  higher substrate application rates, however,  the
film would be growing faster and thus a greater mass
(thickness) would develop before the detachment rate  balanced
the accumulation rate.   Evidence for this may be seen  in
Figure  8  (38) which is  from the work of Zelvar  (45).   The
curve in Figure 7 also  indicates that the detachment  rate
approaches a  very large value at a high biofilm mass,  thereby
suggesting that the fluid shear stress can limit the  maximum
quantity of attached biofilm in a turbulent flow regime  (45).
      Since the biofilm  mass is a function of the thickness,
these results suggest that the detachment rate associated  with
a  given shear stress will increase as the thickness increases,
thereby allowing the mechanism discussed above  to be  modeled.
Unfortunately, things are not that simple because Trulear  and
Characklis (6) have also shown that biofilm density is a
function of  the applied substrate loading rate.  This  means
that the thickness associated with a given mass was greater at
a  lower loading, thereby suggesting that the detachment  rate
cannot  be  expressed as  a simple function of thickness  but  must
be  expressed  in terms of both thickness and density.   What
functional forms should such models take?  Why does this
.relationship  exist?  Were the results influenced by the  type
of  re.actor employed? These and many other questions  must  be
answered before truly mechanistic models can be written
depicting  biofilra thickness.  This has been recognized by
Characklis who listed such questions in a recent review  (38).
One can only  hope that  work is continuing on such matters.
      Because  until recently relatively little was known  about
biofilm detachment, the vast majority of the models for  bio-
films have assumed a constant film thickness consistent with
the type of process under consideration.  Such an assumption
will probably work well for a fluidized bed biofilm reactor
because the biofilm thickness can be controlled and because
the fraction  of particles removed to control thickness is
small.   Furthermore, as long as the recycle ratio is  kept
above 2.0, the bed may  be considered to be completely  mixed
with respect  to the soluble compounds (46), and thus  all
                            361

-------
particles experience  the  same  reactant  concentrations.  It may
not be a good assumption  for other  fixed-film processes,
however, because the  concentration  gradients  within them could
cause wide variations in  film  thickness.   This could prevent
full films from developing  near  the outlet.   Thus,  in general,
it would be better to have  some  way of  estimating the biofilm
thickness.
     The major attempt  to model  biofilm thickness is that of
Rittmann and McCarty  (27,47) who have done so by assuming that
a steady state biofilm  is one  in which  growth would just  be
balanced by cellular  decay  so  the observed yield would be
zero.  Thus there is  no explicit term for attrition or
detachment in their model.  Rather, "it is assumed  that the
total amount of biofilm mass is  just equal to that  which can
be supported by the substrate  flux. The  steady-state-biofilm
thickness can then be computed by equating the available  and
maintenance energy rate..." (47).   This assumption  is probably
a reasonable one for  the  situation  for  which  they developed
their model, i.e., for  very low  substrate concentrations  such
as in ground water recharge.   The model did a reasonable  job
of tracking the substrate concentration profile through a
small tower even when intrinsic  parameters were utilized,
although it did not do  as well tracking the biofilm thickness
(27).  This may have  been due  to their  assumed constant
biofilm density, however.   The value of this  model  comes
from its ability to predict the  minimum substrate concentrtion
attainable in a fixed—film  system.   It  has been applied,
however to a broad range  of reactor configurations  which would
operate with electron donor and  acceptor  concentrations much
different from the ones for which it was  developed  (48).
Arcuri and Donaldson  (49) criticized the  basic assumption of
the model, stating that other  mechanisms  of cell loss would be
important in most biofilm reactors. This criticism certainly
appears to be valid.
     Recognizing that the steady-state  biofilm concept is
limited to a particular situation,  Rittmann (50) extended it
to incorporate detachment by shear  stress, liis analysis  was
based upon the data of  Trulear and  Characklis (6) and
incorported the concept that the detachment rate depended upon
the film thickness and  mass as well as  upon the shear stress.
He asserted that the  basic  steady-state biofilm model could be
employed for a broad  range  of  cases by  recognizing  that the
biomass decay rate, b,  could be  replaced  by a combined factor,
b1, which includes both decay  and attrition by fluid shear.
                           362

-------
From analysis of  the data  of Trulear  and Characklis (6) he
concluded that the attrition portion  of  b'  would be a function
of the shear stress alone  for  biofilms  less than 30 pm thick  .
but would be a function of  both  the shear stress and the
thickness for thicker  films.   This approach appears to be
quite reasonable, given the limited data available.  As point-
ed out previously, however, there  is  a  need for more work on
the subject since Characklis himself  raises questions about
how the detachment rate changes  with  fluid shear stress and
biofilm thickness (38).  When  the  researcher who develops data
indicates that more needs  to be  known about the relationships
involved, it could be  argued that  the development of mathema-
tical functions depicting  those  relationships is premature.
Furthermore, it should be  recalled from  the previous discus-
sion that the detachment rate  is an apparent function of the
biofilm density as well.   Rittmann (50)  did not incorporate
this, thereby giving another reason for  viewing his relation-
ships with caution.
     Andrews and Tien  (14)  have  developed a model for biofilm
growth on activated carbon  particles  that is similar in con-
cept to the steady-state biofilm model  of Rittmann and
McCarty.  Although they state  that their decay term "accounts
for both the basal metabolism  (cell maintenance energy) of the
bacteria and for wash-off  of cells from  the film" it is
assumed to be a constant and is  not a function of film thick-
ness, biomass density, turbulence, etc.   Thus the biofilm
thickness aspect of their  model  is essentially the same as
that of Rittmann and McCartyTs model  and the comments made
about it are equally applicable.
     In contrast to the steady—state  approach taken by Ritt—
mann and McCarty, Howell and Atkinson (8) modeled biofilm
thickness from the dynamic  point of view, i.e., they modeled
sloughing.  In their model  they  allowed  the film thickness to
increase over time by  assuming that no  continual detachment
occurred so substrate  removal  would" result  in accumulated cell
mass.  As the thickness increased  the substrate concentration
profile changed until  eventually the  concentration in the in-
terior of the film was too  low to  sustain the cells, thereby
allowing lysis to occur, leading to sloughing.  Recognizing
that there is a certain amount of  randomness associated with
sloughing, they arranged the model to take  that randomness in-
to account.  They then applied their  model  to a trickling fil-
ter and investigated the time-dependent  performance.  Because
of the sloughing the effluent  substrate  concentration always
varied in a dynamic manner  showing that  the nature of the bio-
                           363

-------
films  in  such reactors is in part responsible  for their
dynamic behavior.
     Because  of  the importance of film  thickness to the proper
modeling  of fixed-film reactors it is important  that accurate
(both  conceptually and mathematically)  models  be developed.
As seen,  however,  there are still many  questions to be answer-
ed.  A reasonable  start has been made but  in the opinion of
this author,  a much greater effort is still  required.   As will
be seen in later sections, many sophisticated  solution tech-
niques have been applied to biofilm models,  but  almost all
have been applied  to films of constant,  arbitrary thickness.
Thus it would appear that the questions  regarding biofilm
thickness should be resolved before more effort  is expended on
new, general,  biofilm models.
Biofilm Density
     Because  the  rates of reaction are  a  function of the mass
of microorganisms present, the density  must  be  coupled with
the thickness  and area to allow computation  of  the reaction
      120


      105


      90


      T5


      6O


      45


      30


      IS


       O
«*
           IOO 200  3OO 4OO BOO  600 700  800 9OO IOOO  MOO 1200 I30O  MOO ISC

                       MEAN FILM THICKNESS, MICRONS
figure 9.  Effect  of  biofilm thickness on the  density  of the
           biofilm growing on a rotating drum.   (From  Hoehn
           and Ray (4)).
                           364

-------
rate.  Although  it has  generally been assumed that the density
is constant and  independent  of  film thickness,  there is
evidence that this is not  the  case.  The first  to discover
that the density of a biofilm  depends upon its  thickness were
Hoehn and Ray (4) who obtained the results shown in Figure 9.
There it can be  seen that  the  density reached a maximum value
at a thickness consistent  with the active film thickness.
They postulated  that the changes in density were due to
variations in the microbial  populations within  the film.  The
maximum density  was thought  to  represent the tight packing
which would exist in the aerobic layer whereas  the . lower
densities were thought  to  be due to the lysis of cells in the
anaerobic region.  Shieh et  al.  (44) and Mulcahy and LaMotta
(51) have observed similar reductions in density with
increasing thickness, although they observed no region of
increasing density.  The latter  authors developed empirical
equations which  were subsequently used to calculate bioraass
concentrations in their model  for the fluidized bed biofilm
reactor (10,51).  Trulear  and  Characklis (6) also observed
changes in biofilm density,  but  their observations cause one
to ask whether thickness is  the  correct parameter to which to
relate density.  This is because the density in their films
approached a maximum value as  the substrate loading rate was
increased at a constant shear  stress.  It was seen previously
that the film thickness associated with a given shear
stress was also  a function of  the substrate loading, thereby
raising the question of whether  it is the thickness or the
loading (i.e. the net growth rate) that influences the
density.  Furthermore,  within  the limits of film thickness in
their studies, no decrease in  biofilm density was observed.
     There are many factors  which could be responsible for
changes in density, from the lysis postulated by Hoehn and Ray
(4) to the changes in culture  morphology observed by Trulear
and Characklis (6), and more fundamental research will be
required to delineate the  mechanism and its importance.  In
the mean time it is unclear  whether changes in  density should
be incorporated  into models  and  if so,  how they should be
formulated.  If  the observed decreases  in density are due to
lysis or culture changes in  the  interior (i.e.,  nonactive)
zones, should they be included  in models in which substrate
removal only occurs in  the active layer?  In other words, if
the sole purpose of the biofilm  density in the  model is to
obtain the reaction rate,  should a constant value be used?
If, on the other hand,  the density is required  to predict
other factors, such as  the bed  height in a fluidized bed
                          365

-------
model, then should  a  variable density be used?  These and
other questions must  be  answered before more exact models can
be developed.
Biofilm Composition
     Although  there  have  been  several observations of changes
in microbial composition  with  a  biofllm (4,6,38,52,53) most
models treat the  biofilm  as  if it were homogeneous throughout.
This is because most models  have sought to predict the fate of
one constituent such as organic  matter or ammonia nitrogen.
Nevertheless,  as  Alleman  and Veil (52) have noted
"»..fixed—film communities  likely include discrete microbial
strata, with divergent metabolic and  diffusional
characteristics.  Appropriate  refinements to these models may
therefore be necessary to insure their validity and utility".
While it may be some time before this is necessary for models
simply depicting  the removal of  soluble organic matter, it is
already necessary for models which depict the fate of both
carbon and nitrogen in biofilms.  One such model is that of
Mueller et_al» (23) which predicts the amount of carbon
removal, nitrification, and  denitrification in an RBC.  Carbon
removal is assumed to occur  by aerobic metabolism as long as
sufficient oxygen is present in  the film, but will occur by
denitrification when the  oxygen  concentration gets
sufficiently low  in the presence of a carbon source and
nitrate nitrogen.  Nitrification occurs as long as sufficient
ammonia nitrogen  and oxygen  are  present and the ratio of
heterotrophs to autotrophs at  any depth within the film is set
equal to the ratio of their  growth rates.
     The work  of  Bryers (54) represents the most ambitious
attempt at modeling spatial  profiles  within biofilms.  His
model can predict the profiles of heterotropic, Nitrospmonas
spp.  and Nitrobacter spp. within biofilms housed in a CSTE
receiving a constant input.  It  also  considers substrate
profiles for NH|", NO^, NO",  02 and acetate.  A finite
element technique is used to integrate dual substrate limiting
rate expressions  over both  time  and distance, thereby showing
changes within a  biofilm  as  it builds up.
     In general,  however, relatively  little work has been
performed to assess the composition of biofilms in general,
much less to look at how  they  might change with depth within a
given film.  One  might imagine,  however, that such information
                           366

-------
«ould be very useful  to the  development  of  a better under-
standing of how individual substrate  components might behave
in a fixed-film reactor.

TRANSPORT AND REACTION WITHIN  BIOFILN

     Once the electron donor and  acceptor have been trans-
ported up to the biofilm:liquid interface they must be carried
into the biofilm where the reactions  will occur.  These events
occur simultaneously  and  thus  the  concentration profiles of
the two constituents  in the  film  will reflect their relative
rates.  Since the reactive .capability of the biofilm (i.e.,
its overall average reaction rate) depends  upon the nature of
those concentration profiles,  the  heart  of  any biofilm model
is the conceptualization  and mathematical expression of the
simultaneous transport and reaction events.  Consider for the
moment the biofilm depicted  in Figure 2a.  Even though the
cells are held together in a complex  geometric arrangement and
have some sort of spatial distribution,  the majority of the
models assume that they are  uniformly distributed throughout
the film.  Because of the gelatinous  character of the biofilm
matrix it is thought  that convective  transport contributes
little to the movement of reactive constituents within the
film and that the electron donor  and  acceptor reach the organ-
isms by diffusion, which  is  characterized by Pick's law:

     N = DedC/dx                                        (4)

Unlike Eq. 1, in which the diffusivity was  given as DW, the
free diffusion coefficient in  water,  the coefficient is given
as an effective diffusivity  De, which reflects the fact that
the diffusion in the  biofilm will  generally be retarded
because it must occur through  the  gelatinous matrix.  If a
mass balance on a reactive constituent is then performed
around a differential element  of  steady-state film of constant
microbial composition, the result  is  an equation which is
almost universally used to model  reactions  within biofilms:
-D A
  e
          dx
                +  e!
                  D  A dC
              x
dx
           rAAx =0
                           x+Ax
in which A is  the  total  surface  area normal to the direction
of diffusion,  x.   The  key  elements which must be inserted into
Eq. 5 are the  effective  diffusivity, De» and the reaction
                           367

-------
rate expression, r, because  these  determine the nature of the
resulting concentration gradient and  overall reactivity.

Diffusion

     Given the importance  of  diffusion to the modeling of
fixed-film reactors there  is  surprisingly little agreement
about how the presence of  the biofilm influences the
diffusivity.  This is due  in  part  to  the fact that the
character of a particular  biofilm  depends upon the type of
organisms growing in it (55)  but it  is probably also due to
the many different techniques that have been used to estimate
the coefficient.
     The most direct method  of estimating De is to measure a
concentration gradient through a biofilm and to couple it with
the flux of material into  the film to allow direct computation
of D£.  This has been done by Bungay  and associates for oxy-
gen diffusion into laboratory-grown  (56) and actual trickling
filter films (3).  The coefficient for laboratory—grown films
was approximately 80 percent  of the  value in water whereas the
coefficient for field-grown  films  was about 35 percent.  It is
likely that the differences  in the values reflect differences
in the cultures residing in  the two  films.
     Another direct technique is to  place a film in a special
chamber which allows a component to  diffuse through it,
thereby allowing measurement  of the  flux.  Then from knowledge
of the film thickness the  diffusivity may be estimated.  Three
investigators have used this  technique (20,55,57),  Williamson
and McCarty (20) measured  the diffusivities of ammonia,
nitrite, nitrate, and oxygen  through  films  which had been
formed by filtration of dispersed  nitrifying bacteria onto
supporting membrane filters.   The  values were all in excess of
80% of the values in water.   Matsoti  (55) and Pipes (57) grew
mixed cultures of bacteria on glucose in completely mixed
reactors, concentrated then  by centrifugation, and formed them
into films by spreading them  onto  a  template with a spatula.
The biofilm was then sandwiched between two membrane filters
prior to placement in the  diffusion  apparatus.  Pipes (57)
grew his organisms at different carbon-to-nitrogen ratios and
found that the diffusivity of glucose ranged from 6 to 60
percent of the value in water, depending upon the growth con-
ditions.  Matson (55) not  only varied the carbon-to-nitrogen
ratio but also varied the  specific growth rate.  He found that
both parameters influenced the diffusivity  and that the value
for glucose ranged from 10 to 30 percent of the value in water
                           368

-------
whereas the value for oxygen  ranged from 20 to 100 percent.
Since the experimental reactors  displayed different
macroscopic characteristics it was  speculated that the most
important factor determining  the  diffusional characteristics
could have been the particular microbial species in residence.
     A third technique that has  been employed for estimating
De is to grow biofilms in  a fixed film reactor,  measure its
performance under a variety of conditions,  and then evaluate
De by curve fitting the  model under consideration to the
experimental data.  Using  a fluidized bed Andrews and Tien
(14) found the effective diffusivity of valeric  acid to be 34
percent of the value in water whereas Wang (15)  (in the same
lab) found it to be 67 percent.   Wang (15) also  found the
effective diffusivity of oxygen  to  be approximately 10 percent
of the free diffusion value,  although he stated  that the
uncertainty associated with the  number was expected to be
high.  Mulcahy et al (10,58)  calculated De for nitrate for
cells growing on a rotating disk  reactor and found it to be
approximately 50 percent of the  value in water whereas Jansen
and Kristensen (13) found  that it varied from 30 to more than
100 percent for films grown in a rotating annular reactor,
depending upon film thickness.   Although it was  apparent that
the value of De increased  as  the  film thickness  increased,
the finding of values in excess  of  the free diffusivity
suggested errors in the  estimation  of the reaction rate
constants which were ultimately  used to calculate the
diffusivity.  Furthermore  even though the variation in De
with film thickness is consistent with the variations in
biofilm density discussed  earlier,  these results illustrate
the dangers in computing coefficients from assumed models.
     Because of the difficulties  associated with direct
measurement of the diffusivity,  most modeling studies have
used assumed values.  Because of  their previous  work (20),
subsequent studies by Williamson  and his students (19,59,60)
and by McCarty and his students  (27,28,47,48,61) have assumed
effective diffusivities  of 80 percent of the free values for a
large number of substances.   The  modeling work done at
Manhatten College (22,36)  assumed a similar value based upon
that same work as well as  upon an analysis of the oxygen
profiles developed by Whalen  et  al  (62).  Harris and Hansford
(63) assumed that the effective  diffusivity of glucose was
equal to the value in water because of the results of Atkinson
and Daoud (64) and Atkinson and  How (65).  They  also used a
value equal to that in water  for  oxygen, but this time their
justification was that the spread in the reported values made
                          369

-------
it impossible  to  obtain  a reasonable estimate.
     From the  preceding  discussion it is apparent that there
is no consensus concerning the effects of Mofilms upon the
diffusive transport  of reactive species.  Furthermore, it
appears that this  lack of consensus is due to variations
within the biofilms  caused by  growth conditions, predominant
microbial populations, thickness,  etc.  Thus, while there can
be no doubt that  additional well defined and controlled
studies are needed,  perhaps the most logical approach to
modeling at the present  time is to just include the
diffusivity in one of the dimensionless groups that must be
evaluated during  experimentation (66).
Reaction Rate Expressions
     Fixed-film processes  are  generally used for one of three
purposes:  to remove  soluble organic matter, to convert
NflJ-N to N03-N (nitrification)  and to convert NOj-N to
N£ (denltrifIcation).   In  some  cases more than one of these
may be accomplished  in  a single reactor,  but in all cases two
soluble, transporting components are necessary for the
reactions to occur.   These are  an electron donor and an elec-
tron acceptor.  In processes focusing on the removal of solu-
ble organic matter,  that organic matter serves as the electron
donor and oxygen serves as the  electron acceptor.  (In anaero-
bic fixed-film reactors some other constituent will serve as
the electron acceptor.  The situation is complicated, however,
by the nature of the  microbial  interactions involved and thus
it will not be considered  herein).  When nitrification is the
objective, NHT—N serves as the  electron donor and oxygen
again serves as the  acceptor.   Some nitrification models seek
to also account for  the production and subsequent oxidation of
NC>2~N to NO-j—N but the  bulk consider only the oxidation of
     .  Finally, when denitrification is the objective,
   ^N serves as the  terminal electron acceptor and some form
of organic matter generally serves as the donor; the focus is
generally on the fate of the NO^-N, however.  Consideration
of these processes suggests that they can be generalized by
writing the reaction  rate  expressions in terms of the
concentrations of the electron  donor, CQ, and the electron
acceptor, CA»  That  approach will be taken herein.
     It is now widely recognized that in the most general case
the reactions within  a  biofilm  may be controlled by the
                           370

-------
concentrations of  both  the  electron donor and the electron
acceptor.  If the  concentration  of  one is much higher than the
other however, then  only  one  constituent controls.  This
latter situation has  been assumed to exist by most modelers,
and thus most of the  models have been written in terms of only
one constituent.   Thus  let  us  first examine the basic single-
substrate rate equations  and  then expand them to the two-
substrate case, which will  serve as a more general model.
     Cell growth and  substrate oxidation are generally con-
sidered to be coupled reactions, i.e.,'substrate removal
occurs because of  cell  growth.  The proportionality constant
is the true growth yield, Yg.   Furthermore, the rate of cell
growth is proportional  to tne  cell  concentration or density
within the film, X^:
     r_ =
where   is  the  specific  growth rate,  T
the substrate removal  rate:

     -rs =  q Xf
                                       	i
         (6)

Likewise with


         (7)
where q is  the  specific  substrate removal rate, I  , which
is related  to the  specific  growth rate by
     q =   p/Y
             g
         (8)
These definitions assume  that  all  substrate utilization is
channeled into  cell  synthesis  and  that cell maintenance needs
are met by decay.  Another  approach would be to assume that a
portion of the  substrate  was  channeled directly into cell
maintenance.  Although  there are differences in the fundamen-
tal mechanisms  employed by  the two models both yield the same
result and can  be considered  to  be equivalent (30).  In the
few models where cell maintenance  energy needs have been con-
sidered, the growth/decay concept  has been employed.  Conse-
quently, it will be  used  here  as well.
     A multitude of  models  could be (and have been) written to
depict the relationship between  the specific growth rate of
bacteria and the concentration of  a single limiting nutrient,
since all such  models are strictly empirical (30).  Conse-
quently, this review will be  limited to the two most widely
used ones:  Monod (68)
                           371

-------
    y =
    V
    K + C
Blackman (69)
        y c
            for C < 2K;    y = y  for C > 2K
                                m       —
           (9)
                                                    (10)
In these models y  is the maximum specific growth rate and K
is the saturation constant.  A plot of these models is depict-
ed in Figure 10.  There they are shown in dimensionless form:
Monod
    JL- =
    y    l+c/K
     m
Blackman
— = ^ £ for C/K < 2;    i^- = 1 for C/K > 2
                                                    (11)
                                                        (12)
     m
    1.0 -
                              m
                 Blackman
                          Monod
                           6      8
                             C/K
                                    10
12
14
Figure 10.
        Dimensionless plots of Monod and Blackman kinetic
        models for substrate limited growth of bacteria.
        (Adapted from Badar (67)).
                         372

-------
The Monod model has been widely  used  to  depict  the removal of
soluble organic matter and  the oxidation of  NH^~N in fixed-
film reactors under the assumption  that  the  electron acceptor
is present in unrestricting concentration.   The Blackman model
has found extensive use in  the modeling  of  denitrification in
fixed-films when the  concentration  of both  the  electron donor
and acceptor are high.
     As discussed previously, because of the concentration
gradients within the  biofilm, it  is likely  that the
concentrations of both the  electron donor and the electron
acceptor could limit  the rates of reaction.   Thus, it would be
desirable to have a general model which  handles all types of
doublesubstrate limitation.  As  Bader (67)  has  pointed out,
however, "this becomes rather difficult  since two separate
schools of thought exist about the  nature of growth with two
limiting substrates,  and there is insufficient  experimental
data to support either school.   In  fact, it  is  doubtful that
sufficient experimental evidence will be developed in the near
future".  Thus a review of  the current state-of-the-art of
fixed film modeling must incorporate  the two philosophies,
which have been labeled noninteractive and  interactive.
     6
  CO
  co
  o
N
1(
DP
od
K

1 1 1
netics
M/jjnT
' 	 0.80-



** f*
	 0.40
L 	 . 	 . 	 . 	 ,-0,20
1


31
ac
,k
man Kinetics
M/pm



t 1 1 1 I I
) 2 4 6 0 2 4 6
Cd/Kd Cd/Kd
     2  -
Figure  11.  Plots  of  lines  of  constant dimensionless specific
            growth rate  (p/ym)  as  a function of the
            dimensionless  concentrations  of electron donor
            (CD/KD) and  electron acceptor (CA/KA) for
            noninteractive  models  using Monod and Blackman
            kinetics.   (Adapted from Bader (67)).
                          373

-------
     A noninteractive  model  is  based upon the concept that the
specific growth rate of  an organism can only be limited by one
substrate at a time.   Therefore,  the specific growth rate will
be equal to the lowest rate  that  would be predicted from the
separate single-substrate models.  For the Monod model, this
may be written:
p  _
p_
                          D
                              C,
          1+C.
             D/KD
                    for  ^ < ~
      m
          CA/KA
          I+CA/KA
K.
                   for
K
                                                        (13)
where the subscript D  refers  to  the  electron donor and A
refers to the acceptor.   Similar equations  may be written for
Blackman kinetics.  Graphs  of  constant  dimensionless specific
growth rate (p/pm) as  a  function of  dimensionless substrate
concentrations (Cp/Kj)  and C^/K^) are shown  for the two
types of kinetics in Figure 11.
     An interactive model is  based upon the assumption that if
two substrates are present  in  less than saturating concentra-
tions, then both must  affect  the overall specific growth rate
of the organism.  One  type  of  interactive model may be con-
structed by multiplying  two single-substrate limited models
  ra
  a
  O
        Monod Kinetics
                                 Blackman Kinetics
                                                  M/Mtn
                                                  • i.oo
                                                  • O.75
                                                  • O.SO
                                                  ' O.25
            2     4
             Cd/Kd
                                     2     4
                                      Cd/Kd
Figure 12.
            Plots of lines of  constant  dimensionless  specific
            growth rate  (p/)^)  as  a function of the
            dimensionless concentrations  of  electron  donor
            (CC/K£)) and  electron acceptor (C^/K^)  for
            interactive  models  using Monod and Blackman
            kinetics.  (Adapted from Bader (67)).
                          374

-------
together, as shown below  for  the  "double  Monod model":


                                                        (14)
      m
A similar equation could  be written  for  Blackman kinetics.
Plots of these models are  shown  in Figure  12.   Howell and
Atkinson (70) have utilized a  slightly simpler  model proposed
by Bright and Appleby (71) and have  found  that  for parameter
values likely in the processes employed  in wastewater treat-
ment there would be little difference in the results obtained
with it and with the double Monod model.
     To summarize, there  are six categories of  models which
could be used to depict the specific growth rate,  y, of  the
microorganisms in the biofilm: single-substrate, Monod (SSM);
single-substrate, Blackman (SSB); noninteractive double-
substrate, Monod (MDSM);  noninteractive  double-substrate,
Blackman (NDSB); interactive double-substrate,  Monod (IDSM);
and interactive double-substrate, Blackman (IDSB).
     The complete reaction rate  expression for  removal of the
electron donor in a biofilm may  be obtained by  combining Eqs.
7 and 8, yielding

     -r  =  yX,/Y                '                       (15)
       s      f  g
and then substituting the  appropriate equation  for y into the
result (i.e., Eq. 14).  The complete reaction  rate expression
for net growth of cells must combine two loss  terms with Eq, 6
for growth:

     r  =  pX- - bX, - r                                (16)
      x      f     f    a
where the term bXj represents  the decay  of cells for main-
tenance purposes and the  term  ra is  the  loss by attrition.
The rate constant b has been taken to have a fixed value or
has been considered to be  a function of  the concentration of
the electron acceptor (22):
         b'CA                                           (17>

     b = Kl+c7
          A  A
                           375

-------
     The  rate  of  utilization of the electron acceptor will be
related stoichiometrically  to the rates of oxidation of the
electron  donor for  cell  synthesis and the decay of the cell
mass:
                     bX
       r  =
                                                        (18)
        A   Y  A       *V
             gA       b

where Y~^ is a conversion  factor relating the mass of
cells formed to  the  amount  of  electron acceptor utilized for
cell synthesis and Y^ is a  factor relating the mass of cells
lost to decay  to  the mass of  electron acceptor utilized for
decay.  If oxygen is the electron acceptor, Yg^ will be
related to the true  growth  yield, Yg (mg cells/ing COD
removed) by:

               Y
      Y   = 1-B?                                        (19)
       §A       g

where 8 is the oxygen equivalence of the cell material (often
taken as 1.25  mg  02  or COD/mg  cells or 1.44 mg02 or COD/
mg volatile solids)  (30).   Yj,  is just given by

      Y.  = 1/6                                         (20)
       b
When N03-N serves 'as the electron acceptor, Y-^ is given
by:

            2.86Y
                CS

and YJJ by

      Yb = 2.86/B                                       (22)

where Y~ and  0 are  on  a  COD  basis.   It is assumed in all of
the above that biodegradable COD is  used to express the con
centration of the electron donor.   If NH^-N served as the
electron donor appropriate conversion factors would be
required in all of  the equations.
                           376

-------
     Having delineated the equations  for  the  reactions in the
biofilm we now have established a  framework which can be used
to categorize fixed—film models.   The  following questions
should be asked in the categorization:
     (1)  Which type of reaction rate  expression is used?
     (2)  If a single-limiting substrate  model  is employed,
          what is the limiting material,  the  electron donor  or
          the electron acceptor?
     (3)  Is the utilization  of the nonlimiting material cal-
          culated?
     (4)  Is cell decay included in the electron acceptor
          balance?
     (5)  Is the film thickness an input  to  the biofilm model
          (either by assumption or by  interfacing with the
          process model) or is it  an  output  from the biofilra
          model?
     (6)  Is a distination made between thick and thin films?
          (A thin film is one in which both  reactants pene-
          trate to the support:biofilm interface).  Such a
          distinction is necessary in  establishing the boun-
          dary conditions and solution techniques for some
          models but is not needed with others.
     Table I presents a review of  some recent models in terms
of these questions.  Examination of the table reveals that
many models have similar characteristics.  When two or more
references appear on the same line, the models  in them are
very similar, both with respect to the rate  equations employed
and the details involved.  When two lines  have  the same en-
tries, the models on them have similar characteristics but
differ materially from one another in  one  or  more ways which
are not reflected by the questions.   Not  surprisingly, single-
substrate models constitute the bulk  of those which have
appeared in the literature, primarily  because of the evolu-
tionary nature of research.   It is now known  that both the.
electron donor and the electron acceptor  can  be important
determinants of the performance of a  fixed-film process and
thus the more recent models have sought to account for the
effects of both.  However, both interactive  and noninteractive
double-substrate limited models have  been employed, reflecting
the two philosophies discussed earlier.   It  appears" to this
author, however, that use of  the noninterative  model is less
straight-forward because it requires  identification of the
limiting substrate within the biofilm before  the correct equa-
tion can be chosen.  Furthermore,  it  is possible for the
limiting substrate to change  with  depth.   Such complications
do not exist with the interactive  models,  although they re-
quire solution of simultaneous differential  equations.  Many
of the models require that the film thickness be an input,
                           377

-------
             Table  I.   CHARACTERISTICS  OF  8IOFILM RATE EQUATIONS
Model
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Type of
Reaction
Rate
Expression!
SSH
SSH
SSH
SSH
SSH
SSH
SSH
SSH;SSB
SSB(1)
SSB(1)
SS8
SSB(0>
NDSH
NOSH
NDS8(Q)
IOSH
IDSH
IDSH
Limiting
Haterial2
ED
ED
ED
ED
ED
ED
ED
ED
ED
ED
EA
EA "
EO,EA
ED,EA
ED.EA
ED.EA
ED.EA
ED.EA
Consideration
of Nonlimiting
Material
Utilization?3
No
No
No
No
No
No
No
No
No
Yes
No
No
N/A
N/A
N/A
N/A
N/A
N/A
Cell Decay
in Electron
Acceptor
Balance?'
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
No
No
No
No
No
No
No
No
Yen
Film
Thicknes
Input
Input
Output
Output
Input
Output
input
Input
Output
Output
Input
Input
N/N
Input
Input
Output
Input
Input
Distinction
J3* between
Thick and
Thin Fiima?5
Yea
N/N
N/N
Yea
Yea
Yea
Yes
N/N i No
N/N
N/N
Yes
Yea
Yea
Yes
Yes
Yea
N/N
N/N
References
17, 72
30, 31
8
70
28
47
61
73
14
15
11
9,10,12,51
19
25, 59
13
70
63
22,23,36
1  SSH, aingle-substrate, Monodi SSB,  single-substrate,  Blackman,  (0)  s zero order only,
     (1) - first order only; NDSH, noninteractive  double-substrate,  Honod;  NDSB(D),
     noninteractive double-substrate,  Blackman,  zero  order  only;  IDSM,  interactive
     double-substrata, Honod.
*  ED s electron donor; EA = electron  acceptor
'  N/A means that the question ia not  applicable to the  particular reaction rate expression
     esuployed.
   Is the value of the film thickness  an  input or  an  output?  N/N  means that it is not
     necessary to know the thickness to proceed.
   N/N means that it is not necessary  to  make a  distinction to solve the equations
	employed.	
                                             378

-------
either by assumption  or  by  interfacing with the process model.
Others, however, provide  the  film thickness as an output from
the biofilm model, either by  involving some type of steady-
state assumption or by solving  the  dynamic case.  With some of
the models knowledge  of  the film thickness is required to
establish the solution technique which will be employed.
Prior knowledge of the film thickness  is particularly critical
when a zero-order rate expression is employed because the
point of substrate disappearance in the film must be identi-
fied to have the proper  boundary condition for solution of the
differential equation.
     In summary, it is apparent  from Table I that there are
many ideas about how  the  rate equations for biofilms should be
written.  No doubt those  ideas will continue to develop and
change as we learn more  about the processes.

SOLUTION TECHNIQUES

     In the preceding sections we have established that trans-
port of materials both up to  and through a biofilm can have a
significant effect upon  the rates of reaction achieved by that
film.  This has an important  impact upon the way that models
of reactors containing biofilms  must be solved.  Consider for
the moment a plug-flow reactor with biofilm distributed along
its length.  In modeling  that reactor  our primary interest
will be in the change in  substrate  concentration axially with-
in it; in other words we  want to know  how reactor length in-
fluences performance.  We realize,  however, that more than one
substrate concentration  exists at any  axial position within
the reactor.  Returning  to  Figure 5 we see that the concentra-
tion in the bulk fluid is higher than  that existing at the
biofilm:liquid interface, and furthermore, that the concentra-
tion at the interface exceeds the concentration surrounding
the organisms within  the  film itself,  all because of the
necessity for transporting  materials from the bulk fluid
through the biofilm.  This  means that  the model for our reac-
tor must combine mass balance equations in the axial direction
with mass balance equations in a direction perpendicular to
that axis (i.e., like Eq. 5).  The  terms in these equations
must reflect both reaction  and  transport.  In the axial direc-
tion, transport will  be  primarily by fluid flow and reaction
will be limited to that  caused by organisms being carried with
the fluid.  In the direction  perpendicular to flow, transport
will be by eddy diffusion in  the liquid film and by diffusion
within the biofilm.   The  bulk of the reaction, however, will

-------
                                 \
be caused by  the  organisms  within  the biofilm.  Although the
complexity of  the  resulting equations will depend upon the
particular characteristics  of the  process being modeled, it
will be necessary  to  solve  the two equation sets simultaneous-
ly.
     Three approaches have  been used for solving the equations
in the overall process  model: direct, indirect and with an
effectiveness  factor.  In the direct approach appropriate
numerical techniques  are  employed  to solve the two sets of
equations simultaneously.   As a consequence, the entire set
of equations must  be  solved every  time the model is employed
to investigate a  new  condition.  In the indirect approach, Eq.
5, which depicts  reaction and transport within the biofilm, is
solved for various  concentrations  of reactants in the bulk
liquid and the result is  expressed as the flux of material
into the biofilm  (which is  equal to its removal rate from the
bulk fluid) as a  function of the bulk fluid concentrations,
transport characteristics,  etc. This flux relationship can
then be used during solution of the process equations in an
iterative manner,  i.e.,  the concentration leaving the control
volume by fluid flow  is  assumed and the corresponding reaction
rate is determined  from the flux:bulk concentration relation-
ship.  The flux is  then  used to calculate the concentration
leaving the control volume  and the procedure is repeated until
the two concentrations  agree.  Even though an iterative proce-
dure is utilized,  the indirect technique is more efficient
because the second  order  differential equation resulting from
Eq. 5 need only be  solved once to  establish the flux:bulk con-
centration relationship  and then that relationship can be used
with any process  model.   The effectiveness factor approach is
similar to the indirect  approach but results in a somewhat
more general solution.   The effectiveness factor is defined
simply as the  ratio of  the  actual, observed reaction rate in
the presence of mass  transport limitations to the theoretical
rate in their  absence (i.e.,  the intrinsic reaction rate)
(30).  As such it  becomes a correction factor that can be
applied to the reaction  rate as calculated from the intrinsic
kinetics at the bulk  substrate concentration, thereby convert-
ing that rate  into  the  actual rate occurring in the presence
of mass transport  limitations both up to and through the bio-
film.  The second  order  differential equation resulting from
Eq. 5 is solved with  the  appropriate boundary conditions to
obtain the concentration  gradients through the biofilm asso-
ciated with various bulk substrate concentrations.  The aver-
age reaction rate  is  then obtained by integrating over the
entire biofilm depth  and is divided by the intrinsic reaction
                          380

-------
rate at the corresponding bulk  substrate  concentration to get
the value of the effectiveness  factor.  By  doing  this  for a
large number of conditions, effectiveness  factors can  be ob-
tained as a function of  the various  parameters  describing
transport to and through the  film.   Once  this  effectiveness
factor relationship has  been  determined it  can  be used with
any type of process model.  Details  of  the  procedures  required
to obtain effectiveness  factor  relationships are  given else-
where (30, 66).  Like  the indirect  technique,  once the effec-
tiveness factor relationship  is known,  the  second order
differential equation  arising from Eq.  5  need  not be solved
again to solve the process model.  Rather  the  approach to the
model solution is quite  similar to  the  indirect approach dis-
cussed above.  Let us  now categorize the  models in Table 1 in
terms of the solution  technique employed.

Direct Technique

     Six of the models in Table 1 were  solved  by  direct tech-
niques: numbers 4, 9,  10, 16, 17, and 18.   Howell and  Atkinson
(70) used both a single-substrate Monod model  (#4) and an
interactive double-substrate  (electron  donor and  electron
acceptor) Monod model  (#16) to  determine  the active film
thickness associated with various concentrations  of substrate
and oxygen at the biofilm:liquid interface.  When the  effects
of both the electron donor and  acceptor are being considered
an equation like Eq. 5 must be  written  for  each component,
with appropriate reaction rate  expressions  substituted into
each.  For this more general  case,  an interactive model was
utilized.  Taking the  limit as A* approaches zero yields two
second-order differential equations  which  must  be solved
simultaneously.  These form a two point boundary  value problem
which is inconvenient  to solve.  However,  by regarding the
film thickness as an unknown  and by  assuming coupling  between
the removal of the electron donor and the  electron acceptor
(i.e., b in Eq. 19 was set equal to  zero) Howell  and Atkinson
were able to convert the problem into an  initial  value problem
which could be readily solved.  Solutions were  then obtained
for a number of interface concentrations,  yielding graphs
which showed how those concentrations influenced  the active
film thickness.  They  also solved the equations for the
situation where only the elector donor was  rate limiting by
setting K^ equal to zero, thereby making  the reaction  rates
zero-order with respect  to the  concentration of the electron
acceptor.
                          381

-------
     Andrews and Tien  (14) used  first—order substrate—limited
kinetics (#9) to model  biofilm growth  and  adsorption In a CSTR
containing activated carbon  granules.   The electron donor (and
adsorbate) was valeric  acid  and  the  electron acceptor was ni-
trate.  A direct solution was used because the assumption of
first—order kinetics and the absence of external mass transfer
resistance made it  possible  to obtain  an explicit solution to
the second-order differential equation which results from Eq.
5.
     Wang (15) extended the  work of  Andrews and Tien (14) by
considering biofilm growth and adsorption  in a fluidized bed
reactor.  Although  the  kinetics  were again taken to be first
order with respect  to  the concentration of electron donor
alone, the utilization  of electron acceptor was accounted for
by stoichiometry (assuming no decay).   In  addition, the pre-
sence of two electron acceptors  (oxygen and nitrate) was
accounted for so that  the biofilm was  divided into two re-
gions, aerobic and  anoxic.   As a result the system model con-
tained a large number of simultaneous  equations which were
solved numerically.
     Harris and Hansford (63) incorporated their biofilm model
(#17) into a process model for a vertical  biofilm with a thin
liquid film flowing over it.  Their  biofilm model was written
in terms of both the elector donor and acceptor with an inter-
active double substrate Monod equation like Eq. 14, resulting
in two simultaneous second-order differential equations.  The
equations were directly coupled, however,  because b in Eq. 19
was assumed to be zero.  Only the situation without recircula-
tion of fluid around the film was considered by breaking the
vertical biofilm up into a number of sequential sectors.
Starting with the first  sector the concentration of electron
donor entering in the  liquid phase,  CQJ, was known and the
concentration leaving  (C^) was  assumed.  By knowing the
liquid flow rate through the sector  it was then possible to
calculate the substrate removal  rate which must equal the flux
of substrate into the biofilm.   The  flux of electron acceptor
was then calculated from stoichiometry. Knowing the fluxes,
Op avg, q[, and the external mass transfer coefficients
made it possible to calculate the concentrations at the bio-
film: liquid interface,  Cp and CA-  These  provided one set
of boundary conditions  for the simultaneous differential equa-
tions which were solved numerically  to obtain the fluxes.
These fluxes were compared to the external fluxes and the pro-
cedure was repeated until they agreed.  The known output from
the first sector then became the input into the second sector
                         382

-------
and the procedure was  repeated  on  down  the  vertical face of
the biofilm.  The results were  then  given as  concentration
profiles down the reactor.
     Mueller etal. used  the  sector  technique to model the
performance of an RBC  (22,23,36) as  well as a trickling filter
(36),  They also used  an  interactive double-substrate Monod
model (//18) but unlike the others  they  took cell decay into
account when writing their rate equation for  the electron
acceptor.  To simplify the determination of the  concentration
gradient into the biofilm they  also  broke it  up  into sectors.
The biofilm model was  coupled with the  process model by equa-
tions depicting external  mass transfer, diffusion,  'etc., and
the entire system model was solved by finite-difference tech-
niques.  This biofilm  model is  one of the most complete, tak-
ing into account carbon oxidation, nitrification,  and denitri-
fication.  Perhaps as  a consequence,  less detail has been pro-
vided in the literature about the  solution  techniques
employed.
     A review of the models that have been  solved by direct
techniques reveals that all of  the interactive double— sub-
strate Monod models fall  into this category.   This  is probably
because of the large number of  parameters required,  which make
it more expeditious to obtain a direct  solution  for a specific
application than to try to develop the  dimensionless groups
required for either the indirect or  the effectiveness factor
approaches.  The use of the direct technique  limits their
flexibility, however,  and requires a relatively  large effort
to obtain a solution for  a new  situation.   The other models
which were solved by direct techniques  also were quite compli-
cated but in one case  an  explicit  solution  was possible.
Nevertheless, it appears  that direct solutions to complete
process models have been  limited to  specific  problems for
which general solutions are difficult.

