EPA/625/1-86/021
October 1986
Design Manual
Municipal Wastewater Disinfection
U.S. Environmental Protection Agency
Office of Research and Development
Water Engineering Research Laboratory
Center for Environmental Research Information
Cincinnati, OH 45268
-------
Notice
This document has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
publication. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
-------
Contents
Chapter Page
1. Introduction , , ............. 1
1.1 General 1
1.2 Purpose and Objectives 2
1.3 Scope 2
1.4 How to Use This Manual 3
1.5 References 3
2. Need for Disinfection Technologies 5
2.1 Need for Disinfection 5
2.2 Disinfection Criteria 7
2.3 Projected Applications of Disinfection Technologies 8
2.4 References 9
3. Disinfection Alternatives and Options 11
3.1 General Considerations 11
3.2 Selecting a Disinfection Alternative 11
3.3 Chlorination 13
3.4 Chlorine Dioxide 17
3.5 Bromine Chloride ,17
3.6 Ozone 18
3.7 Ultraviolet Light 18
3.8 References 19
4. Kinetics and Hydraulic Considerations 21
4.1 Disinfection Kinetics 21
4.2 Mixing and Contactor Hydraulics 23
4.3 References 28
5. Halogen Disinfection 31
5.1 Coverage 31
5.2 History of Halogen Disinfection 31
5.3 Chemistry and Physical Characteristics of Disinfectants 34
5.4 Analysis of Disinfectant Residuals 50
5.5 Kinetics of Microbial Inactivation 52
5.6 Process Options 55
5.7 Design Coordination 57
5.8 Safety and Occupational Health Considerations 80
5.9 Operation and Maintenance Considerations 82
5.10 Case Studies 84
5.11 References 88
-------
Contents (Cont'd)
Chapter Page
6. Ozone Disinfection 97
6.1 Introduction 97
6.2 Ozone Properties, Chemistry and Terminology 97
6.3 Process Flow Schematics 107
6.4 Ozone Equipment Design Considerations 114
6.5 Ozone Disinfection Process Design Considerations 139
6.6 Safety 151
6.7 References 153
7. Ultraviolet Radiation 157
7.1 Introduction 157
7.2 Disinfection of Wastewaters by Ultraviolet Radiation 164
7.3 Process Design of UV Wastewater Disinfection System 184
7.4 UV Disinfection System Design Example 216
7.5 System Design and Operational and Maintenance
Considerations for the UV Process 223
7.6 References ± 245
IV
-------
Figures
Number Page
1 -1 Sequence of Manual's use 3
3-1 Framework for evaluating site-specific wastewater
disinfection requirements 12
4-1 Chick's Law and deviations 22
4-2 Representation of the residence time distribution (RTD) 23
4-3 Representation of pulse and step inputs and resulting outputs 24
4-4 Techniques for experimentally determining the RTD curves of
UV reactors 25
4-5 Examples of RTD curves for various flow characteristics 26
4-6 Relationship of C-curve to E/uX for small degrees of dispersion 27
4-7 Dye test shows effects of gas flow rate on plug flow
characteristics through a bubble diffuser ozone contact basin 28
5-1 Vapor pressure of liquid saturated chlorine gas 35
5-2 Vapor pressure over liquid bromine chloride 38
5-3 Effect of increased chlorine dosage on residual chlorine and
germicidal efficiency 41
5-4 Effect of chlorine or hypochlorite dose on pH of settled
wastewater 42
5-5 Proposed kinetic mechanism for the breakpoint reaction 45
5-6 Distribution of bromamine species as a function of pH and N:Br
molar dose ratio 47
5-7 Elements of halogen disinfection systems with optional
dechlorination 55
5-8 Schematic of Chloromat™ (Ionics, Inc.) electrolytic
hypochlorite cell 58
5-9 Chlorine expansion chambers 60
5-10 Chlorine manifold and switchover system 61
5-11 Schematic of fixed and variable orifice ejectors 62
5-12 Ejector sizing curve 63
5-13 Frictional losses in solution piping 64
5-14 Ejector sizing curve 65
5-15 Headless thru spray nozzle diffusers 66
-------
List of Figures (Cont'd)
Number Page
5-16 Nomograph for design of multiple perforated diffusers 67
5-17 Details of a submerged weir mixing structure 69
5-18 Details of a hydraulic jump mixing structure 69
5-19 Scale diagram of jump as designed 71
5-20 Residence time distribution functions for contact basins 72
5-21 Types of baffled contact chambers 73
5-22 Vaned serpentine contactor design 73
5-23 Dissipation of chlorine residual and point of sampling for control... 75
5-24 Definition sketch for rectangular contactor 79
5-25 Ton container mounting trunions 81
6-1 The direct reaction of ozone with solutes and a hydroxide ion
(or radical) catalyzed decomposition reaction, leading to reactive
intermediates, compete for ozone 99
6-2 Effluent total coliform concentration versus total residual
oxidants and residual ozone , 102
6-3 Effluent fecal coliform concentration versus off-gas ozone
concentration 102
6-4 Effluent fecal coliform concentration versus product of off-gas
ozone concentration times time 102
6-5 Simplified ozone process schematic diagram 103
6-6 Ozone disinfection process gas and liquid flow diagram 108
6-7 Diagrams showing feed-gas flow of typical ozone disinfection
processes 108
6-8 Oxygen requirement for ozone disinfection compared to oxygen
requirement for activated sludge 109
6-9 Example treatment schemes using ozone disinfection 110
6-10 Fecal coliform survival for rotating biological contactor effluent,
screened effluent, anaerobic lagoon effluent, and strong
wastewater 111
6-11 Effect of water quality and performance criteria on ozone
dosage requirement 112
6-12 Total coliform concentration versus transferred ozone dosage
for various effluents 113
6-13 Total coliform reduction versus log transferred ozone dosage
for nitrified effluents 113
6-14 Cross-section view of prinicipal elements of a corona discharge
ozone generator 114
-------
Figures (continued)
Number Page
6-15a Schematic diagram of a typical power supply to an ozone
generator 115
6-15b Schematic diagram of an ozone producing cell, a "Dialectric" 115
6-16 A free flow of electrons in the discharge gap causes various
reactions with the oxygen molecule 115
6-17 Ozone formation occurs when the voltage level is sufficient to
create a free flow of electrons within the discharge gap 116
6-18 Schematic diagram of three power supply systems typically
used for ozone generation 117
6-19 Typical ways for varying voltage and frequency to an ozone
generator 117
6-20 Specific energy consumption versus ozone concentration for
an air fed ozone generator 120
6-21 Example ozone generator mapping curve using air feed-gas ...... 120
6-22 Specific energy consumption versus ozone concentration for
an oxygen fed ozone generator 121
6-23 Details of a horizontal tube, voltage controlled, water cooled
ozone generator ; 122
6-24 Details of a vertical tube, voltage controlled, water cooled
ozone generator 123
6-25 Details of a vertical tube, frequency controlled, double cooled
ozone generator 123
6-26 Details of an air-cooled, Lowther plate type ozone generator 124
6-27 Example low pressure air feed-gas treatment schematic 125
6-28 Example high pressure air feed-gas treatment schematic 126
6-29 Example nominal pressure air feed-gas treatment schematic 126
6-30 Diagram of a heat-reactivated desiccant dryer with internal
heating coils 128
6-31 Schematic of a heat-reactivated desiccant dryer with external
heating equipment 129
6-32 Pressure swing (heat-less) high pressure desiccant dryer in
purging mode 131
6-33 Ozone transfer efficiency decreases as applied ozone dosage
increases and as ozone demand of the wastewater decreases 132
6-34 An increase in the gas to liquid ratio causes a decrease in ozone
transfer efficiency 133
6-35 Schematic of a 3-stage, bubble diffuser ozone contact basin .. 135
vii
-------
Figures (continued)
Number Page
6-36 Schematic of a turbine mixer ozone contactor 137
6-37 Example diagram of a thermal destruct unit with a heat
exchanger 138
6-38 Specific energy consumption versus off-gas temperature rise
through the thermal destruct unit 138
6-39 Example diagram of a thermal/catalyst ozone destruct unit 139
6-40 Dose response curve for nitrified effluent at Marlborough .142
6-41 Example curve showing the effect of different X-axis
intercepts on transferred ozone dosage requirement 143
6-42 Example curve showing the effect of different slopes on
transferred ozone dosage requirement 143
6-43 Specific energy consumption for a typical air-fed ozone
generation system , 145
6-44 Design example projected ozone production rate for various
operating conditions .149
6-45 Human tolerance for ozone 152
7-1 General description of UV design 158
7-2 Example of closed vessel UV reactor with flow parallel to lamps... 160
7-3 Schematic of quartz UV unit in Vinton, IA 161
7-4 Schematic of quartz UV unit in Suffern, NY 161
7-5 Example of open channel unit at Pella, Iowa with flow directed
perpendicular to lamps 162
7-6 Schematic of quartz UV unit in Albert Lea, MN , 162
7-7 Schematic of open channel, modular UV system 163
7-8 Example of UV system Teflon tubes 163
7-9 Electromagnetic spectrum 174
7-10 Relative germicidal effectiveness as a function of wavelength .... 174
7-11 Relative abiotic effect of UV on E. coli, compared to relative
absorption of ribose nucleic acid 175
7-12 Example of DNA and UV damage to DNA 175
7-13 Schematic representation of the effects of photoreactivation 176
7-14 Effect of particulates on UV disinfection efficiency 185
7-15 The rate K increases with increasing intensity for a given
residence time 186
viii
-------
Figures (continued)
Number Page
7-16 Example of RTD curve developed for unit 2 at Port Richmond by
the step input method 189
7-17 Relationships of velocity, length, and dispersion 190
7-18 Log-log plot of head loss against velocity for unit 2 at Port
Richmond indicating transition from laminar to turbulent flow
regime 191
7-19 Estimates of Reynold's number for 8.9 cm diameter Teflon
tubes 191
7-20 Inlet and outlet considerations for submerged quartz systems 193
7-21 Schematic of bioassay procedure for estimating dose and
intensity 196
7-22 Example of bioassay analysis of commercial UV system to
determine dose 198
7-23 Lamp geometry for point source summation approximation of
intensity , 200
7-24 Illustration of the intensity field calculated by the point source
summation method 202
7-25 Schematic of uniform and staggered uniform lamp arrays 203
7-26 Schematic of concentric and tubular lamp arrays 204
7-27 Effect of Teflon system sizing on the power requirement
efficiency 205
7-28 Uniform lamp array intensity as a function of the reactor UV
density and UV-absorbance coefficient 205
7-29 Staggered uniform array intensity as a function of UV density
and UV absorbance coefficient 205
7-30 Concentric lamp array intensity as a function of UV density
and absorbance coefficient 205
7-31 Tubular array intensity as a function of UV intensity and UV
absorbance coefficient 206
7-32 Effect of centerline spacing on intensity for tubular arrays of the
same UV density 206
7-33 Calculated intensity as a function of UV density for different
lamp array configurations 206
7-34 Example for deriving an estimate of the residual fecal coliform
density associated with particulates as a function of
suspended solids 209
7-35 An example for deriving an estimate of the inactivation rate for
fecal coliforms as a function of the calculated average intensity... 210
ix
-------
Figures (continued)
Number Page
7-36 An example of the comparison of disinfection model estimates
to observed effluent fecal coliform densities 210
7-37 Correlation to estimate the spherical absorbance coefficient
from direct unfiltered absorbance coefficient 214
7-38 Comparison of inactivation rate estimates from several
wastewater treatment plants 214
7-39 Estimation of Np from several plants 215
7-40 Photoreactivation effects for total and fecal coliform at
Port Richmond 216
7-41 Example of calculating the limiting U and X on the basis of
head loss ; '....- 220
7-42 Predicted performance as a function of loading for
design example 221
7-43 Effect of bulb wall temperature on the UV output of a low
pressure mercury arc lamp 225
7-44 Nominal lamp output as a function of arc length 226
7-45 Measurement and analysis technique for estimating the total
UV output of a lamp 226
7-46 Sketch of lamp monitoring setup 227
7-47 Energy sinks in UV reactor 228
7-48 Approximation of average lamp UV output at 253.7 nm with
time for quartz systems, accounting for lamp aging and surface
fouling 230
7-49 Estimate of Teflon transmittance by use of a UV detector 231
7-50 Test setup to conduct actinometry experiments 231
7-51 Example of chemical actinometry tests to determine Teflon UV
transmission 232
7-52 Effect of wall thickness as determined by chemical
actinometry 232
7-53 Example of radiometer intensity readings as a function of UV
absorbance at Port Richmond ,, 234
7-54 Schematic of in-place chemical cleaning system at Suffern, NY ... 236
7-55 Comparison of ultrasonic cleaning performance at Suffern, NY ... 238
7-56 Estimate of labor requirements for the operation and
maintenance of UV systems 244
-------
List of Tables
i ' . •
Number Page
2-1 Typical Composition of Domestic Wastewaters 6
2-2 Typical Influent Concentration Ranges for Pathogenic and
Indicator Organisms 6
2-3 Microorganism Reductions by Conventional Treatment
Processes 6
2-4 Secondary Effluent Ranges for Pathogenic and Indicator
Organisms Prior to Disinfection 7
2-5 Number of Wastewater Treatment Plants by Flow Capacity .... 8
2-6 Summary of Wastewater Treatment Processes in the United
States 8
3-1 Major Factors in Evaluating Disinfectant Alternatives 13
3-2 Applicability of Alternative Disinfection Techniques 14
3-3 Technical Factors and Feasibility Considerations 14
3-4 Compilation of Department of Transportation Accident Data 16
3-5 Percent Distribution of Chlorine Shipped by Transportation
Mode and by Shipment Weight 16
3-6 Breakdown of Chlorine Shipments by Transportation Mode and
Container , 16
3-7 Accident Rates per Metric Ton-Km 17
5-1 Early Geographic Distribution of Chlorination Facilities 33
5-2 Development of Chlorination Installations for Wastewater
Treatment 33
5-3 Physical Properties of Chlorine 34
5-4 Thermodynamic Functions of Free Chlorine Species 36
5-5 Equilibrium Constants for Free Chlorine 36
5-6 Physical Properties of Chlorine Dioxide 37
5-7 Comparison of Properties of BrCI and Br 39
5-8 Ratio of Dichloramine Combined Chlorine to Monochloramine
Combined Chlorine as a Function of pH and Applied Molar Dose
Ratio (Equilibrium Assumed) 44
"•.r
5-9 Summary of Kinetics of HOCI" and OCI" Reduction by
Miscellaneous Reducing Agents 46
5-10 Physical Properties of Sulfur Dioxide 48
xi
-------
List of Tables (continued)
Number Page
5-11 Chick-Watson Parameters for Microbial Inactivation by Chlorine ... 53
5-12 Parameters in the Collins et al. Model Describing Wastewater
Coliform Inactivation by Chlorine 55
5-13 Computation of Length to Jump (Lj) 70
5-14 Physical Dimensions of Chlorine Gas Containers 80
5-15 Gas Phase Chlorine Concentrations Evoking Specific Effects 81
5-16 Neutralization Requirements for Chlorine Containers 82
5-17 Troubleshooting Guide 85
5-18 O&M Schedule, Sacramento Regional Wastewater Treatment
Planl: 88
6-1 U.S. Municipal Wastewater Treatment Plants Using Ozone ........ 98
6-2 Properties of Pure Ozone 99
6-3 Solubility of Ozone in Water 100
6-4 Terminology for Measured Ozone Parameters 104
6-5 English Unit Equivalents for Ozone Concentration 105
6-6 Terminology for Calculated Ozone Parameters 105
6-7 Moisture Content of Air for Air Temperatures from
-80° to 40°C 106
6-8 Atmospheric Pressure at Different Altitudes 106
6-9 Effect of Short-Circuiting on Disinfection Peformance 134
6-10 Reported Design Applied Ozone Dosages for Various Wastewater
Treatment Plants 141
6-11 Reported Operating Applied Ozone Dosages for Various
Wastewater Treatment Plants 141
6-12 Summary of Dose/Response Curve Slopes and Intercepts for
Various Ozone Disinfection Research Studies 142
6-13 Ozone Disinfection System Criteria for Design Example Problem .. 147
6-14 Transferred Ozone Dosage Calculations for Design Example 149
7-1 Municipalities That Have Received I/A Funds for Designing
and/or Constructing UV Disinfection Facilities 164
7-2 Summary List of Facilities in the U.S. or Canada Utilizing UV
Disinfection Which are in Design 165
7-3 Summary List of Facilities in the U.S. or Canada Utilizing UV
Disinfection Which are Under Construction 168
7-4 Summary List of Facilities in the U.S. or Canada Utilizing UV
Disinfection Which are in Operation 170
7-5 Summary of UV Installations in U.S. in Operation, Construct,
or Design Phase 173
xii
-------
Tables (continued)
Number Page
7-6 Summary of Reynolds Number Estimates for Different Lamp
Configurations 193
7-7 Examples of Low Pressure Mercury Arc Lamp Specifications 196
7-8 Wastewater Treatment Plants Which are Sources of
Wastewater Characterization Data 212
7-9 Initial Bacterial Density Before Disinfection 213
7-10 Treated Effluent Characteristics from Several Wastewater
Treatment Plants 213
7-11 Example UV Disinfection System Design Criteria 217
7-12 Estimate of Intensity and Rate K for Design Example 220
7-13 Calculations of Performance on the Basis of Loading for the
Design Example ,. 222
7-14 Estimation of Reactor Performance Requirements for the
Design Example 222
7-15 Sizing Calculation for the Design Example 222
7-16 Reactor Sizing Requirement for the Design Example 223
7-17 Effects of Fouling on the UV Transmittance of Quartz 231
7-18 UV Transmittances of New and Used Teflon as Determined by
Chemical Actinornetry , 232
XIII
-------
-------
Acknowledgments
Many individuals contributed to the preparation and review of this manual.
Contract administration was provided by the U.S. Environmental Protection
Agency, Water Engineering Research Laboratory (WERL), Cincinnati, Ohio.
Authors:
Enos L. Stover, Oklahoma State University,
Stillwater, Oklahoma
Charles N. Haas, Illinois Institute of Technology,
Chicago, Illinois
Kerwin L. Rakness, Process Applications, Inc.,
Fort Collins, Colorado
0. Karl Sensible, HydroQual, Inc.,
Mahwah, New Jersey
Project Officer:
Albert D. Venosa, EPA—WERL,
Cincinnati, Ohio
Technical Peer Reviewers:
Karl E. Longley, California State University—Fresno,
Fresno, California
Louis A. Ravina, Riddick Associates, P.C.,
Tappan, New York
C. Michael Robson, Camp, Dresser and McKee,
Louisville, Kentucky
R. Rhodes Trussell, James M. Montgomery Engineers,
Pasadena, California
Other Reviewers:
Edward J. Opatken, EPA—HWERL,
Cincinnati, Ohio
Denis J. Lussier, EPA—CERI,
Cincinnati, Ohio
Orville Macomber, EPA—CERI,
Cincinnati, Ohio
James F. Wheeler, EPA—OMPC,
Washington, DC
John Maxted, EPA—OMPC,
Washington, DC
Alan B. Hais, EPA—OMPC,
Washington, DC
Dennis R. Ohlmansiek, Emery Chemicals, Inc.,
Cincinnati, OH
George C. White,
San Francisco, California
XV
-------
Sidney Ellner, Ultraviolet Purification Systems, Inc.,
Bedford Hills, New York
G. Elliott Whitby, Trojan Technologies, Inc.,
London, Ontario, Canada
L. Joseph Bollyky, International Ozone Association,
Norwalk, Connecticut
Ronald L. Laroque, Hanken Environmental Systems, Inc.,
Scarborough, Ontario, Canada
Carl W. Nebel, PCI Ozone Corporation,
West Caldwell, New Jersey
Frederick C. Novak, Metcalf and Eddy, Inc.,
Boston, Massachusetts
Other Contributors:
Enos L. Stover gratefully acknowledges the assistance of
Brent W. Cowan for project engineering assistance.
Charles N. Haas gratefully acknowledges the assistance of
two of his students in compiling and editing portions of the
chapter on halogenation and dehalogenation, Sandaram B.
Karra and Kirankumar V. Topudurti.
Kerwin Rakness gratefully acknowledges the assistance of
Bob A. Hegg of Process Applications, Inc., in data develop-
ment and document review, and the assistance of Robert C.
Renner of Process Applications, Inc., in document review.
O. Karl Scheible gratefully acknowledges the assistance of
Maureen Casey, Wilfred Dunne, and William Leo in the
development and analysis of data.
XVI
-------
Chapter 1
Introduction
1.1 General
Chlorination has long been the accepted and pre-
ferred method of disinfection for both water and
wastewater in the United States. Major factors in the
implementation of chlorination were efficiency and
low cost when compared to other means of disinfec-
tion. New considerations that are now being exam-
ined very closely are the environmental and biota
impacts of chlorination versus attainment of public
health. The U.S. Environmental Protection Agency
recognized the adverse effects of chlorination of
wastewaters and reported the following conclusions
in its Task Force Report (1):
• disinfection requirements (i.e., the need) should be
evaluated on a case-by-case basis with considera-
tion of beneficial uses and criteria.
• chlorine and subsequent residuals are extremely
toxic to aquatic wildlife.
• chlorine and organic compounds form chloro-
organics and are potentially toxic to man.
In effect, the policy and summary conclusions have
not changed since 1976. The agency policy essen-
tially states:
• disinfection should not be required in those
instances where significant benefits are not dem-
onstrated.
• prospective disinfection benefits must be weighed
against the environmental risks and costs.
• chlorine should be considered only when there are
public health hazards to control.
• alternative disinfection methods and/or dechlor-
ination must be considered when and where public
and aquatic health and/or life impacts co-exist.
With the enactment of the Clean Water Act, its
amendments and new assertions on site specific
criteria, there is no longer a "generic" disinfection
procedure. The realization that chlorine is toxic to
aquatic life and reacts with precursors of trihalo-
methanes (THMs) and other chlorine oxidizable and
substitutable compounds in wastewater effluents
has caused considerable concern. As a result, the
existing practices of disinfection have fallen under
close scrutiny. Disinfection studies, chemical char-
acterizations, and other evaluations are ongoing even
nowto assess more accurately the impact of present-
day disinfection practices.
That a specific disinfection process has been reported
as troublesome at a particular facility could very well
be the result of many factors, some or all of which
may not have been considered prior to engineering
and installation. Critical factors such as disinfection
type, tank configuration, contact time, conditioning
and pre-conditioning criteria, power costs, mainte-
nance requirements, and others may not have been
fully understood or evaluated. Recent information
released by the EPA and the General Accounting
Office(GAO) has indicated that many of the municipal
wastewater treatment plants in the United States
have not been meeting their effluent discharge
limitations (2). An estimated 18,000 municipal waste-
water treatment plants were in operation or under
construction by the end of 1979. More than half of
these plants were not functioning as designed with
respect to BOD, TSS, and fecal coliform removal.
Reasons for plant problems were site-specific and
complicated with long-term violations determined to
be due to a combination of problems, as follows:
operation and maintenance (O&M) deficiencies.
equipment problems.
infiltration and inflow.
industrial waste overloads.
design deficiencies.
With respect to design deficiencies, four basic types
of design problems that have been implicated include
the following:
• limited state-of-the-art during the design phase.
• lack of expertise.
• sampling errors prior to plant design.
• time and funding constraints.
All these problems have been responsible for design
deficiencies associated with wastewater disinfection
facilities. Limited state-of-the-art during the design
phase has been a real problem associated with the
design of alternative disinfection processes, such as
ozone and ultraviolet light. With the increasing need
for wastewater treatment plant upgrading modifica-
tions, including more stringent disinfection criteria.
-------
elimination of design deficiencies and expanding the
state-of-the-art for process design become very
important considerations.
The importance and need for a comprehensive design
document for municipal wastewater disinfection is
apparent. A single source information document does
not exist today that will assist and direct the design
engineer, and other concerned parties, in an educated
and informed decision, selection, and design proce-
dure for the best disinfection alternative process
suited to site specific constraints. A major tool that is
lacking in the area of disinfection is a concise and
unbiased design manual dealing with the disinfection
alternatives, and the design of the most effective
alternative for the application desired. The purpose of
this manual is to bridge this gap and present the
engineer with comprehensive design guidance for
the implementation of the most efficient and cost
effective disinfection method needed for a particular
site.
1.2 Purpose and Objectives
Over the past 14 years the EPA has actively pursued
an extensive program to investigate alternatives to
chlorination of wastewaters through internal and
externally funded research and development. This
research has produced a wealth of information on
disinfection alternatives from literature surveys,
laboratory studies, pilot plant studies, and full-scale
investigations.
The EPA contractors, grantees, and project officers
involved in this extensive effort have developed
results and experience in design and operation of the
most applicable disinfection alternatives. The EPA
disinfection program has advanced to the point where
the wealth of information and expertise developed on
disinfection alternatives can be compiled into a single
document for evaluation of process alternatives and
design of the selected disinfection process. The
primary objective of this document is to provide a
comprehensive process design manual for waste-
water disinfection to be used by design engineers,
regulatory and review agencies, and owners and
operators of disinfection processes. The development
of this document has been based on both the
theoretical and practical application of process design
criteria for the implementation of disinfection tech-
nology in municipal treatment facilities. The informa-
tion in this process design manual for disinfection has
been developed from the following sources:
• Literature
• Extensive experience of the individuals preparing
the manual from participation in disinfection
research and development.
• Active involvement and participation by the EPA
project officers involved in the disinfection pro-
gram.
• Information and results developed from the EPA
internal efforts and externally funded research
programs.
• Communications with investigators and equipment
manufacturers.
• Plant site visits and discussions with operating
personnel.
The disinfection alternatives that are the subject of
this manual and of which sufficient information
exists for design purposes are halogenation/dehalo-
genation (including chlorine, bromine chloride, and
chlorine dioxide), ozonation, and ultraviolet irradia-
tion. In the future the acceptable risks associated with
the various disinfection alternatives and the levels of
disinfection required may be refined and thus create
disinfection criteria other than those in present use.
The design approaches presented in this manual are
flexible and applicable to the various levels of
disinfection presently required as well as those that
may be required in the future by covering the range of
no detectable coliforms up to any desired level in
wastewater effluents.
1.3 Scope
The first part of this manual presents an overview of
the disinfection process, the types of disinfecting
agents, and the advantages and disadvantages of
each.The manual then discusses in separate chapters
how to design each of the primary disinfection
alternatives. Although many more alternative dis-
infection methods have been identified, only those
considered in this manual are cost effective and
presently being implemented. The following criteria
were used to select the disinfection processes
discussed: technical feasibility, flexibility, reliability,
complexity, safety, costs, environmental impacts, and
hazardous material impact or formation.
The primary thrust of this manual is to present
thorough design guidelines on ozone, ultraviolet
light, chlorination, and chlorination/dechlorination
facilities. Bromine chloride and chlorine dioxide are
discussed in the manual but not to the level of detail
as the previous alternatives. Bromine chloride ap-
pears to be as flexible and effective as chlorine;
however, some questions still remain regarding the
use of bromine chloride, such as equipment reliability
and future chemical cost. Full-scale operating data
and a proven track record for bromine chloride
disinfection are still lacking. Chlorine dioxide is a
proven bactericide and virucide and has certain
features making it attractive for drinking water
treatment. However, for wastewater disinfection,
chlorine dioxide is not so attractive, principally
because of its high cost. Also, its persisting residual is
toxic and may require removal through chemical
reduction prior to discharge of the disinfected ef-
fluent.
-------
Each of the respective chapters presents a brief
historical background of the development and use of
the particular 'disinfectant of interest. The chapters
then review process chemistry and disinfection
kinetics. Analytical measurement methodology is
discussed relative to both wet chemistry and instru-
mentation analysis, including disinfection process
control concepts. Then the actual process design
factors, considerations, and experience are presented
along with case histories, operation and mainte-
nance, and safety considerations.
1.4 How to Use this Manual
After discussing the need for wastewater disinfec-
tion, along with disinfection criteria in Chapter 2, the
manual in Chapter 3 discusses the disinfection
alternatives and options considered feasible for
municipal wastewater disinfection. A qualitative
screening procedure is presented in Chapter 3 for
evaluating and selecting an appropriate disinfection
technology for a specific application. Examples are
presented where each of the primary disinfection
technologies is selected. In Chapter 3, the predom-
inant advantages and disadvantages of each of these
disinfection alternatives are also discussed.
After a disinfection alternative has been selected for a
specific application, the manual user can proceed
directly to the appropriate chapter on design of that
technology. Halogen disinfection is discussed in
Chapter 5, ozone disinfection is discussed in Chapter
6, and UV disinfection is discussed in Chapter 7.
However, before the user proceeds to the appropriate
design chapter, he may want to review Chapter 4 on
disinfection kinetics and hydraulic considerations.
Chapter 4 presents a general overview of kinetic
considerations, mixing requirements, and contacting
requirements that apply to each of the disinfection
alternatives. The suggested sequence to follow for
use of this disinfection design manual is presented in
Figure 1.1.
After the appropriate disinfection technology has
been selected for design, the user can proceed
through the specific design chapter indicated in
Figure 1.1. Each design chapter includes a brief
history, overview, and application discussion of the
respective disinfection technology. More detail on the
fundamental chemical and kinetic aspects of each
disinfectant is presented, compared to the general
overall discussion in Chapter 4. Specific aspects of
mixing and contactor hydraulic considerations are
also presented as they relate to the specific disinfec-
tion technology under discussion. Following this,
each chapter presents an in depth discussion of
equipment design considerations, including materials
of construction, similarities and differences among
available equipment suppliers, and factors to be
aware of in specifying equipment. Process design
Figure 1 -1. Sequence of Manual's use.
Need for Disinfection
Technologies
(Chapter 2)
' Disinfection Alternatives
and Options
(Chapter 3)
Kinetics and Hydraulic
Considerations
(Chapter 4)
procedures are then presented, detailing the mechan-
ics of how to size and specify equipment based on
dose requirements, disinfectant demand, NPDES
permit limitations, and cost minimization. The latter
section of each chapter is the heart of the manual, but
the previous sections provide essential technical
back-up for making the proper design decisions.
1.5 References
1. Disinfection of Wastewater Task Force Report.
EPA-430/9-75-012, U.S. Environmental Pro-
tection Agency, Washington, DC, 1976.
2. Morrison, A. GAO Finds Massive Failure of
Wastewater Treatment Plants. Civil Engineer-
ing ^(W? 4, 1981.
-------
-------
Chapter 2
Need for Disinfection Technologies
2.1 Need for Disinfection
Population increases and much greater demands for
water supply and water recreational uses within the
past 15 to 20 years have significantly increased the
opportunity for human exposure to wastewaters
being discharged into the environment. Natural
safeguards, such as dilution and distance or time
before contact or use, have been reduced due to the
large volumes of wastewater being discharged and
the number of discharge locations. Domestic waste-
waters carry human pathogens excreted in the fecal
discharges of infected individuals. Even treated
effluents can affect sources of domestic water supply,
recreational waters, and shellfish growing areas.
Disinfection is necessary to reduce transmission of
infectious diseases when human contact is probable.
Chlorine in the past has been used almost universally
as the disinfectant for wastewaters. However, studies
have shown that chlorine and its by-products can be
toxic to aquatic life, repel and deny spawning grounds
to anadromous fish, and decimate fish larvae and
other forms of life. Out of these concerns, questions
have arisen relative to both disinfection needs in
general and disinfection with chlorine. Although the
need for disinfection is site specific, in general,
disinfection is considered to be a beneficial unit
process and required for most discharge applications.
The organisms of greatest concern in human expo-
sure to wastewater-contaminated environments are
the enteric bacteria and viruses and the intestinal
parasites. Diseases that are spread via water con-
sumption and/or contact can be severe and some-
times crippling. Bacterial diseases such as salmo-
nellosis (including typhoid and paratyphoid fevers),
cholera, gastroenteritis from enteropathogenic
Escherichia coli, shigellosis (bacillary dysentery) and
viral diseases caused by infectious hepatitis virus,
poliovirus, coxsackieviruses A and B, echoviruses,
reoviruses, and adenoviruses may be contracted by
contact with or by consumption of wastewater
contaminated water supplies (i.e., potable and/or
recreational).
The alternative of discharging wastewater that has
not been disinfected allows discharging of pathogenic
organisms and other resultant hazards posed to
humans. Pathogenic organisms, by definition, cause
disease in human beings. Waterborne transmission
of these disease-causing organisms can occur via
four pathways(1):
• direct ingestion of untreated water.
• direct ingestion of treated drinking water.
• ingestion of aquatic food species infected with
pathogens absorbed from contaminated waters.
• invasion resulting from skin contact with contam-
inated water.
The first three pathways are sometimes classified as
the fecal-oral route. The second pathway described
above occurs when a drinking water treatment
system fails or the integrity of the water distribution
system is violated. The fourth pathway is likely to
result in skin, mucous membrane, or urinary tract
infections but is seldom implicated in gastrointestinal
illness in the United States. The risk of disease by
exposure to wastewater effluent in recreational
water, especially non-disinfected effluent, is not well
established on epidemiological grounds; however,
recent work has demonstrated a cause-effect rela-
tionship via this pathway (2).
Wastewater treatment plants historically discharge
their effluents to natural receiving streams that are
often tributaries of larger recreational bodies of water
or that are used as water supply sources by down-
stream communities. City potable water supplies are
often extracted from these tributaries or lakes,
physically and chemically treated and distributed to
customers. The only protection that the recreational
users receive is the hope that the wastewater was
adequately disinfected prior to discharge. So long as
disinfection guidelines and standards are being met
at the sources, public safety and water quality will be
protected. Factors that influence and potentially bias
this type of rational thinking are equipment design,
operator training, equipment dependability, operator
attention, and others.
It is significant that infectious hepatitis has main-
tained a level of 50,000 to 60,000 cases per year in
the United States (3), while typhoid fever dropped
from 2000 cases in 1955 to300 in 1968 (4). There are
more than 100 viruses excreted in human feces that
have been reportedly found in contaminated water, of
which any one could cause a waterborne disease.
-------
A wide variety of enteric pathogens including viruses,
bacteria, and parasites is known to occur in all
community-derived wastewaters. Infective dose
studies with a variety of enteric organisms have been
, conducted overthe past 30 yearsin human volunteers
(5). The widest dose range required to produce a
response was found with the bacterial agents.
Salmonella spp. required the largest dose with the
ingestion of 105 to 108 cells needed to produce a 50
percent illness rate. In contrast, three species of
Shigella produced illness in a significant percent of
subjects dosed with 10 to 100 cells. Protozoan
infections have been produced with Entamoeba coli
and Giardia lamblia dosed in gelatin capsules at the
level of 1 to 10 cysts. Enteric viruses have produced
infection at low dosage levels via oral ingestion,
inhalation, and conjunctiva! exposure. These studies
have clearly shown that specific enteric organisms of
all three classifications, i.e., bacteria, viruses, and
animal parasites, can produce infections at relatively
low exposure levels. Available data are insufficient to
evaluate the actual health hazards that exist for
individuals exposed to wastewater subjected to
various degrees of treatment and dilution; however,
the data indicate thai: enteric pathogens can cause,
infections at exposure levels found in undisinfected
wastewater.
Sobsey reported the viral content of wastewater
(United States origin) ranged from 2 to 3 tp more than
1000 infectious unite/100 ml sample (6). Peak
periods were observed to occur in late summer and
early fall. Most enteric virus concentrations have
been isolated in heavily polluted surface waters, but
Berg and coworkers were able to detect enteric
viruses in the Missouri River having fecal coliform
concentrations as low as 60/100 ml (3). Although
very little quantitative information is available with
respect to the concentration levels of enteric viruses
in the United States surface and groundwaters, much
evidence leads to the indication that wastewaters are
a primary source.
It is difficult to accurately identify general municipal
wastewater characteristics due to differences in
locations, water uses, seasonal variations, diurnal
variations, etc. However, it is important that the
design engineer have a knowledge of the wastewater
characteristics for which he is designing a disinfec-
tion system. He must have a knowledge of the
composition of the conventional parameters as well
as the concentrations of the pathogenic agents or
indicator organisms for which he is designing the
disinfection system. The influent pathogen or indi-
cator organism concentration is a critical parameter
for design of any of the disinfection technologies.
Typical compositions of raw wastewater through
various levels of treatment are summarized in Table
2-1. The total or fecal coliform (indicator organism)
level is a critical design parameter that should be
determined on a site specific basis where possible.
The total bacterial population of human feces has
been estimated to reach a density of 101? organisms
per gram (3). The density range of fecal coliforms in
human feces has been estimated at 106 to 109 per
gram with total coliforms estimated at 107to 109per
gram. Hubley et al. (1) presented an assessment of
the concentration ranges of certain organisms in
domestic wastewater, reduction through primary and
secondary treatment, and estimated secondary
treated effluent concentrations, as summarized in
Tables 2-2, 2-3, and 2-4. If one assumes that the
pathogenic organisms are removed in proportion to
Table 2-2. Typical Influent Concentration Ranges for
Pathogenic and Indicator Organisms (7) ,
Nurnber/100 ml
Organism
Minimum
Maximum
Total Coliforms
Fecal Coliforms
Fecal Streptococci
Virus
1,000,000
340,000
64,000
0.6
49,000,000
4,500,000
10,000
Table 2-3. Microorganism Reductions by Conventional
Treatment Processes (8)(9)
Microorganisms
Total coliforms
Fecal coliforms
Shigella sp.
Salmonella sp.
Escherichia coli
Virus
Entamoeba histolytica
Primary
Treatment
Removal
(%)
<10
35
15
15
15
<10
10-50
Secondary
Treatment
Removal
(%)
90-99
90-99
91-99
96-99
90-99
76-99
10
Table 2-1. Typical Composition of Domestic Wastewater
Water Quality
Raw Wastewater
Primary Effluent
Secondary
Filtered Secondary
Nitrified
Filtered Nitrified
BOD
mg/l
200
130
20
12
7
5
TSS
mg/l
200
100
20
5
10
5
Total N
mg/l
35
30
20
18
18
18
Total
Coliforms
#/100 ml
•|07-108
107-108
105-106
104-105
104-10S
104-105
Fecal
Coliforms
#/100ml
106-107
106-107
104-.1Q5
103-105
103-105
103-105
-------
Table 2-4. Secondary Effluent Ranges for Pathogenic and
Indicator Organisms Prior to Disinfection
Number/100 ml
Organism
Minimum
Maximum
Total Conforms
Fecal Coliforms
Fecal Streptococci*
Viruses
Salmonella sp.
45,000
11,000
2,000
0.05
12
2,020,000
1,590,000
146,000
1,000
570
*Assuming removal efficiencies for fecal streptococci similar to
the fecal coliform removal efficiencies
the indicator organisms (total and/or fecal conforms),
conventional treatment of domestic wastewaters
without disinfection cannot be considered sufficient
for removal and control of human pathogens where
beneficial uses and body contact occur.
2.2 Disinfection Criteria
In 1972 Congress enacted the Federal Water Pollu-
tion Control Act (PL 92-500) (amended in 1977, Clean
Water Act, PL 95-217), to "restore and maintain the
chemical, physical and biological integrity of the
nation's waters." Effluent standards were to be
imposed on industrial and municipal wastewater
dischargers based on the limits of current technology.
Today, secondary treatment is the minimum level
required for municipal wastewater treatment plants.
The Secondary Treatment Information Regulation
was promulgated by the EPA in 1973 as Part 133,
Title 40 of the Code of Federal Regulations (40 CFR
133). Included in the definition of secondary treat-
ment was a fecal coliform limitation, but this was
omitted from the CFR in July 1976, thereby delegating
to the individual states, the responsibility to establish
site specific water quality criteria.
Municipal wastewater treatment plants often dis-
charge into potable water sources ancl recreational
bodies of water. The disinfection practices for potable
water have helped control waterborne disease out-
break from such sources; however, wastewater
disinfection through the destruction of pathogenic
agents, provides a barrier to possible waterborne
disease before the wastewater is released into the
environment. Public health and protection of the
aquatic and human environments should be the
overriding considerations affecting disinfection re-
quirements. With the site specific constraints and
considerations, such as water quality, seasonal
versus year round disinfection requirements, etc.,
each individual state has established its own waste-
water disinfection policies. A review of the types of
disinfection criteria of various states follows.
As a result of the 1924-1925 typhoid epidemic
resulting from consumption of contaminated shell-
fish, and because of several other less severe
incidents related to shellfish contamination, a na-
tional shellfish sanitation program was established
by the United States Public Health Service to protect
the quality of coastal shellfish beds. The 1964
National Shellfish Sanitation Workshop defined the
bacteriological water quality for approved shellfish
areas (those areas where shellfish may be taken for
direct marketing) as having a total coliform mean
Most Probable Number (MPN) not exceeding 70 per
100 ml. In 1977 a fecal coliform median concentration
of 14 MPN per 100 ml with no more than 10 percent
of the samples exceeding 43 MPN per 100 ml was
recommended for approved shellfish waters by the
EPA. These fecal coliform levels were based on an
extensive total coliform to fecal coliform correlation
study and were subsequently accepted for waste-
water effluent discharges into shellfish waters.
The two effluent guidelines or standards most
commonly used by the various states and territories
in the United States have been a total coliform value
of 1000 per 100 ml and a fecal coliform limit of 200
per 100 ml. The microbial guideline for primary
contact recreational waters as adopted by many
states requires that the fecal coliform content shall
not exceed a geometric mean of 200 per 100 ml and
that no more than 10 percent of the total number of
samples taken during any 30 day period shall exceed
400 fecal coliforms per 100 ml.
Some states have adopted even more stringent
disinfection requirements than 14 MPN per 100 ml
for mean effluent fecal coliform levels. Since No-
vember 1976, the Maryland Department of Health
and Mental Hygiene has required all facilities plans to
provide for disinfection to total coliform levels of 3
MPN per 100 ml unless a special exemption is
granted by the department (10). This high level
disinfection is applicable to both coastal and inland
waters but is geared toward protection of the state's
intensive shellfish industry. The California State
Department of Health "Uniform Guidelines for
Sewage Disinfection" incorporates consideration of
dilution, receiving water quality, and beneficial uses
in disinfection requirements, which results in varying
coliform standards for different discharge situations
(11). The standard for nonrestricted recreational uses
of wastewater and shallow ocean discharges in close
proximity to shellfish areas specifies a 7-day median
total coliform value of 2.2 per 100 ml or less at some
point in the treatment process. The 2.2 total coliform
per 100 ml standard applies to those discharge
situations where reclaimed wastewater (i.e., 100
percent wastewater effluent) is impounded for body
contact recreation activities, where wastewater is
discharged to ephemeral streams with little or no
dilution, and where body contact recreation has been
designated as a beneficial use of the stream.
The various states have fecal coliform standards
ranging from less than 2.2/100 ml up to 5,000/100
-------
ml and total coliform standards from less than
2.2/100 ml up to 10,000/100 ml. Some states have
seasonal disinfection requirements, and Illinois, at
the time of writing this manual, has no bacterial
standards for certain, designated non-contact waters.
These disinfection criteria have been established
relative to discharge stream water quality with the
most common standards; being fecal coliform limits of
200/100 ml to 1,000/tOO ml. Over 45 states have
multi-level standards for disinfection relative to the
discharge stream water quality criteria, with 200
fecal coliforms per 100 ml as the most common
standard. At least 15 states have both fecal and total
coliform disinfection criteria. At least 15 states also
require effluent disinfection levels of 14 fecal coli-
forms per 100 ml for discharge into shellfish waters,
and 9 states have more stringent disinfection stan-
dards than those imposed on shellfish water dis-
charges.
As Indicated by Cabelli (2), the development of water
quality guidelines, standards, and associated health
effects for recreational waters, such as disinfection
requirements, has historically followed a pattern
characteristic of many such efforts to control pollu-
tion; the establishment of health and ecological
effects. The first step has been the development of
guidelines and standards dictated largely by attain-
ment with the Best Control Technology Available
(BAT). The guidelines have been based upon limited
epidemiological and ecological evidence in many
cases, and little, if any, data quantifying the risk in
relation to the level of the pollutant in the environ-
ment. The second step has been the modification of
the guidelines and standards on the basis of detect-
able risk with a limited quantity of data. The last step
is then the development of guidelines based upon
acceptable risk, which requires an epidemiological or
ecological data base broad enough to mathematically
model the relationship of some measure of water
quality to the risk. As more studies are conducted and
more data become available, the effluent disinfection
guidelines may change. This design manual has been
prepared, with the realization that different disinfec-
tion criteria are presently required and still others
may be required in the future. Therefore, the design of
the various disinfection alternatives has been
presented with this flexibility in mind, as well as the
considerations of different water qualities. For ex-
ample, this manual can be used to design disinfection
facilities for primary treatment through advanced
wastewater treatment. It can be used for effluents
from lagoons, activated sludge, trickling filters,
oxidation ditches, and rotating biological contactors,
as well as other unit processes.
Attainment of the disinfection guidelines can only be
achieved by the disinfection process, which, from a
disease prevention standpoint, is the most important
unit process in the wastewater treatment system.
Disease transmission via the aquatic route including
recreational water, drinking water; and seafood from
polluted water has been and continues to be a
problem. To the extent that the wastewater disinfec-
tion process mitigates that problem asserts for the
continued use of that practice in this country.
2.3 Projected Applications of Disinfection
Technologies
A summary of the 1 984 EPA wastewater treatment
plant needs survey has been prepared (12). A
breakdown of the number of municipal wastewater
treatment plants by flow capacity in 1 984 and those
projected for the year 2000 is presented in Table 2-5.
Table 2-5. Number of Wastewater Treatment Plants by Flow
Capacity (12)
Percent of Total
Number of Wastewater
Flow Capacity Facilities Flow
m3/d (mgd)
380
(0-0.1)
380 - 3,800
(0.1 - 1.0)
3,800 - 38,000
(1.0- 10.0)
>38,000
Total
1984
5,032
6,962
2,833
551
15,378
2000
8,416
8,313
3,255
687
20,671
1984
0.7
7.4
25.7
66.2
2000
0.9
6.9
24.6
67.6
The total number of facilities broken down by type of
treatment system in 1984 and projected for the year
2000 is shown in Table 2-6. An evaluation of the
number of facilities and quantity of flow treated by
municipal wastewater treatment facilities with pri-
mary treatment through advanced wastewater treat-
ment shows that approximately 41 percent of the
total wastewater flow presently receives secondary
treatment, with similar requirements in the year
2000. Approximately 39 percent of the total flow
presently treated receives greater than secondary
Table 2-6. Summary of Wastewater Treatment Processes
in the United States (12)
Number of Processes
Type of Process 1984 2000
Lagoons
Activated Sludge
Trickling Filter
Land Treatment
Oxidation Ditch
Rotating Biological Contactor
Total Design Flow, m3/d (mgd)
7,500
5,690
2,463
926
741
347
1.3 X 108
(35.9 x 109)
12,210
8,275
2,570
1,454
.1,215
291
1.6x 10s
(43.2 x 109)
8
-------
treatment, with around 51 percent expected to require
this level of treatment in the year 2000.
In 1984 at least 80 percent of the total wastewater
flow treated required secondary or more stringent
levels of treatment. This value is expected to increase
to over 90 percent of the total flow treated in the year
2000. Most of this increase will be in requirements
for advanced secondary treatment with the number of
facilities increasing from around 3,000 in 1982 to
around 6,400 facilities in the year 2000. The per-
centage of flow treated that requires secondary
treatment will remain about the same; however, an
estimated increase of around 5,000 newfacilities has
been projected. These more stringent requirements
for treated effluent quality are related to concerns for
protection of public health and the aquatic environ-
ment. These concerns will most certainly place
greater emphasis on disinfection requirements and
alternatives to chlorine disinfection. Advanced
wastewater treatment facilities are excellent candi-
dates for alternative disinfection with ozone and
ultraviolet irradiation. Due to the health hazards and
environmental concerns with chlorination, there
appears to be an excellent market potential for
alternative disinfectants, especially ozone and ultra-
violet irradiation, over the next 15 years.
Many of the existing treatment facilities are required
to disinfect their effluents, although the degree of
bacterial reduction varies from plant to plant. How-
ever, there are many facilities that presently do not
disinfect their effluents prior to discharge. Due to the
present concerns and future requirements for more
stringent effluent qualities, more stringent disinfec-
tion criteria relative to both microorganism control,
disinfectant residual, and biotoxicity or bioassay
testing requirements can be expected. Many of the
existing facilities may be required to upgrade their
present disinfection systems to include dechlorina-
tion or install alternative disinfection processes. New
facilities being built can be expected to require proper
disinfectant residual controls or alternative disinfec-
tion processes.
Over 400 facilities have been designed for chlorina-
tion/dechlorination, and approximately 50 percent of
these facilities are presently in use. At the time of
writing this manual, there were 24 wastewater
treatment plants in the United States confirmed to be
using ozone disinfection, 20 water treatment plants
using ozone either for disinfection or chemical oxida-
tion, and 30 drinking water plants in Canada using
ozone (R.G. Rice, Personal Communication, 1985).
An estimated 125 wastewater treatment facilities
were reported to be using ultraviolet light disinfection
in 1984. As of March, 1985, 33 municipalities with
ultraviolet light disinfection had been granted funds
by the EPA under the Innovative and Alternative (I/A)
Technology Program.
2.4 References
When an NTIS number is cited in a reference, that
reference is available from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
(703) 487-4650
1. Hubley, D., et al. Risk Assessment of Waste-
water Disinfection. EPA-600/2-85/037, NTIS
No. PB85-188845, U.S. Environmental Protec-
tion Agency, Cincinnati, OH, 1985.
2. Cabelli, V.J. Health Effects Quality Criteria for
Marine Recreational Waters. EPA-600/1 -80-
031, NTIS No. PB83-259994, U.S. Environmen-
tal Protection Agency, Cincinnati, OH, 1980.-
3. Berg, G. Indicators of Viruses in Water and Food.
, Ann Arbor Science, Ann Arbor, Michigan, 1978.
4. White, G.C. Disinfection: The Last Time of
Defense for Potable Water, JA WWA 67(8):410,
1975.
5. Akin, E.W. Infective Dose of Waterborne Path-
ogens. In: Municipal Wastewater Disinfection,
Proceedings of Second National Symposium.
EPA-600/9-83-009, NTIS No. PB83-263848,
Cincinnati, OH, 1983.
6. Sobsey, M.D. Enteric Viruses and Drinking
Water Supplies. JAWWA 67(8):414, 1975.
7. Health Risks Associated with Wastewater Treat-
ment and Disposal Systems, State of the Art
Review. EPA-600/1-79-016a, NTIS No. PB-
300852, U.S. Environmental Protection Agency,
Cincinnati, OH, 1979.
8. Okun, D. and G. Ponghis. Community Waste-
water Collection and Disposal. World Health
Organization, Geneva, 1975.
9. Craun, G.F., et al. Waterborne Disease Out-
breaks in the United States. JAWWA 68:420,
1976.
10. Maryland Water Quality Standards Policy, Oc-
tober 1978.
11. Crook, J. Wastewater Disinfection in California.
In: Wastewater Disinfection-The Pros and Cons.
Proceedings WPCF Pre-conference Workshop,
New Orleans, LA, Water Pollution Control
Federation, 1984.
12. 1984 Needs Survey: Report to Congress. EPA
430/9-84-011, U. S. Environmental Protection
Agency, Washington, DC, 1985.
-------
-------
Chapter 3
Disinfection Alternatives and Options
3.1 General Considerations
Since municipal effluents are an identifiable source
of pathogenic contamination and disinfection pro-
cesses themselves can create hazards to human
health and the aquatic environment, the decision to
disinfect or not disinfect is a very complicated one
that must be made on a site-specific basis. It is
therefore, impossible to establish universal policies
on wastewater disinfection requirements. Resolution
of the need for municipal wastewater disinfection at a
particular site involves the investigation of receiving
water uses and the associated risks to human health,
an assessment of the options that are available for
control of fecally-contaminated discharges, and an
evaluation of the environmental effects that control
measures may create. Figure 3-1 presents an ap-
proach for the type of rationalization that can be
involved in assessing the need for, and consequences
of, disinfecting municipal wastewaters (1).
In general. Figure 3-1 indicates that human health
risks are the initial concern, and upon establishing
the level of risk involved and the potential for reducing
or eliminating the risk, the environmental considera-
tions determine the applicability of the proposed
control measures. Development of an option, chlori-
nation or an alternative disinfectant, that satisfies
both the human health and environmental concerns
at a specific site is the next step.
3.2 Selecting a Disinfection Alternative
As many as 25 disinfection alternatives could be
considered and have been previously identified from
the literature without regard for physical or opera-
tional constraints (2). The major factors that must be
considered when evaluating disinfection alternatives
are summarized in Table 3-1 (3). The first four factors,
effectiveness, use cost, practicality, and pilot study
requirements relate to the disinfection process itself.
