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 |