Indirect Technique

     In contrast to the direct  technique in which the biofilm
and process models are solved together, the indirect technique
employs a generalized  solution  form  for the biofilm model to
arrive at specific solutions  for particular process models.
The generalized solution  for  the biofilm model often takes the
form of a family of curves, although simplified  equations have
also been employed.  The  process model  is then solved by using
the generalized biofilm model in an  iterative fashion.  Four
of the biofilm models  in  Table  I (numbers 6,  7,  13,  14) were
solved directly to arrive at  generalized solutions  which
                            383

-------
could subsequently  be  used  to  solve a number of process
models.
     One of  the earliest  biofilm models to be solved in a man-
ner which makes it  available for use In the indirect technique
is that of Williamson  and McCarty (19) (#13 in Table I).
Because it is a noninteractive double-substrate Monod model,
solutions are presented for only one limiting constituent.
Selection criteria  are provided for determining which consti-
tuent (ED or EA) is  rate  limiting,  although solution is
restricted to the situation where a single constituent is
limiting throughout  the entire film depth.  When Eq. 5 is
applied to a single  limiting constituent one second order dif-
ferential equation  results.  To solve the equation they made
use of the fact that the  concentration and concentration gra-
dient of the limiting  constituent approach zero at a depth
corresponding to the active film depth.  They used a Runge-
Kutta finite difference technique starting at an interior
point where  the concentration  of limiting constituent was set
equal to a small, nonzero value.  Computation then proceeded
in small steps toward  the biofilm surface, with the concentra-
tion and gradient of the  limiting constituent being calculated
at each step.  When  the concentration equaled or slightly ex-
ceeded a preset interface concentration, C*, the computation
was stopped and the  flux  was calculated as the product of the
effective biofilm diffusivity  and the concentration gradient
just inside the biofilm.  The  results were presented as graphs
of active film thickness  and limiting constituent flux as a
function of C*.  Plots were prepared for five different values
of K (including zero)  and each plot contained seven curves for
different values of  the group  jjmXjDe/Yg.  These curves
can be used to solve any  fixed-film process model.  To incor-
porate external mass transfer  effects, the flux and the bulk
substrate concentration are assumed and the value of C* is
determined.  Using C*  and the  appropriate graph, the internal
flux is determined  and compared to  the assumed value.  If they
agree, the flux is  correct  and the  removal rate associated
with the known bulk  concentration is known.  If not, a new
flux is assumed and  the procedure repeated.  While this solu-
tion technique makes it possible to model processes without
recourse to complex  numerical  techniques, the indirect solu-
tion provided by the graphs is limited in the number of para-
meter values considered.  Furthermore, one must determine
beforehand whether  the electron donor or acceptor is limiting.
These limit  the model's utility.  Nevertheless, this model
served a useful purpose as  the starting point from which other
                           384

-------
models have been developed.
     Williamson has  continued  to  work with the noninteractive
double—substrate Monod model with his latest  effort being that
with Meunier (25)  (Model #14 in Table I).   The basic biofilm
model is similar in  concept to the preceding  one but the solu-
tion approach is different.  Using the technique of Chung
(59), the second order differential equation  arising for Eq. 5
was integrated once  by assuming  that the  concentration
approaches zero within the film,  thereby  giving an equation
for the substrate  concentration gradient.   Multiplication of
the value of the gradient at the  biofilm:liquid interface con-
centration, C*, by the effective  biofilm  diffusivity, De»
results in the flux  associated with C*.   This flux can ulti-
mately be expressed  in terms of  the bulk  concentration, C ,
through knowledge  of  the pass  transfer characteristics.  This
technique is only  applicable when the concentration of the
limiting component approaches  zero within the biofilm and thus
the solution is limited to what are called "thick" or "deep"
biofilms.  As seen earlier, most  practical wastewater systems
fall within this category.  Furthermore,  the  solution is only
valid when a single  constituent is limiting throughout the
entire film.  Because of this; and because there will be some
range of bulk fluid  concentrations over which the limiting
constituent changes  within the biofilm, Meunier and Williams
(25) have expressed  their, biofilm model solutions in the form
of operating diagrams which can be used to solve specific pro-
cess models (60).  These operating diagrams show substrate
flux as a function of C^/CQ, the  ratio of.the bulk fluid
concentrations of  electron acceptor and electron donor.  In
developing the diagrams., CS. is fixed and  the  biofilm model
is solved for various C^ concentrations.   This can be done
both for the region  where the  electron donor  limits throughout
the biofilm (which gives a single value of the flux for the
fixed Cj} value) and  for the region where  the  electron accep-
tor limits throughout the film (which gives a flux value for
each value of G£).   Two curves are obtained when these
values are plotted on the operating diagram and these curves
are connected by extrapolation to obtain  the  flux in the
region where the limiting constituent changes within the bio-
film.  This must be  done for a number of  Qp values to
generate the complete operating diagram.   Because specific
values for the kinetic and mass  transport parameters must be
assumed to generate  the operating diagrams, each diagram is
specific for a given biofilm process.  Once it has been gener-
ated, however, the performance of that process can be evaluated
                           385

-------
under a large number  of  conditions  without resolving the
differential equations.   Furthermore,  because the concentra-
tions of both the electron  donor  and acceptor are incorporated
into the operating diagram,  no  further consideration need be
given to which is limiting  while  utilizing the diagram.
     Rittraann and McCarty (61)  also used the integration tech-
nique of Williamson and  Chung  (59)  to  solve a single-substrate
Monod biofllm model (#7).   The  same general solution approach
was utilized but because only  one limiting constituent was
considered they were  able  to present their results in dimen-
sionless form, thereby increasing the  generality of their" plot
of flux versus bulk substrate  concentration.  The parameters
in their plot were effective diffusivity and active depth,
both in dimensionless form.  A.S a consequence their curves can
be applied to any combination  of  kinetic and mass transfer
parameters.  The major limitation,  however,  is that they are
limited to thick films because  of the  use o£ the integration
technique.  From inspection  of  their curves  they developed
simplified equations  to  depict  them, thereby facilitating
their use in the solution of a  broad range of models for pro-
cesses which contain  thick  biofilms.
     The majority of  the biofilm  models  have been solved for a
biofilm thickness which  is  either assumed or is a coupling
point with the process model.   Ritt matin  and McCarty (47), how-
ever, extended the solution  techniques of the previous model
to one for a steady—state biofilm,  i.e.,  one in which cell
growth is just balanced  by  decay  (#6 in  Table I).  In a
steady-state situation there is a unique film thickness asso-
ciated with each bulk substrate concentration.  When that
thickness is "deep",  the concentration of substrate reaches
zero at some interior point.   When  it  is "shallow" a finite
substrate concentration  remains at  the support:biofilm inter-
face.  Two solution techniques  were utilized to generate the
plot of flux versus bulk substrate  concentration, depending
upon whether the film was deep  or shallow.  Using the steady-
state assumption, the film  thickness,  Lf, was calculated for
an assumed flux and the  deep film technique  (61) was used to
get the bulk substrate concentration associated with that
flux.  Because of the need  for  growth  to balance decay in a
steady-state biofilm  there  will be  some  minimum bulk substrate
concentration, C^^   required  to  maintain a  steady—state film.
When that concentration  is  reached  the flux into the film will
be zero and no film will be maintained.   This means that film
thickness will vary from zero  at  tL,£n  to some maximum deter-
mined by the maximum  substrate  concentration.  Furthermore
                          386

-------
this means that the  flux  into  a  shallow,  steady-state biofilm
will vary from zero  at C  min to  the  deep  value at some higher
bulk substrate concentration.  The  fluxes associated with C"
values between C  min and  that for  the  thinnest "deep" film
were calculated in the following manner.   First, a value for
the flux was assumed and  the corresponding film thickness was
calculated from the  steady-state assumption.   This film thick-
ness was then divided into a finite—difference grid and the
steady—state concentration profile  in the biofilm was solved
for (subject to the  boundary condition  that there be no flux
into the solid surface) by an  implicit,  finite difference
technique.  To start the  routine the value of C* was taken to
be the value that gives the deep solution for the flux.  The
profile was then used to  get the average  reaction rate within
the film by numerical integration.   This  average reaction rate
was compared to the  initially  assumed flux and if they did not
agree the procedure  was repeated by  assuming  a new value for
C*.  When the two fluxes  agreed,  knowledge of the external
mass transfer characteristics  and C* allowed  computation of
the bulk concentration C" associated with the flux.  Repeti-
tion of this procedure resulted  in  a plot of  flux versus C"
which was continued  until  it intersected  the  plot for the deep
biofilm.  The plot was made in dimensionless  coordinates which
incorporated all kinetic  and mass transfer parameters except
the external liquid  film  thickness,  which was employed as a
parameter.  As in their previous model  (61),  they then
developed a simplified equation  to  facilitate use of the model
for solving various  process models.   There is a unique curve
associated with each decay rate  since it  determines the
steady-state film thickness.   It should  be recalled that the
major criticism of this model  was the assumption that decay is
the only mechanism removing the  biofilm,  but  that Rittmann
(50) has shown how removal by  shear  stress may be incorpor-
ated, thereby allowing additional interfacing with a broader
range of process models.
     Examination of  the models that  fall  into this category
reveals that both noninteractive double-substrate and single
Monod models have been employed.  As we  will  see in the next
section, single Monod models can be  handled just as well, if
not better, by the effectiveness factor  technique because it
allows more parameters to  be included and simplifies the solu-
tion technique somewhat.  Thus one  must  question whether the
indirect technique is the  best to use.   This  is particularly
true for the noninteractive double-substrate  Monod model
because the oeprating curves developed were unique for a given
                            387

-------
set of kinetics  and  mass  transfer parameters.   If those plots
could be arranged  as  dimensionless  plots  their utility would
be extended.  Whether this  can easily be  done  is not yet
apparent.

Effectiveness Factor  Technique

     The majority  of  the  models in  Table  I (#1,2,3,5,8,11,12)
have been presented with  effectiveness factor  techniques and
all are single substrate  models.  The first worker to apply
this approach to the  modeling  of  fixed-film biological reac-
tors was Atkinson  and his book (66) should be  consulted for
the details of how the effectiveness factor curves were devel-
oped.  Generally,  however,  numerical techniques are used to
solve the biofilm  model directly  and the  results are used to
determine the effectiveness factor  as a functon of various di-
mensionless groups reflecting  the kinetic and  mass transport
characteristics  of the system.  Atkinson  and his coworkers
have limited their effectiveness  factors  to transport within
the biofilms so  that  the  substrate  concentration at the bio-
film: liquid interface must  be  known or must be calculated from
knowledge of the external mass transfer resistance.  Such an
effectiveness factor  is called an internal effectiveness
factor (30).
     Atkinson and  Howell  (17)  used  the internal effectiveness
factor technique to model substrate removal in a trickling
filter with single—substrate Monod  kinetics.  A mass balance
was written over a liquid element prependicular to the reactor
axis, resulting  in a  first-order  ordinary differential equa-
tion which equates the flux to the  biofilm:liquid interface
with the flux into the biofilm.   The mass transfer coefficient
approach (Eq. 3) was  used to model  the flux to the biofilm and
the Monod equation in terms of the  interface substrate concen-
tration, C*, was multiplied by the  internal effectiveness
factor to compute  the flux into the film.  Algebraic manipula-
tion allowed the differential  equation to be rewritten in
terms of C*, thereby  giving an integral equation relating C*
to the axial position in  the reactor.  Numerical solution then
gave C* as a function of  axial position and knowledge of the
flux and the external mass  transfer coefficient at each posi-
tion allowed computation  of the bulk concentration, C .
Through the dimensionless groups  the effectiveness factor is
given as a function of both the film thickness and C* so these
dependencies had to be accounted  for during the numerical
solution.  Although  this  approach could be used directly by
other investigators  to model trickling filters under a broad
                           388

-------
range of conditions, Atkinson  and Howell  (17)  used their model
to investigate a. variety of  limiting  conditions  and to write
simplified analytical procedures for  those  cases,  thereby
facilitating computations.
     Even though Atkinson  and  Howell  (17) only used their
solution technique for a trickling  filter,  Rittmann and
McCarty (28) used it to develop  relationships  between the bulk
substrate concentration and  the  substrate removal  rate by bio-
films of any thickness.  Their results  were presented as plots
of dimensionless flux (removal rate)  as a function of dimen—
sionless bulk substrate concentration with  dimensionless film
thickness as a parameter.  By  so doing,  they used  the internal
effectiveness factor technique to develop information which
could be used in the indirect  technique with bulk  substrate
concentrations.
     Howell and Atkinson (8) also used  the  internal effective-
ness factor technique to model sloughing  in a  trickling
filter.  In this case, however,  they  assumed that  external
mass transport was not limiting  so  the  interface substrate
concentration was equal to the bulk concentration, thereby
simplifying the solution.  The filter was modeled  as a series
of completely mixed elements and a  dynamic  equation was used
in which film thickness within an element was  allowed to
increase with time.  Film  growth and  substrate removal were
modeled by the Monod equation  with  the  bulk substrate concen-
tration and the effectiveness  factor.  Integration was per-
formed over a fixed time interval,  thereby  allowing the film
thickness in each element  to increase.  At  the end of each
interval, a sloughing criterion was checked in each element
and within each one meeting  it,  the film was sloughed, leaving
a new thin film thickness.   Integration again  proceeded for-
ward in time until the next  time interval,  when the criterion
was again checked in each  element.  The  results  were used to
investigate how sloughing  introduces  variation into the per-
formance of a trickling filter.
     Grady and Lira (30,31) have  also  used  the  effectiveness
factor approach with the single-substrate Monod  model, but
unlike the previous examples they used  an overall  effective-
ness factor which accounts for both internal and external mass
transport limitations.  The  overall effectiveness  factor was
derived by Fink et al. (74)  for  immobilized enzyme catalysts
and was solved by a transformation  which  permitted the rewrit-
ing of the two-point boundary  value problem as an  initial
value problem.  In this case the effectiveness factor was
given as a function of a modified Thiele modulus (which
                            389

-------
relates  the maximum  reaction  rate  to the maximum internal
diffusion rate) with the  Sherwood  number (which relates exter-
nal transport  to  internal transport) as a parameter.  Although
the general solution was  presented in graphical form,  empiri-
cal equations were given  for  certain regions to facilitate
numerical analysis.   They then  developed models for both
trickling filters and RBC's  in  which the substrate removal
rate was expressed as a function of  the bulk substrate concen-
tration  and the overall effectiveness factor.   Since both the
Thiele modulus and the Sherwood number depend  upon the biofilm
thickness that dimension  serves as a link with the process
model.
     Jennings  et  al.  (73) numerically solved the second order
differential equation resulting from Eq. 5 for both the
single—substrate  Monod and the  single substrate Blackman
models.  The reaction rates obtained were divided by the in-
trinsic  rates  for the two rate  expressions to  develop  curves
of overall effectiveness  factors as  functions  of a number of
variables.  Their intent  was  to see  how those  variables influ-
enced the effectiveness of the  biological reactions and thus
no attempt was made  to develop  an  all—inclusive effectiveness
factor plot like  that developed by Fink et al  (74). Neverthe-
less, the results were very useful in determining the  condi-
tions likely to maximize  reaction  rates.  They were subse-
quently  used to model a submerged  filter.
     Finally La Motta and coworkers  (9-12,51)  have used the
effectiveness  factor approach extensively in their modeling of
fluidized bed  biofilm reactors.  In  all cases, however, only
an internal effectiveness factor was used, under the assump-
tion that external mass transfer resistance was not important.
Single—substrate  Blackman kinetics was employed which  enabled
the development of explicit equations representing the effec-
tiveness factor for  both  zero-order  and first-order kinetics.
These equations were then coupled  with the intrinsic reaction
rates in the process  model to allow  prediction of performance
under a  large number  of conditions.   Figure 1  illustrated this
coupling.
     Effectiveness factor techniques have a long history in
the field of heterogeneous catalysis and have  been beneficial
in the modeling of fixed-film biological reactors containing a
single limiting component.  They are particularly advantageous
where a  large  range  of parameter values are likely to  be en-
countered and  can be easily coupled  with process models
through  the biofilm  thickness and  the external mass transfer
coefficient.  Consequently, they appear to be  more broadly
applicable than the  indirect  technique for which graphs have
                          390

-------
only been given over a  restricted  range  of  parameter values.
No application of them  has  been  made  to  double-substrate
limited models, however,  and  for that situation more progress
has been made with the  direct  and  indirect  techniques.   There
is no theoretical reason  why  effectiveness  factors could not
be developed for interactive  double-substrate limited models,
although they are likely  to be complex and  may not be amenable
to two dimensional plots  like  those used for  single substrate
models.  Nevertheless,  the  general utility  of the effective-
ness factor approach to the modeling  of  complex processes is
sufficient to encourage the development  of  overall effective-
ness factors for interactive  double—substrate models.  Perhaps
the work that is underway in  the modeling of  double-substrate
limited immobilized enzmes will  provide  guidance in the way to
approach the problem (34,75).

CRITIQUE AND RECOMMENDATIONS

     Having reviewed the  characteristics of a number of bio-
film models one question  remains:  How good  are they?  This is
a difficult question to answer.   When used  in the simulation
of various fixed-film processes,  all  give results which are
qualitatively similar to  observed  performance.  Furthermore,
when the parameters are calibrated for a particular situation
(i.e. reactor type, flows,  nature  of  electron donor and accep-
tor, etc.) all do a reasonable job of tracking experimental
data.  Thus in one sense  all  of  them  are good for at least the
limited situations for  which  they  were derived*  It will be
recalled, however, that the purpose of this review was to
evaluate mechanistic models and  mechanistic models should be
capable of predicting performance outside of  our experience.
How well will the models  do that?  To answer  that we must look
again at each of the component parts  and ask how good they
are.
     First, consider transport in the liquid phase.  It is
evident that external mass  transport  limitations can and do
occur and thus any mechanistic model  of  broad utility must
include them.  If they  happen to be insignificant in a parti-
cular process application this insignificance will be reflect-
ed in the model solutions if  the model is properly construct-
ed.  It makes no difference whether external transport is
modeled with a diffusivity  and a stagnant film thickness (Eq.
I) or with a mass transfer  coerfficient  (Eq.  3) since both
lead to the same result.  What is  unknown,  however, is the
fate of the external mass transfer resistance as turbulence
                           391

-------
becomes large.  With  the  exception of the original work of
Williamson and McCarty  (19,20)  and its subsequent use by
Famularo et al. (22)  and  Mueller et al. (23), all models have
treated the biofilm:liquid  interface as if it were analogous
to the interface  between  a  flowing fluid and a solid support.
Is this an accurate picture?  Or does external resistance to
mass transfer continue  to exist even at high velocities
because of the pseudohomogeneous character of the interface?
If the latter is  true,  there  are likely to be few circum-
stances in which  the  interface  substrate concentration is
equal to the bulk fluid concentration, thereby making a basic
assumption of many of the models invalid.  This is an area
needing further study and is  perhaps one to which microprobe
technology could  be applied with beneficial results.
     Another important  link between the biofilm model and the
process model is  the  biofilm  thickness, because that thickness
is an important determinant of  the concentration profiles
which develop within  the  biofilm.   Unfortunately, it is still
unclear what controls that  thickness.  Rittmann and McCarty
(47) have presented the concept of a steady-state biofilm in
which cell growth is  just balanced by cell decay and this
appears to be a useful  concept  for biofilms growing in
environments with low substrate concentrations, such as in
aquifers receiving recharge by  treated effluents.  Such a
situation is unlikely,  however, in other environments so
Rittmann (50) has  extended  the  concept to a film in which loss
is by attrition as well as  by decay.  How then does one handle
the attrition rate?   The  work of Trulear and Characklis (6)
and Zelvar (45) have  shown  that the rate depends both upon
fluid shear stress and  the  mass of biofilm present.  To be
useful for modeling purposes  it would be better to relate the
attrition rate to  thickness rather than mass but this can only
be done directly  if the density is constant.  Evidence by
Hoehn and Ray (4), Muleahy  and  LaMotta (51) and Trulear and
Characklis (6), however,  all  suggest that the biofilm density
is influenced by  the  thickness, but both the mechanism and the
functional relationship are unclear.  Thus while it is
apparent that the  thickness of  a biofilm will be determined by
a balance between growth  and  loss  by attrition and decay, it
is not apparent how the rate  of attrition should be modeled.
More fundamental  experimental work is needed in this area.
     Almost all biofilm models  assume steady-state biofilms of
some sort.  However,  sloughing  is  a well known phenomenon
although its mechanisms are unclear.  Only Howell and Atkinson
(8) have attempted to model sloughing, but their model
                           392

-------
includes a number of simplifying assumptions, including one
which limits biofilm loss  to  sloughing alone.  Nevertheless,
their work indicates that  the irregular loss of biofilm by
sloughing can have a major  impact  upon performance.  The
magnitude of that impact,  however, will depend upon the fre-
quency with which sloughing occurs (which will depend upon the
net growth rate at the  biofilra)  and the thickness of the film
left after sloughing.   If  the remaining film thickness is
greater than the usual  active thickness then the impact of
sloughing would be small,  whereas  if it were smaller, the
impact would be larger.  Our  ability to model this phenomenon
depends upon knowledge  of  the attrition rate discussed in the
preceding paragraph and the characteristics of the remaining
film.  Very little work has been done  on the latter.  Thus it
appears that a good deal more experimental work is needed
before this important aspect  of  fixed-film reactors can be
adequately modeled.
     The variation of density with thickness was discussed
above with regard to its importance to the modeling of attri-
tion.  Such variations  are also  important because they influ-
ence the quantity of biomass  present within the biofilm.  As
seen in Eqs. 6 and 7 the rates of  cell growth and substrate
removal both depend upon the  amount of biomass present.  While
the majority of models  assume that the density is independent
of depth so the mass is directly proportional to thickness,
the evidence cited above has  shown that this is not the case.
This constitutes an important weakness in most existing
models.  The key question,  however, is whether changes in den-
sity occur within the active  film  thickness or only in the
regions beyond which no significant transport occurs.  If the
latter case exists it may  be  adequate to model the reaction
rate expressions with a constant density term.  If the former
is true, it will be necessary to use a variable density to
accurately reflect the  reaction  rates.  Again, additional
experimental work is needed to resolve this.
     The two main determinants of  the concentration profiles
within the biofilm are  the rates of transport and reaction.
Although considerable effort  has been expended on evaluations
of diffusion coefficients  within biofllms there is little con-
sensus in the  literature regarding the magnitude of the
retardent effects which may be attributed to the slime
material within the film.   This  is a major weakness of current
modeling efforts.  There appears to be two possible causes for
these variations: experimental techniques and variations in
film microbial composition.  As  discussed in detail earlier
                            393

-------
almost every investigator  has  a unique way of measuring rates
of diffusion within films.  Many of  these require formation of
an artificial film and  it  would appear that the exact condi-
tions existing during formation of a film would determine its
diffusive characteristics.  Thus it  is not surprising that
diffusivities measured  in  films formed by filtration (20)
differ from these measured in  films  formed by spreading
(55,57).  Futheremore,  there is evidence that diffusivities in
laboratory films are higher than those in field films (3,56).
It appears that the more direct the  technique for measuring
the diffusivity and the fewer  the assumptions involved in its
computation, the more likely the values are to be correct.
This suggests that microprobe  techniques offer the best poten-
tial for determination  of  how  various physical factors affect
internal diffusivities.  Certainly more work is needed in this
area.
     From the review of the results  obtained with the various
reaction rate models in Table  I there can be no doubt about
the fact that the transport and utilization of both the elec-
tron donor and the electron acceptor are important to the per-
formance of a fixed-film reactor. This suggests that unless
evidence to the contrary is overwhelming, double-substrate
limited models should be employed.   However, as seen in Table
I, two-thirds of the listed models are single substrate
models.  Thus, unless care is  taken  to ensure that they are
only applied in circumstances  where  only one component is
limiting throughout, these  models are likely to give predicted
performance which is not in eonforrnance with reality.  With
regard to the dual-substrate limited models the literature is
divided as to whether they  should be interactive or noninter-
active.  Furthermore, as pointed out by Bader (67), there is
not yet sufficient evidence to allow conclusive determination
of which is of the more general utility.  Nevertheless, con-
sideration of the circumstances under which each type of model
is likely to be valid (67)  and evaluation of the data of Ryder
and Sinclair (76) suggests  that an interactive model is more
likely to be correct for situations  in which electron donor
and electron acceptor are  the  two limiting components.  When
this is coupled with the fact  that a noninteractive model pro-
duces discontinuities in the solution (i.e., regions of limi-
tation must be identified  a priori), it would appear that an
interactive model should be employed unless there is conclu-
sive evidence that the  noninteractive model is mechanistically
more accurate.  As far  as  the  form that the model should take
(Monod or Blaekman) there  is no conclusive evidence in either
direction.  The arguments  for  each are the same in this con-
                           394

-------
text as they have been  for  cell  growth in general because the
rate equations should reflect  intrinsic kinetics.  The main
argument for one over another  in that  context has been one of
mathematical convenience  (30).   If  that same argument is applied
here, one would favor Monod over Blackman kinetics because it is
a continuous function which avoids  discontinuties.  Certainly
more work on intrinsic  kinetics  under  double—substrate
limitation is needed to resolve  this issue.
     Finally, as far as solution techniques  are concerned a
number of numerical procedures have been employed in direct
solutions and to develop  the graphs or effectiveness factor
charts for the other techniques.  Direct solutions offer per-
haps the most straight  forward approach to modeling of a fixed-
film reactor.  They have  the drawback,  however, of being
complicated and therefore of being  unlikely  to be used by
anyone other than the developer.  Thus  for wide—scale study of
fixed-film reactors it  would appear that either the indirect or
effectiveness factor approaches  offer  the most utility.  Of
those two, the effectiveness factor approach appears to be more
useful because its dimensionless  groupings allow more para-
meters to be considered simultaneously.  Furthermore, since the
kinetic and mass transfer coefficients  and the film thickness
are incorporated into the solution  in  a way  which allows them
to serve as links with  the  process  model, an effectiveness
factor solution to the  biofilm model can be  developed while the
questions regarding these items  are being resolved.  Thus it
appears to this author  that the  next step in the development of
mechanistic biofilm models  of  broad utility  in process modeling
should be the development of effectiveness factor relationships
for interactive double  substrate  limiting kinetics.  Since the
solutions to the complex  two-point  boundary  value"problems need
be made only once, they can be made with few simplifying
assumptions, even if the  required numerical  solution are not
very efficient.  Once complete effectiveness factor solutions
are available, however, then extensive  sensitivity analyses can
be run, resulting ultimately in  simplified effectiveness factor
charts which have little  likelihood of  being incorrect because
of unwarranted simplifications.
     In conclusion, it  is clear  that we have not yet achieved
a complete and general mechanistic  model for biofilms which
can be used to simulate the performance of a broad range of
fixed-film processes.   It should  be recalled,  however,  that a
major goal of mechanistic modeling  is  to increase understanding.
                           395

-------
The fact that the large number of events  occurring within  bio-
films is now widely recognized is evidence  for  the attainment
of that goal.  Compared to the situation  which  existed  twenty
years ago a great deal of knowledge has been  obtained and  a
great deal of progress has been made.  Today, we  have a good
idea of what we don't know and therefore  we can design  the
experimental programs required to gain that knowledge.   With
the renewed interest in fixed-film processes  evident today even
greater energies can be brought to bear upon  the  problem and
the remaining gaps in knowledge can be filled.

ACKNOWLEDGEMENT

     The author would like to thank the large number of people
who allowed him to read manuscripts which have  not yet  appeared
in print.

REFERENCES

1.   Shieh, W. K. and Mulcahy, L. T., Discussion  of Fluidized
     Bed for Biological Wastewater Treatment, Journal of the
     Environmental Engineering Division,  ASCE, Vol. 107,
     293-295, 1981.

2.   Atkinson, B. and Fowler, H. W., The  Significance of
     Microbial Film in Fermenters, Advances in Biochemical
     Engineering, Vol. 3, 221-277, 1974.

3.   Chen, Y. S. and Bungay, H. R., Microelectrode Studies of
     Oxygen Transfer in Trickling Filter  Slimes,  Biotechnology
     and Bioengineering,. Vol. 23, 781-792,  1981.

4.   Hoehn, R. C. and Ray, A. U., Effects of  Thickness  on
     Bacterial Film, Journal, Water Pollution Control
     Federation, Vol. 45, 2302-2320, 1973.

5.   LaMotta, E. J., External Mass Transfer in a  Biological
     Film Reactor, Biotechnology and Bioengineering, Vol.  18,
     1359-1370, 1976.

6.   Trulear, M. G. and Characklis, W. G.,  Dynamics of  Biofilm
     Processes, Presented at the 53rd Annual  Conference of the
     Water Pollution Control Federation,  Las  Vegas, Nev.,  Oct.
     1980 and accepted for publication in the Journal.
                           396

-------
7.   Castaldi, F. J. and Malina, J. F., Jr., Velocity-dependent
     Reaction Rates in a Slime Reactor, Journal, Water
     Pollution Control Federation, Vol. 54, 261-269, 1982.

8.   Howell, J. A. and Atkinson, B., Sloughing of Microbial
     Film in Trickling Filters, Water Research, Vol. 10,
     307-315, 1976.

9.   Shieh, W. K., A Suggested Kinetic Model for the Fluidized
     Bed Biofilm Reactor (FBBR), Biotechnology and
     Bioengineering. Vol. 22, 667-676, 1980.

10.  Mulcahy, L. T., Shieh, W. K., and LaMotta, E. J., Kinetic
     Model of Biological Denitrification in a Fluidized Bed
     Biofilm Reactor (FBBR), Progress in Water Technology, Vol.
     12 (Toronto), 143-157, 1980.

11.  Mulcahy, L. T., Shieh, W. K., and LaMotta, E. J.,
     Simplified Mathematical Models for Fluidized-Bed Biofilm
     Reactor, Water - 1980,AIChE Symposium Series, Vol. 77,
     No. 209, 273-285, 1981.

12.  Shieh, W. K., Mulcahy, L. T., and LaMotta, E. J., Unified
     Design Basis for the Fluidized Bed Biofilm Reactor, Paper
     presented at the Second World Congress of Chemical
     Engineering and the Ninth Congress of the Inter-American
     Confederation of Chemical Engineering, Montreal, Canada,
     Oct. 1981.

13.  Janson, J. la C. and Kristensen, G. H., Fixed Film
     Kinetics - Denitrification in Fixed Films, Report No.
     80-59, Uept. of Sanitary Engineering, Technical University
     of Denmark, 1980.

14.  Andrews, G. F. and Tien, C., Bacterial Film Growth in
     Adsorbent Surfaces, AIChEJournal, Vol. 27, 396-403,
     1981.

15.  Wang, S.-C, P., The Interaction between Adsorption and
     Microbial Growth in Biological Activated Carbon (BAC)
     Processes, Ph.D. Thesis, Syracuse University, Syracuse,
     N.Y., 1981.

-------
16.  El Amin, M. A., A Design Procedure for the Packed Bed
     Biological Reactor, Ph.D. Thesis, University of Wales, U.
     K., 1974.

17.  Atkinson, B. and Howell, J. A., Slime Holdup, Influent
     BOD, and Mass Transfer in Trickling Filters, Journal of
     the Environmental Engineering Division, ASCE, Vol. 101,
     585-605, 1975.

18.  Atkinson, B., Busch, A. W., Swilley, E. L. and Williams,
     D. A., Kinetic, Mass Transfer and Organism Growth in a
     Biological Film Reactor, Transactions, Institution of
     Chemical Engineers, Vol. 45, 257-264, 1967.

19.  Williamson, K. and McCarty, P. L., A Model of Substrate
     Utilization by Bacterial Films, Journal, Water Pollution
     Control Federation, Vol. 48, 9-24, 1976.

20.  Williamson, K. and McCarty, P. L., Verification Studies of
     the Biofilm Model for Bacterial Substrate Utilization,
     Journal, Water Pollution Control Federation, Vol. 48,
     281-296, 1976.

21.  Welty, J. R., Wicks, C. E. and Wilson, E. E., Fundamentals
     of Momentum, Heat and Mass Transfer, John Wiley & Sons,
     Inc., New York, N.Y., 1969.

22.  Famularo, J., Mueller, J, A, and Mulligan, T., Application
     of Mass Transfer to Rotating Biological Contactors,
     Journal, Water PollutionControl Federation, Vol. 50,
     653-671, 1978.

23.  Mueller, J. A., Paquin, P. and Famularo, J., Nitrification
     in Rotating Biological Contactors, Journal, Water
     Pollution Control Federation, Vol. 52, 688-710, 1980.

24.  Levieh, V.  G., Phys iochemica1 Hydrodynamics, Prentice
     Hall, New York, N.Y., 1968.

25.  Meunier, A. D. and Williamson, K. J., Packed Bed Biofilm
     Reactors: Simplified Model, Journal of the Environmental
     Engineering Division, ASCE, Vol. 107, 307-317, 1981.
                          398

-------
26.  Snowdon, C, B. and Turner, J. C. R., Mass Transfer in
     Liquid Fluidized Bed of Ion Exchange Resin Beads,
     Proceedings, International Symposiumon Fluidization,
     599-608, 1967.

27.  Rittmann, B. E. and McCarty, P. L., Evaluation of
     Steady State Biofilm Kinetics, Biotechnology and
     Bioengineering, Vol. 22, 2359-2373, 1980.

28.  Rittmann, B. E. and McCarty, P. L., Substrate Flux into
     Biofilms of any Thickness, Journal of the Environmental
     Engineering Division, ASCE, Vol. 107, 831-849, 1981.

29.  Jennings, P. A,, A Mathematical Model for Biological
     Activity in Expanded Bed Adsorption Columns, Ph.D. Thesis,
     University of Illinois, Urbana, IL, 1975.

30.  Grady, C. P. L., Jr., and Lim, H. C., Biological
     Was t e wa ter Tre at ment: Theory and Applications, Chapter 14,
     Marcel Dekker, Inc., New York, 1980.

31.  Grady, C. P. L., Jr. and Lim, H. C., A Conceptual Model of
     RBC Performance, Proceedings: First National Symposium/
     Workshop on Rotating BiologicalContactor Technology, Vol.
     II, E. D. Smith, R. D. Miller and Y. C. Wu (Eds.),
     829-859, 1980.

32.  Mixon, F. 0. and Carberry, J. J., Diffusion within a
     Developing Boundary Layer - A Mathematical Solution for
     Arbitrary Velocity Distribution, Chemical Engineering
     Science, Vol. 13, 30-33, 1970.

33.  Wilson, E. J. and.Geankoplis, C. J., Liquid Mass Transfer
     at Very Low Reynolds Numbers in Packed Beds, Industrial
     and Engineering Chemistry Fundamentals, Vol. 5, 9—14,
     1966.

34.  Dahodwala, S. K., Humphrey, A. E. and Weibel, M. K., Pore
     Diffusion Model for a Two-Substrate Enzymatic Reaction:
     Application to Galactose Oxidase Immobilized on Porous
     Glass Particles, Biotechnology and Bioengineering, Vol.
     18, 987-1000, 1976.
                          399

-------
35.  Brian, P. L. T. and Hales, H. B., Effects of Transpiration
     and Changing Diameter on Heat and Mass Transfer to
     Spheres, AIChE Journal, Vol. 15, 419-425, 1969.

36.  Mueller, J. A., Paquin, P. and Faraularo, J., Mass Transfer
     Impact on RBC and Trickling Filter Design, Water-1979,
     AIChE Symposium Series, Vol. 76, No. 197, 250-270, 1980.

37.  Charpentier, J. D., Recent Progress in Two Phase
     Gas Liquid Mass Transfer in Packed Beds, Chemical
     Engineering Journal, Netherlands, Vol. 11, 161-181, 1976.

38.  Characklis, W. G., Fouling Biofilm Development: A Process
     Analysis, Biotechnology and Bioengineering, Vol. 23,
     1923-1960, 1981.

39.  Kornegay, B. H. and Andrews, J. F., Kinetics of Fixed-Film
     Biological Reactors, Journal, Water Pollution Control
     Federation, Vol. 40, R460-R468, 1968.

40.  Tomlinson, T. G. and Snaddon, D. H. M., Biological
     Oxidation of Sewage by Films of Microorganisms,
     International Journal Air and Hater Pollution, Vol. 10,
     865-881, 1966.

41.  Bungay, H. R., Whalen, W. J. and Sanders, W. M.,
     Microprobe Techniques for Determining Diffusivities and
     Respiration Rates in Microbial Slime Systems,
     Biotechnology and Bioengineering, Vol. 11, 765-772, 1969.

42.  Button, P. M., Langley, D. F., Warner, K. E. and Park, K.,
     Oxitron System™ Pluidized Bed Wastewater Treatment
     Process: Application to High Strength Industrial
     Wastewater, Proceedings, 34th Industrial Waste Conference,
     1979, Purdue University, Ann Arbor Science Publishers,
     Inc., Ann Arbor, Mich., 450-461, 1980.                    :

43.  Tsezos, M. and Benedek, A., A Method for the Calculation
     of Biological Film Volume in a Fluidized Bed Biological
     Reactor, Water Research, Vol. 14, 689-693, 1980.

44.  Shieh, W. K., Sutton, P. M. and Kos, P., Predicting
     Reactor Biomass Concentration in a Fluidized Bed System,
     Journal, Water Pollution Control Federation, Vol. 53,
     1574-1584, 1981.
                          400

-------
45.  Zelver, N., Biofilm Development and Associated Energy
     Losses in Water Conduits, M. S. Thesis, Rice University,
     Houston, TX., 1979.

46.  Shieh, W. K., Comments on Biologically Active Fluidized
     Beds, Biotechnology and Bioengineering, Vol. 23, 2145-
     2147, 1981.

47.  Rittmann, B. E. and McCarty, P. L., Model of Steady-State-
     Biofilm Kinetics, Biotechnology and Bioengineering, Vol.
     22, 2343-2357, 1980.

48.  Rittmann, B. E. and McCarty, P. L., Design of Fixed-Film
     Processes with Steady-State Biofilm Model, Progress in
     Water Technology, Vol. 12, (Toronto), 271-281, 1980.

49.  Arcuri, E. J. and Donaldson, T. L., Biofilm Kinetics,
     Biotechnology and Bioengineering, Vol. 23, 2149, 1981.

50.  Rittmann, B. E., The Effect of Shear Stress On Biofilm
     Loss Rate, Biotechnology and Bioengineering, (accepted).

51.  Mulcahy, L. T. and LaMotta, E. J., Mathematical Model of
     the Fluidized Bed Biofilm Reactor, Report No. Env. E.
     59-78-2 of the Environmental Engineering Program, Dept. of
     Civil Engineering, University of Massachusetts, Amherst,
     1978.

52,  Alleman, J. E. and Veil, J. A., Observations of Attached
     Biofilm Composition and Structure Using Scanning Electron
     Microscopy, Paper presented at the 1981 Annual Meeting of
     the American Institute of Chemical Engineers, New Orleans,
     LA, Nov. 1981.

53.  Alleman, J. E., Veil, J. A. and Canaday, J. T., Scanning
     Electron Microscope Evaluation of Rotating Biological
     Contactor Biofilm, Water Research, Vol. 16 (in press),
     1982.

54.  Bryers, J. D., Swiss Federal Institute for Water Resources
     and Water Pollution Control, Dubendorf, Switzerland,
     Personal Communication, 1982.
                          401

-------
55.  Matson, J. V. and Characklis, W. G., Diffusion into
     Microbial Aggregates, WaterResearch, Vol. 10, 877-885,
     1976.

56.  Bungay, H. R, and Harold, D. M., Simulation of Oxygen
     Transfer in Microbial Slimes, Biotechnology and
     Bioengineering, Vol. 13, 569-579, 1971.

57.  Pipes, D. M., Variations in Glucose Diffusion Coefficients
     through Biological Floe, M. S. Thesis, Rice University,
     Houston, TX., 1974.

58.  Mulcahy, L. T., Shieh, W. K., and LaMotta, E. J.,
     Experimental Determination of Intrinsic Denitrification
     Kinetic Constants, Biotechnology andBioengineering, Vol.
     23, 2403-2406, 1981.

59.  Williamson, K. J. and Chung, T. H., Dual Limitation of
     Substrate Utilization Kinetics within Bacterial Films,
     Paper presented at the 49th National Meeting of the
     American Institute of Chemical Engineers, Houston, Texas,
     March 19, 1975.

60.  Meunier, A. D. and Willtams-on,--K-*-J.t,4 Packed Bed Biofllm
     Reactors: Design, Journal of the EnvironmentalEngineering
     Division, ASCE, Vol. 107, 319-337, 1981.

61.  Rittmann, B. E. and McCarty, P. L., Variable-Order Model
     of Bacterial—Film Kinetics, Journal of theEnvironmental
     Engineering Division, ASCE, Vol. 104, 889-900, 1978.

62.  Whalen, W. J., Bungay, H. R. and Sanders, W. M.,
     Microelectrode Determination of Oxygen Profiles in
     Microbial Slime Systems, Environmental Scienceand
     Technology, Vol.  3, 1297-1298, 1969,

63.  Harris, N. P. and Hansford, G. S., A Study of Substrate
     Removal in a Microbial Film Reactor, Water Research, Vol.
     10, 935-943, 1976.

64.  Atkinson, B. and Daoud, I, S., Diffusion Effects with
     Microbial Films, Transactions, Institution of Chemical
     Engineers, Vol. 48, 245-254, 1970.
                          402

-------
65.  Atkinson, B. and How, S. Y.,  The Overall Rate of Substrate
     Uptake (Reaction) by Microbial Films - 2. Effect of
     Concentration and Thickness with Mixed Microbial Films,
     Transactions, Institution of  Chemical Engineers, Vol. 52,
     260-268, 1974.

66.  Atkinson, B., Biochemical Reactors, Pion Limited, London,
     1974.

67.  Bader, F. G., Analysis of Double—Substrate Limited Growth,
     Biotechnology and Bioengineering, Vol. 20, 183-202, 1978.

68.  Monod, J., The Growth of Bacterial Cultures, Annual Review
     of Microbiology, Vol. 3, 371-394, 1949.

69.  Blackman, F. F., Optima and Limiting Factors, Annals of
     Botany, Vol. 19, 281-295, 1905.

70.  Howell, J. A. and Atkinson, B., Influence of Oxygen and
     Substrate Concentrations on the Ideal Film Thickness and
     the Maximum Overall Substrate Uptake Rate in Microbial
     Film Ferraenters, Biotechnology and Bioengineering, Vol.
     18, 15-35, 1976.

71.  Bright, H. and Appleby, M., The pH Dependence of the
     Individual Steps in the Glucose Oxidase Reaction, Journal
     of Biological Chemistry, Vol. 244, 3t>25-3634, 1969.

72.  Atkinson, B. and Abdel Rahman Ali, M. E., The
     Effectiveness of Biomass Hold-up and Packing Surface in
     Trickling Filters, Water Research, Vol. 12, 147-156,
     1978.

73.  Jennings, P. A., Snoeyink, V. L. and Chian, E. S. K.,
     Theoretical Model for a Submerged Biological Filter,
     Biotechnology and Bioengineering, Vol. 18, 1249-1273,
     1976.

74.  Fink,  D. J., Na, T.-Y. and Schultz, J. S., Effectiveness
     Factor Calculations for Immobilized Enzyme Catalysts,
     Biotechnology and Bioengineering, Vol. 15, 879-888, 1973.

75.  Engasser, J.-M. and Hisland,  P., Diffusional Effects on
     the Heterogeneous Kinetics of Two-substrate Enzyme
     Reactions, Journal of Theoretical Biology, Vol. 77,
     427-440, 1979~.