The fifth factor, potential adverse effects, relates to
effects of the disinfectant on the receiving water and
other environmental concerns and considerations.
Evaluation and thorough consideration of all the
criteria listed in Table 3-1 relative to practical,
physical, and operational constraints of municipal
wastewater disinfection reduces the available alter-
natives to chlorination, hypochlorination, chlorina-
tion/dechlorination with sulfur dioxide, chlorine
dioxide, bromine chloride, ozone, and ultraviolet light.
To properly evaluate and select alternative disinfec-
tion systems two levels of review are required. In the
first level of review, a number of non-monetary
factors are considered. This qualitative assessment is
comprised of three primary components, including
the previously described technical factors, environ-
mental impacts, and safety. In order to assess the
disinfection alternatives with respect to their non-
monetary factors, a qualitative matrix approach, as
shown on Table 3-2, can be used. A relative ranking of
the alternatives based on this qualitative assessment
can also be made, as shown in Table 3-3. The ranking
scale is based on a scale of one to five, with one
indicating the least impact or best degree of con-
fidence. From these types of analyses, the number of
appropriate alternatives can be narrowed, and some
alternatives may be completely eliminated.
The remaining acceptable alternatives can then be
evaluated in a second more detailed level of review. In
this second level of review, a preliminary design can .
be developed, cost estimates performed, and an
economic analysis comparing the alternatives on an
equitable basis can then be evaluated. Detailed
capital and operation and maintenance costs can be
developed for each alternative disinfection system.
Capital costs include structures, process equipment,
major auxiliary equipment, special foundation re-
quirements, electrical and instrumentation, site
work, miscellaneous process and piping, construc-
tion contingencies, engineering, project administra-
tion, and interest during the estimated period of
construction. The operation and maintenance costs
are annual costs and include labor, electrical power,
chemicals, routine equipment maintenance, and
materials and supplies. The specific details required
for performing the second level of review are ad-
dressed in the respective design chapters for each of
the disinfection alternatives.
The predominant advantages and disadvantages of
these disinfection alternatives are well known and
commonly cited in the literature. Some of the more
obscure elements have not been emphasized or have
been considered secondary or insignificant. A brief
-------
Figure 3-1. Framework for evaluating site-specific wastewater disinfection requirements.
Review Initiation
Is receiving water used as a
drinking supply?
(public, private)
No
Yes
r*
Is receiver used for primary
contac recreation. Shellfish
propagation/Harvesting.
Agriculture. Industry?
-*•
Mo
Yes
I
Does the discharge impact on
the quality of intake water?
(Dilution. Distance. Background
quality. Other sources)
No
Yes
Is there any other reason
to disinfect?
Yes
No
Does the discharge impact on
the quality of water at the
point of use (potential use)?
(Dilution. Distance. Background.
Other) I
No I Yes ..
Evaluate the feasibility of
distinfecting the discharge
on a seasonal basis
-Water use
-Target levels
-Effluent quality
1
Evaluate the feasibility of
disinfecting the discharge with
chlorine/chlorine compounds
-Target levels
-Effluent quality
-Organics
-Water treatment
1
Does the use of chlorine for
wastewater disinfection pose
a hazard to human health?
Yes
Is there a potential for significant
chlorine induced toxicity to aquatic
life? (TRC limits. Chlorinated
compounds)
No
Yes
Eliminate Chlorine
Chlorine
disinfection
acceptable
Evaluate alternate disinfection
technologies: Dechlorination
techniques. Alternate methods of
disposal (Tables 3-2 and 3-3)
Select method
of protection
(Prepare
documentation
™p»—»r—.for regulatory «
agency/public
hearings etc.
J
Conduct
monitoring and
surveillance
(effluent/
receiver)
Annual reviews
and cost
summaries
review of the pertinent factors associated with each
of the alternatives is presented in the following
sections of this chapter.
From Tables 3-2 and 3-3, it may be possible to select
an alternative disinfection system, as demonstrated
in the following situations. Examples follow where
ozone, ultraviolet light, and chlorination/dechlorina-
tion alternatives were selected for different specific
applications.
The cost of installing and operating an ozone dis-
infection process is dependent on many variables.
Major factors include size, flexibility, local construc-
tion costs, energy costs, energy efficiency and power
consumption, ozone dose requirements, and site
specific constraints. The cost for producing ozone is
normally higher than the alternative disinfection
methods; however, the other advantages associated
with ozone disinfection (Tables 3-2 and 3-3), may
outweigh cost considerations in some cases. The
following example is cited where ozone would be
selected as the disinfectant of choice based on site
specific constraints; however, this situation may be
unique since ozone would be cost competitive.
12
-------
Table 3-1. Major Factors in Evaluating Disinfectant
Alternatives
Effectiveness - Ability to achieve target levels of
selected indicator organisms
- Broad spectrum disinfecting ability
- Reliability
Use-Cost - Capital cost
- Amortization cost
- Operating and maintenance cost
- Cost of special wastewater pre-
treatment
Practicality - Ease of transport and storage, or
ease of on-site generation
- Ease of application and control
- Flexibility
-Complexity
- Ability to predict results
- Safety considerations
Pilot Studies Required
• Dose requirements
• Refine design details
Potential Adverse Effects - Toxicity to aquatic life
- Formation and transmission of un-
dersirable bio-accumulating sub-
stances
- Formation and transmission of
toxic, mutagenic, or carcinogenic
substances
A large tourist community has selected the oxygen
activated sludge process as the secondary treatment
process of choice. The treatment plant is to be located
next to the municipal utility district power plant with
electrical costs of $0.025/kWh. The following design
conditions were established due to the stringent
effluent discharge criteria:
Average Flow = 378,000 mVd (100 mgd)
Peak Daily Flow = 568,000 mVd (150 mgd)
Effluent BOD = 5 mg/l
Effluent TSS = 5 mg/l
Effluent NH3-N = 1 mg/l
30-Day Geometric Mean Fecal Conforms =
200/100 ml
Maximum Daily Fecal Coliforms = 400/100 ml
Maximum Chlorine Residual = 0.02 mg/l
Dissolved Oxygen = 6.0 mg/l
Due to the recreational nature of the community and
the use of the downstream water, major considera-
tions included effective virus destruction, Giardia cyst
destruction, and color removal. Maintaining the
pristine image of the service area and protection of
the quality of the downstream water supplies were
the primary reasons for the choice of ozone disinfec-
tion. The total cost for ozone disinfection represented
a small part of the total plant costs, and in this unique
situation, at the design ozone transferred dose of 5.0
mg/l, ozone disinfection was cost competitive with
chlorination/dechlorination and ultraviolet disinfec-
tion. The safety concerns for handling and transport-
ing chlorine through the recreational community
were also resolved with the selection of an ozone
disinfection process.
In a second community, ultraviolet disinfection was
chosen over chlorination/dechlorination and ozona-
tion when the following factors were given high
priority:
• effectiveness in destroying bacterial and viral
pathogens.
• safety precautions for transport, storage, and
application.
• environmental impacts, especially as relates to
halogenated organics formation and chlorine
residual toxicity.
• equipment reliability and ease of maintenance.
• process reliability and simplicity.
• process control flexibility.
• overall construction, operation, and maintenance
costs.
The treatment process consists of activated sludge,
final clarification, chemical addition for phosphorus
removal, and filtration prior to disinfection. The
following design conditions applied to this specific
case:
Average Flow = 7,600 mVd (2 mgd)
Peak Daily Flow = 13,200 mVd (3.5 mgd)
Effluent BOD = 10 mg/l
Effluent TSS = 10 mg/l
Effluent NH3-N = 1.0 mg/l
EffluentPO4-P=1.0mg/l
30-Day Geometric Mean Fecal Coliform = 200/100
ml
Maximum Daily Fecal Coliforms = 400/100 ml
Maximum Chlorine Residual = 0.1 mg/l
Ultraviolet light was considered to be the most
desirable of the three disinfection,alternatives con-
sidered with respect to non-monetary factors. Ultra-
violet light systems were considered to be simple to
operate and maintain relative to ozone systems, yet
provide the same advantages for overall disinfection
capacity relative to bacteria, spores, and virus inacti-
vation. Ultraviolet light was the safest alternative
since it is a physical disinfection process, and its use
eliminated the handling, transportation, and storing
of toxic, hazardous, and corrosive chemicals. Due to
the size of the plant and assurance of good effluent
quality especially with respect to the suspended
solids levels, ultraviolet light was chosen as the
disinfectant of choice over both ozone and chlorina-
tion/dechlorination. Ultraviolet light disinfection was
cheaper than ozonation and cost competitive with
chlorination/dechlorination in this particular applica-
tion.
The third specific case cited here involved a com-
munity in the northern part of the United States
13
-------
Consideration CI2 CI/deCI2
Size of plant all sizes all sizes
Applicable level of all levels all levels
treatment prior to
disinfection ^
Equipment good fair to
Reliability good
Process Control well fairly well
developed developed
Relative Complexity of simple to moderate
Technology moderate
Safety Concerns yes yes
Transportation substantial substantial
on site
Bactericidal good good
Virucidal poor poor
Fish Toxicity toxic non-toxic
Hazardous By-products yes yes
Persistent Residual long none
Contact Time long long
Contributes Dissolved no no
Oxygen
Reacts with Ammonia yes yes
Color Removal moderate moderate
Increased Dissolved yes yes
Solids
pH Dependent yes yes
O&M Sensitive minimal moderate
Corrosive yes yes
Table 3-3. Technical Factors and Feasibility Consider-
ations
Considerations CI2 CI/deCI2 BrCI CIO2 O3 UV
Flexibility 2 2 2222
Reliability 1 2 3232
Complexity 2 2 2332
Effectiveness 2 2 1112
Pilot Studies 1 1 4433
Rating based on scale of 1 to 5, with 1 indicating best degree of
confidence.
where seasonal disinfection was required during the
early spring and summer months. In this particular
application there was a concern for chlorine residuals
BrCI CIO2 O3 UV
all sizes small to medium to small to
medium large medium
secondary secondary secondary secondary
? ? fair to fair to
good good
problematic no developing developing
experience
moderate moderate complex simple to
moderate
yes yes no no
substantial substantial moderate minimal
good good good good
fair to good good good good
slight to toxic none non-toxic
moderate expected
yes yes none no
expected
short moderate none none
moderate moderate moderate short
to long
no no yes no
yes no yes (high no
pH only)
? yes yes no
yes yes no no
yes no slight no
(highpH)
moderate ? high moderate
yes yes yes no
and associated fish toxicity; however, the discharge
stream was not used as a water supply downstream.
The following design conditions apply to this plant:
Average Flow = 3,800 mVd (10 mgd)
Peak Daily Flow = 5,700 mVd (1 5 mgd)
Cft li i^n+ Qfin — QA rv\n /I
bttluent bULJ - oU mg/i
EffluentTSS = 30mg/l
30-Day Geometric Mean Fecal Col if orms= 200/1 00
ml
Maximum Daily Fecal Coliforms = 400/1 00 ml
Chlorine Residual = 0.1 mg/l
14
-------
Due to the seasonal disinfection requirements, capital
costs, and operation and maintenance costs, chlori-
nation/dechlorination was chosen as the design
alternative. The non-monetary advantages typically
associated with ozone and ultraviolet light over
chlorination/dechlorination did not outweigh the
advantage of chlorine disinfection in this particular
application.
As indicated in the previous examples, site specific
constraints and criteria will set guidelines for selec-
tion of appropriate disinfection alternatives. Important
considerations include factors such as the size of the
plant and effluent discharge quality requirements.
The more stringent the effluent requirements, the
more feasible alternative disinfectants such as ozone
and ultraviolet light become. Capital and operation
costs are obviously important factors; however, the
cost of any of the disinfection options is a very small
part of the total system costs.
3.3 Chlorination
Chlorine has not been required by EPA for inclusion in
States' nonconventional pollutant standards, but
approximately 15 states have taken steps to develop
specific criteria for determining the impact and
adverse effects of chlorine on aquatic life. Most states
have not adopted site-specific criteria to determine
the need(s) and adverse effects of disinfection using
chlorine. As a result, there has been limited success
in ensuring that public health and aquatic wildlife are
adequately protected.
Current developments consist of revised water quality
standards regulation and guidance documents to
ascertain the appropriateness of existing water
quality standards and assess use attainability. EPA's
Office of Research and Development has completed a
draft criteria .document for chlorine with directions on
how to apply the criteria and the chemistry and fate of
chlorine in natural receiving bodies of water It is
anticipated that guidance for establishing chlorine
effluent limitations for NPDES permits will be pub-
lished as Well as other documents to help encourage
states to consider public and wildlife health issues.
The Advanced Technology (AT) review policy requires
the evaluation of chlorine toxicity and the construc-
tion and operation of dechlorination facilities where
chlorine will exceed EPA criteria.
However, due to the toxicity of chlorine residuals at
extremely low concentrations (11 to 19 yug/l) and the
relatively high limit of detection of chlorine residual
test procedures (50 to 100 /ug/l), it is difficult to
control chlorine-induced toxicity in the receiving
stream. Therefore, use of alternative disinfection
processes should be considered where aquatic
toxicity is the overriding concern.
Today, Chlorination is the most used disinfectant at
water and wastewater treatment plants in the United
States. Chlorine reacts very rapidly when mixed with
water, and both hydrolysis and ionization occur.
Environmental factors such as temperature, pH,
alkalinity, suspended solids, chemical oxygen demand
(COD), and nitrogen containing compounds influence
the effectiveness of chlorine disinfection. Chlorine
reacts rapidly with ammonia and certain organic
compounds to form chloramines and chlorinated
organic compounds. The combined chloramines are
lower in germicidal value compared to the free
chlorine residuals, with organic chloramines in
wastewater offering significantly lower germicidal
value than inorganic chloramines.
The use of chlorine disinfection of wastewater can
result in several adverse environmental impacts,
especially due to toxic levels of total residual chlorine
in the receiving water and formation of potentially
toxic halogenated organic compounds. Chlorine
residuals have been found to be acutely toxic to some
species of fish at very low levels. The chlorine
residuals are stable and can persist for many hours at
toxic levels. Other toxic or carcinogenic chlorinated
compounds can bioaccumulate in aquatic life and
contaminate public drinking water supplies.
Chlorine is normally handled in steel containers of 68
kg (150 Ib) cylinders up to 82 metric ton (90 ton)
railroad cars. Chlorine is an extremely volatile and
hazardous chemical, and proper safety precautions
must be exercised during all phases of chlorine
shipment, storage, and use.
Hubley et al. conducted an extensive analysis of the
transportation risks of chlorine (4). A summary of
some of their findings is presented. A compilation of
the Department of Transportation (DOT) accident
reports for chlorine transportation from 1971 to 1980
is summarized in Table 3-4. These reports have
included information on number of deaths or injuries
and amount of chlorine released. The data in Table
3-4 are summarized by railroad, truck, and barge
shipment. The major accident at Youngstown, Ohio in
February 1978, has been broken out separately.
Chlorine movements for 1972, obtained from the U.S.
Bureau of Census, in metric ton-kilometers by
transportation mode and by shipment weight are
presented in Table 3-5 (5).
A breakdown of the information in Table 3-5 for
chlorine shipments by transportation mode and
container is presented in Table 3-6 (4). The informa-
tion in Table 3-4 was averaged to obtain estimates of
deaths, injuries, and property damage per year.
The percentage breakdown by transportation mode
and container in Table 3-6 was then used to calculate
15
-------
Table 3-4. Compilation of Department of Transportation Accident Data (January 1971 to December 1980) (4)
Railroad
Railroad
Excl.
Youngstown
Truck:
Cylinders
to 114 kg
0.911 metric
ton containers
Tanker
trucks
Barge
Number of
Accident
Reports
72
71
14
4
2
2
Deaths
8
0
0
0
0
0
Injuries
247
87
60
15
71
3
Property
Damage
($)
1.1 x 106
22,500
8,000
23,500
15,000
0
Amount
Released
(kg)
1.4X105
99,300
574
245
23
*
•Information not available prior to 1976.
Table 3-5. Percent Distribution of Chlorine Shipped by
Transportation Mode and by Shipment Weight
(metric ton-kilometers) (4)
Percent
A. Transportation Mode
Rail
Truck (combines DOT motor carrier
and private truck data)
Water
84.9
15.0
0.3
TOOTZ
B. Shipment Weight
Under 454 kg
454 to 4,500 kg
4,500 to 13,600 kg
13,600 to 27,200 kg
27,200 to 40,900 kg
40,900 kg and over
0.1
5.0
6.3
6.1
2.5
80.2
Table 3-6. Breakdown of Chlorine Shipments by
Transportation Mode and Container (4)
Annual Average
Percentage metric ton-km
(for 1972) (for 1971 to 1980)
Rail
Truck
Cylinders to 114kg
0.91 metric ton cylinders
Tank Truck
Barge
84.9
0.1
5.0
9.9
0.3
TU02
1858
2.2
109
219
6.4
2194.6
a corresponding breakdown by ton-km, assuming a
yearly total of 2191 million ton-km.
The results of this analysis, shown in Table 3-6, were
then used to normalize the accident data per ton-km
as shown in Table 3-7. It can be observed in Table 3-7
that the accident rate for truck-transported cylinders
was consistently higher than the other categories
listed.
Chlorination systems are reliable and flexible, and the
equipment is not complex. It is relatively easy to apply
and control in wastewater treatment, and low use
cost is often a great advantage for chlorine. Even
whendechlorination is required, it is still normally the
lowest cost alternative in most cases.
Hypochlorination refers to the use of solid (calcium) or
liquid (sodium) hypochlorite compounds as the dis-
infecting agent. The active compounds are the same
as with chlorination, primarily monochloramine in
wastewater effluents. The mechanism of bacterial kill
is the same as with liquid or gaseous chlorine.
Adverse environmental impacts associated with the
use of chlorine are also applicable with sodium
hypochlorite. Concentrated solutions from 10 to 15
percent chlorine must be stored in rubber-lined steel
or fiberglass storage tanks. Sodium hypochlorite is a
hazardous and corrosive material, but it will not
volatilize to a toxic gas as liquid chlorine does. The
required equipment is relatively simple and easy to
operate and maintain. The primary advantage of
hypochlorination over chlorination is the increased
safety in transporting, storing, and handling of
chemicals; however, chemical costs per unit of free
chlorine are generally much higher.
Dechlorination, normally with sulfur dioxide, has
been used to reduce the environmental impacts and
concerns associated with chlorination or hypochlori-
nation. Other dechlorinating agents such as other
sulfite reducing compounds or activated carbon have
been used; however, costs normally make these
options prohibitive. The levels of total chlorine
residuals can normally be reduced to below levels
that are toxic to aquatic life. The potential for
formation of halogenated organics may be reduced;
however, it appears that many halogenated organics
are formed rapidly upon chlorine addition, and the
application of sulfur dioxide will probably not affect
these compounds.
Dechlorination facilities are similar in most respects
to chlorination systems. Sulfonators meter gaseous
sulfur dioxide and are similar to chlorinators. The
amount of sulfur dioxide applied is normally based on
16
-------
Table 3-7. Accident Rates Per Metric Ton-KM (4)
Deaths
Injuries
Property
Damage
($)
Chlorine
Released
(kg)
Railroad
Railroad
excluding Youngstown
4.3 x 10~10
0
1.4X10-8
4.7 x 1Q-9
6.0 x 10-B
1.2 x 1
-------
special pretreatment prior to disinfection; however,
bromine chloride use for wastewater disinfection is
relatively new and a good data base and track record
are presently not available. Bromine chloride disinfec-
tion has been proven effective in field-scale applica-
tions for wastewater, but problems with the liquid
feed equipment have occurred. Brominated organics
such as bromoform and mixtures of chlorinated and
brominated organics can be expected to be formed.
3.6 Ozone
Ozone is an unstable gas that is produced when
oxygen molecules are dissociated into atomic oxygen
and subsequently collide with another oxygen mole-
cule to produce ozone. The ozone molecule can co-
exist as a gas with other gases such as air or oxygen,
or it can dissolve in a liquid such as water. Ozone is an
extremely reactive oxidant and a very effective
bactericide and virucide. Over 40 full-scale facilities
have been constructed in the United States, and an
extensive amount of research and development has
been expended to develop ozone disinfection of
wastewaters.
Unlike some of the other chemical disinfecting agents
previously discussed, ozone can exert beneficial
impacts on the environment. Since ozone decompos-
es rapidly to oxygen after application, the dissolved
oxygen levels in the treated effluent can be elevated
significantly, often to saturation levels. In most, if not
all cases, the need for effluent reaeration to meet
required dissolved oxygen water quality standards
can be eliminated. Ozone residuals can be acutely
toxic to aquatic life; however, since ozone dissipates
rapidly/ residuals are normally not found in the
effluent by the time it reaches the receiving water.
Ozonation has been, shown in some instances to
produce toxic mutagenic and/or carcinogenic com-
pounds, but little is presently known about these
organic byproducts (6). Ozone is believed to present
fewer potential environmental and health hazards
than chlorine.
Due to the instability of ozone, it must be generated
on-site from air or oxygen carrier gas. The most
efficient method of producing ozone today is by the
electric discharge technique, which involves passing
the air or oxygen carrier gas across the gap of
narrowly spaced electrodes under a high voltage. Due
to this expensive method of producing ozone, it is
extremely important that the ozone is efficiently
transferred from the gas phase to the liquid phase.
The two most often used contacting devices are
bubble diffusers and turbine contactors. With the
bubble diffusers, deep contact tanks are required.
Ozone transfer eff iciencies of 85 percent and greater
can be obtained in most applications when the
contactor is designed properly. The contactors must
be covered to control the off-gas discharges. Since
any remaining ozone would be extremely irritating
and possibly toxic, the off-gases from the contactor
must be treated to destroy the remaining ozone.
Ozone destruction is normally accomplished by
thermal or thermal-catalytic means.
An ozonation system can be considered to be
relatively complex to operate and maintain compared
to chlorination. The process becomes still more
complex if pure oxygen is generated on-site for ozone
production. Ozonation system process control can be
accomplished by setting an applied dose responsive
to wastewater flow rate (flow proportional control), by
residual control, or by off-gas control strategies.
Ozone disinfection is relatively expensive, with the
cost of the ozone generation equipment being the
primary capital cost item, especially since the equip-
ment should be sized for the peak hourly flow rate as
with all disinfectant technologies. Operating costs
can also be very high depending on power costs, since
ozonation is a power intensive system. The important
criteria for design include maximum transfer effi-
ciency in the contactor to maximize ozone utilization
and minimize applied dose and power consumption
requirements along with efficient ozone generation
equipment design. Equipment design is presently
provided by the equipment manufacturers, and they
are improving and updating their equipment to
improve production efficiencies and reduce the
associated operating costs.
3.7 Ultraviolet Light
The effectiveness of ultraviolet light as a bactericide
and virucide has been well established. It is a physical
disinfecting agent compared to the previously dis-
cussed chemical agents. Radiation at a wavelength of
254 nm penetrates the cell wall and is absorbed by
the cellular nucleic acids. This can prevent replication
and cause death of the cell. Since ultraviolet light is
not a chemical agent, no toxic residuals are produced.
Although certain chemical compounds may be altered
by the radiation, the energy levels used for disinfec-
tion are too low for this to be a significant cause for
concern.
Major advantages of ultraviolet light are its simplicity,
lack of impact on the environment and aquatic life,
and minimal space requirements. There is a negli-
gible likelihood of producing harmful chemicals in the
wastewater. Required contact times are very short,
on the order of seconds rather than minutes. The
equipment is simple to operate and maintain, but
fouling of the quartz sleeves or Teflon tubes must be
dealt with on a regular basis. Fouling is normally
handled by mechanical, sonic, or chemical cleaning.
High suspended solids concentrations, color, turbid-
ity, and soluble organic matter in the water can react
with or absorb the ultraviolet radiation reducing the
disinfection performance. High levels of wastewater
18
-------
disinfection (e.g., 2.2 total coliforms per 100 ml) will
be difficult to achieve with ultraviolet disinfection.
Relationships between effluent quality, effectiveness,
and use costs have become better defined in recent
months. Total costs appear to be competitive with
chlorination. The major operating costs are power
consumption and annual replacement of the ultra-
violet lamps. The process has been proven to be
applicable with over 120 installations in the United
States in small to medium size treatment plants. This
process is considered to be an effective alternative to
chlorination. Increased popularity and lowered costs
have occurred due to improvements in modern lamp
and system designs, increased competition, and
improved reliability and simplicity of operation.
3.8 References
When an NTIS number is cited in a reference, that
reference is available from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
(703)487-4650
1. Municipal Wastewater Disinfection in Canada-
Need and Application. Department of National
Health and Welfare, Ottawa, Canada, 84-EHD-
82.
2. Krause, T.L., et al. Disinfection: Is Chlorination
Still the Best Bet. Presented at the 53rd Annual
Conference of the Water Pollution Control
Federation, Las Vegas, Nevada, September
1980.
3. Tonelli, F.A. and K.W. Ho. Evaluating Disinfec-
tion Alternatives. Presented at the Joint Pollu-
tion Control Association of Ontario and Ontario
Ministry of Environment Seminar Current Ap-
proaches in Wastewater Treatment, April 1978.
4. Hubly, D. et al. Risk Assessment of Wastewater
Disinfection. EPA/600-2-85-037, NTIS No.
PB85-188845, U.S. Environmental Protection
Agency, Cincinnati, OH, 1985.
5. Bureau of the Census, U.S. Government, Depart-
ment of Transportation, Volume 3 (Commodity
Transportation Survey), 1972.
6. Stover, E.L., et al. Chlorine Vs Ozone at
Marlborough, Massachusetts: Disinfection and
Mutagenic Activity Screening. Water Chlorina-
tion Environmental Impact and Health Effects,
Volume 4, Book 2, (Jolley, ed.) Ann Arbor
Science, Ann Arbor, Michigan, 1983.
19
-------
-------
Chapter 4
Kinetics and Hydraulic Considerations
4.1 Disinfection Kinetics
4.1.1 Natural Die-Off
The natural processes of dilution, physical removal,
and die-off or inactivation can reduce pathogen
levels. The impacts of dilution on the concentration of
organisms in a wastewater discharged into a stream,
river, or lake can be estimated. Several mathematical
models have been proposed for estimating bacterial
die-off, and these models are mostly based on first
order kinetics. For example, the first order model for
die-off in streams is (1):
N = N0e
<-kt)
(4-1)
where:
No = the initial concentration of microbes dis-
charged into the stream
N = the concentration of the microbes t time
units after discharge into the stream
k = rate constant
The first order model for die-off in standing water
bodies (i.e., lagoons, lakes, etc.) is as follows:
N = N0/(1+ktd) (4-2)
where:
N0= the concentration of microbes in the water
body's inflow
N = the concentration of microbes in the water
body's discharge
td = the hydraulic detention time in the water
body based on the water body's discharge
k = rate constant
The rate constant k can be determined from a die-
away study of typical lakes or streams in a planning
area. Rate constants may also be found in the
literature (2).
Johnson, et al. (3) found that coliform die-away or
removal rates can be as much as 16 times higher in
summer months when compared to winter months.
Results of a study on the Logan City, Utah lagoon
system indicated that the summer coliform decay rate
coefficient was 0.5 per day, and that the winter
coliform decay rate coefficient was 0.03 per day.
Based on these results and using Equation 4-1, a
hydraulic residence time of at least 23 days in the
summer months would be required to reduce an
influent coliform level of 107 organisms/100 ml to an
effluent coliform level of 102 organisms/100 ml.
Redesign of lagoon systems to meet stringent effluent
coliform disinfection criteria appears to be econom-
ically impractical, especially because of the low die-
away rates during winter months. The only remaining
alternative consists of adding disinfection processes
to the final lagoon effluents.
4.1.2 Die-Off In the Presence of a Disinfectant
In cases where natural die-off is not sufficient to
prevent the potential for humans to ingest pathogenic
organisms, disinfection should be required prior to
discharge. Disinfection is a time-dependent process.
The ultimate outcome of bacteria and virus destruc-
tion is the result of a series of physical, chemical, and
biochemical actions that can be approximated by
simple kinetic expressions. It must be pointed out that
although the kinetic descriptions are simple, applica-
tion of these kinetics cannot be used universally. Site
specific conditions at one site may create problems
with precision and accuracy in the use of an empirical
relationship that was found effective at a different
site.
The information needed for the design of a disinfec-
tion system includes knowledge of the rate of
inactivation of the target organism(s) by the disin-
fectant. In particular, the effect of disinfectant
concentration on the rate of the process will deter-
mine the most efficient combination of contact time
and disinfectant dose to use. The major precepts of
disinfection kinetics were first enunciated by Chick
(4), who recognized the close similarity of microbial
inactivation by chemical disinfectants to chemical
reactions.
Chick stated that "disinfection is a gradual process,
without any sudden effects, and if the disinfectant is
sufficiently dilute to admit a reasonable time being
taken for the process, the reaction velocity can be
studied by enumerating the surviving organisms at
successive intervals of time." Therefore, for a given
number of organisms and chemical disinfectants, the
21
-------
rate of disinfection (number of organisms inactivated/
volume-time) can be described by:
dN
dt
= kN
(4-3)
where:
dN
dt
= rate of change in organism popu-
lation
k = organism die-off rate constant
N = number of surviving organisms per
unit volume at any given time
Equation 4-3 expresses the rate of die-off of micro-
organisms as an empirical first order kinetic model
and is commonly referred to as Chick's Law. Chick's
.Law is presented graphically in Figure 4-1. Diver-
gences from exponential decay are commonly ob-
served, and it is recognized that many factors can
cause these deviations, such as changes in disin-
fectant concentration with time, differences in resist-
ance between individual organsims of various ages in
the same culture, existence of clumps of organisms,
occlusion of organisms by suspended solids, etc.
Watson analyzed data with varying concentrations of
disinfectant and demonstrated a definite logarithmic
relationship between concentration of disinfectant
and the mean reaction velocity (5). He proposed the
following equation to relate the rate constant of
inactivation to the disinfectant concentration:
k * k'C"
(4-4)
where:
C - disinfectant concentration
n = coefficient of dilution
k' = corrected die-off rate constant, presumably
independent of disinfectant concentration and
organism concentration
The process of disinfection is influenced by temper-
ature, and the Arrhenius equation can be used to
predict temperature effects when direct heat-kill is
not a significant factor:
kr =
B
(T-20)
(4-5)
where:
kf = rate constant at some temperature T (°C)
kao = rate constant at 20°C
B = empirical constant
Little is known about disinfection efficiency at
elevated temperatures, but for agents such as ozone,
a significant reduction would occur due to the lower
efficiency of ozone mass transfer and greater ozone
decay.
The observation has frequently been made that
inactivation of organisms in batch experiments, even
when disinfectant concentration is kept constant,
does not follow the exponential decay pattern pre-
dicted by Equation 4-3. Two common types of devia-
tion are noted, as shown in Figure 4-1. In addition to
the linear Chick's Law decay, the presence of
"shoulders" or time lags until the onset of disinfec-
tion, is often observed (Curve 2 in Figure 4-1); Also,
some organisms and disinfectants exhibit a "tailing"
wherein the rate of inactivation progressively de-
creases (Curve 3 in Figure 4-1). In some cases, a
combination of both of these behaviors can even be
observed.
Figure 4-1. Chick's Law and deviations.
-1
z
£-3
Curve2
-Presence of "Shoulders
Curve 3
-Presence of "Tailing Off"
0 1 2 345 6
Time
The presence of "shoulders" can be accounted for if it
is assumed that multiple targets within each organ-
ism must each be damaged by independent dis-
infectant molecules prior to kill (6). An alternative
explanation is that diffusion through microbial outer
layers, or binding of disinfectant to the cell is a
necessary precursor to inactivation (7). The "tailing"
survivor curve may be explained (in the absence of
22
-------
changes in disinfectant concentration with time) by
either inherent differences in sensitivities of organ-
isms present, by nonuniformity in a spatial sense
(e.g., some organisms protected by enmeshment in
solids), or by the induction (chemical or biological) of
resistance in survivors over the time course of
exposure to disinfectant (8). The tailing phenomenon
has been explained by Poduska and Hershey (9), who
assumed different subpopulations with differing
sensitivities, and by Horn (10), who developed a
flexible, but highly empirical kinetic formulation
based on modifying Equations 4-3 and 4-4 to the
following form:
dN
dt
= -kNtmC"
(4-6)
Depending upon the value of m, both "shoulders" and
"tailing" may be depicted by Equation 4-6.
4.1.3. Summary of Kinetic Considerations
In reality, the rate of bacterial kill generally increases
or decreases with time. This can be attributed to
several factors such as oxidation, complexation, poor
mixing, cell resistance, dispersion, clumping, and
others. Deviations from Chick's Law can generally be
described by modifying the first-order expressions. It
must be emphasized that all the reported disinfection
kinetic equations are empirical in nature. Unfor-
tunately, Chick's Law does not accurately predict
coliform numbers as a function of dose in real world,
continuous flow systems, and therefore, the kinetics
of disinfection, as with any process, must be deter-
mined experimentally. As a result, pilot studies are
required for effective design of any disinfection
system. But pilot studies are expensive and many
times not conducted. Therefore, the need exists for
compiling all the available data and information on
disinfection to address this problem.
4.2 Mixing and Contactor Hydraulics
In dealing with the hydraulic behavior of any con-
tinuous flow reactor, the ideal situation is one in
which the velocity profile is known at any point within
the reactor. This most often is not possible; the flow
characteristics are instead partially described by the
residence time distribution (RTD). If one considers the
time each element resides in the reactor, the fre-
quency distribution of these times when plotted
against the time forms the residence time distribution
curve. This is presented graphically in Figure 4-2. By
normalizing the distribution, the RTD can be repre-
sented in such a way that the area under the curve is
equal to unity:
/ Edt = 1
o
where E = the residence time distribution.
(4-7)
Figure 4-2.
Residence
. Time
Distribution
E
Presentation of the residence time distribution
(RTD).
RTD Curve (E-Curve)
Edt =
One should note that in these discussions, it is
assumed that the system is at steady-state, that there
are no reactions taking place (conservative tracer),
and that the density of the liquid is uniform through-
out the reactor.
Given an E curve, the system mean residence time
may be determined using:
6 = / tEdt
o
(4-8
It is noted that this definition is similar to that for a
moment of inertia. However, knowing the value for 6,
an infinite number of E curves may still exist, differing
in the spread of actual residence times. It can readily
be shown that 8 is identical to V/Q in systems such as
those considered here (11). The determination of the
system RTD provides information on the numerical
spread of actual residence times, otherwise referred
to as the dispersion.
Thus, the evaluation of a specific reactor relies on the
experimental determination of RTD for that reactor.
This can be accomplished by a number of procedures;
subsequent analysis of the RTD curves may be used
to characterize mixing behavior of the unit.
4.2.1 Experimental Determination of the
Residence Time Distribution Curve
The means to develop the RTD of a reactor is generally
referred to as a stimulus-response technique. The
state of the reactor is perturbed in some fashion and
the resulting response is observed. In this case, the
stimulus is the input of a conservative (non-reactive)
tracer to the fluid entering the vessel and the
response is the time record of the tracer leaving the
vessel. The method by which the tracer is introduced
to the fluid will influence the type of response seen at
the outlet of the vessel. Typically, direct experimental
tracer analyses entail two types of inputs. The first is
the pulse input in which the tracer is injected in a
relatively short period of time to the influent. This is
ishown on Figure 4-3(a); the response is referred to as
a C-curve.
23
-------
Figure 4-3. Representation of pulse and step inputs and
resulting outputs.
Tracer
Concentration
C
Pulse Input
(a)
Output
(C-curve)
0 Time
Tracer Step input
(b)
(elevate concentration)
Output
(F-curve)
Time
_ Step input
/ (shut-off tracer)
Time
The second method is the step input. In this case, the
tracer is introduced as a constant input, i.e., the fluid
is adjusted from one steady-state concentration level
to a new steady-state concentration level. The
response is the time record of the concentration in the
effluent reaching this adjusted level, as shown on
Figure 4-3(b) and (c). This response is referred to as an
F-curve. The step input can be applied in either
direction: on Figure 4-3(b), it is shown as an input to
increase the concentration; the tracer can also be
applied as a steady input and then shut-off, with the
response as illustrated by Figure 4-3(c). The E-curve
(Figure 4-2) and the C-curve are equivalent:
E = C
(4-9)
The F-curve can be transformed to the C-curve by
taking its derivative:
dF/dt = C
(4-10)
The utility of this equation will be demonstrated in a
later discussion.
Experimentally, the development.of the RTD curve is
fairly straight-forward. A substance that is conserv-
ative in the carrying fluid is injected into the influent
of the reactor, either as a pulse or as a step input.
Suitable tracers are typically a salt (NaCI or LiCI) or a
fluorescent dye. These are then measured directly in
the effluent by rapidly taking aliquots for subsequent
analysis or by direct instream measurements (e.g.,
conductivity). For RTD determinations in chlorine or
ozone systems, batch sampling techniques suffice
because of the long systehn mean residence times.
The nature of typical UV system designs presents
logistical problems to the experimental procedures
for the tracer studies. These arise from the very short
average residence times typical of most UV systems,
and in some cases to the difficulty of injecting the
tracer at an appropriate point in the approach to the
reactor. The average residence times generally
encountered in these systems are between 1 and 50
seconds; at the shorter times it is difficult to collect
samples manually at frequent enough intervals to
construct a good time record of the effluent concen-
tration. High frequency automatic samplers and a
very rapid injection of the tracer at the upstream
portion are generally required. Johnson and Quails
(12,13) described such a system for a pulse injection
and sampling method. The method is generally
appropriate where there are distinct inlet and outlet
pipes on the UV reactor (see Figure 4-4(a)). The tracer
is injected quickly into the inlet pipe: the time of
injection must be small with respect to the average
residence time of the reactor and there must be
adequate mixing of the tracer and wastewater prior to
entry into the reactor. Samples should be drawn
immediately after exiting the reactor; alternatively,
conductivity can be measured directly if salt is used as
the tracer.
A step injection method was demonstrated in the Port
Richmond study (14). The reactors in this evaluation
were open vessels, in effect simulating open channel
flow. Thus, there were no distinct inlet or outlet
constrictions in which to inject or measure the tracer.
Additionally, the times were very short, such that
direct sampling of the effluent would be difficult. The
technique that was developed (Figure 4-4(b)) incorpo-
rated the use of the F-curve; the step input was
generated by first injecting a steady-state stream of
tracer into the unitand then shutting off the injection.
A conductivity probe was used to search for the point
of maximum concentration on the exit side of the
lamp battery. Note that the probe can also be used to
scan the entire exit plane in order to define the
conductivity profile and location of the plume as it
exits the lamp battery.
Once the probe is situated and fixed at the center of
the plume, a steady-state condition is allowed to
develop at fixed wastewater and salt solution flow
rates. The high frequency output from the probe is
24
-------
Figure 4-4(a). Techniques for experimentally determining
the RTD curves of UV reactors, (closed vessel
analysis)
Pulse Injection
Tracer
Injection
Influent
Reactor
Conductivity Probe
Effluent
Sample Port
Rotating
Sampling Tray
C n-0
Figure 4-4(b). Techniques for experimentally determining
the RTD curves of UV reactors, (open vessel
tracer analysis)
Event Signal
Salt
Solution
-J-,
Metering
Pump
Wheatstone _
Bridge L
-
o
0
o
£
o
o
o
Lamp Battery
o o o o o oo
C(
1
-1 — -
Oscillograph
anductivity Probe
t = 0
Time After Injection
amplified and recorded by an oscillograph in this
case; it is necessary to have a high speed recorder in
order to obtain a readable and undistorted trace in the
short time frames. Once steady-state is indicated by
the recorder, the salt solution pump is shut-off. This
event should be automatically signalled and recorded.
The die-away of the salt tracer is then monitored by
the fixed probe and continuously recorded. Readings
(conductivity) can then be taken of the tracer,
converted to salt concentration (mg/l), and trans-
posed to a plot of concentration against time. For a
particular type of system this can also be repeated at
several points across the plane of the lamp battery.
The step input method is generally applicable to any
type of reactor system. It is particularly useful in
situations where there is a short average residence
time and where direct in-line monitoring of the tracer
is possible.
4.2.2 Analysis of Residence Time Distribution
Curves
There are a number of uses for the RTD as a tool for
design and as a diagnostic to determine the effect of
the hydraulics on the system's performance. The
shape of the curve and the distribution of the area
under the curve will describe much of the hydraulic
Salt
Cone.
C
Trace after shutoff
of step input
Time After Shutoff of Tracer
characteristics of a system and indicate if it conforms
with proper design for a disinfection process.
Consider the RTD curves on Figure 4-5, for three
different flow regimes. The responses shown are all
to a pulse input. The worst case for a disinfection
reactor is shown as a complete mixed flow reactor
(Figure 4-5(a)). A pulse input enters the reactor and is
completely mixed in the total volume of the reactor
instantaneously. This implies that a fraction of the
input will be immediately discharged without any real
time of contact with the disinfectant. This is a wholly
unacceptable condition for a disinfection process.
Recall that the inactivation of microorganisms re-
quires the time element; thus, in a completely mixed
reactor, a significant fraction of the microorganisms
in the wastewater will exit with little chance of being
inactivated.
The ideal case is the pure plug flow (Figure 4-5(b)).
The response to the pulse input is a spike of
essentially zero width. This implies that every element
of the pulse input resides in the reactor for an equal
amount of time. This ideal situation is not achieved in
actual applications; there will be some degree of
dispersion such that the width of the plug flow C-
curve will expand. The objective of the hydraulic
design aspect of a system will be to minimize the
degree of this spread in the RTD curve.
25
-------
Figure 4-5.
Examples of RTD curves for various flow
characteristics.
(Age Distribution)
Flow Characteristics
(a)
Complete Plug
Mixed Flow
Flow
(b)
b
Area = 1'
Artibrary
-------
4.2.2.1 Dispersion Models
The above discussion centered on indices that could
be readily obtained from the RTD to evaluate or
diagnose the flow regime of a specific reactor.
However, recalling that the primary design approach
incorporates the use of a disinfection model, it is
necessary to also evaluate the RTD as a means of
estimating the dispersive characteristics of the
system. To accomplish this, one can write a mass
balance over the reactor for a conservative substance
(at steady-state), where E is defined as the dispersion
coefficient (lengthVtime), u is the linear axial veloc-
ity, and x is the axial distance from the inlet.
B =
E (d2C) . dC
uX
(4-15)
dx*
dx
The dimensionless group, E/uX, is called the dis-
persion number (d), and is a measure of the axial
dispersion in a reactor:
as
JL -— 0; plug flow
uX
§_ —
uX
as __ — °°; complete mix
Levenspiel (11) gives the solution to Equation 4-1 5 for
pulse inputs in relatively low levels of dispersion. First
define the dimensionless parameter, B';
+
B' =
0
which describes time in units of the mean residence
time. In case of small values of d the solution of
Equation 4-15 gives the symmetrical C-curve:
1
exp
-d -
(4-17)
2(?rd)1/2 4(d)
As shown on Figure 4-6, E/uX(ord), is effectively the
Figure 4-6.
Relationship of C-curve to E/uX for small
degrees of dispersion.
68% Total Area
single parameter for the resulting C-curve. The mean
and variance of this curve, C, are:
ffm,c =
= 1
and
= = 2(
uX
= 2d
(4-18)
(4-19)
Thus, if the RTD closely approximates a normal
distribution, the parameter d = (E/uX) can be obtained
from the C-curve. It can be shown that minimal error
«5 percent) will be incurred by the above estimate
when:
d<0.01
In cases where a significant degree of dispersion
exists, (i.e., d > 0.01), causing the C-curve to skew,
Levenspiel gives the mean and variance for both
closed and open vessels. Note that "closed" vessels
are those in which fluid enters and exits by plug flow.
In open vessels, flow is undisturbed at the boundaries
of the vessel. For closed vessels the mean and
variance are:
0Vn,c = Tm-c = 1
6
= 2
L)2(1-e-uX'E)
uX uX
-2 (
(4-20)
(4-21)
um»c —
0.1
A distinct advantage of the use of the dispersion
number for the characterization of mixing behavior is
27
-------
that several predictive models are available. In the
chapter on chlorination, correlation for dispersion in
pipes and baffled contact chambers will be presented.
There are experimental measurements reported for
UV and ozone systems; however, these have yet to be
reduced to simple correlations.
The characterization of mixing in an ozone contact
basin is more* complex and is dependent on the type of
contactor used. For example, with turbine contactors
a high degree of mixing is a necessary part of the
process of dissolving ozone into the wastewater. At
the same time, the violent mixing helps bring the
organisms into contact with the ozone. On the other
hand, back-mixing ca;n increase d, and care must be
taken' to ensure that good baffling is provided to
prevent short-circuitmg.
With bubble diffuser ozone contact basins the mixing
is much less violent than with turbine contactors.
Even here, however, basin back-mixing in each stage
of the contactor due to the flow of gas in that stage is
sufficient to change the liquid flow characteristics
from a plug flow toward a complete mix pattern. The
tendency toward complete mix operation with in-
creasing gas flow for bubble diffuser ozone contact
basin was demonstrated by Venosa and Opatken (16}
by performing dye tests with and without gas flow
(Figure 4-7).
Figure 4-7. Dye test shows effects of gas flow rate on plug
flow characteristics through a bubble diffuser
ozone contact basin.
O Gas Flow = OL/min
D Gas Flow = 20 L/min
A Gas Flow = 80 L/min
4 68
Time, minutes
Each stage of a bubble diffuser contactor would
exhibit similar characteristics; therefore, a similar
potential for short-circuiting would exist. Also, each
stage of other high mixing type reactors, such as the
turbine mixer reactor, would exhibit characteristics of
a complete mix reactor. To minimize the effect of
short-circuiting, multiple stages that are positively
isolated from each other should be provided.
4.3 References
When an NTIS number is cited in a reference, that
reference is available from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
(703) 487-4650
1. Hubly, D. et al. Risk Assessment of Wastewater
Disinfection. EPA/600-2-85-037, NTIS No.
PB85-188845, U.S. Environmental Protection
Agency, Cincinnati, OH, 1985.
2. Berg, G. IndicatorsofViruses in Water and Food.
Ann Arbor Science, Ann Arbor, Michigan, 1978.
3. Johnson, B.A., et al. Waste Stabilization Lagoon
Microorganism Removal Efficiency and Effluent
Disinfection with Chlorine. EPA-600/2-79-018,
NTIS No. PB-300631, U.S. Environmental Pro-
tection Agency, Cincinnati, OH, 1979.
4. Chick, H. An Investigation of the Laws of
Disinfection, J. Hyg., 8, 92, 1908.
5. Watson, H.E. A Note on the Variation of the Rate
of Disinfection with Change in the Concentra-
tion of the Disinfectant. J. Hyg. 8, 536, 1908.
6. Chang, S.L., Modern Concept of Disinfection. J.
Sanit. Eng. Div., ASCE (97):689-707, 1971.
7. Haas, C.N. Rational Approaches in the Analysis
of Chemical Disinfection Kinetics. Chemistry in
Water Reuse, Volume 1, p. 381-399 (ed. W.
Cooper), Ann Arbor Science, Ann Arbor, Michi-
gan, 1981.
8. Cerf, O. Tailing of Survival Curves of Bacterial
Spores. J. Appl. Bacteriol. 42(1), 1977.
9. Poduska, R.A. and D. Hershey. Model for Virus
Inactivation by Chlorination, JWPCF 44:738,
1972.
10. Horn, L.W. Kinetics of Chlorine Disinfection in
an Ecosystem. J. Sant. Eng. Div., ASCE 90
(SA1): 1983-1994, 1972.
11. Levenspiel, O. Chemical Reaction Engineering.
Second Edition, John Wiley and Sons, New
York, NY, 1972.
28
-------
12. Johnson, J.D. and R.G. Quails. Ultraviolet
Disinfection of Secondary Effluent: Measure-
ment of Dose and Effects of Filtration. EPA-
600/2-84-160, NTIS No. PB85-114023, U.S.
Environmental Protection Agency, Cincinnati,
OH, 1984.
13. Quails, R.G. and J.D. Johnson. Bioassay and
Dose Measurements in Ultraviolet Disinfection.
Applied and Environmental Microbiology
(45):872, 1982.
14. Scheible, O.K., et al. Ultraviolet Disinfection of
Wastewaters from Effluent and Combined
Sewer Overflows, Draft Report submitted to
U.S. Environmental Protection Agency, WERL,
under Cooperative Agreement No. CR807556,
1984.
15. Rebhun, M. and Y. Argaman. Evaluation of
Hydraulic Efficiency of Sedimentation Basins. J.
Sanit. Eng. Div., 91 (SA5):37, 1965.
16. Venosa, A. and E.J. Opatken. Ozone Disinfec-
tion-State of the Art. In: Proceedings Pre-Con-
ference Workshop on Wastewater Disinfection,
Atlanta, GA, Water Pollution Control Federation,
1979.
29
-------
-------
Chapter 5
Halogen Disinfection and Dechlorination
5.1 Coverage
This chapter covers the use of chlorine, chlorine
dioxide, bromine chloride and related compounds for
wastewater disinfection. In addition, design of sys-
tems for the dechlorination of wastewaters will be
discussed.
The material in this chapter may be used in several
ways. Sections 5.2 to 5.6 present an overview of the
history of chlorination and the other halogens, and
the fundamental chemical and kinetic aspects of
these materials as disinfectants. The chemical as-
pects of dechlorination are also reviewed.
Section 5.7 covers in detail the specification and
design of the various components comprising a
halogen disinfection system, and a dechlorination
system. The engineer may proceed directly to this
section, referring to material in Sections 5.2 to 5.6 as
background. For application to a particular case under
consideration, the engineer may also use the design
example presented at the end of Section 5.7 as a
checklist in the construction of other calculations.
Section 5.8 treats safety considerations in the use of
halogens, to which the design engineer must pay
particular attention. Section 5.9 and the following
material summarize important O&M considerations
that should be addressed in the design stage, and
relate performance experience at several full-scale
treatment plants.
5.2 History of Halogen Disinfection
5.2:1 Chlorine and Hypochlorites
While chlorides in the form of salts were known by
the ancients, the first preparation of chlorine gas,
then known as "dephlogisticated marine acid air" is
credited to Scheele who achieved this result from the
reaction of manganese dioxide with hydrochloric acid
in 1774. However, it was not until 1803 that this
material was regarded by Davy as a chemical element,
and general acceptance of Davy's hypothesis did not
occur until 1815 (1).
Medical germicidal applications of the new compound
soon followed, with hospital disinfection equipment
installed as early as 1823, and with the use of
chlorine in surgical applications by Semmelweis in
1826 (1). The first public health or environmental
application of chlorine appears to have been its use as
a prophylactic agent during the European cholera
epidemic of 1831 (1), although Baker{2) asserted that
Javelle water (chlorine gas dissolved in an alkaline
potassium solution) was used in France for waste
treatment as early as 1825.
The earliest U.S. reference to chlorine as a dis-
infectant occurs in 1832, when the known germicidal
properties of chlorine for control of disease epidemics
were summarized by Averill (3). It is particularly
noteworthy that these results were obtained approx-
imately 50 years before the advent of the germ theory
of disease.
The first formal recognition of chlorine or chlorine
compounds for wastewater treatment occurred in
1854, by the English Water Commission (2), although
not until 1884 was wastewater in England actually
chlorinated (4).
During the last years of the nineteenth century,
electrolytic generation of chlorine and hypochlorites
became sufficiently competitive with the chemical
oxidation synthesis routes to spur the use of chlorine
compounds for disinfection. The growth of chlorine
disinfection for both water and wastewater applica-
tions occurred simultaneously.
The three earliest applications of chlorine as a
wastewater disinfectant used patented on-site elec-
trolytic generation. In 1892, operations commenced
at both Hamburg, Germany and Brewster, New York
(4).
The electrolytic generation of chlorine for wastewater
treatment was covered by the patenting of the Woolf
Process in 1893 (5), which superseded the earlier
Powers process for the chemical production of
chlorine by a method similar to Scheele's original
procedure. Many variants on the electrolytic process
were reported up to and including the work of Rideal
in 1908.
Systematic investigations of the efficiency of chlorine
disinfection of wastewater were carried out in the
first decade of the twentieth century. Three research
groups reported on studies conducted almost con-
temporaneously.
31
-------
During 1907 and 1908, Phelps conducted laboratory
and field investigations of chlorination using chloride
of lime. Field chlorination at Boston, Baltimore and
Red Bank, NJ led to the conclusion that a dose of
several milligrams per liter chlorine and 15 minutes
contact time resulted in effective microbial reduc-
tions. Furthermore, Phelps asserted that it was
infeasible to attain complete^ microbial reduction
using chlorination. In his wo'rk, Phelps used the
coliform organism as an indicator of disinfection
efficiency. As a justification for this, he demonstrated
that these organisms had equivalent sensitivity to the
typhoid bacillus in wastewater chlorination (6).