-------
76.  Ryder, D. N. and Sinclair, C. G., Model for the Growth of
     Aerobic Microorganisms under Oxygen Limiting Conditions,
     Biotechnology and Bipengineering, Vol. 14, 787-798, 1972.
                           404

-------
               INVESTIGATION OF SOME PARAMETERS
                      IN RBC MODELING
     Khalil Z. Atasj., Department of  Civil  Engineering
     University of Michigan

     Jack A. Borchardt,  Department of  Civil  Engineering
     University of Michigan
INTRODUCTION
     Although the RBC system has been studied and used on
many full scale treatment plants, a great deal of research is
still needed in order to better define its optimum design
and operational characteristics.  Implementation of this con-
cept requires a more thorough knowledge of the kinetics of
substrate utilization by a fixed-film in the form of a rotat-
ing disc system.  Investigation and application of RBC kin-
etics has suffered from the inherent complications involved
in simultaneous processes such as liquid film mass transfer,
diffusion and reaction within the biofilm.
OBJECTIVE AND SCOPE
     It is the purpose of this research to further study
the kinetics of the RBC process for carbonaceous substrate
removal using a synthetic sewage.   The interest is to concen-
trate on the mechanism of the reaction and its rate order,
both observed and intrinsic.
                              405

-------
     This paper details part of the experimental work of a
long range project, the final results of which will be pub-
lished at a later date.  The overall objective is the develop-
ment of a simplified, practical approach to design.  In
essence the extremely complicated kinetic expressions will
be approached in three steps.  Step one deals with inter-
phase diffusion and surface reaction occuring in series.  In
this step, the problem is dealt with using an external effec-
tiveness factor and a modified Damkohler number to describe
the effect of the mass transfer phenomenon.  This step, even-
tually will be related to the hydrodynamic characteristics
of any specific support system.  Step two deals with intra-
phase diffusion and those reactions which occur simultaneously
within the film.  The intraphase problem will be attacked by
using an internal effectiveness factor and a modified Thiele
modulus.  This step will relate the effect of internal
diffusion and the biochemical reaction taking place together
inside the film.  Finally, step three will relate the external
mass transfer, the internal diffusion, and the substrate oxi-
dation through the use of an overall effectiveness factor.
The ultimate goal is, then, to use this technique for modeling
the RBC process through the use of simple equations in terms
of observable quantities.  This paper will deal only with
step one and the parameters involved with the definition of
the kinetics external to the fixed film.
REVIEW OF PREVIOUS RELATED RESEARCH
     Because more than one phase is involved, the RBC system
is a heterogeneous system where mass transfer, molecular
diffusion, and biological oxidation take place at the same
time in parallel and/or in series.  Accordingly, it is impor-
tant to consider the above phenomena when studying the
different factors that might affect the substrate utilization
rate by the biological film that grows on this rotating sur-
face.  In a simple way, the phenomena that take place when
a biological film is brought into contact with a liquid
containing soluble substrate are as follows:
     1.  Transport of the soluble substrate from the bulk
liquid to the surface of the biofilm (liquid-biofilm inter-
face) ;
     2.  Internal transport of the soluble substrate through
the biological film by diffusional processes;
     3.  Biological oxidation of the soluble substrate by
the biomass in the biofilm;
                            406

-------
     4.  Diffusion of part of the reaction products to the
bulk of the liquid.
     Any kinetic information gathered on the substrate removal
mechanism under the effect of mass transfer and diffusion
will neither give the true or intrinsic kinetics nor the
true mechanism as these are only a part of the above mentioned
effects in some combination.
     Most of the research reported in the literature related
to biofilm kinetics has been carried out on purely laboratory
experimental equipment (4, 19, 21, 23s 24, 26, 31, 35).  As
a result these observations may not have a direct practical
application as they can only with difficulty be transferred
to prototype equipment.
     Little data are available on the intrinsic and overall
rate of substrate utilization within stages of an RBC process.
Because of difficulties encountered when dealing with these
problems in the RBC, most of the investigations were run
using the previously mentioned experimental set up.  A zero
order intrinsic rate was assumed (6,23).  Some assumed Monod
kinetics using that concentration observed in the bulk
liquid (21) while others used first order reaction without
taking into account the mass transport phenomenon (3).
Harremoes (15), in modeling the biofilm as a porous diffusion
model, reached an interesting conclusion, namely that a first
order heterogeneous reaction in a pore will lead to a first
order reaction in terms of the bulk concentration.
     In dealing with the biofilm growing in the RBC process,
Kornegay (22) assumed a homogeneous system with Monod kinetics
in terms of the bulk substrate concentration, with the same
biokinetic constants for all the stages.  The same assumption
was used by others (12, 27, 28) but with the added conditions
that the biokinetic constants change stagewise.  Indeed, the
authors of this paper do agree with the last point above, but
disagree with an observed rate based on Monod kinetics
written in terms of the measurable bulk substrate concentra-
tion unless the mass transport phenomenon is taken into
account.  At a later point, it will be shown that when a
mass transfer phenomenon affects the substrate removal, the
overall rate will no longer follow Monod kinetics.  More
specifically, it will be demonstrated that the observed rate
is first order only when the intrinsic rate exhibits a first
order (pseudo) mechanism.
                            407

-------
Antonie(2), and Stover and Kincannon (30), concluded that
the RBC process follows first order kinetics in terms of the
substrate bulk concentration dealing with an RBC pilot unit
of more than one stage collectively.  Using the same reason-
ing, Harremoes (16) fitted the data presented by Popel (29)
for a seven stage RBC pilot unit into an observed (bulk)
fractional order (half order) and hence suggested that the1
intrinsic rate for BOD consumption is zero order.  It is
important to realize that an RBC plant with several stages
in series behaves as a plug flow reactor although each stage
is a complete mix reactor.  Because of a varying biomass
(flora) stagewise (32) and widely different bulk substrate
concentrations, the use of an overall complete mix technique
aiid the inference of a single kinetic expression for all
stages collectively, is doubtfull or only very approximate.
As a result, this research implies that a kinetic study
should be carried out on each stage separately.
     With respect to the bulk dissolved oxygen within an
RBC reactor, several investigators have stressed the impor-
tance of keeping a minimum bulk D.O. (2 mg/1) to retain
the process efficiency (9, 11, 18, 34). On the other hand,
Hartmann(17) suggested that the bulk D.O may not affect the
efficiency at all.  It has been shown experimentally (7,33)
that the RBC process efficiency can be improved by sealing
the reactors and enriching them with pure oxygen.  Because
of the above conflicting findings, research seems to be
warranted to further elucidate this point.
THEORETICAL BACKGROUND
     The following will be a development of a concept of
the observed rate of substrate utilization by a biological
film.  Some assumptions are made; these are:
     1.  that bacteria are uniformly distributed within
a biofilm
     2.  that the biofilm thickness is uniform
     3.  that the bio-film is at steady-state.  That is,
the density of the biofilm does not change within any
experimental run.
     4.  that the suspended solids in the bulk fluid can
be neglected (they are too low in concentration to have
any marked effect).
     5.  that the mass transport phenomenon can be handled
by assuming a hypothetical liquid film.  In any case the
mass transfer coefficient k  includes the convective and
                            408

-------
 diffusive mass transfer effects.
      6.   that the biofilm is considered to be an "equi-
 accessible" surface;  this has also been called the "quasi-
 stationary" method as developed by Frank Kamenetskii (13)
      7.   that there is a single substance limiting growth;
i.e. the  main substrate providing the carbon
      8.   that the substrate removed is assumed to be consumed
at the surface of the  biofilm.
      9.   that the biofilm area is the same as the disc area.

      Under the steady state, the assumption is made that the
 substrate cannot build up or accumulate at the surface of
 the film.  As a consequence, the rate of substrate supplied
 by the • mass transfer, phenomenon must equal the rate of sub-
 strate utilization in the reaction at the interface.
 Assuming that Monod kinetics prevails . at the surface, and
 denoting  R  as the overall or surface reaction rate, it can
 be stated that:
                R-
                                      sat
      (all terms are defined at the end of this paper)
 Because  S   is unknown and can't  be measured, it is  more
           C3
 convenient to express the rate expressions in terms of
 observable quantities.
      Solving equation (1) above for  S  :
                                       S


   Vf<[ V
                       m                   m
 where  k   = k  x  is defined as the maximum surface reaction
 rate in analogy with Michaelis-Menten enzyme kinetics.
      Substituting the value  Sg  as given by equation (2)
 into either side of equation (1) ,  it can be shown that:
  ,,  ,
  R = k
      max
sat b k
m
k
*V n f\
1 L VJb L^satJ k J T ^^sat^b1
m

5
_ / ON
v j;
C
          (K   +s
            sat   b   k         b  ' sat    k        sat b
                       m                    m
                           409

-------
     From equation 3, it  appears  that Monod  kinetics can
become a pseudo first or  zero  order  depending on the relative
value of Ksa*. as compared to the  substrate concentration.
If Ksa-(- is much, larger  than Ss, then equation (2)  becomes:
              k
      S =  	••- • "2"n? " ' '/i—~~   S,  (f°r  pseudo-first order)
       g      1C  HT  ( iC    / iC  .  D
               ID     max  sat)                           (4)

Under this condition, and according  to  the value of 1% two
regimes (13) might exist:

     1.  A kinetic regime, if  lcm  >>(kmax/ksat) ,  where
         the following  prevails:

                Ss ~ Sb                                  (4a)
or

     2.  A mass transfer  regime,  if  km«(krnax/ksat) , where
         the prevalent  condition  is:

                Ss« Sb                                  (4b)

as a result, and for maximum efficiency,  an  RBC  system should
be operated under case  (1) above.  Since  km  is related,
among other things,  to  the hydrodynatnic characteristics  of
the system, the appropriate kinetic  regime could possibly
be attained depending on  the design  of  the system.
     For pseudo  first  order,  equation  (3) becomes:

           R  =  kosb                                    (5)

where,
     l/k0 - I/km + l/(kmax/ksat)                         (6)

where k  is the observed  first  order reaction rate.

     So, in the presence  of mass  transfer resistance, and
based on the above,  the  following can be  concluded:
                           410

-------
1.  The rate expression, in terms of the observable bulk
    concentration, will not exhibit a Monod type mechanism,
    even when the intrinsic rate is assumed so  (equation  3) .
2.  The observable rate will exhibit a first order mechanism,
    only when the intrinsic rate exhibits a pseudo-first
    order rate in S.  In this case, it is additive
    (equation (5) , (6) ) .  This finding agrees with Harremoes
    (16).
3.  When the intrinsic rate is zero order, then R is no
    longer influenced by the mass transfer phenomenon.

     The above can be simplified by using a dimensionless
concept (5, 10, 20).  For this purpose, the following dimen-
sionless numbers are defined:

                      Ksat

                s =
                    !<_,„      Maximum reaction rate
                      <*5C  55   - ' • "• ..... • •" "W"-* ...... "••nil* n • • - '»"• ....... «, ....... m - II. ill II • ...... i
              Da = — _ - _ —    Maximum mass— transfer rate
where Da stands for the Damkohler number.  The magnitude
of the Damkohler number indicates the significance of the
mass resistance.  Thus:

    If  Da >1     :    a mass— transfer regime prevails
    and if Da <1  :    the reaction is rate limited
By substituting the above dimensionless numbers in the pre-
vious equations, the following results can be obtained:

For Monod kinetics:
        s   =  f   (f  ( 1 + 4Wa2)°*5 -1  )            (7)

     where  a =   Da  + ¥ - 1

For pseudo 1st order kinetics:
     s   =      y                                      (8)
              V + Da
                          411

-------
Finally, defining an external effectiveness factor r)e as the
ratio of the reaction rate in the presence of mass transfer
to the rate which would be obtained with no mass transfer
resistance, that is when  Ss = S^, it can be shown:

For Monod kinetics:
               f + 1
               ¥ + s                                    (9)

    where s is given by equation  (7)

For pseudo 1st order kinetics:

               f
   ne       W + Da                                     (10)

In its application the external effectiveness factor acts
like a correction factor.  It provides for a decrease  in the
reaction rate due to the presence of mass transfer resistance.
Accordingly, the reaction rate can be written as follows:
    For Monod kinetics:    k
                            max
                                    b
                                 Ksat + sb
    For pseudo 1st order:     max     S     n          (12)
                             K         be
                              sat

where n  is given by the appropriate equation.
This approach is much simpler than equation (3) and (5) and
can prove to be advantageous if n  can be represented  grap-
hically in terms of the relevant parameters such as Da and
W.  But, before doing so, the Da number should be clarified.
As can be seen, the Da number is not an observable quantity,
Therefore, a new term must be introduced, Ua, which can be
designated as an observable Damkohler number, such  that:


       Da  =  n  Da                                    (13)
                           412

-------
by algebraic conversion, it can be shown that:

For Monod kinetics:   Da  = 7-—  (1 + Y)              (14)
                            ^m^b
                                ._ u .    -n
For pseudo 1st order kinetics:  Da = v— - - Y         (15)
                                     tC o,
                                      m b

It becomes advantageous to relate analytically r\  to Da.
This has been done and it can be shown that:

For Monod kinetics:  n   =  (     ~
For pseudo-lst order kinetics  n  = 1 -  Da           (17)
                                e
Figure (1) shows a computer plot of n  as a function of Da
and values of  Y as a parameter, for pseudo 1st order reac-
tion kinetics.
     It should be remembered that  V > 1 as equation (17)
holds only when the intrinsic rate is pseudo-first order in
the substrate concentration.  It is very clear from figure
(4) that the kinetic regime becomes well defined at a value
of n  = 1.  At this point the relationship becomes a hori-
zontal line for values of  Da  <1. Contrariwise, the verti-
cal lines resulting at  Da >1  and n «1 represent the mass
transfer regime.  Accordingly, an intermediate region exists
between the above two regimes, where both mass transfer
and kinetics affect the process.  This region is evidenced
by a drastic transition in the slope_of the curves.  It can
be seen also, that any increase in Da above a certain criti-
cal value, would not have any impact on the results, since
the overall reaction is mass transfer limited.  By knowing
Da  and  Y, one can find T\  so that the rate expression can
be expressed by equation (12) in terms of the bulk sub-
strate concentration S, , which is an observable quantity.

EXPERIMENTAL EQUIPMENT, MATERIALS, AND METHODS:
     The experiments detailed in this paper were run on the
first stage of a pilot plant consisting of six stages of
RBC.  Each stage consisted of four discs 2 ft. diameter
fabricated of ultra thin sheets of polyethylene with a
sinusoidal surface configuration that generated a great
deal of turbulence.  This deformation increased the area
                             413

-------
  Kinetic regime
tetO-2
            4567891     2   3 4 5 6 783l(
                    IxttT1
                   OBSERVED DflMKOHLER. OR
Figure 1 •  Bioflira external effectiveness as a  function of  the
           observable Damlcohler number and \hfor  first order
           reaction•                       '

-------
 of the disc by an average of 50% over a. flat disc turbulence.
 The sheets were heat formed spot welded, and provided by
 the FMC corporation.  The rectangular tanks made of plexi-
 glass, were 5 1/2" wide, 11" deep, and 28" long for each "
 stage.  A one inch diameter hole was drilled in the parti-
 tion wall between stages for the flow of sewage.  A concrete ".
 fillet with a triangular cross section of 9" x 9" coated with
 parafin wax, was slipped into each stage to avoid any possi-
 ble shortcircuiting.  The discs were mounted on a stainless
 steel shaft, 3/4" diameter, equipped with a sprocket and
 chain drive which was driven by an AC motor and speed con-
 troller to provide different rotational speeds.  The discs
 were approximately 35% submerged.  The liquid volume was
 about 18. 5£.  The surface area. provided by one stage (one
 module) was about 38 ft^.  Fig (2) shows the KBC reactor and
 Fig. (3) shows a detail of the rotating media.  This con-
 figuration provided a ratio of growth surface area to liquid
 volume, a, equal to 1.92/cm.
      The pilot plant was fed with a synthetic sewage (the
 formulation is shown in) Table (1) .
           Tabl_e_ _! « _Compo_s_i_t:ion j3f_ _the ^sy^nthe tic sewage.

 to 1 liter of tap water* add
Dextrin
NH/C1
T1
MgSO.THoO
Beef Consomme**
184.94 mg
76.43 mg
15.85 mg
8.2 mg
1.05-2.1 ml
 ;1
-------
CTl
                Side view
                               Front view
             Figure 2 «
                               Motor and speed
                               controller unit
An overall view of a biological reactor disc unit
with the FMC media.

-------
Figure 3 .  FMC media detail  (high density  poly-
           ethylene) .  Approximately 1/3 scale,
                    417

-------
     This synthetic sewage provided a waste with the follow-
ing approximate strength:

         BOD5  s   250 - 290  mg/1

         COD   s   300 - 350 mg/1               ;


The above synthetic sewage can be concentrated and diluted
to cover a wide range of organic loading at any specific
hydraulic, loading.
     The influent and effluent was monitored daily with
analyses being performed for suspended solids, volatile
suspended solids, and soluble chemical oxygen demand.  These
analyses were conducted in compliance with Standard Methods
(1975).  Dissolved oxygen was measured with a D.O meter
(YSI model 54ARC) and pH with a pH meter (Corning Model 12).
Both the, D.O. and pH were measured jLn situ as well as the
flow and temperature.  Samples were filtered (for TSS,  VSS,
COD) using glass fiber filter, Type A-E 47 mm diamter, manu-
factured by Gelman Instrument Co.  The biofilm density was
measured by a scraping technique from a measured area and
the scraped biomass volume was measured volumetrically.
Film density estimated in the scraped biomass gravimetrically.
EXPERIMENTAL PROCEDURES
     Prior to the experimental work detailed in this paper,
the pilot plant was operated for three months, approximately,
under the following loading and other physical conditions:
                                         2
       Hydraulic Loading;     1.85 gpd/ft   (based on flat
                                               area)
       Influent, COD:         305-325 mg/1 (BOD:250-268  mg/1)
       Effluent of first stage,
           COD:               80 + 5 mg/1  (BOD: 77-77 mg/1)

       Rotational Speed:      8 rpm        (70-77) mg/1
                           418

-------
The total suspended solids within any stage, had an average
of about 200 mg/1 and the maximum value found in that time
span was 592 mg/1. On this basis, it was decided to neglect
the effect of suspended solids in the substrate removal.
     Because kinetics and rate studies are best performed
on a batch mode, it was decided to run all the experiments
on the first stage of this pilot plant, using the batch
mode during each experiment.  Before each test was under-
taken, the pilot plant was running under steady state with
a well established biofilm.  Under such conditions, the
biofilm density was 28.34 to 30 mg TS/ml and its thickness
was about 1856-2325 u.  A point worth noting is that all
the disc surfaces were covered by slime.
     When running any batch kinetics test, the flow was
stopped, and the reactor outlets were sealed by a rubber
stopper.  Then adding a precalculated volume (usually small)
of a concentrated solution of the synthetic sewage, the test
was started by collecting samples every five minutes for at
least one hour.  The organic content of each sample was
measured by the COD test after filtering.  Hence the COD
values reflect only solubles.  Total suspended solids were
very low, as mentioned earlier, and neglected.  Temperature
in all the runs was about 21 + 0.5°C.  The dissolved oxygen
within the run of any batch test never went below 2.7 mg/1.
These tests were run at two different levels of initial
substrate concentration; 80 mg/1 COD (average) and 500 mg/1
COD; rotational speed was varied from 4 to 10 rpm in an
increment of 2 rpm.
     To study the effect of dissolved oxygen concentration
in the bulk liquid on the substrate utilization rate, it was
decided to run several tests under a condition of zero D.O.
in the bulk liquid.  To remove the D.O. from the liquor,
nitrogen gas was bubbled through the reactor during the
batch operation.  The nitrogen bubbles stripped the D.O.
from the liquid phase.  In any run, the liquid became void
of D.O. within 15-30 minutes.  But even with zero D.O. the
nitrogen gas was kept flowing for an extra 45-60 minutes.
Then, at time zero, a precalculated volume of a concentrated
sewage was added and the test started while maintaining
the flow of nitrogen gas throughout the run.  D.O. was
monitored continuously and was always 0.0 mg/1 except for
a couple of runs where a trace of D.O. was detected.  Those
i;ests with zero D.O. were all conducted with an initial
substrate concentration of 516 mg/1 COD on the average, and
a rotational speed equals to 4 rpm.
                           419

-------
     The maximum resistance  to mass  transfer usually is
exhibited when no  turbulence at  all exists.   It was desirable
to estimate  this value so  that the importance of rotation
(turbulence) on the efficiency of any  type of disc media
would be demonstrated.  To do so, a  slide previously attached
to the discs and covered with blomass  was suspended in 2S,
of substrate with an  average of 530  mg/1 COD.   This system
was controlled as far as temperature,  D.O. and  minimal mix--
ing were concerned.  This slide provided an area of 168 -cm
approximately.  A ratio of surface area to liquid volume,
af, was calculated at  0,084/cm.
     In order to measure the intrinsic reaction rate of the
substrate uptake rate by the biofilm,  two factors had to be
fulfilled:   (1) eliminate  the resistance to mass transfer
and (2) to minimize the internal diffusion problem  (unless
substrate is consumed at the surface layer).  In the field
of catalytic engineering, this has been achieved by using a
semi-batch reactor.  These units share a common feature of
tremendous high fluid flow rates near  the catalyst surface
to minimize mass transfer resistance (generate  a small Da.
But, unfortunately, this requirement presents a limitation
for the RBC system since it would result in a tremendous
amount of shear and biomass sloughing  at high rotational
speeds.  In addition, all rotational speeds above 10.5 rpm
were impossible because there was excessive liquid loss from
the reactor due to tremendous turbulent splashing.  To over-
come such problems, the following procedure was devised.
Some slime was scraped from a measured area (457.Sem^),
then homogenized in a blender for a  few seconds (less than
8 seconds) and then the dispersed biomass was suspended in
a batch reactor.  Oxygen was maintained by aeration.  The
air was provided for both D.O. (minimum was 9.8 mg/1) and
to assure a complete mix regime.   Samples were  collected
every 5 minutes for one hour.  These samples were centrifuged
and the supernatant filtered.  The filtrate, was analyzed
for organic carbon by the COD test.  All biomass from the
centrifugation and filtration process was returned to the
reactor to minimize the loss of bio  active solids.  This
test was an attempt to minimize the mass transfer resis-
tance of the film as well as the internal diffusion if
such effects existed.  Because this  biofilm was suspended,
it had to be related in some way to  the area of the disc.
This was done by calculating a factor QSJdefined as the
ratio of the area scraped to the batch volume.  In this
test the relationship was 0.23/cm.  The TSS in  such tests
                            420

-------
(the suspended biofilm) measured an average of 4900 mg/1.
     Table (2) below summarizes the tests and testing  condi-
tions under which these runs were performed.

 Table 2_._ Tests Run and Tes_ting_ _Condi_tions

     (all are batch reactors; Temp. =' 21°+ 0.5°C


        Initial Substrate              Bulk        Number  of
 ZZEJL    Cone. COD, mg/1     w, rpm  D.C1. mg/1      _ _....Runs

 RBC,           515             4    2.7-2.8           4
 a=1.92/cm      494             6    2.7-2.8           4
                538             8    3.0-3.1           4
                513       '     10    5.6-5.8           4
                120             4    3.8-4.0           4
                 76             8    4.4-4.6           4
                516             4    0.0*              6

 Slide   .       521             —    >10,2            4
 a =0.084/cm

 Suspended
 Film           551                    >9.8            4
 a =0.23/cm


 *With perfusion of nitrogen gas
                         ••  421

-------
RESULTS AND DISCUSSION
     After obtaining the analytical results from the batch
studies by monitoring S ver time, an attempt was made to
fit these data to several kinetic models:  -

  Zero order    -r  = k      (plot S  ver t)
                  S                 D
  1st order     -r  = k S,    (plot Log S,  ver t)
                  S      D              D
  2nd order     -r  = k S^   (plot 1/s  ver t)
                  S      D            D

where the k's are not the same.

The results of all data without exception, fitted the 1st
order model.  In this paper it is impossible to show all
the results but Fig (4) shows some typical plots.  Table
(3) shows the averages of the results using a calculated
reaction rate constant.
     It can be seen from these results that the observed
rate of substrate utilization in this RBC reactor is 1st
order in the measurable bulk substrate concentration, at
least within the range of experimental data obtained.  As
a result:

             R  -  ko  Sfa                  (       (6)

k  is the observed 1st order substrate reaction rate con-
stant in units of LT"1.  This then is related to keo(where
"e" indicates the natural logarithm base, and "o" indi-
cates observed) as follows:

               k0 • ke,o/a

k  ,o is calculated from the slope of the plot of LnS, ver
t by fitting the data into a linear regression model using
the least square estimator.  These results have an impor-
tant implication.  It was demonstrated earlier in this
paper that for an observed rate to be first order, and
because mass transfer is likewise 1st order, it follows
that the intrinsic rate has to be first order (method
of additive or combined resistance, Frank Kamemetskii (13).
In fact it was also demonstrated that the intrinsic  rate
was a 1st order rate.  These data also show that the
                         422

-------
             Table 3.  Summary of Experimental Results
IX)
oo
Initial
Type Cone.,
RBC,
a=1.92 ./cm





Slide
af = 0.084
Substrate
COD, mg/1
515
494
538
513
120
76
516
521
/cm
w, rpm
4
6
8
10
4
8
4
«.

Bulk D.O, , mg/1
2.7-2.8
2.7-2.8
3.0-3.1
5.6-5.8 '
3.8-4.0
4.4-4.6
0.0
>10,2

k o, I/day .t<
35.83 + 2.27
53.97 + 2.57
56.09 + 5.13
64.65 + 2.69
14.30 + 2.51
23.89 + 8.80
34.54 + 3.46
1.24*0.039

_kejD cm
"o a day
18.66
28.11
29.21
33.67
7.45
12.44
17.99
14.76

ka, cm/day
22.68
38.36
40.43
49.51





             For  the intrinsic reaction rate:  (at an initial substrate  concentration of 551 mg/1

             COD) from the plot  (InS) ver  (t)  get the slope: ke,i =-24.2 + 1.15,  I/day
               and as related  to  surface area, where a  =  457.8
                        ki =
24.2

U.23
             2000


= 105.22 cm/day
                                                                  =  0.23   /cm
             *This is the slope of the plot in S ver t, where t = time

-------
   600
   400

   300
                      r=-0.997
   200
   100
          Run B7
          Ct) =4 rpm
Cn
E
p
o
o

i
-H
"H
nS
§
600

400
300

200
   100
                      r=-0.999
          Run BIO
         6J=8 rpm
   500
   400
   300

   200
   100
       0
                    t-0.988
          Run BN16
          6J=4 rpm(zero D.O.)
           10
             20
       Figure 4a.
 30   40   50   60
  Time, minute
Log remaining COD ver.
time for first order
kinetics model.
                      424

-------
500 P




400




300




200



500
 ,400
Q
O
O
£
Q)
  300



  200
  400



  300



  200



  100
                      Run B7

                      ^ =4 rpm
                    Run BIO
                     <^=8 rpm
                    Run BN16
                     <*> =4  rpm
                     (zero D.O.)
      0
              20   30   40
               Time, minute

Figure 4b.  Remaining COD ver.  time
for zero order Kinetics model.
60
                      425

-------
                   Run B7
                   "•> =4 rpm
                      Run BN16
                        =4 rpm
                      (zero D.O.)
0
10
60
70
               20   30   50
               Time, minute
Figure 4c.  I/remaining COD ver. time for
second order kinetics model.
               426

-------
observed rate is 1st order regardless of the rotational
speed, within the range of speeds tested.
     The observed reaction rate constant k  increased with
an increase in w.  Since the data of Table 3 likewise
indicates that km is substantial which implies that mass
transfer is significant, this result should be expected.
Indeed, as  w  increases, the thickness of the diffusion
boundary layer (different than Prandtl hydrodynamic boundary
layer) should decrease.  Levich (25) has demonstrated that
the thickness of the diffusion layer, for a disc fully
submerged in a liquid and rotating around its own axis
is inversely proportional to the square root of w.  This
author estimated that this layer thickness is about one
tenth of Prandtt layer.  As a result, increasing w should
boost the efficiency of substrate removal but only up to
a limit beyond which little improvement would result.  From
the negative point of view a high w would increase the
sloughing rate, and would have a detrimental effect on bio-
chemical removal.  A representative, empirical equation
relating the mass transfer coefficient to w has been
developed:

            .         C7  ,_,0.809                    , _.
            km  =  7.87  (w)                          (13)

where k  is in cm/day and w in rpm.  Fig (5) shows this
relationship.   It is important to realize that this equation
holds only for the media, used in this experiment, that it
can not be used for extrapolation beyond w = 10 rpm, and
finally that it should not be used for scale-up.  One
should expect that km would become independent of w beyond a
certain value.  This equation (or any similarly derived
equation) would enable a designer to estimate the effect of
media geometry as to whether or not the mass transfer
regime would be eliminated at the minimum rotational speeds
expected in the prototype assuming other parameters are not
affected.  Equation (13) does not include the km values
calculated for the slide test or those from the low initial
substrate concentration run for the following reason. The
tests run using the  slide  and its attached biomass, attempt
to simulate a condition where turbulent does not exist.
Such a test does not have a practical value. However, it
helps in showing the maximum transfer resistance.  Beside
that, it is clear that at 4 rpm the observed mass transfer
rate did not show much increase, showing that the mass trans-
fer resistance is still large.
                           427

-------
  100

  90

 £80

^s. 70


 0 60



% 50
 0

"o 40
-H
g 30
o
0)


c 20
(0
                         1     |    1   1  1  1  1  1
                                                         L
                   rcO.962
ftf
&
  10
       I


       1
                  1
                         1
            1    1   1  1   1  1 1
2      3    456769

  Rotational speed, rpm
                                             1D
     Figure  5.
                 Empirical  relationship between the mass

                 transfer coefficient and the rotational

                 speed
                          428

-------
     Where running the test in the RBC using a batch_mode
for two different initial substrate concentrations  (w = 4 rpm,
with COD: 120 mg/1 and 515 mg/1) and  (w = 8 rpm, with COD
76 mg/1 and 538 mg/1), one might have expected some differences
to be evidenced.  In these cases differences were minimal
and as far as the rate mechanism was concerned, its 1st order
dependence on S, was maintained.  To make this point more
clear, it should be recalled that the intrinsic rate is a
biokinetic mechanism.  Generally, the Monod equation is
used:
             -T  =
               s       K   + S
                        sat

where two extreme,cases can be expected;

(1)    low S:  such that K     »S
                          sat

   hence:  -r   = (kx  s  „    ,     ,   ,..       ,      _.
             s     j7~  ) * S    (pseudo-first order in S)
                    sat

(2)  high S, such that K   <550 mg/1
 Sclt y                                  Sett
measured in COD for the carbon compound (Dextrin (C/.H., n^s^ '
Accordingly this biofilm needs a very high .substrate
concentration (above 550 rag/1 as COD) to reach half the
maximum specific growth rate  (p ). This high value for K
                                                          SeiC
is much larger than those reported in the literature
(Grieves, (14).
     Fig (6) shows the effect of K    on p in the Monod
equation.  Another point one might expect is that the
observed rate constant, k , under the same w would have the
                            429

-------
                   increasing  Ksat
        Substrate concentration
Figure 6. Relationship between the
          specific growth rate and
          the substrate concentration
          showing the effect of Monod
          saturation constant on
          Monod equa ti on.
                  430

-------
same value for two different initial substrate concentration
(e.g., 120 and 516 mg/1 COD @ w = 4 rpm).  But it was not
the case.  Two possible justifications could be thought of:
The first is that Monod saturation constant has changed,
but this is very unlikely; the second is that sloughing
occurred between these tests (there was 6 days lap time)
and the biofilm interface texture has changed hence affect-
ing the mass transfer coefficient.
     One striking point was that the dissolved oxygen con-
centration in the bulk fluid did not affect the substrate
removal rate.  Even under zero D.O. conditions, the observed
reaction constant for the 1st order mechanism was not
affected.  This contradicts the generally accepted concept
in the biological wastewater treatment field that a minimum
of 2 mg/1 D.O. in the bulk fluid should be maintained for
successful operation.  This finding implies that the source
of oxygen required by the biofilm for the oxidation of sub-
strate is the surrounding atmosphere and that the bulk of
the liquid plays little or no part in this two foot model.
If this observation could be extrapolated to a prototype
plant for BOD removal, it would not have to be operated
at a certain rotational speed controlled by the bulk D.O.
for a minimum value of 2 mg/1.  Rather, w should be looked
at as the frequency of this system at which any point in
the biofilm should be exposed to the atmosphere for optimum
removal efficiency under a given substrate strength and
loading.  Of importance too would be the additional effects
of increasing k  and decreasing the thickness of the diff-
usion boundary -layer.  The increase in efficiency gained
by enclosing the RBC in an atmosphere enriched with pure
oxygen or increasing the air pressure (Torpey et al., (33)
and Bintanja _e_t jal (7) actually enhances the diffusion
rate by increasing the partial pressure and the concentra-
tion gradient of the oxygen into the biofilm'besides over-
saturating the liquid film attached to the bioflira, as it
emerges from the tank.
                            431

-------
CONCLUSIONS
1.  Bulk D.O., as low as 0.0 mg/1, did not affect the
    substrate removal rate in this pilot plant work.

2,  The observed rate was first order in the substrate
    bulk concentration only when the intrinsic rate was
    also pseudo-first order.  Under this case the observed
    rate ko is related to the mass transfer coefficient
    k  and the surface intrinsic rate as follows
     m
    l/k0 = l/kra 4- 1/k'i, where all values are in unit of
    length per unit of time.

3.  The mass transfer coefficient km was related to the
    rotational velocity by the empirical formula:

           km  =  7.87  (w)0'809                (13)

where km is in em/sec and w in rpm.  This is valid strictly
for this pilot unit and for w values up to 10 rpm.

4.  In this case, the Monod saturation constant, K   , had
    a larger value than 550 mg/1 soluble COD when tne
    carbon source was provided by dextrin (C,H.. ^Or) •  This
    is higher than previously reported values in the
    literature.

5.  It appears that a kinetic study should be done on each
    stage separately and not on the overall system as a
    single unit.  This latter assumption can result in
    an error in detecting the rate order.
                            432

-------
 REFERENCES
 1.   Antonie, R,L.,  Kluge, D.L., and J.H. Mielke, "Evaluation
     of a Rotating Disk Wastewater Treatment Plant, "
     J_. WPCF, 46, 498 (1974).
 2.   Antonie, R.L.,  Fixed Biological Surfaces-Wastewater.  The
     Rotating Biologi cal Contactor, CRC Press, Inc.,  (1976).
 3,   Atkinson, B., Busch, A.W., and G.S. Dawkins, "Recircula-
     tion, Reaction Kinetics, and Effluent Quality in A
     Trickling Filter Flow Model," -J. WPCF, 35, 1307  (1963).
 4.   Atkinson, B., Swilley, E.L. ,,.Busch, A.W., and
     D.A. Williams,  "Kinetics, Mass Transfer, and Organism
     Growth in a Biological Film Reactor," Trans. Inst. Chem.
     Engrs..  45, T257 (1967).
 5.   Bailey,  J.E., and D. F. Ollis, Biochemica_l Engineering
     _Fundainenjta.ls_, McGraw-Hill Book Company, N.Y.1977.
 6.   Baillod, C.R.,  and W.C. Boyle, "Mass Transfer Limita-
     tions in Substrate Removal," J_. S_an Eng^in. Div.^, ASCE,
     96, SA2, 525 (1970).
 7.   Bintanja, H.H.J., Brunsmann, J.J., and C. Boelhouwer,
     "The Use of Oxygen in a Rotating Disc Process," Water
     Res., JJO, 561 (1976)
 8.   Borchardt, J.A., "Biological Waste Treatment Using
     Rotating Discs," Biotech. & Bioengin S_yinp. No, 2, 131
     (1971)
 9.   Borchardt, J.A., Kang,, S.J., and T. H. Chung, "Nitrifi-
     cation of Secondary. Municipal Waste Effluents by Rota-
     ting Bio-Discs," EPA-600/2-78-061, June 1978.
10,   Carberry, J. J., Chemical and Cata_ly_tic Re act ion
     Engineering, McGraw Hill Book Coinp., N.Y. 1976.
11.   Chesner, W.H.,  and A.H. Molof, "Biological Rotating
     Disk Scale-up Design: Dissolved Oxygen Effects,"
     Prog. Water Tech., _9, 811 (1977).
12.   Clark, J. H., Moseng, E.M., and T. Asano, "Performance
     of a Rotating Biological Contractor Under Varying Waste-
     water Flow," J.WPCF, 50, 896 (1978)
13*   Frank Kamenetskii, D.A., Diffusion and Heat Transfer
     in Chemical Kihetics, Plenum Press, N.Y., 1969
14.   Grieves, C.G.,  Dynamic and Steady-State Models for the
     Rotating Biological Disc Reactor, Ph.D., dissertation,
     Clemson University, 1972
                             433

-------
15.  Harremoes, P., "The Significance of Pore Diffusion
     to Filter Denitrification," j;. WPCF, 48, 377 (1976).
16.  Harremoes, P, "Biofilra Kinetics, in Water Pollution
     Microbiology, Vo1. 2, edited by P. Mitchell, J. Wiley
     1978.
17.  Hartmann, H., "Untersuchung Liber die Biologische
     Reinigung von Abwasser mit Hilfe von Tauchtropfkorperan-
     lagen" (Investigation of the Biological Clarification
     of Wastewater Using Immersion Drip Filters).  Band 9
     der Stuttgarter Berichte zur Siedlungswasserwirtschaft
     Kommissionsverlag Munich R. Oldenbourg 1960.
18.  Hitdlebaugh, J.A., and R.D. Miller, "Operational
     Problems with Rotating Biological Contactors",  J_. WPCF,
     _53, 1283 (1981)
19.  Hoehn, R.C., and A. D. Ray, "Effects of Thickness on
     Bacterial Film, " j;. WPCF, 45, 2302 (1973).
20.  Horvath, C., and J.M. Engasser, "External and Internal
     Diffusion in Heterogeneous Enzyme Systems", Biotech.
     and Bioengin, 16, 909(1974).
21.  Kornegay, B.H., and J.F. Andrews, "Kinetics of Fixed
     Film Biological Reactors", 22nd Ind. Waste Cqnf.,
     1967, Part 2, Pudue U., 620 (1967)
22.  Kornegay, B. H., "Modeling and Simulation of Fixed
     Film Biological Reactors for Carbonaceous Waste Treat-
     ment", in: Mathematical Modeling for Water Pollution
     Control Processes",  Edited by T.M. Keinath and
     M.P. Wanieliste Ann Arbor Science, Publisher,
     Ann Arbor, MI (1975)
23.  LaMotta, E.J., Evaluation of Diffusional Resistances
     in Substrate Utilization By Biological Films",  Ph.D.
     Dissertation, University of North Carolina, Chapel
     Hill, 1974
24.  LaMotta, E.J., "External Mass Transfer in A Biological
     Film Reactor", Biotech, and Bioengin,  18, 1359 (1976)
25.  Levich, V.G., Physicochemical Hydrodynamics, Prentice-
     Hall, Inc., 1962.
26.  Maier, W.J., Behn, V.C., and C.D. Gates, "Simulation
     of the Trickling Filter Process", J_. San Engin, Div. ,
     ASCE. j>3_, SA6, 91 (1967)
                           434

-------
27.  Mikula, W.J., Reynolds, J. H., George, D.B., Porcella,
     D, B., and E. J. Middlebrooks, "A Kinetic Model for
     the Treatment of Cheese Processing Wastewater with
     A Rotating Biological Contactor", in Proc. 1st Nat'l
     _S_ymp/Wor k s h o p on Rotating B io 1 og i c a 1 Contractor
     Technology, 491 (1980). Edited by E.D. Miller, et. al.
28.  Pano, A,, Reynolds, J.H., and E.J. Middlebrooks,
     "The Kinetics of a. Rotating Biological Contactor
     Treating Domestic Sewage", in: Proc. 1st Nat'l Sy.mp/
     Workshop on Rotating Biological Contactor Te ch., 449
     (1980) edited by E.D. Miller, et al.
29.  Popel, F., "Leistung, Berechnungund Gestaltung von
     Tauchtropfkorperanlagen "(Estimating, Construction,
     and Output of Immersion Drip Filter Plants), Stufatgarte.r
     Berichte zur Siedlungswasserwirtschaft, 11,
     Kommissionverlag R. Oldenbourg, Munchen 1964.
30.  Stover, E.L., and D.F. Kincannon, "Evaluating Rotating
     Biological Contactor Performance," Water & Sew.  Works,
     123, 88 (1976)
31.  Tomlinson, T. G. , and D.I1.M. Snaddon, "Biological
     Oxidation of Sewage by Films of Microorganisms," AjLr
     and Water Pollut. Int. _J. , 10, 865 (1966) .
32.  Torpey, W.N., Heukelekian, H., Kaplovsky, A.J.,  and
     R. Epstein, "Rotating Disks with Biological Growths
     Prepare Wastewater for Disposal or Reuse", ^J. WPCF, 43
     2181 (1971)
33.  Torpey, W., Heukelekian, H., Kaplovsky, A.JN , and L.
     Epstein, "Effects of Exposing Slimes on Rotating
     Discs to Atmospheres Enriched with Oxygen", in Adv.
     in Water Poll. Res., 405 (1972) Edited by S.Ii. Jenkins.
34.  Welch, P.M., "Preliminary Results of a New Approach
     in the Aerobic Biological Treatment of Highly Concen-
     trated Wastes", 23rd Indus. W_aste Conf., 428 (1968)
     Purdue U.
35.  Williamson, K.J., "The Kinetics of Substrate Utiliza-
     tion by Bacterial Films", Ph.D. Thesis, Stanford U.,
     1973.
36.  WPCF, "Operation of Wastewater Treatment Plants",
     Manual of Practice No, 11, WPCF, 1976
                            435

-------
                      NOMENCLATIJRE t

Symbol                    Definition                  Unit

A       Disc or biofilm area                        I/
a       Ratio of disc area to bulk liquid             1
                       volume                       L
a,;      as a above but for slide                    1.
                                                      1
a       as a above but suspended film               L
 S
Da      Damkohler number                            —
Da      Observable Damkohler number                 —
                                                      	T
k.      intrinsic rate constant (1st order)         LT

k       Substrate mass transfer coefficient            ,
 m      (liquid film)                               LT
                                                      _i
k       Observed rate constant (1st order)          LT

keso    kQ as measured from plot in S,  ver t;

         ke,Q = ko« a                               T

ke,i    k. as measured from, plot in S.vert; ke»i=
         X  k.»a_                                   T
             a.  s

k       Maximum specific substrate utilization        1
        rate in Monod equation                      T

k       Maximum reaction rate in Monod equation;       „  1
 maX      k    = k.x                                ML'V1
           max
                                                      —3
K       Monod saturation constant                   ML
 sat
                                                      -2 -1
R       Overall reaction rate                       ML  T
                                                      -2 -1
r       Substrate removal rate                      ML  T
 s
S,       Limiting Substrate concentration in           _„
 b          bulk fluid                              ML

S       Limiting substrate concentrate at biofilm      „
            interface                               ML
                           436

-------
Symbol                  Definition



s       Ratio of S  to S,
                  s     b

                                                        3
V       Liquid volume in RBC reactor                   L




                                                         -2
X       Biofilm biomass density as TS or TVS           ML



Y       Yield coefficient in Monad equation              —



n.       External effectiveness factor                    —



                                                         -1
]i       Maximum specific growth rate in Monod equation  T



w       Disc rotational speed                          rpm
                           437

-------
             ANALYSIS OF STEADY STATE SUBSTRATE
                 REMOVAL MODELS FOR THE RBC
      David E. Schafer. Camp, Dresser and McKee, Inc.,
      Boston, Massachusetts.

      James C. O'Shaughnessy.  Department of Civil Engi-
      neering, Northeastern University, Boston, Massachusetts.

      Frederic C. Blanc.  Department of Civil Engineering,
      Northeastern University, Boston, Massachusetts.
INTRODUCTION

     This paper evaluates three independently-derived steady
state mechanistic substrate removal models for the rotating
biological contactor (RBC), intended for use by design engi-
neers.  The three models evaluated are:  1) Kornegay's steady
state model for carbonaceous waste treatment;  2) Schroeder's
steady state RBC design equation;  and 3) Grieve's Pseudo-homo-
geneous steady state model.  Utilizing a common data base, each
has been assessed with respect to model calibration, adequacy
of fit, relative influence exerted by various design parameters,
and limitations and restrictions.
MATHEMATICAL MODELING OF THE RBC

     Several empirical models have been developed to predict
the steady state substrate removal in RBC units (1,2,3).  These
models express stage-by—stage removal of substrate as a power
function of major design variables such as hydraulic loading
rate, influent concentration, retention time, surface area,
                                438

-------
.temperature, disc  configuration  and .rotational  speed.
      To  improve RBC  process modeling,  current research  efforts
are being  focused  upon  the development of  a mechanistic  or  de-
terministic model  for substrate  removal.   Although  proposed
mechanistic models for  wastewater  treatment processes also  pos-
sess  empirical qualities, a "true" mechanistic  model is  defined
as one which assists understanding and allows useful, though
not necessarily exact,  extrapolation over  a wide  range  of oper-
ating conditions  (4) .
      Mechanistic modeling of  substrate uptake and cell  growth
in biological systems is highly  complex.   Even  in the simplest
biological reaction, a  multiplicity of cellular reaction mech-
anisms take place.   Adsorption,  enzyme catalysis  and diffusion-
al processes represent  major  functional mechanisms  which can
control  the uptake of a specific substrate (4).
BIOFILMS

     Biologically, each RBC consists of  a  complex  interrelated
population of  predominantly heterotrophic  attached microorgan-
isms.   In general, this attached mieroogranism population will
be  comprised of  aerobic,  facultative and anaerobic bacteria  (5).
In  addition, as  indicated by Kornegay  (6), a  significant popu-
lation  of suspended microorganisms may also be present if the
system  is operated at a long hydraulic retention time.
     With respect to substrate  removal,  the concept of an "ac-
tive" microbial  depth has been  adopted by  several  investigators
(6,7,8,9,10 and  11).  This hypothesis  divides the  total micro-
bial film thickness into  two layers.   The  outermost layer, be-
ing in  direct  contact with the  adhered liquid film, is termed
the active layer.  The "inactive layer", if present, is in di-
rect contact with the support media.                '
     Sanders (9) evaluated active depth  in terms of the "criti-
cal" depth at  which diffusion of oxygen  within the slime layer
becomes limiting.  Tomlinson and Snaddon (10) have also sug-
gested  that the  active layer consists  of the  aerobic microorgan-
ism zone.  However, Atkinson and Davies  (11), Kornegay (6) and
Grieves (7) contend that  the active depth  should be defined
with respect to  the .depth of penetration of a limiting nutrient.
     Estimated values of  active microfilm  depths have ranged
from 27 to 200pm.  To date, no  universally acceptable technique
exists  for the measurement of active depth in any  fixed film
system. Conceptually, substrate removal from the  bulk liquid
phase requires diffusion  of metabolic  reactants into the at-
tached  biofilm,  metabolism by the organisms,  and diffusion of
                                439

-------
the metabolic by-products back through the biofilm and into
either the bulk liquid or the atmosphere.  Since relatively
thick biofilms are employed, significant concentration gradi-
ents, resulting from mass transport resistances, can exist be-
tween the bulk liquid and the active microbial layer (12).
TREATMENT KINETICS

     In 1950, Monod presented an initial mathematical analysis
for cell growth based upon work with batch reactors (13).  His
hypothesis assumes that microorganism growth rate is dependent
upon the concentration of a limiting substrate, which he tested
using a completely-mixed continuous flow chemostate containing
a dispersed culture of microorganisms.  The versatility of the
Monod kinetic relationships in fitting data normally obtained
from a variety of wastewater treatment systems has made it a
logical starting point for modeling the RBC process (6,7,14 and
others).
STEADY STATE MODELS FOR THE RBC

Kornegay's Model

     The mathematical algorithms proposed by Kornegay to simu-
late RBC system performance have been developed under the as-
sumption that ultimate substrate removal is dependent upon mi-
crobial growth and that the entire mass of attached film is not
considered active in the removal of organics.  Additional as-
sumptions are as follows (6) :
     1) complete mixing is achieved in the liquid volume;
     2) organism decay is neglected;
     3) maintenance energy is not included in explicit terms;
and  A) saturation or Monod function coefficients are assumed
        to remain constant during periods of dynamic operation.
     Kornegay's approach to system performance under continuous
flow conditions is illustrated in Figure 1 and expressed by the
following steady state equation (6) :
                         y   2NTT(rl2-r22)Xd(C, )
                          max                P
where:  Vmax is maximum specific growth rate of fixed film or-
ganisms; Kc is half saturation constant; Y is the apparent
yield of fixed film organisms; F is influent flow rate;  Co is
                                440

-------
                                                 Effluent Flow
         FIGURE 1: Definition Sketch for Kornegay's
                   Substrate Removal Models
influent substrate concentration; C^ is reactor substrate con-
centration; N is the number of discs; rl is total disc radius;
r2 is unsubmerged disc radius; X is unit mass of the fixed mi-
crobial film; and d is active film depth.
     Setting the area capacity constant, P, to ymaxXd/Y and the
contact area, A, to 2Nir(rl2- r22) , the removal equation for a
single stage RBC in which suspended growth is negligible be-
comes (6) :

                    F(co-cb> =PA
     Multi-stage operation can be evaluated by setting the in-
fluent concentration of the second reactor equal to the bulk
liquid (i.e., effluent) concentration of the first reactor and
performing successive iterations until all reactor concentra-
tions are known.
                                 441

-------
Schroeder's Model

     Schroeder's steady state design model for the.RBC process
is based upon a theoretical analysis of substrate utilization
by microbial films conducted by Atkinson and his coworkers dur-
ing the late 1960's and early 1970's.  Schroeder has modified
the Atkinson Model for use in municipal wastewater treatment
applications, incorporating the following assumptions (14):
     1) slime  phase diffusion controls overall system perform-
        ance;
     2) no significant concentration gradients exist within the
        adhered liquid film while in the bulk gas phase;
     3) mass transport through a differential element follows
        Fick's Law of Diffusion; and
     4) A plug flow mode of operation is appropriate in model-
        ing the EBC process.
     Schroeder's approach to system performance under continu-
ous flow conditions is expressed by the following steady state
equation (14):

                  1    1        C0   f K* Ag 8 d
               K(-^   7r~~) "*" In. 7T~ =      ™(3)
                 Cb  C0       Cfc        VL

where:  K is the half saturation constant; Cf, is bulk liquid
substrate concentraton; Co is influent substrate concentration;
f is the proportionality factor; K* is the maximum specific
growth rate; As is submerged disc area; 6 is reactor hydraulic
retention time; d is active biofilm depth; and V^ is liquid vol-
ume per disc.
     Multi-stage operation is evaluated using a technique simi-
lar to the Kornegay approach.
Grieves * Model

     Grieves combined both biological growth kinetics and mass
transport resistances in the development of a dynamic substrate
removal model for the RBC.  Grieves adopted a more complex phys-
ical representation for the system than those developed by Kor—
negay or Schroeder by subdividing each disc into pie-shaped seg-
ments as detailed in Figure 2.  Within each segment, substrate
is assumed to be transferred across a biofilm-liquid film inter-
face at a rate directly proportional to the concentration gra-
dient between the two phases.  Major model assumptions are as
follows (7):
                                442

-------
           FIGURE 2: Grieves' Model-Details of the
                     Elements Taken Around the Disc
     1) there is complete mixing'in the reactor, biofilm and
        liquid film;
     2) as the disc leaves the bulk liquid phase, a stationary
        liquid film having an initial substrate concentration
        equal to that of the bulk liquid phase adheres to the
        microbial film;
     3) substrate utilization by an individual microorganism in
        the biological film can be represented by a saturation
        or Monod function;
     4) Monod coefficients are assumed to remain constant dur-
        ing periods of transient operation;
     5) the mass of substrate consumed by the organisms for
        maintenance purposes is negligible when compared with
        that used for growth; and
     6) substrate diffusion in the radial and circumferential
        directions is negligible when compared with diffusion
        into the biological film.
     In the derivation of this model, Grieves has considered
mass balances on substrate in:  1) any segment of the liquid
film exposed on the disc; 2) the biological film exposed on the
disc; 3) any segment of the biological film submerged in the
bulk liquid; and 4) the reactor bulk liquid.  With the addition-
al assumption that first order removal kinetics are appropriate
when substrate concentrations are relatively low, Grieves*
steady state model takes the form (7):
             1 +
)(AS)  + Ff[l - e
                                        {(PO(A )/Ff>
                                                            (4)
                               443

-------
where:  C-^ is bulk liquid substrate concentration; Co is influ-
ent substrate concentration; N is number of discs per reactor;
F is influent flow rate; PJ is (KL)(K{)/(1+K{); Aa is area of
the disc in the air; Ff is liquid film flow rate; As is sub-
merged disc area; KJ is {(ymax) (IF)11"1 (X) (Az) }/{ (Y) (Kc) (n) (KL)} ;
TF is treatability; N is stage number; X is organism density in
segment L,M: Y is organism yield coefficient; Vma-x is maximum
specific growth rate; Kc is saturation constant in Monod equa-
tion; 11 is the effectiveness factor; KL is mass transfer coef-
ficient; and Az is active biofilm thickness.
PILOT PLANT OPERATION

     The three steady state models were compared using simula-
tion results and data collected in a pilot plant investigation
conducted at the Yankee Greyhound Racing, Inc. dog track lo-
cated in Seabrook, New Hampshire.  Septic tank effluent charac-
teristics (used as influent feed to the RBC unit) measured dur-
ing the 60~day pilot plant study are summarized in Table I and
listed in Tables II and III (15).  High nitrogen and total and
soluble BOD values indicate a waste strength roughly three times
that of normal domestic sewage.
                          Table I

      Septic Tank Effluent Characteristics Pilot Plant
      Influent Feed, Throughout the Testing Period (15)

      Parameter                             Range**

      BOD5                                250 - 600
      COD                                 350 - 750
      Suspended Solids                     50 - 200
      M3-N                               100 - 200
      Organie-N                            50 - 100
      N03-N                                 <1.0
      PO^-3                                10 - 20
      Grease and Oil                       50 - 200
      Alkalinity                          250 - 500
      pH                                    6-8

      **A11 values except pH in mg/1.
                                444

-------
              Run
Date
                                                          Table II

                                  Yankee Greyhound Inc.  Dog Track Pilot Plant Data Summary
Temp.
Flow
Rate*
Influent
Cone.**
Stage
Stage
 2**
Stage
 3**
Stage
 4**
45.
•*»
en
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
5-12
5-14
5-19
5-22
5-25
5-27
5-30
5-31
6- 2
6- 4
6- 6
6- 9
6-11
6-15 ,
6-17
6-19
6-21
6-23
12.0
14.5
13.0
11.5
9.5
12.0
16.0
17.5
14.0
12.5
12.5
17.0
18.8
18.0
20.0
19.0
—
—
0.50
0.50
0.67
1.00
0.93
0.80
0.40
0.40
0.37
0.40
1.82
1.30
.0.26
0.25
0.40
0.15
0.15
0.15
212
288
275
223
265
375
235
375
515
250
305
395
470
400
535
365
420
455
84
155
133
123
118
126
68
110
143
68
102
256
88
89
188
39
60
86
29
80
50
53
62
74
37
138
83
42
59
182
20
39
68
45
45
18
16
38
43
19
27
30
28
65
45
11
36
110
16
21
24
29
24
29
14
28
13
16
22
25
27
39
37
19
12
58
9
18
24
8
22
22

-------
                                           Table III

                  Pilot Study Process Loading Factors and Removal Efficiencies
               First Stage Loading Factors
Hydraulic
Loading*
1.80
1.80
2.41
3,60
3.35
2.88
1.44
1.44
1.33
1.44
6.55
4.68
0.94
0.90
1.44
0.54
0.54
0.54
Organic
Loading**
3.183
4.323
5.532
6.695
7.399
9.007
2.822
4.504
5.721
3.002
16.666
15.417
3.669
3.002
6.425
1.644
1.892
2.049
% BOD
Removal
60
46
52
45
55
66
71
71
72
73
67
35
81
78
65
89
86
81
 Run
   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
 *gal/day-sq.ft.
**lb. sol. BOD/day-1000 sq. ft.
Overall Unit Loading Factors
Hydraulic
Loading*
0.45
0.45
0.60
0.90
0.84
0.72
0.36
0.36
0.33
0.36
1.64
1.17
0.23
0.23
0.36
0.14
0.14
0.14
Organic
Loading**
0.796
1.081
1.383
1.674
1,850
2.252
0.706
1.126
1.430
0.751
4.167
3.854
0.917
0.751
1.606
0.411
0.473
0.512
% BOD
Removal
93
90
95
93
92
93
89
90
93
92
96
85
98
96
96
93
98
95

-------
     The pilot plant utilized was a 4-foot, 4 equal stage ro-
tating biological contactor supplied by the Environmental Pol-
lution Control Division of the George A. Hormel Company (EPCO-
HORMEL) of Austin, Minnesota.  This unit provided a total of
1600 square feet of polyethylene surface area for biomass
growth and a liquid volume in the unit of approximately 100 gal-
lons.  This unit was fed continuously by a small submersible
pump suspended between the floating scum layer and the bottom
sludge deposit in one of the secondary septic tanks.  The flow
during the study ranged from 0.15 to 1.82 gallons per minute.
During the course of the study, the pilot unit was operated at
4-stage detention times which ranged from 57 minutes to 695 min-
utes and overall organic loading rates ranging from 0.41 to 4,17
pounds of soluble 5 day BOD per day per 1000 square feet (15),
CALIBRATIONS/RESULTS/SENSITIVITY ANALYSES

Kornegay Model

     Calibration of Kornegay*s steady state substrate removal
model involved the evaluation of two unknown kinetic parameters;
Kc and P,  These values are idealy developed by curve fitting
of actual data.  The rearrangement of Equation 2 as follows:
                                Kc  1  .  1
                   (F)(C0-Cb)   P  Cb   P
                                                            (5)
should plot as a straight line having a slope of KC/P and inter-
cept 1/P when 1/Cjj is plotted against the term on the left side
of the relationship (6).
     In the pilot plant study, approximately 60-80% of the total
BOD reduction occurred in the initial stages of treatment.  Es-
timates for the unknown parameters were therefore established
based upon a least squares analysis of the first and second
stage data, presented in Figure 3.  An initial estimate for the
area capacity constant, equal to the inverse of the Y-axis in-
tercept, is 5.62 Ib, BQB/day-1000 ft2 (a value roughly equiva-
lent to the mean first stage organic loading of 5.72 Ib. BOD/
day-1000 ft2 applied during the waste treatability study).  From
the slope of the line, an initial estimate for the saturation
coefficient was determined to be equal to approximately 150 mg/1.
     Simulation results were obtained through rearrangement of
Equation 5 into its quadratic form with respect to C^ such that:
                                 447

-------
03
                        2000 r
                        1500 -
                    u
                    ,Q
                        1000 -
                    o°  500 -
y - mX + b
y - 1.67X + 178
r - 0.79
                                                                                                  1200
                                                         l/Cb  (fC*Ab.
                         FIGURE 3: Kornegay's Model-Final Plot for Biological Parameter  Estimation

-------
              P) + (F)(KC) - (F)(C0)
    Cb2 +	.__	 (Cb) _ (KC)(CO) = 0    (6)


Equation 6 permitted easy calculation of successively-staged
effluent concentrations for each reactor given an influent
waste strength and flow rate.
     In order to test the model's adequacy of fit, a statisti-
cal least sum of squared error analysis was selected to serve
as the basis for final parameter calibration.  Utilizing this
approach, simulation results were equally weighted for each
stage of treatment.  However, because of the relative magnitude
of the initial stage bulk liquid concentrations, numerical em-
phasis was focused upon the initial stages of treatment where
the majority of organic removal occurred.
     For the initial parameter estimates obtained from Korne-
gay's steady state model, the total sum of squared error for
the 72 simulated values (4 stages on 18 testing dates) was rela-
tively large, equaling 119,564.  However, 44% of the total
error resulted from poor simulation of Run 11, attributed to
the model's pronounced response to flow rate variations.
     Figure 4 depicts the model's sensitivity to biological
parameter estimates with respect to single stage removal effi-
ciency.  This figure served as a guide for additional calibra-
tions as it indicated the relative importance of each parameter
within the given range of waste strengths and flow rates.  By
definition, model sensitivity with respect to the saturation
constant varies with reactor bulk liquid concentrations, parti-
cularly with series treatment applications.  However, as the
magnitude of the slope of each line indicates model results are
slight,ly more dependent upon the value of the area capacity con-
stant, P.
     Utilizing Figure 4, the initial parameter estimates were
systematically adjusted in an effort to improve overall model
simulation.  By increasing the area capacity constant from its
initial value of 5.62 to 6.50 Ib. BOD/day-1000 ft2 and decreas-
ing the saturation constant in the Monod relationship from 150
to 135 mg/1, overall simulation results improved approximately
22%, with a total sum of squared error equal to 92,887.  Again,
poor simulation of Run 11 resulted in the generation of 39.5%
of the total squared error.
     Figures 5 and 6 illustrate single stage model response or
sensitivity with respect to variations in organic loading re-
sulting from an increase in either influent waste strength or
flow rate.  Single stage reactor response was selected for il-
lustration because of the dampening effects produced by multi-
                                   449

-------
                                                                 15 1
cn
o
                              -20
                                        -15
                                                 -10
                                                                 -5 •
                                                                -10 .
                                                                -15 j
                                                                                     10
                                                                                              15
                                                                                                       20
                                                                                I Change In Indicated Parameter
                                       FIGURE  4:  Kornegay Model-Biological Parameter Sensitivity

-------
                      CONCENTRATION SENSJTVJTY
                 25<5678BIB
                Ib. sol. BOD app. / day - .1000 ft  .
FIGURE  5:  Kornegay Model-Single Stage Reactor
           Response at Constant Flow Rate
                        FLOY RATE SENSITIVITY
      o  4

       I
                 Ib.  sol. BOD app.  / day - 1000 ft
FIGURE 6: Kornegay Model-Single  Stage Reactor
           Response at  Constant Waste Strength
                               451

-------
stage simulation.
     Figure 5 was generated by utilizing the final values ob-
tained at a constant flow rate of 0.40 gal/rain.  As shown, cal-
culated first stage organic removal efficiency remained essen-
tially constant below the recommended upper limit on organic
loading of 5 Ib. BOD/day-1000 ft2.  Above this value, predicted
removal efficiency decreased due to the fixed nature assumed
for the saturation coefficient.
     Figure 6 indicates the model's response to flow rate vari-
ations and was generated using a constant influent waste con-
centration of 250 mg/1.  Above an organic loading of 3.5 Ib.
BOD/day-1000 ft2 (i.e., first stage hydraulic loading in excess
of 1.5 gal/day-ft2), first stage simulated removal efficiency
decreased from approximately 70 to 35% as first stage organic
loading approaced 10 Ib. BOD/day-1000 ft2 (i.e., as first stage
hydraulic loading approached 4.7 gal/day-ft2).
Sehroeder Model

     Calibration of Sehroeder *s steady state simulation model
(Equation 3) required the evaluation of four unknown parameters:
K, the saturation coefficient in the Monod equation; K*, the
maximum removal rate constant; f, a proportionality factor; and
d, the active biofilm depth.  F, K* and d appear together in
the group YK = (f)(K*)*d) and the reactor retention time, 6, is
equal to the reactor volume, ¥, divided by the flow rate, F.
Therefore, a modified relationship expressing steady state sys-
tem performance can be stated as:


              K(   '     + ln    =
     However, unlike Kornegay's two-parameter relationship,
Equation 7 does not permit rapid parameter estimation through
standard curve-fitting techniques.  Therefore, it was necessary
to establish initial values for the unknown parameters from the
available estimates contained in the relevant literature.
     Assuming that each stage of the four stage unit acts as an
independent reactor, sensitivity runs were performed by solving
for YK in Equation 7 using measured BOD and flow rate values as
well as values for the remaining system variables.  By varying
the value of the saturation coefficient throughout its recom-
mended range, a new range for the variable YK was determined
and extended from 0.00025 to 0.0005.  Similarly, an appropriate
                                 452

-------
range in values for Schroeder's saturation coefficient was
evaluated and found to lie between 5.0 and 10.0 mg/1.
     Single stage results from this analysis are illustrated in
Figure 7,  Although restrictions similar to those indicated for
Figure 4 apply, model calibration was shown to be essentially
independent of the value for the saturation coefficient, K.  By
setting K equal to its estimated mid-range value of 7.5 mg/1,
final calibration was accomplished by altering the value of YK
in order to obtain the best fit for the measured data.  Based
upon a least sum of squared error calculation utilizing the
measured BOD results, a calibrated value for YK in Equation 7
was determined to be equal to 0.0004 cm/sec.
     The final results of model calibration had a total sum of
squared error for the 72 simulated values equal to 112,036,
roughly 47% of which resulted, again, from poor simulation of
of Run 11.  As with Kornegay's model, this was attributed to a
pronounced response by Schroeder's model to high influent flow
rate values.  Nevertheless, overall first stage simulation re-
sults were essentially good, demonstrating the model's capabili-
ty to predict single reactor removal efficiency over a wide
range of organic loadings.
     The model does, however, tend to over-estimate organic re-
moval during the third and fourth stages of series treatment
applications.  This is particularly evident for runs in which
the unit flow rate dropped below 0.50 gal/min; unfortunately,
the theoretical formulation of this model does not permit the
incorporation of a treatability factor, thereby eliminating any
means to attempt mathematical correction.
     To assess the model's sensitivity to variations in flow
rate and influent waste strength, Figures 8 and 9 were devel-
oped.  Again, single stage reactor response was selected to
•best illustrate model sensitivity.  The linear relationship de-
picted in Figure 8 was generated at a constant flow rate of
0.40 gal/min using the calibrated parameter values and indicates
that predicted removal efficiency is first order with respect to
influent concentration.
     Figure 9 indicates simulation results generated by an in-
crease in organic loading caused by an increase in feed flow,
and suggests that an upper limit on removal efficiency was ap-
proached at a hydraulic loading of 2,5 gal/day/ft2 and that
this limit may be increased by increasing reactor retention
time.
                                  453

-------
      Z Change in Indicated Parameter
                                _ 20
                                U 15
                                i. 10
                                _ 5
                                - -5
                                --10
                                _ -20
FI0UKE 7: Sctiroeder Model-Biological Parameter  Sensitivity
                                    454

-------
                  CONCENTRATION
             lb. sol. BOD app / day - 1000 ft/.

FIGURE  8:  Schroeder Model-Single Stage  Reactor
           Response  at Constant Flow Rate
                    FLOW 8*TE SENSITIVITY
             lb, sol. BOD app. / day - 1000 ft .


FIGURE 9: Schroeder Model-Single  Stage Reactor
           Response at Constant Waste Strength
                           455

-------
Grieves Model

     Grieves has Indicated that initial estimates for the bio-
logical model parameters of Equation 4 can be assumed to fall
within the following ranges (7):
     Parameter               Range                  Units
        KL              0.003 - 1.00                cm/sec
        y                0.02 - 0.54        .         1/hr
        Y                0.26 - 0.64                 	
        Kc                  4-10                  mg/1
        X                   8-20                  rag/ml
        n                   1-15
        Az                 50 - 200                  ym

However, the calculated feasible range in values for the un-
known parameter, Pj, varies by several orders of magnitude.
The selection of individual parameter values would thus prove
to be meaningless, especially since no attempt was made to test
for these parameters.  Therefore, based upon the measured re-
sults and physical characteristics of the unit, the calibra-
tion procedure was initially directed towards establishing nu-
merically feasible ranges for both unknown model parameters, Pj
and Ff.
     The term Ff was evaluated based upon theoretical and ex-
perimental analyses conducted by Zeevalink et. al. (16), Bin-
janta et. al. (17) and Levich (18).  In these investigations,
relationships were developed between disc rotational velocity
and liquid film thickness.  Given a disc peripheral velocity of
1 ft/sec, an appropriate range in values for the term Ff was
found to extend from 6.0 to 8.0 cm3/sec/disc face.
     Through substitution of the mid-range value of Ff into
Equation 4, individual Pj_ values were calculated utilizing a
trial and error approach, obtaining rapid convergence for this
function through parameter modification via the Newton-Rapson
method.  An appropriate range of Pj was found to extend from
0.25 x 10"1* to 1.75 x Mr4 cm/sec.
     Model sensitivity runs revealed that simulation results
were essentially Independent of the value for the rotational
flow rate constant, Ff, throughout its recommended range.
Therefore, further calibration involved the evaluation of the
biological model parameter P, or, more appropriately, a value
for PJ for staged treatment applications.
     This was accomplished by adopting a value of 0.40 m/hr
(0.0111 cm/sec) for the liquid film coefficient, KL, a value
                                456

-------
assumed, and' successfully used, by Grieves for the simulation
of a 10-stage RBC pilot plant treating municipal wastewater (7).
Values for the biological coefficient, Kj, and the treatability
factor, TF, which best fit the measured data were then selected.
Finally, based upon a least sum of squared error calculation,
calibrated parameter values were determined.
     Substitution of the values for KL and Kj reveals a Pj
value for simulation of the overall unit (without the use of a
treatability factor) equal to 1.29 x  I'D"1* cm/sec.  This is
roughly equivalent to the mean first and second stage calcu-
lated PI value of 1.33 x  ID"1* cm/sec and resulted in a total
sum of squared error equal to 69,031 for the 18 runs.  However,
46% of this error resulted, once again, from poor simulation of
Run 11.  The Pi values used during simulation with a treatabili—
ty factor equaling 0.75 for stages 1 through 4 were 1.53 x  IQ""1*,
1.15 x  10"1*, 0.88 x  10-t|» and 0.65 x  10 ~1+.  As with Kornegay's
steady state model, use of a treatability factor improved simu-
lation results in runs having an influent flow rate less than
0.60 gal/day-ft .  Overall simulation results, however, were
only found to improve approximately 1%.
     Figures 10 and 11 illustrate the model's steady state re-
sponse to alterations in influent feed characteristics.  Be-
cause of the simplifying assumption that first order microbial
kinetics govern treatment, simulated substrate removal is inde-
pendent of feed strength, as indicated by the linear relation-
ship depicted in Figure 10.  Although experimental results have
confirmed this hypothesis at low influent waste strength, the
validity of this assumption can not be justified if feed
strength were to increase above approximately 750 mg/1 BOD.
     As can be seen from Figure 11, model response is highly
dependent upon influent flow rate.  However, unlike the results
obtained from the sensitivity analysis performed on Schroeder's
model, an upper limit on BOD removal due to influent flow rate
is not implied.
DESIGN CONSIDEJRATIONS/CONCLUSIONS

     Although each of the steady state models evaluated in this
analysis was found to provide an adequate fit of the measured
data, caution must be exercised in the use of either of these
models for design purposes.  Inherent in their respective deri-
vations are several simplifying assumptions regarding process
performance which, if violated or neglected, can significantly
affect "predicted" system response.  Major factors which must
be considered, particularly with respect to calibrated biologi-
                                457

-------
                             StKSIIJVITt
              Ib, sol. BOD app. / day - 1000 ft^,


FIGURE  10: Grieves  Model-Single Stage Reactor
            Response at  Constant Flow Rate
                      TLOV BiTC SENSITIVITY
                Ib. sol. BOD app / day - 1000 ft .
FIGURE 11: Grieves Model-Single Stage Reactor
            Response at Constant Waste Strength
                            458

-------
cal constants and coefficients, are;
     1)  The impact that dynamic loading of the system will have
        upon system response.
     2)  The impact that scale-up will have.  Preliminary re-
        search indicates that  a 15 to 25% scale-up factor is
        appropriate depending  on the physical characteristics
        of the pilot scale unit (19).
     3)  Media configuration/reactor "shor circuiting" has not
        been modeled explicitly.
     4)  Oxygen limitation has  not been considered in any model.
     5)  Temperature effects, which can alter microbial reaction
        rates, diffusivity coefficients and dissolved oxygen
        concentrations have not been considered.
     6)  In reference to Schroeder's steady state model, speci-
        fication of a given surface area also fixes the surface
        area to volume ratio of the reactor.
     The best overall simulation results were obtained util-
ing the Grieves model for this data set.  However, Kornegay's
model was found to have the simplest calibration methodology.
All of the models were found to exhibit a pronounced response
to flow rate variations.  On a single stage basis, removal ef-
ficiency was negatively impacted above a hydraulic loading of
1.5 gal/ft2/day.
REFERENCES

1.   Antonie, R.L., Welch, F.M., "Preliminary Results of a Novel
     Biological Process for Treating Dairy Wastes", Proceedings
     of the 24th Industrial Waste Conference, Purdue University,
     LaFayette, Indiana, 1969.

2.   Joost, R.H., "Systematation in Using the Rotating Biologi-
     cal Surface Wastewater Treatment Process", Proceedings of
     the 24th Industrial Waste Conference, Purdue University,
     Lafayette, Indiana, 1969.

3.   Weng, C.M., Molof, A.H.,  "Nitrification in the Biological
     Fixed Film Rotating Disc System", Water Pollution Control
     Federation Journal, Vol. 45, 1974.

4.   Boyle, W.C., Berthouex, P.M.,  "Biological Wastewater
     Treatment Model Building Fits and Misfits", Biotechnology
     and Bioengineering, Vol. 16, No. 6, September, 1974.
                                   459

-------
 5.   McKinney, Ross E., Microbiology for Sanitary Engineers,
      McGraw-Hill Book Company, Inc., 1977.

 6.   Kornegay, B.H., "Modeling and Simulation of Fixed Film
      Biological Reactors for Carbonaceous Waste Treatment", in
      Mathematical Modeling for Water Pollution ControlProcesses,
      edited by Keinath, T.M. and Wanielista.M., Ann Arbor Sci-
      ence Publishers, Inc., 1975.

 7.   Grieves, C.G., "Dynamic and Steady State Models for the
      Rotating Biological Disc Reactor", Ph.D. Thesis, Clemson
      University, South Carolina, 1972.

 8.   Williamson, K., and McCarty, P.L., "A  Model of Substrate
      Utilization by Bacterial Films", Water Pollution Control
      Federation Journal, Vol. 48, No. 1, January 1976.

 9.   Sanders, W.M., "Oxygen Utilization by  Slime Organisms in
      Continuous Culture", International Journal ofAir and
      Water Pollution, Vol. 10, 1966.

10.   Tonlinson, T.G., and Snaddon, D.M.H.,  "Biological Oxida-
      tion of Sewage by Films of Microorganisms", In te mat ional
      Journal of Air and WaterPollution, Vol. 10, 1966.

11.   Atkinson, B., and Bavies, I.J., "The Overall Rate of Sub-
      strate Uptake (Reaction) by Microbial  Films", Transactions
      of the Institute ofChemical Engineers,  Vol. 52, No. 3,
      July 1974.

12.   Famularo, J., Mueller, J..A., and Mulligan, T., "Application
      of Mass Transfer to Biological Contactors", Water Pollution
      Control FederationJournal, Vol. 50, No. 4, April 1978.

13.   Brook, T.D., Biology of Microorganisms,  Prentice Hall, Inc.,
      Englewood Cliffs, New Jersey, 1979.

14.   Scroeder, E.D., Water and Wastewater Treatment:   Chapter 9-
      Biological Film Flow Processes, McGraw-Hill Book Company,
      Inc., 1977.

15.   Blanc, F.C., O'Shaughnessy, J.C.,  and  LaRosa, A.P.,  "Treat-
      ment of Racetrack Wastewater Using Rotating Biological Con-
      tactors", NewEngland Water PollutionControl Association
      Journal, Vol. 11, No. 2, October 1977.
                                   460

-------
16.    Zeevalkink,  J.A.,  Kelderman,  P.,  and  Boelhouwer,  C.,
      "Liquid  Film Thickness  in a  Rotating Disc  Gas  -  Liquid
      Contactor",  Water  Research, Vol.  12,  No.  8,  1978.

17.    Bintanja, H.H.J.,  Brunsmann,  J.J.  and Boelhouwer,  G.,
      "The Use of  Oxygen in a  Rotating  Disc Process", Water  Re-
      search, Vol.  10, No.  6,  1976.

18.    Levich, V.G.,  Physiochemical Hydrodynamics,  Prentice  Hall
      Inc., Englewood Cliffs,  New Jersey,  1968.

19,    Wilson, R.W., Murphy, K.L., and Stephenson,  J.P.,  "Scaleup
      in Rotating  Biological Contactor  Design", Water Pollution
      Control Federation Journal, Vol.  52,  No.  3,  March  1980.
                                 461

-------
              MATHEMATICAL MODELING FOR ASSESSING
           DEVELOPMENT AND SLOUGHING OF" FIXED FILMS
           AND THEIR EFFECTS ON WASTE STABILIZATION
      Ju-Chang Huang.  Department of Civil Engineering,
      University of Missouri-Rella, Rolla, Missouri.

      Shoou-Yuh Chang.  Department of Civil Engineering,
      University of Missouri-Holla, Rolla, Missouri.

      Yow-Chyun Liu.  Department of Civil Engineering,
      University of Missouri-Rolla, Molla, Missouri.
INTRODUCTION
     One of the major parameters governing the performance
of any biological treatment system is the food to microorgan-
isms (F/M) ratio.  In an aerobic biological unit, the rate
of organic stabilization is directly proportional to the
quantity of aerobic microorganisms (M) present when the sub-
strate concentration is not limiting.  Therefore, to optimize
the use of any aerobic biological treatment unit, efforts
must be made to keep the aerobic microbial concentration as
high as possible.  This is particularly important when the
substrate concentration is high.  However, in most biological
treatment systems, the upper limit of aerobic microbial con-
centration is normally regulated by the oxygen availability.
At a given oxygen level, only a certain limit of aerobic
microorganisms are maintainable, and the rate of organic oxi-
dation depends on this limit.  If the microbial concentration
is maintained beyond this limit, a portion of the microbial
mass would be of either the facultative or anaerobic type,
which can often develop odorous conditions in a treatment
system.  For example, in a suspended-growth system like con-
                           462

-------
ventional activated sludge process, the level of mixed liquor
suspended solids (MLSS) is generally kept within 4,000 mg/L
to insure that molecular oxygen will penetrate into the center
of biological floes.  However, when aeration is provided with
pure oxygen (patented as "Unox Process")* the maximum MLSS
concentration can be increased to as high as 8,000 mg/L be-
cause of the increased oxygen penetration into biological
floes.  With the increased biomass in the Unox Process, its
organic stabilization rate is greatly increased and its re-
quirement of hydraulic retention can thereby be reduced. This
would, of course, result in a substantial saving in the con-
struction of the aeration tank system.
     The development and maintenance of the biomass in an at-
tached-growth (or fixed biofilms) system is considerably more
complicated in comparison to the suspended-growth unit. This
is because the biofilm development on a surface exposed to
waste flow is the net result of physical transport and bio-
logical growth rate processes.  The processes which contribute
to the overall biofilm accumulation are:  1) diffusion of sub-
strate into the biofilm; 2) diffusion of molecular oxygen in-
to the biofilm; 3) substrate oxidation and growth of the at-
tached microorganisms; and 4) sloughing of the biofilm. Among
these processes, it is reasonable to assume that for a given
substrate compound, the rate of substrate diffusion depends
upon its concentration gradient in the biofilm layer. Similar-
ly, oxygen diffusion rate also depends upon its concentration
gradient.  In an actual biofilm treatment unit (such as rota-
ting biological contactor or RBC, trickling filter, and aero-
bic fluidized bed), under a steady-state condition the rate
of substrate oxidation may be limited by either the substrate
penetration or oxygen diffusion depending on the relative
availabilities of these two substances.  In a biological treat-
ment system exposed to air, the maximum concentration of dis-
solved oxygen in wastewater seldomly exceeds 4 or 5 mg/L while
the substrate concentration may be as high as hundreds or even
thousands mg/L.  Under such a situation, diffusion of mole-
cular oxygen into the biofilm is normally the rate-limiting
step in the waste stabilization process.  For example, in a
model-scale fixed film system, it has been found that for a
glucose substrate with a concentration of 88 mg/L or more,
the rate of organic oxidation is generally limited by the ox-
ygen diffusion rather than by the substrate penetration (1,2).
This type of oxygen limitation in the fixed film systems has
also been observed by other researchers  (3,4,5).  Thus,
all of these seem to suggest that in a biofilm treatment
                             463

-------
system, if the influent BOD5 is well above 100 mg/L, a sig-
nificant portion, or even the majority, of the fixed-film
growth will function under an anaerobic condition.  When
this occurs, the total oxidative capability in such a system
cannot be measured by its total biomass since the anaerobic
biomass does not possess the same level of biological activi-
ties as aerobic bacteria.  Therefore, in order to optimize
the utilization of each supporting surface area in a fixed-
film system, every effort should be made to increase the
"aerobic" portion of the biomass.  This, of course, can be
accomplished by increasing the oxygen availability in the
treatment system.  In fact, in a recent study using pure
oxygen in the RBC operation, Huang and Bates (6) found that
the use of pure oxygen was able to phenomenally increase the
aerobic biomass accumulation on each unit disk surface area.
Unfortunately, that study only demonstrated the qualitative
evidence of the increased aerobic biofilm development; the
quantitative relationship between the oxygen flux and the
aerobic fixed-film accumulation was not established.
     Another important parameter complicating the dynamic
behavior of the fixed film development is the sloughing of
biomass.  Although it is known that sloughing is caused by
the hydraulic shear at the biofilm layer, it is not clear as
to the general frequency and exact location (or interface)
that the sloughing would normally take place.  It is specu-
lated that the biofilm sloughing is most likely to take place
at the aerobic-anaerobic interlayer, where the production of
acidic metabolites by anaerobes is likely to weaken the
binding strength of polysaccharides in the biofilm establish-
ment.
     From the above discussion, it is clear that most of the
fixed film biological treatment system being used today (such
as RBC, trickling filters and aerobic fluidized beds, etc.)
have not been optimized to utilize their valuable surface
areas to support exclusively aerobic biomass due to a lack
of oxygen.  Because of the oxygen limitation, several in-
vestigators (7-11) have found that the rate of substrate
removal cannot be further increased once the effective
thickness of the biofilm reaches a. certain level.  Undoubted-
ly, if the entire layer of biofilm is made of aerobic bac-
teria, the rate of substrate removal should continue to in-
crease with the biofilm thickness as long as the substrate
concentration is not limiting.  On the other hand, if the
biofilm is also composed of anaerobes, then the rate of sub-
strate oxidation may not linearly increase with the thickness
                              464

-------
of the biomass.
     At the present time, our understanding of the dynamic
behavior of the biofilm development and sloughing is quite
meager.  The relationship between the thickness of aerobic
fixed film and the available substrate/oxygen concentrations
has not been quantitatively established.  This paper presents
a rational modelling approach for developing equations which
may be used to predict the development and sloughing of fixed
films under defined conditions.  Also, the specific experi-
mental tests which are required to generate pertinent model-
ling parameters are discussed in detail.
     In order to fulfill the modelling requirement, specific
experimental tests have been designed to generate the follow-
ing data:
     1) to quantitatively relate the substrate removal rate
with both the aerobic and anaerobic biofilm development under
some specially-designed operating conditions; 2) to assess
the impact of substrate and oxygen concentrations on the de-
velopment of biofilm thickness and its impact on waste stabi-
lization rate; and 3) to estimate the attenuation of dissolved
oxygen and substrate concentrations across the biofilm layer
and then to identify the interface at which biofilm sloughings
are most likely to occur.

MODELING APPROACH

     In order to develop fixed—film biological growths in
well defined conditions, several annular reactors need to
be fabricated.  Each reactor will consist of a stationary
outer cylinder and a rotating inner impeller, as shown in
Figure 1.
     The annular reactor has the advantages of providing a
constant shear throughout the stationary supporting surface
as well as allowing direct insertions of oxygen probes and
sampling capillaries during testings.  Therefore, this type
of reactor will .allow generation of experimental data for
the development of a model to correlate the substrate removal
rate with biofilm buildup.  This system will also provide
data to establish the attenuation of substrate and dissolved
oxygen (DO) through the biofilm layer at various substrate
and DO availabilities and then to identify the interface at
which biofilm sloughings are most likely to occur.  A glucose
substrate with adequate minerals and phosphate buffer will
be used as the feed.  The glucose concentration may be ad-
justed to any level in different phases of the experiment
to suit the modeling need.
                               465

-------
    Removable Thin
     Plastic Strip
       D.O. Probes
         Recorders
Sampling
 Capillaries
                                      1/2" S/S Shaft
Substrate
     or N£ Purging
     as Necessary
Removable
 Lid


—•^-Effluent

  Fixed Films
                                                   Mixing
                                                   Impeller
   D.O. Probe
                                                Temp Probe
                                               Fixed Films


                                               _^  Effluent
                                                "^^-

                                                Removable Thin
                                                Plastic Strip
                                    D.O.  Probes
      Figure 1.   A Schematic Diagram of  the Annular Reactor
                             466

-------
     In each testing, the substrate will be added to the an-
nular reactor at a sufficient flow rate to intentionally
maintain a hydraulic detention time of no more than 15 min
so that the growth of suspended biomass can be neglected in
the mathematical modeling.  Before the substrate is added
into the reactor, pure oxygen at various flow rates will be
injected into it and briefly mixed to maintain desirable
dissolved oxygen (DO) levels.  The DO concentrations in both
the mixer unit and the annular reactor will be monitored and
recorded continuously throughout each test.  The speed of
the impeller rotation inside the annular reactor will be
properly regulated, but in no case shall the peripheral
velocity ever exceed 1 ft/sec, which is the upper limit being
used in most full-scale RBC applications.  Because of the
short hydraulic detention inside the reactor, biomass pro-
duction would be limited mainly to the attached biomass.
Hence the variation in suspended solids with time can be di-
rectly attributed to the process of biofilm sloughings.
     The experiment will be initiated by inoculating a small
amount of sewage microorganisms and operating the reactor
in a batch mode until some surface slimes start to develop.
This technique will speed up the initial establishment of
the primary slime layer (12,13,14) in the reactor.  After
the initial primary layer has developed, the reactor will
be switched to the continuous— flow operation with the feed
of a synthetic substrate.  At this stage, the continuous de-
velopment of biofilms and the associated substrate stabiliza-
tion rate will be monitored as frequently as necessary.
The mass balance equation for the substrate in the system is:

     V 2$. - n (c  ^   Ma + Mx   Ms                   f
     V dt - Q (S0-S) -- f- -- Y-   ........ (Eq. 1)
where V = liquid volume in the annular reactor
      S = substrate concentration in the reactor (ML~ )
      t = time elapsed (t)
      Q = feed rate (L3/t)
      S0= influent substrate concentration (ML  )
      Ma= attached biomass growth rate (M/t)
      Mj^ sloughings of attached biomass in the reactor (M/t)
      Ya= attached biomass yield coefficient
      MS= suspended biomass growth rate (M/t)
      Ys= suspended biomass yield coefficient

Because of a short detention time (no more than 15 min) em-
ployed in this study,  the removal, of substrate due to sus-
                             467

-------
pended biomass growth can be neglected in comparison to the
substrate consumption for the attached biomass growth.  Thus,
Eq. 1 may be rewritten as:

     V || = Q  (S0-S) - |j	(Eq. 2)

     The mass balance of the biomass in the reactor, on the
other hand, can be expressed as follows:

     V H = Q  (Xo-X) + Mx	(Eq. 3)

where Xo is the influent suspended biomass concentration and
X is the biomass concentration in the reactor.
     The growth rate of attached biomass can be expressed as:

     Ma = AP ~	, . (Eq. 4)

where A = reactor surface area of the attached biofilm
      P = biofilm volumetric density, and
      Th= attached biofilm thickness
Since the influent suspended biomass concentration is zero,
the rate of sloughing which results in the production of
suspended biomass can be estimated from Eq. 3:

     MX = V || -i- QX	(Eq. 5)

     After substituting Eqs. 4 and 5 for the terms of Ma and
Mx in Eq. 2, the following equation can be obtained:

            =  Q(S0-S) - V |f  Ya - ¥ ff - QX . • • • (Eq. 6)

After Ya has been determined, Eq. 6 can be used to correlate
the substrate removal rate with the biofilm development.
     As the reactor is operated longer and longer, the bio-
film layer inside the reactor will become more and more estab-
lished.  As the biofilm thickness becomes greater, sloughings
will start to occur and suspended solids concentration in the
reactor will increase.  At this stage, the last two terms in
Eq. 6 cannot be neglected any more.  The thickness of the at-
tached biomass will become a function of the sloughing rate.
Thus the effluent suspended solids concentration  (X) and ^
inside the reactor over a short defined test interval
must be determined to calculate the rate of change of bio-
film thickness.  The calculated value will then be checked
against the actual measurement from the inserted  thin plastic
                              468

-------
strip during the course of the experimental study.
     When steady state conditions of Eqs. 2,3 and 6 are
reached, the thickness of the attached biomass, the substrate
concentration and the biomass concentration in the reactor
will be constant.  Equation 6 becomes:

     Q (S0-S) -f^=0.o .............  . (Eq. 7)
                ra
The substrate removal rate for a constant biofilm thickness
can then be calculated as:

     Q (S0-S) =|^   ......... . ......  (Eq. 8)
                Ia
Note that QX is actually the amount of biomass that has been
sloughed off in the reactor and can be expressed as the prod-
uct of specific "yield" rate and the overall attached biomass
(/ii.APTh).  Thus the substrate removal rate can be related to
the thickness of the attached biomass as follows:

     Q (S0-S) = (f^).Th   ... ..........  (Eq. 9)
                 xa
     The specific yield rate, ju, is a function of the sub-
strate concentration, oxygen concentration as well as other
environmental factors.  A model similar to the Monod equation
and Michaelis— Menten relationship may be used:
where J%ax = max:>-mum "yield" rate
      Ks   = Monod half velocity concentration and
      S    = limiting substrate concentration.
After Umax and Ks have been experimentally determined, the
substrate removal rate for a given. Th can be calculated.  The
calculated value will then be compared to the actual measure-
ment in the test.
     It is expected that the active biofilm thickness is
dependent on both the oxygen and the substrate concentrations
in the reactor.  Various oxygen and substrate concentrations
will be employed in the test to evaluate the dependence of
the biofilm thickness on these two parameters and then to
establish the quantitative relationship between the substrate
removal and the biofilm thickness.
     It must be noted that the theoretical considerations re-
presented by the aforementioned equations will hold true only
if the biofilm establishment in the reactor is either com—
                               469

-------
pletely aerobic or anaerobic.  A mixed aerobic-anaerobic bio-
film system will complicate the calculations since their bio-
mass yield coefficients and biofilm densities are not the
same.
     After the biofilm is well developed and the relationship
between the substrate removal rate and the biofilm develop-
ment has been established, the concentration in the influent
feed will be progressively increased to effect the buildup
of a thicker biofilm until it reaches a critical point at
which DO concentration becomes a limiting factor.  At this
point, a complete aerobic condition will not prevail through-
out the biofilm layer.  Thus, some dark-color anaerobic bio-
mass will develop at the biofilm's underlayer and sloughing
will occur at a much greater rate.  At this stage of opera-
tion, five DO microelectrodes as described by Whalen, et al.
(15) and an equal number of capillary sampling tubes will be
inserted into different depths of the biofilm layer from the
reactor's cylindrical wall.  The positions of insertion will
be close together along the removable thin plastic strip so
that at any particular moment, the monitored DO and substrate
concentration profiles can be related to the biofilm thick-
ness.  During each separate testing, the DO and substrate
concentrations in the influent .feed will remain the same,
while in the bulk solution the DO concentration will be
monitored continuously and the substrate concentration de-
termined as frequently as necessary.  The DO monitorings
will be continuously recorded throughout the test period to
evaluate an expected "sigmoidal" pattern of the DO variations
due to periodic sloughings of biofilms.
     The concentration profile of the substrate can be ob-
tained by establishing a mass balance equation for a differ-
ential thickness in the attached biofilm (1,2,16).  A simple
experimental first—order decay equation may be assumed for
the limiting substrate, as follows:
               —V Th
     STh = S±10 Ksin	  (Eq. 11)'

     DOr^ = DOi10~k°Th	o . . . .  (Eq. 12)

where D0j[ and S± are the DO and substrate concentrations at
the biofilm surface; k0 and ks are the attenuation rate con-
stants for DO and substrate concentrations across the biofiM;
and Th is the thickness of biofilm at the point of measure-
ment.  If the substrate is not limiting, the zero order decay
equation will be used:
                              470

-------
     STh = Si-ksTh	 (Eq. 13)

     DOTh = DOi-k0Th	 . (Eq. 14)

where kgand k^ are the decay rate for the substrate and oxy-
gen, respectively.  By keeping the rotating impeller at a
reasonably high speed (so that the Reynolds number is in the
turbulent range), the values of DO-j^ and Si will be close to
those existing in the bulk solution.  Through an adequate
number of repeated determinations, the experimental data
should be able to allow for estimations of k0 and ks.  Also
attention will be given to correlate the interface of slough-
ing with the DO and substrate profiles to establish the most
likely location that the sloughings would normally take place.
From the established ko and ks values, the attached biomass
accumulation in any waste treatment system may be predicted
from the available DO and substrate concentrations using
Eqs. 11 through 14.  The validity of such a prediction will
be verified in the testing by systematically changing the DO
and substrate concentrations in each study.