Kellerman et al. (7). conducted two field studies on the
chlorination of (slow) sand filter effluents in Ohio and
came to conclusions similar to Phelps. In addition,
this latter work estimated costs for chlorination at two
particular plants studied as $7.50 and $16/million
gallons ($2.00 and $4.25/1,000 m3) using chlor-
inated lime. Futhermore, it was concluded that "the
quantity of chlorine immediately absorbed cannot be
estimated from the determination of the oxygen
consumed factor of the sewage effluent," thus
contradicting the hypothesis advanced by Rideal (8).
The third set of initial investigations was conducted
by Clark and Gage (9) at the Lawrence, MA, exper-
iment station. In addition to corroborating the studies
at Ohio and by Phelps, these workers appear to have
been the first to investigate the phenomenon of post-
disinfection regrowth. Their cost estimates for the
use of calcium hypochlorite for disinfection were
$ 1.75 to $3.75/million gallons ($0.45 to $ 1.00/1,000
m3}.
Following these early studies, a rapid introduction of
chloride of lime disinfection into wastewater treat-
ment plants occurred. While a commission of the City
of Milwaukee, Wl recommended, in 1909, that a
primary treatment/chlorination plant be built, the
first major city to install chlorination appears to have
been Philadelphia, PA in October 1910 (10). In
addition, Baltimore, MD commenced chlorination in
February of 1912 and Providence, Rl commenced
chlorination of a 20 MGD (0.9 mVs) stream that same
year (10,11). As of 1913, chlorination facilities using
chloride of I/me were installed at Red Bank, NJ;
Rahway, NJ; Shore Harbor, NJ; Ventnor, NJ; Atlantic
City, NJ; Bridgeton, NJ; Keyport, NJ; Margate City, NJ
(10); and at Pleasantville, NY (9) and College Park, CA
(12).
The rise in chlorine treatment of wastewater was
paralleled by the use of chloride of lime in water
disinfection. The first applications of this technology
occurred in 1908 at Boonton, NJ and Bubbly Creek
(Chicago), and in 1909 at Poughkeepsie, NY (13).
While Phelps, Kellerman, and the Lawrence, MA
group contributed to knowledge concerning the
application of chlorinated lime to wastewater dis-
infection, perhaps the first serious study of the
kinetics of wastewater disinfection, per se, was that
of Avery in 1913 (14). In contrast to prior work, and
particularly, that of Phelps, Avery concluded that it
was important to achieve some minimum amount of
contact time between the points of chlorine addition
and discharge. He suggested that this time be at least
20 minutes. Furthermore, Avery's studies were
significant in noting the relative difficulty of disinfect-
ing septic, as opposed to fresh, wastewater.
The use of solid hypochlorite compounds as dis-
infectants was superseded following the development
of practical and economic means of generating and
dispensing liquid chlorine. In 1909, production of
liquid chlorine commenced at Niagara Falls, NY, by
Electro Bleaching Gas Co. In 1912, Dr. Georg Ornstein
of this company developed and patented a device for
the application and feeding of chlorine in solution
form. Simultaneous to this, C.F. Wallace and M.F.
Tiernan developed and installed a direct feed gas
chlorinator at the Boonton, NJ water works, and went
on to develop solution feed chlorinators. The acquisi-
tion and abandonment of the Ornstein patent rights
by W&T in 1917 finalized the adoption of modern
solution feed liquid chlorination treatment for both
water and wastewater (13).
In wastewater treatment, the liquid chlorine direct
feed process was used as early as 1918 in Millville
,and Camp Merritt, NJ (14). The W&T solution feed
liquid chlorination process was used at New Haven,
CT commencing in 1919, where costs were estimated
at $4/million gallons ($1.00/1,000 m3) (15).
In light of the apparently rapid adoption of chlorination
for wastewater disinfection it is also of interest to
note that reservations were expressed then, some of
which are still relevant to practice today. For example,
as noted above, the issue of regrowth of bacteria
subsequent to chlorination and release was noted by
Clark and Gage at Lawrence (9). In addition, as early
as 1922 (16) reservations were expressed about the
relative resistance of coliforms to chlorine vis a vis
pathogenic bacteria, and the resulting adequacy of
the coliform test as an indicator of disinfection
efficiency.
The genesis of modern chlorination practice is seen in
the experiences at the Easterly and Westerly plants of
the city of Cleveland, OH (17). At these plants, with
average flows of 90 and 30 mgd (3.9 and 1.3 m3/s),
respectively, disinfection using liquid chlorine was
commenced in 1923, using heated water evapo-
rators. In 1929, at Easterly, one-ton container
32
-------
handling systems were installed, while at Westerly,
the 15-ton unit tank car system was placed on line. At
these two plants, over the period 1923 to 1929, costs
averaged $4.09 and $5.50/million gallons ($1.08
and $1.45/1000 m3), respectively. The final stage in
development of modern wastewater chlorination
concepts was the demonstration of the validity of
operational control based on measurement of the
chlorine residual by Tiedeman (18).
The geographic and temporal spread of wastewater
disinfection by chlorine is illustrated in Tables 5-1
and 5-2, which indicate the geographic distribution of
chlorination plants in 1910 and 1916, and the growth
in chlorination from 1910 to 1957.
5.2.2 Chlorine Dioxide
The discovery of chlorine dioxide, produced from the
reaction of potassium chlorate and hydrochloric acid,
is attributed to Davy in 1811 (21). However, it was not
until the industrial scale preparation of sodium
chlorite, from which chlorine dioxide may more
readily be generated, that its widespread use occurred
(22).
Chlorine dioxide has supplemented and supplanted
chlorine as a bleaching agent in pulp and paper
manufacture (22); however, despite early investiga-
tions on the use of chlorine dioxide as an oxidant and
disinfectant (23), its ascendancy in both water and
wastewater treatment has been slow.
As of 1977, 84 potable water treatment plants in the
United States used chlorine dioxide treatment, al-
though only one of these relied upon it as a primary
disinfectant (21). In Europe, chlorine dioxide is used
as either an oxidant or disinfectant in almost 500
potable water treatment plants (21).
While there have been numerous laboratory and pilot
plant investigations of chlorine dioxide disinfection of
wastewater, there does not appear to have been any
full-scale operating experience with this disinfectant
in wastewater treatment. The situation in Europe is
not known in detail; however. White cites one
Table 5-2. Development of Chlorination Installations for
Wastewater Treatment after Laubusch (5, 20)
Wastewater
Treatment
Plants
Year
1910
1916
1934
1940
1945
1948
1957
Total
619
846
3697
5580
5786
6058
7518 „
With
Chlorination
22
55
655
1127
1262
1307
2216
% of Population with
Wastewater Chlorination
Of Those
with Waste
Treatment
2.4
4.6
N/A
35.3
34.2
37.0
49.5
Of Total U.S.
Population
0.12
0.28
N/A
10.9
12.1
12.3
N/A
N/A not available...
unpublished study of chlorine dioxide at a full scale
wastewater treatment plant in France (24).
5.2.3 Bromine Chloride and Bromine
Due to the relative economics of bromine and
chlorine, little interest in the former compound as a
disinfectant was shown until the Second World War,
when shortages of chlorine and increased supplies of
bromine as a byproduct occurred (24). McCarthy
demonstrated that, as a potable water disinfectant,
elemental bromine was of similar effectiveness as
chlorine (25). In a more detailed study, Wyss and
Stockton demonstrated the relative insensitivity of
bromine disinfection to high ammonia nitrogen
concentrations (26). These results were confirmed by
Johannesson (27), who subsequently attributed this
finding to chemical properties of bromamines (28).
While the existence of bromine chloride has been
known since the work of Balard in 1826, final
equilibrium coefficients were not well defined until
the work of Mattrawetal.(29). However, little interest
in the industrial applications of bromine chloride
were evidenced until work by Mills and associates at
Dow Chemical (30-32).
In the mid 1970s, several investigations of the
effectiveness of bromine chloride in laboratory and
pilot plant disinfection of wastewaterswere reported.
Table 5-1. Early Geographic Distribution of Chlorination Facilities (19)
1910
1916
State
California
Connecticut
Maryland
New Jersey
New York
North Carolina
Ohio
Pennsylvania
Texas
No. of
Plants
1
0
2
6
1
0
1
4
6
% of Serviced
Population with
Chlorination
0.7
0
0.7
0.7
0.4
0
0.4
7.0
3.8
No. of
Plants
1
1 .
2
16
1
1
2
5
25
% of Serviced
Population with
Chlorination
0.7
5.9
0.4
38.8
0.2
8.2
1.7
9.9
15.0
33
-------
These included pilot or demonstration studies at
Hampton Roads, VA(33,34), Freedom District(Sykes-
ville). MD (35,36), Hatfield, PA (37), Hawaii (38-41),
and Michigan (42,43).
The number of full scale operating wastewater
disinfection systems using bromine chloride is un-
known, although a number of bromine chloride
installations for power plant condenser biofouling
control exist (24).
5.3 Chemistry and Physical
Characteristics of Disinfectants
The fundamental mechanisms of action of disinfect-
ants and the problems raised in the design of
wastewater disinfection systems can be related, in
many cases, to the physical properties of the dis-
infectants themselves, and their chemical reactions
with other constituents that may be present. In this
section, these properties will be reviewed.
5.3.1 Properties of Disinfectants
5.3.1.1 Chlorine and Hypochlorites
Elemental chlorine CI2 is a gas of density greater than
air at room temperature and pressure. When com-
pressed to pressures in excess of its vapor pressure,
chlorine condenses into a liquid with the release of
heat and with a reduction in specific volume of
approximately 450 fold. Hence, commercial ship-
ments of chlorine are made in pressurized tanks to
reduce shipment volume. When chlorine is to be
dispensed as a gas, it is necessary to supply thermal
energy to vaporize the compressed liquid chlorine.
Table 5-3 summarizes the major physical properties
of chlorine (44). Figure 5-1 summarizes the vapor
pressure of chlorine as a function of temperature.
These data are obtained from a publication of the
Chlorine Institute (45), from which more extensive
tabulations of the physical properties of gaseous and
liquid chlorine may be obtained.
Commercially, sodium hypochlorite and calcium
hypochlorite are also used as sources of chlorine
compounds for wastewater disinfection. The relative
amount of chlorine present in these alternative
sources of chlorine is expressed in terms of "available
chlorine."
The concentration of hypochlorite (or any other
oxidizing disinfectant) may be expressed as available
Table 5-3. Physical Properties of Chlorine (44)'
Liquid
Gas
Affinity for Water
Boiling Point (@ 1 atm)
Color
Corrosivity
Density, Ib/cu ft
Explosive limits (in air)
Flammability
Melting (freezing) Point (@ 1 atm)
Odor
Solubility
Chlorine hydrate
Specific gravity (compared to
4e water)
Relative Vapor Density
(Alr=1)
Vapor pressure
Viscosity
Slight
-34.05°C (-29.3°F)
Clear amber
Extremely corrosive to steel in pres-
ence of small amounts of moisture.
See discussion.
88.79 @ 60°F
(85.61 psia)
Non-explosive
Non-flammable
-100.98°C
(-149.76°F)
Penetrating and irritating
1.468 @32°F and 3.617 atm
See Figure 5-1
0.385 centipoise @ 0°C
0.729 centipoise @ -76.5°C
Slight
Greenish-yellow
Same as liquid
0.2003 @ 32°F and
1 atm
Non-explosive
Non-flammable
Non-flammable
Same as liquid
Below 9.6°C
(49.3°F)
CI2:8H2O, may
crystallize
2.482 @ 32°F and
1 atm.
167.9 micropoise
@ 100°
34
-------
Figure 5-1. Vapor pressure of liquid saturated chlorine gas. (45) (Reprinted by permission of The Chlorine Institute.)
10000
Triple Point
I I i I i I
I I I I
I I I I I I I I
I
I I I
-100 -80
-60
-40 -20
0 20 40
Temperature, °C
60
80
100
120 140 160
chlorine by determining the electrochemical equiv-
alent amount of CI2 to that compound. By Equation
5-1, it can be seen that one mole of elemental
chlorine is capable of reacting with two electrons to
form inert chloride:
CI2 + 2e" = 2 Cf
(5-1)
From Equation 5-2, it can also be noted that one mole
of hypochlorite (OCI~) may react with two electrons to
form chloride:
OCr + 2e~ + 2H* = CI" + H2O
(5-2)
Hence, one mole of hypochlorite is equivalent (elec-
trochemically) to one mole of elemental chlorine, and
may be said to contain 70.91 grams of available
chlorine (identical to the molecular weight of CI2).
Since calcium hypochlorite (Ca(OCI)2) and sodium
hypochlorite (NaOCI) contain two and one moles of
hypochlorite per mole of chemical, respectively, they
also contain 141.8 g and 70.91 g available chlorine
per mole. The molecular weights of Ca(OCI)2 and
NaOCI are, respectively 143 and 74.5, so that pure
preparations of the two compounds contain 99.2 and
95.8 weight percent available chlorine; hence they
are effective means of supplying chlorine for dis-
infection purposes.
Commercially, calcium hypochlorite is available under
a variety of trade names (HTH, Pittchlor, Perchloron)
35
-------
as a dry solid. It is relatively stable in the dry form,
subject to a loss in strength of approximately 0.013
percent/d (46).
Sodium hypochlorite is available in solution form
commercially in strengths of 1 to 16 percent by
weight. It is not practical to provide higher solutions
since chemical stability rapidly diminishes with
increasing strength. At ambient temperatures, the
half-life of sodium hypochlorite solutions varies
between 60 and 1700 days, respectively, for solutions
of 18 and 3 percent available chlorine (46,47).
When either chlorine gas or one of the hypochlorite
compounds is added to a water containing insign-
ificant quantities of Kjeldahl nitrogen, organic mate-
rial, and other chlorine demanding substances, a
rapid equilibrium is established among the various
chemical species in soliution. The term "free available
chlorine" is used to refer to the total of the con-
centrations of molecular chlorine (CI2), hypochlorous
acid (HOCI) and hypochlorite ion (OCI~), each ex-
pressed as "available chlorine" as described prev-
iously.
The dissolution of gaseous chlorine to form dissolved
molecular chlorine is expressible as a phase equi-
librium, and may be described by Henry's Law:
Hypochlorous acid is a weak acid and may dissociate
according to:
HOCI = OCI" + H~
KA = [OCr] [H+]/[HOCI]
(5-5)
The pKA of hypochlorous acid at room temperature is
approximately 7.6 (48).
The three equilibrium constants describing the free
chlorine system are each a function of system
temperature. Morris has provided a correlating
equation for KA as a function of temperature (48):
ln(KA) = 23.184 - 0.0583 T - 6908/T (5-6)
In Equation 5-6, T is specified in degrees Kelvin (°K =
°C + 273).
For engineering purposes the chlorine equilibria in
water may be adequately described using the Gibbs
free energy and enthalpy at 298°K with the assump-
tion of constancy of enthalpy over the practical range
of interest. For each of the important species. Table
5-4 indicates the relevant thermodynamicquantities.
CI2(0) =
: H (mo>l/L-atm) = [Clataqd/Pci, (5-3)
In Equation 5-3, quantities within square brackets
represent molar concentrations, PCI, is the gas phase
partial pressure of chlorine in atmospheres, and H is
the Henry's Law constant.
The dissolved aqueous chlorine is capable of reacting
with water to form hypochlorous acid, chloride ions,
and protons as indicated by:
Clataqi + HaO =
HOCI
[HOCI]
(5-4)
Using the standard thermodynamic relationships
from the data in Table 5-4, the equilibrium constants
for each of the chemical reactions governing free
available chlorine may be calculated. These constants
are summarized in Table 5-5.
Table 5-4.
Species
Thermodynamic Functions of Free Chlorine
Species-298°K (49)
Standard Gibbs Free
Energy of Formation
(G) (kJ/mol)
Standard Enthalpy of
Formation (H)
(kJ/mol)
CI2 (g)
CI2 (aq)
HOCI
ocr
ci-
0
+6.90
-79.9
-36.82
-131.30
0
-23.43
-120.92
-107.11
-166.94
The kinetics of this reaction have been discussed in
detail by Morris and have been found to be extremely
rapid (47).The reaction mechanism appears to involve
elementary reactions between dissolved molecular
chlorine and hydroxyl ions.
The extent of chlorine hydrolysis, or disproportiona-
tion, as described by Equation 5-4, is dependent upon
the pH and the salinity of the solution. The extent of
reaction decreases with decreasing pH and increas-
ing salinity; hence, the solubility of gaseous chlorine
may be increased by the addition of alkali or by the use
of fresh, rather than brackish water.
Table 5-5. Equilibrium Constants for Free Chlorine
(Applicability Approximately 283 - 308°K)
KA = 7.349 x 10~6 exp (-1660.89/T)
KH = 2.581 exp (-2581.93/T)
H = 4.805 x 1Q-6 exp (2818.48/T)
Note: T is in degrees Kelvin.
(mol/l)
(mol2/!2)
(mol/l - atm)
From the use of these equilibrium constants it is
possible to determine the relationship between gas
phase partial pressure of chlorine (Pci2) and solubility
under two particular conditions of importance. If
gaseous chlorine is dispersed into pure water with
36
-------
negligible amounts of demand, buffering potential,
and chloride ion, then the following equation can be
used to determine solubility:
Step 1—We may use the same values for KH and H as
the prior problem, since the temperature is the same.
A value for KA is needed, however.
. S = Pci2H + (KHPci2H)1/3
where S is solubility in mol/l as CI2.
(5-7) From Table 5-5:
To determine the solubility as free available chlorine
in mg/l multiply the latter quantity by 71,000.
However, if a solution of chlorine is prepared using
water with significant buffering potential and/or
chloride, or if a solution is prepared from calcium or
sodium hypochlorite, both of which may contain
chloride, then Equation 5-8 is appropriate. This would
also be applicable to determine the partial pressure of
chlorine above a solution containing the free chlorine
species.
(5-8)
S = Pci,H[1 + (KH/[H+][Cn) + (KAKH/[H+]2[Cr])]
The use of these equations is illustrated by way of two
examples.
Example 1:
A solution of chlorine is prepared by bubbling chlorine
gas at 0.5 atm pressure into pure water until
equilibrium is reached. The temperature is 54°F
(285°K). Determine the solution concentration of free
chlorine.
Step 1—Determine constants at the temperature of
concern (285°K)
From Table 5-5:
KH = 2.581 exp(-2581.93/285) = 3 x10~4mol2/!2
H = 4.805 x 10'6 exp (2818.48/285)
= 0.0948 mol /l-atm
Step 2—Apply Equation 5-7 to determine solubility:
S = (0.50 x 0.0948) + (0.50 x 3 x 10'4 x 0.0948)1/3
= 0.0716 mol/l
= 5,084 mg/l free available chlorine
Example 2:
An effluent is used to make up a chlorine stock
solution. The effluent is chlorinated until it contains
1,000 mg/l free available chlorine. At this point, it
has a pH of 6.0 and a chloride ion concentration of
500 g/m3 (this includes the initial chloride that was
present). What would be the equilibrium pressure of
chlorine in the gas phase? The temperature is 285° K.
KA = 7.349 x 10'6 exp (-1660.89/285)
= 2.16 xKT8 mol/l
Also, since the pH = 6, [H+] = 10"pH = 10"6 mol/l
Step 2—Apply Equation 5-8:
(1,000 mg/l free chlorine/71,000 = 0.0141 mol/l)
0.0141 = PCi,x 0.0948
x[1 + (3 x 10"4)/(10"6 x 0.0141) + (3 x 10~4)
x (2.16 x10-8)/[(1CT6)2x 0.0141]]
0.0141 = PCI, x 2061
PCIZ = 6.84 x 10"6atm
The second example illustrates an important aspect
to chlorine chemistry, namely that adjustment of pH
upwards (in this case to pH 6) results in a substantial
diminution of the equilibrium gas pressure. This is of
practical importance in the proper responses to spills
of hypochlorite, or solutions of chlorine. Adjustment
of pH, by the addition of an alkaline material such as
lime or sodium bicarbonate will reduce the volatility
of chlorine from such spills and thus minimize danger
to exposed personnel.
5.3.1.2 Chlorine Dioxide
Chlorine dioxide (CIO2> is a neutral compound of
chlorine in the +IV oxidation state. It has a boiling
point of 1.1 °C at atmospheric pressure. The liquid is
denser than water and the gas is denser than air (50).
Relevant physical properties of chlorine dioxide are
summarized in Table 5-6 (51).
Table 5-6. Physical Properties of Chlorine Dioxide (50)
Property
Value
Melting point
Boiling point
Density (0°C)
-59°C
n°c
1.640g/ml (liquid)
2.4 g/l (vapor)
Heat of Vaporization
Critical temperature
Vapor Pressure (0°C)
Heat of Solution (0°C)
Color
- Solid
- Liquid
- Gas
27.28 kJ/mol
153°C
0.626 or 0.673 atm
27.61 kJ/mol
Red
Orange
Orange
37
-------
Chemically, chlorine dioxide is considered to be a
stable free radical. Hence, at high concentrations it
may react violently with reducing agents. It is
explosive, with the lower explosive limit in the vapor
phase variously reported as 10 percent (49,51) or 39
percent (50). As a result, virtually all applications of
chlorine dioxide require the synthesis of the gaseous
compound in a dilute stream (either gaseous or liquid)
on location as needed.
The solubility of gaseous chlorine dioxide in water
may be described by Henry's Law, and a fit of the
available solubility data (52) results'in the following
relationship for the Henry's Law constant:
H(mol/l-atm) = [CI02
-------
Table 5-7. Comparison of Properties of BrCI and Br2 and
CI2 (57)
Property Br2 BrCI CI2
Freezing Point
Boiling Point
Density (gas,
20°C, 1 atm)
Vapor Pressure
atm, 20°C
Solubility
(g/100 g H20)
-7.3°C
-58.8°C
3.12 g/l
0.98
3.3
-66°C
5°C
2.34 g/l
3.03
8.5
-101°C
-34°C
1.40 g/l
7.81
0.75
and chlorine. In general, the properties of BrCI are
intermediate between the two halogens, with the
exception of the solubility of BrCI, which exceeds that
of either of the two molecular halogens.
Commercially, bromine chloride is supplied as a
containerized liquid under pressure. Since the vapor
pressure over liquid bromine chloride at ambient
temperatures is substantially in excess of atmos-
pheric (Figure 5-2), the chemical may be withdrawn
under its own pressure.
In both liquid and gas phases, BrCI may dissociate
into its constituent elements according to the follow-
ing:
2 BrCI(g, = Br2(a, + CI2(fl> (5-15)
2 BrClm = Br2 + CI2(I, (5-16)
The equilibrium constant for the gas phase reaction
(Equation 5-15) is given by:
Kg -
BrCI
(5-17)
This equilibrium has been well studied, and the value
for Kg appears to vary only slightly with temperature
and assumes the value of approximately 0.12 at 25°C
(29,30,58,59). No equilibrium constant for Equation
5-16 has been reported; however, it has been
asserted on the basis of unpublished results that liquid
bromine chloride is less than 20 percent dissociated
(30).
Due to the unavoidable presence of molecular
bromine and molecular chlorine in both the gas and
liquid phases of a bromine chloride system, the
thermodynamic significance of the vapor pressure,
boiling and melting point data for BrCI is ambiguous.
In a closed vessel containing BrCI in both liquid and
vapor phases, the vapor phase will be enriched in CI2
and the liquid phase will be enriched in Br2 under
normal conditions. To ensure withdrawal of bromine
chloride containing a 1:1 stoichiometric ratio of the
halogens it is necessary that all withdrawals be made
from dip tubes below the liquid level.
Aqueous solutions of bromine chloride in which the
concentrations of reducing agents are negligible
undergo a series of reactions similar to solutions of
chlorine gas. Bromine chloride may hydrolyzetoform
hypobromous acid (HOBr) and chloride according to
the following equation, which is analogous to Equa-
tion 5-4 for chlorine:
BrCI + H20 = HOBr + H+ + CI" (5-18)
KH = [H+][HOBr][Cr]/[BrCI]
The value of KH at 25°C has been reported as 2.94 x
I0~5 M2/!2, which is approximately fivefold less than
that for molecular chlorine (30,60).
Hypobromous acid (HOBr) is formed on hydrolysis in
preference to hypochlorous acid, since the latter is
able and will readily act to oxidize bromide ion to
HOBr. Hypobromous acid is a weak acid and will
dissociate in accordance with:
HOBr =• H+ + OBr"
KA = [H+][OBr~]/[HOBr]
(5-19)
Using thermodynamic data on hypobromous acid and
hypobromite (51), the following relationship for KA
may be derived for temperatures in the range of
normal operation:
KA = 5.24 x 10~6 exp (-2265/T) (5-20)
In general, the equilibrium constant for HOBr dis-
sociation is lower than that for HOCI at a given
temperature (i.e. the pK for HOBr is higher than that
for HOCI). Hence, the percentage of bromine present
as hypobromous acid at a given pH is greater than
would be the relative abundance of hypochlorous acid
in a system containing chlorine at identical concen-
trations.
Concentrations of bromine chloride can be expressed
in an analogous manner to those for free chlorine.
From Equation 5-21, in comparison with equation (1),
it may be seen that one mole of BrCI is equivalent to
one mole of chlorine:
BrCI + 2e~ = Br~ + CI"
(5-21)
Since one mole of BrCI contains 115.4 grams of
material, there are 0.62 (71 /115.4) grams of available
halogen as chlorine per gram of bromine chloride.
5.3.2 Ons/te Production Chemistry
In the case of chlorine and chlorine dioxide, tech-
nologies exist that enable a wastewater treatment
plant to produce disinfecting chemical at the plant
location itself, rather than relying upon shipment of
externally generated material. In the case of chlorine,
these technologies are alternative to external produc-
tion, while in the case of chlorine dioxide the onsite
39
-------
synthesis of disinfectant is mandatory due to de-
composition discussed above. The chemistry asso-
ciated with the production of these chemicals will
now be discussed.
5.3.2.1 Chlorine ancl Hypochlorites
'Chlorine and hypochlorites have been produced from
the electrolysis of brines and saline solutions-since
the early days of the 20th century (8). The basic
principle is the use of a direct current electrical field to
effect the oxidation of chloride ion with the simul-
taneousand physically separated reduction usually of
water to gaseous hydrogen.
Atthe anode of such an electrolytic cell, one or both of
the following reactions takes place:
CI" = 1/2 CI2(a, + e" E° = 1.36 volts (5-22)
(5-23)
Cl~ + 2OH" = OCI" + H2O + 2e~ E° = 0.90 volts
In Equations 5-22 and 5-23, E° is the standard
electromotive potential at unit activity of all products
and reactants at 298K, and the actual half-cell
voltage changes may depend slightly on concentra-
tion. This dependence may be determined by standard
relationships.
At the cathode of an electrolytic cell, reduction of
water occurs producing molecular hydrogen gas
according to Equation 5-24, written for alkaline
conditions.
(5-24)
H2O + e~ = OH" + 1/2 H2(a, E° = 0.83 volts
Combination of Equations 5-22 and 5-24 or 5-23 and
5-24 yield the two overall possibilities, depending
upon whether the chlorine is withdrawn as gaseous
chlorine evolving at the anode or as hypochlorite in
spent brine.
CI" + H20 = 1/2 CI2(a) + 1/2 H2(a, + OH'
E° == 2.19 volts
(5-25)
cr + H2o = ocr + H2o + 1/2 H2(a)
E° = 1.73 volts
(5-26)
From these thermodynamic relationships, it may be
concluded that for each mole of CI2 (or the equivalent
one mole of OCI") produced, two moles of electrons
are required. Since there are 96,493 coulombs (=
ampere seconds) per equivalent, theoretically 1.93 x
10 coulombs are required per mole of available
chlorine, or 2.72 C/mg available chlorine.
Additionally, from Equations 5-25 and 5-26, the
minimum voltage difference between the plates of an
40
electrolytic cell ranges from 1.73-2.19 volts. The
minimum theoretical energy required is the product
of this and the quantity (2.72 C/mg available chlo-
rine). Thus, 4.7-5.96 Watt-seconds are required to
produce 1 mg available chlorine. This may also be
expressed as 1.31 -1.66 kWh/kg available chlorine.
In actual practice, it is necessary to operate electro-
lytic chlorine generating units at voltages as high as
3.85 volts in order to increase the rates of the
generation reaction. At these overvoltages, however,
additional oxidations such as formation of chlorate,
ohmic heating, and incomplete separation of hydro-
gen from oxidized products with subsequent dissi-
pative reaction, combine to produce system inef-
ficiencies. For typical electrolytic generating units,
current efficiencies (based on actual to theoretical
C/mg) of 97 percent may be obtained along with
energy efficiencies (based on kw-hr/kg) of 58 percent
(49). These efficiencies are related to the physical
configuration of the electrolysis cells, brine concen-
tration, and desired degree of conversion to available
chlorine (61,62).
5.3.2.2 Chlorine Dioxide
Theoretically, chlorine dioxide may be produced by
either the oxidation of a lower valence compound or
reduction of a more oxidized compound of chlorine.
Chlorites (CI02~) or chlorous acid (HCI02) may be
oxidized by chlorine or persulfate to chlorine dioxide,
or may undergo autooxidation (disproportionation) to
chlorine dioxide in solutions acidified with either
mineral or organic acids. Chlorates (ClOal may be
reduced by use of chlorides, sulfuric acid, sulfur
dioxide, or oxalic acid, or electrocherriically to form
chlorine dioxide (51).
For practical purposes in wastewater treatment,
chlorine dioxide is generated exclusively from chlorite
inasmuch as the reductive processes using chlorate
as a starting material are capital intensive and
competitive only at larger capacities (51).
In the acid-chlorite process, sodium chlorite and
hydrochloric acid react according to:
(5-27)
5 NaCIO2 + 4 HCI = 4 CI02 + 5 NaCI + 2 H2O
The resulting chlorine dioxide may be evolved either
as a gas, or removed in solution. Mechanistically, this
process occurs by a series of coupled reactions, some
of which may involve the in situ formation of chlorine,
the catalysis by chloride, and the oxidation of chlorite
by chlorine (50,51,53). In addition, the yield of the
reaction as well as the rate of the process are
improved by low pH values in which both gaseous
chlorine and chlorous acid formation are favored.
Under these favorable conditions, the reaction pro-
-------
ceeds in the order of minutes; however, to achieve
these conditions, excess hydrochloric acid is required.
Alternatively, chlorine dioxide may be produced by
the oxidation of chlorite with chlorine gas according
to:
2 NaCI02 + CI2 = NaCI + 2 CIO2 (5-28)
As in the previous case, low pH accelerates the rate of
this process, as does excess amounts of chlorine gas.
However, if chlorine gas is used in stoichiometric
excess, the resultant product may contain a mixture
of unconsumed chlorine as well as chlorine dioxide.
5.3.3 Disinfection Demand Reactions
In the presence of such dissolved impurities as exist
in wastewater effluents, each of the halogen dis-
infectants may undergo reactions in which they
decompose or transform to less effective chemical
forms. In the case of chlorine and bromine chloride,
these principally involve reactions with ammonia and
amino-nitrogen compounds, while chlorine dioxide
demand may result from reactions with other organic
materials. The nature of these processes will now be
discussed.
5.3.3.1 Chlorine and Hypochlorites
Available chlorine, whether in the form of chlorine
gas or hypochlorite, when added to wastewater, can
undergo a series of dissipative reactions that result in
a loss of disinfectant from the system, or a change in
disinfectant form to a less active chemical species.
The following reactions of chlorine or hypochlorite
may occur (63):
• cyanide may react with chlorine to form chlorides,
bicarbonate, and a variety of oxidized nitrogen
forms;
• ammonia or amino-nitrogen groups may react with
chlorine to form chloramine compounds by the
replacement of a proton by a chlorine atom;
• organic molecules containing unsaturated (double
or triple) bonds may react to form chlorinated
organic molecules;
* a variety of inorganic reducing agents, including
hydrogen peroxide and reduced iron and sulfur
compounds, may consume chlorine by redox
reactions.
Of these processes, probably the most significant
process in determining the fate of chlorine added to
wastewater is the reaction with ammonia or amino-
nitrogen groups. However, the reactions with organic
materials are also of importance in the production of
chlorinated organic byproducts during wastewater
chlorination.
To understand chlorine demand reactions, a further
definition is necessary. As noted above, the term
"free available chlorine" is used to denote the
concentrations of hypochlorous acid pi us hypochlorite
ion expressed on a mass equivalent CI2 basis. The
term "total available chlorine" is defined as the mass
equivalent CI2 contained in all materials that contain
chlorine in an oxidized state. The difference between
total available chlorine and free available chlorine is
defined as "combined available chlorine" and repre-
sents the amount of chlorine that is in chemical
association with various compounds (usually amino-
or ammonia-nitrogen) but that is also capable of
carrying out oxidation. The significance of this
distinction, which will be further discussed in the
section on kinetics, is that the free chlorine forms are
generally more effective disinfectants than the com-
bined chlorine forms.
One example of this behavior, along with a graphical
demonstration of typical chlorine behavior in waste-
water, is given in Figure 5-3 (64). In this depiction, as
the dose of chlorine is increased, the total available
chlorine residual (i.e. remaining in the system after
30 minutes) increases until a dose of approximately
50 mg/l, whereupon residual chlorine decreases to a
very low value, and subsequently increases linearly
with dose indefinitely. The "hump and dip" behavior
is paralleled by the sensitivity of microorganisms to
the available chlorine residual (indicated by the time
required for 99 percent inactivation of Bacillus
met/ens spores). At the three points indicated, the
total available chlorine is approximately identical at
22 to 24 mg/l yet there is a 32-fold difference in
microbial sensitivity.
Figure 5-3.
Effect of increased chlorine dosage on residual
chlorine and germicidal efficiency: pH 7.0,
20°C NH310 mg/l. (64) (Reprinted by permis-
sion of the American Waterworks Association.)
120
Residual
Available Chlorine
20 40 60 80 100 120 140 160 180 200
Available Chlorine Added (ppm)
The explanation for this behavior is founded in the
"breakpoint" reaction between free chlorine and
41
-------
ammonia, which is present in non-nitrified effluents.
At doses below the "hump" in the chlorine residual
curve, only combined available chlorine is detectable
in the wastewater. At doses between the "hump" and
the "dip" in the curve, there is an oxidative destruc-
tion of combined residual chlorine accompanied by
the loss of ammonia nitrogen from the wastewater
(65). In fact, this particular reaction may be used as a
means to remove ammonia nitrogen from waste-
waters. Finally, after the ammonia nitrogen has been
completely oxidized, the residual remaining consists
almost exclusively of free chlorine. The minimum in
the chlorine residual vs dose curve (in this case 70
mg/l) is called the "breakpoint" and denotes the
amount of chlorine that must be added to a waste-
water before a stable free residual can be obtained.
The relative inefficiency of combined chlorine forms
as disinfectants vis a vis free chlorine forms has been
known at least since the work of Enslow (66).
Much of the knowledge of breakpoint reactions during
chlorination arises from studies on potable water
chlorination. In investigating chlorination of drinking
water. Griffin and Chamberlain (67,68) observed that:
• the classical "hump and dip" curve is only seen at
water pH's between 6.5 and 8.5;
• the molar ratio between ammonia nitrogen and
chlorine dose at the breakpoint under ideal condi-
tions is 1:2 corresponding to a mass dose ratio
(ammonia N:chlorine) of 1:10;
• in practice, mass dose ratios of 1:15 may be
needed to reach breakpoint.
In later work with breakpoint chlorination of waste-
waters, Griffin and Chamberlin found that measure-
ment of the 30 minute and 2 hour residuals failed to
produce a "hump and dip" curve, and that only at 18
hours could a classical breakpoint pattern be observed
(69). Furthermore, it was found that even above the
breakpoint, the concentrations of organic nitrogen
were unaffected. However, this work did confirm that
a mass dose ratio of 10 mg chlorine to 1 mg ammonia
nitrogen results in breakpoint at 18 hours in raw,
primary, and secondary (activated sludge) waste-
waters.
The breakpoint reaction may also affect the pH of a
wastewater. Figure 5-4 indicates that, if sodium
hypochlorite is used as the source of active chlorine,
as breakpoint occurs, the pH decreases due to an
apparent release of protons during the breakpoint
process (70). If gaseous chlorine is used, this effect is
obscured by the release of protons by hydrolysis of
gaseous chlorine according to Equations 5-4 and 5-5.
In wastewaters, the chlorine demand, representing
the difference between the applied chlorine dose and
Figure 5-4. Eff ect of chlorine or hypochlorite dose on pH of
settled wastewater.(70) Reproduced from
Journal of the Water Pollution Control Federa-
tion, by permission.
8 -
o.
Hypochlorite Solution
of pH = 11.1
Chlorine Water of pH = 2.0
10 15
Chlorine Dose, mg/l
20
25
the total chlorine residual at a given contact time, is
due to ammonia nitrogen, as well as organic and
inorganic compounds. In early work, Symons et al.
determined that domestic wastewater contains an
average chlorine demand of 45 g/capita-d, including
all sources (this is demand in the raw wastewater and
fails to account for possible changes in quality
through prior unit processes) (71,72). In addition,
superimposed upon this average were significant
fluctuations of various frequencies.
It was found that chlorine demand fluctuated sea-
sonally approximately with temperature. Weekly and
diurnal fluctuations also existed, and the latter were
found to be more variable than flow. During storms,
the chlorine demand was found to increase shortly
after the onset of the storm flow, and then to rapidly
return to baseline levels.
The rate of exertion of chlorine demand in both
potable water and wastewater chlorination has been
the subject of numerous studies. The most systematic
work has been that of Taras (73), who chlorinated
42
-------
pure solutions of various organic compounds and
found that chlorine demand exertion follows:
D = k(tf
(5-29)
where t is the time in hours, D is the chlorine demand,
and k and n are empirical constants. It was found that
n for various substances ranged from 0.02 to 0.29.
In subsequent work, Feben and Taras chlorinated
potable water, and potable water blended with up to
1.5 percent wastewater to a free residual, and found
that the results could be correlated to Equation 5-29
with the value of n determined by the one hour
chlorine demand as follows (74):
n = 0.18 - 0.017 log (D1) (5-30)
where D1 is the one hour chlorine demand.
Lin and Evans extended the significance of Taras'
work by chlorinating secondary effluents to combined
residuals (75). They found that Equation 5-29 could
be used to describe the kinetics of chlorine demand
exertion except that two sets of k and n values were
required. At times from 1 to 12 minutes, the rates of
chlorine demand development were significantly
greater than between 12 and 60 minutes. It was also
found that when calcium hypochlorite was used as a
chlorinating agent, a lower chlorine demand was
obtained as compared with chlorination using dis-
solved chlorine gas.
A more recent kinetic model for chlorine demand
exertion has been developed by Haas and Karra (76):
D = Co (1 - X exp (-kit) - (1 - X) exp (-kat)) (5-31)
In Equation 5-31, X is an empirical constant typically
0.4-0.6, and ki and k2are rate constants, typically 1.0
min"1 and 0.003 min"1, respectively, and C0 is the
chlorine dose in mg/l.
While it has been generally observed that nitrified
effluents that are low in ammonia-nitrogen have
lower chlorine demands in comparison to nonnitrif ied
effluents that have higher chlorine demands, recent
work suggests some anomalies. White et al. (77)
observed that the addition of low levels of ammonia to
nitrified effluents may reduce chlorine demand. This
phenomenon may be due to competitive formation of
inorganic versus organic chloramine compounds,
and will be discussed in greater detail below.
In the presence of ammonium ion, free chlorine may
react in a step-wise manner to form chloramines.
This process is depicted by the following equations:
NH3 + HOCI = NH2CI + H2O + H* (5-32)
NH2CI + HOCI = NHCI2 + H2O (5-33)
NHCI2 + HOCI = NCI3 + H20 (5-34)
These three compounds, monochloramine (NH2CI),
dichlorarnine (NHCI2), and trichloramine (NCU), each
contribute to the total (or combined) available chlorine
residual in a wastewater. Each chlorine atom asso-
ciated with a chloramine molecule is capable of
undergoing reduction to chloride, and in the process
accepting 2 electrons; hence, each mole of mono-
chloramine contains 71 grams available chlorine;
each mole of dichlorarnine contains 2 x 71 or 142
grams; and each mole of trichloramine contains 3 x
71 or 213 grams of available chlorine. Inasmuch as
the molecular weights of mono-, di-, and trichlor-
amine are 51.6, 86, and 110.5, this means that the
chloramines contain respectively 1.38, 1.65, and
2.02 grams available chlorine per gram. However, the
efficiency of the various combined chlorine forms as
disinfectants differs, and thus the concentration of
available chlorine also is insufficient to characterize
process performance.
As Equation 5-32 indicates, the formation of mono-
chloramine is accompanied by the loss of a proton,
which corresponds to the experimental findings of
McKeeetal.(70). The loss of a proton is due to the fact
that chlorination reduces the affinity of the nitrogen
moiety for protons; this was verified experimentally
for ammonium as well as a variety of amines by Weil
and Morris (78).
Under conditions where ammonia nitrogen is in
excess of the chlorine dose (i.e. below the "break-
point") and where the pH is below 9.0 (so that the
dissociation of ammonium ion is negligible), the
amount of combined chlorine in dichlorarnine relative
to monochloramine after Equations 5-32 and 5-33
have attained equilibrium may be given by the
relationship described by McKee et al. (70):
A = BZ/(1-(1-Z(2-Z)B)) -1
(5-35)
In Equation 5-35, A is the ratio of available chlorine in
the form of dichlorarnine to available chlorine in the
form of monochloramine, Z is the ratio of moles of
chlorine (as CI2) added per mole of ammonia nitrogen
present, and B is defined by:
B = 1 -4Keq[H+]
(5-36)
The equilibrium constant in Equation 5-36 refers to
the direct inter-conversion between dichlorarnine
and monochloramine as follows:
H+ + 2 NH2CI =
NHCI2
= [NH4+][NHCI2]/[H+][NHaCI]2
(5-37)
At 25°C, Keq has a value of 6.7 x 10s liters/mole
(70,79).
Using these relationships, it is possible to determine
the equilibrium ratio of dichlorarnine to monochlor-
43
-------
amine as a function of pH and applied chlorine dose
ratio (assuming no dissipative reactions other than
those involving the inorganic chloramines). Table 5-8
summarizes these calculations. As pH decreases and
the CI:N dose ratio increases, the relative amount of
dichloramine also increases.
Table 5-8. Ratio of Dichloramine Combined Chlorine to
Monochloramlne Combined Chlorine as a
Function of pH and Applied Molar Dose Ratio
(Equilibrium Assumed)
Molar
CI2:N
Ratio
.1
.3
.5
.7
.9
1.1
1.3
1.5
1.7
1.9
PH
6
.13
.389
.668
.992
1.392
1.924
2.7
4.006
6.875
20.485
7
.014
.053
.114
.213
.386
.694
1.254
2.343
4.972
18.278
8
1E-03
5E-03
.013
.029
.082
.323
.911
2.039
4.698
18.028
9
0
0
1E-03
3E-03
.011
.236
.862
2.004
4.669
18.002
As the CI:N molar dose ratio increases beyond unity,
the amount of dichloramine relative to monochlor-
amine rapidly increases as well. For the conversion
from dichloramine to trichloramine, the equilibrium
constant given at 0.5 M ionic strength and 25°C
indicatesthatthe amount of trichloramine to be found
in equilibrium with di- and mono-chloramine at molar
dose ratios of up to 2.0 is negligible (79). This is in line
with experimental measurement of the individual
combined chlorine species as a function of approach
to breakpoint, which indicates that detectable tri-
chloramine is not formed until shortly before, or after
the break point itself, and then only in relatively small
amounts (80).
These findings, coupled with the routine observation
of the breakpoint at molar doses below 2:1 (weight
ratios CI2:N below 10:1) indicate that trichloramine is
not an important species in the breakpoint reaction.
Rather, the breakpoint reaction leading to oxidation of
ammonia nitrogen and reduction of combined chlo-
rine is initiated with the formation of di-chloramine.
The kinetics of formation of chloramine species have
been investigated by various researchers. It has been
found that the formation of monochloramine is a first
order process in each of hypochlorous acid and
uncharged ammonia. Solely through kinetic argu-
ments, however, it is not possible to determine
whether this, or a process involving hypochlorite ions
reacting with ammonium cations, is the actual
mechanism of reaction. If the neutral species are
selected as the reactants, then the rate of formation
of monochloramine may be given by (81):
(5-38)
r(mol/l-s) = 6.6x108exp(-1510/T)[HOCI][NH3]
As noted above, hypochlorous acid dissociates into
hypochlorite with a pK of approximately 7.5; analo-
gously, ammonia is able to associate with a proton to
the ammonium cation, with the pK for the latter of
approximately 9.3. For a constant chlorine:nitrogen
dose ratio, the maximum rate of monochloramine
formation occurs at a pH where the product HOCI x
NHs is maximized, which is at the midpoint of the two
pK values or 8.3. At this optimum pH and the usual
temperatures encountered in practice, the formation
of monochloramine attains equilibrium in less than a
second; however, at either a higher or lower pH, the
speed of the reaction slows.
The formation of dichloramine from reactions be-
tween hypochlorous acid and monochloramine obeys
the following rate law (81 ):
(5-39)
r(mol/l-s) = 3.0 x 107 exp (2010/T)[NH2CI] [HOCI]
While this reaction has been reported to be catalyzed
by protons and acetic acid, the effectiveness of these
catalysts is such that they would not be of importance
at the usual concentrations encountered.
Given the situation where free chlorine is contacted
with ammonia, the initial velocity of the monochlor-
amine formation process given by Equation 5-38 is
substantially greater than the velocity of the sub-
sequent formation of dichloramine given by Equation
5-39. Hence, relative to equilibrium levels, there will
be an initial accumulation of monochloramine if large
dose ratios are used until the dichloramine formation
process can be driven. Thus in practice, the ratio of
dichloramine to monochloramine found may be less
than that given by equilibrium analysis as in Table
5-8.
Additional formation reactions of the chloramines are
disproportionation processes, such as Equation 5-37.
However, both Morris (82), and Gray et al. (79) provide
data to indicate that the formation of dichloramine via
monochloramine disproportionation is relatively slow
compared to Equation 5-38 under the usual condi-
tions encountered in wastewater disinfection.
Morris has also studied the reaction of organic
amines to form organic monochloramines (82). The
rate laws for these reactions follow a similar pattern
asthoseforthe inorganic monochloramine formation
process, except that the rate constants are generally
less. In addition, the rate constants for this process
correlate to the relative basicity of the amine reactant.
The mechanism of the breakpoint reaction has been
extensively studied by Morris, and more recently by
Saunier and Selleck (83). The oxidation of ammonia
nitrogen by chlorine to gaseous nitrogen would
theoretically yield a stoichiometric ratio of 1.5 moles
44
-------
of chlorine (CI2) consumed per mole of nitrogen
oxidized according to:
NH3 + 1.5HOCI = N2 + H+ + CP + H20 (5-40)
As noted previously, the observed minimum stoich-
iometric ratio between chlorine added and ammonia
nitrogen consumed at breakpoint is 2:1, suggesting
that more oxidized nitrogen compounds are produced
at breakpoint. Evidence (83) suggests that the prin-
cipal additional oxidized product may be nitrate
formed via:
(5-41)
NH4+ + 4HOCI = N03" + 4 CP + 6 H+ + H2O
As indicated, the production of nitrate from ammonia
results in the consumption of four moles of CI2(or four
moles of HOCI) per mole nitrate formed. Hence,
depending upon the relative amount of nitrate formed
in comparison to nitrogen at breakpoint, between 1.5
and 4.0 moles of available chlorine may be required,
which is consistent with the available data.
The breakpoint reaction consists of a complex series
of elementary reactions, of which Equations 5-40 and
5-41 are the net results. On the basis of extensive
kinetic investigations, Saunier and Selleck proposed
that hydroxylamine (NH2OH) and NOH may be inter-
mediates in this reaction (83). A complete kinetic
scheme for the breakpoint process as proposed by
these authors is given in Figure 5-5; however, there
does not yet appear to be sufficient experience with
the use of this reaction mechanism to justify its use in
wastewater applications.
At chlorine doses below breakpoint, the inorganic
chloramines can decompose by direct reactions with
several compounds. For example, Trofe et al. deter-
mined that monochloramine may react with bromide
ions to form monobromamine, the chemistry of which
will be discussed below (84). If trichloramine is
formed, as would be the case for applied chlorine
doses in excess of that required for breakpoint, it may
decompose either directly to form nitrogen gas and
hyppchlorous acid or by reaction with ammonia to
form monochloramine and dichloramine. Saguinsin
and Morris determined the kinetics of these process-
es, the relative importance of which are dependent
upon solution pH, trichloramine, and ammonia con-
centrations (85).
Organic chloramines may be formed by processes
similar to inorganic chloramine formation, although
usually at lower rates. Pure solutions of amino acids
and some proteins may display breakpoint curves of
identical shape to those of ammonia solutions (86,87).
Organic chloramines may also be formed by the direct
reaction between monochloramine and the organic
amine, and this is apparently the most significant
mechanism of organic N-chloramine formation at
higher concentrations such as might exist at the point
of application of chlorine to a wastewater (88).
Free chlorine residuals may react with other inorganic
compounds likely to be present in wastewaters. Table
5-9 summarizes available data on rates of these
processes (63). These reactions are generally first
order in each of the oxidizing agent (hypochlorous
acid or hypochlorite anion) and the reducing agent.
Notable among these possible reactions are the
reductions in the presence of nitrites, and sulfites.
Nitrites might be present in incompletely nitrified
effluents and react via a complex, pH-dependent
mechanism (89). The reaction with sulfite is used in
dechlorination and will be discussed subsequently.
Figure 5-5. Proposed kinetic mechanism forthe breakpoint reaction. (83) (Reproduced from American Water Works Association
by permission.)
pKb
NH3
HOCI
. NH2CI
HOCI
OH'
_>.NHCI2
NOi
OH"
Y
• NH2OH
K5
HOCI
HOCI
.NOH
NHCIa
->N2
o
o
pKm
pKa
NCI3
NH3OH+
ke
NHsCI
ocr
45
-------
Tablo 5-9. Summary of Kinetics of HOCI and OCI~ Reduc-
tion by Miscellaneous Reducing Agents (63)
Oxidizing
Agent
oci-
oci-
oci-
oci-
HOCI
HOCI
HOCI
HOCI
HOCI
HOCI
Reducing
Agenl:
io3-
oci-
CIO,-
SOs2"
N02-
HCOO-
• Br~
OCN-
HC204-
1-
Oxidation
Product
io4-
CI02-
CI03-
SO42~
N03- ,
H2C03
BrO~
HCO3-, N2
CO2
10-
Logk
(M-1 s1)
25°C
-5.04
-7.63
-5.48
3.93
0.82
-1.38
3.47
-0.55
1.20
8.52
Free chlorine can react with organic constituents to
produce chlorinated organic by-products. Murphy et
al. surveyed the reactivity of many classes of organic
materials, and indicated that phenols, amines, alde-
hydes, ketones and pyrrole groups are all readily
susceptible to chlorination (90). The first rigorous
study of this point was that of Granstrom and Lee,
who noted that phenol could be chlorinated by free
chlorine to form chlorophenols of various degrees of
substitution (54). The kinetics of this process were
dependent upon both phenolate ions and hypochlo-
rous acid. However, if excess ammonia was present
the formation of chlorophenols was substantially
inhibited.
In more recent work, De Laat et al. determined the
rate of reaction of a variety of organic compounds
with hypochlorous acid to form chloroform (91). This
reaction, the prototype of which is the reaction of
acetone with hypochlorous acid to form chloroform as
in the following, is of major interest in potable water
chlorination.
(5-42)
H3CCOCH3 + 3 HOCI = CHCI3 + H3CCOOH
+ 2 H20
Polyhydric phenols are substantially more reactive
than simple ketones in the production of chloroform,
and that the rates of these processes are typically first
order in each of the phenol concentration and the free
chlorine concentration (91). More significantly, it was
observed that the reactivity of these compounds,
many of which are analogous to materials such as
tannins and humic acids that could be present in
effluents, is greater than the reactivity of ammonia
with hypochlorous acid. Therefore, even if sub-break-
point chlorination is practiced, some chloroform may
be formed rapidly prior to the conversion of free to
combined chlorine.
Experimentally, this has been verified by Jolley who
noted that more than 1 percent of the applied chlorine
dose may be converted to chlorinated organic mater-
ials during wastewater chlorination to residuals on
the order of 12 mg/l (92). However the kinetics of this
process were sufficiently slow in that the formation of
organic chlorine compounds was found to increase
with time up to at Ieast4 hours following chlorination.
Similar findings were noted by Chow and Roberts,
who determined that the production of both organic
halogen compounds, and trihalomethanes (chloro-
form and its analogs) specifically was greater in
nitrified than in non-nitrified effluents (93). In break-
point chlorination of effluents, a large spectrum of
individual organic compounds may be isolated from
wastewater effluents at levels of 10 to 500 micro-
grams/I (94).