SUMMARY

     The  purpose of this paper is to present a logical ap-
proach to develop mathematical models for assessing the
fixed-film buildup and sloughings in a biological waste
treatment process and their resultant impacts on the rate of
waste stabilization.  Careful experimental testings are now
being conducted at the University of Missouri-Rolla to gener-
ate pertinent parameters associated with the modeling.  Be-
sides, these testings will also be used to verify the validity
of the proposed models.  It is hoped that with a better un-
derstanding of the fixed-film system, future designs of RBC
and aerobic fluidized-bed biological reactor can be optimized
by eliminating the oxygen availability as the most common
rate-limiting factor.  This would result in a significant
reduction of the reaction time requirement, thus achieving
a corresponding capital saving associated with the tankage
construction.

REFERENCES

1.  Williamson, K., and McCarty, P.L., "A Model of Substrate
    Utilization by Bacterial Films."  Journal of Water Pollu-
    tion Control Fed., 48, 9  (1976).
                              471

-------
2.  Williamson, K., and McCarty, P.L., "Verification Studies
    of the Biobilm Model for Bacterial Substrate Utilization."
    Jour. Water Poll. Control Fed., _48, 281  (1976).
3.  Mehta, D.S., Davis, H.H., and Kingsburg, R.P., "Oxygen
    Theory in Biological Treatment Plant Design," Jour, of
    the San. Eng. Div., ASCE, Vol. 98, SA3, P. 471 (1972).
4.  Owen, D.T. and Williamson, K.J., "Oxygen Limitation in
    Heterotrophic Biofilms," Proceedings of the 31st Purdue
    Industrial Waste Conference, Purdue University, West
    Lafayette, IN. (1976).
5.  Tropey, W., et^ _al_., "Effects of Exposing Slimes on Ro-
    tating Discs to Atmosphere Enriched with Oxygen," Pro-
    ceedings of the Sixth International Conference on Ad-
    vances in Water Pollution Research, P.  405 (1973).
6.  Huang, J.C. and Bates, V.T., "Comparative Performance
    of Rotating Biological Contactors Using Air and Pure
    Oxygen," Jour. Water Poll. Control Fed. Vol. 52, No. 11,
    pp. 2686-2703 (November 1980).
7.  Kornegay, B.H. and Andrews, J.F., "Characteristics and
    Kinetics of Biological Fixed Film Reactors," J. Water Pol-
    lution Control Fed. Vol. 40, R460  (1968).
8.  Maier, W.J. et al., "Simulation of the  Trickling Filter
    Process," Jour, of the San. Eng. Div.,  ASCE, Vol 93, No.
    SA4 (August 1967).
9.  Tomlinson, T.G. and Snaddon, D.H., "Biological Oxidation
    of Sewage by Films of Micro-Organisms," Air and Water
    Poll. Int'l. Jour., Vol. LO, 865 (1966).
10. Hoehn, R.C. and Ray, A., "Effects of Thickness on Bacterial
    Film," Jour. Water Poll. Control Fed.,  Vol. 45, 2302
    (1973).
11. Sanders, W.M., "Oxygen Utilization by Slime Organisms in
    Continuous Cultures," Intl. Jour. Air & Water Poll., Vol.
    10, P. 253 (1966).
12. Baier, R.E., Shafin, E.G. and Zisman, W.A., "Adhesion:
    Mechanisms that Assist or Impede It," Science, 162, 2,
    1360  (1968).
13. Mariappan, M., "Generation of Sulfide in Filled Pipes,"
  •  Ph.D. Dissertation, Dept. of Civil Engr., Univ. of MO-
    Rolla, Rolla, MO (1976).
14. Characklis, W.G., "Biofilm Development  and Destruction,"
    Final Report, Electric Power Research Institute RP 902-1,
    Palo Alto, CA (1979).
15. Whalen, W.J., Bungay, H.R. and Sanders, W.M., "Micro-
    electrode Determination of Oxygen Profiles in Microbial
    Slime Systems," Environ. Sci.  & Tech., Vol. 3, 12
                               472

-------
    (December 1969).
16» LaMotta, E.J., "Internal Diffusion and Reaction in
    Biological Films," Env. Sci.& Tech,, 10, 765 (August
    1976).
                              473

-------
                  EVALUATION OF RBC SCALE-UP
          Yeun C. Wu, Department of Civil Engineering,
          University of Pittsburgh, Pittsburgh, Pa.
          Ed, D. Smith, Environmenatl Division, U.S.
          Army Construction Engineering Research Lab,,
          Champaign, IL.
          Chiu Y. Chen, Department of Environmental
          Engineering, National Chung Hsin University,
          Taichung, Taiwan
          Roy Miller, Environmental Health Branch,
          U.S. Army Environmental Hygiene Agency
INTRODUCTION

     Development and application of Wu's model for the pre-
diction of soluble BOD removal in rotating biological conr
tactor (RBC) vastewater treatment systems have been dis-
cussed in detail elsewhere (1, 2).  The model is capable of
both precisely"estimating the treatment efficiency of RBC
systems and successfully determining the size of the treat-
ment plant if the design conditions such as influent soluble
BOD concentration, vastevater temperature, number of RBC
stages, and % BOD removal requirement are known.  Therefore,
the model is very useful for predicting performance and can
be easily applied for engineering design purposes.

     Presently,  little is known about the applicability of
pilot plant data for full-scale design.  As a result, there
is an essential  need to investigate RBC scal4e-up under vari-
ous operating conditions.  This study was primarily designed
to determine the influence of wastewater temperature on pro-
cess scale-up.  Wu's model is capable of performing this
important task.
RBC MODEL

     Wu's model was developed on the basis of full-scale RBC
data reported by many researchers (3-10).  The model is
given as follows:
                              474

-------
                             a0.5579
                             q
                Exp.0.32N L°-6837 T0.2477

in which

            F = fraction of influent soluble BOD remaining
                in the effluent,  %

            q = surface hydraulic loading,  gpd/ft^

            N = number of RBC stages

           LQ = influent soluble  BOD concentration,  mg/1

            T = wastewater temperature,  °C

     Eq. 1 describes the relationship between I BOD
removal/remaining (F) as function of process variables
including q, LQ> N, and T.  For instance,  the effect of
hydraulic loading, q, on F under  varying influent soluble
BOD concentrations, Lo, and number of RBC  stages, N, at tem-
perature T = 25°C is shown in Figure 1.   It can be seen in
Figure 1 that the F value always  decreases  as q, LQ, and N
increase.  The influence of stage number,  N on F under
differing conditions for q and Lo at T = 25°C is illustrated
in Figure 2,  Obviously Figure 2  shows that the F value
decreases profoundly as a result  of either  decreasing q or
increasing both Lo and N.  However, F changed only slightly
after N was greater than 6.  This result becomes very obvi-
ous when Lo is high and q is low.  The relationship  between
T and F under varying Lo, q, and  N is depicted in Figure 3.
It is apparent from Figure 3 that for all  conditions inves-
tigated here, F appears to become independent of T after T >
15°C.  In addition, Figure 3 also shows  that the influence
of T on F is less significant when both Lo  and N are high
and q is low.

     The reliability and accuracy of this model has  been
extensively studied by using more than eighty data sets
obtained from the operation of six full-scale RBC plants
(2).  The maximum error which results from  the use of Wu's
model was found to be HH 4.64% in  terms of  the efficiency of
BOD removal.
                              475

-------
DEVELOPMENT OF RBC SCALE-UP FACTOR

     Past experience indicates that most operating full-
scale RBC treatment plants were designed according to cri-
teria generated from small-scale pilot plant studies.  It is
unknown whether the pilot scale data are adequate for the
process engineer to physically size the full-scale plant.
An early work of Famularo et al. predicted a 101 reduction
in organic removal in a 4 stage RBC system if the disc size
increaesd from 2 m to 6 m (11).  Further, Murphy and Wilson
have recently demonstrated that the removal of COD is
approximately 15% lower for a 2 m RBC than for 0.5 m RBC.
They propose that an additional 10% increase be made in
scaling up from a 2 m RBC to 3.5 m RBC at 17°C (12).  How-
ever, the effect of temperature on RBC scale-up was not
reported by Famuaro et al. or Murphy and Wilson.

     According to Murphy and Wilson, the inverse relation-
ship between disc size and substrate removal efficiency
could be explained by a combined physical and biological
effect.  They have speculated that as the RBC disc diameter
is increased, the liquid film on the biomass is exposed to
the atmosphere for longer times resulting in greater sub-
strate depletions and lower substrate concentrations in the
liquid layer.  Under conditions of low substrate concentra-
tions, when substrate availability and diffusion is limit-
ing, total removal efficiency declines as disc size
increases.  Another possibility which may produce the
aforementioned result is the operation of the small-scale
pilot unit at a higher rotational speed.  Therefore, the
rate of oxygen transfer from gas phase to liquid phase in
RBC system under the identical hydraulic/organic loading
favors the small unit because of its high rotational speed.

     It is evident from the discussion above that an inves-
tigation of RBC scale-up is necessary for engineering
design, even though some difficulties are encountered due to
a lack of field data and an available mathematical model.
Since this predictive model enables one to correlate the BOD
removal efficiency with the process controlling variables
successfully, the scale-up factor can be determined if both
pilot-scale and full-scale plant data are obtained.  Sixty-
four data sets including influent soluble BOD concentra-
tions, hydraulic loading, wastewater temperature, % BOD
removal, and number of RBC stages produced from seven full-
scale RBC plants, along with sixty-three data sets developed
                               476

-------
from the study of five small-scale RBC units were employed
for the present investigation (13-24).  Actual equations
involved in the development of the scale-up factor are indi-
cated as follows:

                 Fl               Fl
            KL = — =	   	(2)
                 Fr        14.2    q0.5579
                      exp.0.32N L0.6837 T0.2477
                        ^        n       *
and

                            14.2    qO-5579
                      exp.0.32N L0-6837 To.2477
            K2 = — =
                           14.2
                 Fr       o 32N .0.6837 T0.2477
                  r       u'JZW L       T.
                                 o       *
     Lo> 1 >  N, and T in Eqs. 2 and 3 are the system operat-
ing conditions for either a small-scale or a full-scale RBC
plant.  FI represents the measured % BOD remaining and F£ is
the predicted % BOD remaining obtained from the model calcu-
lation at the same conditions as F^.  However, Fr in Eq.  3
is different from F2 because it is calculated at a refer-
enced temperature T* instead of T.  The ratios of F^ to Fr
and F2 to Fr are designated as K^ and K2, respectively.  K2
is theoretically equal to K^, if the results of % BOD
remaining for both field measurement and model prediction
are identical.

     The effect of T* on the relationship between K^ or K2
and T is shown in Figure 4.  It is important to point out
that the theoretical curve (K2 vs T) always passes through a
point where K2 is equal to 1 and T is the same as T*.  K^ is
also a function of T*, that is, K^ increases as a result  of
increasing T*.  But it was decreased as the wastewater tem-
perature T was increased, according to Figure 4.

     The operational curves (K^ vs T) as shown in Figure  4
were constructed using the full-scale RBC plant data.
                               477

-------
Although the K^ values are randomly dispersed, the opera-
tional curves obtained from data analyses utilizing a non-
linear least square method, closely approximate the theoret-
ical curves for different operating conditions.

     Further development of the relationship between Kj_, K£
and T for pilot-scale system (disc size < 6 ft) at T* - 20°C
was made.  The results are illustrated in detail in Figure
5.  A comparison of performance of the pilot-scale RBC with
the full— scale RBC under the same operating conditions is
made with Figure 4-(C) and Figure 5,  The comparison
revealed the operational curve in the former system to be
far below the theoretical curve.  However, the reverse is
found in the latter system.  This result is expected because
the presently employed model was developed based on the
full— scale plant data.

     From the above discussion, it is known that the direct
application of pilot plant data for full-scale design is not
acceptable.  In all cases studied, the K^ value at any par-
ticular temperature, T is always higher in the full-scale
system than in the pilot scale system if the referenced tem-
perature T* is the same.  This phenomenon indicates that the
full-scale RBC plant is less effective if the system is
designed in accordance with the data obtained from a treata-
bility study of a small-scale pilot plant.  As a result of
this observation, the following investigation was aimed to
develop the scale-up factor (SUF).

     Both the operational curves as shown in Figure 4-(C)
and Figure 5 are the lines of best fit, calculated by non-
linear least squares regression with a 95% confidence limit.
From these analyses it is found that at the referenced tem-
perature T* = 20°C, the operational curves for full-scale
RBC plants and pilot-scale RBC plants can be described by
the following two equations:

(K1)Full - 1.5535 - 0.041666T + 0.00075233 T2     ------ (4)

and


-------
depicted in Figure 6.  It is clearly shown in Figure 6 that
the SUF value varies significantly as a function of tempera-
ture.  The scale-up factor increases from 1.067 to 1.227 as
the temperature increases from 3°C to 25°C.  However, a
decrease in SUF was found when the temperature exceeded
25°C.

     It is important to point out that within the tempera-
ture range investigated the maximum scale-up (22.7%) occurs
at T = 25°C and the minimum scale-up (6.7%) takes place at T
= 3°C.  Inhibitory effects due to high temperature begin to
show when T exceeds 25°C.  The relationship between SUF and
T is described as

   SUF = 1.0097 + 0.016206T - 0.00032842 T2       ----- (6)

and is illustrated in Figure 6.

     The following example is given to demonstrate the
method for incorporating the SUF into the full-scale plant
design:

     The experimental data obtained from a small-scale pilot
plant study are (25);

               % Soluble BOD Removal
               Required               = 82% or F = 0.18

               Hydraulic Loading
               in gpd/ft2 (q)         =1.50

               Stage Number (N)       = 4

               Influent Soluble
               BOD Concentration      = 50 mg/1
               Wastewater
               Temperature (T)        = 13.4°C

               Design Flow Rate       = 2 MGD

     Based on the design criteria specified above, the SUF
was calculated by using Eq. 6.  The result is 1.20256.  As
mentioned earlier, the SUF is mathematically defined as:
                                479

-------
                  (Kl)Small


                        14.2    gO.5579
                  exp.0.32N L-     To.2477
         1.1670 - _ * _ S _
                            0.17

For the experimental data:
                           14.2 X qO.5579
         1.1670 - exp.0-32x4(5o)0.6837(13.4)0.2477      	
                                0.17
By solving Eq. 7, the q value is found to be equal to 1.348
gpd/ft2, that is less than 1.50 gpd/ft2 resulting from the
pilot plant study.  The total disc surface required is
1,483,680 ft2 (2,000,000/1.348) instead of 1,333,333 ft2
(2,000,000/1.5).  Additionally, it is important to recognize
that the effluent quality of the full-scale RBC plant will
be slightly less due to the change in hydraulic loading.
The resulting effluent quality is estimated as follows:

                       14.2 (1.348)0-5579
                0.210
Therefore, the soluble BOD remaining in the full-scale BBC
plant effluent is 50 mg/1 x 0.210 = 10.5 mg/1 instead of 50
mg/1 x 0.18 « 9.0 mg/1.

     Additional calculations show the difference in hydaulic
loading with and without the considering scale—up factor and
are listed in Table I.  The table clearly indicates that the
reduction of hydraulic loading is greater when both T and q
used for the operation of pilot plant are higher.
                                480

-------
                            Table I

          Comparison of Hydraulic Loading Calculated
            With and Without the Inclusion  of  SUF**
Operating Parameters
L0(mg/l)
(1)
45
64
60
53
49
57
73
N
(2)
4
4
4
4
4
4
4
m { Q¥* \
(3)
15.1
16.8
20.6
23.9
24.9
17.4
24.4
F
(4)
0.18
0.20
0.23
0.16
0.21
0.18
0.15
Hydraulic Loading
without
SUF
(5)
2.0
3.0
4.0
2.0
3.0
2.0
3.0
with
SUF
(6)
1.54
2.51
3.31
1.58
2.39
1.82
2.13
(5)-(6)
(7)
0.460
0.490
0.690
0.420
0.610
0.180
0.870
**Data from Ref. (25)
CONCLUSIONS

     According to this investigation, when a full-scale RBC
plant design is based on the essential controlling variables
of influent soluble BOD concentration, wastewater tempera-
ture, number of disc stages, surface hydraulic loading, and
% BOD removal requirement, the preliminary design criteria
developed from pilot plant study cannot be directly employed
for design.  A scale—up factor should be used.

     This factor was successfully determined by the model
proposed by Wu et al.  Its relation to wastewater tempera-
ture was mathematically formulated by conducting non-linear
least squares regression analysis on both full-scale and
pilot-scale data previously reported by other investigators.
It is apparent that the process scale-up increases from
                              481

-------
1.067 at T = 3°C to 1.227 at T = 25°C.  However,  a decrease
in the scale—up factor was found when the temperature
exceeded 25°C.

     This study shows the effect of process scale-up on the
selection of hydraulic loading for full—scale design is sig-
nificant when the wastewater temperature and hydraulic load-
ing determined during the pilot plant study are high.

     It is necessary to mention that the results of this
study are valid only for the treatment of municipal wastewa-
ter by mechanical drive RBC and bio-oxidation of carbona-
ceous organic material in the RBC system occurs under oxygen
sufficient conditions.
                              482

-------
 REFERENCES
 1.   Wu,  Y.  C.,  et  al.,  "Modelling  of  Rotating  Biological
     Contactor  Systems," Biotechnology and Bioengineering}
     Vol. 12, p.  2055,  1980.
 2.   Wu,  Y.  C.,  et  al.,  "Design of  Rotating  Biological  Con-
     tactor  Systems,"  In  Press,  Journal of Environmental
     Engineering Division,,  ASCE, November,  1981.
 3.   Antoinie,  R. L.,  and  Koehler,  F.  J., "Application  of
     Rotating Disc  Process to Municipal Wastewater  Treat-
     ment,"  EPA Project  No. 17050 DAM, Autotrol Corporation,
     Milwaukee,  Wisconsin, 1971.
 4.   Sack, W. A., "Evaluation of the Biodisc Treatment  Pro-
     cess for  Summer Camp Application," EPA Project No.
     17010 EBM,  University of West  Virginia, Morgantown,
     West Virginia, 1973.
 5.   Clark,  et  al., "Performance of a  Rotating  Biological
     Contactor  Under Varying  Wastewater Flow,"  Journal  of
     Water- Pollution Control  Federation,  Vol. 50, p.  896,
     1978.
 6.   Hao, 0., et al.,  "Rotating Biological  Reactors Removal
     Nutrient - Part 1,"  Water- and  Sewage Works, Vol. 122,
     p. 10,  1975.
 7.   Antonie, R.  L., "Fixed Biological Surfaces," Wastewater
     Treatment,  CRC Press, Cleveland,  Ohio,  1976.
 8.   Antonie, R.  L., Kluge, D. L. ,  and Mielke,  J. H.,
     "Evaluation of a  Rotating Disc Wastewater  Treatment
     Plant," Journal of Water Pollution Control Federation,'
     Vol. 46, p.  498,  1974.
 9.   Malhortra,  S.  K.,  and Williams, T. C.,  "Performance of
     a Biodisc  Plant in a Northern  Michigan Community," In
     Proceedings of the 48th Annual Conference  of the Water
     Pollution  Control Federation,  Miami, Florida,  1975.
10.   Borchardt,  J.  A.,  "Biological  Wastewater Treatment
     Using Rotating Discs," Journal of Biological Waste
     Treatment,  Wiley Interscience, New York, p. 131, 1971.
11.   Famularo,  J.,  et  al., "Application of  Mass Transfer  to
     Rotating  Biological  Contactors,"  Presented at  the  49th
     Annual  Conference of the Water Pollution Control
     Federation,  Minneapolis, Minnesota,  1976.
12.   Murphy, K.  L., and  Wilson,  R.  W., "Pilot Plant Studies
     of Rotating Biological Contactor  Treating  Municipal
     Wastewater," Report  SCAT-2, Environmental  Protection
     Service, Environmental Canada, July  1980.
                              483

-------
13.  Hltdlebauch, J. A., et al., "Full-Scale Rotating Bio-
     logical Contactor for Secondary Treatment and Nitrifi-
     cation," in Proceedings of the First National
     Symposium/Workshop on Rotating Biological Contactor
     Technology, Champion, Pennsylvania, February 1980.
14.  Dupont, R. R., and McKinney, R. E., "Data Evaluation of
     a Municipal RBC Installation, Kirksville, Missouri," in
     Proceedings of the First National Symposium/Workshop on
     Rotating Biological Contactor Technology, Champion,
     Pennsylvania, 1980.
15.  Sullivan, R. A., et al., "Upgrading Existing Waste
     Treatment Facilities Utilizing the Bio-surf," in
     Proceedings of the First National Symposium/Workshop on
     Rotating Biological Contactor Technology, Champion,
     Pennsylvania, 1980.
16.  Andeson, E. D., "Performance of Full-Scale RBC Plant at
     Ft. Bragg, North Carolina," Personnel Communication,
     the Office of the Chief of Engineers, U.S. Department
     of the Army.
17.  United States Army Environmental Hygiene Agency, "Phase
     1 - Water Quality Engineering Special Study No. 32-24-
     0116—79 Sewage Treatment Plant Evaluation - Summer  Con-
     ditions for Fort Knox, Kentucky," published by the
     Department of the Army, 1978.
18.  Chow, C. S., et al., "Comparison of Full-Scale RBC  Per-
     formance with Design Criteria," in Proceedings of the
     First National Symposium/Workshop on Rotating Biologi-
     cal Contactor Technology, Champion, Pennyslvania, 1980.
19.  Smith, E. D., et al., "Recarbonation of Wastewater
     Using the Rotating Biological Contactor," in Proceed-
     ings of the First National Symposium/Workshop on Rotat-
     ing Biological Contactor Technology, Champion, Pennsyl-
     vania, 1980.
20.  Khan, A. N., et al., "Rotating Biological Contactor for
     the Treatment of Wastewater in India," in Proceedings
     of the First National Symposiun/Workshop on Rotating
     Biological Contactor Technology, Champion, Pennsyl-
     vania, 1980.
21.  Orwin, L. W., and Sieenthal, C. D., "Hydraulic and
     Organic Forcing of a Pilot Plant Scale RBC Unit," in
     Proceedings of the First National Symposium/Workshop on
     Rotating Biological Contactor Technology, Champion,
     Pennsylvania, February 1980.
                               484

-------
22.  Torpey, W. N., et al., "Rotating Biological Disc Waste-
     water Treatment Process - Pilot Plant Evaluation," EPA
     Project No. 17010 EBM, Environmental Protection Agency,
     Washington, D.C., 1974.
23.  Williams, et al., "The Gladstone, Michigan Experience:
     Performance of a 1.0 MGD RBC Plant in a Northern Michi-
     gan Community," In Proceedings of the First National
     Symposium/Workshop on Rotating Biological Technology,
     Champion, Pennsylvania, February 1980.
24.  Griffith, G. T., et al., "Rotating Disc Sewage Treat-
     ment System for Suburban Developments and High Density
     Resorts of Hawaii," Water Resources Research Center,
     Technical Memorandum Report No. 56, University of
     Hawaii, Honolulu, Hawaii, 1978.
25.  Miller, R. D., et al., "Rotating Biological Contactor
     Process for Secondary Treatment and Nitrification Fol-
     lowing -a Trickling Filter," Technical Report No. 7905,
     U.S. Army Bioengineering Research and Development
     Laboratory, Fort Detrick, Frederick, Maryland, 1979.
                                485

-------
 NOTATIONS


  F «  fraction of influent soluble BOD remaining in the
       effluent, %

 Fj^ *  measured value of F, %

 F2 =  predicted value of F, %

 Fr «  predicted value of F at given referenced temperature,

 Lo *  influent soluble BOD concentration, mg/1

 Kl »  ratio of FI to Fr

 K2 -  ratio of F£ to Fr

  N «  number of RBC stages

  q **  hydraulic loading rate, gpd/ft^

SUF »  scale-up factor

  T «  wastewater temperature, °C

 T* *=  referenced temperature, °C
                                 486

-------
            PART V:  SMALL-SCALE/ON-SITE SYSTEMS
             SMALL WASTEWATER TREATMENT SYSTEMS
               USING SOIL PURIFICATION METHOD
 Masaaki Niimi, Director, Soil Purification Center, Ltd.
 Ueki Bldg., 2-41-8 Kabuki-cho, Shinjuku-ku, Tokyo 160, Japan
INTRODUCTION

     Since night soils were used as fertilizer in agricul-
tural land until 1950's, wastewater treatment in rural areas
has generally been neglected until recently.  Wide use of
chemical fertilizers in recent years, however, prompted the
necessity of rural wastewater treatment in Japan since night
soils are no longer used in agricultural land.
     Under these circumstances, Japan Ministry of Agriculture
started a program in 1977 construct small system wastewater
treatment facilities in rural areas, and adopted to promote
soil purification systems as one of the most suitable methods
of treatment.
     The purpose of the paper is to describe unique features
of the soil purification systems developed in Japan and to
discuss construction, operation, and maintenance ,of the
following systems:
     Enhancement of Treatment by Use of Soil Cover
     In this process, soils are not used as a mere construc-
     tion materials, but they are effectively used as a media
     for supporting microbial life.  In actual installations,
     treatment facilities are constructed underground covered
     by soil layer.  Treatment efficiencies are observed to
     be greatly increased by the use of soil cover in these
     instances, and the ground surface can be used for lawn
     area or other uses for esthetic enjoyment.
                         487

-------
     Underground Trench Soil Purification System
     By installing an impermenable sheet under the trench,
     capillary action of soils in horizontal directions is
     enhanced, thus preventing groundwater pollution due to
     enhanced soil purification in aerobic soil zone.

     There have been already approximately 25,000 installa-
tions of our system in Japan.  These systems utilize eco-
system of soilsphere and are suitable for small system
wastewater treatment in rural areas.  These facilities are
low cost and low maintenance wastewater treatment systems
and would not require extensive pipeline networks such as in
a large-scale central wastewater treatment plant.

 1  Oriental Tradition of Recycling Human Waste to Farmland

     In the Eastern countries including Japan, human excrement
has been utilized for agricultural production for as long as
several thousand years, and this practice still survives even
at present, though less commonly.
     In my paper entitled "Do Joker Process"( ) of last
September, the auther quoted the words of two Europeans who
had referred to this Eastern wisdom" to our shame".  One was
Victor Hugo, in "Les Miserables" published in 1862 and the
other was Dr. H. March,    a German who had visited Tokyo,
then called Edo, around the same period.
     For your reference, Victor Hugo and Dr. H. March said
as follows respectively:
    "Paris Pours twenty-four million francs a year into the
water.  That is no metaphor.  She does so by day and by night,
thoughtlessly and to no purpose.  She does so through her
entrails, that is to say, her sewers.  Twenty-five millions
is the most modest of the approximate figures arrived at by
statistical science.
     After many experiments science today knows that the most
fruitful and efficacious of all manures is human excrement.
The Chinese, be it said to our shame, knew it before us."
     How often do we hear our farmers talk about this manure
being preferable to that manure on account of its fertilising
action being 'more lasting;' yet with all our wise provision
for the future, how far are we now behind the Japanese, who
seem to look always to the next harvest only!   As they manure
for each fresh crop, and the term 'fallow' in our acceptation
is entirely unknown to them, they are forced to distribute
their yearly production of manure equally over the entire
area of their land, which can be accomplished only by sowing
                              488

-------
In drills or furrows, and by top-dressing."
     Also, F. H. King stated in Chapter 9 of his "Farmers of
Forty Centuries on Permanent Agriculture in China, Korea and
Japan" (1911)    as follows:
     One of the most remarkable agricultural practices a-
dopted by any civilized people is the centuries-long and well
high universal conservation and utilization of all human
waste in China, Korea and Japan, turning it to marvelous
account in the maintenance of soil fertility and in the pro-
duction of food.
     The same book quoted the words by fur then, Dr. Arthur
Stanley, Health officer of the city of Shanghai, in his
annual report for 1899, as follows:
     "(•••••) while the ultracivilized Western elaborates
destructors for burning garbages at a financial loss and
turns sewage into the sea, the Chinaman uses both for manure.
He wastes nothing while the sacred duty of agriculture is
uppermost in his mind.  And in reality recent bacterial work
has shown that faecal matter and house refuse are best de-
stroyed by returning them to clean soil, where natural
purification takes place.
     The question of destroying garbage can, I think, under
present conditions in Shanghai, be answered in a decided
negative.  While to adopt the water-carriage system for
sewage and turn it into the river, whence the water supply
is derived, would be an act of sanitary suicide.  It is best,
therefore, to make use of what is good in Chinese hygienee,
which demands respect, being as it is, the product of an
evolution extending from more than a thousand years before
the Christian era".
     To my regret, this excellent Eastern wisdom has been
utterly forgotten in present day Japan which has undergone
ultracivilized  'modernization*.
     The words of Socrates  - "A bad law is also a law."
- still survive in Japan too, even today when about 2,400
years have passed since his time.  Ultracivilized modernized
Japan has not only forgotten this excellent wisdom but also,
on the contrary, has enacted a law to prohibit it.  As a
result, she has extended the life of his famous words into
the 20th century.
     Such being the case, I would like to tell you first of
all that our proposed system for purifing or utilizing rain-
water and waste water/sludge by exploiting the power of the
soil has been developed under various strict restrictions
imposed by this "bad law". (3).
                            489

-------
 2  Japanese Laws and Regulations Negating Her Good
    Traditions, and the Development of a New Soil
    Treatment System Under These Restrictions

     Even under the new Japanese law enforced from the 1st
of last June, our position, consistently advocated for years,
that human waste and bath/kitchen waste water from smaller
numbers of persons be treated of jointly was not accepted.
The only one restriction relaxed by the law is that joint
treatment for 51 or more persons is authorized instead of
the previous 100 persons.  Consequently, if law-asiding
citizens wish to make onsite treatment for a small number
(less than 50) of persons, we are obliged to make equipment
in accordance with this bad law.  In other words, human waste
must first be treated independently, and then, the treated
product must be re-treated together with domestic miscellane-
ous waste water from kitchens and baths using additional
equipment.  This situation is also quite different from that
of foreign countries where joint treatment for a small number
of people is authorized.
     Nevertheless, while maintaining such a strict law for
treating human waste, no law for domestic miscellaneous
waste water is maintained in present day Japan, where rivers,
lakes and seas are abandoned to rapid pollution, to .such as
extent that parts of them are being called "dead"5  'Since I
believe that this miserable situation is already known to
many of you experts in waste water treatment, and also that
they are not the direct main subject of this.paper, I would
like to refrain from going into further details.
     However, under the above-mentioned situation in Japan,
our system mentioned below has hitherto been practiced as
follows:
     For the Soil-Cover type,  (l), any process (e.g., Acti-
vated Sludge process)  or any equipment (e.q., either aerobic
or unaerobic) may be used underneath its cover soil.  Where-
as, for the Underground Trench type, @, this process has
been used mostly for treating domestic miscellaneous waste
water and for the Tertiary Treatment since it is restricted
by law.
     Thence, those who emphatically supported type  (2) were
cities, towns and villages which had resisted the bad law of
the Central Government  (the State).  Currently, the number
of such cities, towns and villages exceeds 60, and. their
fine results are discussed at the National Diet A  'Finally,
the Ministry of Agriculture, Forestry and Fisheries, which
                            490

-------
hitherto had not been at all concerned with waste water
treatment has adopted this process for sewerage in rural
areas.  At present, this process is used for as much as 90%
of the rural area sewerage.
     In the following, I would like to explain this system
by showing you figures.
     (By the way, this system is called "Dojo-Joka" in
Japanese, while the magazine in English published by us
called the "Do Joker System" is a pun on the Japanese words.
So, please allow me to use the term "Do Joker System" in
this paper too.)
             Pebble
               CjSIOcm
                                                Pebble
  Net
                                              	Waste water
                 -Sludge
         Figure 1.
                  (6)
Figure 2.
Note 1.  The cover soil shall be of aggregate structure and
         contain much organic substance.

Note 2.  The net mesh shall be fine enough to support the
         soil but coarse enough to allow soil organisms to
         pass through easily, and shall be installed as
         convexly, as possible.
Note 3.  As fillers, such natural products as river pebble,
         volcanic pebble, and crushed stone as well as even
         plastic waste may be utilized.  Their grain size
         should be S^lOcm.
                            491

-------
Note 4.  The larger the balance between HWL and LWL, the less
         excess sludge is produced.  If the water-covered
         portion is made larger as shown in Fig. 4, nitrogen
         will effectively be removed.
Note 5.  In Fig. 1, the filling rate of the filtering material
         shall be determined according to the nature of the
         waste water.  The bigger the filling rate, the
         better the decomposing rate of organic substances
         is but the harder the removal of the excess sludge.

Note 6.  The equipment shown in Fig. 2 may provide tertiary
         treatment by changing the air diffusion method for
         the 2nd and tertiary treatment.  The air diffusing
         pipe is placed either above or under the grate, but
         in either case it should diffuse big air bubbles to
         prevent clogging.

 3  Do Joker System as An aerobic Fixed-Film Biological
    Process

     The equipment shown on Fig. 1 is being utilized as a
settlement tank,septic tank,sludge condensation tub, sludge
storage tank, and a pumping tank, and characteristically gener-
ates no scum on the waste water surface.  If larger pebbles
of 7 ^ 10cm are selected for filling, the sludge filling the
gaps between the pebbles can easily be scooped up by lowering
the water level.  Usually, the pebble layer is as thick as
approx. 50cm.

 4  Do Joker System as Aerobic Fixed-Film Biological Process

     The equipment shown in Fig. 2 is being utilized as
secondary treatment equipment, tertiary treatment equpment,
denitrodizing equipment, and purification equipment for river
water or other slightly polluted water, and is characteris-
tically of slim structure with the pebble layer as deep as
150 'V 300cm  and as wide as approx. 100cm.  (The slim
structure is possible thanks to the non-generation of scum.)
A typical example is the rural sewerage of Wadayama Town
where it is installed under roads.  The grain diameters of
pebbles are 3 ^ 7cm being a bit smaller than anaerobic
filtering materials.  Pebbles of cbout these sizes are se-
lected because the peeling of the biotic film is easily
solved by sending large quantities of air into the diffusing
pipe.  The clogging problem is usually met by using pipes
                             492

-------
diffusing big air bubbles, but,  in  some .cases,, by  combining
an airlift.pump or other oxygen  supplying method other  than
.the air.diffusing pipe system.
     For raw water whose BOD is SOppm or less, the forced
oxygen supply system is omitted.
     We .have consecutively succeeded in the last two years
in hatching and breeding salmon fry by purifying polluted  .
river water (approx. 50 ppm BOD) in Metropolitan Tokyo using
only the equipment shown on Fig. 2.  The hatching/breeding
is scheduled to be continued for another five years.
                          Soil
                           J— Pebble ;
                          	Soil structure / •

                  Figure 3.   Soil Structure\8A
                             493

-------
      Table 1.  Jidayubori Park River Water Analysis Table
Station:  A = Raw Water, B = First Settling Tank, C = Discharge
Dace

11. Nov.


18-Occ.


20. Jan.


12. Feb.


19. Mar.


8. Apr.


7. May


9,Jun.


6. Aug.


lO.Jul.

Station
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
Trans-
parency
30<
30<
30<
4
10
30<
2.5
12.0
30<
12.5
16.0
30<
30<
5
7
30<
30<
30<
30<
30<
30<
30<
30<
30<
30<
30<
30<
30<
30<
30<
PH
1.5
1.4
7.1
7.3
7.4
7.4
7.6
7.5
7.7
7.1
7.3
7.1
7.6
7.5
7.6
7.7
6.8
7.2
7.5
7.2
7.3
7.5
7.0
7.5
8.3
7.6
7.7
7.9
7.1
7.5
BOD
7
13
9
6
5
1
23
13
3
30
25
3
12
13
3
17.4
29.5
34
12.0
9.8
3.3
9.7
8.0
3.8
8.2
7.2
3.3
5.5
4.5
2.9
COD
9
9
7
12
11
6
51
13
5
20
18
7
23
18
7
13.6
14.5
7.2
10.2
9.2
5.8
10.2
10.4
7.6
10.4
10.0
6.8
11
11
6
SS
9
2
9
200
52
8
74.6
40
1
62
29
1
230
54
0
8
8
-
8
9
4
8
11
18
9
11
2
3
6
-
DO
8,4
7.5
8.8
5.0
5.9
10.9
7.1
7.3
11.5
7,2
7.8
9.9
6.T
6.1
10,1
8.8
1.4
8.7
8.4
7.4
9.2
7.6
5.0
9.0
7.0
5.3
5,6
7,7
8.7
8,6
Coliforra
bacillius
940
1,100
58
1,000
1,000
0
560
520
14
2,480
2,000
10
3,200
3,300
6
_
-
-
_
-
-
„
-
-
21,000
7,700
350
17,000
67,000
130
                                494

-------
                         Figure 4-(a)
                                            I Net filler lank(3Qcmx3

                       f,                    2 Net for filtering solids
                         r  V' , >, / V {,.,.,      3 Impertncabk sheet
                         t j ('* c^iLf *  ' / /l  f
                                            5 Pipe

                                            6 Cruwct with dbittcier of 5-8 cut

                                            7 Pulyoihylone nci

                                            8 Karlh miscd with pcarlilc (opilbry sail)
                            0 10 20 30 cm
                         Figure 4-(b) (9)
 5  Utilization of Fluctuation in Water Level within Filtering
    Material of Fixed  Film Biological  Process

     Fig.  3 is an enlarged portion  depicting the  relationship
between  the covering soil and the contact filtering  material
of fixed biotic film under it which is the biggest  feature
of the Do Joker System.
     Now let me explain  its features.
     The first feature is that the  cover-soil layer  on the
equipment is 20 ^ 50cm thick and is of continuous structure
covering both the inside and otuside of the tank.  This is
based on an idea for enabling earthworms or other advanced
                               495

-------
soil creatures to be most skillfully utilized in the waste
water treatment system.  Even creatures living in the soil
space outside the equipment can be utilized for decomposing
the sludge. (See Chapter 1.)
     The second feature is that the boundary between the soil
and the pebbles is of convex structure.  Both ends are sex
at positions lower than the wall top.  This structure allows
the polluted water in the tank to move either into the surface
soil or into the soil layer outside the equipment by capil-
liary action.  (This structure will be explained in the next
chapter.)  Thanks to this structure, we utilize the natural
principle that soil organisms and microbes suited to the
polluted water to be treated, propagate themselves rapidly in
the soil.  One example is the utilization of hemolytic
bacteria in the soil when treating bloody waste water.
     The third feature is the structure which causes fluctua-
tion in the water level within the pebble layer.
     Fig. 3 shows a fluctuation of 5cm in the water level
illustrating that the water level drops due to capilliary
action during the night when the equipment is not in use.
However, if designed according to the long canal system,
there would be a difference in water levels at the inlet and
the outlet due to the resistance of the pebble layer.  It
would then be possible to have a fluctuation of approx. 10 ^
20cm in the water level occur at least twice a day by design-
ing accordingly.
     Further, where the water is supplied intermittently by
a pump under the trench system shown on Fig. 4, if the pump
is installed in the equipment shown on Fig. 2, the LWL can
be lowered indefinitely.  The boundary face between the soil
and the pebbles is made concave for the additional purpose
of allowing the septic gas to go through more easily, and of
utilizing plant roots (especially root hair) as the carbon
source during denitrodization, and of not having the rain-
water flow into the tank.

 6  Underground Trench Soil Purification System

     At a first glance Fig. 4 may be thought to be not much
different from the unarmored trench system, but actually it
is different.  The initial idea of laying impermeable sheet
on the bottom of the unarmored trench in order not to have
the pollutant permete by gravity is based on the following
point.
                             496.

-------
     Organic substances are most effectively decomposed when
the three biotas of vegetable roots (GL to - 100cm), soil
microbes (GL to - 50cm) and soil creatures (GL to - 30cm)
are participating comprehensively.  The fear of groundwater
pollution cannot be removed by the conventional trench system
under which the polluted water and sludge are allowed to
enter into the soil locating it more deeply.
     Furthermore, since there are many cracks, aqueducts, and
big gaps in the soil, the gravity permeatation method by
which the pollutant passed through only the big gaps is not
appropriate due to the fear of probable ground water pollu-
tion.  Only a purifying method making use of capilliary
action which ensures that absolutely no pollutant goes
through big gaps can remove the fear of ground water pollu-
tion.  Because, being quite different from the gravitation
method in which water permeats under positive pressure and
saturation, the capilliary action enables the polluted water
to pass, at a certain planned permeating speed, through the
soil less than 50cm from ground level where the biotic ac-
tivities are active under negative pressure and unsaturated
conditions.
     Also, if compared to the sprinkling system which cannot
treat much water per unit area, 1m of the trench can treat
100£ per day - about five times - thanks to the difficult-to-
dos portion of the trench wall near ground level. A detailed
scientific explanation of this phenomenon, however, has yet
to be clarified.  One thing I Have never found over the past
20 years of our study Is that even polluted water of high
BOD density (as high as 1,000 ppm or more) does causes un-
expected clogging.  I feel this is attributable to action of
earthworms or other large-sized soil creatures, and if
combined with the re-use for lawns, etc. of the domestic
waste water and rainwater, this would become the most
practical water treatment/storage system.
     The surface of the filtering material shown on Fig. 4
is a fixed film biological process under both anaerobic and
aerobic .conditions which, if the system is home-sized, has
successive water level fluctuations 5 ^ 6'times per day.
Though it has not been fully clarified what role this plays
in denitrodization, I presume that, according to the line
meter test using the primary treated water (actual results
were 95% or more COD, SS and 65% T-N(10)), a method for
heightening the removal rate of T-N may be a carbon source
supplying system only.  Since I obtained a 95% removal rate
by adding methyl alcohol, it is a problem in Jpaan^11) as to
                               497

-------
capilliary moisturing trenches are laid out at 3 ^ 4m
intervals, no sprinkling will be necessary for the lawn as
has been attested in California.
     With respect to the relationship between the soil thick-
ness and plants, the utilization of the results of a vegeta-
tion study on artificial ground will suffice.  Even tall
trees may be planted in soil as deep as 80cm.  The growing
speed of crops on this structure gave a test result of 2 ^ 3
times of ordinary soil when waste water from a pig farm had
been used.C  )  No humidity hazard for crops is found when-
ever the waste water level is GL - 60cm according to the
study results.
     The environmental pollution preventive function of this
structure is acknowledged by all those who have seen its
actual results as far as the non-proliferation of odor,
bubbles, human pests and other readily identifiable effects
are concerned.  However, the function acknowledged as most ,
practical is the non-proliferation measures taken against
pathogenic bacteria (viruses) and NOx which are an invisible
environmental pollution problem hard to detect by the senses.
     While the conventional septic activated sludge process
system needs bubble-preventing devices, this system needs no
such thing.  Further, diffusion into the air of fine droplets
caused by exploding bubbles on the water surface is simply
and completely solved by aereating the soil.
     Furthermore, the over-jeneration of NOa and the fear of
N02 diffusion into the air with consequent adverse affects
or the human body, both of which are the biggest demerits of
the F.F.B.P., can be solved through absorption by the soil
and oxidization into NOs.  The biggest reason why the Do
Joker System is used in 90% or more of the rural sewerage
systems in Japan is the completeness of the environmental
pollution preventative measures as such.  Their completeness
is attested by actual examples of its use under a busstop
waitingroom, the lawn of an outdoor eating place, a road, a
flower bed in front of a railroad station, and in the middle
of a housing complex.  Its deodorising system needs no ex-
cessive power, activating carbon, acid, alkali or heating.
Odors from sludge treatment equipment, the covered chamber
of the rotating biological contactor process, and the filter-
ing bed of the sprinkling water could be solved by the Do
Joker System under which the air is pressurized into gaps in
the pebble layer at a pressure as low as that of a ventila-
tion fan.  The design speed is 300m3 per m2 of ordinary
soil.  Before this amount is increased to 1,000m3, the soil
                                498

-------
how to increase the removal rate using natural soap instead
of synthetic cleansers.

 7  Environmental Measures and Environmental Pollution
    Preventive Function

     The point where the Do Joker System a differs from all
other waste water treatment technologies is that environ-
mental measures can be reasonably combined with it, and at
the same time, it can so completely prevent environmental
pollution that no maintenance costs are caused.
     Firstly, with respect to environmental measures, the
surface of the equipment may be utilized.  The point in
common between Figs. 1 through 4 is that a soil layer as
thick as several tens of can covers the equipment.  Therefore,
if care is taken to grow plants over it, the facility itself
will become a green area necessiating no buffer green zone
around it usually, a thin soil layer needs sprinkling
with wates for plants to grow.  This structure, however,
needs no sprinkling at all because the waste water level is
several tens of cm below and its surface is designed so as
to have water supplied by capilliary action.  Usually, the
capiliary water extends as far as approx. 200cm.  Then, if
                              499

-------
nature shall be determined by experiment.  Anyway it is
necessary to satisfy such contradicting functions as
aerability, water preservation and absorptive capacity.
Since they vary greatly according to the soil, construction
is currently performed by determining these factors through
individual experiments.  Care not to make the soil too dry,
and the technique for mixing perlite, comport, etc. are
important.

 8  The Production of Excess Sludge

     A feature of the trickling filter method is that it
produces less excess sludge than the activated sludge method,
the contact aeration method, or the rotating biological
contractor method.  If, in addition to this normal feature,
the F.F.B.P. filter medium in the soil and under the water
surface has a continuous structure as in the Do Joker System
(Figs. 1 to 4), the question is, what biota will be formed?
This is not yet understood in detail.  It is said that in
the F.F.B.P. large-sized Metazoa, which are not found in
activated sludge, live in the membrane to form a wide variety
of biological groups.  The formation of various biological
groups can be easily hypothesized because a net-covered soil
layer of 20 to 50cm thick is over a gravel layer, which is
on the fixed biomembrane in water and about 20cm above the
surface, offering soil organisms an area for living.  Only
a report of rise and fall of soil organisms within the in-
stallations was presented by a Japanese researcher of
earthworms, Yoshio Nakamura, at the Darwin Centenary,Swipo-
sium on Earthwoarm Ecology in cumdria U.K. last yeari  *6ut
the report shows a profile of soil animal ecology different
from that commonly thought, and helps us to understand an
aspect of the complicated ecology.  When there is a shortage
of food, earthwoarms pass through the gravel .layer connected
with the soil, reach the water, take in activated sludge as
food there and return to the soil.  Such an earthworm habit
would be suited to the configuration of the fixed biomembrane.
Installations with the biomembrane are regarded as those
utilizing most effectively the soil animals under natural
ecosystem.  Though the production of excess sludge is not
estimated to account for the proportion to the amount of
eliminated BOD, the amount of excess sludge produced in 132m-
long Do Joker System of rural sewage buried under the farm
road of Wadayama with a population of 250 txersons over 36
months was 18m3 (moisture contents: 98%)\  /En a cold area,
                                500

-------
                         BOD volume load

Note:   1,  A solid line  is drawn based on the results  of experiments
          using 10cm—diameter gravel and 2.2cro-diameter gravel the
          line is modified slightly downward.
       2,  The results of this construction method are about 1,000
          of domestic waste water.   BOD volume load fell mainly
          between 0.3 and 1.Okg/m'/day.
              Fig. 5   BOD,  BOD  volume load and
                        removal ratio  of soil aerobic
                        (10cm—diameter gravel)(   )
                                  501

-------
Haguro  town,  Yamagata Prefecture,  excess sludge  has not been

taken out yet as  long as May,  1979 after installing.
         sa  - v
                                            .:•.».„«.-.:,,-,,«:,,^; .:•,•.:,«„»„,..,:,-,,»:»
                                            r "TPT35™ "™TBTZ3F=STSr:^ST?T*fawflfr?w"l
                               Table  2
  Sampling dace (1980)
Water temperature
Transparency
PH
DO
BOD
COD
T-N
(organ)
T-P
SS
17, Feb.
30 or more
7.2
9.78
3.91
8.7
(0.58)
12.18
2.51
1.0
A, Jun.
30 or more
7.0
8.05
7.53
21.88
(0.75)
17.54
4.52
7.5
19, Aug.
30 or more
7.2
8.73
5.5 .
12.2
(0.62)
12.93
3.54
10.0
7, Oct.
30 or more
7.4
10.41
5.40
11.6
(0.69)
10.25
3.51
4.0
6, Nov.
30 or more
7.4
10.88
0.52
8.7
(0.51)
7.69
3.04
3.6
  Quanlity of  discharged water from Haguro agricultural village sewer system



  Note:  1.  (organ) represents  (Kj-N minus NHi,-N)
                                    502

-------
The long and narrow waterway of installations allows the
filtration ability of the F.F.B.P. to work fully.  The re-
duction of the amount of suspended solid (SS), rapid settle-
ing speed, high rate of conversion to inorganic substances,
promoted possibility of denitrification have been attained
by further expanding the features of the F.F.B.P., depending
on the soil.

 9  Selection of Filter Mediums for Use in the Fixed
    Biomembrane Method

     Industrially produced filter mediums (manufactured
plastic products) are mainly used in Japan, and natural
gravel is now used only at the author's laboratory and the
River Bureau of the Ministry of Construction which is study-
ing low level water.
     Gravel includes river gravel, crushed stones, rapilli
and slag.  Rapilli has the greatest efficiency among these
because of its high porousness and rough surface.  The point
of using natural gravel is to employ less industrial products
from the standpoint-of saving oil and energy, it is not an
attitude of using no industrial products at all.  Drawbacks
in using natural gravel are malodor generated from sludge
accumulating in the pores of gravel, and clogging.  The
problem of malodor has been solved completely by employing
the soil-covering structure described above, but clogging of
the gravel layer still remains to be solved.  Therefore, the
author's studies have concentrated on clogging, and the
following two results have been obtained:

  (a)  Selection standards for gravel size of filter mediums

     The author's experiences show that 7 to 10cm diameter
gravel should be used for primary treatment and anaerobic
filter mediums receiving high levels of sludge, and 3 to 7cm
diameter gravel for secondary and tertiary treatment, and
aerobic filter mediums receiving waste water containing no
toilet paper so that clogging materials can be easily removed
by increasing flow rate, bubbling, and lowering the water
level.  For those receiving river water and turbid water
containing inorganic materials, it is thought to be simpler
to replace the clogged gravel completely rather than to wash
it.  When this construction method is used in river construc-
tion, the diameter of gravel may be 1.0 to 2.0cm or more.
                             503

-------
  (b)  Greater accumulations of sludge mean a wider variety
       of biota

     The scheme for the biomembrane of a trickling filter is
often used to explain the self-purifying function of the
F.F.B.P.  An anaerobic biomembrane is near the surface of
the filter medium and an aerobic biomembrane is on the former
membrane.  The aerobic biomembrane in this combination elimi-
nates malodor, making it practical for use.  However, when
gravel is used for the submerged biofilm, anaerobic sludge
accumulates in places where the aerobic biomembrane should
be formed and breaks the aerobic biomembrane to generate
malodor, so that the method using gravel is absolutely un-
practical.  This fact is coincident with historical fact, in
which the trickling filter method using gravel was replaced
by the activated sludge process because the trickling filter
clogged with floating sludge producing malodor.  When tech-
niques to solve the malodor problem are not available, the
development of manufactured filters on which sludge hardly
accumulates will play a leading part.
     In contrast to the above method, the Do Joker System
completely solves the malodor and sanitary pest problems and
it is concluded that gravel on which sludge which might in-
crease the variety of biota accumulates should be used to
facilltae purification of waste water, self-disintegration
and denitrification.

 10  Advanced Treatment and Countermeasures to
     Trihalomethanes

     The aerobic fixed biomembrane method is used for
secondary treatment.  But, when the long waterway system
shown in Fig. 2 is employed, less BOD volume load results
in the increase in removal ratio of BOD, SS, fat and fatty
oil, and ABS, so that the system can be used for tertiary
treatment.  A removal rate of 95% can be attained by making
BOD load 0.5kg/m3/day.
     The system in Fig. 2 was used for purification of pol-
luted river water with BOD of SOppm according to the results
of experiments at Nogawa, Tokyo (1,500m /day).  Three handred
thousand young salmon were successfully hatched and reared
with treated water, and it was shown that the system was very
effective in removing SS, fat and fatty oil, ABS and colon
bacilli as well as BOD.
                              504

-------
     The system produces treated water with very high trans-
parency and no need for chlorination, so that it is highly
regarded as a counter-measure to trihalomethanes along with
the capillary saturation trench system.
     In my edited book entitled "Do Joker System — Lectures"
and published in December 1980, I used the expression: "death
riding on a pale horse" in conjunction with the formation of
trihalomethanes by chlorine sterilization.  This is because
the Greek word Khloros for "pale" in the "pale horse" which,
in the Book of Revelations of St. John "was allowed to kill
people" is the origin of the word chlorine.  So I quoted two
books entitled "Pale Horse' written by Ropsin and "Look at
a Pale Horse" by Hiroyuki Itsuki these are wellknown novels
in Japan, (in Japan "Pale Horse, Pale Rider" written by
Katherin Ann Porter, is not very famous) and I appeal to my
readers that not ride on a Pale Horse.

 11  Flow Sheet of Ideal DO JOKER SYSTEM

     Necessary elements of an ideal treatment system are
simplicity, low construction costs, easy maintenance requir-
ing no special techniques, production of a small amount of
excess sludge, perfact environmental protection, perfect
pollution prevention, good treated water, and also complete
treatment within a site.
     The following two systems have been developed to meet
the above demands:

  (a)  a combination of the soil settling filtration tank
(Fig. 1) and capillary see page trench system (Fig. 4) uses
alternately two 2m-long trenches per head, requires an area
of 4 to 8m  per head, but needs no aeration power,

  (b)  a combination of the soil settling fitration tank
(Fig. 1) and soil contact aeration tank requires no flow
regulating tank, sludge accumulation tank or final settling
tank, and needs an area of 0.5 to 1,0m2 per head,  The past
long-period records of the systems,  (a) = table 3 and (b)=table 4
will be shown in the following tables.        (16)
     The records are shown in Tables 3 and 4.  Table 3 shows
the records at Shin-Matsuda, and Table 4 at Okutama.  (These
records will be rearranged.)
     In the development of Do Joker System, the obtained
data was far from an estimate because the anaerobic fixed
biomembrane method using gravel was employed for the settling
                               505

-------
filtration tank.  The facility is the same as a part of the
installation shown in Table 2, and very distinctive results
were obtained, so	
                               506

-------
Table 3.  The records of water treated by soil anaerobic fixed biomembrane method
          (settling filtration tank)  and  capillary saturation trench system (ppm)
          (Life Research Report, No.  11,  page 81)
X
Wace







Si








**""*""- — •— — ^__19?8 year
*X. Analysis item$r\

pH 
-------
                                                             Table  3   (Continued)
CJl
o
CO
^••x^^^^^-^^^1978 year
Water^x^ , */*' 	 r*" 	 •.
swii^>Anill5'8ls ltS5r\
So
Temperature of water
pH (ppn)
COD (pp»)
BOD (ppra)
SS (ppm)
DO (ppm)
No, of colon bacilli
(No. /at}
No. of common bacteria
(No./m2)
Chlorine ion (ppffl)
Anmoniacal nitrogen(ppm)
Total nitrogen (ppm)
Nitrite nitrogen (ppm)
Nitrate nitrogen (ppm)
Phsophoric phosphorus
(ppm)
ABC
Total phosphorus (ppm)
Oct. '78
18.2
6.5
92
230
44

7?xl03
110*103

29
38
0.017
0.10
5.7

21
Nov.
14.5
6.7
100
280
32
0
71><103
83X101"








Dec,
14.5
6.6
73
190
63
0
46X101
300xl03

41
49
0.013
0.38
6.3

18
Jan, '79
12.7
6.7
39
100
25<
0.74
140x10
54xl03








Feb.
13.0
6,8
55
120
55
0
89*102
300xl02
63
52
61
0.019
<0.1
7.2
3.1
29
Hnr.
14.5
6.6
82
120
47
0
140*102
120X101








Apr.
18.5
6,5
97
270
78
0,8*
47*10*
140* 10W
61
46
58
0.028
0.13
4.8
2.6
30
May
15.0
6,4
55
130
29
0
92X103
150*103








Jun,
18.7
6,7
39
40
25<
0
77*103
210xl03
42
27
32
0.025

-------
                                              Table 3   (Continued)
^ — ~~ ~-^J978 year
Water\. . T~~ : 	 ,
, , ^x. Analysis items \
sampliiie^> \
W2
Temerature of water
pH (ppm)
COD (ppn)
BOD (ppm)
SS (ppm)
DO (ppm)
No. of colon bacilli
(No. /rot)
Ho. of common bacteria
(No./m«)
Chlorine ion (ppm)
Ainmaniacal nit rogen(ppm)
Total nitrogen (ppm)
Nitrite nitrogen (ppm)
Nitrate nitrogen (ppm)
Phsophoric phosphorus
(ppm)
ABC
Total phosphorus (ppni)
Oct. '78
19.9
6.7
4,5
2<
25<

7
140

<0.13
70
0.037
68
0.029

0.45
Nov.
17.4
6.6
4.4
2<
25<
7.0
11
113








Dec.
15.4
6.6
3.4
2<
25<
5.1
1
76

0.20
62
0.005
61
<0.004

0.10
Jan. '79
11.9
6,9
2.9
2<
25<
6.9
13
250








Feb,
10.4
7,0
3,6
2<
25<
8.9
0
20
60
<0,13
30
0.005
30
0.014
<0.08
0.30
Mar.
11,5
6.6
3,1
2<
25<
7.9
6
80








Apr .
13,5
6,4
3,8
2<
25<
8.9
0
51
47
<0.13
41
0.005
41
0.028
0.49
6.0
May
17,6
6,4
2,7
2<
25<
7.0
51X10
80x10








Jun,
18,5
6,6
2,4
2<
25«
6.5
0
2
35
<0..13
26
0,0054
26
0.016
0.19
0.23
Jul.
20.2
6,7
3,0
2<
25<
6.8
2
110








ftug.
22.2
6,2
4.3
2<
25<
5.4
120
32*10
23
<0.13
40
<0.005
40
0.014
0.28
0.21
Sep.
18.7
6.4
11
2<
25<
4.4
270
67x10








1.  Domestic waste  water  from 4 farmhouses, Actual population  21 persons,
2.  Measured by the Ministry of Agriculture, Forestry and Fisheries (as described  in  the  second chapter, combination
    measurement for less  than 50 persons is usually prohibited,  but this governmental  experiment was conducted under
    the responsibility of  the Ministry of Agriculture,  Forestry  and Fisheries.
3-  Excess sludge was removed every three years without  using any power.
A.  A two-meter-long trench per head was clogged three  years after installation.   An additional 2m-trench is planned
    for alternate use, but at present recovery of the existing trench  is being  investigated.
5.  A feature of  the soil  settling filtration tank is to use rubber sheets  and  beer transporting containers filled
    with gravel for the inner contract filter medium.

-------
cn
i—'
o
                                                              Table 4.    BOD  in  water  treated  by  soil  aerobic  F.F.B.P.  and
                                                                              capillary  saturation  trench  system  (ppm)
U.UI-

1978
Aug.H
Sep. 10
Oct. 16
Nov. 18
Dee. 10
1979
Jon. S
Feb. 13
Mar. 11
Apr. 9
Hay 14
Jun. 4
Jul.14
Aug. 6
Sep. 10
BOII (PR/I)
lalrj r»nt*ct


«ov
180
510
288
164

46
35
32
309
68
135
98
-
322
luteJ contact


9.4
0.7
1.1
1.7
0.6

1.2
6.0
1.3
4.5
2.2
2.8
2.8
-
7.4
Soil


0.8
0.4
0.4
0.7
0.0

0.8
1.2
0.6
1.6
2.1
0.8
1.0
0.2
0.4
Total plionphnrii*. (ngt/t)
Inlel of circu-
lated contact


3.7
6.9
2.5
2.1
2.6

5.0
5.7
4.3
2.0
5.5
5.8
4..'
-
4.6
Outlet of circu-
lated COIU.lCl


l.t
3.4
2.1
1.6
-

-
-
-
0.4
-
-
2.9
-
3,5
Soil


0.2
0.1
0.0
0.0
0.0

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.2
Total nitrogen (nic/l)
Inlvt of circu-
lated contact


35
60
13
12
79

68
59
55
10
80
124
64
-
43
Outlet of circu-
lated contact
aurat Ion tank

32
30
24
24
-

-
-
-
4
-
-
46
• -
30
Soil
leaclute


17
22
22
18
15

17
15
12
11
3
9
18
-
14
                                                     1.  Domestic waste water from the museum  (stand, dining room, public lavatory),  estimated 300 persons.
                                                     2.  Measured by the Waterworks Bureau of  Tokyo.  {according to the Water Pollution Control Agreement)
                                                     3.  At first, BOD at  the outlet of the contact aeration tank was estimated as 60ppm. but only the activated  sludge method was
                                                        available In Japan at that tlrae, so that volume load was calculated based on the activated sludge method.  The renewal ratio
                                                        was unexpectedly  high,  these results  and those shown in Table 3 led to the Rravel filter medium being highly regarded.
                                                     4.  This  Installation Is a  combination of soil settling tank (Fig.  1),  soil contact aeration tank (Fig. 2) and capillary salur.icl
                                                        trench system (Fig. £). and was constructed In 1978.

-------
12  Conclusion

     Examples of application of this process in Japan include
both anaerobic and aerobic F.F.B.P. as stated above, and the
process is applied not only to primary and secondary treat-
ment but also to tertiary treatment equipment.
     Further, the recharge of treated water underground can
be carried out by is same equipment simultaneously and with-
out any other special equipment.  Similarly, some equipment
handles the use of treated water for plants besides treating
sewage.
     With this process, the beauty of a flowering plant which
impresses people who look at it is not impaired in the least
by foul smells or viruses from the equipment.  So, guests
enjoying their meal in a hotel garden on a summer night are
unaware that their own excreta washed away from their rooms
during the daytime are being treated under their feet —
only a few score centimeters from the ground surface.
     This process is used for sludge treatment as well as
sewage treatment.  The faicility shown in Fig. 1 of this
article is used as a sludge concentrating tank (this alone
is a mere storage tank) and the equipment in Fig. 4 is
installed as a supernatant liquid treating facility and
surplus sludge is treated by the combination of the two.
In the ordinary sense of storage, feeding is no longer pos-
sible when the container is full.  But with equipment used
for this process, this is not "the end of the world" but it
is just the beginning, because it is provided with the
capillary seepage trench of Fig. 4.  In other words, the
feeding of surplus sludge is continued everyday even when
the tank is full.  Naturally, the supernatant liquid that
overflows everyday is equal in quantity to the sludge that
is fed in.  This supernatant liquid is treated by the capil-
lary seepage trench and the concentration of sludge continues
in the main tank.
     In the case of some equipment in Shibukawa City, Gumma
Prefecture, 477m  of surplus sludge was fed in and 120m  was
applied to mulberry fields as a liquid fertilizer during the
two months from May 1979.  This means that the sludge that
was fed in was concentrated to about 1/4 by the capillary
seepage trench.  The city, which handles night soil treatment
for about 60,000 people, has installed four machiens since
the installment of Machine No. 1 in 1979.  When three more
are constructed, bringing the total to seven, it will no
longer have to use expensive petroleum to incinerate surplus
sludge.
                            511

-------
     F.H. King, who was quoted in Chapter 1, refers in his
two papers quoted by the author to Dr. Oskar Kellner's
analysis conducted in Japan about a century ago on the ferti-
lizer composition of human waste.  Tadashi Niimi, developer
of this process, graduated from the university at which
Dr, (Kellner) had taught long after he had left.  At the
same university, the author's father was taught by Dr. Masuji
Akiba about the capillary siphon movement of water in the
soil and developed the equipment shown in Fig. 4 of this
article about 15 years ago.  With a chracteristically Japanese
sentiment, the author cannot but see strange ties among those
three persons to whom the same university was the stage.
     Since the main theme at this conference is the Fixed-
Film Biological Process, the author did not mention the fact
that rain water is recharged underground by the Do Joker
System.  The successful percolation of as many as 41 tons/
day of rain water by a trench of only 10m is the clearest
proof that, whereas vertical percolation of water is difficult
due to fill-up, water seeps in the horizontal direction with-
out any fill-up.  In his January 1982 letters to the Minister
of Construction and the governor of Kanagawa Prefecture, the
mayor of Zama City of that prefecture reported that he would
adopt the above-mentioned Do Joker System for the combined
purpose of underground nourishment by rain water, counter-
measures against river flood and overflow, and prevention
of land subsidence by excessive pumping, taking advantage of
the rain water percolator installed in the relatively shallow
ground.  "Relatively shallow ground", as referred to in these
letters, is a factor that is most important to the Do Joker
System.  (One will do well to recall that digging a deep
well and injecting rain water deep undergound with the object
of recharging it to the ground is a common practice every-
where in the wrold.)
     To appeal the importance of this "relatively shallow
ground", the author and my group call this part soilsphere
or pedosphere — rather than simply calling it soil in as
much as they regard it as the "abode of living things" with
the greatest biological density on the earth.
     In this sphere, different forms of living animals,
micro-organisms and plants mix together and, with the par-
ticipation of sewage, "a perfect circulation of the forces
of nature", as first quoted from Dr, Maron, comes into
existence.  This "links of the chain" is still a mysterious
world which, regrettably, has not yet been sufficiently
clarified.  Out process is named the "Do Joker System" as a
                             512

-------
pun on the Joker of the playing cards and the Japanese words
"Dojo Joka" (soil purification) referring to our process.
It is hoped that this announcement by the author will serve
as an opportunity for people particularly in the sectors of
civil engineering and sanitary engineering to become inter-
ested in the "Links of the chain" in the pedosphere.
     It is regrettable that this article could not describe
details as the emphasis was placed on the wide-ranging ap-
plication of our system.  As regards contact materials, for
instance, the article could not cover the use of empty cans
instead of gravel or industrial wastes smaller than water
in specific gravity or seawead-like strings moving freely .
in the sewage to improve existing facilities, e.g., aeration
tank.  We hope that detailed reports on these can be pre-
sented in the future.
                            513

-------
Reference

 1.  Wagner, W., Die Chlnesisehe Landwirtsehaft, 1926

 2.  Niimi, M., "Do Joker Process" Do Joker System, Vol. 6,
     No. 10, Sep. 1981 (in English), p, 20^21.

 3.  Notification No. 1292 of the Ministry of Construction.

 4.  "The Yomiuri Shimbun", 22, Feb., 1982.

 5.  Sepcial Committee on Environment of the House of
     Representatives, 22, Nov. 1977.
     Same Committee of the House of Councilors, 22, Mar., 1978,
     Same Committee of the House of Representatives, 13, Jun.,
     1978.
     Question to the President of the House of Councilors
     from a member, 19, Nov., 1980.
     Answer to the President of the House of Councilors from
     the Prime Minister, 28, Nov., 1980.
 6.  Niimi, T., Arimizu, T., Soil Purification of Sewage—
     General, Do Joker Center Ltd., Oct., 1977, p, 35.
 7.  ibid.
 8.  ibid., p. 36
 9.  Yahata, T., "Wastewater Treatment Through Surface Soil"
     Do Joker System, ibid., p. 5.
10.  Amada, T. et. al., "The Application of Underground
     Piping Method to the Field of Water Purification Civil
     Engineering Journal, .Public Works Research Center,
     Vol. 23, No. 10, Oct., 1981, 0, 21.

11.  Studies by Aida, T., University of Ibaragi.
12.  Nakamura, Y., "Colonization by Earthworms of Niimi
     Waste Water Treatment Trenches".
13.  Namikawa, J., " New sewage system between men, earth and
     water" Quarterly Garden City, Japan Institute for'Com-
     munity Affairs, Vol. 3, No. 1, Jan., 1982, p. 28V32.

14.  Suzuki, N. et. al., The Report from the Yamagata
     Prefectural Institute of Public Health, 1981.
15.  Yano, Y., "Application of Self-Purification to River
     Water Quality Control          ", Journal of Water and
     Waste, The Industrial Water Institute, Vol. 24, No, 1,
     Jan., 1982. p. 21.
                           514

-------
16.  "Rural Life Research           ", General Research
     Center for Rural Life, No. 11, Mar,, 1980, p. 81^82.

17.  Inchinohe, M,,  "Renovation of Weste Water Effluent by
     irrigation of Forest Land", Journal of Water and Waste,
     ibid., p.  22,
                             515

-------
      A NEW FIXED-FILM SYSTEM COVERED BY SURFACE SOILS
      Tsutomu Arimizu,Forestry and Forest Products Research
      Institute, Ministry of Agriculture, Fishery and Forestry
      Japan
INTRODUCTION

      The acceptance of a new fixed-film system covered by
surface soils have been in progress last ten years for the
treatment of a wide-range of low and high strength of biologi-
cal wastewater, because in its very simple process removal rate
of BOD,COD, total nitrogen and SS are extremely high with very
few sludge production and few input of particular energy, and
without daily operation and maintenace effort and skill, in
the same area of the conventional wastewater treatment
processes.

DEVELOPMENT 10F THE SYSTEM
      It is interesting to note that the prototype of this
system with the name of Do Joker System( hereafter it will be
abbreviated to DJS ) came from studies of drain field whiqh we
call trench and has been very common in the United States.
      In a DJS aerated surface soils are made much use of with
capillary water having the mean infiltration rate of 0.65 gpd/
sq.ft in the case of the eastern United States soils(l). One
of our succesful experiments that has been carried out at a
                              516

-------
public waste disposal site in Gifu city showed that wastewater
there with BOD of 20,000 mg/1 was purified to 2 mg/1 by this
trench. In many trenches which are working well sludge disposal
has not been made more than last ten years. However,' I would
like to call your attention that our trench as shown in Fig.l
is quite different from darin field in some essential portions.
           Aerated surface
                soil
             Net
          Gravle
         Waste water
                —-—'
         Perforated
           pipe,
      Evaporation
     Evapo-transpiration
   Capillary water
                    Impermeable
                      base
Net
 Figure 1. Typical Trench system

       After .the stage of trench DJS advanced to replace a
septic tank as shown in Fig.2.
                         Sand filter  Aerated surface soil
          Influent
          Clarifierl
                                  e,rll  Gravel layer
                                  Perforated
                                   plate
                                                    Effluent
 Figure 2, Small DJS

       Along with them DJS was developed to provide more than
tertiary wastewater treatment by adding it after activated
sludge process to reduce BOD average from 20 mg/1 to one or
two mg/1,  or total nitrogen and SS average to one or two mg/1,
eliminating phosphorus by making use of trench simultaneously
that was added after the process, in order to meet the heavy
requirements set to control water pollution in a number of
lakes in this country.
                               517

-------
       It is since 1978 when on the basis of more than 20,000
cases of experience including trench DJS reached the stage to
aim an independent and large scale facilities in wastewater
treatment under a program of the Ministry of Agriculture,
Fishery and Forestry to introduce sewage system to local large
farmers' community.

STRUCTURE OF PRESENT DJS

       This system is one of the packed-bed processes covered
by surface aerated soils so that fixed-film system can develop
its potential ability completely through its good contacts
with gas and liquid.

Temperature

       In the beds of irregular and randomly packed granular
particles, heat conductivity is much large than that of liquid
and the effective conductivity of it is an average with few
deviation, being independent of radius of the particles. The
aerated surface soils over it can contribute to not only keep
in it high temperature essential for fixed-film system under
extremely cold winter, minimizing thermal fluctuations and
absorbing offensive odor, but also supply soil organisms hav-
ing strong capacity to purify wastewater, eliminate pathogens
and digest sludges, to the filter, preventing, emission and
airborne spread of pathogens.
       When outside am~bient  temperature was 1 C, the effluent
temperature was 10 C in a case to be mentioned here.Everything
outside was frozen, DJS  could work well.

Ecological composition

       Although DJS has a simple  structure, it provides  an
attractive habitate for  a more wide-range of microorganisms
and  for some animals  than others.
       The bacterial  flora  in a DJS consists of both Gram-
negative  bacilli  derived  from wastewater and Gram-positive
from  soils which  are  generally much more active than Gram-
negative. They compete directly with  fungi.
       Fungi are  also present in  the  filter beds  and occasion-
ally  dominate  the primary stage of  the  process. But population
of bacteria and  fungi is  controlled by  protozoa living in  the
beds  which compete with  nematodes and rotifers in  the  second-
ary  stage of the  process. These small metazoa  sometimes
harvest bacteria  and  fungi  while  processing  the solid
                              518

-------
materials. Mold mites are also present in it which feed on
fungi in localized anaerobic portions of the process but they
are preys of beetl mites and springtails.- Large metazoa are
present in it. They display functional roles in the recycling
and communications of all types of organic debris. Adult flies
of many types and species of beetles are the transporting agent
of bacteria, fungi, protozoa and mites. At the tertiary stage
of the process earthworms flourish as the first or second
descomposers in waste materials including sludges(2).
       In this way the ecological community of DJS has much
more complicated and rich food chains including not.only
aquatic but also terrestrial microorganisms which will lead to
maintain or increas overall stability(3). Considering the case
of bicultural, two-stage, high-rate activated sludge process
which consists of a simple food chain mainly between bacteria
and protozoa, this highly developed ecological composition of
DJS contribute to extremely small sludge production as well as
excellent and stable effluent quality from the process(4).

Hydraulic conditions

       Another typical feature of a DLS consists  in hydraulic
conditions.
       At first, clarifier ahead of contact aeration tank or
contact basin holds the maximum quantity of liquid between
doses, which reaches  a filter over the clarifier with bottom
feed as shown in Fig.3. Then each cycle provokes  a chain
reaction of flow, down the filter and clarifier,  smoothing out
any variations in BOD loading and above all eliminating scums.
       A filter of basin which comes next to the  clarifier is
similar to  the anaerobic filter with bottom feed  and is
completely  submerged by dosing in the waste which reaches also
aerated surface soils as in the case of clarifier through syn-
thetic net which prevents fall of soil particles  into the fil-
ter as shown in Fig.4. Then immediately capillary upflow takes
place in the same way with that of trench. Between dosing
liquid flows down through the filter to make aerobic condi-
tions there with reduced loading, which will prevent clogging
and increas plant capacity and efficiency.
       In the upflow and downflow packed-bed, liquid does not
completely cover the outer surface of the porous media with
biofilm and the part covered by gas contributes to reaction
through absorption and desorption, producing aerobic and an-
aerobic conditions there. In the beds a radial variation in
resistance  to flow may cause appreciable maldistribution and
                               519

-------
    GL ///////////7///////7/A   £
                            -Gravel
                           Wastewater
                             surface
                           Perforated
                             plate
                            Aeration
                             pipe
                          Sludge
                •—-Capillary
                   upflow
                 Net
 Figure 3. Clarifier
Figure 4. Coniac-t basin
cross—flow of repetitive  aerobic and anaerobic cycles in time
and space, without  input  of  particular energy. Intermittent
application of liquid  in  a DJS  is to provide alternate periods
of aerobic and anaerobic  conditions  everywhere in the attached
growth reactor all  the times(5).

Filter media

       Filter media used  in  a DJS are crushed stones, blast
furnace slag, discarded cans and synthetic products which are
specially manufactured for wastewater treatment,

DESCRIPTION OF RECENT  DJS

       The cross-section  and horizontal section of a recent
DJS constructed  in  1979 at Haguro-cho, Yamagata prefecture,
northern part of Japan, by the  Ministry of Agriculture,
Fishery and Forestry,  are shown in Fig.5 and Fig.6,respective-
ly.
       The feed  to  the filter comes  from a nearby community
with the population of 700 and  a. wastewater flow of 150 m /day
(  39,600 gal/day ) with  a strength  of 200 mg/1 BOD- and TSS
has been treated. Recirculation has  not been used.  They do
do not use oxygen gas.  The total volume of all filters are
714.5 m ( 188,770 gal/day )  with the total filter surface area
of 20,403  sq.m  ( 219,536 sq.ft ).  The results of chemical
analysis at each filter are  shown in Fig. 7 when the retention
time was 60 hours in total,  on  August 11, 1981.
                               520

-------
       frlmlty
              tear

                                       S Capillary net
                        'Irst contact bos;
                             '-rrfSsrfi
                       C&lorise contact basi
       [Excess sludge
 Figure  5, Cross-section of  DJS at  Haguro
  IDfluent ~f


Effluent —
                    First
                      clarif cr
                          r=m,
                          M
                               O    O
                             Interned!ate cXcriTi.
                                     ^
                              Final clarifler
                               0
                              Final clarific
Figure  6.  Horizontal  section  of DJS  at Haguro
          Elfccclvo volund>-

  Figure 7 v  Results  of  Chemical Analysis  of Flow  in DJS
                                       521

-------
       For the efficient nitrification organic loading up to
25 lb/BOD5/1000-ft3-day was employed with gravel media filter
but surface loading are maintained at 0.03 gpm/ft^. This will
be one of the reasons why organic nitrogen removal efficiency
was very high, reaching 96 per cent during the period of opera-
tion and sludge production was few.
       At Haguro plant operation and maintenance have been made
by some community people who had no skill before the plant was
constructed. As far as changes of quality of treated water in
Fig. 7 are concerned, excessive aeration had been made as in
other areas instead of frequent changes of aerobic and anaero-
bic conditions. From the experience of many other DJS plants,
BODc, TSS and total nitrogen average will be able to reach one
or two mg/1 in the course of time.

DESIN RELATIONSHIPS

       Design relationships related to attached growth bio-
logical treatment processes can be applicable to design of DJS
except for that of surface loading (6).

COSTS

       Although  relationships between costs and effluent
quality is not clearcut, investment per capita for DJS itself
    less than $1,000 at the time of construction.
CONCLUSION

       A DJS is a valuable and successful experiment which has
displayed the possibility of solving many problems of biolog-
cal wastewater treatment by the fixed-film biological process-
es. If a DJS could be added to after any wastewater treatment
processes in question or such processes were converted to DJS,
the situations concerned will undoubtedly be much improved.
       Mathmatical modeling has been under way out tnis model
must be different much from the conventional ones in that in-
stead of solving at the expense of creating new problems, all
problems of internal and external environments must be solved
towards increasing overall stability with healing process in
an automatic way(3) ,

REFERENCES

       1. KropfjF.W., et al., "Equilibrium Operation of Subsurfa-
          ce  Absorption Systems," Journal of WPCF. Sept .1977,
                              522

-------
pp.2007-2016.
James,A,,"An Ecological Model of Percolating Filters
," in "Mathmatical Models in Water Pollution Control
," edited by James, A,,John Wiley & Sons, N.Y.,1978
pp.303-318.
Goldsmith,E.,"Thermodynamics or Ecodynamics," The
Ecologist, Vol.11, No.4,1981,pp.178-195.
Dixit,N.S.S.."Bicultural, Two-stage, High-rate Acti-
vated Sludge Process," Illinois Institute of Techno-
logy, PhD.Dissertation, 1976.
Sharma,B.,et al.,"Nitrification and Nitrogen Removal
" Water Research, Vol.11,1977,pp.897-925.
Benefield, L.D., et al,"Biological Process Design
for Wastewater Treatment," Prentice-Hall,Inc.,Engle-
wood Cliffs,N.J.,1980,
                   523

-------
        STUDY OF FIXED-FILM BIOLOGICAL CONTACTORS
  FOR RECREATIONAL AREA WASTEWATER TREATMENT APPLICATION
        Calvin P.C.  Poon.   Department of Civil and
        Environmental Engineering,  University of
        Rhode Island, Kingston,  Rhode Island.

        Edgar D.  Smith.   Department of the Army,
        Construction Engineering Research Laboratory,
        Champaign,  Illinois.