5.3.3.2 Bromine Demand
The reaction of free bromine (hypobromous acid and
hypobromite ion) follows many of the same pathways
as the reactions involving hypochlorous acid and
hypochlorites. The key differences in chemistry
between free chlorine and free bromine, as outlined
by LaPointe et al are as follows: (95)
• formation of bromamines is rapid in comparison
to chloramines;
• except for the irreversible oxidation of ammonia
in the breakpoint reaction, the bromamine
system is adequately described purely by equil-
ibrium relationships;
• at normal pH and Br:N dose ratios, the major
form of bromamine is dibromamine (NHBra)
versus monochlorarnine; and
• at low pH values and relatively high Br:N dose
ratios, tribromamine may be of significance,
whereas trichloramine almost always is insig-
nificant.
The comparative rate of formation of monobrom-
amine versus monochlorarnine has been studied by
Wajon and Morris (96). These results indicate that the
monobromamine is formed 14 to 20 times more
rapidly than monochlorarnine at pH 7 and equivalent
molar concentrations of free halogen and ammonia.
On this basis, it is satisfactory to describe the
predominance of various bromamine compounds by
the use of an equilibrium diagram in which pH and
N:Br molar dose ratio are the two master variables.
Such a diagram is reproduced in Figure 5-6 (97).
The oxidation of ammonia nitrogen by free bromine
residuals during the break point process follows the
analogous reaction as in the case of chlorine. As
described by Equation 5-43,1.5 moles of hypobromous
acid are consumed by the reaction with one mole of
ammonia. This corresponds to the equivalent of 1.5
46
-------
Figure 5-6. Distribution of bromamine species as a function
of pH and N:Br molar dose ratio (97). (Repro-
duced by permission, American Society of Civil
Engineers.)
moles of molecular bromine per mole ammonia
nitrogen (97).
(5-43)
2NH4+ + 3HOBr - N2 + 3Br' + 3H20 + 5H+
There do not appear to have been any detai led studies
on the mechanism of this reaction in real waste-
waters, or the production of more oxidized nitrogen
forms, such as nitrate, which the reports noted above
indicate may be produced during breakpoint chlo-
rination.lt is reasonableto suspect that theamount of
bromine consumed in the breakpoint reaction would
be somewhat greater than that indicated by Equation
5-43 due to the possible existence of these other
processes.
One factor that must be mentioned is that combined
bromine residuals (inorganic bromamines) possess
disinfection efficiencies comparable to free bromine
residuals. Thus, while the ammonia-bromine reac-
tions do exist and do occur in wastewater disinfection,
their significance in governing process performance
is somewhat less than the comparable ammonia-
chlorine reactions.
The kinetics of decomposition of bromamine residuals
have been studied in some detail in pure laboratory
systems (98,99). Both of these studies suggest a
mechanism of the bromine breakpoint process that
may involve the elementary reaction between di-
bromamine and tribromamine acting via a series of
complex intermediates. One consequence of this
process appears to be that high ammonia concentra-
tions may reduce the rate of the breakpoint reaction
due to a reduction in the amount of tribromamine
present (98). In addition, the decomposition of
dibromamine may lead to the reappearance of free
bromine (which could, in turn, form additional
amountsof combined bromine) via thelollowing (98):
(5-44)
NHBR2 + NBr3 + 2H20 = N2 + 3H+ + 3Br" + 2HOBr
Comparatively little information exists on the forma-
tion of organobromine compounds during wastewater
bromination. However, it is now well known from
water treatment practice that transient free bromine
may react to form bromoform and mixtures of
chlorinated and brominated compounds, and it would
be prudent to anticipate that these reactions would
also occur during wastewater bromination.
5.3.3.3 Chlorine Dioxide Demand
The reaction of chlorine dioxide with material present
in effluents resulting in chlorine dioxide demand
appears to be less significant than in the case of
bromine and chlorine. Rather, the dominant causes of
loss of chlorine dioxide from wastewaters may be the
direct reactions with water and interconversions to
chlorite and chloride, as outlined above.
In one of the only studies directly applicable to
wastewater, Roberts et a I'. (100) and Chow and
Roberts (93) determined the chlorine dioxide demand
of wastewater effluents and the formation of halo-
genated organics and trihalomethanes. It was found
that at equal mass doses (i.e. 10mg/l chlorine versus
10 mg/l chlorine dioxide), the chlorine dioxide demand
of a wastewater (expressed as electron equivalents)
was less than the chlorine demand of that waste-
water. In addition, no trihalomethanes were formed
from the application of chlorine-free chlorine dioxide
(prepared via the acid-chlorite process, rather than
the chlorine-chlorite process), and the amount of
halogenated organics produced was 10 to 20 times
less than those produced by equal mass doses of
chlorine.
It has been established that, at usual concentrations,
ammonia nitrogen, peptone, urea, and glucose have
insignificant chlorine dioxide demand in 1 hour
(101,102).
Masschelein reviewed the reactions of various
classes of organic materials with chlorine dioxide,
and only the following would appear to be of
significance to wastewater applications (51):
1. ClOa may oxidize tertiary amines to secondary
amines and aldehydes, with the formation of
chlorite. For trimethylamine, this reaction is first
order in chlorine dioxide and the amine, and has
a rate constant of approximately 100,000 liters/
mol-s.
2. Ketones, aldehydes, and alcohols may be oxi-
dized to acids. While Masschelein cites this as a
significant reaction, Somsen indicates that the
rate of ethanol reaction with chlorine dioxide is
insignificant at pH 7, while carbonyl compounds
do exhibit reactivity (103).
47
-------
3. Phenols and phenol derivatives may react with
chlorine dioxide to form oxidized and chlori-
nated products.
4. Sulfhydryl amino acids, particularly cystine,
may be oxidized to cysteic acid.
Wajon et al. have recently studied the reaction of
phenols with chlorine dioxide under dilute conditions
typical of water and vvastewater disinfection applica-
tions (1 04). The reaction's stoichiometry is 2 moles of
chlorine dioxide consumed per mole of phenol (or
hydroquinone) consumed. Products formed included
chlorophenols, aliphatic organic acids, benzonqui-
none, and (in the case of phenol) hydroquinone. The
mechanism appeared to include the possible forma-
tion of hypochlorous acid as an intermediate which
would thus chlorinate, and the rate of this process
was found to obey the following law:
(5-45)
r = 2{kb + kaKa/[H"]) [Phenol] [ClOa] (mol/l-s)
Where Ka is the acidity constant for the phenol, and kb
and ka are the rate constants. For phenol, these were
0.24 and 2.4 x 107 l/mol-s, respectively, while for
hydroquinone, these rate constants were respectively
39,000 and 6.5 x 109 l/mol-s.
5.3.4 Dechlorination Chemistry
In certain circumstances, it is desirable to reduce the
chlorine residual in a disinfected wastewater prior to
discharge. In order to do this, it is necessary to contact
the chlorinated wastewater with a substance that
reacts with, or accelerates the rate of decomposition
of, the residual chlorine. While many compounds may
perform this function, including thiosulfate, hydrogen
peroxide, ammonia, sulfite/bisulfite/sulfur dioxide,
and activated carbon, only the latter two materials
have been widely used for this purpose in either
water or wastewater treatment (105). In this section,
the chemistry of the reaction between residual
chlorine and the latter two substances will be
discussed.
5.3.4.1 Sulfur (IV) Compounds
Sulfur dioxide, and its aqueous dissolution products,
sulfite and bisulfite ions, are reduced sulfur com-
pounds with an oxidation state of +4. They may react
with free or combined chlorine residuals to form
oxidized sulfur products, principally-sulfate, and
hence convert available chlorine to chloride.
Commercially, S(IV) used for dechlorination is sup-
plied principally as sulfur dioxide under pressure as a
gas in equilibrium with its liquid. Sulfur dioxide is an
irritant gas, with a water solubility 20 times greater
than chlorine gas, and with a gas density relative to
air of 2.26 (106). Sulfur dioxide gas is supplied in
cylinders similar to those used for the supply of
gaseous chlorine.
Table 5-10 summarizes physical properties of sulfur
dioxide at various temperatures.
Table 5-10. Physical Properties of Sulfur Dioxide (106)
Temperature
-40
-29
-18
-6
4
16
27
Liquid
Density
(g/ml)
1.530
1.510
1.480
1.450
1.420
1.390
1.360
Vapor
Pressure
(atm)
—
—
0.15
0.81
1.74
3.83
Solubility
(g/D
—
—
—
199.8
130.5
87.6
Upon addition to water, sulfur dioxide forms sulfur-
ous acid (HaSOs), which may lose, successively, two
protons, according to the following reactions:
HaSO3 = H+ + HS03" (5-46)
HS03~ = H+ + S03~2 (5-47)
The pK for the second dissociation (Equation 5-47) is
7.2 at room temperature (107), and hence sulfur (IV)
is present in most wastewater effluents as a mixture
of both sulfite and bisulfite ions. In some circum-
stances, sulfur (IV) may be applied as sodium sulfite
or bisulfite, and the resulting equilibrium mixture is
established via the above equations; however, the
form of application of sulfur (IV) does not, in and of
itself, affect the nature of the subsequent dechlori-
nation reactions.
The reaction between sulfur dioxide and free chlorine
is such that one mole of sulfur dioxide reacts with one
mole of either free chlorine or monochloramine via
the following stoichiometric equations (105):
S02 + H20 + HOCI = 3H+ + CI" + S04
_(5-48)
(5-49)
oc-> -2
SO2 + 2H2O + NH2CI = NH4+ + 2H+ + CI" + SO4
The rate of reaction between free chlorine and sulfur
(IV) compounds appears to be quite rapid. This rate
has been found to be first order in each of sulfite and
free chlorine concentration. However, Lister and
Rosenblum (108) have indicated that the kinetics are
dependent on OCf concentration, while Srivastava et
al. (109) proposed that HOCI is the reactive species.
48
-------
i The rate laws given by these investigations are,
i respectively:
• r - 2.7 x 109 exp(-3773/T) (5-50)
: x [OCr][SO3"2](M/l-s)
j r = 1.2 x 1015 exp(-7851/T) (5-51)
i x [HOCI][SO3~23(M/l-s)
i At 25°C and 0.2 mM sulfite concentrations (16 mg/l
' as sulfite), these rate laws give half-lives of 0.4 and
j 0.8 s, respectively. Thus, it appears, for free chlorine,
that the rate of reaction is sufficiently rapid so as to be
regarded as instantaneous (80).
Surprisingly, there does not appear to have been any
direct measurements made of the rate of reaction.of
sulfur (IV) species with inorganic combined chlorine.
However, based on the results reported for organic
chloramines, it may be expected that the dechlorina-
tion of combined chlorine residuals is slightly slower
than free chlorine. Stanbro and Lenkevich reported
that the sulfite dechlorination of organic chloramine
residuals follow kinetics first order in both total
reduced sulfur and chloramine concentrations (110).
For the monochloramine derivatives or methylamine,
N-alpha-acetylysine, alanine, leucine, and alanyl-
alanylalanine(N-chloro derivative), at pH 7 and 25°C,
the measured second order rate constants were
4,333, 2,167, 1,167, 2,333, and 83 liters/mol-s. For
the peptide chloro derivative, in particular, this is suf-
ficiently slow to suggest that small amounts of some
organochloramines may not be dechlorinated com-
pletely in some sulfur (IV) processes, particularly
where extremely short contact times are used.
One side reaction that may occur during sulfur
dioxide dechlorination, and that may be of concern for
the receiving water quality, is the deoxygenation of
the effluent. This occurs by the following reaction
stoichiometry (105):
O2 + 2S02 + 2H20 = 4H+ + 2SO4~2 (5-52)
This reaction, which is used during the standard
oxygen transfer tests (111) is catalyzed by a variety of
trace metals, including copper, cobalt, iron, cerium,
and manganese, and is inhibited by ethanol, glycerol
and mannitol. In uncatalyzed systems, at sulfur (IV)
concentrations below 0.02 M it has been found to be
first order in reduced sulfur and zero order in oxygen,
with a first order rate constant of 0.005 s"1 at pH 7.35,
and increasing rates as pH increases (105). Thus,
unless catalyzed, the rate of deoxygenation is much
less than the rate of sulfur dioxide dechlorination. In
pilot studies, it has been noted that no significant
oxygen depletion occurred until sulfur dioxide over-
doses exceeded 50 mg/l (112). However, the degree
to which catalysis of Equation 5-52 might occur in
dechlorinated effluents has not been thoroughly
explored.
5.3.4.2 Activated Carbon Dechlorination
Activated carbon is a material prepared by controlled
combustion and oxidation of an organic material. This
section will describe the chemical properties of
activated carbon peculiar to dechlorination applica-
tions. However, it should be noted that the use of
activated carbon in wastewater treatment may simul-
taneously achieve other objectives, such as the
reduction of total organic carbon to low levels, or the
specific removal of biologically refractory trace pol-
lutants.
The reduction of free chlorine residuals by activated
carbon at low loadings has been found to be due to an
actual reaction of the chlorine with the carbon
surface, rather than merely by a catalytic decomposi-
tion (113). Products of this reaction include chloride
and surface-bound oxides of carbon that evolve as
carbon monoxide and carbon dioxide, as well as
water, when the spent carbon is thermally regener-
ated. Hence, the dechlorination process using acti-
vated carbon represents a consumptive use of the
activated carbon itself.
At higher loadings, once the surface of the carbon has
accumulated surface oxides (denoted by CO*), reduc-
tion of free chlorine residuals may continue to occur.
This has been postulated to be due to the catalysis of
free chlorine decomposition to chloride and chloric
acid according to (114):
3HOCI = 2HCI + HCIO3
(5-53)
The reaction of combined chlorine residuals with the
carbon surface is more complex, and differs between
mono and dichloramine. For fresh carbon, before a
large amount of surface oxides have formed, mono-
chloramine is reduced by the carbon surf ace (denoted
by C*) to form surface oxides (which are converted to
carbon monoxide and dioxide when thermally regen-
erated) according to (115):
(5-54)
NH2CI + H20 + C* = NH4+ + CP + CO*
After some surface ^oxides have formed, they may
react with monochloramine by the additional reaction
(115):
(5-55)
2NH2CI + CO* = N2 + 2H+ + 2CI" + H20 + C*
Reaction 5-55 serves to reduce the concentration of
ammonia nitrogen present in a wastewater, and this
process has been suggested as a means for the
removal of ammonia nitrogen from effluents.
49
-------
Dichloramine residuals appear to react primarily by
the formation of nitrogen gas and surface oxides by
(115):
(5-56)
2NHCI2 + H2O + C* = N2 + 4H+ + 4CI" + CO*
The kinetics of monochloramine-carbon dechlorina-
tion occurring by Equations 5-54 and 5-55 have been
studied in detail (116). Since the rate of monochlor-
amine decomposition is accelerated by the formation
of surface oxides, the overall process is autocatalytic,
and for low effluent monochloramine residuals,
prolonged contact times are necessary—for example,
for a 2 mg/l monochloramine residual, 16 minutes
contact in a packed bed carbon contactor are neces-
sary to reduce the residual below 0.3 mg/l.
The reaction between dichloramine and carbon is
faster than that between free chlorine and carbon,
and much faster than that between monochloramine
and carbon. In kinetic investigations, Kim et al.,
(117,118) determined that the reduction of dichlor-
amine occurred by Equation 5-56 in parallel with a
catalytic breakpoint reaction, given by:
NH4* + 3NHCI2 = 2N2 + 6CI" + 7H+ (5-57)
Hence, dechlorination of predominantly dichloramine
residuals will also produce significant degrees of
ammonia nitrogen removal.
Since, as noted above, in sub-breakpoint chlorination,
monochloramine predominates, it would appear that
the design of activated ciarbon beds for dechlorination
would be limited by the reduction of monochloramine
residuals, which are the more slowly reacting form,
particularly when final chlorine residuals below 0.1
mg/l were desired.
Only a small number of wastewater treatment plants
practice GAC dechlorination. Other than the models
of Kim et al. (118) and Kim and Snoeyink (116), little
guidance for design is available.
5.4 Analysis of Disinfectant Residuals
The monitoring and control of halogenation process-
es requires that a means be available for the analysis
of residual disinfectant, which, as previously noted, is
one major variable governing microbial inactivation.
Ideally, the analytical method should be simple,
rapidly performed, and not subject to analytical
interferences by other constituents likely to be
present in the wastewater.
5.4. / Chlorine
Chlorine may be analyzed as total residual, including
all chemical forms capable of functioning as oxidizing
agents (free chlorine plus inorganic chloramines), or
it may be differentiated as free chlorine and combined
chlorine. Each of the latter two fractions may be
further subdivided; free chlorine into HOCI and OCI~,
and combined chlorine into mono-, di-, and tri-chlor-
amine (as well as organic chloramine compounds). In
most cases, however, the concentration of total
residual chlorine is sufficient to allow adequate
process monitoring.
All currently available methods for the analysis of
total residual chlorine rely upon the oxidation of
iodide to form molecular iodine, which then can be
analyzed. The chemistry of this process, which is
favored at low pH, is as follows:
HOCI + H* + 2\~ = I2 + CI" + H20 (5-58)
OCF + 2H+ + 2I~ = I2 + CI" + H20 (5-59)
NHaCI + 2H+ + 2I~ = NH4+ + I2 + CI" (5-60)
NHCI2 +3H+ + 41" = NH4 + 2I2 + 2CI" (5-61)
NCI3 + 4H+ + 6I~ = NH4+ + 3I2 + 3CI" (5-62)
At a pH of 4.0, potassium iodide will react with all of
the above forms of chlorine residual to produce
iodine.
At this point, the analysis can be conducted in two
general manners. First, in the so-called forward
methods, the released iodine may be directly meas-
ured by titration using either a starch colorimetric
indicator or amperometric measurements for end-
point determination. In the reverse methods, simul-
taneously with the production of iodine in Equations
5-58 to 5-62, reductant is added. By measuring the
amount of this reductant that has been oxidized
during the analysis, the amount of total residual
chlorine may be determined.
The two most common forward titration methods
involve the use of starch indicator in conjunction with
sodium thiosulfate (Na2S2O3) titrant and the use of
amperometric titration equipment in conjunction with
phenylarsine oxide titrant.
In the first case, sodium thiosulfate is added to the
sample until the reddish iodine color is almost
discharged. At this point, starch indicator is added,
which forms a strong bluish complex with iodine, and
sufficient additional titrant is added until the color
disappears. The amount of sodium thiosulfate that
has been added is equivalent to the amount of iodine
that had been released, and this in turn is equivalent
to the total residual chlorine.
The second procedure substitutes the amperometric
titrator for the starch indicator solution (119). This
instrument measures the current developed across
electrodes maintained at a constant potential suf-
ficient to reduce free halogen residuals. In the forward
50
-------
titration, phenylarsine oxide (PAO) titrant is added
until there is no further change in the indicated
current; this is the point at which all halogen residual
has been reduced, and the amount of PAO added is
equivalent to the total residual chlorine.
In the reverse titration, the amount of consumed
reductant (either thiosulfate or PAO) is measured by
titration with standard iodine solution using either
the starch indicator or amperometric end-point
detection methods. Alternatively, the reductant used
may be a compound that changes color with oxidation,
and thus the amount of residual chlorine may be
determined either colorimetrically or by titrating the
color with reductant until disappearance. This latter
procedure is exemplified by the use of DPD (N,N-
diethyl-p-phenylenediamine) or Leuco Crystal Violet
(LCV).
In the DPD procedure, DPD reagent is added prior to
iodine addition in the reverse titration. Oxidized DPD
is reddish in color, whereas DPD itself is colorless;
thus the amount of residual chlorine may be deter-
mined by a spectrophotometric determination of the
amount of oxidized DPD produced. This method is
widely available in a field kit form. Alternatively, the
oxidized DPD may be titrated with ferrous ammonium
sulfate reductant.
The LCV method is a colorimetric assay based upon
the vivid bluish color exhibited by oxidized LCV.
For each of the above procedures, detailed methods
may be found in Standard Methods (111). All colori-
metric and visual titrimetric methods are subject to'
interferences in highly colored or turbid wastewaters.
Oxidized manganese and chromium, as well as other
materials capable of converting iodide to iodine,
interfere with all of the above methods. Silver or Cu(l)
ion interfere with the amperometric methods.
In previous years, orthotolidine reagent had also been
used for total chlorine determinations in wastewater.
However, a number of studies have indicated that the
OT method is substantially less satisfactory than
either the amperometric or starch iodide titration
methods (120-126). In addition, orthotolidine is a
suspect carcinogen and therefore, on the basis of
these two reasons, this method is no longer approved
(111).
The amperometric and starch-iodide forward titration
methods for total chlorine residuals have been found
to produce comparable results (122). For waste-
waters, it is recommended that the total chlorine
reverse titrations be used. In the forward methods,
free iodine residuals are in contact with the waste-
water for a period of ti me during the completion of the
titration. During this time, there may be a loss of
halogen residual due to exertion of demand on the
part of the wastewater. In the reverse titrations, any
iodine produced is rapidly removed by reaction with
the reducing reagent (thiosulfate, PAO, LCV or DPD),
and thus there is little opportunity for demand
exertion.
Experimental studies comparing these methods,
however, have tended to yield disparate results,
which appear to vary with the quality of the waste-
water being chlorinated. Browning and McLaren
(127) developed a modified starch-iodide forward
titration (MSI), which was found to yield results
closely comparable to the reverse amperometric
method and greater than the forward amperometric
method, and reverse starch-iodide methods. In con-
trast, Collins and Deaner (128) found that the MSI
procedure consistently produced lower chlorine
residual readings than did the reverse starch-iodide
method, and that the extent of this difference
increased with the degree of pollution of the effluent
(e.g. BOD).
Lin et al. compared the reverse starch iodide titration
(SIB), the DPD method, the LCV method and the MSI
titration using three wastewaters. These authors
concluded that the MSI procedure generally, but not
always, yielded results less than SIB and LCV
methods, but that the results were indistinguishable
from the DPD procedure. Furthermore, following
chlorine addition, in some cases, the total chlorine
residual measured by the DPD and LCV procedures
increased with time; the MSI procedure did not
produce such anomalous results. On this basis, Lin et
al. suggested that the MSI procedure was preferable
to either DPD or LCV methods (129).
Recent data suggest that all of the methods used for
total chlorine residual may produce false negative
readings if sulfite or similar reducing agents are
present(110). This situation may occur when dechlor-
ination using S(IV) compounds is practiced, partic-
ularly if organic chloramines are also present. In this
case, it is possiblethattotal chlorine residuals actually
exist in the effluent, but that when the pH is lowered
during the initial addition of iodide, the reaction
between S(IV) and chlorine residuals may be more
rapid than the reaction with iodide. Thus, chlorine
may be deemed to be absent when it is in fact present.
Two additional methods for the analysis of total
residual chlorine in wastewater have been reported;
however, neither of them has yet appeared in
Standard Methods. Jenkins and Baird (130) utilized a
polarographic iodine electrode to measure the iodine
released following acid-iodide addition to waste-
water. This method is essentially a forward titration
procedure using a direct electrode analysis of iodine
production. It was found that this method was equal
to the SIB method in precision and accuracy at total
SI
-------
chlorine residuals above 1 mg/l; below 1 mg/l, the
iodine electrode procedure was more precise than the
SIB procedure, which is not recommended for
residuals below 1 rng/l (111). The principal inter-
ference with the electrode iodometric assay occurred
in effluents with more than 50 mg/l BOD—in these
cases, the iodine progressively declined to zero prior
to a stable electrode reading, presumably due to
exertion of iodine demand.
Liebermann et al. (131) investigated the use of the
syringaldizine reagent for total chlorine determina-
tion following acid-KI addition, with a colorimetric
assay. This test yielded a linear response to 10 mg/l
total chlorine, and was found to be comparable to the
forward amperometric method. Syringaldizine forms
the basis of the FACTS test for free chlorine which
has been accorded tentative status in the current
edition of Standard Methods (111).
In certain cases, particularly in breakpoint chlorina-
tion for nitrogen removal, or the chlorination of
nitrified effluents, it may be desirable to measure free
chlorine apart from total chlorine. The iodometric
starch-iodide, amperometric, DPD, and LCV methods
may be modified to be relatively selective for free
chlorine. Further details on these methods may be
obtained in Standard Methods (111). It should be
noted that the determination of small concentrations
of free chlorine in the presence of large amounts of
total chlorine (as might be found immediately prior to
the breakpoint itself) is an extremely difficult ana-
lytical problem (132).
5.4.2 Bromine
The analysis of bromine residuals, as might be
produced from BrCI treatment, proceeds directly
analogous to chlorine residuals. Free or combined
bromine can oxidize iodide to iodine, and any of the
methods for total chlorine can be used to measure
total bromine. If mixtures of chlorine and bromine
residuals are present, it becomes impossible to
distinguish between the two by iodometric methods.
5.4.3 Chlorine Dioxide
For the determination of chlorine dioxide in waste-
waters, it is particularly necessary to differentiate this
compound from chlorite (which is produced as a
reaction product) and free and combined chlorine
(which may arise if the chlorine-chlorite process is
used to produce chlorine dioxide). Both the DPD and
the amperometric methods may be modified to
determine chlorine dioxide apart from total chlorine
(111); if the total chlorine methods are used, they Will
tend to indicate the sum of total chlorine, chlorine
dioxide, and a portion of the chlorite. Roberts et al.
(100) indicated that, as for total chlorine, use of the
reverse titration procedures produces results in
excess of the forward titration methods.
In addition to the above procedures, Knechtel et al.
(133) determined that acid chrome violet K (ACVK)
reagent is suitable for the direct colorimetric deter-
mination of chlorine dioxide in wastewaters. ACVK is
not subject to interference by free or combined
chlorine, chlorite, or nitrite, and results in waste-
water were found to be comparable to determinations
using electron spin resonance., However, if large
amounts of turbidity are present, a pretreatment via
centrifugation is necessary to prevent optical inter-
ference.
Similarly, Wheeler et al. (134) proposed chlorophenol
red (CPR) as a selective indicator for chlorine dioxide;
this compound is suitable for use in a direct colori-
metric or a forward titrimetric assay. No interferences
have been reported by chlorite, chlorate, free chlorine,
Cr(VI), Fe(lll), or Mn(V).
5.5 Kinetics of Microbial Inactivation
The information needed for the design of a disinfec-
tion system includes knowledge of the rate of
inactivation of the target, or indicator, organism(s) by
the disinfectant. In particular, the effect of disinfect-
ant concentration on the rate of this process will
determine the most efficient combination of contact
time (i.e. basin volume at a given design flow rate) and
dose to use.
Chick's Law and deviations therefrom are discussed
in Section 4-1, and the reader is referred to that
section for the historical development of disinfection
theory. The discussion that follows here is a practical
and logical outgrowth of that theory.
In the disinfection of a wastewater by halogens the
concentration of disinfectant changes with time, and,
particularlyduringthe initial moments of contact with
either chlorine or bromine, the chemical form of the
halogen undergoes a rapid transformation from the
free to the combined forms. Since C in Equation 4-6
would thus not be a constant, typically wastewater
disinfection results obtained in batch systems exhibit
"tailing," the degree of which may depend on the
chlorine demand and the ammonia concentration of
the system (135). It is more critical to determine what
the chlorine residual is, rather than the'chlorine dose
in these systems. According to Heukelekian and
Smith: "...the control of chlorination cannot be based
on the dosage of chlorine, because neither for sewage
from different sources, nor for a sewage from a given
source, but at different times, does a constant dosage
of chlorine produce a constant number of coliform
organisms per ml in effluent." (136)
As a result, the analysis of the kinetics of microbial
inactivation in wastewater effluents by disinfectants
has often been accomplished by empirical models
deviating substantially from Chick's Law. For example,
52
-------
McKee et al., chlorinated primary effluent, and found
that his data could be described by (70):
log(N) =
bC0)
(5-63)
where N is the viable microorganism concentration
and, C0 is the chlorine dose; in this equation, the
contact time is held constant. McKee et al. also
proposed the following equation as a means to
extrapolate inactivation results from onecontact time
to another (70):
= (t/a)n
(5-64)
In Equation 5-64, Nt and Na are the microorganisms
remaining at time t and time a, respectively (chlorine
dose fixed), and n' is a constant.
The following equation, similar in concept to McKee's
model, was developed to describe inactivation in
various chlorinated effluents (137,138):
N/No = (bCt)~
(5-65)
In Equation 5-65, a and b are constants, and C is the
chlorine residual remaining at time t. At values of the
product (Ct)< 1/b, by definition, N/N0 equals unity.
The use of this equation will be further discussed
below.
While it is recognized that the presence of solids in
wastewater effluents may harbor and protect en-
meshed microorganisms from the action of disinfect-
ants, few methods are available for the quantitative
consideration of these phenomena.
However, it has been shown that organic solid
materials present in actual wastewaters can render
enmeshed microorganisms some protection from a
measurable chlorine residual (139,140).
Analysis of data on the inactivation of a variety of
microorganisms by free and combined chlorine
suggests that the combined Chick-Watson law(Equa-
tion 4-4 substituted into Equation 4-3) provides a
satisfactory description of the disinfection process.
Values for k' and n are a function of the micro-
organism, the pH, and the temperature of the system,
and also differ depending upon whether free or
combined chlorine species are used as disinfectants.
Table 5-11 'summarizes a number of values of the
Chick-Watson parameters for microbial inactivation.
In general, microbial inactivation increases with a
decrease in pH for either free or combined chlorine
residuals. An increase in temperature also increases
inactivation. At a given temperature and pH, free
chlorine residuals are more effective than combined
chlorine residuals. These data may underestimate
the resistance of microorganisms in wastewater.
However, the degree of underestimation is unknown.
Table 5-11. Chick- Watson Parameters for Microbial
Reference Organism
141 £ co//
Aerobacter
aerogenes
Pseudomonas
Pyocyanea
Salmonella
typhi
Shigella
dysenteriae
142 £ co//
/W/crococcus
pyogenes var.
aureus
Inactivation
PH
8.5
9.8
10.7
8.5
9.8
10.7
7
8.5
9.8
10.7
7
9.8
7
8.5
7
7
7
by Chlorine
Temp.
(°C)
FREE CHLORINE
2-5
2-5
2-5
20-25
20-25
20-20
20-25
20-25
20-25
20-25
2-5
2-5
20-25
20-25
20-25
25
25
k'
lnmg~nmin~1
10.9
1.18
0.279
30.6
5.91
1.30
1.39X104
312
2.13
0.738
7.87 X1012
0,962
8.15X106
2.45X1 04
9.07 xlO7
4.02
3.32
n
1.2
1.29
1.11
1.46
1.34
0.79
3.78
2.74
1.26
0.711
7.26
0.76
4.07
1.78
4.92
.801
1.10
143
Bacillus metiens
10
6
7
8
9
20
20
20
20
20
.00577
.029
.0219
.0209
.008
.483
1.24
1.18
1.12
0.99
53
-------
Table 5-11. Continued
Reference
144
145
Organism
Poliovirus
Type I
(Mahoney)
Poliovirus
Type I
(Brunhilde)
PH
9.35
12.86
10
10
10
6
6
6
6
6
10
Temp.
(°C)
20
20
30
35
50
2
10
20
30
20
20
k'
lnmg-nmin-1
.0086
.0015
.00324
.0044
.0075
11.28
12.78
30.12
75.12
12.20
6.5
n
1.04
.58
.868
1.0
1.26
.766
.818
.615
.608
0.69
3.23
146
147
Naegleria
7.2 25
COMBINED CHLORINE
.171
.786
Mycobacterium
phlei
M. fortuitum
Candida
parapsilosis
E. coll
148 E co//
Aerobacter
Aerogenes
Shigella
dysenteriae
Salmonella
typhi
Pseudomonas
pyocyanus
7
7
7
7
7
7
7
8.5
9.5
6.5
7.0
7.8
8.5
9.5
10.5
6.5
7
7.8
8.5
9.5
10.5
6.5
7.0
7.8
8.5
9.5
10.5
7
8.5
9.5
6.5
7.0
7.8
8.5
9.5
6.5
7.0
7.8
8.5
9.5
20
20
5
20
5
20
35
35
35
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
2-6
2-6
2-6
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
.226
6.21x10-"
5.68 X10~4
6.85X1 0-3
.656
1.01
.084
.0109
2.48X10-5
.483
.316
.193
.0854
.049
.0125
.363
.241
.095
.0715
.0358
.00809
.821
.55
.341
.151
.064
.0301
.0902
.0182
6.8X10-"
.491
.290
.211
.113
.0417
.44
.301
.174
.102
.0483
.545
3.51
2.68
2.38
1.16
0
1.39
1.52
13.3
1.07
1.04
1.18
1.125
1.37
2.27
1.19
1.35
1.18
.917
1.16
1.7
1.3
1.15
1.32
1.02
.995
1.52
1.32
1.67
6.26
1.13
1.84
1.07
1.16
.878
1.27
1.44 '
1.55
1.01
1.05
The use of the parameters in Table 5-11 for the
description of wastewater microbial inactivation
entails the assumption that the microorganisms for
which k' and n have been measured are identical (in
54
-------
terms of sensitivity towards disinfectant) to those
likely to be present in wastewater to be disinfected.
Unfortunately, all of the data summarized in Table
5-11 are based upon the use of pure, laboratory
strains of microorganisms. These may have different
sensitivities as compared with indigenous waste-
water organisms.
As an alternative approach, the model of Collins et al.
(Equation 5-65) has been used to describe inactiva-
tion of coliforms indigenous to wastewater. Table 5-
12 summarizes values for the a and b parameters in
Equation 5-65 which have been determined for the
iriactivation of coliforms indigenous to wastewaters.
Table 5-12. Parameters in the Collins et al. Model Describing
Wastewater Coliform Inactivation by Chlorine
(177)
System
Palo Alto— 1978
Palo Alto— 1979— Unfiltered
Palo Alto— 1979— Lab Filtered
Dublin — San Ramon Lab
Dublin — San Ramon Field
San Jose
b
l/mg-min
.193
.452
1.190
1.54
0.598
0.246
a
3.15
2.22
2.10
1.62
1.79
2.82
Note - Palo Alto and San Jose are non-nitrified secondary efflu-
ents. Dublin is nitrified activated sludge.
Use of the Chick-Watson and the Collins et al. models
for the estimation of microbial inactivation in waste-
water chlorination are described in Section 5.7.3.
5.6 Process Options
Disinfection systems using chlorine, chlorine dioxide,
or bromine chloride, with or without dechlorination
share certain common elements. For each of these
elements, several possible options exist for the
system designer to consider. By specifying the
potential options, the many available halogenation
disinfection systems may be enumerated.
In this section, these elements and options will be
summarized. In the following section, the method-
ology for detailed sizing and specification of each
option element will be discussed.
The general block diagram of a halogen disinfection
system, with optional dechlorination, is shown in
Figure 5-7. In general, the disinfection system
contains four elements and the dechlorination system
contains three elements. Flows of information are
depicted by dashed lines and flows of material are
depicted by solid lines.
The disinfection subsystem elements are chemical
generation and/or supply, mixing, contacting and
control. In the generation/supply element the dis-
infecting agent is maintained and fed continuously. In
the mixing element, the incoming wastewater is
blended with the disinfectant. In the contacting
element, the blended wastewater is held for a
sufficient period until the desired microbial inactiva-
tion has been attained. The control element uses
Figure 5-7. Elements of halogen disinfection systems, with optional dechlorination.
Disinfectant
Generation
and/or
Supply
Dechlorination
Chemical
Supply
Contro
Control
Dechlorination Subsystem
(Optional)
Hydraulic Flow
Information Flow
-------
information about incoming wastewater flow and
residual disinfectant to adjust the operating param-
eters of the generation and supply element so as to
maintain consistent disinfectant performance.
In the dechlorination subsystem, the dechlorination
chemical supply element functions to provide a feed.
The mixing elements and the control elements have
identical functions as in the case of the disinfection
subsystem. In dechlorination, there is generally no
substantial contacting element (other than, perhaps,
the final outfall pipe).
For the halogen generation/supply subsystem, a
number of choices exist as to the option to be
selected. These may be enumerated as follows:
1. Choice of chemical
A. Chlorine Gas
1. Method of Withdrawal
(a) Gas
(b) Liquid/external evaporation
2. Method of feeding
(a) Solution
(b) Direct Gas
B. Sodium Hypochlorite
1. Commercial supply
2. Onsite generation
C. Calcium Hypochlorite
1. Solid Feed
2. Solution Feed
D. Chlorine Dioxide
1. Method of Generation
(a) Chlorine-chlorite process
(b) Acid-chlorite process
E. Bromine Chloride
For chlorine gas, the choice of gas or liquid withdrawal
is dictated primarily by the size of the disinfection
system to be designed. It is impractical to use gas
withdrawal in very large systems due to limitations on
safe withdrawal of gas from ton containers of
chlorine. This decision between solution and direct
gas feed processes is also based upon the size of the
plant; in all but very small wastewater treatment
plants (less than 1 MGDjl, or if chlorine is only to be
used on an occasional bas is, the solution feed process
is generally used.
When sodium hypochlorite is to be used as the
disinfecting agent, the chemical maybe purchased in
bulk as a liquid, or may be produced continuously
onsite by electrolysis of brine. The latter process
becomes feasible at remote locations, or at waste-
water treatment plants where inexpensive sources of
brine are available (i.e., plants on ocean coast lines).
The use of calcium hypochlorite is confined primarily
to smaller wastewater treatment plants due to
economics of chemical supply. In these systems, the
chemical may be fed directly as a solid (proprietary
tablet feed chlorinators) or may be prepared as a
slurry/solution and fed in a similar manner as sodium
hypochlorite. Solution feed of calcium hypochlorite
has inherent disadvantages in that the addition of
water to calcium hypochlorite will produce a calcium
carbonate sludge that will foul surfaces and storage
tanks.
The choice among gaseous chlorine/sodium hypo-
chlorite/calcium hypochlorite is governed primarily
by economic considerations in conjunction with
inherent safety and handling hazards with the use of
gaseous chlorine. In wastewater treatment plants
serving large municipalities, it may be advantageous
to use sodium hypochlorite, despite higher unit costs,
to minimize risks associated with the transport of
liquid chlorine through populated urban areas.
In chlorine dioxide systems, the primary design choice
in the generation/ supply subsystem is the type of the
generation process. As discussed above, the utiliza-
tion of chlorite is greater in the chlorine/chlorite
process than in the acid-chlorite process. However, in
the former process, the possibility for the production
of chlorine dioxide containing chlorine at low levels
exists, and thus some of the deleterious aspects of
chlorine residuals and byproduct reactions may exist.
For BrCI systems, liquid withdrawal with external
evaporation is necessary to prevent chemical dissoci-
ation and enrichment of gaseous chlorine, which
would occur in gas phase withdrawal. Solution feed
systems for BrCI plants are universally used.
In the mixing subsystem, the primary design options
are the use of devices (static mixers or hydraulic
jumps) that dissipate hydraulic head and the use of
mechanical mixers (jet mixers, impeller mixers, etc.)
that require external motive power such as pumps or
motors. In plants where there is little available head
between the influent and effluent (such as may exist
in wastewater treatment plants serving municipal-
ities with relatively flat topography), the use of the
former processes may be infeasible. In general,
however, the former processes offer an advantage in
that their performance is insensitive to any power
fluctuations or outages that may occur, and thus they
are inherently more reliable.
The contacting subsystem may consist of a separate
contact basin or the outfall pipe itself. The latter
option reduces capital costs; however, it necessitates
the existence of a sufficiently long outfall pipe, which
may not exist at all plants.
The control subsystem serves to produce an effluent
in which a consistent effluent quality is attained at a
minimum chemical dose. Three options exist: manual,
56
-------
or "open-loop" control (relying upon operator inter-
vention), simple flow-proportional feed-forward con-
trol, and true closed-loop feedback control (relying
upon continuous sensing of halogen residual, usually
in conjunction with feed-forward flow proportional
control). The last option provides minimum chemical
utilization, but is more capital intensive. In very small
plants, where dechlorination is not practiced, manual,
or feed-forward control is sufficient. However, in
plants subject to effluent constraints on chlorine
residual, or in very large plants where small decreases
in chemical dosage result in large economic savings,
the last option may be most desirable. Also, in
dechlorination plants, the consumption of dechlori-
nating chemical will be a function of the chlorine
residual leaving the chlorine contact system, and
thus the use of true feedback control will help to
minimize costs associated with the dechlorination
process.
For the dechlorination sub-system, the options for
chemical supply and generation are as follows:
Choice of chemicals
A. Sulfur Dioxide
1. Method of Withdrawal
(a) Gas
(b) Liquid/External evaporation
B. Sulfites or Bisulfites
The choice of sulfur dioxide versus sulf ites/ bisulf ites
is dictated by considerations similar to those that
govern the selection of gaseous chlorine versus
calcium or sodium hypochlorites. If sulfur dioxide is
used, selection of gas or liquid withdrawal is governed
by considerations identical to those that pertain to the
selection between these options for the use of
gaseous chlorine. It is virtually universal to use
solution feed processes for sulfur dioxide dechlo-
rination processes.
The types of mixing processes that may be used for
dechlorination are identical to those used for chlo-
rination. Due to the rapidity of the dechlorination
reaction, and the absence of deleterious side re-
actions, it is considered unnecessary to use separate
contacting systems for dechlorination, since the short
residence time in the mixing and outfall structures
generally suffices for the reduction of halogen
residuals to desired values.
The choice of process control options for dechlor-
ination is identical tothatforhalogenation processes,
with one exception. In wastewater treatment plants
that discharge effluents into low dilution receiving
streams, the use of manual control will result in
overdosing of dechlorination chemicals into the
wastewater. This may produce deoxygenation of the
effluent and the receiving stream, with undesirable
results. Hence, in some plants, even where simpler
control systems for chlorination are used, flow-
proportional, or feedback control for dechlorination
may be necessary to minimize the possibility for
deoxygenation in the effluent and the receiving
stream.
5.7 Design Coordination
As previously indicated, an overall halogen disin-
fection system may be subdivided into several
components. In this section, the procedures for sizing
and specifying each of these components are dis-
cussed.
5.7.1 Disinfectant Generation and Handling
In the case of chlorine or bromine chloride this
subelement includes possible on-site generation (in
the case of hypochlorites), storage, metering, and
mixing of the disinfectant for delivery to the point of
application. In the case of chlorine dioxide, the items
included are identical to the above, except on-site
generation is mandatory.
5.7.1.1 Chlorine and Hypochlorites
The first decision to be made by the designer is
whether gaseous chlorine or liquid sodium hypo-
chlorite is to be the chemical used for disinfection. For
small wastewater treatment plants, generally up to
100,000 gpd (378 mVd), proprietary solid hypochlo-
rite systems are available. However, these are
generally not used in larger plants.
The basis for deciding between chlorine gas and
sodium hypochlorite as a disinfecting chemical is
primarily one focusing upon risk of chlorine transport
versus additional costs of sodium hypochlorite. Only
in rare instances will the delivered cost of NaOCI
solutions per unit available chlorine be less than the
delivered cost of gaseous chlorine.
If sodium hypochlorite is used as the chlorinating
agent, it may be purchased in strengths of 5 to 15
percent, or it may be generated onsite. If purchased,
storage tanks for sodium hypochlorite solution are
necessary. It is recommended that the size of these
storage tanks should equal the amount of disinfectant
required for a time equal to the shipping time from the
vendor plus a 15 day emergency reserve against
strikes and transportation problems (149). At smaller
wastewater treatment plants, the reserve will be
higher due to the necessity of purchasing minimum
loads of solution to achieve reasonable chemical
costs. The concentration of available chlorine in
NaOCI solutions diminishes with time, and allowance
must be made for this in computation of reserve
requirements. A loss of 0.031 percent/d from 10
percent NaOCI solutions, and 0.075 percent/d from
15 percent NaOCI solutions have been reported (46).
57
-------
Example: A wastewater treatment plant has an
average daily flow of 0.116 mVs (2.6 MGD) and an
average chlorine dose of 5 mg/l is necessary.
Chlorine is to be supplied in the form of 10 percent
NaOCI solution. The shipping time for the vendor is 3
days. Determine the required storage capacity assum-
ing a 15-day emergency reserve.
Step 1. Hourly chlorine requirement = (0.116
mVs) (5 g/m3) (3600 s/hr) (24 hr/d) = 50 kg/d
Step 2. Daily NaOCI volume = (50 kg /d)/ (1007
kg/m3) = 0.5 mVd
Step 3. Storage volume = (3 d + 15 d) (0.5
m3/d) = 9 m3
Step 4. Correct for hypochlorite decay. (18 d) (0.031
percent/d) = 0.56 percent loss in 18 d.
(9 m3)(10 percent)/(10-0.56 percent) = 9.5 m3
Therefore a storage tank of at least 9.5 m3 (2,500 gal)
would be necessary.
Onsite Generation. For onsite generation of sodium
hypochlorite, several systems are available. All re-
quire a source of DC electricity (rectification facilities
may be included as part of the generation package). In
addition, a source of brine is necessary, and partic-
ularly if seawater or rocksalt is used, pretreatment of
brine for the removal of calcium carbonate and iron
(which precipitates in the alkaline side of the elec-
trolysis cell) may be required.
In one such system (Chloromat-TM, Ionics Inc.),
electrodes are separated by two membranes. One
membrane is a cation exchanger permitting only the
passage of cations (such as sodium), and in this
manner, chlorine gas is evolved, which is then
immediately dissolved to form hypochlorite solution.
In addition, caustic (NaOH) and hydrogen gas are
produced as by-products. The hydrogen gas must be
flared off to minimize fire hazards. A schematic of this
electrolytic system is shown in Figure 5-8.
Individual cells are manifolded together to produce
hypochlorite generation capacities of 9-150 kg/hr
(500-8,000 Ib NaOCI/d). Capital costs have been
estimatedforthe electrolytic units, exclusive of power
rectification and brine pretreatment, and fit the
following equation (150):
(5-66)
Capital Cost ($, 1974) = 10386 (kg NaOCI/hr)0'66
Replacement of the cell anode and membranes has
been estimated as a contribution to operating costs to
be $4.40-$ 13.20/1000 kg NaOCI (150). Power
requirements vary with the desired product strength
of hypochlorite and are 3.5-5.5 kWh/kg (1.6-2.5
kWh/ Ib) available chlorine, increasing with increas-
Figura 5-8.
Schematic of Chloromat™ (Ionics, Inc.) electro-
lytic hypochlorite cell (Reproduced by permis-
sion of the Water Pollution Control Federation).
Chlor
_ Spent Brine
I
I
I
I
_L
Anode
Na+
Me
Brine —
mbrane
1 1
- -+
^ Caustic
Hydrogen
l, u
~ Cathode
-».
_ _
r-
_
Diaphragm
ing strength (Ionics, Inc.); sodium chloride require-
ments are 2 kg/kg available chlorine without brine
recycle, and 1.75 kg/kg with brine recycle. Doan and
Haimes presented a case study of the application of
this system to a 24 MGD (1 mVs) wastewater
treatment plant at Amherst, NY (151). In that case,
brine pretreat ment with sodjum carbonate and
caustic (some of which may be recovered from the
electrolysis operation) to less than 10 mg/l (as
calcium carbonate) hardness and less than 1 mg/l
iron was needed since brine was produced from
rocksalt. Capital costs for the entire system (electro-
lytic apparatus, power rectification, and brine pre-
treatment) were estimated at $1,400,000 (1978
dollars).
Electrolytic hypochlorite generating equipment suit-
able for on-site use is also manufactured by Engelhard
Corporation (152) and Diamond Shamrock (61).
If onsite hypochlorite generation is used, short term
NaOCI storage (several days) may be desirable to
permit maintenance of the electrolytic units, and to
eliminate the need for continuous operation. In
wastewater treatment plants serviced by electric
utilities with off-peak billing schedules, it may also be
desirable to operate the electrolytic generation equip-
ment only during offpeak hours and to store NaOCI
solution for use during peak periods.
Gaseous Chlorine Supply. When chlorine gas sup-
plied in cylinders or containers is to be used as the
disinfecting chemical, it is important to decide upon
the necessary inventory in the same manner as
outlined for hypochlorite. In addition, the working
inventory should be increased by an amount equal to
the number of chlorine containers simultaneously
58
-------
under service. However, an inventory above that
minimum calculated may be desirable since the
shipping cost of bulk chlorine in the form of cylinders
or containers is a function of the number of cylinders
or containers shipped, and hence a larger working
inventory may enable the treatment plant to take
advantage of quantity price discounts.
Chlorine is generally supplied either in 150 Ib
cylinders, ton containers, or larger rail or truck
tankloads. In such containers, liquid chlorine exists in
a pool at the bottom of the container and is in
equilibrium with gaseous chlorine in the container
vapor space.
The chlorine may be withdrawn for use from ton
containers either as a gas or as a liquid. If it is
withdrawn as a gas, it may be directly piped to the
ejector, which produces a chlorine solution in water.
If it is withdrawn as a liquid, it is necessary to allow
the liquid chlorine to vaporize in an external evapor-
ator prior to its dispersal in water.
There is a maximum rate of safe withdrawal of
gaseous chlorine from cylinders or containers. This
limit exists because, as gaseous chlorine is with-
drawn, liquid chlorine must vaporize to maintain
equilibrium within the container. As this vaporization
occurs, heat must be withdrawn from the surround-
ings. If the rate of gas withdrawal is too high, the rate
of heat abstraction from the surroundings will not be
sufficient to prevent cooling of the chlorine to low
temperatures. At a sufficiently low temperature, the
vapor pressure of chlorine will be below the exit
pressure of chlorine gas from the cylinder, and thus
no further flow will occur. To prevent this, a maximum
gas withdrawal rate is permissible. This is dependent
upon surrounding temperature and exit gas pressure.
For a 21 °C (70°F) external temperature and 2.4 atm
exit pressure, the maximum gas withdrawal rate from
a 68 kg (150 Ib) cylinder is 0.8 kg (1.75 lb)/hr, and
from a 910 kg (1 ton) container is 6.8 kg (15 lb)/hr
(44). For liquid withdrawal, no such cooling occurs
(since the heat of vaporization is supplied by an
external evaporator), and the maximum withdrawal
rate is set primarily by the hydraulics of chlorine flow
through the fittings; it is 90 kg (200 lb)/hr and 180 kg
(400 lb)/hr, respectively, for the two types of con-
tainers.
In general, a liquid withdrawal system will be used
when the number of parallel containers necessary to
maintain the average daily chlorine flow while
obeying the maximum rate of safe withdrawal per
container becomes unwieldy. For example, if 90 kg
(200 lb)/hr of chlorine are to be supplied from ton
containers, at least 14 ton containers must be
simultaneously manifolded (90 kg/hr-e-6.8 kg/hr/
container). In addition, treatment plants using bulk
rail or truck shipments generally practice liquid
withdrawal. Benas presents a detailed case study of
bulk chlorine storage utilized at a 2.8 mVs (65 mgd)
and a 1.3 mVs (30 mgd) treatment plant in San
Francisco, CA (153).
If liquid withdrawal is practiced, the contents of a
single container are piped to an evaporator. Multiple
evaporators and containers may be used; however, it
is undesirable to manifold several cylinders or
containers each with liquid withdrawal due to the
possibility of liquid transfer between containers.
In piping systems where liquid chlorine is transported,
particular attention must be given to preventing the
possibility of heating of the pipeline. If this occurs
(such as by exposure to sunlight), the liquid flowing
full in the pipeline will expand and could hydro-
statically rupture the line. To guard against this
possibility, an emergency expansion chamber (or
chambers) equivalent to at least 20 percent of the
volume contained in the liquid chlorine pipeline must
be provided, and should be protected by a rupture disk
rated at less than the maximum bursting pressure of
the supply line. Figure 5-9 details this provision. In
addition, pipelines should be insulated to minimize
heating. Further information on the design of chlorine
pipelines may be found in Chlorine Institute Pamphlet
#60(154).
While it is desirable to minimize the length of liquid
chlorine pipe, to reduce the risk of leakage in the
event of pipeline breaks, instances exist wherein this
may be difficult. Cameron reported on the use of a
490 m (1,600 ft) liquid chlorine pipe handling a flow of
1,365 kg (3,000 lb)/d of chlorine supplied to the City
Island Plant at Atlantic City, NJ (155).
Chlorine evaporators are usually supplied as integral
units by various suppliers and consist of a pressure
vessel surrounded by a heating bath (hot water or
thermostated electrical) that serves to convert the
incoming liquid chlorine into a gaseous chlorine
product stream. Evaporators are inherently high in
energy requirements, and manufacturers' literature
should be consulted for such details. Design and
safety information for evaporators are available from
the Chlorine Institute (156). In particular, pressure
relief devices leading to gas absorption tanks and
super heating of effluent gas are necessary. It is
recommended that an absorption tank (of caustic or
lime solution) capable of neutralizing all liquid
chlorine in the supply line plus 70 kg (150 Ib)/
evaporator be provided, and also that the exit gas be
11 °C (20°F) higher in temperature than the boiling
point at the exit pressure (157). This latter require-
ment will prevent reliquification in the subsequent
gas piping as the pressure of the gas is reduced.