        Vicki A.  Strickier.  Department of Civil and
        Environmental Engineering,  University of
        Rhode Island, Kingston,  Rhode Island.
INTRODUCTION

     A survey was conducted in 1981 to evaluate the type and
performance of existing wastewater treatment facilities in
U.S. Army Corps of Engineers Civil Works (CE) recreational
areas (1).   It was found that septic tank-leaching field or
septic tank-sand filter systems for subsurface discharge are
by far the most used treatment systems, followed by extended
aeration and lagoon systems.  Occasional high suspended
solid (SS)  concentration in lagoon effluents is not un-
common since dispersed growth and algal cells do not settle
well.  Upsets of extended aeration treatment plants are
experienced by many recreational areas from time to time
resulting in biochemical oxygen demand (BOD) and SS concen-
trations higher than the acceptable limits.  This phenomenon
is typical of an extended aeration process which has
dispersed growth leading to poor settling in the final
                            524

-------
clarifier.
     A prime concern of applying a suspended growth biologi-
cal reactor to the treatment of organic wastewater is the
occurence of periodic shock of hydraulic and/or organic
loads.  A successful control of the suspended culture
population in the reactor by sludge return requires a skill-
ful operation.  Even so, a washout of the suspended culture
occurs quite often with hydraulic shock loads.  The
previously mentioned survey identifies the flow fluctuations
at the CE recreational areas being
weekend ,.-     _.    i xn ,_  -,c n  holiday ,,     ..    , no
	 flow ratxo = 1.62 to 15.0, 	. , •* flow ratio = 1.93
   t i   -i. _L,V/W i_O.I_ O.W   O. • W£- W*J -I,—* • V «     ff
weekday        _.        ,          weekday
   m en    j offseason day r-     ^.     « -.-. ^  rt c
to 27.50, and 	—~	*• flow ratxo = 0.17 to 0.5.
                 weekday
These flow fluctuations apparently present an operational
problem to numerous extended aeration and oxidation pond
;reatment facilities in recreational areas.  It is believed
that because of the simpler operational requirement of a
fixed-film biological contactor and its ability to retain its
aiological culture with hydraulic and/or organic loading, a
rotating biological contactor (RBC) will lend itself a
favorable alternative to suspended growth reactors in
recreational area sewage treatment.
RBC TREATMENT KINETICS
     Under a steady hydraulic^and organic loading condition
of 2.5 to 7.5 g soluble BOD/m .d (approximately 0.5 to
1.5 Ib SBOD/1000 ft2.d), a RBC is able to remove consistently
from 57 to 90 percent of the soluble BOD (SBOD).  Outside of
this range of loading, the percentage removal is definitely
lower at lower loadings (2).  It is not certain that the
same percentage of SBOD/removal can be^maintained'at' steady
high loadings higher than 7,5 g SBOD/ra ,d, (Figure 1).  It
appears that with limits and with any given SBOD loading, a
lower percentage of removal can be expected when the influent
SBOD concentration is lower, particularly when the. influent
las already received some degree of biological treatment.
Ct is noted that below the loading of 7.5 g SBOD/m .d, the
affluent SBOD concentration is consistently below 20 mg/1,
   which 33 to 70 percent is made up of soluble nitrogenous
30D.
                           525

-------
     Successful nitrification of wastewater using RBC have
been demonstrated.  O'Shaughnessy et al (3) show 81 to 96
percent NEL-N removal from secondary effluent by RBC when

optimal pH and alkalinity are under control.  Beyond a load-

ing of 4.0 g NH3-N/m2.d (0.8 lb/1000 ft2.d) however, the

percentage is decreased significantly.  With approximately
the same range of NH--N loading (0.2 to 4.0 g/m.d), Zenz et
al (4) report 70 to  94 percent NH -N removal, while Reh (5)
report 85 percent removal.  For the nitrification of primary
effluent where the process is sensitive to organic loading
and the subsequent sloughing of nitrifying biofilm, Zenz et
al (4) report from 20 to 95 percent NH,-N removal within the
range of 0.25 5o 2.0 g NH,-N/m .d      (0.05 to 0'.40
          9
lb/1000 ft .d).  On the other hand, Poon et al (6) show 50
                                                         2
percent removal within the range of 0.1 to 0.65 g NH^-N/m .d

loadings to 83 percent removal for up to 2.8 g NH^-N/m .d
loading (Figure 2).
     In a suspended growth complete-mix reactor a Monod or
Michaelis-Menten enzyme kinetics is applicable.  The
kinetics assumes a hyperbolic saturation phenomenon with a
graduate change in reaction rates.  Biofilm kinetics
however may involve three distinct regions with abrupt
transition in the order of the bulk reaction from one
region to the other shown in Figure 3 according to
Harromoes (7).  Data from Kornegay et al (8) indicate that
from 0 to 65 y biofilm thickness, glucose  fully penetrates
the biofilm, resulting in zero-order reaction or the rate
of glucose removal increases proportionally to the film
thickness for a given glucose concentration.  For biofilm
thickness greater than 65 y, the reaction  rate becomes
constant corresponding to a partly penetrated biofilm.  The
same data also show that a zero-order reaction rate can be
obtained for a biofilm thickness of 200 y  if the glucose
concentration  is 1300 mg/1.  Below this concentration, only
a half-order reaction rate is obtained.  LaMotta's work  (9)
shows zero-order reaction rate obtainable  with a biofilm
thickness of  10 .y when the glucose concentration is 5.2
mg/1 and 70 y when the glucose concentration is increased
to 200 mg/1.
     In the study with steady loads,  (2),  the biofilm
thickness is  not measured.  Instead, the biofilm of a unit
                         526

-------
surface area is collected and its dry weight is measured.
Knowing the moisture content and the density of the dry
biofilm, the thickness of the biofilm on the media is
calculated.  The average biofilm thickness for the 1st, 2nd,
3rd and 4th stages of the RBC are respectively 250p, 180y,
120y and 102y.  It is apparent that an influent BOD from 40
to 186 mg/1 in the study do not fully penetrate the biofilm,
resulting in a reaction kinetics ranging from half-order to
first-order.
     Assuming that the biological contactor is rotating in
an ideally mixed compartment, the following mass balance
equations can be written:
     Half order
                 dC                            i,
               v -_£  = QC      Q.C  + A(-kj C  2) = 0
                 dt     x n-1   H  n     v  %a n
          or   Q(C  ,-C  )/A = ki C ^ = ^               (1)
               H  n-1  n'      %a n     *ga
     First Order
               Q(C   - -  C  )/A =  kn  -C  =  r-             (2)
               x  n-1    n       la n     la
 in which Q  is  the  flow  rate;  C   is  the  substrate  concentra-
 tion at the n  stage;  C   .  is     the substrate concentration
 at the  (n-1) stage;        A is  the  surface  area of  the
 rotating media at  stage n;  k,    and  k,   are  respectively the

 half-order  and first  order rate constants; and r^a and r^&
 are respectively the  half-order and first-order rates of
 substrate, removal.  Plotting  the substrate,  removal  rates
 versus  C  2  or  C should yield a straight  line, the  slope of

 which  is  the rate  constant kj,  or k,   according to
 equations 1 and 2.  Using  the total BOD data (non-settled,
 carbonaceous BOD and  nitrog^enOjUS BOD combined) such plots
 yield  a lq  value  of  0.93  g^/m^.d (correlation coefficient
 r = 0.63)^a and a  k,  value of  0.08 m/d (r  = 0.64).  If
                    xa
 SBOD data are  used,  the plots yield a k^ value of  0.75
                           527

-------
        (r = 0.68) and a k   value of 0.09 m/d (r = 0.69).
     In the same study (.2) ,  NH,C 1 is. introduced into the
sewage to create a very high NEL-N loading for a period of
10 days.  Figure 4 shows that a relatively constant rate of
                                              2
removel of soluble NFL-N is reached at 2.8 g/m .d when the
                                      2
soluble NH«-N loading is about 5, 0 g/m .d or the soluble
NH«-N concentration is about 50 mg/1.  The S-shaped curve
therefore suggests that the removal rate could be first-
order initially, changed to half-order as the soluble NK_-N
concentration increases and finally reaching the maximum
or the zero-order with very high NH,-N concentrations.  The
result also suggests that where nitrification primarily
takes place in the third and fourth stages of the RBC, a
soluble NH~-N concentration at or higher than 50 mg/1 is able
to penetrate fully a biofilm of 102 to 120p thick.  One can
not take advantage of the maximum removal rate in the design,
however, because the percentage of removal is lower and the
effluent would be unacceptable.
SIMULATED RBC STUDY IN RECREATIONAL AREAS
     A simulated study is carried out in laboratory using a
4-stage RBC 0.5m in diameter with a total area of 23.3m
       2
(250 ft ) of media.  The purpose of this study is to investi-
gate the effect of shock loads typical of recreational areas
on the treatment performance of the RBC.
     Three series of experiments are conducted.  The first
series uses a synthetic sewage of the following composition:

            Glucose          100 mg/1
            Bacto-peptone    55 rag/I
            Fed,            0.35 mg/1
MgS04.7H20       62.5 mg/1

                 92 mg/1

                 22 mg/1
            NH.C1            92 mg/1
            KHLPO,           8.4 mg/1
                   .7H0     37.5 mg/1
                           528

-------
This sewage simulates that of a recreational area facility
where urine is the major component of the wastewater.
Facilities for short term visits (visitors center, swimming,
boating, hiking, etc. but no camping) usually have waste-
water of this characteristic relatively weak in BOD but high
in NH..-N compared to a typical municipal sewage.  At first a
                                              32
steady hydraulic load of 460 liter/day (0.02 m /m .d or
          2
0.5 gpd/ft ) is maintained over a period of time.  After a
steady performance is reached, load fluctuation simulating
the frequency of use by visitors is applied with 16 hours of
                3  2
high load (0.06m /m  .d) followed by 8 hours of normal load

(0.02 in3/ni2.d).  This 24 hour cycle is repeated for 2 days
in this series of experiments.  Figure 5 shows that the RBC
consistently produces a very low BOD effluent under the
fluctuating load- condition.  BOD removal is 92 percent,
comparing to that of a control experiment (a steady normal
load maintained for a period of several days) of 86.4
percent.  Ammonia nitrogen in the effluent as depicted by
Figure 6, is high throughout the experiment despite the
fact that 3.9 to 9.2 mg/1 of NO -H repeatedly shows up in

the effluent.  However, there is a significant removal of
the organic-N from' the synthetic wastewater.  The conversion
of organic-N to NH--N adds to the already high NH_-N concen-

tration, resulting in very high NH--N loadings to the RBC
unit.  This may explain the phenomenon of low percentage of
NH_-N removal and strong nitrification taking place at the
    same time.  This phenomenon is unique and reflects the
special characteristics of a recreational area wastewater or
a similar wastewater with high NtU-N and organic-N concen-
trations.   If a more complete nitrification is desirable,
more stages- or additional media can be added to  the RBC unit
to reduce the NHL-N  loading.
     The other  two  series of experiments use a  synthetic
sewage  similar  to the  one aforementioned except  that glucose
is increased to 300  mg/1, NH.C1 concentration remains
relatively  the  same  and Bacto-peptone is eliminated.  The
sewage  is stronger  in  BOD and contains a relatively high
nitrogen concentration.  The  strength is equivalent  to that
of a typical municipal wastewater.•  It simulates the
                           529

-------
characteristics of a sewage from a recreational area with
camping, shower and laundry facilities.  Both experimental
series start with a steady normal load for a long period of
time.  In one series, this period is followed by 18 hours
of high load (approximately 3 times the normal load) and
then 6 hours of normal load.  The (3Q-1Q) cycle is repeated
twice in the experiment.  The other series is similar except
that the high load is approximately 4 times the normal load
(4Q-1Q) series.  As shown in Figure 7, the effluent SBOD
concentration is relatively stable at or below 17 mg/1
under the fluctuating load condition (4Q-1Q).  Even with a
short term shock (almost 3 times the high BOD load or 10-12
times the normal BOD- load) that is applied to the RB.C by
mistake, the effluent SBOD concentration is only .24 mg/1
and the unit recovers quickly once this unusually high load
is eliminated.
     Although most engineers use the SBOD parameter in
monitoring RBG performance, the result of the 3Q-1Q aeries
as depicted in Figure 8 indicates the importance of total
BOD rather than SBOD in monitoring the effluent quality.
Again the RBC is able to produce a good quality effluent
under the fluctuating load condition.  However, towards the
end of the second high-load period the. effluent BOD is
increased to 28 mg/1.  After a short period of recovery the
effluent BOD is further increased to 36 mg/1.  The increase
of the effluent BOD coincides with biofilm sloughing
initially from the first-stage and later on from the
second-stage.  When sloughing occurs in either one of the
first two stages, the biological solids do not settle well
even though the overflow rate of the clarifier is low at
20.4m3/m2.d (500 gpd/ft2).  The suspended solid (SS) concen-
trations corresponding to the effluents with 28 mg/1 and
36 mg/1 total BOD are respectively 23 and 134 mg/1.  It is
expected that the -effluent 'SS contributes to 'some effluent
BOD, making the effluent SBOD values lower than the
respective 28 and 36 mg/1 values.  The effluent quality
expressed in SBOD concentration in effect would be accept-
able under the fluctuating load condition.  An implication
of this finding is that for a small RBC treatment facility
in a recreation -area the period of sloughing could yield a
higher  total BOD and therefore a poorer effluent quality.
The frequency of sloughing is not monitored in this study.
Consequently it is not known how often a lower quality
effluent occurs.  It should be noted that this problem is
                             530

-------
greatly minimized in larger RB.C facilities because only a
small fraction of the entire media would experience sloughing
at any given time.
     To test the kinetics of BOD removal, removal rates are
plotted versus effluent concentration or (effluent concen-
tration)"5.  When the data of all three series are put
together with the normal loads as one group and the high
loads as another, such plots yield a k,   value of 0.54
 \,  \,                                 **&•
g2/m .d (r = 0.37) and a k,  value of 0.1 m/d (r•= 0.55) for
                                       l^  l*
normai loads, but a k^  value of 1.56 g2/m2.d (r = 0.60) and

a k   value of 0.23 m/d (r = 0.56) for the high loads.  This
indicates that first-order kinetics is more applicable to
the normal loads (lower effluent BOD concentration) and half-
order kinetics a better fit with high loads (higher effluent
BOD concentration).  The finding is in conformity with the
fixed-film reactor kinetics of Harremoes (7).
     Ammonia nitrogen removal is low in these two series of
experiments with high influent BOD concentrations.
Percentage of removal is 36.2% for the (3Q-1Q) series and
30% for the (4Q-1Q) series.  Only trace amount of nitrate is
detected in the effluents, indicating that nitrification can
not be estiablished in these high and fluctuating BOD load
conditions.  The NH~-N removal is due to biofilm synthesis

alone as NH» stripping is unlikely at the wastewater pH of
4.8 to 6.0.  It should be remembered that nitrification
takes place in the (3Q-1Q) series with fluctuating but low
BOD load condition even though nitrification is not
complete.  The incomplete nitrification is partly due to
insufficient media area for the high NH_-N load and partly
due to the relatively unfavorable pH (4.8-6.0) of the
simulated recreational area wastewater.  Because of the
unsuccessful nitrification In fluctuation load conditions,
investigation of nitrification kinetics is omitted from this
work.
SUMMARY
     Under a steady load condition, the biofilm thickness on
all 4 stages of a -RBC unit is indirectly measured and	
                            531

-------
calculated to be 250, 180 120 and 102 P respectively.  Zero-
order reaction kinetics is not to be expected since the sub-
strate (BOD) does not fully penetrate the biofilm.  The data
indicate that both first-order and half-order kinetics apply
equally well for SBOD loadings within the range of 0 to 8.0
g/m2.d (0 to 1.6 lb/1000 ft2.day) using NH3~N loadings from
0 to 5.0 g/m .d and beyond.  The data suggest that the
removal rate could be first-order followed by half-order and
the.n reaching zero-order when the NH,-N concentrations and
loadings "are high.  A full penetration of the biofilm 102 to
120 p thick is possible when NH_-N concentration is 50 mg/1
or above.
     Two studies are conducted, one with a synthetic sewage
relatively weak in BOD but strong in organic-N and NH3~N,
and the other strong in BOD and NH_-N, simulating
two different recreational area wastewaters.  BOD removal is
good under fluctuating load conditions.  Three problems are
identified in the application of RBC for the treatment of
this special waste.  One" is that despite the nitrification
taking place when the weak BOD wastewater is treated, a
great deal more media is required if near complete nitrifi-
cation is desired since the NH3~N loading is high.  Secondly

nitrification can not be established when the wastewater is
strong in BOD probably due to the sloughing of nitrifiers.
Thirdly,sloughing of biofilm from small RBC facilities
periodically yields effluents with high total BOD even
though the SBOD concentration may be acceptable.
ACKNOWLEDGEMENT
     This study was supported by funds provided by the U.S.
Army Construction Engineering Research Laboratory, Champaign
Illinois.  The guidance and advice provided by John
Cullianane, Jr. of Waterways Experiment Station, and Glenn
Hawkins, Chief Engineer, Department of the Army, are
appreciated.
                            532

-------
REFERENCE
1.  Poon, C.P.C., and Smith, E.D., "Rotating Biological
    Contactor Technology Evaluation for Civil Works Recrea-
    tional Area Application,"  First Report to Construction
    Engineering Research Laboratory, Department of the Army,
    July, 1981.
2.- Popn, C.P.C., Chin, H.K., Smith, E.D., and Milkueki,
    W.J., "Upgrading With Rotating Biological Contactors for
    BOD Removal,"  J. Water Pollution Control Fed., Vol. 53,
    No. 4, April, 1981.  pp. 474-481.
3.  O'Shaughnessy, J.G., et al. , "Nitrification of Municipal
    Wastewater Using RBC," Proc. 1st National Sypm/Workshop
    on RBC Technology, Vol. 2, February, 1980, pp. 1194-1219
4.  Zenz, D.R., et al., "Pilot Scale Studies on the Nitrifi-
    cation of Primary and Secondary Effluents Using RBC at
    the Metropolitan Sanitary District of Greater Chicago,"
    Proc. 1st National Sypm/Workshop on RBC Technology,
    Vol. 2, February, 1980, pp. 1222-1246.
5.  Reh, C.W., et al., "An Approach to Design of RBC for
    Treatment of Municipal Wastewater," Paper presented at
    the ASCE National Environ. Engr. Conf., Nashville,
    July, 1977.
6.  Poon, C.P.C., Chin, H.K., Smith, E.D., and Mikueki, W.J
    "Upgrading with RBC for Ammonia Nitrogen Removal," J.
    Water Pollution Control Fed., Vol. 53, No. 7, July, 198]
    pp.  1158-1165.
7.  Harremoes, P., Biofilm Kinetics, Chap. 4, Water Pollu-
    tion Microbiology, ed. R. Mitchell, Vol. 2, John Wiley
    &  Sons.             '
8.  Kornegay, B.H., Andrew, J.F., "Characteristics and
    Kinetics of  Biological Film Reactors," Fed. Water
    Pollution Control Admin., Final Report, WP-011811, 1969

9.  LaMotta, J.,  "Internal Diffusion and Reaction  in
    Biological Films," Environ. Science and Tech., Vol. 10,
    No.  8, August, 1976, pp.  765-769.
                                533

-------
en
OJ
                                       7.5
                                     T3
                                    «
<  5.0

O

Ui
tr


rf
O
m
CO

   2.5
                                                                                                          100% Removal Line
                                                                                                      Reh Autotrol Manual
                                                                                                    Autotrol Manual
                                                           2.5
                                                                                                             10
                                       5.0              7.5


                                   SBODj LOADING, g/m'day


                       FK3URE 1.  Relationship between soluble BOD5 ramoval and loadkig

-------
                                             I
                                             z
                                             I
                                             I
                 r*-cHN aTanios)

ABp-ziu/B'Noiivoui«j.nN jo aiva
                535

-------
en
CO
cr>
                                        QC

                                        UJ

                                        <
                                        oc


                                        0
<
UJ
DC
                                                K=ra/koa

                                                   /
                                                   /1st order

                           l/2 order
                                                                                         zero order
                                              ----- 1st order
 1/2 order r,/k0a-



- zero order ra/ko««1
                                                               1                   2

                                                                        CA=C*k1a/k0a


                                      FIOUM S. Dimensionless plot of reaction versus concentration in the bulk HquW

-------
                                               o

                                               *
                                               o
                                               03
                                               111
                                               E

                                               O
                                        O
            N-CHN aiarnos)
- ui/B-NOIAVOUIUilN JO 3i¥U J.INH
             537

-------
Norm*! load
Experiment
Mart* 	 *•
Hydraulic toad m3/mjd
BOO toad, f BOD/m2d
influent BOD conctntratton, mo/1
Normal
load
S hr».

0.02
1.7
87.3
High
toad
16 hr».

0.069
6.6
92.0
Normal
load
8hr».

0.02
1.7
86.3
High
load
10hra.

0.060
8.1
87.0
Normal
load
8 tir*.

0.02
1.7
86.0
cn
co
CO
Z
O
                                                     £20
                                                     u
                                                     O
                                                     Z
                                                     o
                                                     o

                                                     o  10
                                                     a
                                                     (0
                                                     i-
                                                     z
                                                     u
                                                     3
                                                     U.
                                                     U
                                                                               24
                                     32
48
56
                                                                                 TIME, bars

                                                                               (Not to scale)
                                        FWURE 6.  RBC affluent SBOD oonc«rtretioo under fhctuatkia load oondlllon - tow influent BOD concentration

-------










en
OJ












Normal load
Experiment
starts — »•

Hydraulic Ipad m3/m2d
NHj-N load gNHj-N/m'd
Influanl NHa~N mg/l
concentration

o 20
z E

M "aT
ll
t <
g C 10
u! w
u. o
UJ Z
O
O 0
Normal
load
8 hr».


0.020
0.46
23.4

High
load
16 hr».


0.058
1.73
28.1

Normal
load
0 hn.

\
0.02
0.34
17.2

High
load
16 hrs.


0.069
1.33
22.3

Normal
load
8 hrs.


0.02
0.48
24.6

o
o
o

o







i i ! 1 I
0 8 24 32 48 52
TIME, hours
(Not to scale)
FIGURE 6. RBC affluent NH3-N concentration under fluctuating load condition - low
          Intluant BOO concentration

-------










en
o













Normal
load
Experiment
• tarts-*
Hydraulic load m*/m2d
BOD load g BOD/m'd
Influent BOD mg/l
concantratlon
_
a
E
a - 20
So
z<
UJ fit
=; H
3 Z 10
U. Ill
u. u
Ul Z
o
u

Normal
load
1 hr.

0.014
3.4
107


High
2.6 hre.

0.078
42.3
639

o
load
16 hra.

0.074
7.6
103


Normal load
4 hra.

0.016
1.0
121


4 hr*.

0.010
2.7
143


High
2.6 hr».

0.044
17.8
233


load
16 hra.

0.076
13.8
181


Normal load
3 hra.

0.010
3.6
100


10 hr*.

0.014
4.3
341












-
0 0
o
0
o o


o



1 1 1 1 1 1 1 1
° 3.5 10.5 23.5 27.5 30 44 49 68
TIME, hours
(Not to scale)
FIGURE  7.  RBC effluent SBOD concentration under fluctuating load condition - high Influent BOD
           concentration, (4Q-1Q) series

-------
Normal
load
Experiment
atarta— »•
Hydraulic load mj/m*d
BOD load g BOO/m*d
Influent BOO ma/1
concentration
Z*
O
1 "
| 20
z
§ ^ 15
a E
i 10
z -
iu o
3
> t 0
u
Normal
load
2hre.

0.020
4,1
203


High
2hra.

0.060
10.8
181


load
16.Shra.

0.082
8.6
136


Normal load
3.5hrt.

0.020
3.2
160


4hra.

0.060
10.7
177


High
2.6hrs.

0.066
12.8
108

o
load
19.6hr«.

0.067
11.2
167


Normal load
2.6hr».

0.022
4.6
206 o


IShrs.

0.022
3.9
177


-
o
o
o o
o
i i rj l ] | If
4 20.5 24 28 30.5 44.5 50.5 68.
5
TIME, hours
(Not to scale)
FIQURE B.  RBC effluent total BOD  concentration under fluctuating load condition  -
           high Influent BOD concentration, (3Q-1Q)aerl*a

-------
       START-UP AND SHOCK LOADING CHARACTERISTICS OF A
         ROTATING BIOLOGICAL CONTACTOR PACKAGE PLANT
     Farley F. Fry.  Department of Civil Engineering,
     Virginia Polytechnic Institute, Blacksburg, Virginia.

     Tom G. Smith.  C.M.S. Rotordisk Limited, Mississauga,
     Ontario, Canada.

     Joseph H. Sherrard.  Department of Civil Engineering,
     Virginia Polytechnic Institute, Blacksburg, Virginia.
INTRODUCTION

     Rotating biological contactors have become increasingly
popular as a wastewater treatment alternative within the last
several years.  Because this method of treatment is relatively
new, design and operation procedures are still being developed.
Information that is currently lacking for use in design and
operation includes a) the time needed for development of
sufficient biofilm mass to insure an acceptable effluent
quality, and the progress of organic removal and development
of nitrification during the start-up period, and b) the shock
loading response of an RBC to increases in hydraulic and
organic loading.
     The purpose of this investigation is to provide start-up
and shock loading response data for an RBC pilot plant re-
ceiving primary effluent from a municipal wastewater treatment
plant.  Data reported in this study are abstracted from a
more comprehensive study performed by Fry (1).  Due to space
limitations only a small portion of this study are reported
herein.
                          542

-------
BACKGROUND INFORMATION

     A review of pertinent literature pertaining to the start-
up and shock loading response of RBC's is presented in this
section to provide a background for understanding the observa-
tions and analysis of results which follow.

Start-Up Characteristics of RBC's

     In general, little information is available on start-up
characteristics of RBC units.  References which are encoun-
tered are brief, incomplete and incidental in nature because
the research was not focused on start-up.  With the issuance
of so-called "tiered" National Pollutant Discharge Elimination
System (NPDES) permits, the time required to attain steady-
state conditions has become very important.  For example, the
assimilative,capacity of a body of water receiving effluent
from a wastewater treatment facility may require the operation
of nitrification facilities only during the summer months.
If the RBC process is used for nitrification purposes, what
length of start-up time is required for the units to discharge
acceptable quality effluent?  Obviously start-up character-
istics, especially the length of time required to discharge
acceptable effluent quality, is desirable information.
     Information is available on two aspects of the start-up
period.  One aspect is the establishment of the biomass.
According to one authority the attached biofilm generally
ranges from 2 to 4 mm thick one week after start-up (2).
Although useful as a general guide, this statement does not
consider the effects of varied organic and hydraulic loads.
In another study, 9 days were required to establish a thin
layer of biomass covering the entire outside of the media (3).
This observation was made in a study using RBC units to up-
grade trickling filter effluent.
     Establishment of an observable biomass required two
weeks in an RBC treatability study for phenol-formaldehyde
resin wastewater (4).  Pjr\or to introducing the industrial
wastewater, domestic primary effluent was fed to the RBC at
the rate of 1.6 gal/ft /day.  The phenol-formaldehyde resin
waste was introduced after the biomass was established.
Average organic strength of the primary effluent was not
provided.
     Based on the preceding studies it is apparent that a
measurable or observable biofilm will result 1 or 2 weeks
after start-up begins.  At the current time it is not known how
other factors such as wastewater characteristics and loading
                          543

-------
rates affect the development of the biomass.  It should be
noted there was no distinction made between heterotrophic
and autotrophic growth.
     Another important aspect of the start-up period which re-
lates to the time required to achieve steady-state operating
conditions, was evaluated by the Ontario Ministry of Environment  (5)
During the start-up period 300 gpd of raw domestic sewage was
fed to a five stage unit at the average rate of 1.05 Ib BOD_
(5-day Biochemical Oxygen Demand)/1000 ft /day.  After three
weeks of operation acceptable effluent quality (i.e. 15 rag/I
BOD- and 15 mg/1 SS) was discharged.  Several more months of
operation were required for a comparable growth in the last
two stages.  Effluent concentrations of BOD,, and ammonia-
nitrogen (NFL-N) remained the same.
     Nitrification of a high strength ammonia waste by use of
RBC units was examined by Lue-Hing e_t al. (6).  Sludge lagoon
supernatant diluted by 50% with water was introduced to an
RBC unit for 10 days of batch aeration.  After batch aeration,
the eight stage pilot plant RBC unit was continuously fed
diluted supernatant with a 12 day hydraulic detention time.
Following three weeks of continuous flow operation, effluent
nitrate-nitrogen (NO«-N) approximately equalled the total
Kjeldahl nitrogen (TKN) removed.  Typical 1KN removal was
approximately 600 mg/1 and influent BOD averaged approximately
100 mg/1.
     Trinh (7) reported the acclimation of an RBC unit (in
terms of BOD, removal) within two weeks with a loading rate
of 7.3 kg/100 ra /day (1.5 lb/1000 ft /day).  This investi-
gator compared the performance of an extended aeration
activated sludge package unit with an RBC package unit using
domestic waste from an isolated work camp.
     An RBC pilot unit required approximately three weeks to
reach steady-state conditions using primary municipal effluent
in a study conducted by Srinivasaraghavan et al. (8).  Soluble
organic2loading ranged from 0.5 to 1.2 Ib SBOD5 (Soluble BOD5)/
1000 ft /day.  Since nitrification did not occur in any phase
of this study, steady-state operation was based on organic
substrate removal.
     Based on the examples cited above it appears that 2 to
3 weeks of operation are required for an RBC to attain steady-
state operating conditions.  This appears evident not only in
terms of BOD,, but also for nitrification when the IKNiBOD,. ratio
is very large.
                             544

-------
Shock Loading Characteristics of RBC's

     Statements concerning the excellent ability of  RBC  units
to successfully handle shock organic and hydraulic loadings
are frequently encountered.  Wu et al.  (9)  for  example,
noted a chief advantage of an RBC is the ability to  resist
organic and hydraulic loads.  These statements  are generally
based on one or two characteristics of  RBC  plants.   One  im-
portant characteristic is the ability to retain the  attached
biomass when exposed to large hydraulic shocks.
     Welch observed this ability in one of  the  first investi-
gations of RBC units in the United States  (10).  Welch focused
his attention on  the response of a two-stage  RBC to  different
variables.  Variables included concentrations of synthetic feed,
disc speed, hydraulic residence time, intermediate settling and
sludge recycling.  Data on shock loading characteristics were not
presented, but it was observed that the process did  not  experi-
ence biological upsets encountered as with  the  activated sludge
process.
     An analysis  of phenol-formaldehyde resin wastewater treat-
ment by an RBC process found effluent Chemical  Oxygen Demand
(COD) values to be a function of the influent COD concentra-
tion.  More importantly the pilot plant RBC units functioned
effectively "under varying climatic and loading conditions and
exhibited excellent stability in withstanding periodic shock
-loadings" (4).
     Trinh (7) reported that the biological slime of an  RBC
system weathered  shock loads without sloughing  and' produced
consistent effluent quality.  However,  diurnal  flow  variations
caused a slight deterioration of effluent quality.   These
comments were based on a study comparing an extended aeration
activated sludge  process with a full-scale  RBC  system.
     Researchers  in California also reported  a  stable biomass
(11).  Municipal  primary effluent was used  to investigate  the
response of an RBC pilot plant to increases in  hydraulic load-
ing rate.  Over a 15 day period, the feed rate  was increased
from 6 gpm (1.1 gal/ft /day) to 70 gpm  in 5 steps.   SBOD,-
removal remained  relatively constant within the RBC  while  the
hydraulic loading was up to 1,040% of design  values  and  organic
loading up to 370% of design values.
     Later, the RBC received a two-fold hydraulic peak on  the
first day, a three-fold hydraulic peak  on  the second day,  a
four-fold hydraulic peak on day three and  five-fold  hydraulic
peak on day four.  These daily peaks were  timed to include
an increasing organic load in the municipal primary  effluent.
The result was a  significant increase in both organic and
hydraulic loading rates.  With soluble  organic  loading in-
creases of up to  700%, the total mass of soluble Total Organic
Carbon  (TOC) removed increased although admittedly effluent
soluble TOC increased considerably.  Of major importance was
                             545

-------
the lack of operational difficulties encountered in contrast
to occasional biological "washout" encountered in a suspended
growth system.
     It appears conclusive that RBC units are more resistant
to the loss of biological solids than suspended growth systems.
Perhaps the key words for describing this attribute are
"biofilm stability."
     Another important characteristic is the alledged ability
of RBC units to produce a consistent and acceptable quality
effluent while exposed to shock loads.  A main objective of
a study conducted by the Ontario Ministry of Environment in
1973 (5) was to determine the performance of a full-scale
RBC system under intermittent•feed conditions.  During a 6
week period raw sewage was fed to the unit at the rate of
320 gph for 2 consecutive days«per week.  The average organic
loading of 0.92 Ib BOD/1000 ft /day for each consecutive 2
day period was previously determined to be the approximate
maximum capacity of the system for the continuous feed phase.
When compared to the data from the continuous feed period,
little difference was found in terms of organic removal
efficiency.  Noticeable evaporation losses were evident in the
two central stages which were isolated from the primary and
secondary clarifiers.
     Kinner and Bishop (12) reported similar findings while
investigating saline RBC microbial populations.  The RBC units
were set-up at a sewage pumping station in Durham, New Hampshire
and received the diurnal loading characteristic of a small
town.  An effluent SBOD below 30 mg/1 was consistently
observed.
     Srinivasaraghavan et al. (8) evaluated the effect of
diurnal flow variations with no impairment of SBOD,- removal
efficiency.  For this study primary municipal effluent was fed '
to a four-stage, air-driven, pilot plant RBC unit.  The RBC
was 10 feet long, 10.4 feet in diameter and was preceded by a
10.4 foot diameter aerated wet well.  In the diurnal flow
phase the organic loading rate ranged from 0.47 to 0.78 Ib
SBOD.-/1000 ft /day, which was typical of other phases of the
study.  Nitrification did not occur in any phase of the study. •
The diurnal flow pattern consisted of periodic four-fold
hydraulic increases.  Unfortunately, the time between simu-
lated diurnal peaks was only 20 minutes.  It is probable the
large detention time of the wet well and RBC dampened or
eliminated all effects of the 20 minute diurnal cycle,
     In contrast, Dupont and MeKinney (13) after studying the
performance of a municipal RBC installation in Kirksville,
                          546

-------
Missouri, found treatment efficiency was reduced as a result
of variable hydraulic loadings.  These workers evaluated
monthly reports of the treatment plant and not the RBC unit
alone.  Reduced treatment efficiency was attributed to reduced
contact time within the RBC units and hydraulic surges on the
final clarifiers.
     The results of a study by Poon et al. (3) agree with the
Kirksville study.  Trickling filter effluent was fed to a pilot
plant RBC treatment system (including primary and secondary^
clarification) at the moderate hydraulic rate of 0.045 m /m /day
(1.1 gal/ft /day).  As expected, the trickling filter effluent
supplied a low SBOD,- influent concentration.  The RBC system
was exposed to a series of hydraulic shocks ranging from 120
to 220% of the steady-state loading.  Effluent SBOD  from the
RBC system increased rapidly as the hydraulic shocks increased.
An organic shock was simulated by coupling a high hydraulic
feed rate with a moderate SBOD .  Total SBOD5 removal actually
improved but effluent quality deteriorated significantly.
     Using a laboratory scale two stage RBC unit combined with
primary, intermediate and secondary clarification Antonie (14)
examined treatment during intermittent flow conditions.  To
simulate an industrial wastewater flow cycle, synthetic waste-
water was introduced only during the regular eight hour working
day.  Performance was generally consistent throughout the eight
hour period with the exception being a delayed response period
in percent COD reduction for during the first several of hours.
Continued 'sloughing of the biofilm during the night led to a
five-fold increase in mixed liquor suspended solids (MLSS).
     To reduce this problem the author repeated the experiment
but maintained a low wastewater flow and reduced disc revolu-
tions per minute (RPM) overnight.  Instead of a delayed response
period of COD removal, the COD removal initially was greater than
steady-state operation.  The author noted in an actual application
this could be accomplished by recycled effluent.  Antonie con-
cluded that intermittent flows could be effectively treated by
the RBC process as long as a low .wastewater flow was maintained
between cycles.
     In addition, Antonie evaluated the RBC system with the
intermediate clarifier bypassed under varying flow conditions.
During each day the treatment system was exposed to periods of
decreasing, increasing and constant flow.  The COD concentration
remained constant with only the flow rate changing.  Overall
performance in terms of COD reduction actually improved over
steady-state performance.
     In the final phase of testing, Antonie investigated the
                           547

-------
response of a 10 stage RBC unit without clarification to
hydraulic surges.  The 60 gallon unit was fed a synthetic waste
with strength of 500 mg/1 COD.  Shock loads of 500 gph for 6
minutes, 750 gph for 4.5 minutes, and 1000 gph for 3 minutes
were used.  In all cases the severe hydraulic surge drastically
impaired percent COD reduction although the total mass of COD
removed increased dramatically.  The unit required one hour to
return to steady-state conditions.  Even though effluent quality
was impacted it is important to note deleterious effects on the
biofilm were absent.
     This study by Antonie was probably the best and most com-
prehensive study of shock loadings currently available.  However,
Antonie neglected to examine the effects of pure organic shocks
and did not include nitrification in his research.
     In a well conducted study, Stover and Kincannon (15) found
that nitrification was more easily inhibited than COD removal.
By using a synthetic waste of known composition, nitrification
and carbon oxidation could be carefully monitored.  The steady-
state hydraulic loading was 0.5 gal/ft /day with respective COD
and NH_-N concentrations of 250 mg/1 and 27.6 mg/1.  Complete
nitrification was achieved during this study.  On two separate
occasions the workers introduced quantitative shock loads to the
RBC.  The unit was exposed to two-fold and four-fold shocks.
The percent COD removal remained relatively constant in all
instances. _In contrast, percent NH_-N remaining increased while
effluent NO -N concentrations decreased.  The authors attributed
the depressed nitrification rate to possible intermediary
metabolic by-products resulting from the increased heterotrophic
growth rates.

MATERIALS AND METHODS

     This research was conducted to determine the start-up
characteristics of a full-scale RBC unit and to determine the
response of the same unit to controlled shock loadings.  In
this section descriptions and details of the procedures and
methods used to attain these goals are provided.

Rotating Biological Contactor

     In 1978 CMS Rotordisk Limited of Mississauga, Ontario,
loaned an S5 Rotordisk unit to the Department of Civil Engi-
neering, Virginia Polytechnic Institute and State University,
for research purposes.  Primarily intended for small commercial
establishments and single family dwellings, the S5 Rotordisk is
designed to treat 600 US gallons per day.  The fiberglass unit
                           548

-------
includes the rotorzone (compartment containing the rotating
disks), subjacent primary clarifier and secondary clarifier
shown in Figure 1.  The primary and secondary clarifiers had
respective detention times of six and four hours.  These
detention times do not include ample space reserved for
sludge storage.
     Support for the biofilm is provided by 500 square feet of
high density polyethylene 3/8 inch mesh divided into four
stages.  A 1/4 horsepower motor provided power for continuous
rotation at three RPM (approximately 0.5 ft/sec tip speed).
Wastewater enters the primary clarifier, flows under the
rotorzone and enters the first stage through a slot located
in the opposite corner.   A smaller slot is provided at the
bottom of the first stage to provide the recirculation of some
aerated wastewater into the primary clarifier.  The flow proceeds
through the four stage RBC unit in a serpentine manner finally
exiting to the secondary clarifier and eventually discharges with
gravity flow utilized throughout the unit.  A baffle in the
secondary clarifier inhibits the discharge of floating solids.
     The S5 Rotordisk was placed next to a primary clarifier
at the Blacksburg and Virginia Polytechnic Institute Sanita-
tion Authority Stroubles Creek Wastewater Treatment Plant, near
Blacksburg, Virginia.  Approximately 30,000 people in the
Blacksburg vicinity are served by this treatment plant.  This
population includes approximately 21,000 students engaged
in studies at Virginia Tech.  Very few industries are within the
service area so the wastewater is primarily of domestic origin.
     An ECO C-15 Centrichem Pump was purchased by the Depart-
ment of Civil Engineering to supply primary effluent to the
package plant.  Three C-clamps on the discharge hose provided
effective and economical flow rate control.  C.M.S. Rotordisk
Ltd. supplied a pin timer manufactured by Hydro-Aerobics
International, Inc. of Milford, Ohio for positive pump control.
The pin timer provided 24 hour off/on pump control in 15
minute increments.  Additionally, the pin timer provided the
capability of turning the pump off any or all days of the week.
Operation of the motor for disc rotation was independent of
the pin timer.

Sample Collection Points

     Throughout the entire research period grab samples were
collected from the water surface at the same locations in the
                         549

-------
FIGURE i    PROFILE AND PUN VIEW OF S5 ROTORDISK
                             550

-------
package plant.  Samples were collected in the primary clarifier
near the inlet pipe, in each of the four RBC unit stages and
in the secondary clarifier on the discharge side of the baffle.
A sample was collected from the secondary clarifier instead of'
the discharge because of intermittent flow.  After three and
one-half weeks it became apparent the sample collected in the
primary clarifier of the rotordisk was not necessarily reflective
of the influent composition.  Therefore an additional grab sample
was regularly collected from the clarifier supplying primary
effluent to the Rotordisk.  This sample is referred to as the in-
fluent to avoid confusion.  The influent has been labeled sample
collection point A, with the Rotordisk primary clarifier referred
to as collection point B.  Stages 1 through 4 of the rotorzone are
denoted by C, D, E and F respectively.  The letters G and H are
used to designate the sample collected in the secondary clarifier.

Operations

     Operations began on May 8, 1981 when the package plant
was filled with primary effluent.  A normal feedrate of 480 pgd
was maintained during all phases of research.  One exception
was the introduction of hydraulic shock loadings.  The desired
flow rate was achieved by using the pin timer to alternately
turn the pump on for 15 minutes and off for 30 minutes.

"Start-up

     The first set of samples were collected at 8:00 a.m. on
May 9.  Additional sample sets were collected every other day
at 8:00 a.m. until the end of the start-up phase on June 22.
Samples were quickly transported to laboratory facilities on
the Virginia Tech campus.  Suspended solids (SS) and total
alkalinity determinations were immediately conducted after which
all samples were acidified and cooled to 4 C for later analysis
of COD, NH3~N, NO -N, and Organic Nitrogen (Org-N).  Dissolved
oxygen (DO; concentrations were recorded at each sample col-
lection point in the Rotordisk after collection of the samples.

Hydraulic Shock Loading

     Three separate hydraulic shocks were conducted to evaluate
the response of the unit.  The initial hydraulic shock was
applied on July 17, 1981.  Starting at 6:00 a.m. samples were
                       551

-------
collected at 8:00 a.m. the pump was operated continuously for
eight hours.  This created a three-fold increase with an eight
hour duration.  Each set of samples was immediately placed in
an ice cooler for preservation and DO concentrations were
recorded at each sample collection point.  After the final set
of samples were collected at 6:00 p.m., all samples were trans-
ported to the laboratory where they were preserved and cooled
to 4 C following analysis of SS and alkalinity.  To monitor
recovery, samples were collected the following morning at
8:00 a.m. and again on June 20 at 8:00 a.m.  All samples were
later analyzed for COD, NH -N, and NO~-N.
     The entire procedure was repeated on July 25, 1981 with
the duration of the hydraulic shock extended to 10 hours.
Samples were not collected at 6:00 a.m. and 10:00 a.m.
Otherwise, procedures were the same as the previous hydraulic
shock.  In addition, monitoring of the recovery period was
extended to seven days after the shock.  Once again DO was
recorded on site while SS, alkalinity, COD, NFL-N, and NO_-N
were determined at a later time.
     Another eight hour hydraulic shock test was conducted on
August 11, 1981 to examine reproducibility.  Seven sets of
samples were collected at two hour intervals from 6:00 a.m.
until 6:00 p.m. on the day of the increased hydraulic loading.
Samples were collected at 8:00 a.m. on August 12, 14 and 16
to monitor the return to steady-state conditions.  All other
collection and analytical procedures remained the same.

Organic Shocks

     To examine the effects of an organic shock without an
increased hydraulic loading, Kroger Incorporated (Cincinnati,
Ohio) Nonfat Dry Milk was added to the primary clarifier of
the Rotordisk on two separate occasions.  A step feed increase
was produced by thoroughly stirring the milk into the primary
clarifier.
     The first organic shock was conducted on August 18, 1981.
Prior to the addition of two pounds of milk, samples were
collected and DO recorded at 6:00 and-8:00 a.m.  Normal pump-
ing routines of 480 gpd were maintained.  Samples were collected
every two hours until 2:00 p.m. when it became visibly apparent
the organic removal capacity was grossly exceeded.  Samples were
collected the following morning at 8:00 a.m. and again at
8:00 a.m. on August 21 and 23.  All samples which could not be
immediately transported to the laboratory were placed in an ice
chest.  Upon arrival at the laboratory SS and alkalinity were
                          552

-------
evaluated.  Afterwards all samples were acidified and stored
at 4°C until later analysis for COD, ML-N, and NO -N.
     The entire organic shock procedure was duplicated on
August 25 with two differences; 1.2 Ibs. of nonfat dry milk
was used, and samples were collected every two hours from
6:00 a.m. until 6:00 p.m.

ANALYTICAL PROCEDURES

     Unless otherwise stated, each parameter evaluated for
this investigation was determined in accordance with Standard
Methods for the Examination of Water and Wastewater (16).
     All dissolved oxygen concentrations were measured by
means of a Yellow Springs Instrument Company, Inc. (Yellow
Springs, Ohio) Model 54 Oxygen Meter.
     Unfiltered chemical oxygen demand determinations were
performed on all samples by use of the dichromate reflux
method as prescribed in Standard Methods
     The procedures found in Section 208.C of Standard Methods
were utilized to measure suspended solids.  All samples were
filtered through 5.5 cm glass fiber filters  (Grade 934 AH,
Fisher Scientific Company, Clifton, New Jersey).  All weight
measurements were made by use of Mettler Instrument Corporation
(Princeton, New Jersey) balance Model AC100, Model H 10 or
Model H 18.
     Total alkalinity determinations were performed on all
samples by titration to a pH of 4.5.  A Fisher Scientific
Company Accumet Model 120 or Corning Glass Works  (Corning,
New York) Model 7 pH meter was used to measure the pH.
     Unfiltered ammonia-nitrogen and organic-nitrogen concen-
trations, were determined in accordance with  Standard Methods.
After distillation and digestion, the acidimetric method was
used to determine NH_—N and Org-N concentrations.
     Unfiltered nitrate-nitrogen determinations were made in
accordance with the Brucine method presented in Standard Methods.
A Bausch & Lomb Incorporated (Rochester, New York) Spectronic
100 was used to measure absorbance.

COMPUTER GRAPHICS

     Data from the start—up, hydraulic shock and  organic shock
phases is presented in three-dimensional graphs.  The three
variables presented are sampling location, time and parameter
concentration.  All three-dimensional graphs were drawn by
use of the Surface II Graphics System (17).  Each graph was
                            553

-------
plotted by the perspective block diagram mode of the Surface
II program.  To enhance the view of the diagram, 30  was selected
as the angle of the observation point above the horizon.  To
reduce distortion from convergence of lines, the distance from
the center of the block diagrams to the point of observation was
assigned the value of 10,000 grid units.  As a result the
perspective block diagrams appear as conventional three-
dimensional plots.  Finally, the diagrams were placed at a 25
azimuth to aid the viewer.  An azimuth of -155  was utilized,
when it was desirable to view a diagram from the reverse side.
     Difficulties were encountered when plotting the data from
the organic and hydraulic shock exercises.  Variables contained
in the Surface II program could not be adjusted to accommodate
the transition from two hour sampling intervals to 48 hour
sampling intervals.  Using SAS (18) multiple linear regression,
Intermediate sampling values were generated to eliminate this
problem.

RESULTS AND DISCUSSION

     The goals of this research were to determine the start-up
characteristics of a full-scale RBC and to examine the response
of the same unit to controlled shock loadings.  The RBC pack-
age plant contained a primary clarifier, four stages of discs,
and a secondary clarifier.  Samples were collected from seven
locations ranging from the influent, the Rotordisk primary
clarifier and through each stage of the rotorzone into the
secondary clarifier.  The influent and primary clarifier are    >
respectively referred to as sample collection points A and B
while the four stages of discs are labeled C through F.  A
final collection point in the secondary clarifier is designated,
by the letters G and H.
     Samples were analyzed for DO, COD, SS, nitrogen forms, and
alkalinity.  After careful consideration three-dimensional plots
were selected to illustrate the trends revealed by the data.
In this section the three-dimensional plots will be analyzed
and discussed.  Due to space limitations only eight representa-
tive three-dimensional graphs will be illustrated.  Emphasis
will be placed on observable trends rather than quantities con-
sumed or generated.  A trend analysis such as this will be more
applicable to other RBC treatment systems.  In fact, the chief
advantage of three-dimensional graphs is the easy perception of
surface trends.  For the purposes of this study this feature
compensates for the difficulty encountered when trying to read
precise values from the graphs.
                          554

-------
     The three-dimensional graphs for the hydraulic and organic
shock loading experimets utilize supplemental multiple linear
regression data.  These supplemental or intermediate values
were merely intended to reflect general trends of the actual
data without replacing any actual data.

START-UP

     A slight bacterial growth was observed 24 hours after
start—up.  Two days later an increased biofilm thickness was
observed, but a COD reduction trend did not begin until the
fourth day.  Chemical oxygen demand profiles slowly changed
until 20 days passed.  During the next 10 days effluent COD
values were uniform and the reduction of COD concentrations
occurred primarily in the first two stages of the rotorzone.
     At the same time, changes in the appearance of the bio-
film occurred.  Initially growth on the discs was brown to
grey-brown in color.  This remained true for approximately
three weeks of operation.  By this time growth was greater
on the first two stages while the last two stages began acquiring
a reddish brown appearance.  In addition, growth in the first
two stages became filamentous.
     An examination of Figure 2 reveals an average influent DO
concentration of 0.5 mg/1 rising to a peak of approximately
7.0 mg/1 in stages 3 and 4 of the rotorzone.  The DO concen-
tration decreased to an average of 4.5 mg/1 in the secondary
clarifier..  The graph reveals that peak DO values began de-
creasing on the tenth day and decreased for 18 days.  At this'
point respective DO concentrations in stage 4 and the secondary
clarifier are 3.2 and 2.3 mg/1.  During the first 9 days this
trend resulted from increasing COD reduction whereas the last
half.coincided with the start of nitrification.
     As can be seen from Figures 3 through 5, nitrification
began slowly after 18 days with vigorous activity recorded six
days later.  Influent NH_-N concentrations consistently ranged
from 17 to 22 mg/1 until the end of the start-up period.
Concentrations in the effluent started decreasing rapidly after
20_days and remained close to zero after 36 days.  On day 18,
NO,,—N was detected in the secondary clarifier and fourth stage
of the rotorzone (Figure 4).  Two days later concentrations had
increased to over 12 mg/1 and were detected in the third stage
of the rotorzone.  After 30 days effluent N0_—N concentrations
slowly decreased while N0_—N appeared in stage 2.
     Influent total alkalinity concentrations shown in Figure 5
averaged 165 mg/1 as calcium carbonate (CaCO ) with' a range
from 151 to 180 mg/1 as CaCO .  After 28 days effluent
                           555

-------
           11
                      22
                   TIME, DAYS     33
FIGURE 2    DISSOLVED OXYGEN CONCENTRATIONS FOR THE START-UP PERIOD
                                                                 30
                                                             SAMPLE
                                                            LOCATION
FIGURE 3
AMMONIA-NITROGEN CONCENTRATIONS
FOR THE START-UP PERIOD
                               556

-------
FIGURE
                   TIME
             MITRATE-HITR06EN CONCENTRATIONS
             FOR THE START-UP PERIOD
                                                             SAMPLE
                                                            LOCATION
                                                              SAMPLE
                                                             LOCATION
                   TIME, DAIS
FIGURE 5     ALKALINITY CONCENTRATIONS FOR
             THE START-UP PERIOD
                                  557

-------
alkalinity concentrations averaged 105 mg/1 as CaCCL with
approximately 60 mg/1 consumed.  Nitrification was confined
to the last two stages where a distinctive red-brown growth
was present.  Sufficient DO concentrations were present for
nitrification throughout the start-up period.
     On June 11, 33 days after testing began, final examina-
tions were concluded at Virginia Tech.  An exodus of most of
the 21,000 students changed the character of the wastewater.
With the exception of SS and alkalinity, influent concentrations
were reduced.  As a result, peak DO levels began increasing
and sloughing occurred in the second stage, accompanied by a
gradual change of the biofilm to a red-brown color.  Afterwards
COD reduction occurred only in the first stage while nitrifica-
tion migrated forward to the second stage._ The graphs also
indicate lower effluent COD, NH.-N, and NO_-N concentrations
while effluent alkalinity values increased.  When samples were
collected on the last day, sloughing was negligible and the
color change appeared complete.  By this time, the second stage
had changed to a red-brown color and the first stage changed
from a grey-brown to a grey-white color.  It is suspected that
septic conditions in the sludge zone of the primary clarifier
stimulated the growth of sulphur organisms in the first stage.
     Suspended solids and Org-N concentration trends did not
change during the start-up period.  As a general rule SS and
Org-N concentrations decreased as the wastewater passed through
the treatment "system.
     In summary, it may be stated that the attached biofilm
was present within 24 hours of start-up and steady-state opera-
tions in terms of COD removal and nitrification were definitely
achieved in 44 days.  This statement is tempered by the sudden
change of wastewater characteristics during the start-up
period.  Consistent, effluent COD concentrations were observed
after 20 days while nitrification appeared to approach steady-
state conditions at 30 days.  The latter observation can not
be confirmed because of the change in wastewater characteristics.
These statements are all based, on observation at a lightly-
loaded RBC package plant and may not be true under other
conditions.

HYDRAULIC SHOCK LOADINGS

     This phase consisted of three separate hydraulic shock
loadings.  On each occasion the hydraulic flow rate was tripled
for a specific period of time.  As expected the response char-
acteristics were similar on each occasion.
                             558

-------
First 8-Hour Hydraulic Shock Loading

     A three-fold step increase in hydraulic loading was im-
posed on the RBC for 8 hours.  A diurnal pattern was exhibited
by the influent COD and NH_-N concentrations.  Therefore the
RBC unit was exposed to increased concentrations coupled with
a tripled flow rate.  Dissolved oxygen concentrations decreased
until two hours after the shock loading ended.  This can be
attributed to an increased biological activity and/or decreased
hydraulic detention time.
     An interesting progression of high COD concentrations was
observed in the first 8-hour hydraulic shock.  The high COD
concentrations first appeared in the primary clarifier and
stage one of the rotorzone.  Two hours later the peak appeared
in the second stage, disappeared at the eight hour sample, but
reappeared in the third stage only to disappear at 12 hours.
Because floating and rising sludge had been previously observed -
in the start-up phase, it might appear that the increased flow
rate disturbed the primary clarifier.  However, SS concentra-
tions do not corroborate this hypothesis.  Therefore it appears
logical that the biofilm could not quickly accommodate the large
influx of soluble organic matter.  In addition to the peak values,
slowly increasing COD concentrations may be observed in the first
and second stages.  Interestingly, effluent COD concentrations
did not increase during the hydraulic shock loading.
     As in the start-up period SS concentrations tended to de-
crease as the wastewater flowed through the RBC package plant.
Generally, SS concentrations in the unit did increase towards
the end of the hydraulic shock loading.  Visual observations of
increased turbidity confirmed this pattern.
     In contrast to COD removal, nitrification was greatly
inhibited.  Effluent NKL-N and alkalinity concentrations in-
creased immediately while NO--N formation decreased.
Nitrification began to recover within 24 hours and was fully
recovered within 74 hours.

Ten-Hour Hydraulic .Shock Loadings

     Trends of the 10-hour hydraulic shock are very similar
to those of the first shock.  Influent COD concentrations were
typical of a diurnal pattern whereas NH,,-N and alkalinity
concentrations were relatively constant.  Dissolved oxygen
concentrations decreased during the shock but began increasing
as the hydraulic shock ended.
     Chemical oxygen demand concentrations started to slowly
increase throughout the unit six hours after the shock began.
                           559

-------
The increases are first noticeable in the second stage and
soon appeared in the remainder of the treatment unit.  Effluent
COD concentrations had only begun to increase when the intensive
sampling period ended.  Chemical oxygen demand removal returned
to normal within 24 hours.
     Suspended solids fluctuated considerably even when the
unit was exposed to normal flow conditions before imposing the
shock load.  During the increased hydraulic loading, SS
fluctuations intensified throughout the entire treatment system.
     Nitrification was quickly and significantly impacted by
the 10-hour hydraulic shock loading.  At the end of the shock,
alkalinity consumption, NO_-N formation and NH_-N were
barely noticeable.  In this case nitrification recovered quickly
and appeared normal after 24 hours.

Second 8-Hour Hydraulic Shock Loading

     The second 8-hour hydraulic shock resembled the first
8-hour hydrualic shock.  The influent COD concentrations ex-
hibited a diurnal pattern.  As the increased hydraulic loading
began, DO concentrations decreased until the shock ended, there-
after DO concentrations increased.  DO concentrations in-
creased immediately when flow rates returned to normal.
Chemical oxygen demand concentrations are presented in
Figure 6.  Influent COD concentrations fluctuated from 110
to 154 mg/1 with an average value of 134 mg/1.  Effluent
values consistently less than 25 mg/1 with only one excep-
tions occurring between 8 and 21 hours.  Once again the last
three stages were not capable of removing excess COD.  In
contrast to the first 8-hour hydraulic shock, effluent COD      '
concentrations increased significantly.  At the same time
SS concentrations were also larger than for the first 8-hour
hydraulic shock.  This indicates the larger COD concentra-
tions were probably caused by increased turbidity from the
primary clarifier.  The sudden decrease of SS and COD
concentrations after the shock ended affirms this belief.  As
with the other hydraulic shock loadings, increased turbidity
was visually observed but was not accompanied by biofilm
sloughing.
     Total alkalinity concentrations are presented in Figure 7.
Influent alkalinity concentrations averaged 148 mg/1 as CaCO
with a narrow range of 142 to 157 mg/1 as CaCO,..  Under normal
                            560

-------
                                                                     180
 FIGURE 6
                   TIME, HOURS   12
 CHEMICAL OXYGEN DEMAND CONCENTRATIONS
 FOR THE SECOND 8-HOUR HYDRAULIC SHOCK LOADING
FIGURE 7
                  TIME, HOURS
ALKALINITY CONCENTRATIONS FOR THE
SECOKD 8-HOUR HYDRAULIC SHOCK LOADING
                                                                    160
                                                                    120
                                                            CO
                                                            s
                                                              D
                                                               SAMPLE
                                                              LOCATION
                                   561

-------
conditions alkalinity consumption averaged 65 mg/1 as CaCO,
leaving an average effluent concentration of 83 mg/1 as CaCO.,.
As expected, nitrification was significantly inhibited by the
second 8-hour hydraulic shock loading.  Immediately after the
shock began NH-—H concentrations increased dramatically in the
latter stages of the RBC while NCL-N concentrations decreased
significantly.  In addition, alkalinity consumption decreased
to 20 mg/1 as shown in Figure 7.  Recovery began as soon as the
flow rate returned to normal with complete recovery within 24
hours.  The recovery rate compares favorably with the 10-hour
shock loading response but is faster than the original 8-hour
hydrualic shock.

Summary

     In general, COD removal was only moderately affected by
the hydraulic shocks.  As expected COD removal was affected
least by the shorter 8-hour shock loadings.  Under normal
operating conditions the last three stages of the rotorzone
removed very little COD.  These stages were incapable of
removing larger COD concentrations for short term hydraulic
shock loadings.
     Nitrification was more inhibited than COD removal.  In
each instance nitrification was quickly inhibited to a
significant degree.  Recovery was slower than that of COD
removal with a minimum of 24 hours required.
     Peak DO concentrations declined during each hydraulic
shock loading.  The decrease began when the flow rate in-
creased and recovered when the flow rate returned to normal.
Sufficient DO concentrations were available for COD removal
and nitrification at all times.
     Biomass stability was excellent throughout the hydraulic
shock loadings.  Unusual or excessive sloughing did not occur
as evidenced by SS concentrations.
     Data from the two 8-hour hydraulic shock loadings does
not indicate excellent reproducibility.  In general the re-
sponse of the Rotordisk to the first 8-hour shock loading was
less severe.  Nitrification and COD removal inhibition were
greater in the second test but restoration of full nitri-
fication was quicker than the first 8-hour experiment.  This
phenomenon can not be adequately explained.

ORGANIC SHOCK LOADINGS

     The organic shock loading phase consisted of two large
step increases in organic loading.  A normal flow rate of
                           562

-------
480 gpd was maintained during each test.•

First Organic Shock Loading

     During the first' organic shock, influent COD remained
fairly consistent.  DO concentrations decreased slightly in
the first 12 hours of sampling whereas the decreases encoun-
tered with the hydraulic shocks were much larger.
     Chemical oxygen demand increased tremendously when pow-
dered milk was added to the primary clarifier.  Six hours
later the large influent COD values were only slightly smaller
and remained constant throughout the rotorzone.  A normal
COD concentration profile was encountered when samples were
collected at 26 hours.
     The addition of powdered milk to the primary clarifier
did not cause a general increase in SS'concentrations.
Fluctuations 'and peak concentrations were found as usual, but
with greater magnitude.  A recognizable SS concentration
trend could not be found.
     Nitrification was seriously inhibited by the organic shock
loading.  When compared to the hydraulic shock loadings, the
response differed in two significant ways.  First the inhibi-
tion of nitrification occurred gradually.   Alkalinity consumption
and effluent NtL-N concentrations increased slowly while NCL-N
production decreased at a slightly more rapid rate.  Secondly,
when the 26 hour sample t was collected, nitrification was
barely evident.  In comparison, nitrification began to revive
when the hydraulic shocks ended, and with the exception of the
first 8-hour shock, were completely recovered in 24 hours.
Full recovery from the powdered milk occurred within 72 hours.

Second Organic Shock Loading

     Trends for the second organic shock closely resembled those
of the first shock.  As envisioned, the effects were less severe
because less powdered milk was added to the primary clarifier.
Dissolved oxygen concentrations decreased slightly for the
first 4 hours but soon recovered.
     Changes in COD concentrations are presented in Figure 8.
Influent COD values fluctuated widely from 108 to 174 mg/1,
but appear small when compared to the high value of 480 mg/1
recorded in the primary clarifier.  The addition of powdered
milk to the primary clarifier caused high COD concentrations
                         563

-------
throughout the unit.  When the intensive sampling period ended
at 12 hours, COD concentrations were decreasing throughout
the unit except for the secondary clarifier.  Secondary clari-  '.
fier effluent concentrations increased from 45 to 159 mg/1,
but declined to near steady-state conditions within 24 hours.
     In contrast to the hydraulic shock loadings a general
increase in SS concentrations was not encountered.  A recog-
nizable pattern was not found, except for the general decrease
from the influent to effluent.
     As shown in Figure 9, effluent NO,.-N had an initial average
of 10 mg/1 and slowly decreased to approximately 5.0 mg/1.
This demonstrates that nitrification was reduced by the second
organic shock loading.  In a manner similar to the first organic_
shock, nitrification was slowly inhibited but recovery was more .
rapid.  Although nitrification did not completely recover within
24 hours, it recovered quicker than for the first organic shock.'
Eecovery was completed within 72 hours as indicated by the
raw data.
     The second organic shock was the only shock loading which  ;
affected the stability of the biofilm.  Six days after the
shock was applied severe sloughing was observed in the second
stage and the first stage had changed from a grey-white color
to grey—brown color.  This color change is the exact opposite   !
of what occurred in the latter part of the start-up period.     .
Since this study did not include biofilm examinations, it is
impossible to conclusively state what caused the color change.

Summary

     Chemical oxygen demand concentrations increased dramatically
throughout the unit during the organic shock loading testing.
The last three stages of the rotorzone were Incapable of re-
ducing COD concentrations.  This observation was also found
during the hydraulic shock loading studies.
     Nitrification was inhibited by both organic shock loads.
The inhibition of nitrification occurred gradually and recovered
slowly in comparison to the response of nitrification to the
hydraulic shock loads.
     Dissolved oxygen concentrations declined only slightly
during the organic shock loadings.  This directly contrasts the
sharp declines encountered in the hydraulic shock loadings.
It is speculated that this can.be attributed to a change in
the organic constituents of the wastewater, i.e., the biofilm
was not acclimated to the change in organic composition.
                            564

-------
565

-------
     Biofilm stability was affected by the second organic
shock loading.  The sloughing was observed six days after the
package plant was exposed to the increased organic loading.
This occurred approximately 72 hours after the data indicated
the treatment system had returned to normal steady-state
operations.  A satisfactory explanation cannot be offered
for this phenomenon.

CONCLUSIONS

     Start-up characteristics of a full-scale RBC unit and the
response of the same unit to controlled shock loadings were
examined in this study.  Based on an analysis of the results
obtained the following conclusions are drawn.

     1.   Start-Up
          a)  Growth of the biofilm began within 24 hours.
          b)  The autotrophic biofilm was easily identified
              by a distinct non-filamentous red-brown color.
          c)  Twenty days were required to achieve steady-state
              conditions in terms of COD removal.
          d)  Approximately 30 days were needed for nitri-
              fication to approach steady-state conditions.
              This statement is based on observable trends
              before wastewater characteristics suddenly
              changed in early June.

     2.   Controlled Shock Loadings
          a)  Hydraulic shock loads depressed DO concentra-
              tions due to decreased hydraulic detention
              times and/or increased biological activity.
          b)  The response to hydraulic shock loadings were
              not very reproducible.
          c)  A DO depression did not occur for the organic
              shock loads.  Most likely this can be attributed
              to a change in the organic constituents of the
              wastewater.
          d)  Nitrification was more easily inhibited by shock
              loads than was COD removal.
          e)  Nitrification was inhibited immediately and re-
              covered more quickly from hydraulic shock load-
              ings when compared to organic shock loadings.
          f)  The attached biofilm was not adversely affected
              by the shock loads.
                            566

-------
ACKNOWLEDGEMENTS

     The authors would like to acknowledge the following
persons and organizations for their help in completing
this research.
     The staff of the Blacksburg and Virginia Polytechnic
Institute Sanitation Authority for their assistance and
cooperation during the course of this study.
     C.M.S. Rotordisk Limited for providing the equipment
essential for this research.
     Tammy E. Altizer and Donna K. Mann for their patience
in typing this manuscript.
     Kenneth D. Farrar for assistance and advice with the
computer graphics presented in this paper.

REFERENCES

1.   Fry, F. F, , "Start-Up and Shock Loading Characteristics
     of a Rotating Biological Contactor Package Plant."
     Thesis submitted in partial fulfillment of requirements
     of Master of Science Degree, Virginia Polytechnic
     Institute and State University (1982).

2.   Antonie, R. L., "Fixed Biological Surfaces-Wastewater
     Treatment."  CRC Press, Cleveland, Ohio (1976).

3.   Poon, C. P. C., Chin, H. K., Smith, E. D., and Mikucki,
     W. J., "Upgrading Trickling Filter Effluents with a
     RBC System."  Proceedings:  First National Symposium/
     Workshop on Rotating Biological Contactor Technology,
     (1980).

4.   Bracewell, L. W., Jenkins, D., and Cameron, W., "Treat-
     ment of Phenol-Formaldehyde Resin Wastewater Using
     Rotating Biological Contactors."  Proceedings:  First
    ' National Symposium/Workshop on Rotating Biological
     Contactor Technology,  (1980).

5.   Ahlberg, N. R., and Kwong, T. S., "Process Evaluation
     of a Rotating Biological Contactor for Municipal Waste-
     water Treatment."  Research Paper No. W2041, Wastewater
     Treatment Section, Pollution Control Planning Branch,
     Ministry of Environment, Ontario, Canada (November
     1974).
                            567

-------
6.   Lue-Hing, C., Obayashi, A. W., Zenz, D. R., Washington, B.,
     and Sawyer, B, M., "Nitrification of a High Ammonia Content
     Sludge Supernatant by Use of Rotating Discs."  Proc. 29th
     Ind. Waste Conf., Purdue University, 245  (1974).

7.   Trinh, D. T., "Exploration Camp Wastewater Characterization
     and Treatment Plant Assessment."  Report No. EPS 4-WP-81-1,
     Environmental Protection Service, Environment Canada,
     Ottawa, Canada (1981).

8.   Srinivasaraghavan, R., Reh, C. W., and Lilegren, S.,
     "Performance Evaluation of Air Driven RBC Process for
     Municipal Waste Treatment."  Proceedings:  First
     National Symposium/Workshop on Rotating Biological
     Contactor Technology, (1980).

 9.  Wu, Y. C., Smith, E. D., and Gratz, J., "Prediction of
     RBC Plant Performance for Municipal Wastewater Treatment."
     Proceedings:  First National Symposium/Workshop on
     Rotating Biological Contactor Technology, 887, (1980).

10.  Welch, F. M., "Preliminary Results of a New Approach in
     the Aerobic Biological Treatment of Highly Concentrated
     Wastes."  Proc. 23rd Purdue Ind. Waste Conf., 428,
     (1968).

11.  Orwin, L. W., and Siebenthal, C. D., "Hydraulic and Organic
     Forcing of a Pilot Scale RBC Unit."  Proceedings:  First   ;
     National Symposium/Workshop on Rotating Biological Con-
     tactor Technology, 119, (1980).

12.  Kinner, N. E., and Bishop, P. L., "High Salinity Waste-
     water Treatment Using Rotating Biological Contactors."
     Proceedings:  First National Symposium/Workshop on
     Rotating Biological Contactor Technology, 259, (1980).

13.  Dupont, R. R., and McKinney, R. E., "Data Evaluation of
     a Municipal RBC Installation, Kirksville, Missouri."
     Proceedings:  First National Symposium/Workshop on
     Rotating Biological Contactor Technology, 205, (1980).
                            568

-------
14.  Antonie, R. L., "Response of the Bio-Disc Process to
     Fluctuating Wastewater Flows."  Proceedings of the 25th
     Purdue Ind. Waste Conf., 425, (1970).

15.  Stover, E. L., and Kincannon, D. F., "One Step Nitrifica-
     tion-Carbon Removal,"  Water_ & Sewage Works, 122, 66,
     (June 1975).

16.  Standard Methods for the Examination of Water and Waste-
     water, 14th Edition, Washington, D.C., American Public
     Health Association (1976).

17.  Sampson, R, J., "Surface II Graphics System."  Kansas
     Geological Survey, Lawrence, KS (1978).

18.  Barr, A. J., Goodnight, J. H., Sail, J. P., Blair, W. H.,
     and Chilko, D, M., SAS User's Guide, SAS  Institute Inc.,
     Raleigh, NC,  (1979).
                           569

-------
     UPGRADING WITH SUBMERGED BIOLOGICAL FILTERS
     OryalQ. Matteson.  Mid-South Distributor,
     Jacksonville, Alabama.
     It may sound like Utopia, but it is now possible to mat-
erially upgrade any aerobic wastewater treatment system by
just adding three types of very simple devices to the second-
ary aeration and settling tanks, with the work done in-house
with off—the—shelf materials.  It makes no difference how big
the systems are or of what type.  And, the cost of doing this,
related to a gpd basis, is low for a very small system and
nominal for a big system.  Further, if the primary treatment
component is then eliminated, except for non-organic grit re-
moval, the total overall treatment will be even better than
it was and much less expensive.  Digester loads will be de-
creased, and in many cases the digesters can be eliminated,
as can tertiary treatment.
     This observation is not based on just theory or on the
results of bench-type experiments.  A working system incor-
porating these techniques and devices has been operating in
Jacksonville, Alabama, for over eight years, with rotifers
clearly visible in its clarifier, consistently producing an
effluent of 10 BOD5/SS mg/1, + or -, even under periods of
forced extreme overloads.  Although this is an extended aera-
tion unit housed in a 1000 gallon tank serving a single home,
the techniques and devices employed are equally adaptable to
any type of aerobic treatment unit or system regardless of
its design or size.  Also, as the gpd rate increases,
                              570

-------
the cost—benefits ratio increases in geometric progression.
     In the first forty-some years after the activated sludge
process was developed in England in 1913, any improvements
made came from operators.  Just so did my concept and its
application come from an operator, me, and not from a design
engineer or a research laboratory.  I started on this route
trying to solve special problems I found in the early 70 *s
while handling what was then the most effective package ex-
tended aeration secondary treatment unit on the market.  How-
ever, the manufacturer finally stopped making it because of
operational problems:  it frequently clogged up.
     Problems with wastewater treatment systems, even in the
best designed and best operating situation, basically come
from widely fluctuating growth patterns of the organisms,
which cause oscillation and continuous imbalance.
     However, as few systems are designed to meet the needs
of or to take advantage of the natural capabilities of the
wastewater treatment organisms, most systems do not fall into
the category of "best designed."  Also, as operators usually
understand so little of the biological/biochemical aspects
of their treatment systems, of whatever type, few systems
can be classified as "best operated."
     Let's see what such established experts as Ross McKinney
and W.W. Eckenfelder have had to say on this matter.
     Ross E. McKinney, in the 1962 edition of his book, Micro-
biology for Sanitary, Engineers, said, "Fundamental microbiol-
ogy offers the means for the sanitary engineer to base the
design of biological waste treatment systems.  It is impor-
tant for the engineer to realize that all microbial systems
operate on the same general biochemical principles and that
the differences between the various biological systems lie in
the environment imposed by the mechanical aspects of the sys-
tem" (1).
     W. Wesley Eckenfelder, Jr., at one of the sessions he
presented at Vanderbilt University in 1971, said, "Waste-
water treatment systems should be designed so that the bugs
would be very happy and thus eat, reproduce and die at a
great rate" (2).
     McKinney also wrote that the activated sludge process is
the most versatile of the biological treatment processes;
that activated sludge is simplicity personified; that no
other treatment process has more advantages or disadvantages;
that the chief disadvantage with it lies in the lack of under-
standing of the basic process by both design engineers and
plant operators; that the design of any biological waste
                           571

-------
treatment system can be made properly only if the designer
has a thorough understanding of the microbiology and the bio-
chemistry of the process; that engineers never consider micro-
biology in the design of waste treatment systems and that the
sanitary bacteriologist is not interested in the design of
treatment systems; and that activated sludge is a pure bio-
logical process and yet biology never entered into its de-
sign or operation until the past few years (3). •
     McKinney's comments in 1962 seemed to indicate that he
then expected that we today should find great improvement in
the design of systems.  To his observations I say, amen; to
his prediction, I would have to say, wishful thinking.  Think
about it; how many municipal or package systems would you say
were designed by a sanitary biologist?  What percentage of
the classes offered for BS degrees in engineering or biochem-
istry or in schools for plant operators, and how many of the
questions on operators' tests, are on the biology or biochem-
istry of waste treatment?
     Apparently McKinney did not think his predictions had
come through by 1972 for the paper he presented at a confer-
ence in Atlanta is filled with comments such as, "The recom-
mended design criteria for activated sludge systems employed
by the various state regulatory agencies clearly demonstrates
the lack of concern for the biological factors affecting act-
ivated sludge." Or, the statement, "At best the design engi-
neer gives lip service to the fact that activated sludge is a
biotreatment process" (4).  He did comment that by then young
engineers were getting some instruction in the "why" as well
as the "how."  He said that most research scientists and
university professors have attempted to make the biological
process more, rather than less., complicated; that there is
going to have to be a drastic change in the philosophy and
attitude of everyone involved.  He contended that operators
have failed to apply basic biological concepts to understand-
ing how their biosysterns should be operated, and that the
design engineers have been of no help.
     I liked what he said about it being necessary to make
the system so simple that everyone can understand it.  Par-
ticularly I liked his comment to the effect that if you have
a system designed around sound biological principles the mi-
crobes will run their part with little operator attention,
and that such attention, mainly directed to control of the
MLSS through balanced sludge wasting, can be designed to
operate automatically.
     Incidentally, these comments of McKinney and Eckenfelder
                          572

-------
all happen to be made in relation to the activated sludge
process.  However, I am confident that they would agree that
they apply to all types of aerobic/anerobic systems/processes
as well.
     McKinney's closing comment takes the prize:  "If we are
to make real progress in solving the current water pollution
problems, we are going to have to recognize that we must de-
sign and construct as simple systems as possible to minimize
problems."
     Well, statements like those of McKinney and Eckenfelder,
and others, and their implications, and a review of all types
of conventional systems (including the ones I had been hand-
ing) are what brought me to develop these new applications of
old techniques and devices.  I set out to apply the common-
sense knowledge gained by hands-on experience to marry what
mechanics, biochemistry and microbiology I acquired to meet
the ultimate demands of wastewater treatment.  I tried out
my ideas until I got the results I wanted, then I got patent
rights, and I now suggest that these techniques and devices
be used to upgrade every aerobic system, of whatever type or
size or state of operation or development.
     The application is new, or I would not have been able
to patent it.  The possibilities these techniques and de-
vices offer are news to the wastewater treatment business or
everyone would have already used them.  They are now avail-
able to everyone as license to use the features of the pat-
ent described herein can be obtained for an extremely small
fee, to match with minimal costs related to the on-the-job
construction and installation of the devices.
     The Environmental Protection Agency puts out a big loose
lea'f publication titled, Process Design Manual for Upgrading
Wastewater Treatment Plants (5).  It emphasizes that only
the most effective design and operation of treatment facil-
ities, with the latest techniques, will meet the future water
quality objectives, and that it is essential that this new
technology be incorporated into the contemporary design of
waste treatment facilities to achieve maximum benefits.  I
am pleased to note that although it seems to contradict the
title, they do recognize that the term "upgrading" should
also apply to systems on the drawing board or in the manu-
facturing process.
     Unfortunately, but, as I expressed earlier, to be ex-
pected, the emphasis in the Foreword and throughout its
content is on "engineering."  There is very little included
on ways to make the "bugs" happier that is not directly
                             573

-------
related to a pump, pipe, or tank, and even when the struc-
tural or mechanical feature is directed to making the "bugs"
happier the "whys" of it are not provided.  Microbiology and
biochemistry seem to have no place in the upgrading process.
     Also emphasis is placed on removing pollutants by the
use of chemicals that, although effective for the purpose,
also always have a collaterally adverse impact on the sys-
tem, rather than in the possible use of cultured varieties of
specific microorganisms products such as LS-1471, BPS-202Q,
or GSHGD-1, or of enzymes such as Septictrine, or when appli-
cable, of the use of the non-toxic algaecide, Cutrine-Plus.
     A most disturbing feature is that, even though it is a
loose—leaf manual, and thus easy to update, there seems to
have been no advances in upgrading techniques of note since
1974, as the EPA manual 1 received in January 1982 seems to
be the same as the one I received in 1974.
     In categorizing the reasons for upgrading, the manual
lists meeting more stringent treatment requirements, and in-
creased hydraulic or organic loads, and to overcome improper
plant design or operations.  All are quite valid needs.  But,
they seem to have ignored more basic reasons .for upgrading
which are common to all systems, even those working just fine:
to just improve system performance, perhaps to get "more
bang" for the "total bucks" invested; or, even though all is
working fine, to upgrade to improve the system's capability
to meet potential shock loads, or to simplify procedures, or
to reduce capital or 0 & M costs.
     When I talk about upgrading I'm addressing any or all
needs, for systems in trouble and for those working OK, for
those in place or those yet to be.
     Incidentally, did you ever examine the wastewater treat-
ment patents in the Search Room of the Patent Office in Ar-
lington, Virginia?  It is an interesting experience that I
recommend to any student of the science of wastewater treat-
ment.  It is a must, I think, for anyone who is thinking
about trying to get an idea patented.
     Even if all you do is look at the pictures, one thing
that will strike you is the complexity of so many of the de-
vices.  You keep expecting to see Rube Goldberg's name on
them.  But then, the devices on operating treatment systems
are complicated—a lot of those patent ideas got incorporat-
ed into systems on the market.  You might think the objec-
tive is to establish a cult that believes that if it is not
complicated it can't do the job; with a creed that simple
ideas or things won't work.
                           574

-------
     I am here to attack and dispel that belief.  My creed is
to keep it simple and inexpensive.  -So, what I have to offer
here to meet the objectives of McKinney and Eckenfelder, et
alii, and I am sure yours as well, are inexpensive' simple de-
vices, simply applied.  That is, they are simple in the sense
of being easy to do and understand and operate.  They are
simple to construct and install and simple to care for.  They
are,also very inexpensive to construct and install, have no
moving parts, and require no 0 & M efforts or costs.
     If your objectives are just to prevent things getting
into the secondary clarlfer, or to upgrade the microbiology
and biochemistry aspects of treatment in activated sludge,
trickling filter or .oxidation ditch systems far more than is
possible in conventional systems, or to increase the organic  .
or hydraulic capacities, or to eliminate some or all of the
features of primary treatment except non-organic grit removal,
or to.reduce final clarifier or digester or trickling filter
or oxidation ditch loads, or to reduce digester problems, or
even to eliminate the digester phase, then you place a series
of "permeable retaining members" in each aeration tank.
     If your objectives are to control velocities and cur-
rents coming into the clarifier element and to accelerate
settling and to concentrate sludge both as to content and
location far more .than is possible in conventional systems,
then you install a "permeable deflection member" in the ,
clarifier or settling tank.                •  •
     If your objectives are to produce a highly clarified ef-
fluent with very low BOD/SS, maintain a state of quiescence
prior to discharge and at the same time to be able to keep
up a return sludge/skimmer rate far- greater than ever pos-
sible in conventional systems, and to also maintain DO
levels in the clarifier capable of supporting animals such
as protozoa and rotifers, then you place .a "permeable re-
straining member" in the clarifier or settling chamber.
     Of course the use of the retaining members in the aera-
tion tanks will materially add to the effectiveness of the
deflection and restraining members in the clarifier, and
vice versa.
     Why, or how, do my devices produce these results?  Let
us for the moment refresh our thinking on the fundamental
microbiology and biochemistry of wastewater treatment,'
     Metcalf & Eddy state:  "By controlling the environment
of the microorganisms, the decomposition of wastes is speeded
                          575

-------
up.  Regardless of the type of waste, the biological treat^
merit process consists of controlling the environment required
for optimum growth of the microorganisms involved...Effective
environmental control in biological waste treatment" is based
on an understanding of the basic principles governing the
growth of microorganisms" (6).
     Without getting too technical, let us review what they
are,
     Bacteria (single-cell plants) grow in a pattern of com-
petition in mixtures of species,each organism and each spec-
ies competing with the others.  The prime factor is competi-
tion for food, with the dominate strains surviving.  Which
are dominate depends upon the type of nutrients available,
the DO, temperature and pH.   In aerobic systems with a pro-
per balance of nutrients the bacteria species which survive
are those that can oxidize the organic matter completely to
carbon dioxide and water.  Both aerobic and faculative bac-
teria will be found in aerobic treatment systems; the facu-
lative use the free oxygen as long as it is available.  Bac-
teria absorb the nutrients, produce and use enzymes to speed
up the processes, metabolize  the organic and inorganic com-
pounds and produce energy and protoplasm, thus producing new
bacteria (usually by splitting into two cells).  For this as
well as for mobility and to just stay alive they require oxy-
gen.  If the oxygen is free (available in water) they produce
energy faster and more efficiently, thus absorbing food fast-
er than if they have to make oxygen out of the wastes.  If
lots of nutrients are available, then available dissolved
oxygen is the principal limiting factor to organic loading—
increase the available oxygen and you increase the eating
rate.  Two of the most critical nutrients for growth are
nitrogen and phosphorous, another is carbon; nitrogen defi-
cient nutrients stimulate filamentous fungi over bacteria,
which prevents good settling.
     Not all organisms are beneficial.  If the DO goes down
below 0.5 nag/1 the faculative bacteria (those which can use
free oxygen or produce it from the wastes, and which always
take the free if it is available) start to metabolize aner-
obically.  At this stage filamentous microorganisms (strict
aerobes) can still use the low rate free oxygen and they
start to predominate; they also dominate when the critical
nutrients of nitrogen and phosphorus are deficient, or at
low pH.  These organisms keep the floe from compacting. Fil-
amentous microorganisms also  tend to predominate over long
                          576

-------
periods when waste food Is absent because they can use the
cell wall material of dead bacteria for food, which natural-
ly the bacteria can not use.
     It is easier to provide food than oxygen, for water does
not take up oxygen easily; turbulence and contact time are
needed.  The idea is to have small bubbles, which have more
surface area for contact than do large ones, bounced around
and broken up, so that the water area which surrounds the bub-
ble and thus is oxygenated will move aside and allow -unsatur-
aged water (oxygen deficient) to contact the bubble.  So, you
need turbulence.  You also need contact time.  If you get the
turbulence by increasing air pressure, or velocity, you lose
transfer efficiency as the contact time is reduced; if you
have too little pressure the bubbles can be too small for ef-
fective transfer or too slow to be able to break the liquid
film which is resistant to the passage of oxygen.  Time of
contact is usually controlled by velocity and vertical depth
(distance traveled).
     The metabolism or growth pattern of bacteria, individu-
ally or as a mass, involve these phases.
     The Lag phase is the time required for them to become
acclimated to a new environment, which could extend for hours
or days.  A surge of food in the morning after, the drop dur-
ing the night, or re-entering the aeration tank from the clar-
ifier usually produces the Lag phase.  We want to cut this
Lag time away down by assuring no low-food periods and reduc-
ing the holding time in the clarifier.
     The Log Growth phase is a period of constant growth, in-
dividual and mass, when there is always more food .than bac-
teria, and the only thing that holds them back is their indi-
vidual capacities to eat and reproduce, and available oxygen.
We want to promote this phase by providing balanced nutrients
all the time and accelerating the oxygen uptake of the water.
These bacteria will handle organic and hydraulic shocks.  How-
ever, these bacteria are too active .to floe as they do not
stick together and thus do not settle well.
     The Stationary or Declining Growth phase is a time when
food and bacteria balance out to a level population matching
growth and death, and then start to have death rates exceed  •
growth.  We have to have this phase, but want to shorten this
part of the cycle as it is less productive than Log Growth.
     The Endogenous or Log Death phase is when food gets pro-
gressively scarcer and the organisms metabolize their own pro-
toplasm without replacement, keeping the mass constant to the
food, which is mainly nutrients from dead cells.  Bacteria are
                             577

-------
not cannibals, they don't eat each other, they are scavengers.
As their energy level drops these bacteria floe.  We have to
isolate this phase to get concentrated sludge.  We also need
to create and hold to this condition in isolation near the
effluent point,  (Self-metabolism is constantly occurring to
some degree in each phase.)  These organisms are very sus-
ceptible to hydraulic/organic shock.
     The relationship between the plants and the animals is
the secret of success in biological systems of any sort.
     Animals, such as worms, snails and crustaceans, eat the
waste and start it on its way to faster oxidation.  The bac-
teria and other plants such as fungi and slimes eat the ani-
mals* wastes and materials coming in with the influent.  Ani-
mals such as the various types of protozoa and rotifers eat
the bacteria; they must have DO equal to or higher than that
for aerobic bacteria.  As bacteria populations develop the
protozoa appear to eat the bacteria, and some organic matter,
and some eat each other.  Some types predominate in the Log
Growth environment and others take charge through the Declin-
ing Growth and Log Death phases, depending on the numbers of
bacteria and their activity (energy level), and the energy
levels required by the different types of protozoa.  The ro-
tifer, a multicell animal, is a strict aerobe and thus re-
quires a higher DO level than the others.  Rotifers eat the
bacteria as well as any small organic particles, such as the
residue from bacteria cells which bacteria cannot process,
You will have rotifers only if the water has low organic con-
tent, so if you have rotifers you have a highly efficient
aerobic biological process.  Some rotifers are macroscopic
and can be seen without magnification.
     All these animals preying on the bacteria keep the bac-
teria population in balance.  As the bacteria population gets
too low then the animals start to die off in proportion to
their available food.  The animals are never able to eat all
the plants or other animals nor ever die off completely so
the cycle is never stopped.
     Treatment is never complete in an aerobic unit unless
bacteria and animals are in proper balance.
     The challenge is to operate the system so that it always
has food available to the bacteria to control a smooth growth
pattern preventing imbalance, and yet to also have a semi-
starvation condition to achieve flocking, and then also to
maintain an aerobic effluent staging area practically void of
organic materials in order to assure a low BOD/SS effluent.
     Further, for high quality treatment it is necessary to
                             578

-------
metabolize materials and compounds that are slow to oxidize,
inorganics, and those which will change from an inorganic to
an organic state with time.  In conventional systems most of
these substances go out with the effluent, some not register-
ing in the BOD test but producing an eventual DO demand on
the receiving waters.  To just meet typical effluent stand-
ards it is necessary to hold the bacteria (activated sludge)
in the system for several days (6-15) striving for sufficient
sludge age (mean cell residence time, or MCRT).  For conven-
tional systems this means MCRT to meet prescribed standards.
However, high quality treatment requires sufficient MCRT to
completely metabolize all materials and compounds (20 days, •
90-95% oxidation).  Nitrification requires at least 6-10 days
MCRT as the nitrifying bacteria have a very slow growth rate.
     Also, the objective is to hold hard-to-oxidize solids
such as grease, hair, seeds, or rubber in the aeration tank.
This is not generally accomplished in conventional systems.
These materials either pass out with the effluent or settle
out in the aeration or secondary clarifier tanks, where they
produce rising and bulking sludges and keep the bacteria out
of the Log Death phase.
     Secondary clarifiers have to be capable of controlling
velocity and turbulence to permit settling and clarification,
to produce concentrated sludge, to be able to remove the
sludge fast enough to keep it from turning anerobic, to keep
the bacteria in condition to minimize the shock when they re-
turn to the aeration tank, and to support protozos and roti-
fers.  In current designs these objectives can seldom be a'ch—
ieved:  sludge is not concentrated; it becomes anerobic; Lag
periods can last several days; it is impessible to achieve
anything like quiescence in the effluent holding area; DO
levels are low.
     When all these most complicated challenges are met we
have, except for pH and temperature control, succeeded in
meeting the objectives of McKinney, et alii:  we have a sys-
tem based on the concept of controlling the environment of
the plants and animals.  But, in order for everyone to be able
to have a system with such an environment we need to change
the engineering of conventional systems, whether in operation
or on the drawing board.
     It can be done now, using my devices.
     Opinions I will offer on the effectiveness of these de-
vices and the effect they have on wastewater treatment will
seem to some to be iconoclastic.  That is good because we
need to jolt many of our sanitary engineers and biochemists
                             579

-------
(remember, no new pages to the EPA manual on upgrading systems
since 1974).  Also, analysis of the possibilities presented
and the mechanics, microbiology and biochemistry involved in
changes of this nature can be a fertile field for laboratory
and operational evaluations.  Further, the possibilities
which are created for new designs for total systems and their
various elements, as well as conversions of existing systems,
by the use of these devices may be a boon to engineers as
well.  We may have opened a veritable treasure house of pos-
sibilities for the waste treatment industry I
     Let us consider what these devices do, and how they
create a system based on biological/biochemical principles,
and Xtfhat changes they can effect, and why.
     Let's follow the sewage through the treatment system.
     Conventional systems use racks, mechanical or otherwise,
mechanical screens and grinders of various types in primary
treatment to prevent materials from getting any further be-
cause they will upset the procedure, clog pumps, pipes and
equipment, and cause delays and generate the need for costly
repairs or replacements.  Primary clarifiers, and sometimes
skimming and preaeration tanks separate solids and liquids.
Capital, energy, and 0 & M costs for these items are high.
     It is not necessary to have all these machines, tanks,
pipes and pumps to accomplish the purposes for which they are
used.
     My "permeable retaining devices," installed in the aera-
tion tank do a better job and cost practically nothing.  But,
even when used in conjunction with all the primary system's
apparatus, they still meet the challenge of providing the
means to materially upgrade any treatment system's operation
at practically no cost.
     The "permeable retaining devices" installed in the aera-
tion tank operate, to put it simply, as do nets or filters.
Framed, they are installed so as to extend from side to side
and from the bottom to above the water level.  They are made
of any inater-ial which is impervious to the wastewater, e.g.,
treated metal, plastic, synthetic fibers.  The mesh in the
network of materials may be formed by any means, such as
punching, weaving, braiding, molding, into whatever size or
shape is desired.  The retaining element can be in any con-
figuration, such as that of a fish net or a furnace air fil-
ter.
     Several devices are installed per tank; the more used
the more effective the results—mainly because each device
functions both as an habitat for organisms and to create
                             580

-------
water turbulence.  The buildup of the materials caught and the
s