59
-------
Figure 5-9. Chlorine expansion chambers (reproduced from Chlorine Institute Pamphlet #60,1982 by permission).
Alternative "A"
1/4V
Product
Line
(Supports may be
Necessary but
are Not Shown)
Tee or
Welded Outlet
(As Applicable)
Item
1
2
3
4
5
6
7
8
9
10
11
Name of Part
Primary Expansion Chamber
(Note 1 )
Secondary Expansion Chamber
(Note 2)
Rupture Disc (400 psi) (Note 3)
Pressure Indicator or Alarm
Switch, 1 /2" Conn. (Note 4)
Tee (Note 5)
Reducing Tee (Note 5)
Elbow (Note 5)
Reducing Elbow (Note 5)
Union (Note 5)
Valve (Note 5)
Plug (Note 5)
Alternative "B"
3/8V
Product
Line
Notes:
1. Capacity—20% of line volume
2. Capacity—10% of line volume
3. 400 psi setting suitable for many
systems, setting must not exceed
system design pressure
4. Liquid-filled protective diaphragm
optional
5. Fittings shall be forged carbon steel,
3000 Ib cwp
For gas withdrawal systems, two major design options
exist. These have been referred to as pressure/
vacuum/pressure (PVP) and all vacuum systems
(158). In the former system, chlorine gas is withdrawn
from the supply containers into a pipe at greater than
atmospheric pressure, whereupon it flows to a
chlorinator at which point the pressure is reduced
below atmospheric until the chlorine flows to the
ejectors. In the all vacuum system, the chlorine gas is
immediately expanded to a pressure below atmos-
pheric, eliminating the initial pressure line. In a PVP
system, the initial pressure line introduces some
60
-------
potential risk of gas loss in the event of line rupture;
however, it is easier to manifold multiple cylinders
than in the all vacuum case.
In gas withdrawal systems, two major concerns are
temperature in cylinders and supply lines, and
impurities in the gas. Filters should be placed inline
as close as possible to the chlorine cylinder or
manifold to remove liquid droplets and solid impur-
ities(SO). It is desirable that the chlorine gas lines and
subsequent connections be warmer than the chlorine
cylinders or containers to prevent liquification of
chlorine in the downstream lines. In an all vacuum
system (as opposed to a PVP system) this possibility is
minimized since the gas pressure is reduced con-
siderably below the vapor pressure of saturated
chlorine gas at the existing temperature. In PVP
systems, it is also desirable to have duplicate piping
and manifold systems to facilitate cleaning.
Gas cylinders manifolded together should be at
similar temperatures. This may deserve special
consideration in wastewater treatment plants located
in areas of extreme temperature variability. Other-
wise, there will be a transfer of contents between
containers, which could lead to over-filling and
container rupture. Nevertheless, multiple ton con-
tainers (or gas cylinders) may be readily manifolded to
provide easy switchover of cylinders upon emptying.
One such configuration is indicated in Figure 5-10
(80). Pressure reducing valve 1 (PRV 1) is initially set
at a higher pressure, e.g. 375 kPa (40 psig), than PRV
2, e.g. 238 kPa (20 psig). Initially, gas will flow only
from the cylinders in header A. When the pressure in
these cylinders approaches 20 psig, flow will com-
mence from header B. The valves connecting the
individual cylinders to header A can now be shut, and
the cylinders replaced (since 20 psig is below the
vapor pressure of chlorine at usual ambient temper-
atures, relatively little chlorine will be remaining in
these cylinders). Upon replacement, valves a, b, c, and
d can now be switched such that the header
containing the "old" cylinders is connected to the
375-kPa (40-psig) pressure reducing valve.
In both liquid and gas withdrawal systems, it is
desirable to have weighing scales for each gas
cylinder under service. Since chlorine liquid and gas
are at equilibrium at a constant pressure given a fixed
temperature, pressure will only drop from a cylinder
upon exhaustion of all of the contained liquid and will
thus provide little warning to the operator of imminent
emptying. Thus, the only reliable means of inventory
control, as well as verification of chlorine consump-
tion, is weight. A variety of weighing devices exist;
however, these should permit the weighing of
cylinders or containers without lifting.
The chlorine gas from the cylinder manifold, or in the
case of a liquid withdrawal system, from the evapo-
Figure 5-10. Chlorine manifold and switchover system
(after White, 1972 reproduced by permission
of Van Nostrand Reinhold).
Manifold A
Manifold CI2 _Anr,ci
Valves Gage P.i-40psi
Chlorine
Supply
System
(Can be Any
Size Container) •*
Manifold B
Chlorine Pressure
Reducing Valves
P2 = 20 psi
rator, flows to a chlorinator. The function of this
chlorinator is to regulate the flow of chlorine gas, and
to couple with the control elements so that this
chlorine gas flow rate (and ultimately the chlorine
dosage to the wastewater) varies with the chlorine
demand.
Chlorinator Design and Hydraulics. Chlorinators are
sized at a variety of gas flow rates. Typical upper limit
gas flow rates are 100, 200, 500, 1,000, 2,000,
4,000, 6,000 and 8,000 Ib chlorine/d (1.9, 3.8, 9.4,
18.8, 37.5, 112.5 and 150 kg/hr). The lower limit to
flow is 1 /20 of this upper limit. If there are frequent
conditions necessitating a chlorine dose below
1/20th of the maximum, it may be necessary to have
two chlorinators in parallel, with a second used to
regulate gas flow under low flow conditions.
In PVP systems, the lines between gas supply and the
chlorinator should be on an uphill grade, to help
minimize liquid chlorine carryover. Typically the exit
pressure reducing valve at the gas source is set at 172
kPa (25 psia). In ton containers, it is customary to
install a blind drip leg prior to the chlorinator as a final
trap against liquid chlorine and debris passage, which
could damage chlorinator components.
The output from the chlorinator in a PVP system is a
regulated chlorine gas flow stream to the ejector
where an aqueous solution is produced (however,
often the ejector may be physically housed close to or
in the same mounting as the chlorinator). The chlorine
ejector (also variously called the eductor or injector) is
a device that conveys the chlorine gas into the
converging region of a jet of high velocity water. The
diverging water downstream of the point of gas
injection produces a vacuum, which draws the
chlorine gas into solution and serves to convey the
gas stream from the pressure reducing valve at either
the gas manifold or the evaporator. Figure 5-11
illustrates the typical configuration of ejectors with
fixed water flow rates, and with variable water flow
61
-------
rates set by a variable orifice. Additional configura-
tions are available wherein the gas flow rate may be
controlled by a variable orifice on the gas inlet pipe.
The inlet water for the ejector may be either city
water, or, more commonly, chlorinated secondary
effluent. As will be indicated below, the inlet water
supply to the ejector must be at relatively high
pressure, and therefore if secondary effluent is used,
a booster pump is required. In all but the smallest
wastewater treatment facilities it will be less expen-
sive to provide this booster pump for secondary
effluent ejector water supply rather than to rely upon
purchased city water.
Important considerations for ejector sizing include
(157):
1. The ejector water supply should be no less than
that required to produce a chlorine solution of
less than 3,500 mg/l to minimize the possibility
of chlorine gasi volatilizing from the feed solu-
tion.
2. The ejector back pressure (sum of static and
dynamic pressure losses between the outlet
side of the ejector and the mixing device) should
be at least 1.2 m (4 ft) of water (1.7 psi) to
prevent chlorine release in the solution line.
3. The ejector should be installed as close as
possible to the mixing point to minimize the time
lag in the solution line and the back pressure at
the ejector.
4. The vacuum line carrying gas between the
chlorinator and the ejector should have a
pressure drop of less than 3.8 cm (1.5 in) of
mercury (0.7 psi).
Specific information needed to size and specify the
ejector is the following (Fischer & Porter):
• maximum capacity of the chlorinator;
• inlet water supply pressure (this can be assumed,
with several values, for the purpose of determining
ejector water booster pump capacity);
• back pressure at the ejector outlet;
• elevations of the ejector and the diffuser; and
• distance between ejector and diffuser, and any
piping bends or constrictions.
Figure 6-11. Schematic of fixed and variable orifice ejectors. (Reproduced courtesy of Fischer and Porter.)
Chemical
Solution
•Diaphragm
Back-Flow
Check Valve
Gas.
•Back-Flow
Ball Check
Valve
•Either One
or Both,
as Required
Throat
Orifice
(Nozzle)
Water
or
Process
Liquid
Chemical
Solution
r
*Diaphragm
Back-Flow
Check Valve
\
•Back-Flow '
Ball Check
Valve
I /
Water or
Process — +•
Liquid __^»
1,
/
/
•y
f
)
— Throat
r Orifice
(Nozzle)
' Tapered
Plug
, ^ Hand
" " Wheel
•Either One
or Both,
as Required
Fixed Orifice Ejector
62
Variable Orifice Ejector
-------
Ejector sizing procedures will be illustrated by
example. Assume a maximum chlorine application
rate of 18.8 kg/hr(1,000 Ib/d) has been calculated. It
will also be assumed that the booster pump supplying
f eedwater to the ejector can provide a pressure of 515
kPa (60 psig)). At the point of mixing (the outlet side of
the diffuser nozzles) the pressure is 1.8 m (6 ft) of
water or 17.9 kPa (2.6 psi). The diffuser is located 1.5
m (5 ft) higher than the ejector. There are 100 ft (30.5
m) of equivalent pipe (pipe length plus length
equivalents of fittings obtained from standard hydrau-
lic handbooks) between the ejector outlet and the
diffuser. It is assumed that the headloss through the
diffuser itself is 13.8 kPa (2 psi) or 1.4 m (4.6 ft).
For one possible ejector (Fischer & Porter, Figure 5-
12), manufacturer's data are available for acceptable
combinations of inlet and back pressure and water
flow rate. Since the inlet pressure is assumed at 515
kPa (60 psig), from Figure 5-11, the acceptable back
pressure is 273 kPa (25 psig) and the acceptable
water f lowrate is 2.8 l/s (45 gpm). Now it is necessary
to verify that the system backpressure is below the
maximum acceptable (273 kPa). The backpressure is
the sum of the following elements:
9 elevation of diffuser with respect to ejector (A);
• back pressure on diffuser outlet (B);
• friction loss through diffuser (C); and
« friction loss in piping (D).
From the assumptions, A + B + C = 5ft + 6ft + 4.6 ft =
15.6 ft or 6.8 psig. In order to calculate D, the pipe
diameter needs to be specified. Assume that 1 -V4 in ID
Schedule 80 PVC pipe is used. Frictional losses may
be estimated using the graph in Figure 5-13. At a
water flow of 45 gpm in 1-1/4 in PVC Schedule 80 pipe,
the frictional loss is 37 ft/100 ft. Therefore the total
Figure 5-12. Ejector sizing curve (Reproduced by permission of Fischer and Porter).
200
180
160
140
120
100
£
-------
Figure 5-13. Frictional losses in solution piping (Reproduced courtesy of Fischer & Porter Co.).
44
30
20
10
100
50
30
_ s
if 20
8
110
3
I 5
3
2
2V2
I I I I I I
2 3
10 23 5 100
Water Flow Rate, gpm
1000 235
Rubber Hose
PVC Pipe Sch. 80
PVC Pipe Sch. 40 •
system backpressure (A+B+C+D) is 53 ft or 260 kPa
(23 psig). Since this is below the maximum acceptable
backpressure, the design is satisfactory. At this point,
a different inlet assumed pressure may be tried to
investigate the trade-offs inherent in various booster
pumps. It is also necessary to determine the transit
time in the solution piping, which is necessary, as will
be noted later, to ensure that unreasonably long
control lags are not present in the system.
The flow rate in the solution pipe is 2.8 l/s (45 gpm).
Assuming that the entire pipe is straight, the volume
of a full 100 ft pipe of 1 -14 in ID is 24.5 I (0.85 cu ft).
Therefore, the residence time in the line is 0.14
minutes (8.5 sec) which is reasonably short. A larger
diameter line, while serving to reduce back pressure,
would also have increased this system delay time. As
will be noted, the total of the lag time in the solution
line, the time for analysis, and the residence time
between mixing and sample collection for analysis in
the contactor should be under several minutes for
best performance when residual control or compound
loop control is practiced.
As a more complex example, a second manufacturer
(Wallace and Tiernan Division of Pennwalt, Figure
5-14) produces ejector specifications in which the
relationship between water flow rate and injector
pressure is more variable than in Figure 5-12. For the
above example, at an ejector inlet pressure of 515 kPa
(60 psig) and 273 kPa (25 psig) backpressure, a water
flow of 15.5 l/s (245 gpm) is required for this
particular ejector. Losses A, B and C will remain as
above (15.6 ft). If schedule 80 PVC pipe is used, from
Figure 5-12, to keep the total backpressure below 25
psig (273 kPa) (and, in other words, to keep the
frictional losses in the solution line below 231 kPa(19
psig), 3 inch (76 mm) ID pipe must be used. In this
case, the residence time in the solution line is 9 sec,
or nearly identical to the lag obtained with the
preceding ejector. In this case, however, the neces-
sary ejector water flow rate is substantially higher,
which may necessitate the use of more costly
pumping equipment and larger operational pumping
costs. It should be noted that for both chlorine
solution piping itself and the ejector, maximum
pressure limits exist above which material failure
becomes possible. Generally these pressures are in
excess of 100 psig (790 kPa); however, they decrease
at above-ambient temperature. Prudent practice
dictates that the inlet water supply pressure be
maintained belowthat specified by the ejector or pipe
manufacturer.
At this point, knowing the flow rate of chlorine
solution and the desired maximum headless through
64
-------
Figure 5-14.
300
275 -
Ejector sizing curve (Reproduced by permis-
sion, Wallace & Tiernan Division, Pennwalt
Corp.).
1000lbCI2/d
25
50 75 100
Injector Pressure, psi
125 150
Note: Extrapolations and interpolations
should not be used.
the diffuser structure, one can design the diffuser
itself. For small solution flows, a single spray nozzle
on the end of the solution pipe or hose may be
adequate. Figure 5-15 can be used to determine the
headloss for this design as a function of hose
diameter.
For the solution flow rates used in the example case,
however, the use of a diffuser with multiple perfora-
tions is more likely. Such a system consists of a pipe
lateral of diameter equal to the solution pipe itself,
with multiple outlet holes. This diffuser pipe may be
suspended horizontally or vertically in a pipe or in a
rapid mixing chamber or device. Figure 5-16 provides
a nomograph for the estimation of headlosses
through perforated pipe diffusers. In the present
example, given a solution flow of 2.8 l/s (45 gpm), a
diffuser with a headloss of less than 13.7 kPa (4.6 ft)
is desired.
A vertical line at 1.4 m (4.6 ft) headloss is constructed
to the point of intersection with the loss of head curve.
At this point of intersection, a horizontal is con-
structed to the left. Each intersection of this horizontal
with an orifice diameter represents suitable combina-
tion of chlorine solution flow and orifice diameter
necessary to achieve the design headloss. The
number of orifices needed is obtained by dividing 2.8
l/s (45 gpm) (the total solution flow) by the flow per
orifice and rounding up to the nearest whole number.
Thus, 6 !/2-in orifices would be needed; 5 9/16-in
orifices, etc.
Figures 5-15 and 5-16 could also be used to design
the diffuser system for a sodium hypochlorite feed
unit in which the flow rate of sodium hypochlorite is
known. For example, to feed 18.8 kg/hr (1,000 Ib/d)
of chlorine using a NaOCI solution containing 10 percent
available chlorine:
1000 Ib/d
100g/l
x 454 g/lb = 4,540 l/d
= 1,200 gpd
would be needed. This flow rate is equal to 0.05 l/s
(0.83 gpm). Hence, from Figure 5-15, if a Vi-in hose
spray nozzle is used, a headloss of approximately 8.1
kPa (2.7 ft) would result.
Materials Compatibility. In the design of piping and
handling systems for both chlorine and hypochlorite,
attention must be given to the corrosive nature of
both of these materials. Factors involved in the
handling of liquid chlorine have already been dis-
cussed. For dry chlorine gas, copper, iron, and steel
pipe are satisfactory (44). For hypochlorite solutions
(including the solution piping from the chlorine
ejector), rubber, ceramic, glass, Tyril, saran, PVC,
vinyl and Hypalon are all suitable materials (159).
5.7.1.2 Bromine Chloride Supply
Systems for the supply and handling of bromine
chloride are similar to those for the supply and
handling of gaseous chlorine. BrCI is supplied in
cylinders of 68 kg (150 Ib) capacity and containers of
1,365 kg (3,000 Ib) capacity. Since the vapor phase
above liquid BrCI is relatively enriched in molecular
CI2 due to gas phase dissociation, it is necessary to
withdraw BrCI from either type of container from the
liquid phase.
The vapor pressure of BrCI is less than that of gaseous
chlorine at identical temperatures. This means that,
65
-------
Figure 6-16. Headloss thru spray nozzle diff users. (Reproduced courtesy of Wallace & "Pieman Division, Pennwalt Corp.)
2 345
Flow in gpm (Under Water Discharge)
8 9 10
while similar liquid evaporators may be used for BrCI
service, it is necessary to attain a higher degree of
superheating of the BrCI product gas to prevent
reliquifaction. White recommends that 30°F super-
heat be provided to the product gas (24).
Bromochlorinators (analogous to chlorinators) and
solution ejectors may be designed on a similar basis
to those used for chlorine service. The only substantial
differences are that materials specified for chlorine
service may not necessarily be suited for BrCI service.
It is recommended that piping should be black iron or
carbon steel, but not cast iron (56). Other suitable
materials for BrCI service are tantalum, nickel-molyb-
denum alloy, lead, silver, platinum, glass ceramic and
some plastics (56). Early pilot plant studies.noted
frequent clogging of BrCI evaporators (42). This
problem persists in BrCI evaporators, due, apparently,
to substantially greater impurities existing in BrCI as
compared to Cla. As a result, at least one manufac-
turer has withdrawn its designs for BrCI evaporation
systems, and is currently testing direct liquid BrCI
feeding equipment (T. Zeh, Capital Controls Co.,
personal communication).
Similarly, design of ejectors and diffuser structures
for the resulting bromine chloride solution follows
principles noted above for gaseous chlorination
systems.
5.7.1.3 Chlorine Dioxide Supply
It is necessary to generate chlorine dioxide on a
continuous basis for use as a disinfectant. Although a
few European potable water treatment plants use the
acid-chlorite generation process (21), the most com-
mon synthesis route for disinfectant ClOa generation
is the chlorine-chlorite process.
In the chlorine-chlorite process, sodium chlorite is
supplied as either a solid powder or a concentrated
solution. A solution of chlorine gas in water is
produced by a chlorinator-ejector system, of design
similar to that used in chlorination. The chlorine-
water solution and a solution of sodium chlorite are
simultaneously fed into a reactor vessel, typically
PVC, 36-42 inches (91-107 cm) in height and 8
inches (20 cm) in diameter, packed with Raschig rings
to promote mixing (21). From Equation 5-28, one
mole of chlorine is required for two moles of sodium
66
-------
Figure 5-16. Nomograph for design of multiple perforated diffusers (Reproduced courtest of Wallace & Tiernan Division,
Pennwalt Corp.)
Head Loss thru Orifice, feet
567
10
11
12
\J l +J \J t / i
-fTeTTe IF , "is "2
1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1
12 16 20 24 28 32
Chlorine Solution Flow, gpm
36
40
44
48
chlorite—or 0.78 part CI2/part NaCI02 by weight. It
has been found, however, that for this reaction to
proceed to completion it is necessary to reduce the pH
below that provided by the typical chlorine-water
solution produced by an ejector. At 1:1 feed ratios by
weight, only 60 percent of the chlorite typically reacts
(21).
To provide greater yields, several options exist. First,
it is possible to produce chlorine-water solutions in
excess of 3,500 mg/l using pressurized injection of
gas. In this case, however, there will be an excess of
unreacted chlorine in the product solution, and the
resultant disinfectant will consist of a mixture of
chlorine and chlorine dioxide. The second option
67
-------
(160), consists of acid addition to the chlorine and
chlorite solutions; a 0.1 M HCI/M chloride addition
enabled the production of a disinfectant solution of 95
percent purity in terms of chlorine dioxide, and
achieved a 90 percent conversion of chlorite to
chlorine dioxide. A third process, developed by CIFEC
(Paris, France), involves recirculation of the chlor-
inator ejector discharge water back to the ejector inlet
to produce a strong chlorine solution (5-6 kg/m3)
(0.04-0.05 Ib/gal), typically at pH below 3.0, and in
this manner to increase the efficiency of chlorite con-
version (21,24). It has been reported that this last
option is capable of producing 95 to 99 percent pure
solutions of chlorine dioxide (24).
In any of these generation processes, the additional
equipment needed (over and above a chlorinator-
ejector system) includes the reactor vessel, metermg
pumps for the chlorite (and, if used, acid) solutions,
and ancillary piping.
In the acid-chlorite method of generation, sodium
chlorite and hydrochloric acid are used as reagents.
There appears to be substantially less operating
experience with this process in water or wastewater
treatment (21,24); however, it appears to be con-
ceptually simpler than the chlorine-chlorite process.
In one plant in Europe (21), a batch reactor is used to
mix the reagents in amounts necessary to produce a
20 percent final chlorine dioxide concentration; 15
minute reaction time has been found to be sufficient.
There are also reported to be continuous generation
processes available for use with the acid-chlorite
method (24).
Due to the instability of strong chlorine dioxide
solutions, it is not practical to store disinfectant
solution for any significant period of time. Hence, the
rate at which the chlorine dioxide is generated must
be capable of being coupled in some manner directly
to process sensors (a. g., flow meters or residual
analyzers). This is generally done by varying reagent
flow rates.
In addition, with both of the generation processes, a
chlorine dioxide solution is produced directly, rather
than a chlorine dioxide gas stream. Since chlorine
dioxide gas at high concentrations may produce an
explosive mixture, its handling as a solution is
preferable. This chlorine dioxide solution may be
pumped directly to a diffuser or may be diluted in a
solution ejector (in which the strong solution is sent
to a vacuum ejector similar to that used for gaseous
chlorine application for the purpose of dilution) and
then piped to a diffuser.
5.7.2 Mixing Systems
The necessity for rapid mixing of the disinfectant
solution exiting the diffuser ports with the bulk
wastewaterflow was graphically shown by Heukelek-
ian and Day (126). In their studies, the vertical
gradient in distribution of chlorine residual in a
contactor downstream of a simple dropped hose
diffuser was measured and was found to be sub-
stantial unless adequate turbulence was present at
the point of application of the chlorine solution.
Unless mixing is present, there will exist zones of
high and low disinfectant concentration, and the
surviving microorganisms present in the zones of low
concentration will act to diminish the process effi-
ciency.
Longley determined that the degree of mixing at the
point of disinfectant application has a pronounced
effect upon the initial rate of inactivation of a variety
of microorganisms over the range of G (root mean
square velocity gradient) values of 100-105s~1 (161).
White recommends rapid mixing with a G value of
500-1,000 s"1 and a residence time of 5-15 seconds
(24).
5.7.2.1 In-Line Diffuser
Four major options for mixing exist. In the first case,
the diffuser may be placed in the center of a pipe
running full at a turbulent Reynolds number. For this
system, the mixing intensity may be computed from
hydraulic relationships as follows:
Computation of G value in pipes:
The metric form of Manning's equation may be
written as:
S = (nv/R2/3)2
(5-67)
where S is the hydraulic gradient (dimensionless), n
is Manning's coefficient, v is the flow velocity in m/s,
and R is the inside pipe diameter assuming a full-
flowing pipe (or hydraulic radius) in m.
Camp and Stein's relationshipfor the G value is given
as:
G = (P/uV)1/2
(5-68)
where P is the power dissipation (Watts), V is the
volume over which such dissipation occurs and u is
the viscosity.
For a closed conduit in which power dissipation
occurs by internal fluid friction, the power dissipation
is given in terms of the headless (hi) by:
P = SwghiQ
(5-69)
where Sw is the density of water, g is the gravitational
acceleration, and Q is the volumetric flow rate. By
multiplying S obtained from Equation 5-67 by the
length of pipe (L), the headless (he) may be obtained.
The power dissipation then obtained from Equation
5-69 is substituted into Equation 5-68, and the
68
-------
volume is obtained via the geometric relationship for
a volume of a cylinder. The resulting relationship for
the G value in a pipe flowing full is given by:
G = (SwgSv/u)1/2
(5-70)
Finally, to ensure turbulent flow, it is necessary to
check the Reynolds number from:
Re = DvSw/u
(5-71)
If Re is in excess of 2,000, turbulent flow may be
assumed.
Using these relationships, given the design flow, the
velocity may be computed for various pipe diameters.
If Re (given by Equation 5-71) indicates turbulent
flow, and if the G value, given by Equations 5-67 and
5-68 is in the acceptable range, then the design is
satisfactory. The headless may be computed directly
from Equation 5-67.
Figure 5-17. Details of a submerged weir mixing structure.
Flow
CI2 Diffuser
5.7.2.2 Submerged Hydraulic Structures
The second mixing option involves the use of a
hydraulic structure at which turbulence is induced, as
the point of application of chlorine solution. Two such
possible structures are the submerged weir and the
hydraulic jump (24).
A submerged weir consists of an open channel in
which a rectangular projection arises from the
channel base in such a manner that both the free
liquid surfaces above and downstream from the weir
are at a greater elevation than the weir itself. The
diff user inlet should be located at a point of maximum
turbulence downstream from this structure. Figure
5-17 details the design of a submerged weir mixing
structure.
The major variables necessary to design the sub-
merged weir mixing structure are the distance
between the weir and the point of maximum turbu-
lence, at which the diff user should be located, and the
anticipated headloss across the weir.
The hydraulic jump following a chute and a sharp
change in channel slope may also be used as a mixing
device. Figure 5-18 details the general characteristics
of a hydraulic jump used as a mixing structure. To
design such a structure it is necessary to*ensure that
shooting flow (i.e. flow at a sub-critical depth) occurs
at the bottom end of the chute, and that a hydraulic
jump is located downstream of the diff user but within
the downstream channel.
Figure 5-18. Details of a hydraulic jump mixing structure.
^- L'---f-M
b = Width
Cla
Diffuser
So = tan a
Sa = tan/8
The various computations for designing a hydraulic
jump mixing device are outlined below. For more
detailed discussions, standard hydraulic texts should
be consulted (162).
Step 1—Decide upon the entering jump Froude
number.
The Froude Number (F) is defined by Equation 5-72,
where g is the gravitational acceleration (9.8 m/s2), Q
is the volumetric flow rate, y4 is the liquid depth and b
is the channel width.
F = (Q/by4)/(gy4)1/2
(5-72)
F must be greater than 1 for a jump to occur
(supercritical flow), and Chow indicates best per-
formance in terms of stability and insensitivity to
downstream conditions for F between 4.5 and 9.0
(162).
To illustrate these calculations, a hydraulic jump will
be designed to serve as a mixing device for a
wastewater of average flow 26.5 MGD (1.16 m3/s).
For a Froude number of 6.0, and assuming a channel
width (w) of 1 m, the value of y4from Equation 5-72 is
0.156 m. From the depth of the water immediately
upstream of the jump, the final depth immediately
downstream (y4) may be determined by:
2v1/2v
(yi/y4) = 0.5 (-1 + (1 + 8FT/Z)
Step 2—Determine chute slope (S0) and y3.
(5-73)
For the hydraulic jump to form downstream of the
termination of the chute, it is necessary that the depth
at the chute outlet, ya, be some value less than the
depth immediately prior to the hydraulic jump (y4).
69
-------
One can assume a chute slope and calculate y from
Manning's equation (Equation 5-67) or vice-versa.
Both the hydraulic radius and the flow velocity are
functions of ya.
Assuming a value of y3 of 0.12 m (0.4ft), the hydraulic
radius equals 0.0968 m (wy3/(w + 2y3j), and the flow
velocity equals 9.67 m/s (Q/wy3). From Equation 5-
67, assuming a Manning's coefficient of 0.015 (typical
of finished concrete), the frictional headless (or
required channel slope for uniform flow at the
indicated depth) is 0.473 (i.e., the channel must have
a rise of 0.473 m/m horizontal). Hence, Sc must be
0.473, and thus the angle of inclination is 28°.
Step 3—Determine length until jump (Lj).
The water depth increases from the base of the chute
(ya) until the onset of the jump (y4) due to frictional
losses that reduce the energy (and since the flow at
these points is subcritical, increase the depth). To
determine the length of travel between the chute exit
and the jump onset, numerical integration of the
energy conservation equation in conjunction with
Manning's equation must be performed.
This process will be illustrated in tabular form. The
basic relationships are Manning's equation (Equation
5-67) and the following rearrangement of Bernoulli's
equation accounting for changes in channel elevation
and frictional losses:
-12
= (da - d,) + (vj - Vf)
2g(hei2 - 812)
(5-74)
In Equation 5-74,1_12 is the channel length between
two points (1 and 2) having indicated water depths (d)
and velocities (v), arid between which there is a
channel floor slope of heia (dimensionless rise/run)
and a frictional headloss of 812 (dimensionless).
The computation of L\ is executed in Table 5-13. First,
the range of depths between y3 and y4 is subdivided
into a number of intervals arbitrarily (however, the
number of intervals will influencethe precision of the
calculation) as indicated in column (1). In columns (2)
and (3), the hydraulic radius and the velocity, respec-
tively, associated with each depth (at the known flow
of 1.16 m3/s and width of 1.0 m) are computed.
Column (4) represents the frictional slope (S) as-
sociated with each depth, and is obtained by the use
of Equation 5-67 at each point. Column (5) represents
the arithmetic average of the frictional slope in the
range defined by an upstream and a downstream
depth. Finally, column (6) tabulates the calculated
length incrementfrom Equation 5-73 associated with
the change in depth—a channel bottom slope of 0.01
has been assumed; however, this has only a very
minor effect upon the result. By addition of each of the
length increments, a total length from the chute
bottom to the initiation of the jump (Lj) of 6.12 m is
obtained.
Table 5-13. Computation of Length to Jump (Lj). (Assumed
HE = 0.01)
(1)
d(m)
0.12
0.13
0.14
0.15
(2)
RH(m)
0.0968
0.1032
0.1094
0.2154
(3)
v (m/s)
9.67
8.92
8.29
7.73
(4)
S
0.473
0.369
0.296
0.239
Total LJ
(5)
Savg
0.421
0.333
0.267
0.226
(6)
L12 (m)
1.71
1.68
1.74
0.99
B7T2"
Therefore, in the example case, the headloss dis-
sipated in the jump is 1.68 m. By application of
Equation 5-69, the power dissipation equals 19.1 kW.
It is not readily feasible to estimate the G value, since
the volume of the liquid contained within the jump is
not easy to obtain.
Step 4—Determine length of jump (LH).
The ratio of jump length to height after the jump
(LH/y) has been observed to be a function of the
Froude number alone. For Froude numbers between
4.5 and 9, this ratio is relatively constant at 6.15
(162). In the example problem, since y has been
determined to be 1.25 m, the value for LH is 7.69 m.
Step 5—Determine minimum chute length Lc.
A sufficient length of chute is needed so that the
water velocity attained at the base of the chute is its
normal value as predicted by Manning's equation. At
the onset of the chute, the depth of flow may be
assumed to be equal to the critical depth (yc), predicted
from:
yc = (Q2/gw2)1/2
(5-75)
Thus, for the example under discussion, the initial
depth at the top of the chute, or yc, equals 0.52 m.
It is necessary that the length of this chute be
sufficient so that the depth changes from 0.52 m to
the design value of 0.12 m at the chute base. The
slope of the chute bottom (heia) is 0.473. Equations
5-68 and 5-73 may be used in the manner in which
they were applied in Table 5-13 to estimate the
minimum length of this transition. By this procedure,
it can be calculated that the length of the required
chute is 6.99 m.
Step 6—Compute energy dissipation in jump.
The amount of energy transferred to the fluid in the
hydraulic jump, and thus available for mixing, maybe
given by the difference in the energy of the fluid
immediately prior to and following the jump. Chow
gives the headloss in the jump as (162):
hi =
(5-76)
70
-------
Step 7—Repeat computations at low and high flows.
Before the above design can be accepted, it is
necessary to determine that an acceptable hydraulic
jump exists in the given physical system under
anticipated low and high flow conditions, and fur-
thermore that the location of the diff user is upstream
of the jump in all likely flow situations. By using the
design specifications (chute slope and length, chan-
nel width, downstream slope and length) with the
expected low and high flows, the values for position of
the jump under all circumstances can be calculated
by application of the foregoing equations.
Once these computations are completed, the hydrau-
lic profile of the design may be sketched, and the
static headloss computed. A scale diagram of the
jump illustrating diffuser location is given in Figure
5-19.
5.7.2.3 Mechanical Mixer
The third mixing option involves the use of a
mechanical mixer, such as a propeller or turbine
mixer, in conjunction with a small residence time
mixing chamber containing the disinfectant diffuser.
If a G value is specified (for example, 500-1,000/sec
as recommended by White (24)), then Equation 5-68
can be used to determine the mixing power to be
imparted to the fluid per unit volume.
For example, for the 26.5 mgd (1.16 mVs) waste-
water flow described above, if a G value of 750 s~1 is
specified, and the viscosity is assumed to be 0.001
kg/m-s, then from Equation 5-68, a power per unit
volume of 866 W/m3 is required. If the rapid mixing
chamber has a residence time of 10 seconds, and
hence a volume of 11.6 m, then the required mixing
power to be imparted to the fluid equals 1,385 W. To
determine the actual electric energy consumption for
the drive motor for the necessary mixing device, it is
necessary to consult manufacturer's specifications
as to the efficiency of conversion of electrical to
mechanical mixing energy. Similarly, differing geo-
metries for mixer design exist, and the necessary
motor speed will be a function of the specific propeller
or turbine utilized.
5.7.2.4 Jet Mixer
One additional mixing configuration exists in which
the chlorine solution (or in some cases, the gaseous
chlorine) coming from the ejector is introduced into a
large jet, of similar overall design to a chlorine ejector.
This jet mixer carries a substantial volumetricfraction
of the influent wastewater through a high velocity
nozzle and the disinfectant solution (or the disin-
fectant gas stream) is injected into the vacuum
produced.
This jet mixing system is of proprietary design;
however, it has been claimed that a G value in a jet
tube of 10 s~1 or more, and a dimensionless tube Gt
product of 1.5-15 produces satisfactory mixing (163).
The jet mixer may be mounted directly in the entrance
region of the contact chamber, and a booster pump
used to provide the motive power. In one such field
test, it was shown that the required dosage of
chlorine was less with jet mixing than with poor
mixing(164); however more recent tests suggest that
such reductions in chlorine dose, and in any neces-
sary dose of sulfur dioxide required for dechlorination
may be insufficient to compensate for the additional
energy costs of jet mixing (165).
Of the mixing systems, the most common is the use of
an in-line diffuser, followed by a rapid mixer, then
hydraulic jumps and finally jet mixers. In-line dif-
fusers are particularly applicable to small plants.
However at flows greater than 10 MGD it becomes
difficult to design a fully flowing turbulent section of
pipe. Both in-line diff users and hydraulic jumps share
the advantage of-requiring no direct external power
input. However, hydraulic jumps are inapplicable
unless upstream equalization of flows is practiced,
since the horizontal location of the jump is extremely
sensitive to flow variations.
Figure 5-19. Scale diagram of jump as designed.
Scale: I 1 = 1 meter
Static
Headloss
~2.4m
Chute
Increasing
Depth
Ja Development
of Jump
1.75m Below
Base of
Chute Inlet
71
-------
5.7.5 Contacting Systems
The role of the chlorine contact system is to permit
sufficient time to elapse for adequate disinfection to
occur. A characteristic measurement of time is the
mean hydraulic residence time, 6. However, it is also
necessary to ensure that the bulk of the wastewater
has had the opportunity to remain in the contact basin
prior to release. Figure 5-20 depicts the frequency
distribution of fluid residence times in three basins of
identical 6. The spread of residence times can be
characterized by the dimensionless dispersion index,
d, which decreases as the range of individual
residence times decreases. See Section 4.2 for a
theoretical development of residence time distribu-
tion functions.
Figure 5-20.
Residence time distribution functions for
contact basins.
1.0
§
oc-S
o
I'
8= Mean Residence Time
Time = V/Q
< da < d3
1.0
t/O
It has been shown that the performance of a chlorine
contact chamber declines substantially (i.e., the
effluent microbial concentrations increase) when the
dispersion in the contactor increases (166), and this
finding has been verified experimentally (167-169).
There appears to be a point of diminishing return, and
Trussell and Chao (166) propose that d = 0.01 is a
practical lower limit to be attained to optimize the
hydraulic performance of chlorine contact chambers
in wastewater disinfection.
Once the average residence time for the contactor
has been determined, from the kinetics of microbial
inactivation discussed above, several options exist for
the configuration of the contactor. When disinfection
is the final treatment process (in other words, when
no dechlorination is to be practiced), and a sufficiently
long distance is available, the outfall pipe itself may
be used to provide contact. In other cases, it is
necessary to design a separate tankfor contacting. In
either case, the volume to be contained in the
contactor may be determined from the equation V =
Q.8, where Q is the design flow, 8 is the design contact
time, and V is the volume of liquid to be contained in
the contactor.
For pipe contactors, diameter influences the headloss
in the pipe, which may be determined via Manning's
equation, the flow velocity in the pipe, and the
dispersion. The velocity in the pipe must be kept
above the scour velocity for particles to avoid sedi-
mentation. Generally a 1 ft/s(0.3 m/s) flow velocity
will suffice, or the designer may use the scour
relationship of Camp to estimate the minimum
velocity. The dispersion in a pipe may be estimated by
the use of the following equation (166):
d = 89,500 f3'" (D/L)
tO.859
(5-77)
In the above equation, f equals the Darcy Weisbach
friction factor for pipe flow, which is obtained as a
function of the pipe Reynolds number and the relative
roughness from a Moody diagram (170), D is the pipe
diameter, and L is the length of the pipe. This equation
is valid for the case of a straight pipe alone, Without
bends or other flow disturbances.
If it is necessary to'build a separate contact basin, it is
beneficial to construct a basin, either rectangular or
annular, in such a manner that baffling is present to
provide the longest possible pathway for flow in order
to minimize the dispersion. Figure 5-21 illustrates
various possible configurations for baffling in contact
chambers.
In a baffled contact chamber, the flow velocity
(defined by the length of the flow path divided by the
mean residence time) should be greater than the
scour velocity to minimize particle deposition. The
dispersion in rectangular baffled contact chambers
(either end-around or over and under baffles) may be
given by either Equation 5-78 (166) or Equation 5-79
(157):
d = 0.14 K/(L/W)
d = 1.15(L/W)"1'13
(5-78)
(5-79)
In Equation 5-78, K is a coefficient of nonideality,
found to vary between 2.3 and 15.8 for real con-
tactors. In both cases, L is the length of the flow path
and W is the width between parallel baffles.
The headloss in baffled contactors is due to frictional
losses with the bordering surfaces and dissipative
losses associated with changes in flow direction.
Frictional losses may be evaluated as in the case of
72
-------
Figure 5-21. Types of baffled contact chambers (174).
Figure 5-22. Varied serpentine contactor design of Louie
and Fohrman (172).
Baffles
Cross Baffled Tank
(May be Horizontal
or Vertical)
Longitudinal
Baffled Tank
Annular Ring Surrounding
Secondary Clarifier
any open channel flow system. Losses due to velocity
changes may be estimated using the following (171):
hi = ((n + Dv,2 + n v22)/2g
(5-80)
In this equation, v-\ and v2 are, respectively, the flow
velocities between parallel baffles and through the
baffle slots, and n is the number of baffles. It is
possible to reduce the headloss associated with these
velocity changes by adding flow redirectors. Louie
and Fohrman evaluated a number of such configura-
tions, and concluded that the vaned serpentine design
in Figure 5-22 provided minimum headloss and
satisfactory reduction in dispersion (172).
It has also been suggested that the performance of
baffled rectangular contact chambers may be im-
proved by the use of air scour. This has been
demonstrated in scale models (173); however, little
full-scale quantitative data exist to evaluate this
effect. In addition, such scouring would increase
dispersion, possibly to a deleterious degree.
The effluent weir from a contact chamber should
extend along the full length of the exit channel from
the flow channel. Marske and Boyle (174) found that
Effluent
Flow
Redirecting
Vanes
Influent
basins with full sharpcrested weirs performed better
than partial length Cipolleti (i.e. trapezoidal) weirs.
Although there has been little direct work in this area,
to minimize the potential for short-circuiting asso-
ciated with entrance effects, the inlet to contact
chambers should be designed to introduce flow over
as large a fraction of the width as possible, and with
minimum velocity. Alternatively, a single pipe inlet
impinging upon a momentum-absorbing baffle may
be used.
Contact chambers, like many wastewater treatment
unit processes, will occasionally foul with a microbial
slime. Therefore, at least two parallel contact cham-
bers should be used so that one can be drained and
cleaned. The contact chamber should have provision
for draining, preferably by gravity, and be in proximity
to a high pressure water hose for cleaning.
The required size of a contact chamber is a function of
the chlorine dose to be used, the nature of the
wastewater to be disinfected, and the required
amount of disinfection to be achieved. For a first
approximation, manufacturer's data may be consulted
for the doses of chlorine to be used in typical
situations. As a more exact estimate of the effect of
contact time on required chlorine dose, information
presented earlier on the kinetics of microbial inacti-
vation and the decay of chlorine residual may be used.
This will be illustrated in the design example to be
presented below.
73
-------
5.7.4 Process Control
The function of the process control element is to
ensure that there is consistent disinfection perform-
ance despite fluctuations in the quality and quantity
of wastewater influent to the disinfection process.
Due to the scarcity of experience with chlorine
dioxide and bromine chloride as wastewater disin-
fectants, only process control of chlorination will be
discussed.
The simplest control strategy involves eitHer the use
of flow equalization basins prior to chlorination or
extremely conservative design factors. If a sufficiently
large equalization basin exists, e.g., a lagoon with
multiple day residence time, upstream of disinfection,
the wastewater quality and quantity entering the
chlorination process will be relatively uniform. Thus,
the performance of the chlorination process will be
reasonably constant. However, very few treatment
plants are able to provide this degree of equalization.
Alternatively, the dose of chlorine or hypochlorite
applied to the wastewater may be made sufficiently
high to ensure that the high flows with the greatest
chlorine demands are adequately disinfected. This
will ensure that low flows, or flows of lesser chlorine
demand, will be at least adequately treated.
For high wastewater flows, it will be impractical to
provide equalization or uneconomical to overdose
with chlorine. In thts case, the amount of chlorine
must be varied in some manner so that the dose
applied is just sufficient to provide the required
degree of disinfection. This maybe done manually, by
the operator, one or more times per day. The feed rate
(kg chlorine per hr) from the gas chlorinator (or the
metering pump rate from the hypochlorite storage
tank) is adjusted in accordance with the flow and the
chlorine demand, usually to attain a target chlorine
residual. Typically, this adjustment may be made
once per shift.
For larger treatment plants, the savings on chlorine
utilization will justify some degree of automatic
control. In addition, for treatment plants using
dechlorination, automatic control may further be
justified by savings on dechlorination chemicals.
The simplest form of automatic control that can be
used is flow proportional control. In this case, the
instantaneous chlorine application rate is maintained
directly proportional to the instantaneous flow rate
coming into the disinfection process, resulting in a
constant chlorine dosage. If the disinfection influent
flow is controlled by a pumping station upstream, the
chlorine application rate may be set by a signal from
the wastewater pumps to the chlorinator. This
modification is referred to as additive rate control.
Generally the signal from the pumps is a standard
4-20 ma DC signal (the typical electrical control
signal) and may be used to vary the orifice opening on
a gas chlorinator vacuum line (with the flow being
linear in orifice area) or the metering purnp speed on a
hypochlorinator to provide a relatively constant
chlorine dose. One drawback of this strategy is the
slight nonlinear behavior of flow rate versus pump
horsepower due to intervening friction losses in the
line between the pumps and the chlorine contact
chamber. In addition, there will be a lag time in
process response due to the hydraulic residence time
in the system between the pump station and the point
of chlorine application. During this lag, there will be
under or over application of chlorine as flow de-
creases or increases, respectively.
A second form of flow proportional control uses a flow
meter (e.g., magnetic in-line, or venturi) to sense the
flow entering the chlorine contact chamber. Due to
the problem of lags, the plant raw wastewater
flowmeter should not be used to supply this control
signal. The output of the flowmeter may be a 4-20 ma
DC signal, or a vacuum or pressure pneumatic signal.
This signal is then applied to either the orifice on the
gas chlorinator or the metering pump on a hypo-
chlorinator to provide a constant dose.
While flow proportional control of either type ac-
counts for a major source of variation in chlorine dose
requirement, it fails to provide adjustment in chlorine
dose with variation in wastewater chlorine demand.
For example, if the chlorine demand varies by a factor
of 2 over the course of a day it will be necessary to
overdose during times of low chlorine demand, even
if flow proportional control is used.
To circumvent this difficulty, it is, in principle, possible
to measure the chlorine residual after the initial high
chlorine demand has been exerted (2 to 5 minutes
following rapid mix) and use an automatic measure-
ment of this residual to control the chlorine applica-
tion rate. This strategy, feedback residual control, is
generally not practiced alone, except when flow
variations are negligible, due to problems of hydraulic
lags in the system.
Instead the flow is most commonly measured using
an automaticflow transducer, and the residual after 2
to 5 minutes is measured using an automatic chlorine
residual analyzer. These signals are then combined to
provide a constant residual. In this manner, the
variation in flow is used as the major signal for dose
adjustment, and the variation in residual is used to
fine tune the dose. This option is designated as
compound loop control.
Two major types of compound loop control exist. One
type, typified by many gas chlorinators such as those
manufactured by Wallace and Tiernan Division of
PennWalt Corporation, uses the flow signal to vary
the size of the gas orifice, and the residual signal
(modified by a square root module) to vary the
74
-------
pressure differential across the orifice. The square
root module takes the incoming signal and converts it
to its square root—this is done because the flow rate
is linear in the square root of pressure drop, and not in
pressure drop itself. The second type of compound
loop control, implemented in many chlorinators
manufactured by Fischer and Porter and Capital
Controls, electronically combines the two signals
from the flow meter and the analyzer into a single
signal to a chlorine valve, which may regulate either
the orifice area or the pressure differential across a
fixed orifice.
In either case of compound loop control, either or both
signals from the flow and residual sensors may be
electric (4-20 ma DC), pressure pneumatic (3-15 psig
is a common industry standard) or vacuum (5-55
inches of water is common). If pneumatic or vacuum
control signals are used, a separate system is needed
to produce the vacuum or pressure that drives the
control signal lines. In addition, direct digital com-
puter control, using either dedicated distributed or
centralized computers, is possible.
Critical elements in the design of compound loop
control are the time lags inherent in the residual
measurement system, and delay times between
application of the chlorine residual signal to the
control device and the attainment of,the control
action. These time lags include those due to the
sample transit time inthe sampling lines, the lagtime
between the ejector and the point of application, and
the residence time in the contactor prior to sampling.
The calculation of the lag time in the line between the
ejector and the point of application may be calculated
as described above once the ejector is sized. The lag
time in analysis is of the order of 1-2 minutes. The lag
time inthe sampling linesissetbythevelocityandthe
length of the line. This velocity should be maintained
above 3 m/s (10 ft/sec) to minimize fouling (157). By
maintaining the analyzer as close as possible to the
chlorine contact basin, this lag may be minimized.
The lag time between the point of application and the
sampling line should be set so that the rapid initial
decay has been completed. This usually occurs 3 to 5
minutes following chlorine application (Figure 5-23).
Due to the existence of lags in the chlorine residual
sampling lines, it may be necessary to incorporate a
delay timer on the residual control loop. This insures
that no additional adjustment in chlorine dose occurs
following a preceding adjustment until sufficient time
has elapsed for the wastewater receiving the new
chlorine dose to be analyzed for its response. This
timer should be adjustable and is ordinarily "tuned"
during start up of a treatment plant to minimize
control instability.
Roop presented a comprehensive review of waste-
water chlorine control using the compound loop
Figure 5-23. Dissipation of chlorine residual and point of
sampling for control.
C/C,
Most Rapid Decay
Completed in C0 1 -5 min.
Control Sample Should
be Withdrawn Here.
system (175). For situations in which the rapid phase
of chlorine demand is prolonged (5 to 15 minutes), a
second chlorine analyzer is recommended at the far
point to trim the signal from the first sampling
location.
The analyzers used for residual may use either the
amperometric, or the colorimetric or titrimetric DPD
methods, as described above. In either case, manu-
facturers' designs of continuous analyzers incorpo-
rate in-line filters to minimize solids deposition. For
continuous analyzers, special reagent solutions are
used that differ somewhat from manual analytic
techniques in that proprietary additions are incorpo-
rated that are said to reduce fouling inherent in
continuous systems. The maintenance requirements
for continuous analyzers are higher than manual
analyzers, and at least daily standardization is
required to prevent drifting in the automated analyzer.
5.7.5 Dechlorination
From Equations 5-48 and 5-49, one mole of sulfur
dioxide can dechlorinate one mole of either free
chlorine or combined chlorine. On a mass basis,
therefore, 64/71 grams of SOa are required to
dechlorinate one gram of chlorine, either in the free
or the combined form. On a practical basis, about 1
gram of SOa is required per gram of chlorine.
Sulfur dioxide used for dechlorination may be sup-
plied in either 68 kg (150 Ib) cylinders or 1,365 kg
(3,000 Ib) containers. Railroad tank cars for bulk
sulfur dioxide shipment are also available. The
discharge from sulfur dioxide containers may be
taken from the liquid or gaseous phases (106).
For gaseous withdrawal, the maximum safe rate of
withdrawal on a continuous basis at 21 °C (70°F) is
0.9 kg (2 lb)/hr for a 68-kg (150-lb) container, or 11.4
75
-------
kg (25 lb)/hr for a "ton" container (106). To attain
higher rates of gaseous withdrawal, it is acceptable to
immerse the containers or cylinders in a liquid bath,
or to surround the containers with strip heaters; in
both cases, only the lower 50% of the container
should be subject to heating, and heating should be
limited to 325°K (106). As in the case of chlorine, the
pressure at which sulfur dioxide is withdrawn should
be sufficiently below the saturation vapor pressure to
provide superheating and prevent reliqu if action (157).
For liquid withdrawal, the maximum rate of removal
is limited by hydraulic considerations to 135 kg (300
fb)/hr) from either type of container (106), although it
is possible to exceed this by pressurization, or
"padding" with dry air or nitrogen up to 515 kPa (60
psig) (157). With liquid withdrawal systems, external
evaporators, of similar overall design to those used
for chlorine service, are required.
It may be necessary to manifold multiple cylinders of
sulfur dioxide for supply to larger dechlorination
systems. Considerations for manifold design are
similar to those used for gaseous or liquid chlorine
systems.
Piping and materials used for chlorine service are
generally satisfactory for sulfur dioxide service (157),
although the physical systems used for chlorine
should not be used for sulfur dioxide, or vice versa,
prior to thorough cleaning, to prevent potentially
explosive reactions from occurring. In liquid piping
systems, expansion chambers are required, as for
liquid chlorine service, to prevent hydrostatic expan-
sion rupture from occurring (106).
The gaseous sulfur dioxide exiting from the cylinder
or the evaporator miay be dispersed into water using
vacuum ejector systems of similar design to chlorine
ejectors. Design curves for chlorine ejectors must be
corrected for different flow rates (on a mass basis) for
sulfur dioxide versus chlorine. This is done by
multiplying the specified flow rate for a chlorinator by
0.95 (which represents the ratio of the density of
sulfur dioxide to that of chlorine). For example, if a gas
chlorinator is designed to deliver 45 kg (1,000 lb)/d
chlorine at maximum span, it can deliver 43 kg (950
lb)/d of sulfur dioxide.
The mixing at the point of application of sulfur dioxide
to the full wastewater flow should be conducted in an
intensive mannerto allow complete dechlorination of
all fluid elements. Injection into the center of a
turbulent fully-flowing pipe has proven satisfactory,
with complete mixing being attained 10 pipe diam-
eters downstream of the injection point (176).
The kinetics of the reaction between sulfur dioxide
and chlorine are sufficiently rapid to preclude the
necessity for a separate contact chamber prior to
discharge of the effluents (109). If a contact chamber
is used, the absence of a chlorine residual may allow
for the growth of a microbial slime layer containing
coliforms, and possibly other organisms. While
prevention of this slime growth does not appear to be
feasible, some regulatory authorities (e.g., the State
of California) permit the coliform standard to be
attained at the chlorine contactor effluent, despite an
increase in bacterial numbers during dechlorination
(112).
Control of sulfur dioxide dechlorination is generally
more difficult than chlorination. According to Chen
and Gan: "simple feedforward sulfur dioxide feed
control system(s) [i.e., systems which rely solely on a
measurement of influent volumetric flow rate] (are)
inadequate for most dechlorination installations"
(112). This is due to the variation in the chlorine
residual at the outlet from the chlorine contact
chamber.
Most commonly, control systems for dechlorination
consist of either straight flow proportional or feed-
forward control based on the product of influent flow
and chlorine residual. In dechlorination installations,
it is necessary to maintain two automatic chlorine
residual analyzers—one 3 to 5 minutes from the point
of application of chlorine used for chlorine dose
control, and the second located at the contact
chamber outlet, used for control of the sulfonator
(112). Feedback control using measured chlorine
residual following sulfur dioxide addition is not
currently feasible due to the instability of currently
available analyzers measuring low (below 0.1 mg/l)
or zero chlorine residuals.
Two types of feedback control systems may be used to
circumvent the problems associated with continuous
measurement of low chlorine residuals. In the first
method, which may be described as split dechlorina-
tion, a flow proportional feedback chlorine residual
system (i.e. compound loop control) is used to produce
an effluent that is dechlorinated to a low residual, but
well within the range of commercially available
chlorine analyzers—this is usually 90 percent de-
chlorination of the incoming residual. A subsequent
dechlorination step, using simple feedforward con-
trol, may then be used to produce a chlorine-free
effluent.
The second process uses a chlorine analyzer that is
biased with the addition of a constant flow of a side
stream containing a constant chlorine concentration
(produced by a chlorinator or hypochlorite feed system
separate from that used for chlorination) to the
sample being analyzed. The setpoint on this analyzer
is controlled to maintain the chlorine residual sub-
sequent to dechlorination at the concentration im-
posed by the biasing stream. In this case, classical
compound loop systems may be used (112).
76
-------
One modification of the biased feedback control
strategy has recently been devised (177). In this
approach, the chlorine residual immediately prior to
dechlorination is measured (Cpre). Equal flows of the
influent to and effluent from the dechlorination
process are blended and the residual measured using
a second analyzer (Cmixed). The chlorine residual
leaving the dechlorination basin may then be calcu-
lated as 2Cmixed-Cp,e, and this value used as a control
signal.
This system has been in use in Metro Seattle and is
reportedly successful at maintaining a residual below
0.067 mg/l. The principal disadvantage of this
approach is the relatively higher complexity of the
control system, and the potential for biological slime
accumulation in the dechlorinated effluent sample
lines; however, this system eliminates the need for a
separate chlorinator used to bias the sulfonator
analyzer.
In the absence of such control systems, it will be
necessary to overdose the wastewater with sulfur
dioxide, which may result in deoxygenation of the
effluent. In this case, reaeration of the dechlorinated
wastewater may be necessary to produce compliance
with an effluent DO requirement. In the presence of
one of the suggested dechlorination control systems,
such deoxygenation and reaeration requirements
have been found not to occur (112).
5.7.6 Design Example
The principles of design of chlorination and dechlor-
ination systems will be illustrated by reference to the
common design example. A chlorination system is to
be designed to treat a non-nitrified activated sludge
effluent. The design flows are 7.5 mgd (mean) and 15
mgd (peak); initial flows are 3.5 mgd (mean) and 7.5
mgd (peak). The influent to the disinfection system is
expected to contain a geometric mean of 500,000
fecal coliforms/100 ml, with a peak of 2,000,0007
100 ml. The effluent coliform standards are a 30-day
geometric mean of 200 fecal coliforms/100 ml, and a
7-day geometric mean of 400/100 ml. Other effluent
parameters are a 30-day mean for both BODs and
suspended solids of 15 mg/l and a maximum daily
value of 30 mg/l for each of these parameters. It is
also desired to dechlorinate such that the final
effluent chlorine residual is less than 0.05 mg/l.
Step 1—Determine the average desired fractional
survival.
From the conditions of this problem, the desired
fractional survival under average conditions is 200/
500,000, or 4 x 10~4. This will be used as an initial
basis for design.
Step 2—Characterize the decay of chlorine residual.
This must be done experimentally, or by resort to
information from the literature.
For the conversion of free chlorine to combined
chlorine in wastewaters where the CI:N ratio is far
below breakpoint, a first order decay relationship may
be assumed:
Cf = Co exp (-kit)
(5-81)
In Equation 5-81, Cf is the free chlorine concentration,
C0 is the chlorine dose, and is the apparent first order
decay constant, which may be given by (81):
(5-82)
ki(s-1} = 9.7 x 1010CNKDN[H+]exp(-1510/T)
([H+] + KKA)([H+] + KDN)
In Equation 5-82, CN is the concentration (in M/l) of
ammonia nitrogen, KA is the dissociation constant of
HOCI (given by Equation 5-6), and KDN is the
dissociation constant of the ammonium cation. T is
the Kelvin temperature. With an influent ammonia
concentration of 10 mg/l as N (=0.7 mM), a mean
temperature of 15°C (288°K), and an influent pH of
7.5, ki equals 193 s"1.
The resultant combined chlorine, once formed,
decays. It has been found that the kinetics of this
decay may be described by two parallel first order
reactions (76):
Cc = Co [x exp (-kat) + (1 - x) exp (-k2't)] (5-83)
In Equation 5-83, x, k2, and k2' are constants
characteristic of the wastewater. For a variety of
wastewaters, it has been found that values of 0.3,
1.67x10~2s~1 and5x10~5s~1, respectively are typical
(76). Since the rates of these two decay reactions are
disparate, Equation 5-83 may be simplified to:
Co = Co (1 - x) + Co x exp (-k2t) for t < 1 /k2 (5-84)
Cc = Co (1 - x) exp (-k2't) for t > 1 /k2 (5-85)
With the particular constants assumed. Equations
5-81,5-84, and 5-85 may be written as the following,
with t in minutes:
Cf = Co exp (-11,580t) (5-86)
Cc = .7C0 + .3C0 exp (-t) for t < 1 min (5-87)
Cc = .7C0 exp (-.003 t) for t > 1 min (5-88)
Step 3—The chlorine decay relationships may now be
substituted into the kinetic relationships describing
microbial inactivation.
Method 1: Using Equation 5-65, with parameters
from Aieta and Roberts (178), the following may be
written:
N/No = (0.25Ctr*'8a
77
(5-89)
-------
Using either Equation 5-87 (fortimes below 1 minute)
or Equation 5-88 (for times greater than 1 minute) to
substitute C as a function of time into Equation 5-89,
an equation relating contact time, chlorine dose, and
inactivation may be obtained. In other words, it may
be shown that:
(5-90)
N/No = (.175 Co t + .075 C01 exp (-t))~2'821 < 1 min
N/No = (.175 Co t exp (-.003))~2-821 > 1 min (5-91)
Therefore, since N/IM0 has been specified, if a value of
t is assumed, C0 may be calculated. Hence, for times
of 10, 15, 30 and 60 minutes, C0(the chlorine dose)
equals 9.45, 6.39, 3.34 and 1.83 mg/l, respectively;
the effluent residual chlorine concentrations may be
determined from Equation 5-88 to be 6.42,4.28,2.14
and 1.07 mg/l, respectively.
Method 2: From the Chick-Watson equation, assum-
ing inactivation by free and combined chlorine is
additive, the following may be written:
dt
(5-92)
The chlorine demand relationships may be substi-
tuted into this equation, and it can be integrated
numerically.
Fort>k2, the result may be obtained as an analytical
integral in the following form:
-In (N/No) = (kiC0n(/kinf) (1 -
+ (kcC0nc (1 - x)
(5-93)
For the inactivation data for E. coli at room temper^
ature and relatively neutral pH, values for kf, kc, nf and
nc of 30, .085, 1 .46, and 1 .25, respectively, may be
assumed (Table 5-11). Using these, and the prior
assumed constants, the following equation is ob-
tained:
-In (N/No) = 2.86x1Cr3C0
1.46
(5-94)
+ 14.4C01-25 - 14.5C01-25e-aco375t
Equation 5-94 relates N/N0, C0, and t, and since the
first parameter has been specified, given a value for t,
the requisite value of C0 may be calculated. For t
equals 10, 15, 30 and 60 minutes, C0 equals 10.03,
6.92, 3.85, and 2.26 mg/l, respectively. Using these
doses and contact times, from Equation 5-88 the
effluent chlorine residuals may be calculated to be
6.81, 4.63, 2.46, and 1.32 mg/l, respectively.
Step 4—By two methods, suitable combinations of
residence time and chlorine dose have been deter-
mined which will satisfy the 30 day coliform standard
under average conditions. Based on this information,
a contact chamber will be sized. It is now necessary to
determine, for a specific volume contact chamber, the
performance under various extreme conditions to
ascertain the maximum chlorine dose likely to be
necessary. While this sensitivity analysis can be
performed using either Method 1 or Method 2 (i.e..
Equation 5-91 or 5-94), the approach will be illus-
trated using Method 2 (Equation 5-94), which, based
on the computations in Step 3, for this situation,
appears to be more conservative.
Based on step 3, it is anticipated that a 60-minute
contact tank will be most suitable (obviously, the
following can be repeated with different initial
assumed tank sires to generate economically opti-
mized designs). For the design mean flow, this would
require a volume of 1,200 m (312,500 gal). Rounding
up, the design volume of 320,000 gallons is selected.
At the design mean 'flow and influent coliform
concentration, this would require a chlorine dose of
2.26 mg/l. However, at higher flow rates and coli-
forms, the chlorine dose must be increased if the
effluent is to be in compliance with the coliform
standard. Based on the assumed tank volume, at each
combination of chlorine dose, flow rate, and influent
coliform concentration, the anticipated effluent coli-
form and total residual chlorine concentration maybe
computed. For example, at a dose of 3.7 mg/l, for the
given tank, the following are to be expected:
Assumed
Flow
avg design
max design
avg design
max design
Influent
Coliforms
avg
avg
max
max
Effluent
Coliforms
(#/100ml)
0.1
197.4
0.6
789.7
Effluent
Total
Chlorine
Residual
(mg/l)
2.15
2.36
2.15
2.36
This is in compliance with the standards under all
conditions except for the combination of maximum
design flow and maximum influent coliform concen-
trations. For design purposes, it may either be
assumed that these two factors are unlikely to occur
coincidentally more than once in 7 days (and hence
the 7-day standard will be complied with), or the dose
necessary to produce compliance at this extreme may
be computed.
At this point, the chlorinator(s) that is(are) to be
specified must be capable of a maximum output of 3.7
mg/l at the maximum design flow of 15 mgd, or 462
Ib/d. This is close to 500 Ib/d, which is a commonly
available chlorinator size; to provide system redun-
dancy, two such units would be specified. Alter-
natively, the overall plant design might use two
78
-------
parallel chlorination systems, each sized to half the
design capacity; in this situation, two 250 Ib/d units,
plus one spare 250 Ib/d unit would be specified.
Step 5—Based upon a 320,000 gal volume, the
specific contact basin may now be designed. The two
particular options considered will be the use of a
single baffled tank, and the use of a long pipe.
Option 1—Long pipe contact:
If a 2-m (78-in) diameter pipe running full is used,
the required pipe length may be given by appli-
cation of the formula of a volume of a cylinder.
Thus, a run of 389 meters is necessary to provide
the required volume. The velocity in the pipe at
mean flowjs 389 m/3600 s = 0.11 m/s (21
ft/min). This should be checked against the
specific scour velocity of the particles expected to
be present to ensure minimum sludge deposition.
The dispersion in this pipe may be estimated
using Equation 5-77. The pipe Reynolds number
is given by Re = VDR/u (where R is the density of
water and u is the viscosity). For turbulent flows,
the friction factor T is given by the following:
1/f°'5 = 2 logic (D/e) + 1.14 (5-95)
In Equation 5-95, e is the absolute pipe roughness
(0.002 inches is typical for concrete). Thus f =
0.0094. Substituting in Equation 5-87 along with
the pipe diameter and the length, the dispersion
number is determined as 4.9 x 10~5, sufficiently
low that the plug flow assumption is justifiable.
The headless may be estimated using Manning's
equation. With an 'n' value of 0.015, the hydraulic
grade is calculated as 9.767 x 10~7 at mean
design flow, and 2.127 x 10~7 at peak design flow.
Therefore, the estimated headless is 0.038 cm
and 0.151 cm, respectively, under these two
conditions.
Option 2—Serpentine contact:
A rectangular baffled contactor with a depth of 2
m is used. The sketch in Figure 5-24 illustrates
one configuration providing the required resi-
dence time.
The flow path of the system is40x7 = 280 m.The
L/W ratio is 280/2.16 = 129. The velocity (at
average design flow) between parallel baffles is
0.076 m/s, and through the baffle slots is 0.082
m/s. This should be checked against the specific
scour velocity as in the previous option. From
Equation 5-78, the estimated 'd' value is 0.0047
(for K = 1), again indicating reasonable approach
to plug flow. Finally, the headloss is computed
Figure 5-24. Definition sketch for rectangular contactor.
1.4m •**•
{1.4m
from Equation 5-80 as 0.41 cm at average design
flow and 1.65 cm at peak design flow.
Step 6—Based on conditions at maximum design
flow, assuming peak chlorine dose of 3.7 mg/l, the
chlorine residual leaving the contact basin is esti-
mated as 2.36 mg/l (see table under step 4). Thus, at
peak design flow, sulfonation capacity (assuming 1
mg S02 per mg chlorine residual) of 15 mgd x 2.36
mg/l x 8.34 = 295 Ib/d is needed. From this
information, the sulfonator(s) can be specified, keep-
ing in mind the desired redundancy.
Step 7—Annual chemical requirements can be
estimated using average design conditions. As noted,
at mean design flow and coliform concentrations, a
chlorine dose of 2.18 mg/l would be satisfactory, and
would impart a residual exiting the contact chamber
of 1.27 mg/l (using Equation 5-88). Thus, the average
annual chlorine requirement under design conditions
would be 365 x 7 mgd x 2.18 mg/l x 8.34 = 46,300
Ib/yr (say 24 tons/year). The average annual re-
quirement for sulfur dioxide under design conditions
(assuming 1 mg SO2/mg chlorine residual) would be
365 x 7 mgd x 1.27 mg/l x 8.34 = 27,000 Ib/yr (say 14
tons/yr). This information could then be used to
design the chemical handling facilities.
If chlorine and sulfur dioxide are obtained in ton
containers, and if gaseous withdrawal is utilized, the
average chemical consumption is 2.4 kg (5.3 lb)/hr
and 1.4 kg (3.1 lb)/hr, respectively, for chlorine and
sulfur dioxide. Both of these are substantially less
than the maximum safe rate of supply from ton
containers. Hence, there is no need to manifold more
than two cylinders together (one in service and one
awaiting) for either chlorine or sulfur dioxide. It would
be reasonable for the treatment plant to anticipate
ordering 2-ton containers of chlorine and 1-ton
container of sulfur dioxide every four weeks. A safe
inventory would be 3 containers of chlorine and 2
containers of sulfur dioxide (this would provide for a
28-day reserve for transportation, plus excess for
emergencies), and provision should be made for
79
-------
storage of 2 empty chlorine containers and 1 empty
sulfur dioxide container.
Step 8—A detailed cost estimate would be made at
this point, based on actual catalog costs, or prior bid
data. For the purpose of estimating chemical con-
sumption and annual O&M costs during the period
prior to attainment of actual design flow, the chlorine
dose to achieve inactivation at various interim flows
can be computed (using Equation 5-94), and inte-
grated over an estimated flow-duration curve. Addi-
tionally, other alternative design contactor volumes
may be assumed, and the design computations begun
again commencing with step 4.
5.7.7 Economics
The estimation of chlorination, chlorine dioxide or
dechlorination costs is highly site specific. Particular
concerns include the necessity for separate chlorine
contact basins (versus the use of an effluent channel
as a contactor), the site-specific chemical costs, and
the necessary chlorine dosages. However, prelim-
inary rough estimates of process alternatives may be
developed from available literature data based on
field experience.
Geisser et al. developed cost equations based on their
studies on the disinfection of combined sewer
overflows using either chlorine or chlorine dioxide
(179). Other useful sources of chlorination and
dechlorination cost data are Chen and Gan (112), who
have developed estimates for sulfur dioxide dechlo-
rination systems, and Gumerman et al. (180), who
provide cost estimates for drinking water chlorination
systems.
The design engineer may also use standard estima-
tion manuals for the costing of contact basins and
physical structures, plus manufacturers' quotations
on equipment (chlorinators, mixers, etc.).
5.8 Safety and Occupational Health
Considerations
Halogen disinfectants have certain properties that
must be considered at the design stage for the
protection of operating personnel from risks that may
arise during the plant life. Much detail is available in
regard to necessary safetyfeatures to be incorporated
in wastewater disinfection facilities using gaseous
chlorine or hypochlorite as disinfecting agents. Little
information is available in regard to adequate pre-
caution using other halogens.
5.8.1 Physical Site Layout
For plants using gaseous chlorine, general space
requirements may be estimated as follows (157):
• single chlorinator facilities of less than 200 Ib/d
(90 kg/d) capacity require at least 64 ft2 (6 m2) of
space for the chlorinator and ancillary equipment;
• for plants with two chlorinators and feed rates up
to 180 kg (400 lb)/d, 15 m2 (160 sq ft) of area are
needed for the chlorinators; and
• for each extra chlorinator above two, an additional
15 m2 (160 sq ft) area should be provided.
In addition to the above areas needed for the
chlorinator and evaporator modules, space is also
required for the chlorine containers or cylinders being
used to feed the system (or, in the case of sodium
hypochlorite plants, the hypochlorite storage tanks)
and space for inventory and empty cylinders or
containers. The size of gaseous chlorine containers is
given in Table 5-14, and, in conjunction with esti-
mates of required inventory, working supply, and
storage of empties, can be used to develop site area
estimates for chlorine storage.
Table 5-14. Physical Dimensions of Chlorine Gas Containers
Diameter Length
Capacity (inches)3 (inches)
150 Ibs (70 kg)
2000 Ibs (910 kg)
10.25 to 10.75
30
53 to 56
79.75 to 82.5
a150 Ib cylinders should be used and stored upright, thus requir-
ing approximately 0.6 ft2 (0.06 m2) (plus separation area) per
cylinder, while ton containers are used and stored in the horizon-
tal position, requiring approximately 17 ft2 (1.6 m2) (plus separa-
tion area) per container.
The chlorine storage room (which may or may not be
separate from the room containing the chlorinators)
should be isolated from any other process by use of a
separate building, or by use of an isolated room
bounded by fire resistant walls (20). If the chlorine
cylinders are physically separate from the chlorin-
ators, and if the gaseous chlorine withdrawal method
is used, it is necessary that the temperature of the
room containing the chlorine gas cylinders be kept
below the temperature of the room containing the
chlorinator in order to prevent reliquifaction in the
lines between the chlorine supply and the chlorin-
ators. For chlorine installations practicing liquid
withdrawal from either ton or bulk containers,
outdoor storage is acceptable. Climatic considera-
tions will dictate necessary site specific details, such
as provision for prevention of icing of valves and
shielding containers from direct sunlight.
The chlorine supply room must have at least 2 means
of egress, with doors opening outwards from the
room. The structural designer must consider the dead
load of full chlorine containers in his calculations. If
ton containers are to be used, floor mounted trunions
or scales (Figure 5-25) are needed. Furthermore, for
ton containers, the use of overhead crane of at least
2-ton capacity in conjunction with a special lifting bar
is needed to facilitate container movement, and
therefore the physical site layout must carefully
consider ceiling heights. It is highly desirable that the
80
-------
Figure 5-25.
Storage
Roller
Ton container mounting trunions. (Courtesy
Force Flow Equipment.)
Trunnions ..^.
Vi Dia.
Length = 21" Bolt „„
Width = 4" \
Capacity - 4000 Ib
Container
weight - 1 b IDS each /
ASTM A7 S
Epoxy Finis
+ ;j>
J***} 3" Dia. Bushed Beam
hed i. is» ,i /
Storage
Cradles
Length = 21"
Width = 4"
Capacity = 4000 Ib
Va" DiaJ
'Bolt
Container
Weight = 25 Ibs each
ASTM A7 Steel
Epoxy Finished
chlorine container(s) in use be placed on a scale or
load cell to have a positive record of remaining
disinfection.
Chlorine cylinders of 68-kg (150-lb) capacity are
emptied in a vertical position, secured to a wall or a
sturdy upright by means of a band or chain clamp to
the upper portion of each cylinder.
Inthe chlorine storage room, ventilation must be such
to assure a complete air change in 1-4 minutes.
Generally, this is achieved by use of an exhaust fan
near the floor of an outside wall (20); such fan must
have a switch outside the chlorine room itself as well
as inside. No ductwork, shafts, or other potential
sources of gas travel should exist between the
chlorine supply room and any other portion of the
plant. The exhaust fan should vent at a level above
neighboring buildings, trees, etc., to afford high
dilution of any contained contaminant gases with the
atmosphere.
No turpentine, ether, ammonia (except for small
amounts associated with leak detection equipment),
finely divided metals, or other flammable materials
are permissible within the chlorine storage room (20).
5.5.2 Leah Detection
In gaseous chlorine installations, the major safety
and health concerns involve the possibility for
chlorine leakage to occur from a cylinder, a valve, or
piping. The threshold limit value (TLV) for worker
exposure to chlorine in air is 1 ppm by volume as an
8-hr time weighted average (181). Other atmospheric
chlorine concentrations of interest are presented in
Table 5-15 (20). It is particularly noteworthy that the
•minimum chlorine concentration detectable by odor
is greater than the above specified TLV. Therefore, in
order to provide for the continuous sensing of
chlorine leakage, it is necessary to rely upon some
chemical or electronic device. These may be of
several types.
Table 5-15. Gas Phase Chlorine Concentrations Evoking
Specific Effects (20)
Response Concentration (ppm v/v)
Minimum odor threshold
Minimum 1-hour no serious effect
level
Throat irritation
Coughing
30-minute, 1-hour danger level
3.5
4.0
15.1
30.2
40 to 60
The older type of continuous chlorine sensors rely
upon a version of the iodometric chlorine detection
procedure, using starch/iodide or other chemically
impregnated paper and measuring the change in
color that results when vapor containing chlorine
passes over this material. A more modern version of
this principle is used in a sensor that mea$ures the
current required to electrochemically reduce the
chlorine present in a gas (Mine Safety Appliances,
Pittsburgh, PA). A third principle used in continuous
chlorine gas analyzers uses the change in electrical
conductivity of a gas that occurs as chlorine concen-
trations increase (International Sensor Technology,
Irvine, CA).
Once the vapor phase analyzer has signaled the onset
of a chlorine leak, it is necessary to determine the
cause of such event and proceed towards its repair.
Personnel entering the room in which the ambient
chlorine levels have exceeded the alarm limit should
be clothed in protective equipment (gloves^ breathing
device, suit). Using portable continuous chlorine
sensors, the location of highest chlorine concentra-
tion may be found, and thus the site of the leak
inferred. Alternatively, it is possible to locate the site
of a gaseous chlorine leak by spraying a solution or
the vapors of a solution of strong ammonia (com-
mercial, not household strength). A white cloud of
NrUCI precipitate will form in the region where the
chlorine leak exists (181).
81
-------
5.8.3 Emergency Responses
All breathing devices used for personal protection in
chlorine installations; must be U.S. Bureau of Mines
approved, and should be sterilized after use if it is
anticipated that they may be used by another
individual. The "buddy" system should be employed
in situations where it is necessary to go into a room
containing high vapor concentrations of chlorine.
Approved breathing devices may be categorized as
follows (20):
Industrial Canister Mask ("gas masks"). This is only
suitable for atmospheric chlorine concentrations
below 1 percent and oxygen concentrations above 16
percent, and only for short duration use. It should not
be relied upon for use during leaks, and rigid
adherence to canister replacement time is required.
Self-Contained Breathing Apparatus. These may use
a portable cylinder of oxygen or air, or a chemical
generation system ifor continuous production of
oxygen. For the latter type of system, it is necessary to
wait until oxygen production has commenced prior to
entry into a contaminated area. No oxygen system
should be used for entry into closely confined areas
(storage tanks) or where the danger of sparks or fire
exists.
Positive Pressure Blower Mask. A hose is used to
supply breathable air from either a remote air tank or
a remote compressor. Use is permissible only if
immediate safe escape from the contaminated area is
possible in the event of a failure in the air supply
system. If a compressor is used, its intake should be at
least 6 ft (1.8 m) above grade to prevent inadvertent
contamination with chlorine emanating from the area
under investigation.
Personnel should be trained in the use of such
breathing devices, and, in particular, in the location
and management of chlorine leaks while using such
protective equipment.
When a chlorine leak is located, if possible, the
chlorine cylinder or container should be turned so
that the defect is shifted at the upper portion of the
container. This will assure that gaseous (rather than
liquid) chlorine leakage occurs, and, eventually, the
leak may be self-limited by evaporative cooling
described earlier. All facilities handling gaseous
chlorine should have available the appropriate Chlo-
rine Institute emergency kit, and be familiar with its
use. Devices in this kit may be used to patch or repair
the leaking container, valve, or line.
If the gaseous or liquid chlorine that leaks can be
collected, it is possible to neutralize it using alkaline
solutions. Table 5-16 describes the required chem-
icals and water necessary to neutralize the contents
of 68-kg (150-lb) and 910-kg (2,000-lb) chlorine
containers. In the event of fire in association with a
chlorine leak, no water should be used, since this will
increase the corrosivity of the mixture. Non-water
based extinguishers should be used until the chlorine
leak is stopped.
Table 5-16.
Neutralization Requirements for Chlorine
Containers
Container
Capacity
150 Ib
2000 Ib
100 Percent
Caustic
188 Ib
in 60 gallon
2500 Ib
in 800 gallon
Soda Ash
450 Ib
in 150 gallon
6000 Ib
in 2000 gallon
Hydrated
Lime
188 Ib
in 188 gallon
2500 Ib
in 2500 gallon
Should any persons come in contact, either by
inhalation, or eye or skin contact, with chlorine
resulting from a leak, the following first aid measures
may be taken, prior to consultation with a physician
(156):
General. Remove the person to an uncontaminated
area, and remove contaminated clothing, washing
any parts of the body exposed to chlorine with water.
Inhalation. If breathing has ceased, commence
artificial respiration, When breathing recommences,
or if breathing has not stopped, administer oxygen.
Keep the person warm and at rest.
f ye Contact. The eye should be flushed with water for
15 minutes, holding the eyelids apart to get complete
irrigation. In the design stage, eyewash basins should
be provided.
Skin Contact. Wash the exposed parts with soap and
water. It would be desirable to provide in the facility
design provision for an emergency shower. Installa-
tions using sulfur dioxide should adhere to safety
precautions regarding handling of a nature similar to
those used in chlorine handling. At the design stage,
it is particularly useful to specify fittings, valves, etc.,
for SOa service that are compatible with Chlorine
Institute emergency kits.
5.9 Operation and Maintenance (O&M)
Requirements
In the start-up of chlorination systems initially, or
after a shut-down, a careful procedure must be
followed to minimize the likelihood of chlorine leaks.
Sepp and White outline the following procedure
when gaseous chlorine is used (157):
1. Check all joints for proper gasketing and check
that all supply valves are closed;
82
-------
2. Check ejector for proper vacuum;
3. Place automatic chlorinators in manual mode,
and set feed rate for about 25 percent of
maximum;
4. Open one chlorine cylinder slightly, and check
all joints for leaks. If no leaks are found the
remaining number of required cylinders can be
opened;
5. If a leak occurs, immediately close the cylinder
valve, open all other valves fully and increase
the chlorine feed setting on the chlorinator to its
maximum value. When all chlorine has been
ejected, repair the leak, and return to step (1);
6. Check performance at maximum rated capacity.
The following are likely causes of capacity loss:
a. insufficient ejector vacuum,
b. vacuum line leaks,
c. insufficient feed water pressure,
d. solution line friction losses too high due to
too small pipe diameter,
e. air or gas binding in solution line.
The following modifications to the above must be
made when the liquid chlorine withdrawal method is
used (157):
1. Dry the entire system by heating water in the
evaporator bath and passing dry air (-40°C dew
point) through the evaporator and all supply
lines from the cylinders to the chlorinators.
Several hours may be required for this step.
2. Bring the evaporator water to the design tem-
perature. Start the ejector water flow. Start up
system with gas phase withdrawal to check
leaks as above. If no leaks are found, then liquid
withdrawal may be commenced.
The maintenance requirements for the chlorine
supply system are as follows (157):
1. Flexible copper tubing reliability must be per-
iodically checked. This can be done by bending
the lines slightly. Any "screeching" is indicative
of corrosion, and the tubing must be replaced.
2. Minute leakage from lines and fittings, undetect-
able by odor or standard leak detection pro-
cedures, may be ascertained by inspecting for
signs of moisture accumulation or metal dis-
coloration, both of which are signs of incipient
leak development.
3. In some cases, a one inch flat file may be used to
reface chlorine cylinder valves to ensure more
precise seating.
4. Evaporator vessels should be inspected for
sludge accumulation every year or each 200
tons of chlorine. If superheating capability
declines, this may be indicative of either sludge
accumulation or heater failure. Piping and
connections to the evaporator should be in-
spected every six months. Sacrificial anodes
used for corrosion control in the evaporator
water bath should be inspected, and, if neces-
sary, replaced, every six months.
5. The sludge accumulation in the evaporator may
be cleaned by flushing with cold water until the
effluent water runs clear. At this time, the
vessel can also be inspected for pitting, and, if
this is severe, replaced. Before placing the
evaporator back into service, it must be dried by
holding the external water temperature at 82°C
(180°F) and maintaining a vacuum of 25 inches
of mercury (85 kPa) for 24 hours.
6. The chlorine gas filter should be inspected every
six months, at which time the filter element
should be replaced. The sediment trap should be
washed and dried at this time, and the lead
gaskets disposed and replaced.
7. The chlorine pressure reducing valves may be
cleaned of any deposits with a soft cloth, or, in
more severe cases, with isopropyl alcohol or
trichloroethylene. The valve spring should be
replaced every two to five years.
The chlorinator system itself is subject to the fol-
lowing maintenance requirements (157):
1. The ejectors should be disassembled and
cleaned every six months, and iron and man-
ganese deposits removed with muriatic acid.
2. Booster pumps are subject to similar main-
tenance requirements as any other pump.
3. The chlorinator rotameters and floats should be
removed and cleaned every six months, and the
metering orifice inspected. All valve stems and
seats should be cleaned and inspected once a
year, and all valve springs should be replaced
every two years.
The chlorine analyzer is probably the single system
component requiring most careful attention. The
following operation and maintenance requirements
are essential (157):
1. Sample lines must be inspected daily for solids
accumulation and filters and screens cleaned
daily. This is particularly important in lines from
dechlorination systems that lack residual, which
could reduce line fouling.
83
-------
2. In amperometric systems, electrodes should be
cleaned weekly or biweekly, or more frequently
if erratic or drifting calibrations occur.
3. The effluent line plH should be checked daily
with pH paper. If it is not within the range of 4.5
to 5.0, a stronger buffer solution may be
required.
4. Periodically, preferably at least daily, the ana-
lyzer reading should be compared with the
results from a manual analysis. A control chart
graphing the difference in results as a function
of day should be maintained. If there is a
consistent deviation, the span control on the
analyzer may be adjusted on a weekly basis. A
trend to increasing deviation may be indicative
of instrument malfunction or sample line or
filter fouling.
5. The adequacy of Kl addition may be checked by
adding slightly more Kl than recommended to
the feed buffer solution. If there is a change in
cell response, this is indicative of insufficient Kl.
Continue increasing the Kl concentration until
no further change in the analyzer output occurs.
Gulp and Heim present a useful guide to trouble-
shooting of chlorination systems (182). This is
reproduced in Table 5-17.
5.10 Case Studies
In this section, several treatment plants that use
chlorination and dechlorination will be described,
indicating their design details, operating experience,
and features of their facility that influence their ability
to achieve satisfactory performance.
5.10.1 Stony Brook Regional Sewerage A uthority
Treatment Plant #1, Princeton, NJ
This treatment plant is located in Princeton, New
Jersey and services a metropolitan area consisting of
the Borough and Township of Princeton, West
Windsor Township, Hopewell Borough and Township,
Pennington Borough, and South Brunswick Town-
ship, NJ. Wastewater is collected in separate sewers,
and the service area is essentially 100 percent
domestic. Within the service area is the campus of
Princeton University.
The current treatment pliant was completed in 1977.
The average daily design tflow is 11 mgd (482 l/s), and
the current average daily flow is 5 mgd (219 l/s). The
discharge is to the Millstone River, a tributary to the
Raritan River.
Permit requirements include a BOD of 8 mg/l, a 2
mg/l ammonia nitrogen requirement, a suspended
solids requirement of 10 mg/l, a maximum chlorine
residual of 0.05 mg/l, and a bacteriological standard
of 20 fecal coliform/100 ml (enforced year round).
The bacteriological standard is enforced at the
chlorine contact tank effluent. In addition, a minimum
dissolved oxygen requirement of 6.0 mg/l is enforced
at the final effluent.
The treatment system consists of primary and pre-
liminary treatment, activated sludge, separate stage
biological nitrification, multi-media filtration, chlo-
rination, dechlorination with sulfur dioxide, and
reaeration.
The filtered, nitrified effluent routinely contains less
than 1 mg/l suspended solids, and 1-3 mg/l total
Kjeldahl nitrogen (occasional spikes of 10-20 mg/l
occur). Nitrites have not been found in this effluent.
The disinfectant used is gaseous chlorine, supplied in
ton cylinders. Chlorine usage is 6.5 kg/h (350 Ib/d),
on average (average dose of 8.4 mg/l) and two
cylinders may be manifolded for withdrawal. The
chlorine supply system was originally designed for
liquid phase withdrawal, and the plant has evapora-
tors; however, due to the low hydraulic loading, gas
withdrawal has been found to be satisfactory, and no
problems with reliquifaction have been noted. Ap-
proximately 1 cylinder in 30 is defective, and an
inventory of about 6 cylinders is maintained, with
new orders received when only two full cylinders
remain (this is about a 10 day supply).
The chlorine solution is dispensed through solution
feed chlorinators. The plant was originally designed
to operate on a residual control mode; however, the
operators noted that the analyzers were difficult to
maintain, and that the control system was unstable
(possibly due to a 200-500 ft length of sampling line,
or to the location of the point of sampling prior to the
end of the rapid decay phase of chlorine demand). The
chlorination process is currently operated in manual
mode, with residual monitored 3 times/shift at the
end of the chlorine contact chamber. A residual of
1.5-2 mg/l is generally used as a control point (some
of this residual is free).
The chlorine solution is applied at a separate stilling
well through submerged nozzles. Two parallel, baf-
fled (3 baffles of the end-around type) contact
chambers are used. The contact time at design
average flow is 18 minutes, making the current
contact time at average flow about 40 minutes. Slime
accumulation has occurred, and the contact tanks
must be cleaned by high pressure water hoses every
3-4 months. Some difficulty in cleaning these tanks
due to lack of provision for isolated clean outs for
wasting of the loosened debris has been noted.
The dechlorination system uses gaseous sulfur
dioxide, also supplied as "ton" containers, with gas
withdrawal. The S02 utilization rate is 89 kg (197
84
-------
Table 5-17. Trouble Shooting Guide Adapted (181)
Indicators/Observations and Action
1. Low chlorine gas pressure at chlorinator.
A. Reduce feed rate and note if pressure rises appreciably after short period of time.
(i) so, it is likely that more cylinders must be manifolded together to avoid exceeding the maximum safe withdrawal rate.
(ii) if not, but if icing or cooling effect on lines continues, it is likely that there is a stoppage or flow restriction between the
cylinders and chlorinators. Disassemble the header system, locate the blockage, and clean with solvent.
2. No chlorine gas pressure at chlorinator.
A. Visually inspect to verify that chlorine cylinders are connected, and that they are not empty.
B. Inspect the pressure reducing valve for plugging or damage, and repair, if necessary, following emptying chlorine gas from
system be certain that a sediment trap is upstream of the valve.
3. Chlorinator will not feed any chlorine.
A. Visually inspect the chlorinator valve stem and seat for dirt, and clean if necessary; precede the valve with a sediment trap.
B. Measure the temperature in the chlorine cylinder storage area. If this area is warmer than the chlorinator room, reduce the
temperature (shading, ventilation).
4. Chlorine gas escaping from chlorine pressure reducing valve (CPRV).
A. Place ammonia bottle near termination of CPRV vent line to confirm leak.
(i) disassemble valve and diaphragm, repairing if necessary. Inspect chlorine supply system for moisture intrusion.
5. Inability to maintain chlorine feed rate without icing of chlorine system.
A. Reduce feed rate to about 75 percent of evaporator capacity. If this eliminates the problem, then there is an insufficiency of
evaporatory capacity.
B. Inspect CPRV cartridge and flush and clean if necessary.
6. Chlorination system unable to maintain water-bath temperature sufficient to keep external CPRV open.
A. Remove and replace water bath heating element.
7. Inability to obtain maximum feed rate from chlorinator.
A. Check chlorine gas pressure; if inadequate, remove and replace empty cylinders.
B. Check water pump on injector for clogging, and clean, if necessary, with acid.
C. Check for leaks in the vacuum relief valve; disassemble and replace all springs.
D. Check for leaks in the vacuum lines, joints, gaskets associated with the chlorinator system by using an ammonia solution, or
moistened starch-iodide indicator paper. Repair all leaks, and replace all leaking gaskets, tubing, etc.
8. Inability to maintain adequate chlorine feed rate.
A. Inspect water feed pump and overhaul, if indicated. If a turbine pump is used, try closing the needle valve to maintain proper
discharge pressure.
9. Wide variation in chlorine residual produced in effluent.
A. Check chlorine meter capacity against plant flow meter.
(i) replace with higher chlorination capacity meter.
B. Check automatic controls, and request manufacturer's service, if indicated.
C. Check for solids accumulation in the contact chamber, and clean, if necessary.
D. Check zero and span of flow control device on chlorinator.
(i) re-zero and span the device in accordance with manufacturer's instructions.
10. Chlorine residual analyzer recorder controller does not control chlorine residual properly.
A. Inspect electrodes for fouling, and clean if necessary.
B. Check loop-time, and reduce, if necessary, by one of the following actions:
(i) move injector closer to point of application.
(ii) increase velocity in sample line to analyzer cell.
(iii) move cell closer to sample point.
(iv) move sample point closer to point of application.
C. Check that sufficient Kl is being added for the anticipated amount of chlorine residual, and increase Kl if indicated.
D. Check pH of analyzer cell, and replace buffer solution, or increase buffer strength if the pH is out of range.
E. Disconnect analyzer cell and apply a simulated signal to recorder mechanism. If recorder works, contact authorized service
personnel for repair of analyzer cell.
F. Analyze a series of samples taken after the point of mixing manually for chlorine residual under conditions of constant chlorine
feed rate. If there is a wide variation in residual, inadequate turbulence at the point of mixing is indicated—enhance mixing
efficiency.
G. Check that the rotameter tube on the chlorinator is in the proper range, and replace, if necessary.
85
-------
lb)/d. Based on a 2 mg/l chlorine residual, this is 2.4
mg SOz/mg Cla. The sulfonator was designed for
manual control and is controlled by monitoring the
influent chlorine residual 3 times/shift. No particular
problems with the sulfonator system were noted.
There is no data available on the dissolved oxygen at
the effluent from the sulfonator mixing point; how-
ever, the plant was designed with reaeration due to
the effluent DO permit condition. The operating
personnel only use the aerators sporadically, and
believe that the necessity for reaeration stems from
the oxygen demand from nitrification rather than
from any demand from sulfur dioxide.
Both the chlorine and sulfur dioxide containers are
stored in a single room isolated from the other rooms
in the plant. Temperature control and ventilation is
provided. Necessary safety equipment, including self-
contained breathing systems, and leak detectors are
on-hand. The two cylinders each of sulfur dioxide and
chlorine which are in use are mounted on hydraulic
load scales. The operating personnel feel that the
necessary handling and safety features appropriate
to chlorine are also appropriate to sulfur dioxide. No
corrosion or leak problems, other than occasional
defective cylinders, have been noted.
Due to the substantial overcapacity, no significant
down times have been recorded, except for an
instance of failure of the sulfonator due to water
intrusion. Spare vacuum regulator parts and valve
springs are maintained. However, there is no periodic
maintenance program.
The operating personnel estimate that 3 hr/d of
operator time are required for the manual sampling
and dose adjustments for the chlorinators and
sulfonators, and about 0.5 hr/wkfor cylinder switch-
overs. Some complaints as to the lack of manufacturer
training for maintenance and troubleshooting of the
chlorinators and sulfonators were voiced.
The 1983 annual chemical costs for the chlorination-
dechlorination system were $14,000 for the chlorine
($0.24/kg) and $6,500 for sulfur dioxide ($0.20/kg).
5.10.2 Southeast Water Pollution Control Plant,
City and County of San Francisco, CA
This treatment plant is located in San Francisco, CA
and services a portion of the combined sewer system
of the City and County of San Francisco. The flow is
predominantly domestic.
The current treatment plant was completed in 1981.
Design average flow for secondary treatment is 85
mgd (3,700 l/s) with a peak flow of 210 mgd (9,200
l/s). Current flows are 20-200 mgd (875-8,750 l/s).
Discharge is to San Francisco Bay. Permit require-
ments are monthly average BOD/SS of 30/30 and a
5-day median total coliform requirement of 240/100
ml (with a geometric mean coliform requirement over
30 days of 200 fecal coliform/100 ml). The bacterio-
logical requirements are met at the chlorine contact
chamber exit, although problems with high counts
after dechlorination have been noted. The instan-
taneous maximum chlorine residual is 0 mg/l.
Treatment consists of preliminary and primary treat-
ment, closed tank oxygen activated sludge (non-
nitrifying), chlorination and dechlorination. Chlorina-
tion is achieved using 14 to 15 percent sodium
hypochlorite stored in fixed storage tanks. An approx-
imate six day storage capacity exists. Chlorine dose is
generally 12 mg/l.
While compound loop control hypochlorinators are
installed, they have not been used, since it is believed
(by the operators) that before the secondary treatment
process went on line, the use of primary effluent
fouled up the analyzers. Additionally, the automatic
control system was originally designed for gaseous
chlorine service, and only mid-way through construc-
tion were modifications made to switch to hypo-
chlorite service. Due to more pressing operational
difficulties, the shakedown of the automatic disinfec-
tion control system has not yet been accomplished.
Flow-paced manual chlorine control is practiced in
which the residual at the end of the contact chamber
is measured every hour, and feed rate adjusted to
maintain a 6-8 mg/l residual in dry weather and 3-5
mg/l in wet weather.
The hypochlorite is supplied to diffusers, and mixing
at the point of application is supplied by two 30-hp
(22-kW) turbine mixers. The chlorine contact system
consists of an underground channel (2 identical
parallel channels, each 680 m (2,200 ft) long, 3 m (10
ft) high and 2.8 m (9 ft) wide) providing 50 minutes
contact at 85 mgd(3,700 l/s). No solids accumulation
or sliming problems have been noted.
The dechlorination system uses sodium bisulfite
supplied as a 22 percent (as S02) solution. Tankage
for 1 week's supply exists. The dose of dechlorinating
chemical is also controlled in a flow-paced manual
mode after hourly analyses just after the point of
sulfite addition to provide about a 1-2 mg/l excess
dose. Mixing at the point of application is provided
using two 22-kW (30-hp) mixers.
The plant appears to have substantial problems in
achieving automatic chlorine control. However, these
are traceable to the initial use of the chlorination
equipment for treatment of primary effluent, and the
need to modify the disinfection/dechlorination sys-
tem from gaseous chlorine and sulfur dioxide to
sodium hypochlorite and sodium bisulfite late in the
design and construction process.
86
-------
5.70.3 Sacramento Regional Wastewater
Treatment Plant, Sacramento, CA
This treatment plant is located in Sacramento County,
CA and services a metropolitan area including the
City and County of Sacramento and the City of
Folsom. Except for a small portion of the downtown
Sacramento area, the facility services a separate
sewer collection system. About 10 percent of the flow
is due to industrial sources, primarily canneries,
which contribute a high BOD, low nitrogen waste-
water.
The current treatment plant was completed in 1982.
The design flow is 136 mgd (6,000 l/s) (daily average)
with a design hourly peak flow of 240 mgd (10,500
l/s). The current daily average flow is 130 mgd (5,700
l/s), with a peak hourly flow of 240 mgd (10,500 l/s).
The discharge is to the Sacramento River.
Permit requirements include a BOD monthly average
of 30 mg/l, 30 mg/l monthly average suspended
solids, monthly average total coliform MPN of 23/100
ml, with a daily maximum of 500/100 ml, a maximum
chlorine residual of 0.1 mg/l. The bacteriological
standard applies at any location at the choice of the
facility, and is currently achieved at the chlorine
contact tank effluent.
The treatment system consists of primary and pre-
liminary treatment, closed tank oxygen activated
sludge (non-nitrifying, although an average of 0.4
mg/l nitrite is present in the secondary effluent),
chlorination and dechlorination.
The secondary effluent contains an average of 12
mg/l suspended solids, 9 mg/l ammonia, and 22
mg/l total Kjeldahl nitrogen.
Gaseous chlorine is supplied in 90 ton tank cars
delivered by rail. The chlorine dose is 13 mg/l for an
average daily use of 260 kg/hr (14,000 lb)/d (giving a
12-day life for a tank car). Two tank cars are located
on site, with one in service and one full tank car
awaiting use, or on order. Standby equipment to use
ton containers also exists. The tank car loading
facilities are located away from the chlorinator
building, and liquid chlorine is withdrawn into an
intermediate storage tank with a level sensor. The
drop in the liquid level sensor in the storage tank is
used as a signal for switching to another tank car. No
load scales exist on the tank cars and this is
recognized as a design deficiency—the plant com-
puter control system maintains a cumulative log of
chlorine withdrawals from the tank car as a means of
determining when the tank car is nearly exhausted.
Withdrawal of chlorine is accomplished with a
compressed air padding system.
From immediate storage, chlorine is fed to evap-
orators, and the gaseous chlorine is then fed to
chlorinators. Chlorinators are controlled using a
compound loop with residual trim algorithm. The
residual at the end of the contact system is main-
tained at 9 mg/l.
The chlorine solution is applied through diffusers at a
point where two two-speed 36/72 kW (35/70 hp)
turbine mixers are located. The initial residual
analyzer sampling point is located a short distance
(probably less than 1 minute) from this point of
mixing. The contact system consists of a 3,000 m
(10,000 ft) pipe 2.6 m (8.5 ft) in diameter, that has
never been inspected (in approximately 1 year of
service) for slime or solids accumulation; however it
has not appeared to be a problem. The second
analyzer, providing the residual trim signal, is located
at the end of this pipe.
The dechlorination system uses gaseous sulfur
dioxide. The SO2 is supplied either by truck or by rail
and off loaded into a fixed 150 ton storage tank. The
fixed storage tank was modified such that its fittings
are compatible with the Chlorine Institute emergency
kit specifications. There is no load scale on the
storage tank, and inventory is monitored using the
control computer system. Withdrawal of liquid S02 is
accomplished under compressed air padding, and the
compressor system is separate from that used for the
chlorine withdrawal system. Sulfur dioxide dose is
typically 1 mg/mg chlorine residual, plus an excess of
2 mg/l for a daily use of 5,450 kg (12,000 Ib).
The sulfur dioxide is fed to evaporators that supply
sulfonators. Gaseous sulfur dioxide under vacuum is
then piped 3,000 m (10,000 ft) to the end of the
chlorine outfall, where it is mixed in ejectors and
then, using four circumferentially-mounted pipe
diffusers, injected to the wastewater flow. The control
system is compound loop using a biased chlorine
residual signal using a hypochlorite biasing system.
The set point is at a 2 mg/l sulfur dioxide excess,
partially due to the requirement for maintenance of
an undetectable chlorine residual, and partially due to
the inability to achieve precise sulfonation control in
the long vacuum line (a lag of 5 minutes in the SOa
vacuum supply line was cited). No problems with
dissolved oxygen depression or pH were cited,
although continuous pH sensors are used.
No problems with slime growth or recontamination
subsequent to dechlorination were expressed, al-
though the total coliforms measured after dechlor-
ination are somewhat erratic.
The plant has been fully operational for only about
one year, so evaluations of corrosion and machine
reliability are difficult. Some problems noted to date
are the corrosion of the trim chlorine analyzers from
the chlorine vapors emanating at the end of the
87
-------
contact pipe, the breakdown of PVC solution in 1.
sunlight (painting of PVC is suggested), and the lag
time in the sulfonation control system noted above.
The dechlorination analysis system is cleaned and
calibrated twice daily, and it is estimated that total 2.
operator time required for operation of the chlorina-
tion/dechlorination system is 4 hr/d. Table 5-18
summarizes the overall O&M schedule for the major 3.
components of the chlorination-dechlorination sys-
tem at Sacramento.
5.11 References
When an NTIS number is cited in a reference, that 4.
reference is available from:
National Technical Information Service
5285 Port Royal Road 5.
Springfield, VA 22161
(703) 487-4650
Belohlav, LR. and McBee, E.T. Discovery and
Early Work. In: Chlorine: Its Manufacture,
Properties and Use. ACS Monograph #154,
Reinhold Publ. Corp., New York, NY, 1962.
Baker, J.C. Use of Chlorine in The Treatment of
Sewage. Surveyor 69, 241, 1926.
Averill, C. Facts Regarding the Disinfecting
Powers of Chlorine. Letter to Hon. J.I. Degraff,
Mayor of the City of Schenectady. SS Riggs
Printer, Schenectady, NY, 1832.
Gascoigne, G.B. Chlorination of Sewage and
Sewage Plant Effluents. Sewage Works Jour-
nal 3:38-49, 1931.
Laubusch, E.J. State Practices in Sewage
Disinfection. Sewage and Industrial Wastes
30(10):1233-1240, 1958.
Table 5-18. O&M Schedule,
Component
Sacramento Regional Wastewater Treatment Plant
Action
CI2 leak indicator
CI2 pressure reducing shutoff
valves
SOj ejector
NaOCI biasing pump
Chlorine Unloading platform
Air padding system
Chlorine room exhaust fan
Air padding dryer
Chlorinators and sulfonaton;
CI2 and SO2 emergency
expansion tanks
Flash mixers
Sulfonator ejector
Hypochlorite mix tanks
Chlorine analyzers
Chlorine evaporators
Adjust sample flow rates (M)
Remove and clean valve seats (SA)
Check and clean (if necessary) (M).
Check and adjust V-belts (Q). Overhaul (A).
Exercise emergency baths and all valves (M). Check for leaks (M). Check: alarm switch
lights, emergency repair kit maintenance tools and masks, catwalks and rails, safety chains,
area lighting, tank chocks, safety signs, and padding air quick discharge (M).
Check oil level, air intake, drain condensate trap, separator (W). Tighten bolts (M). Check
pressure reducing valves (M). Change oil frame oil (Q). Clean cooling coil (Q). Inspect intake
air cleaner and valves (Q). Check-motor lubrication (A),
Lubricate fan bearings (M). Check V-belt tension (M). Check fan wheel (Q). Check fan motor
and bearings, V-belt alignment (SA).
Check purge rate and temperature, pressure and flow rate, filter pressure drop, solenoid
valves, cycle timer (M). Check outlet dew point and blow down relief valves (Q). Change air
filters (SA). Inspect exhaust mufflers (SA). Inspect dessicant, check valve seats, and
solenoid valves (A).
Remove organic residues with wood alcohol and inorganic residues with hydrochloric acid
(A).
Visual inspection (A).
Check motors (M).
Check motor condition and lubrication, pumps packings, and pump lubrication and condi-
tion (M).
Clean (M).
Check and clean sample intake lines, electrodes, reagent reservoir and sample pH (D).
Check calibration, sample flow, and liquid level (W). Clean constant head tank (W).
Clean and inspect vaporization chamber (A).
VI UXSt IIIW WU^JWIOIVSI *> Wl^ailUiiUM lop^^l vuf^isi ic-uil^ 11 v*i iui I lu^i \f—!/•
(D) •« daily; (W) =•= weekly; (M) = monthly; (Q) = quarterly; (SA) = semiannually; (A) = annually.
88
-------
6. Phelps, E.B. Disinfection of Sewage and
Sewage Filter Effluents with a Chapter on the
Putrescibility and Stability of Sewage Efflu-
ents. U.S. Geol. Surv. Water Supply Paper,
229, 1909.
7. Kellerman, K.E., et al. The Disinfection of
Sewage Effluents for the Protection of Public
Water Supplies. U.S. Dep. Agr., Bur. Plant Ind.,
Bulletin #115, 1907.
8. Rideal, S. Application of Electrolytic Chlorine
to Sewage Purification and Deodorization in
the Dry Chlorine Process. Trans. Faraday
Society 4:179-206, 1908.
9. Disinfection of Sewage and Sewage Filter
Effluent. Engineering Record 67:14, 1913.
10. Hooker, A.H. Chloride of Lime in Sanitation.
John Wiley & Sons, New York, NY, 1913.
11. Phelps, E.B. The Chemical Disinfection of
Sewage. American Journal Public Health,
2:72-86,1912.
12. Nikirk, F.A. Disinfection of Sewages and its
Success in a Small City. Municipal Engineer-
ing 46:478, 1914.
13. Faber, H.A. How Modern Chlorination Started
The Story of the Solution Feed Process as it
Began Forty Years Ago. Water and Sewage
Works 99:45-50, 1952.
14. Wigley, C.G. Disinfection of Sewage. Munic-
ipal Journal Public Works. 47:292-293,1919.
15. Chlorination of Sewage. Municipal Journal
46:266,1919
16. Editorial. Is Chlorination Effective Against all
Waterborne Disease? JAMA 78:283, 1922.
17. Ellms, J.W. and Pond, G.T. Sewage Disinfec-
tion. Municipal Sanitation, 1:266-268,1930.
18. Tiedeman, W.D. Efficiency of Chlorinating
Sewage Tank Effluent. Engineering News-
Record 98, 1927.
19. Porges, R. United States Sewage Treatment
Practices During the Early Twentieth Century.
Sewage and Industrial Wastes 39:13-21,
1957.
20. Laubusch, E.J. Safe Handling of Chlorine. In:
Chlorine: Its Manufacture, Properties and Use.
ACS Monograph #154, Reinhold Publ. Corp.,
New York, NY, 1962.
21. Miller, G.W., et al. An Assessment of Ozone
and Chlorine Dioxide for Treatment of Munic-
ipal Water Supplies. EPA 600/8-78-018, NTIS
No. PB-288196, U.S. Environmental Protec-
tion Agency, Cincinnati, OH, 1978.
22. Rapson, W.H. From Laboratory Curiosity to
Heavy Chemical. Chemistry Can. 18(1 ):25-31,
1966.
23. Aston, R.N. and J.F. Synan. Chlorine Dioxide
as a Bactericide In Waterworks Operation.
Journal New England Water Works Assoc.
62:80-94,1948.
24. White, G.C. Disinfection of Wastewater and
Water for Reuse. Van Nostrand Reinhold, New
York, NY, 1978.
25. McCarthy, J.A. Brand CIO as Water Disin-
fectants. Journal New England Water Works
Assoc. 58:55-68, 1944.
26. Wyss, O. and Stockton, J.R. The Germicidal
Action of Bromine. Arch. Biochem. 12:267,
1947.
27. Johannesson, J.K. Bromamines. Part I. Mono-
and Di-bromamine. Journal Chemical Society
2998-3001,1959.
28. Johannesson, J.K. Anomalous Bactericidal
Action of Bromamine. Nature 181:1799-1800,
1958.
29. Mattraw, H.C., et al. Equilibrium Constant,
Infrared Spectrum, andThermodynamic Prop-
erties of Bromine Chloride. Journal Chemistry
Phys. 22:1117-1119, 1954.
30. Mills, J.F. and Schneider, J.A. Bromine Chlo-
ride: An Alternative to Bromine. Ind. Eng.
Chem. Prod.Res. Devel. 12(3): 160-165,1973.
31. Mills, J.F. and Oakes, B.D. Bromine Chloride,
Less Corrosive Than Bromine. Chemical Eng.
Aug. 6, 1973, pp. 102-106.
32. Mills, J.F. Disinfection of Sewage by Chloro-
bromination. ACS Div Water Air Waste Chem-
istry 13(1): 106, 1973.
33. LeBlanc, N.E., et al. Disinfection Efficiency and
Relative Toxicity of Chlorine and Bromine
Chlorine in an Estuarine Environment. Special
report in Applied Marine Science and Ocean
Engineering #206, Virginia Inst. of Marine
Sci., 1978.
34. LeEUanc, N.E. and McEuen, J.M. A Compar-
ison of Bromine Chloride and Chlorine as
89
-------
Wastewater Disinfectants in Pilot Plant Stud-
ies. Water Chlorination, Environ. Impact and
Health Effects 2:637-650,1978.
35. Greene, D.J, Wastewater Disinfection with
Bromine Chloride. Presented at the Chesa-
peake Water Pollution Control Association/
Water and Waste Operators Association of
Maryland, Delaware and District of Columbia
Joint Annual Meeting, 1979.
36. Greene, D.J. An Alternative Wastewater Dis-
infectant. Water, July 1981.
37. Keswick, B.H. et al. Comparative Disinfection
Efficiency of Bromine Chloride and Bromine
for Poliovirus.JAWWA 70(10):573-577,1978.
38. Keswick, B.H., et al. Evaluation of Bromine
Chloride as n Disinfectant of Poliovirus Sus-
pended in Water Flowing Thru a Model Contact
Chamber. Presented at 19th Ann. ASM Mtg.,
1979.
39. Keswick, B.H. et al. Comparative Disinfection
by Bromine Chloride and Chlorine of Viruses
and Bacteria in Wastewater. JWPCF 52 (10):
2581-2518,1980.
40. Keswick, B.H. et al. Comparative Disinfection
of Poliovirus by Bromine Chloride and Chlorine
in a Model Contact Chamber. Water Research
16{1):89-94, 1982.
41. Ward, R.W., et al. Disinfection Efficiency and
Residual Toxlcity of Several Wastewater Dis-
infectants, Volume I. Grandville, Michigan.
EPA-600/2-76-156, NTIS No. PB-262245,
U.S. Environmental Protection Agency, Cin-
cinnati, OH, 1976.
42. Ward, R.W., et al. Disinfection Efficiency and
Residual Toxicity of Several Wastewater Dis-
infectants, Volume II. Wyoming, Michigan.
EPA-600/2-77-203, NTIS No. PB-274498,
U.S. Environmental Protection Agency, Cin-
cinnati, OH, 1977.
43. Laubusch, E.J. Chlorine. Public Works 94:
511-521,1963.
44. Chlorine Institute, Properties of Chlorine in SI
Units. The Chlorine Institute, New York, NY,
1981.
45. Laubusch, E.J. Hypochlorites. Public Works
94{6}:106-109,192, 1963
46. Baker, R.J. Characteristics of Chlorine Com-
pounds. JWPCF 41:482, 1969.
47. Morris, J.C. The Mechanism of the Hydrolysis
of Chlorine. Journal American Chemical Soci-
ety 68, 1946.
48. Morris, J.C. The Acid lonization Constant of
HOCI From 5 C to 35 C. Journal Phys.
Chemistry 70, 1966.
49. Downs, A.J. and Adams, C.J. The Chemistry of
Chlorine, Bromine, Iodine and Astatine. Per-
gamon Press, Oxford, 1973.
50. Noack, M.G. and Doerr, R.L. Chlorine Dioxide,
Chlorous Acid and Chlorites. Kirk-Othmer
Encyclopedia of Chemical Technology, 3rd ed.,
HF Mark et al., eds., 5:612-632, 1979.
51. Masschelein,W.J. Chlorine Dioxide. Ann Arbor
Science. Ann Arbor, Ml, 1979.
52. Equilibrium Solubility of Chlorine Dioxide in
Water. Tech. Info Sheet #2. Tappi 44(8): 160A,
1961.
53. Gordon, G., et al. The Chemistry of Chlorine
Dioxide. Progress in Inorganic Chemistry
15:201, 1972.
54. Granstrom, M.L and Lee, G.F. Rates and
Mechanisms of Reactions Involving Oxychloro
Compounds. Public Works 88(12):90-92,
1957.
55. Medir, M. and Giralt, F. Stability of Chlorine
Dioxide in Aqueous Solution. Water Research
16:137-139, 1982.
56. Dow Chemical Co. Bromine Chloride Hand-
book, 1979.
57. Mills, J.F. Chemistry and Disinfectant Proper-
ties of Bromine Chloride. In: A.J. Rubin, ed.
Chemistry of Wastewater Technology. Ann
Arbor Science, Ann Arbor, Ml, 1978.
58. Cole, L.G. and Elverum, G.W. Thermodynamic
Properties of the Diatonic Halogens from Spec-
troscopic Data. Journal of Chemical Physics
20, 1952.
59. Greenwood, N.N. Physicochemical Properties
of the Interhalogen Compounds. Revs. Pure
Applied Chem, Australia 1, 1951.
60. Mills, J.F. Interhalogens and Halogen Mix-
tures as Disinfectants. In: Johnson, ed., Dis-
infection: Water and Wastewater. pp. 113-
114, Ann Arbor Science, Ann Arbor, Ml, 1975.
61. Bennett, J.E. On-Site Generation of Hypo-
chlorite Solutions by Electrolysis of Seawater.
AlChE Symp. Ser. 74(178):265-269, 1978.
90
-------
62. Michalek, S.A. and Leitz, F.B. On Site Genera-
tion of Hypochlorite. JWPCF44(9):1697-1712,
1972.
63. Wojtowicz, J.A. Chlorine Monoxide, Hypochlor-
ous Acid and Hypochlorites. Kirk-Othmer
Encyclopedia of Chemical Technology, 3rd Ed.,
H.F. Mark et al., eds., 5:580-611, 1979.
64. Weber, G.R., et al. Effect of Ammonia on the
Germicidal Efficiency of Chlorine in Neutral
Solutions. JAWWA 32:1904-19I2, 1940.
65. Taras, M.J. Effect of Free Residual Chlorina-
tion on Nitrogen Compounds in Water.
JAWWA 45:47-61, 1953.
66. Enslow, L.H. Ammonia Chlorine Reactions
and Chloramine Production in Water and
Sewage Treatment. Contract Record and Eng.
Rev. 45:291-295, 1931.
67. Griffin, A.E. and Chamberlin, N.S. Some
Chemical Aspects of Breakpoint Chlorination.
Journal New Eng. Wat. Wks. Assn. 55(3),
1941.
68. Griffin, A.E. and Chamberlin, N.S. Relation of
Ammonia Nitrogen to Breakpoint Chlorination.
American Journal Public Health. 31:803-808,
1941.
69. Griffin, A.E. and Chamberlin, N.S. Exploring
the Effect of Heavy Doses of Chlorine in
Sewage. Sewage Works Journal 17:730-742,
1945.
70. McKee, J.E., et al. Chemical and Colicidal
Effects of Halogens in Sewage. JWPCF 32:
795-819,1960.
71. Symons, G.E., et al. Reduction of Bacteria in
Sewage by Chlorination. Water Works and
Sewerage 85:983-988, 1938.
72. Symons, G.E., et al. Variation in the Chlorine
Demand of Buffalo Sewage. Sewage Works
Journal 13:249-264, 1941.
73. Taras, M.J. Preliminary Studies on the Chlor-
ine Demand of Specific Chemical Compounds.
JAWWA 42:462-474, 1950.
74. Feben, D. and Taras, M.T. Studies on Chlorine
Demand Constants. JAWWA 43(9):22-31,
1951.
75. Lin, S. and Evans, R.L. Chlorine Demand Study
of Secondary Sewage Effluents. Water and
Sewage Works 121(1 ):35, 1974.
76. Haas, C.N., and Karra, S.B. Kinetics of Waste-
water Chlorine Demand Exertion. JWPCF
56(2):170-173, 1984.
77. White, G.C., et al. Problems of Disinfecting
Nitrified Effluents. Presented at the ASCE
Nat'l Conf. on Environ. Eng., 1981.
78. Weil, I. and Morris, J.C. Equilibrium Studies on
N-Chloro Compounds. Journal American
_ Chemical Society 71:312-313, 1949.
79. Gray, E.T., etal. Chloramine Equilibria and the
Kinetics of Disproportionation in Aqueous
Solution. ACS Symp. Ser. #82: Organbmetals
and Organometal loids, Occurrence and Fate
in the Environment, 1978.
80. White, G.C. Handbook of Chlorination. Van
Nostrand Reinhold Co., New York, NY, 1972.
8.1. Morris, J.C. and Isaac, R.A. A Critical Review
of Kinetics and Thermodynamic Constants for
the Aqueous Chlorine-Ammonia Systems.
Water Chlorination, Environmental Impact
and Health Effects 4(1 ):49, 1983.
82. Morris, J.C. Kinetics of Reactions Between
Aqueous Chlorine and Nitrogen Compounds.
In: S.D.Faust, ed. Principles and Application of
Water Chemistry, John Wiley, New York, NY,
pp. 22-53, 1967.
83. Saunier, B.M. and Selleck, R.E. The Kinetics of
Breakpoint Chlorination in Continuous Flow
Systems. JAWWA 71:164-172, 1979.
84. Trofe, T.W., et al. Kinetics of Monochloramine
Decomposition in the Presence of Bromine.
Environ. Sci. and Tech. 14:544, 1980.
85. Saguinsin, J.L.S. and Morris, J.C. The Chem-
istry of Aqueous Nitrogen Trichloride. In: J.D.
Johnson, ed. Disinfection: Water and Waste-
water, pp. 277-299, Ann Arbor Science, Ann
Arbor, Ml, 1975.
86. Baker, R.W.R. Studies on the Reaction Be-
tween Sodium Hypochlorite and Proteins. I.
Physicochemical Study of the Course of the
Reaction. Biochemical Journal 41:337-342,
1947.
87. Wright, N.C. The Action of Hypochlorites on
Amino Acids and Proteins. The Effect of Acidity
& Alkalinity. Biochemical Journal 30:1661-
1667,1936.
88. Isaac, R.A. and Morris, J.C. Rates of Transfer
of Active Chlorine Between Nitrogenous Sub-
stances. Water Chlorination: Environ. Impact
and Health Effects 3, 1980.
91
-------
89. Cachaza, J.M. Kinetics of Oxidation of Nitrite
by Hypochlorite in Aqueous Basic Solution.
Can. Journal Chemistry 54:3401 -3406,1976.
90. Murphy, K.L., et al. Effect of Chlorination
Practice on Soluble Organics. Water Research
9:389,1975.
91. De Laat, J., et al. Chloration de composes
organizes: Dernande on chlore et reactivite vis
a vis de la Furmation des trihalometha Inci-
dence de I'azole ammoniacal. Water Research
16:143-147,1982.
92. Jolley, R.L Chlorine Containing Organic Con-
stituents in Chlorinated Effluents. JWPCF
47:601-618,1975.
93. Chow, B.M. and Roberts, P.V. Halogenated
Byproduct Formation by Chlorine Dioxide and
Chlorine. Journal of the Environmental Engi-
neering Division, ASCE107(4):609-618,1981.
94. Glaze, W.H. and Henderson, J.E. Formation of
Organochlorine Compound from the Chlorina-
tion of a Municipal Secondary Effluent. JWPCF
47:2511,1975.
95. LaPointe, T.F., et al. Kinetics of Tribromamine
Decomposition. In: J.D. Johnson, ed. Disinfec-
tion: Water and Wastewater, pp. 301-338,
Ann Arbor Science, Ann Arbor, 1975.
96. Wajon, J.E. arid Morris, J.C. Bromamination
Chemistry: Rates of Formation of NH Br and
Some N-Bromamino Acids. Water Chlorina-
tion: Environ. Impact and Health Effects
3:171-181,1980.
97. Johnson, J.D. and Overby, R. Bromine and
Bromamine Disinfection Chemistry. Journal
of the Sanitary Engineering Div., ASCE97
(SA5):617,1971.
98. Cromer, J.L., et al. Dibromamine Decomposi-
tion Kinetics. In: A.J. Rubin, ed, Chemistry of
Wastewater Technology, Ann Arbor Science,
Ann Arbor, 1978.
99. Inman, G.W., et al. Kinetics of Nitrogen
Tribromide Decomposition in Aqueous Solu-
tion. Inorg. Chemistry 15:3037-3042, 1976.
100. Roberts, PA/., et al. Chlorine Dioxide for
Wastewater Disinfection: A Feasibility Evalua-
tion. EPA-600/2-81-092, NTIS No. PB81-
213357, U.S. Environmental Protection Agen-
cy, Cincinnati, OH, 1981.
101. Ingols, R.S. and Ridenour, G.M. Chemical
Properties of Chlorine Dioxide in Water Treat-
ment. JAWWA 40:1207-1227, 1948.
102. Sikorowska, C. Influence of Pollutions on
Chlorine Dioxide Demand of Water. Gaz.
Woda. Tech. Sanit., 35, 12, 4645, 1961.
103. Somsen, R..A. Oxidation of Some Simple
Organic Molecules with Aqueous Chlorine
Dioxide Molecules. I. Kinetics. Tappi 43:154-
156,1960. II Reaction Products, Tappi 43:157-
160, 1960.
104. Wajon, J.E., et al. Oxidation of Phenol and
Hydroquinane by Chlorine Dioxide. Environ.
Scj. and Tech. 16:396, 1982.
105. Snoeyink, V.L. and Suidan, M.T. Dechlorina-
tion by Activated Carbon and Other Dechlor-
inating Agents. In: J.D. Johnson, ed., Disinfec-
tion: Water and Wastewater. Ann Arbor
Science, Ann Arbor, Ml, 1975.
106. Laubusch, E.J. Sulfur Dioxide. Public Works
94(8):117-121, 1963.
107. Scott, W.D. and Hobbs, P.V. The Formation of
Sulfate in Water Droplets. Jour. Atmos. Sci.,
24,54,1967.
108. Lister, M.W. and Rosenblum, P. Rates of
Reaction of Hypochlorite Ions with Sulfite and
Iodine Ions. Can. Journal Chemistry 41(12):
3013-3020,1963.
109. Srivastava, R.D., et al. Kinetics of Reaction of
Sodium Hypochlorite and Sodium Sulfide by
Flow Thermal Method. Ind. Eng. Chem. Fund
19:207-209,1980.
110. Stanbro, W.D. and Lenkevich, M.J. Slowly
Dechlorinated Organic Chloramines. Science
215:967-968,1982.
111. Standard Methods for the Examination of
Water and Wastewater, I6th edition. APHA,
AWWA and WPCF, Washington, DC, 1985.
112. Chen, C. and Gan, H.B. Wastewater Dechlor-
ination: State of the Art Field Survey and Pilot
Studies. USEPA Contracts 14-12-150 and 68-
03-2745, Final report, 1981.
113. Puri, B.R., et al. Interaction of Charcoal with
Chlorine Water. Journal Indian Chemical
Society 35:181-184, 1958.
114. Puri, B.R., et al. Interaction of Charcoal with
Chlorine Water II. Journal Indian Chemical
Society 37:171-176, 1960.
115. Bauer, R.C. and Snoeyink, V.L. Reactions of
Chloramines with Active Carbon. JWPCF
45(11 ):2290-2301, 1973.
92
-------
116. Kim, B.R. and Snoeyink, V.L. The Monchlor-
amine GAC Reaction in Adsorption Systems.
JAWWA 72(8):488-490, 1980.
117. Kim, B.R., et al. Removal of Dichloramine and
Ammonia by Granular Activated Carbon.
JWPCF 50(1), 1978.
118. Kim, B.R., et al. Analysis of Models for
Dichloramine Removal by Activated Carbon in
Batch and Packed Bed Reactors Using Quasi-
linearization and Orthogonal Collocation Meth-
ods. Water Research 12(5):317-326, 1978.
119. Marks, H.C., et al. Amperometric Methods in
the Control of Water Chlorination. Analytical
Chemistry 19:200-204, 1947.
120. Marks, H.C., et al. Amperometric Titration of
Residual Chlorine in Sewage. Water and
Sewage Works, May, 1948.
121. Marks, H-C. and Joiner, R.R. Determination of
Residual Chlorine in Sewage. Analytical
Chemistry 20:1197,1948.
122. Heukelekian, H., et al. Disinfection of Wastes
by Chlorination. Ind. Eng. Chem. 45:1004-
1008, 1953.
123. Day, R.V., et al. Residual Chlorine Methods
and Disinfection of Sewage. Ind. Eng. Chem.
45:1001-1004,1953.
124. Feng, T.H. Behavior of Organic Chloramines in
Disinfection. JWPCF 38:614-628, 1966.
125. Henderson, W.L Control of Sewage Chlor-
ination by Amperometric Titration. Sewage
and Industrial Wastes 24:1467-1472, 1952,
126. Heukelekian, H. and Day, R.V. Disinfection of
Sewage with Chlorine. III. Factors Affecting
Coliform Remaining and Correlation of Ortho-
. lidine and Amperometric Chlorine Residuals.
Sewage and Industrial Wastes 23:155-163,
1951.
127. Browning, G.E. and McLaren, F.R. Experiences
with Wastewater Disinfection in California.
JWPCF 39(8):1351-1361, 1967. "
128. Collins, H.F. and Deaner, D.G. Sewage Chlor-
ination Versus Toxicity, Dilemma. Journal of
the Environmental Engineering Division, ASCE
99(SA6):761-772, 1973.
129. Lin, S., et al. Selection of a Method to
Determine Residual Chlorine in Sewage Ef-
fluents. Water and Sewage Works 118(11):
360-364, 1971.
130. Jenkins, R.L. and Baird, R.B. Determination of
Total Chlorine Residuals in Treated Waste-
waters by Electrode. Analytical Letters 12(A2):
125-141,1979.
131. Liebermann, J., et al. Development of the
FACTS Procedure for Combined Forms of
Chlorine and Ozone in Aqueous Solution.
Environ. Sci. and Technol. 14:1395-1400,
1980.
132. Cooper, W.J., et al. Comparison of Several
Instrumental Methods for Determining Chlo-
rine Residuals in Drinking Water. JAWWA
74(10):546-552, 1982
133. Knechtel,J.R.,etal. Determination of Chlorine
Dioxide in Sewage Effluents. Analytical Chem-
istry 50(2):202-205, 1978.
134. Wheeler, G.L., et al. A Rapid Microdetermina-
tion of Chlorine Dioxide in the Presence of
Active Chlorine Compounds. Microchemical
Journal 23:160-164, 1978.
135. Olivier!, V.P., et al. Inactivation of Virus in
Sewage. Journal of the Sanitary Engineering
Div., ASCE97(SA5):661, 1971.
136. Heukelekian, H. and Day, R.V. Disinfection of
Sewage with Chlorine. II. Method and Uni-
formity of Distribution of Chlorine in Sewage
Plants. Sewage and Industrial Wastes 23:47-
53,1951,
137. Selleck, R.E., Collins, H, and White, G.C.
Kinetics of Bacterial Deactivation with Chlo-
rine. Journal of the Environmental Engineer-
ing Division, ASCE 104:1197-1212,1978
138. Collins, H.F., et al. Problems in Obtaining
Adequate Sewage Disinfection. Journal of the
Sanitary Engineering Division, ASCE 97(SA5):
549-562, 1971.
139. Stagg, C.H., et al. Chlorination of Solids
Associated Coliphages. Progr. Water Technol.
10(12):381-387, 1978.
140. Hejkal, T.W., et al. Survival of Polioviruses
Within Organic Solids During Chlorination.
Appl. Environ. Microbiol. 38:114, 1979.
141. Butterfield, C.T., et al. Influence of pH and
Temperature on the Survival of Conforms and
Enteric Pathogens When Exposed to Free
Chlorine. U.S. Pub. Health Rep. 58:1837-1866,
1943.
142. Ozgumus, N.H. Evaluation of Chemical and
Bacteriological Methods of Determining Ger-
93
-------
micidal Activity of Chlorine. Ph.D. Thesis,
Mich. State Coll., Lansing, Ml, 1950.
143. Rudolph, A.S. and Levine, M. Factors Affecting
the Germicidal Efficiency of Hypochlorite
Solutions. Iowa State College Eng. Exp. Sta.
Bulletin 150, 1941.
144. Floyd, R., et al. Inactivation by Chlorine of
Single Poliovirus Particles in Water. Environ-
mental Science and Technology 13:438-442,
1979.
145. Sharp, D.G. and Leong, J. Inactivation of
Poliovirus I, Brunhilde Single Particles by
Chlorine in Water. Appl. Environ. Microbiol.
40:381 -385,1980.
146. Chang, S.L Resistance of Pathogenic Naegler-
ia to Some Common Physical and Chemical
Agents. Appl. Environ. Microbiol. 38:36-38,
1978.
147. Severin, B.F. Inactivation of Proposed Indicator
Organisms by Inorganic Chloramines. M.S.
Special Problem Report, Dept. of Civil Eng.,
Univ. of Illinois at Urbana, 1975.
148. Butterfield, C.T., et al. Influence of pH and
Temperature on the Survival of Conforms and
Enteric Pathogens When Exposed to Chlorine.
U.S. Pub. Health Rep. 61:157-192,1946.
149. Scott, W.J., et al. Recommended Procedures
in the Use of Chlorine at Water and Sewage
Plants. JAWWA 45:1060-1074,1953.
150. Fiedler, R.P. On-site Caustic Chlorine Genera-
tion for Water Disinfection. JAWWA 66(1 ):46-
49,1974.
151. Doan, R. and A.S. Haimes, Hypochlorite Gen-
eration at a Water Pollution Control Plant.
Public Works 109(1):48-51, 1978.
152. Baur, F. Contract Granted for Sodium Hypo-
chlorite Generator. Water and Sewage Works
119:76, 1972.
153. Benas, B. Bulk Storage of Liquid Chlorine at
San Francisco Sewage Plants. Water and
Sewage Works 100:381 -385,1953.
154. Chlorine Institute, Chlorine Pipelines. Pam-
phlet #60,2nd ed. The Chlorine Institute, New
York, NY, 1982.
155. Cameron, G.W. Chlorination-Plant Problems
at City Island Plant of the Atlantic City Sewage
Company. Sewage Works Journal 19:603-
606. 1947.
156. Chlorine Institute, Chlorine Vaporizing Equip-
ment. Pamphlet #9, ed. 2, rev. 2. The Chlorine
Institute, New York, NY, 1979.
157. Sepp, E. and White, G.C. Manual for Waste-
water Chlorination and Dechlorination Prac-
tices. California Department of Health Serv-
ices, 1981.
158. Cbnnell, G. F. and Fetch, JJ. Advances in
Handling Gas Chlorine. JWPCF 41:150-155,
1969.
159. Chemical Aids Manual for Wastewater Treat-
ment. EPA 430/9-79-018, U.S. Environmental
Protection Agency, Washington, DC, 1979
160. Jordan, R.W. Improved Method Generates
More Chlorine Dioxide. Water and Sewage
Works 127(10):44, 1980.
161. Longley, K.E. Turbulence Factors in Chlorine
Disinfection of Wastewater. Wat. Res. 12:
813-822, 1978.
162. Chow, V.T. Open-Channel Hydraulics. Mc-
Graw Hill Book Co., New York, NY, 1959.
163. Mandt, M. Jet Disinfection Improves Contact
Chamber Performance, Cuts Costs. Water and
Sewage Works 125:447-453,1978.
164. LaGrosa, J.J., et al. Wastewater Chlorination
by Jet Diffusion. April 1978.
165. Huang, J.Y.C., Warriner, R. andNi, N.S.N. Pilot
Tests of Chlorination Facility for Disinfecting
Secondary Effluent JWPCF 57(7):777-784,
1985.
166. Trussell, R.R. and Chao, J. Rational Design of
Chlorine Contact Facilities. JWPCF 49(4):659-
667, 1977.
167. Hart, F.L. Modifications for the Chlorine Con-
tact Chamber. Jour. New Engl. Water Pollut.
Control Assoc. 13(2): 135-151, 1979.
168. Hart, F.L. Improved Hydraulic Performance of
Chlorine Contact Chamber. JWPCF 51(12):
2868-2875,1979.
169. Hart, F.L. and Vogiatzis, Z. Performance of
Modified Chlorine Contact Chamber. Journal
of the Environmental Engineering Division,
ASCE 108:549, 1982.
170. Metcalf & Eddy. Wastewater Engineering, 2nd
ed. McGraw Hill, New York, NY, 1979.
94
-------
171. Fair, G.M., etal. Elements of Water Supply and
Wastewater Disposal. John Wiley & Sons,
New York, NY, 1971.
172. Louie, D.S. and Fohrman, M.S. Hydraulic
Model Studies of Chlorine Mixing and Contact
Chambers. JWPCF 40:174, 1968.
173. Kothandaraman, V. and Evans, R.L Hydraulic
Model Studies of Chlorine Contact Tanks.
JWPCF 44(4}:625-633, 1972.
174. Marske, D.M. and Boyle, J.D. Chlorine Contact
Chamber Design-A Field Design Evaluation.
Water and Sewage Works 70(1), 1973.
175. Roop, R.N. Evaluation of Residual Chlorine
Control Systems. JWPCF 49:1591-1603,
1977.
176. White, G.C. Chlorination and Dechlorination,
A Scientific and Practical Approach. JAWWA
60:540,1968.
177. Finger, R.E., et al. Development of an On-Line
Zero Chlorine Residual Measurement and
Control System. Presented at the 56th Annual
WPCF Conference, 1983.
178. Aieta, E.M. and Roberts, P.V. Disinfection with
Chlorine and Chlorine Dioxide. Journal of the
Environmental Engineering Division, ASCE
109:783-799,1983.
179. Geisser, D.F., et al. Design Optimization of
High-Rate Disinfection Using Chlorine and
Chlorine Dioxide. JWPCF 51(2):351-357,
1979.
180. Gumerman, R.C., et al. Estimating Water
Treatment Costs. EPA-600/2-79-162, NTIS
No. PB80-139819, U.S. Environmental Pro-
tection Agency, Cincinnati, OH, 1979.
181. Manual of Practice No. 4. Water Pollution
Control Federation, Washington, DC, 1976.
182. Gulp, G.L and Heim, N.F. Field Manual for
Performance Evaluation and Trouble Shooting
at Municipal Wastewater Treatment Plants.
EPA-430/9-78-001, U.S. Environmental Pro-
tection Agency, Washington, DC, 1978.
95
-------
-------
Chapter 6
Ozone Disinfection
6.1 Introduction
6.1.1 General
The planning, design, construction and operation of a
wastewater treatment process involves at least five
separate parties: owner, design engineer, contractor,
equipment manufacturer, and regulatory agency.
Because the use of ozone for wastewater disinfection
is relatively new, the responsibility for development
of process and equipment design has been largely the
equipment manufacturer's. However, many design
features are not related to a given manufacturer's
piece of equipment, like process flexibility considera-
tions, contact basin size and configuration, and
influent water quality characteristics. The informa-
tion presented in this manual is intended to provide
necessary design capability and understanding to all
parties so that each party can be properly involved in
the development of an ozone disinfection system.
During the development of this design manual, site
visits to seven operating ozone installations were
conducted in order to obtain information on operating
experiences at existing facilities. The information
obtained from these site visits is incorporated in the
manual.
6.1.2 History of Ozone
A history of ozone has been documented by others
(1,2). This work was based on surveys of operating
ozone installations in Europe and the United States
and is briefly described.
Experiments conducted in 1886 showed that ozon-
ized air "will effect the sterilization of polluted
water." The first drinking water plant to use ozone
was built in 1893 at Oudshoorn, Holland. The French
studied the Oudshoorn plant, and after pilot testing
constructed an ozone water plant at Nice, France in
1906. Because ozone has been used at Nice since
that time, Nice is often referred to as "the birthplace
of ozonation for drinking water treatment."
In Europe there is a strong commitment to attain a
water of the highest chemical quality. Currently,
there are over 1,000 European drinking water plants
that use ozone at one or more points in the treatment
process.
In contrast to the widespread use of ozone for water
treatment in Europe, very few European wastewater
ozone disinfection systems exist. Currently, there are
more ozone disinfection systems in use at U.S.
wastewater plants than at U.S. water plants or
European wastewater plants.
The first U.S. wastewater plant to use ozone for
disinfection was Indiantown, Florida, which began
operation in 1975 (3). By 1980 about 10 wastewater
treatment plants using ozone for disinfection had
been constructed. Unfortunately, some of these
earlier ozone disinfection facilities have chosen to
abandon ozone disinfection for one or more of the
following reasons: excessive high cost of operation,
equipment problems, excessive maintenance cost,
and inability to attain performance objectives without
major modifications (4). Despite these early set-
backs, many facilities have proposed to utilize ozone
for wastewater. A list of wastewater plants that have
used or are using ozone disinfection is presented in
Table 6-1 (5).
6.2 Ozone Properties, Chemistry and
Terminology
Ozone is a molecule that can co-exist with air or high
purity oxygen, or can dissolve in water. It is a very
strong oxidizing agent and a very effective disinfect-
ant.
6.2.1 Ozone Properties
Ozone (Oa) is an unstable gas that is produced when
oxygen molecules are dissociated into atomic oxygen
and subsequently collide with another oxygen mole-
cule (6). The energy source for dissociating the
oxygen molecule can be produced commercially or
can occur naturally. Some natural sources for ozone
production are ultraviolet light from the sun and
lightning during a thunderstorm.
Ozone may be produced by electrolysis, photochem-
ical reaction, radiochemical reaction, or by "electric
discharge" in a ga;s that contains oxygen (7). The
electric discharge principle has been used in most
commercial applications and in all known water and
wastewater treatment applications. The electric
discharge method is presented in this manual.
97
-------
Table 6-1 . U.S. Municipal Wastewatar Treatment Plants Using Ozone (5)
Primary
Feed Purpose of Startup
Location Gas Ozone Date
Indiantown, FL
Woodlands, TX
Upper Thompson Sani-
tation District, CO
Hunter, NY
Harriman, NY
Chino Basin, CA
Palo Alto, CA
Collegeville, MN
Mahoning County, OH*
Hunter Highlands, NY
Cotter Gasville, AR
Springfield, MO*
Bull Shoals, AR
Sebring, FL
Maryland City, MD
Oak Ridge, NY
Norton AFB, CA
Carmel, NY
Potomac Heights, MD
Murphreesboro, TN*
Pensacola, FL*
Hercules, CA
Marion, NY
Brookings, SD
Concord, NC*
Delaware County, OH
Frankfort, KY
Ocean City, MD*
Madisonville, KY*
Little Valley, NY
Yaphank, NY
Granby, CO
Rocky Mount, NC*
West Knoxville, TN
Vail, CO
Hagerstown, MD*
Olympia, WA*
Indianapolis, IN
Belmont Plant*
Southport Plant*
Twining, NM
Somero, NY
Alburndale, FL
air
air
air
air
air
air
air
air
02
air
air
02
air
air
air
air
air
air
air
02
02
air
air
air
02
air
air
02
02
air
air
air
02
air
air
02
02
02
02
air
air
air
disinfection
disinfection
disinfection
disinfection
disinfection
suspended solids
organics
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
phenols
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
&BOD
flotation
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection
disinfection ,
disinfection
disinfection
disinfection
disinfection
disinfection
1975
1976
1977
1977
1977a
1978
1978
1978
1978
1978
1978a
1978a
—
—
1980a
1980a
1980"
1980a
1980a
1980
1980a
1980a
1980a
1980a
1980a
1980a
1980a
1980a
1981
1981
1981
1982a
1982
19823
1983
1983
1983
1983
1983
1983
1984
Average
mgd
0.5
1.5
1.5
0.1
0.36
5
4
0.22
4
<1
1
30
0.19
0.52
0.58
0.12
0.25
1
0.2
8
20
0.4
0.125
6
25
1.5
7
12
5
0.28
0.12
<1
40
2
2.7
8
'14
120
125
0.095
0.1
2
Flow Rate
mVd
1,890
5,680
5,680
368
1,360
18,930
15,100
830
15,100
<3,780
3,780
113,600
720
1,960
2,200
450
950
3,780
760
30,280
75,700
1,510
470
22,700
94,600
5,680
26,500
45,400
18,900
1,060
450
<3,680
151,400
7,570
10,200
30,300
53,000
454,200
473,200
360
368
7,570
'Plants operating per 1982 EPA survey
•Plants using oxygen activated sludge process
Ozone is used in relatively low concentrations in
water and wastewater treatment applications. The
properties of pure ozone are presented for general
background information in Table 6-2 (7).
, i
6.2.1.1 Ozone Color and Odor
At ordinary temperatures ozone is a blue gas, but at
typical concentrations its color is not noticeable
unless it is viewed through considerable depth (7).
Ozone has a very distinct odor and it owes its name to
its odor. The word ozone is derived from the Greek
word 'ozein', which means to smell. Ozone can be
detected at concentrations of only 2x10"5 to 1x10~4
g/m3(i.e.,0.01 toO.OSppm by volume) (1). The ability
to smell ozone at small concentrations is considered a
safety feature because an ozone odor can be detected
before ozone related health considerations develop.
The present allowable 8-hour exposure concentra-
98
-------
Table 6-2. Properties of Pure Ozone (O3) (7)
melting point, °C
boiling point, °C
critical temperature, °C
critical pressure, atm
critical volume, cm3/mole
density and vapor pressure of liquid
Temperature, °C
-183
-180
-170
-160
-150
-140
-130
-120
-110
-100
density of solid ozone, g/cm3, at 77.4°K
viscosity of liquid, cP3, at 77.6°K
at 90.2°K
surface tension, dyn-cm, at 77.2°K
at 90.2°K
parachlor* at 90.2°K
dielectric constant, liquid, at 90.2°K
dipole moment, debye
magnetic susceptibility, cgs units, gas
liquid
heat capacity of liquid from 90 to 150°K
heat of vaporization, kcal/mole, at -111.9°C
at -183°C
heat and free energy of formation
gas at 298.15°K
liquid at 90.15°K
hypothetical gas at 0°K
Density, g/cm3
1.574
1.566
1.535
1.504 .
1.473
1.442
1.410
1.378
1.347
1.316
-192.5 ± 0.4
-111.9±0.3
-12.1
54.6
111
Vapor Pressure, Torr
0.11
0.21
1.41
6.73
24.8
74.2
190
427
865
1605
1.728
4.17
1.56
43.8
38.4
75.7
4.70
0.55
0.002 x 10~6
0.150
CP = 0.425 + 0.0014 (T-90)
3410
3650
AHf, kcal/mole
34.15
30.0
34.74
AGf, kcal/mole
38.89
*M,1/4 (D - d) where M = molecular weight; T = surface tension; D = liquid density; d = vapor density
tionis2x10~5g/m3(i.e.,0.1 ppm by volume), which is Figure 6-1.
2 to 10 times higher than the concentration at which
ozone can be smelled. Additional safety considera-
tions for ozone are discussed in Section 6.6
6.2.1.2 Ozone Stability
The stability of ozone is greater in air than in water but
is not excessively long in either case. The half-life of
residual ozone in water is reported to range from 8
minutes to 14 hours depending on the phosphate and
carbonate concentration of the water (8). With no
phosphates or carbonates and the water adjusted to I
pH 7.0 with sodium hydroxide, the half-life was 8 I
minutes. Hoigne and Bader (9) found that ozone will \
react directly with solutes in the water, and that ^.
hydroxide ions and hydroxyl radicals will provide a N
catalyst for the decomposition of ozone into inter- \
mediate compounds that are also reactive, such as
peroxide ions and hydroxide ions. Their findings are
summarized in Figure 6-1.
The direct reaction of ozone with solutes (Mi,
Ma) and a hydroxide ion (or radical) catalyzed
decomposition reaction, leading to reactive
intermediates, compete for ozone (9).
03-
t
i
H02-
-------
Hoigne and Bader's results suggest that ozone
disinfection is influenced by raw water chemistry
characteristics, in addition to the more well known
influences of wastewater pollutants. The water
chemistry influences are generally not utilized in
ozone system design, except as developed in pilot
plant studies of the specific wastewater to be
disinfected. However, the water chemistry influences
are important to keep in mind when a comparison is
made of ozone disinfection performance at different
facilities.
The residual ozone concentration in water is de-
creased rapidly by aeration or agitation of the liquid
(7). A wastewater sample with residual ozone must
not be agitated or collected after a period of agitation
has occurred, or the measurement of residual ozone
will be inaccurate. Also, because ozone may be
released from the water, a plant effluent that contains
a high residual ozone concentration that is used as a
source of non-potable water, may release excessive
amounts of ozone and contaminate the ambient
environment.
The stability of ozone in air or oxygen is significantly
affected by the temperature of the gas. In a clean
vessel at room temperature the half-life of ozone may
range from 20 to 100 hours (7). At 120°C (248°F) the
half-life is only 11 to 112 minutes and at 250°C
(482°F) only 0.04 to 0.4 seconds. This characteristic
of ozone is important for design because cooling of
the ozone generators is necessary. Also, good room
ventilation is necessary in case an ozone leak occurs,
and ozone contained in the off-gas must be destroyed.
6.2.1.3 Ozone Physical Characteristics
Gaseous ozone is explosiveat an ozone concentration
of 240 g/m3{20 percent wt in air) (7). Fortunately, the
maximum gaseous ozone concentration typically
found in water or wastewater ozone disinfection
systems does not exceed 50 g/m3 (4.1 percent wt in
air). If, however, a medium that can adsorb and
concentrate ozone is inappropriately located in the
system, then explosive ozone concentrations could
develop.
concentrations has been developed, as shown in
Table 6-3.
H=Y/X
(6-1)
where:
Y = partial pressure of the gas above the liquid
atmospheres
X = molar fraction of the gas in the liquid at
equilibrium with the gas above the liquid
H = Henry's law constant (varies with tempera-
ture), atm/mole fraction
Table 6-3. Solubility of Ozone in Water
Water
Temperature
(°C)
0
5
10
15
20
25
30
0
5
10
15
20
25
30
0
5
10
15
20
25
30
0
5
10
15
20
25
30
Henry's
Constant
Atm/Mole
1,940
2,180
2,480
2,880
3,760
4,570
5,980
1,940
2,180
2,480
2,880
3,760
4,570
5,980
1,940
2,180
2,480
2,880
3,760
4,570
5,980
1,940
2,180
2,480
2,880
3,760
4,570
5,980
Ozone
Concentration
mg/l ppm-vol
12.07
12.07
12.07
12.07
12.07
12.07
12.07
18.11
18.11
18.11
18.11
18.11
18.11
18.11
24.14
24.14
24.14
24.14
24.14
24.14
24.14
36.21
36.21
36.21
36.21
36.21
36.21
36.21
6,044
6,044
6,044
6,044
6,044
6,044
6,044
9,069
9,069
9,069
9,069
9,069
9,069
9,069
12,088
12,088
12,088
12,088
12,088
12,088
12,088
18,132
18,132
18,132
18,132
18,132
18,132
18,132
Ozone
Solubility
mg/l
8.31
7.39
6.50
5.60
4.29
3.53
2.70
12.47
11.09
9.75
8.40
6.43
5.29
4.04
16.62
14.79
13.00
11.19
8.57
7.05
5.39
24.92
22.18
19.50
16.79
12.86
10.58
8.09
Note: The concentration of the ozone gas is determined at a
standard temperature of 68°F (20°C) and a standard pres-
sure of 1 atmosphere (101 kPa).
6.2.1.4 Ozone Solubility
Ozone solubility in water is important because ozone
disinfection is dependent upon the amount of ozone
transferred to the wasnewater. Henry's law relative to
ozone systems states that the mass of ozone that will
dissolve in a given volume of water, at constant
temperature, is directly proportional to the partial
pressure of the ozone gas above the water (10). Using
Equation 6-1 the maximum solubility of ozone in
water at various temperatures and feed-gas ozone
6.2.2 Ozone Chemical Reactions
Ozone is a very strong oxidizing agent, having an
oxidation potential of 2.07 volts (1). Ozone will react
with many organic and inorganic compounds in the
wastewater. These reactions are typically called
"ozone demand" reactions. They are important in
ozone disinfection system design because the reacted
ozone is no longer available for disinfection. Waste-
waters that have high concentrations of organics or
inorganics may require high ozone dosages to achieve
700
-------
disinfection. It is very important to conduct pilot plant
studies on these wastewaters during ozone disinfec-
tion system design in order to determine the ozone
reaction kinetics for the level of treatment prior to
ozone disinfection.
In most instances, the oxidation reactions produce an
end product that is less toxic than the original
compound (1,2). Numerous studies have been com-
pleted describing the reactions with ozone and
various inorganic and organic compounds (11,12). A
brief summary of these reactions is presented.
6.2.2.1 Reactions with Inorganic Compounds
The inorganic compounds that most commonly react
with ozone in a wastewater treatment plant are
sulfide, nitrite, ferrous, manganous, and ammonium
ions. Other reactions may also occur if the waste-
water characteristics are affected by an industrial
contribution or by in-plant recycle loads. Ozone
reactions with various inorganic compounds have
been analyzed by several researchers (2). These
reactions are summarized below:
Sulfide. The degree of oxidation of sulfide depends
upon the amount of ozone used and the contact time.
Organic sulfides will oxidize to sulfones, sulfoxides
and sulfonic acids at slower rates than the sulfide ion
itself. The sulfide ion will oxidize to sulfur, to sulfite
and to sulfate.
Nitrogen Compounds. Organic nitriles, nitroso com-
pounds, and hydroxylamines will be oxidized de-
pending on the amount of ozone used and the contact
conditions. The oxidation reaction of ammonia is first-
order with respect to the concentration of ammonia
and is catalyzed by OH" over the pH range 7-9(13). At
an initial ammonium concentration of 28 mg/l as N, a
pH of 7.0, and a contact time of 30 minutes, an 8
percent reduction of ammonium was reported. At a
pH of 7.6, a 26 percent, 8.4 a 42 percent, and 9.0 a 70
percent reduction was observed. Narkis reports that
total oxidation of organic nitrogen and ammonia was
never achieved, even at a pH of 12, and at a pH of 6
nitrates were not produced (14).
Nitrite ion is oxidized very rapidly to nitrate ion. This
reaction can have a significant effect on ozone
disinfection capability when incomplete nitrification
occurs. Venosa reported that as much as 2 mg/l of
ozone was required to oxidize 1 mg/l of nitrite-
nitrogen (15).
Iron and Manganese. The reaction with ozone and the
ferrous and manganous ions will form an insoluble
precipitate. The ferrous ion will be oxidized to ferric,
which will react with OH" to form an insoluble
precipitate. Similarly, manganeous ions will form
manganic ions which will react with OH" to form a
precipitate.
Cyanide. Toxic cyanide ions are readily oxidized by
ozone to the much less toxic cyanate ion. At low pH,
cyanate ion hydrolyzes to produce carbon dioxide and
nitrogen.
6.2.2.2 Reactions with Organic Compounds
An in-depth analysis of the reactions with ozone and
various organic compounds were developed by
several investigators. These reactions were described
by Miller et al. (1), and are summarized below.
Aromatic Compounds. Phenol reacts readily with
ozone in aqueous solution. Oxalic and acetic acids are
relatively stable to ozonation in the absence of a
catalyst such as ultraviolet light or hydrogen peroxide.
Cresols and xylenols undergo oxidation with ozone at
faster rates than does phenol. Pyrene, phenanthrene,
and naphthalene oxidize by ring rupture. Chloro-
benzene reacts with ozone slower than does phenol.
Aliphatic Compounds. There is no evidence that
ozone reacts with saturated aliphatic hydrocarbons
under water or wastewater treatment conditions.
There is no evidence that ozone oxidizes trihalo-
methanes. Ozone combined with ultraviolet radiation
does oxidize chloroform to produce chloride ion, but
no identified organic oxidation product. Unsaturated
aliphatic or alicyclic compounds react with ozone.
Pesticides, Ozonation of parathion and rnalathion
produces paraoxon and malaoxon, respectively, as
intermediates, which are more toxic than are the
starting materials. Continued ozonation degrades the
oxons, but requires more ozone than the initial reac-
tion. Ozonation of heptachlor produces a stable
product not yet identified. Aldrin and 2,4,5,-T are
readily oxidized by ozone, but dieldrin, chlordane,
lindane, DDT, and endosulfan are only slightly
affected by ozone.
Humic Acids. Humic materials are resistant to
ozonation, requiring lengthy times of ozonation to
produce small amounts of acetic, oxalic, formic and
terephthalic acids, carbon dioxide, and phenolic
compounds. Ozonation of humic materials followed
by immediate'chlorination (within eight minutes) has
been shown to reduce trihalomethane formation in
some cases. Ozonized organic materials generally are
more biodegradable than the starting, unoxidized
compounds.
6.2.3 Ozone Disinfection Reactions
Transfer of ozone into the wastewater is the first step
in meeting the disinfection objective, since ozone
must be transferred and residual oxidants produced
before effective disinfection will occur (16). Once
transferred, the residual oxidants, such as ozone,
hydroxide, or peroxide, must make contact with the
organisms in order for the disinfection action to
707
-------
proceed. Therefore, similar requirements and kinetic
relationships used for chlorine disinfectants can also
be used for ozone disinfection.
Contact time has been the subject of much contro-
versy for ozone disinfection. Studies have shown that
effective disinfection can occur at contact times as
short as one minute (17), but most existing ozone
disinfection systems have contact times that are 10 to
15 minutes. Effective ozone disinfection is due to the
combined results of high transfer efficiency, good
mixing, adequate contact time, and minimal short-
circuiting in the contactor. These factors are all
interrelated, thus isolating contact time as a particular
variable during kinetic a nalysis has not been practical.
As a result, kinetic relationships in terms of trans-
ferred ozone or residual concentrations have been
utilized.
The transferred ozone dosage versus disinfection
performance kinetic relationship is further discussed
in Section 6.5.1.2. The relationship between dis-
infection performance and residual oxidants is shown
in Figure 6-2 (17). These types of relationships can be
used graphically or empirical equations can be
developed and used to predict disinfection perform-
ance at various dosage rates. All these relationships
assume that effective gas/liquid contacting occurs.
Studies have shown that ozone disinfection can also
ba related to contactor off-gas ozone concentration
Figure 6-2. Effluent total coliform concentration versus
total residual oxidants and residual ozone (17).
(15,18,19). Additional work has shown that it was
possible to empirically relate effluent fecal coliform
concentration to the product of contactor off-gas
concentration and liquid contact time (20). The data
shown in Figures 6-3 and 6-4 indicate that a better
correlation occurred for the product of off-gas ozone
concentration times time than occurred for off-gas
ozone concentration alone. Additional information
regarding the use of off-gas ozone concentration as a
process control parameter is presented in Section
6,3.3.
Figure 6-3. Effluent fecal coliform concentration versus
off-gas ozone concentration (20).
rrio6
5
8 io5
1
I1°4
|io3
•£ •
o
^10*
c 10
-------
6.2.4 Ozone Terminology
Throughout this manual the most common units of
expression are used. The standard temperature and
pressure are 20°C(68°F) and 101 kPa(1 atmosphere),
respectively. All gas volume values are corrected to
standard temperature and pressure, unless otherwise
noted.
6.2.4.1 Measured Ozone Parameters
A simplified line diagram of a wastewater ozone
disinfection system is shown in Figure 6-5. The
system consists of the four components of feed-gas
preparation, ozone generation, ozone contacting and
ozone destruction. The feed-gas source may be either
air or high purity oxygen. In unique applications
recycled oxygen has been utilized. The feed-gas flow
rate (d) is usually measured before the ozone
generator so more precise instrumentation may be
used without the added expense of providing ozone
resistant materials.
The concentration of ozone in the feed-gas (Yi) is
measured before the ozone contact basin. In the
contact basin most of the ozone is transferred to the
wastewater and reacts with the ozone demanding
constituents in the wastewater or decomposes to
hydroxyl radicals (9). Some of the transferred ozone
may not react or decompose and will exist in the
contactor as residual ozone (C2>. Ozone not trans-
ferred exits the contact basin in the off-gas flow(Ga).
The off-gas flow rate (G2) can be quite different from
the feed-gas flow rate (GT). When oxygen is used as
the feed-gas, from 5 to 10 percent of the oxygen may
dissolve in the wastewater and cause the off-gas flow
rate to be lower than the feed-gas flow rate. If an
exhaust blower is used to pull the off-gas out of the
contact basin, then the off-gas flow rate could be
higher than the feed-gas flow rate due to air getting
into the tanks through cracks, etc. The off-gas flow
rate may be measured before or after the ozone
destruct unit since the two flows are identical. The
ozone concentration in the off-gas (Y2) is measured
prior to the ozone destruct unit and the ozone
concentration in the exhaust-gas (Y3) is measured
after the ozone destruct unit. The wastewater flow
into (Li) and out of (L2) the contact basin is identical,
and may be measured at either point. The influent
coliform concentration (No) is sampled prior to the
Figure 6-5. Simplified ozone process schematic diagram.
G3 Y3
Ozone
Destruction
G2 Y2
Feed Gas Preparation
• Oxygen Production
• Oxygen Storage
• Air/Oxygen Treatment
G,
Ozone
Generation
'G, Y,
r*
It,
w
Ozone
Contact
Basin
N
U C2
No
Legend:
Gi = Feed-Gas Flow Rate (mVmin)
YT = Feed-Gas Ozone Concentration (g/m3)
G2 = Ga = Off-Gas and Exhaust-Gas Flow Rate (mVmin)
Y2 = Off-Gas Ozone Concentration (g/m3)
Y3 = Exhaust-Gas Ozone Concentration (g/m3)
Li = L2 = Wastewater Flow Rate
Ci = Residual Ozone Concentration in Wastewater Influent (mg/L)
C2 = Residual Ozone Concentration in Wastewater Effluent (mg/L)
No = Influent Coliform Concentration (#/100 ml)
N = Effluent Coliform Concentration (#/100 ml)
103
-------
ozone contact basin and the effluent coliform con-
centration (N) is sampled after the contact basin.
The metric units of expression and English unit
conversion factors for typical ozone related measure-
ments are presented in Table 6-4. The English unit
conversions for ozone concentration in the gas are
dependent upon the standard temperature and
pressure conditions selected. The conversion factors
shown are for a standard temperature of 20°C (68°F)
and a standard pressure of 101 kPa(1 atmosphere). If
other standards are used, then the conversion factors
may be changed using the formulas shown in Table
6-4. The English unit equivalents for typical ozone
concentrations encountered in ozone disinfection
facilities are presented in Table 6-5.
6.2.4.2 Calculated Ozone Parameters
Ozone measurements described in the previous
section can be used to calculate ozone production (P),
applied ozone dosage (D), ozone transfer efficiency
(TE), and transferred ozone dosage (T). The termi-
nology, symbols, units of expression and formulas for
the calculated ozone parameters are shown in Table
6-6.
The formulas shown assume that metric units of
expression are used for the individual parameters.
The most common English units of expression are
percent wt for Yi and Y2, scfm for Gi and G2,mgdfor Li
and L.2, and Ib/d for P. To calculate P, the gas flow rate
is typically converted to Ib/d and is multiplied by the
ozone concentration in percent wt.
P in Ib ozone/day = d in Ib/d * Yi in % wt/100
The applied ozone dosage, in English units, is equal to
the ozone production in Ib/d divided by liquid flow in
mgd divided by a conversion factor of 8.34 Ib/gal.
D in mg/l = P in Ib ozone/d/Li in mgd/8.34
Table 6-4. Terminology for Measured Ozone Parameters
Parameter Symbol
Metric Units
English Conversions
Flow Measurements
Feed-gas
Off-gas
Exhaust-gas
Wastewater
Ozone Concentration
Measurements
Feed-gas
Off-gas
Exhaust-gas
G,
G2
Y,
Y2
Y3
m3/min
m3/min
rrvVmin
m3/min
g/m3
g/m3
g/m3
cfm x 0.02832
cfm x 0.02832
cfm x 0.02832
mgd x 2.629
gpm x 0.003785
Air Feed-gas
%Vol x 19.96
%Wtx 12.10
ppm Vol X 0.001996
ppm Wt. x 0.001210
Oxygen Feed-gas
%Vol x 19.96
%Wtx 13.35
ppm Vol x 0.001996
ppm Wt x 0.001335
Note: The ozone concentration conversion factors are based upon a standard temperature of 20°C and standard pressure of
1 atmosphere. The conversion factors may be changed and the ozone concentration may be calculated using the formulas
shown Below:
Ozone concentration in g/m3 = (%Vol)(W)(10)
= (MHW)(1000)/((M-48) + (48007% Wt))
= (ppmVol)(W)(0.001)
= (ppm Wt)(W)(M/48)(0.001)
where:
%Wt and % Vol are expressed as percent and not as a decimal (i.e., 1% and not .01)
M = Gram molecular weight of the feed-gas (Assumed)
Oxygen = 32 g/imole
Air = 29 g/imole
W — Maximum weight of ozone per unit volume of gas at standard temperature and pressure (see example calculation below)
Example calculation at 1 atmosphere standard pressure:
Ozone gram molecular weight = 48 g/mole
Molar volume = 0,08205 LAnole/°K
Therefore:
Molar volume at temperature of 20°C = (273.15 + 20°K)* (0.08205 L/mole°K)
= 24.503 L/mole
W = (48 g/mole)/(24.053 L/mole) = 1.996 g/L
704
-------
Table 6-5. English Unit Equivalents for Ozone Concentration
Standard Pressure 14.696 psi (101 kPa)
Standard Temperature 68°F (20°C)
Gram Molecular wt. of air 29 g/mole
Gram Molecular wt. of oxygen 32 g/mole
Gram Molecular wt. of ozone 48 g/mole
Molar Volume 0.08205 L/mole/°K
Metric
Wt Ozone
toVol
Gas
(g/m3)
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
English
Air
(%)
0.17
0.33
0.50
0.66
0.83
0.99
1.16
1.32
1.48
1.65
1.81
1.98
2.14
2.30
2.46
2.63
2.79
2.95
3.11
3.27
3.44
3.60
3.76
3.92
4.08
Percent Weight
Oxygen
(%)
0.15
0.30
0.45
0.60
0.75
0.90
1.05
1.20
1.35
1.50
1.64
1.79
1.94
2.09
2.24
2.39
2.53
2.68
2.83
2.98
3.12
3.27
3.42
3.57
3.71
Percent
Volume
(%)
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
1.10
1.20
1.30
1.40
1.50
1.60
1.70
1.80
1.90
2.00
2.10
2.20
2.31
2.41
2.51
Air
(ppm)
1,658
3,313
4,967
6,618
8,267
9,914
11,559
13,201
14,842
16,480
18,116
19,750
21,382
23,012
24,640
26,265
27,889
29,510
31,129
32,746
34,362
35,974
37,585
39,194
40,801
PPM Weight
Oxygen
(ppm)
1,503
3,004
4,503
6,001
7,498
8,993
10,486
11,978
13,469
14,958
16,446
17,932
19,417
20,900
22,381
23,862
25,340
26,818
28,294
29,768
31,241
32,712
34,182
35,651
37,118
PPM
Volume
(ppm)
1,002
2,004
3,007
4,009
5,011
6,013
7,015
8,018
9,020
10,022
11,024
12,026
13,029
14,031
15,033
16,035
17,038
18,040
19,042
20,044
21,046
22,049
23,051
24,053
25,055
Table 6-6. Terminology for Calculated Ozone Parameters
Parameter Symbol
Units
Calculation Formula
Ozone Production
Applied Ozone Dose
Transfer Efficiency
Precise
Approximate
Transferred Ozone Dose
P
D
TE
TE
T
g/min
mg/l
mg/l
G, x Y
100*(Y1-Y2)/Y1
Y, x TE * d/L,
To accurately calculate ozone TE several parameters
must be measured (See Table 6-6). However, a
simplified and acceptable estimation can be obtained
if the off-gas flow rate is assumed to approximate the
feed-gas flow rate, as shown in Table 6-6.
The error that occurs using the approximate TE
calculation is less than two percent for most design'
situations (i.e., actual TE is usually greater than 85
percent and off-gas flow rate is usually within 10
percent of the feed-gas flow rate). In most applications
the approximate calculation is adequate and elimi-
nates the need for installing a separate flow meter for
off-gas flow measurement. However, when precise
data are required, for example for verification of
design specifications, the precise method of calcu-
lating TE should be used.
6.2.4.3 Auxiliary Measurements and Calculations
Other important parameters in ozone terminology are
temperature, pressure, energy, power, moisture
content, and dew point. Moisture content and dew
705
-------
point of the feed-gas are most important when air or
recycled oxygen is u$>ed because the feed-gas mois-
ture content must be reduced to very low levels in
order to prevent damage to the internal components
of the ozone generator. Moisture removal is typically
not required when high purity oxygen is the feed-gas,
because its moisture content is usually very low.
The relative moisture content of the feed-gas is called
dew point temperature, or "dew point." Dew point is
the temperature at which a gas (at a specific pressure
condition) is saturated with water. If the temperature
of the gas decreases or pressure increases from those
conditions, the water will condense. Calculating dew
point is an important element in ozone system design
because the size of some of the air treatment
equipment is a function of the feed-gas dew point
temperature.
The relationship between moisture content in air and
dew pointtemperature at standard pressure is shown
in Table 6-7(21). The data is for the range of dew point
temperatures typically encountered in ozone process
Table 6-7. Moisture Content of Air for Air Temperature
from -80 to 40°C (21)
design. The moisture content increases dramatically
as the dew point temperature increases. For example,
the initial moisture content of an air feed-gas with a
standard dew point of 20°C (68°F) is 17.7 g/m3 (1.11
lb/1,000 ft3), whereas the desired moisture content
of the treated feed-gas is only 0.008 g/m3 (0.000507
lb/1,000 ft3) (dew point temperature of-76°F(-60°C)).
The dew point temperatures shown are for a standard
pressure of 101 kPa(14.7 psi), and must be adjusted
for operating pressure conditions. The precise calcu-
lation for dew point and moisture content is by means
of psychrometric formulas or by use of psychrometric
charts (21). However, the calculation may be approx-
imated by an inverse proportional relationship of
absolute pressures (gauge plus atmospheric pres-
sure). A summary of atmospheric pressures at
different elevations is shown in Table 6-8 to assist in
the determination of absolute pressure.
In ozone process design the moisture loading to the
desiccant dryer should be calculated. The desiccant
dryer is used to reach the required dryness of the
feed-gas, and is a very important piece of equipment
in the air treatment process. An example calculation
of moisture loading to the desiccant dryer is shown in
Example 6-1.
Air Temperature
ro
-80
-75
-70
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
5
10
15
20
25
30
35
40
Moisture
\A/oirthit
(°F) Ib H20/lb air
-112
-103
-94
-85
-76
-67
-58
-49
-40
-31
-22
-13
-4
5
14
23
32
41
50
59
68
77
86
95
104
Air weight/volume is
sure
"Air weight/volume =
bg/m3^lb/1,000ft3*
Weight8
Volums
Ib/ft3
0.0000003168 0.07526
0.0000007713 0.07526
0.000001640 0.07526
0.000003342
0.000006743
0.00001311
0.00002464
0.00004455
0.00007925
0.0001381
0.0002344
0.0003903
O.OOOS731
0.001020
0.001606
0.002485
0.003788
0.005421
0.007(558
0.01069
0.01475
0.0201)6
0.027511
0.03673
0.0491 1
corrected to
' 1,205 g/m3
16.012
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
0.07526
Moistureb
Content
Ib HaO per
1,000ft3
0.00002384
0.00005805
0.0001234
0.0002515
0.0005074
0.0009866
0.001854
0.003352
0.005964
0.010393
0.017640
0.029373
0.050665
0.07676
0.1209
0.1870
0.2851
0.4080
0.5763
0.8045
1.1100
1.5171
2.0552
2.7642
3.6959
68°F and 1 atmosphere pres-
Example 6-1. Calculation of Moisture Loading to
Desiccant Dryer
Assume that a plant is at an elevation of 91 5 m (3,500
ft) above sea level (atmospheric pressure is 89.1 kPa
(12.93 psi)); the feed-gas flow rate is 600 scfm (17
mVmin); the gauge pressure is 93 kPa (1 3.5 psig); the
inlet air
temperature
is 30°C
humidity is 60 percent; and
temperature is cooled with an
(68°F). Determine the
Table 6-8.
ambient
Atmospheric Pressure
Altitude
Feet
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
5,000
6,000
7,000
8,000
9,000
10,000
Meters
0
152
305
457
610
762
914
1,067
1,219
1,372
1,524
1,829
2,134
2,438
2,743
3,048
(86°F); the
relative
the compressed air
after-cooler to 20°C
air moisture
content;
at Different Altitudes.
Atmospheric Pressure
psi
14.70
14.43
14.16
13.91
13.66
13.41
13.17
12.93
12.69
12.46
12.23
11.78
11.34
10.91
10.50
10.10
kPa
101.325
99.49
97.63
95.91
94.18
92.46
90.80
89.15
87.49
85.91
84.32
81.22
78.19
75.22
72.39
69.64
Atm
1.000
0.982
0.964
0.947
0.930
0.912
0.896
0.880
0.864
0.848
0.832
0.802
0.772
0.742
0.714
0.687
1O6
-------
the compressed, cooled feed-gas moisture content;
and the moisture loading of the feed-gas to the
desiccant dryer.
Ambient air moisture content. From Table 6-7 the
moisture content of the air at standard pressure and a
temperature of 30°C (86°F) is 32.1 g/m3 (2.055
lb/1,000 ft3). This value must be adjusted for ambient
pressure and relative humidity conditions, and may
be approximated as follows:
2.055 lb/1,000 ft3 * 14.70 psia/12.93 psia *
0.60 = 1.401 lb/1,000 ft3
or
24.29 g/m3 * 1 atm/.88 atm * .60 = 16.56 g/m3
Compressed, cooled feed-gas moisture content. From
Table 6-7 the moisture content of the feed-gas at a
temperature of 20°C (68°F) is 17.8 g/m3 (1.110
lb/1,000 ft3). The absolute pressure is 182 kPa (26.43
psia) (i.e., 13.5 psi + 12.93 psi). The approximate
moisture content is:
1.110 lb/1,000 ft3 * 14.70 psia/26.43 psia =
0.617 lb/1,000 ft3
The moisture content of 9.88 g/m3 (0.617 lb/1,000
ft3) is the maximum amount of water the feed-gas can
"hold" at 20°C (68°F) and 182 kPa (26.43 psia). This
moisture content is 41 percent of the moisture
content of the ambient air. The remaining 59 percent
of the moisture will condense out of the air stream
and will be removed through a "moisture trap."
Although not part of the question, the standard dew
point of the compressed air can be approximated
using Table 6-7 as about 10°C (50°F).
Moisture loading to the desiccant dryer is approx-
imated as follows:
0.617 lb/1,000 ft3 * 600 scf m * 60 min/hr/1,000 =
22.21 Ib/hr
or
9.88 g/m3 * 17 m3/min = 167.9 g/min
6.3 Process Flow Schematics
The ozone disinfection process consists of a liquid
and a gas flow scheme, as shown in Figure 6-6. The
feed-gas is typically air or high purity oxygen. A small
amount (from 1 percent wt in air to 4 percent wt in
oxygen) of the feed-gas is converted to ozone within
the ozone generator. The ozone containing feed-gas
is directed to a contact basin where it combines with
the wastewater. Typically, a high percentage of the
ozone, but only a small percentage of the total feed-
gas flow, is transferred to the liquid in the contact
basin. The disinfection action proceeds in the contact
basin.
The gas and liquid flows separate after the contact
basin. The off-gas is treated to remove the un-reacted
ozone and is reused, recycled, or discharged to the
atmosphere. The disinfected wastewater is dis-
charged to the receiving water. Any residual ozone in
the wastewater is reacted or is reverted back to
oxygen.
6.3.1 Feed-Gas Flow Schematic
Ozone disinfection processes are typically distin-
guished by the type of feed-gas used. The most
common types are air-fed and oxygen-fed. Oxygen-
recycle systems have been used on some occasions.
A diagram showing the components of these proc-
esses is presented in Figure 6-7. The air-fed system is
most common in plants where oxygen is not available.
Oxygen-fed systems are typically used in conjunction
with an oxygen activated sludge treatment system,
where the unused oxygen from the ozone disinfection
system is used in the biological treatment process.
The quality of the feed-gas is critically important for
the electric discharge ozone generators. The feed-gas
must be oil-free, particle-free, and dry. To achieve
these characteristics the air-fed (and oxygen-recycle)
systems must pre-treat the gas to remove moisture
and particulates, and oil, if present. The once-through,
oxygen-fed system typically pretreats for particulates
only, because the oil and moisture content of the high
purity oxygen from the oxygen production facilities is
negligible.
The ozone generation equipment and power require-
ments are about 50 percent lower for an oxygen-fed
system than for an air-fed system because approx-
imately twice as much ozone is produced with high
purity oxygen at a given input of power to the
generator. However, this lower cost of producing
ozone is often offset by a higher cost of obtaining high
purity oxygen to feed the generator. On the other
hand, when the oxygen can also be used in the
biological system the economics are more favorable
using high purity oxygen.
The ozone disinfection contacting units are similar for
all feed-gas systems shown in Figure 6-7. In the
contact basin some disinfection action occurs through
direct contact of the microorganism with the ozone in
the gaseous phase; however, "effective" disinfection
appears to exist only when residual oxidants or an
ozone residual is present (16,17). In order to obtain
residual oxidants or residual ozone, a sufficient
amount of ozone must be transferred to the waste-
water.
Several investigators have documented that an
excellent correlation exists between the log-trans-
ferred ozone dosage (log(T)) and log-coliform survival
ratio (iog(N/No)) (17,18,22-25). Stover (17) reported
707
-------
Figure 6-6. Ozono disinfection process gas and liquid flow diagram.
Gas Flow Stream Liquid Flow Stream
Feed-Gas
I
Ozono
Generation
Wastewater
I
I
Chemical,
Physical,
and/or
Biological
Treatment
Ozone
Disinfection
Contact
Basin
To Off-Gas
Treatment
and
Discharge,
Recycle
or Re-Use
L_ ..
Wastewater Effluent
To Discharge
Figure 6-7. Diagrams showing feed-gas flow of typical ozone disinfection processes.
Air Fed
Air
Treatment
Ozone
Generator
Ozone
Contacting
Ozone
Destruction
Vent
Oxygen Fed
High Purity
Oxygen
Ozone
Generator
Ozone
Contacting
Ozone
Destruction
Oxygen
Activated
Sludge
Vent
Oxygen Rec/cle
Dew Pont
Treatment
j
Ozone
Generator
> i
up C
Ozone
Contacting
xygen
Ozone
Destruction
».
* Vent
Oxygen
Activated
Sludge
Recycle Oxygen
Vent
708
-------
that this relationship exists irrespective of the type of
feed-gas used; thus, the decision to use air or oxygen
as the feed-gas is typically based on economic
considerations of providing the oil-free, particle-free
dry gas to the generator and not on the disinfection
performance capability associated with the feed-gas.
A common application for oxygen-fed ozone disinfec-
tion is in conjunction with an oxygen activated sludge
process where the oxygen is reused. Only about 10
percent of the oxygen in the feed-gas is "used-up" in
the ozone process; thus, 90 percent or more is
available for the biological treatment process. The
major design consideration is to balance the oxygen
requirements of both processes. This can be ad-
dressed by balancing the oxygen gas (G-i) to waste-
water liquid (l_i) flow rates to both processes.
An example approach to balancing the gas to liquid
ratios to both processes is shown in Figure 6-8. The
first step is to estimate the applied BOD concentration
to the oxygen activated sludge process. The activated
sludge oxygen gas to wastewater liquid ratio can then
be approximated by estimating the applied oxygen
requirement. The example in Figure 6-8 shows that
the activated sludge Gi/l_i ratio is 0.135 for an
applied BOD concentration of 120 mg/l and an
oxygen requirement of 1.5 kg O2/kg BOD. The
example also shows that the activated sludge gas to
liquid ratio balances with the ozone process Gi/Li
ratio at a feed-gas ozone concentration of 40 g/m3(3
percent wt) and an applied ozone dosage of 6 mg/l,
assuming 10 percent loss in the ozone contactor.
The Gi/L| ratio required for the ozone system is
shown as 10 percent greater than required for the
activated sludge process, because about 10 percent
of the oxygen is used up in the ozone process through
transfer of ozone and oxygen to the wastewater. For
the example in Figure 6-8, the adjusted Gi/l_i ratio for
the ozone system is 0.15. Knowing the Gi/l_i ratio for
ozone, either the applied ozone dosage can be
determined for a given feed-gas ozone concentration,
or the required feed-gas ozone concentration can be
estimated for a given applied ozone dosage. If the
oxygen balancing analysis indicated that significantly
more oxygen was required for ozone disinfection
because the ozone concentration was too high, then
an oxygen recycle process could be provided. If the
analysis indicated that more oxygen was required for
the activated sludge process, then an oxygen bypass
pipe around the ozone disinfection system could be
provided.
Figure 6-8. Oxygen requirement for ozone disinfection compared to oxygen requirement for activated sludge.
10% Adjustment
for Oz Loss
40 80 120 160
Applied BODS Concentration, mg/L
200
246
Applied Ozone Dosage, mg/L
10
109
-------
When oxygen is not used in the biological portion of
the wastewater treatment process, then the cost of
an air-fed ozone process may be compared with the
cost of a total oxygen-recycle system to determine the
most economical alternative. The main disadvantages
of the oxygen-recycle system were the cost of
obtaining oxygen-rich feed-gas and the additional
attention required to handling an oxygen-enriched
gas. The main advantages were the reduced cost of
the ozone generation equipment and the reduced
energy consumption for ozone generation. Most
oxygen activated sludge plants in the United States
that use ozone use the once-through, oxygen-fed
system. As indicated in Table 6-1, all wastewater
treatment plants that use ozone disinfection, but do
not use oxygen activated sludge, use the air-fed
system.
The ozone destruct units shown in Figure 6-7 are
fundamentally the same for all feed-gas systems. The
size of the ozone destruct unit is based on the volume
of gas directed to the unit.
6.3.2 Wastewater Flow Schematic
Ozone will disinfect to a very high degree if a
sufficient amount of ozone is applied to the waste-
water and a properly designed disinfection contact
basin is provided (17). However, there appears to be
economic considerations for the ozone disinfection
process associated with the type of wastewater
treatment processes selected prior to the ozone
disinfection system. The main effect of the waste-
water treatment processes used is the impact on the
amount of ozone required to obtain the desired
disinfection level.
Wastewater treatment schemes prior to ozone dis-
infection that have been evaluated by various re-
searchers are shown in Figure 6-9 (17,24,25). These
schemes range from simple "fine screening" to
biological systems followed by chemical treatment.
Given and Smith et al. (24) evaluated ozone disinfec-
tion using the effluent from four types of wastewater
treatment process schemes, including fine screening
effluent, rotating biological contactor effluent, an-
aerobic lagoon effluent, and a "strong" waste flow
stream. The coliform survival results obtained over a
range of transferred ozone dosages is shown by the
dose/response curves in Figure 6-10.
The major factor influencing the amount of ozone
required to achieve a desired reduction in fecal
Figure 6-9. Example treatment schemes using ozone disinfection (17,24,25).
Preliminary Treatment
Fine
Screening
Ozone
Disinfection
Secondary Treatment
Biological
Ozone
Disinfection
Tertiary Treatment—Filtration
Biological
Ozone
Disinfection
Advanced Treatment
Biological
Carbon
Adsorption
Ozone
Disinfection
770
-------
Figure 6-10. Fecal coliform survival for rotating biological
contactor effluent, screened effluent, anaero-
bic lagoon effluent, and strong wastewater
(24).
JL = / C" \"2
N0 \1.8/
0.7
u
n = 331
r = -0.89
11.8 3.4
12.5
I
5 10
Ozone Utilized, mg/L
50
.999
Initial Conditions
Type of Waste
Parameter RBC Screened Lagoon Strong
FC
TEMP
TURB
pH
BOD
SS
VSS
6.5 x10"
8.9
5
7.9
13
13
8
1.3x10"
8.6
42
7.8
95
102
57
3.8 x105
6.9
67
7.6
92
121
47
2.8x10e
5.4
151
—
—
1010
«20
coliform concentration was the "initial ozone de-
mand" of the wastewater, which was 0.7, 1.8 , 3.4 ,
and 12.5 mg/l for the RBC, screened, anaerobic
lagoon and "strong" effluents, respectively (24). The
slopes of the dose/response curves varied from -2.9:1
to -4.6:1. Some difference in the slope may be due to
the type of wastewater treated, but the data set f or al I
wastes except the screened waste was too limited (13
to 16 data points) to confirm a cause for the variation
in slope.
Gan et al. (25) evaluated ozone disinfection per-
formance of an activated sludge effluent and three
tertiary treatment schemes following activated sludge
treatment including: coagulation, sedimentation, and
filtration (Scheme A); coagulation and filtration
(Scheme B); and carbon adsorption (Scheme C). The
carbon adsorption data set was further divided into
Schemed and Scheme C2. Schemed had a nitrite-
nitrogen concentration greater than 1.0 mg/l and
Scheme C2 a nitrite-nitrogen concentration less than
1.0 mg/l. The major factors influencing the amount of
ozone required for disinfection were the influent
dissolved chemical oxygen demand (DCOD), the
influent nitrite-nitrogen concentration, andthetarget
or desired effluent coliform concentration. The effect
of the target effluent coliform concentration was
evaluated by comparing the ozone dosage required to
meet the former EPA standard of 200 fecal coliforms
per 100 ml andthe more stringent California standard
of 2.2 total coliforms per 100 ml. The results are
shown in the bar graph in Figure 6-11.
The carbon adsorption effluent met the more stringent
California standard 100 percent of the time at an
ozone dosage of 6 mg/l. The other pre-treatment
effluents did not meet this high coliform standard
even at ozone dosages of 10 mg/l. The major effect of
carbon adsorption pre-treatment was to reduce the
DCOD concentration of the wastewater and thus
reduce its ozone demand.
The less stringent disinfection standard of 200 fecal
coliform/100 mL was achieved 100 percent of the
time for all wastewater treatment schemes evaluated,
except when the nitrite-nitrogen concentration was
greater than 1.0 mg/l (Scheme C1). The high nitrite-
nitrogen concentration of the wastewater influent to
the contact basin was caused by an upset in the
activated sludge plant. The ozone reacted with the
nitrite causing an increase in ozone demand and a
significant reduction in disinfection performance (25).
The specific relationships between disinfection per-
formance and the DCOD and nitrite-nitrogen con-
centrations are shown in Figure 6-11. At a given
ozone residual in the contact basin effluent of 2.0
mg/l, the best disinfection performance was achieved
with an influent nitrite-nitrogen concentration less
than 0.15 mg/l and a DCOD concentration less than
12 mg/l. Based on these results, to keep the ozone
dosage less than 10 mg/l, tertiary treatment of the
wastewater may not be necessary to meet the former
EPA standard (25). However, tertiary treatment to
reduce the ozone demand due to DCOD and nitrite-
nitrogen appears necessary to meet the stringent
California standard of 2.2 total coliforms per 100 ml.
Stover et al. (17) also evaluated the ozone dosage
requirements to meet two different disinfection
standards of 2.2 and 70 total coliforms per 100 ml.
The evaluation was completed for three different
wastewater treatment schemes consisting of a
filtered activated sludge secondary effluent, a nitri-
fied effluent, and a filtered nitrified effluent. The most
significant factors affecting the dosage requirement
were the organic quality of the wastewater as
111
-------
Figure 6-11. Effect of water quality and performance criteria on ozone dosage requirement (25).
•o
CO
m
m
g
-------
Stover et al. (17) evaluated the effect of the organic
concentration of the wastewater influent to the
contact basin by comparing the transferred ozone
dosage requirement for secondary and nitrified
effluents to obtain similar effluent total coliform
concentrations. For both disinfection standards eval-
uated, significantly more transferred ozone dosage
was required for the filtered secondary effluent (COD
of 40 mg/l) than was required for the nitrified
effluents (COD of 20 mg/l).
Stover also showed that a more stringent effluent
total coliform concentration standard caused a signif-
icant increase in the transferred ozone dosage
required. For example, Figure 6-12 shows that for the
filtered nitrified wastewater treatment scheme five
times more transferred dosage was required to meet
the 2.2 per 100 ml standard (15 mg/l) than was
required to meet the 70 per 100 ml standard (3 mg/l).
Stover a na lyzed the effect of the inf I uent tota I col if orm
concentration by comparing the disinfection results
for the nitrified and filtered nitrified effluents. The
data in Figure 6-12 show that less ozone was required
to reach a given effluent coliform concentration for
the filtered nitrified effluent than was required for the
nitrified effluent. However, these data do not consider
the fact that the influent total coliform concentration
of the filtered, nitrified wastewater was lower. The
inf I uent total coliform concentration was incorporated
in the analysis in Figure 6-13, where the log reduction
of total coliform (Log N0/N) is plotted against the log
transferred ozone dosage. The results indicate no
difference in performance. Stover's conclusion was
that the benefit of filtration in removing total coliform
may be a more significant factor in reducing the ozone
dosage requirement than is the benefit of removal of
suspended solids, at least within the small range of
suspended solids tested (average 2.5 mg/l for the
filtered nitrified and 5.7 mg/l for the nitrified ef-
fluents).
Gan et al. (25) also addressed the subject of filtration
for suspended solids removal prior to an ozone
disinfection process. The results of Can's comparison
between a high (23 mg/l) and a low (11 mg/l)
suspended solids influent to the ozone disinfection
contact basin are shown in Figure 6-11, where total
coliform survival is shown as a function of contact
column or stage in the disinfection contact basin. As
shown, the total coliform survival ratio was lower
(i.e., better kill of conforms) when the suspended
solids concentration was lower. However, the effect
of suspended solids diminished as.disinfection pro-
gressed through the six-stage contact basin. The
effect of suspended solids was greatest in the first
two stages, and minimal effect was evidenced by the
sixth stage. Gan concluded that the removal of
suspended solids (via flotation) that occurred in the
first two stages "indicates that removal of suspended
solids prior to ozonation may not be essential to
achieve disinfection of wastewater."
Figure 6-12. Total coliform concentration versus transfer-
red ozone dosage for various effluents (17).
10"
E
8
c
3
O
u
E
o
I 10=
o 70
§
2.2
10°
Filtered Nitrified
Nitrified
LogY = 0.121X+2.179
r = 0.63
LogY = 0.146X+3.12
r = 0.78
Filtered Secondary
Log Y = 0.088X + 3.47
r = 0.77
7 14 21 28 35
Transferred Ozone Dose, mg/L
42
Figure 6-13.
Total coliform reduction versus log transferred
ozone dosage for nitrified effluents (17).
6
5
c
o
o
•§ 4
IB '
DC
E
£ 3
1
£
o>
o
-1 1
0
u Nitrified
° Filtered Nitrified
Log Y = 2.51 Log X + 0.76
r = 0.76
0.50
(3.2)
1.00
(10.0)
1.50
(31.6)
2.00
(100.0)
Log Transferred Ozone Dosage (mg/l)
773
-------
Venosa et al. (26) also studied the effect on disinfec-
tion of filtered arid non-filtered secondary effluent.
The data indicated that improved disinfection ef-
ficiency occurred as a result of filtration, but was
more related to the reduction in total chemical oxygen
demand (TCOD) 1:han to removal of TSS. The con-
clusion was that filtration to remove TSS may not be
necessary when Ihe TCOD concentration is low.
The EPA Water Engineering Research Laboratory has
also evaluated several secondary treatment plant
effluents to determine the relationship between
ozone dosage and total coliform reduction (27). The
most significant factor influencing the ozone dosage
requirement to achieve a desired effluent total
coliform concentration was the TCOD concentration
• of the effluent. At five plants where the TCOD of the
secondary effluent was less than 40 mg/l, a total
coliform concentration of 1,000 per 100 ml could be
achieved with ozone dosages between 4 and 7 mg/l.
However, when Meckes et al. (27) evaluated a plant
which treated a significant amount of industrial
waste (TCOD = 74 mg/l), a dosage greater than 12
mg/l was projected in order for the process to meet
the 1,000 total coliforms per 100 ml limit.
Based on the data presented above, it appears that
there is no technical basis for excluding the use of
ozone following any treatment scheme. However,
depending upon the type of wastewater treated
and/or the effluent disinfection requirement, the
wastewater treatment scheme may be an important
economical consideration. A summary of the issues
to consider when selecting the liquid flow schematic
prior to ozone disinfection is presented below:
1. Required Effluent Target
• To meet the former EPA sta ndard of 200 fecal
coliforms per 100 ml, tertiary treatment may
not be necessary.
• To meet more stringent standards, such as 14
fecal coliforms per 100 ml, tertiary treatment
should be considered.
• To meet a standard of 2.2 total coliforms per
100 ml, advanced treatment unit processes
prior to the ozone disinfection process may be
required.
2. Influent Coliform Concentration
• Coliform removal is a function of transferred
ozon.e dosage; thus, the influent coliform
concentration will affect the amount of ozone
dosage required to meet specific effluent
criteria.
• Treatment processes that reduce the influent
coliform concentration (such as filtration) will
decrease the ozone dosage required to
acheive a specific effluent standard.
3. Wastewater Quality Characteristics
• The ozone demand of the wastewater signif-
icantly increases the ozone dosage require-
ments. A plant with a large industrial
contribution may have a large ozone dosage
requirement. Pilot testing to establish ozone
dosage requirements in these plants is highly
recommended.
• Incomplete nitrification and a high concen-
tration of nitrite-nitrogen will significantly
increase the ozone demand and thus the
ozone dosage requirement. The nitrite-nitro-
gen concentration preferably should be less
than 0.15 mg/l to optimize disinfection
performance.
6.4 Ozone Equipment Design
Considerations
Ozone generation is an established process, but its
use in wastewater disinfection is relatively new. In
this section the current state-of-the-art equipment
design considerations to develop an ozone disinfec-
tion system that consistently achieves desired levels
of performance is discussed.
6.4.1 Ozone Generation Equipment
The basic components that comprise an ozone
generator are depicted in Figure 6-14. The com-
ponents include an electrical source that supplies
high voltage, alternating current across a discharge
gap where the oxygen containing feed-gas passes; a
dielectric material that prevents electrical short-
circuiting; and a heat removal mechanism that
prevents rapid decay of the ozone molecule back to an
oxygen molecule. Heat removal is required because
85 to 95 percent of the electrical energy supplied to
the ozone generator produces heat (1,7).
6.4.1.1 Ozone Generation Theory
Electrical power used to generate ozone is received
from a voltage regulator, and in some generators a
frequency regulator is also used. The altered current
Figure 6-14. Cross-section view of principal elements of a
Corona discharge ozone generator.
Heat Removal
.^Electrode
-Dielectric
Generator Discharge
Containing Ozone
""Electrode
High
Voltage
AC Feed-Gas
Alternatingjcontaining
Current I O
Power
Source
Corona
"Discharge Gap"
Heat Removal
774
-------
is supplied to a number of ozone generation cells
connected in parallel, as shown in Figure 6-15. Each
cell acts as a capacitor, as illustrated in Figure 6-15b.
The capacitance of the cell is a function of the width of
the gas space and the electrical conductivity of the
dielectric material. Because the dielectric material
(glass or ceramic) is the major component of the cell,
the ozone generating cell is often called a "dielectric."
Ozone is produced when the ionization potential of
the dielectric is reached. The system's electrical
characteristics before and after the ionization poten-
tial is reached are quite different. Prior to the
dielectric reaching its ionization potential the voltage
is insufficient to allow an electric discharge. When
the ionization potential is reached a flow of electrons
(corona) will occur across the discharge gap and the
electrical circuit will be completed. The minimum
voltage required to meet the ionization potential of a
dielectric is about 10,000 volts (1,7,28).
When the voltage is greater than the ionization
potential of the dielectric the electrons travel from
one electrode to the other electrode within the
discharge gap and collide with the oxygen molecules
in their path. Upon collision the reactions noted in
Figure 6-16 occur. The number of ozone molecules
formed will vary from none at all to a maximum of two
for every free electron discharged. The number
formed is highly dependent upon the temperature of
the ozonized gas. At higher temperatures the ozone
rapidly decomposes back to oxygen. A major con-
sideration in the design of an ozone generation
system is cooling of the ozone generator.
An alternating electrical current must be used in
ozone generation; thus, the voltage will cycle above
and below the ionization potential of the cell (1).
Ozone production occurs when the voltage is greater
than the ionization potential. Ozone production is
terminated in the portion of the cycle when the
voltage is below the ionization potential, as shown in
Figure 6-17. The amount of time ozone is formed is
dependent upon the frequency of the power supply.
The number of free electrons discharged is a function
of the applied, peak voltage and its electrical fre-
quency. If the peak voltage is significantly higher than
the ionization potential of the dielectric, a propor-
tionately greater number of electrons will be released.
The result is an increase in the ozone production rate
if the temperature of the ozonized gas is acceptable.
Similarly, an increase in the frequency of the power
supply increases the amount of time that the ioniza-
tion voltage will be reached, which also increases the
ozone production rate if the temperature of the
ozonized gas is acceptable.
Figure 6-15a. Schematic diagram of a typical power supply
to an ozone generator (1).
PS
R
Electrical Regulation of
Power Voltage/Frequency
ll--
Ozone
^- Generation
Cells
Connected
in Parallel
<-- -II"
Ground
Figure 6-15b. Schematic diagram of an ozone producing
cell, a "Dielectric" (1).
EMF
High
Tension Electrode
Dielectric
i Ground
Low
Tension Electrode
Discharge Gap
Figure 6-16. A free flow of electrons in the discharge gap
causes various reactions with the oxygen
1 molecule (1).
<* Results in (a)
O ih\
(c)
\;
O—O = Oxygen Molecule
• —*• = Free Electron
= Ozone Molecule
6.4.1.2 Design Considerations for Ozone
Generation Equipment
The amount of ozone produced by an ozone generator
is affected by the physical characteristics of the
equipment, the power supply to the generator; the
775
-------
Figure 6-17. Ozone formation occurs when the voltage
level is sufficient to create a free flow of
electrons within the discharge gap (1).
Ionized
Q.
|°
£
u
eft
5
i . Period of
xr^H
/
/
\
» 7T/2
Build Up
to lonlzation
Threshold
i
*" Deionization x^1 \
Waveform of /
^ — Power Source /
A
\
\
3/2 7T
V
^___j'
*~~~ —
A"
Correspondingly, in the Discharge Gap:
o-o * Oxygen
•-»• = Free Electron & Direction of Travel
°cf> = Ozone
"
x>6 *>
0-0%
g
'I
£
u
2
CD
£
Is'?
r i ^
c?
o
N C
o-
g>
^T C
•o'S S
u > g
"•§ S
!!!
|si
•J25
.2 « o
ID O
.a o g
.2 u. c/j
^1
§o o-o
O
3)
£•
O
U)
S O
&Z
c> u
<->£
c: ^J
•Si Q)
si »
'c: g
,c> S
CI Q
t » 2~°
* T °a>
o^ t '<&
•o
o
u
ID
•doc
0) _ ID
fell
S -1 CD
?..i s
*C ro fl)
«£ e
o E to
Q. 2. W
<$>?<&
O-OOo^1
0) U)
°°- 2
E g-
o E
0 8
C 0)
0 Q
So
c a
.2 E
0) O
O (O
.......
5/277-
S^C^1
i t ^ i
Cycle _
Repeats
\
37T
moisture content and dust content of the feed-gas;
the temperature of the ozonized gas, a nd the feed-gas
oxygen content. Each factor is further discussed.
Physical Characteristics of Ozone Generator. The
relationship of the factors affecting the ozone produc-
tion rate of the generator dielectric is shown in
Equation 6-2 (7).
P 4k f Va [Cd(Vo-Va) - VaCd] (6-2)
where:
P = Ozone production
Va = Voltage applied to the discharge gap
Vo = Extreme or peak value of the applied voltage
Cd = Capacitance of dielectric
Ca = Capacitance of the discharge gap
f = Frequency of current supply
k = Constant
The most important point in the equation is that the
ozone production rate increases with frequency and
the square of the voltage. Frequency or voltage
changes are used to adjust the ozone production rate
for a given ozone generator. The production rate also
increases with a greater number of dielectrics in
service. Because of these relationships, almost any
desired production rate can be supplied by the ozone
equipment manufacturers.
The dielectric constant and the size of the discharge
gap (typically 2 to 3 mm) between the high and low
tension electrodes also affect the ozone production
rate (7). These factors are typically defined by the
individual ozone generator manufacturer and are
proprietary items. The dielectric constant and width
of the discharge gap may vary slightly from generator
to generator, even with generators of the same
manufacturer. A very slight change in either param-
eter has a large influence on the ozone production
rate. As such,, the ozone production rate may vary
slightly for similar sized ozone generators, and from
generator to generator.
Ozone Generator Power Supply. Three primary ways
in which the power is supplied to the ozone generator
are shown in Figure 6-18 (1). Method A is a low
frequency (typically 60 hertz), variable voltage system.
This system is most common because high voltage
transformer technology preceded the technology in
high frequency transformers. Method B is a medium
frequency (up to 600 hertz), variable voltage power
supply system. This process has been used to
increase the production rate of installed ozone
equipment (1). Method C is a variable high frequency,
constant voltage unit. Method C is used by various
ozone generatormanufacturersin lieu of the variable
voltage process. No one method has a clear advantage
over the other.
The most common ways in which voltage and
frequency to the ozone generator are controlled are
shown in Figure 6-19 (1). The voltage controlled
system uses single phase power, which will cause an
imbalance in the amperage of a three-phase power
supply system unless three or multiples of three
ozone generators are operating simultaneously.
However, this capability is seldom available; thus, the
impact of an unbalanced electrical load on the overall
plant three-phase power supply should be evaluated.
To balance the electrical load a Scott transformer is
typically used (29).
The power supply to the ozone system is typically at a
frequency of 60 cycles and potential of 480 volts. For
a variable voltage system the voltage may be in-
creased by applying the line current to one of a series
of tappings on the primary side of the potential
transformer, thereby changing the transformer ratio
and the resultant secondary voltage (1). Another
method is to use a variable autotransformer to feed
776
-------
Figure 6-18. Schematic diagram of three power supply
systems typically used for ozone generation
(1).
Figure 6-19. Typical ways for varying voltage and frequency
to an ozone generator (1).
Method A
Fixed Line Low Frequency (60 Hz), Variable Voltage
Ozone
Generator
PS,
VT
Power Powerstat
Source or Variable
Transformer
HTT
' 3
-I
High Tension
Step-Up
Transformer
Method B
Fixed Medium Frequency (600 Hz), Variable Voltage
PS
Power
Source
FC
VBL
HTT
,-T-, Ozone
Frequency Variable High Tension
Converter Transformer Step-Up
Transformer
Method C
Fixed Voltage Variable Frequency
PS
VFC
Power Variable
Source Frequency
Converter
HTT
High Tension
Step-Up
Transformer
"p Ozone
'Generator
Ozone
Generator
Line Voltage
at 60 Hz
Auto
Transformer
(Sometimes
Connected to a
Powerstat)
Inductance
to Correct
Power Factor
High Tension
Transformer
Variable Voltage
Series
Inductance
FourThyristers
Adjust Rectified
Power and Change
to Desired Frequency
(This is Where Control
/f~ is Exercised)
Rectifier Bridge
(Convert AC to DC)
High Tension
Transformer
Ozone
Generator
Variable Frequency
the primary side of the main transformer (1). When a
variable frequency system is used both the frequency
and voltage must be increased. The voltage must be
increased to a level above the ionization potential of
the dielectric (about 10,000 volts). The frequency is
increased up to a maximum of 2,000 hertz (1).
Because the ozone generator uses high voltage and in
some cases also high frequency electrical current,
special electrical design considerations must be
implemented. For example, special insulation must
be provided for the electrical wire; a cool environment
for the high voltage transformers should be provided;
and the electrical transformers should be protected
from ozone contamination due to minute ozone leaks
that could occur on a periodic basis.
The electrical considerations for an ozone system
should receive special attention. For example, a
number of problems have been reported with dry-type
potential transformers (1). Oil-cooled transformers
apparently have performed more reliably. In view of
the dependence of ozone generation on high fre-
quency or high voltage electrical energy, the ozone
generator supplier should be responsible for design-
ing and supplying the electrical subsystems. How-
ever, the specifications should require that the
frequency and voltage transformers be high quality
units designed for ozone service. The ozone generator
supplier should be requested to provide a record of
successful electrical equipment performance.
Another item to consider in the design of the ozone
generation system is power factor. An operating
ozone generator can decrease the power factor to 0.3
to 0.5, depending on the generator setting (29).
Corrections will normally be cost effective, since
utilities that supply electrical power typically impose
penalties for a low power factor.
Power factor is the ratio between the apparent power
(kW) measured by a watt-hour meter and actual
power (kVA) measured in terms of voltage and
amperage. This relationship is shown in Equation
6-3.
Power Factor = Apparent power/Actual power
pf = kW/kVA (6-3)
The power factor is UNITY when the voltage and
current of an alternating current power supply are
"in-phase" with each other, for example in a purely
resistive circuit like a heating element. In a purely
777
-------
capacitive circuit, like an ozone generator that has not
reached the ionization potential of the dielectric, the
voltage and current are 90 degrees out of phase. In
this case the power factor is ZERO. For a generator
producing ozone the voltage and amperage will be
somewhere between 0 and 90 degrees out of phase;
thus, the power factor will be less than 1.0. The actual
powerfactor will vary depending on the power supply
to the ozone generator and the amount of electrical
resistance developed within the electrical circuit.
The power factor may be corrected by installing
inductors in the electrical circuit or by using the
inductance created by the operation of motors in
other areas of the treatment plant. However, caution
must be exercised when using other plant equipment
for power factor correction, because of the variable
operating conditions of the equipment from hour to
hour and from day to day. The inductance of other
equipment should only be used to control the low
power factor of the ozone generator when consistent
equipment operation can be assured.
Moisture and Dust Content of the Feed-Gas. The
moisture content of the feed-gas to the ozone
generator has a two-fold influence on ozone produc-
tion. A high moisture level not only decreases the
ozone production rate, it also increases the rate of
contamination of the ozone generator dielectrics.
Contamination occurs whether the feed-gas is high
purity oxygen or air. If oxygen is the feed-gas,
hydrogen peroxide isformed in the presence of water
vapor and forms deposits on the dielectrics; these
deposits can be removed by scrubbing with soapy
water (7).
If air is the feed-gas, about one mole of nitrogen
pentoxide (NaOs) will develop for every one hundred
moles of ozone formed (30). Nitrogen pentoxide can
decompose to nitrogen dioxide (NO2), which interferes
directly with ozone output (30). When nitrogen
pentoxide is in the presence of water vapor, nitric acid
will form (7). The nitric acid will be deposited on solid
surfaces inside the ozone generator and piping, and
will enhance corrosion of metal surfaces (7). In
addition, the nitric acid can create a heat sink on the
glass or ceramic dielectrics and will increase the
potential for dielectric breakage.
In normal practice (i.e., air dew point less than -50°C
(-58°F)), about 3g to 5g of nitric acid are formed for
every 1,000 g of ozone formed (30). Nitric acid can be
easily removed by scrubbing with soapy water. This
taskshould be considered as routine maintenance on
a 12-month interval. Significantly greater quantities
of nitric acid will form when the dew point temper-
ature of the air is higher than -50°C (-58°F). If a
malfunction occurs and excessive moisture is dis-
charged to an operating ozone generator, the di-
electrics must be cleaned before the generator is
placed back into service.
A high moisture content of the feed-gas not only
causes damage to ozone generator dielectrics and
increased generator maintenance, but also will
decrease the ozone production rate at a given power
setting of the generator. Ozone production begins to
decrease when the dew point temperature exceeds
-51 °C (-60°F)(31,32). To prevent the negative effects
of moisture content on ozone production and gen-
erator maintenance, several authors suggest a min-
imum dew point temperature of the feed-gas.
Manley and Niegowski (7) suggest that the feed-gas
moisture content should be less than 0.02 to 0.03
grams of water per cubic meter of air (i.e., dew point
temperature less than -53°C (-63°F)). Varas et al. (33)
recommend that the dew point temperature be less
than -60°C (-76°F). Robson (34) suggests that less
maintenance and prolonged dielectric life occurs
when the dew po'int temperature of the feed-gas is
-60°C (-76°F) or lower. Gerval (35) states that the
"highest value to be guaranteed for dew point should
be -50°C (-58°F)." Chapsal (36) recommends that the
dew point temperature range between -50 and -80°C
(-58 and -112°F). Damaz (37) recommends a dew
point of -60°C (-76°F) to protect the generator equip-
ment.
. Based on these sources it appears that the dew point
temperature should not be higher than -50°C, in
order to achieve maximum output from the ozone
generator. However, in order to reduce maintenance
requirements and prolong dielectric life, a dew point
temperature equal to or less than -60°C appears
warranted. To achieve either dew point condition,
extensive feed-gas treatment is required for air-fed
and oxygen recycle ozone generation systems.
Damaz (38) presents other design considerations to
prevent moisture contamination of the ozone gener-
ator. Moist air back-flow from the contact basin
should be avoided by installing a leak-proof check
valve in the piping from the ozone generator to the
ozone contact basin. Also, the ozone generator must
be purged with dry feed-gas prior to start-up of the
ozone generator. Most operators purge with dry feed-
gas for a minimum of eight hours prior to start-up of
the ozone generator. It should be noted that this time
period does not include the time required to obtain the
dry feed-gas from the dessicant dryers.
Dust and organics in the feed-gas can also create
operating problems with the ozone generator (35).
The dust can collect on the dielectrics, decrease
generator efficiency, increase dielectric stress and
cause unnecessary dielectric breakage. A filter
should be installed prior to the ozone generator to
118
-------
capture the dust and organics. According to Gerval
(35), the filter(s) should be able to remove 99 percent
of the particles greater than 1 micron in diameter and
98 percent of the particles greater than 0.4 micron in
diameter. Rakness et al. (18) also recommend two
filters in series, the first a 1 micron filter and the
second a 0.3 micron filter. It is emphasized that filters
should be installed for oxygen as well as air feed-gas
systems.
Ozone Generator Cooling. Ozone production increas-
es when the temperature of the ozonized gas is
minimized. In addition, reduced heat build-up will
increase life expectancy of the dielectrics. The feed-
gas flow rate and the heat removal capability of the
ozone generator are factors influencing the temper-
ature of the ozonized gas. The feed-gas flow rate is
typically established by the design engineer through
the selection of the maximum ozone concentration.
The heat removal capability of the ozone generator is
governed by the ozone generator equipment manu-
facturer and by the temperature of the cooling
medium.
Because cooling is a major aspect of the energy
efficiency of an ozone generator, the cooling methods
utilized by the ozone generator manufacturers are
highly competitive. The discussion of generator
cooling presented in this manual addresses general
design considerations. The various ozone generator
manufacturers must be contacted to determine the
specific cooling requirements for their individual
ozone generators.
Depending on the type of ozone generator, cooling is
accomplished with either water, oil or freon plus
water, or air. In order to optimize electrical energy
efficiently, the cooling water temperature should be
20°C (68°F) or less. Approximately 3 to 4 liters (0.75
to 1.0 gal) of 20°C (68°F) cooling water is required for
each gram of ozone produced. Different ozone
generators will vary on the amount of cooling water
required. A common design parameter is that the
temperature rise of the cooling water should not
exceed 5°C (41 °F). It is noted that some ozone
generators have operated successfully with cooling
water temperatures as high as 40°C (104°F). How-
ever, the design engineer must be aware that ozone
production capacity is lowered, electrical energy
efficiency is reduced, and dielectric glass is stressed
during operation with warm water temperatures.
Equipment capacity must be de-rated for operation
with warm temperatures.
The cooling water source can be potable water or
good quality non-potable water. The cooling system
may be closed-loop or once through and discharge.
Typically, potable water in a closed-loop cooling
system is used with the potable water re-cooled with
water from the plant non-potable water supply. The
ozone generator cooling water is often treated to
obtain "boiler" quality water, in order to prevent
scaling or corrosion in the cooling loop. For example,
water with a high chloride concentration has been
reported to "attack" stainless steel. A cost-effective
analysis can be completed to determine if it is
economically feasible to recycle the cooling water, or
to simply purchase, use and discharge the ozone
generator cooling water. In instances where the
source of cooling water is too hot, a refrigeration
cooling system has been used (1).
Air cooled ozone generators require that the gener-
ator room ventilation system be an integral part of the
ozone system design. Experience from operating
plants indicates that the cooling air must be dust-free
and oil-free in order to prevent electrical short-
circuiting on the high voltage electrical supply
system. The heated air may be used for plant heating,
but care must be taken to avoid ozone contamination
of the plant buildings in case of ozone leakage.
Feed-Gas Oxygen Content and Flow Rate. The feed-
gas flow rate and oxygen content can be used to
change the ozone concentration from the ozone
generator, and thus change the rate of ozone
production. When the oxygen concentration of the
feed-gas is greater, a higher concentration of ozone
can be obtained for a given generator power setting.
The ozone concentration increases because the
potential increases for collisions between the elec-
trons released across the discharge gap and the
oxygen molecules within the gap. In some cases
oxygen enriched air has been used to upgrade the
capacity of an existing ozone generator (39). However,
the most common use of an oxygen-fed ozone
disinfection system is direct use of high purity oxygen
and re-use of the oxygen in the activated sludge
process.
A lower feed-gas flow rate will increase the ozone
concentration. However, the specific energy con-
sumption will be greater because of less cooling
capability by the air flow. This relationship is shown in
Figure 6-20 for a typical air-fed ozone generator. As
shown, the-specific energy consumption gradually
increases from an average 12.1 Wh/g (5.5 kWh/lb) at
an ozone concentration of 4.8 g/m3 (0.4 percent wt)
to an average 23.2 Wh/g (10.5 kWh/lb) at an ozone
concentration of 30.2 g/m3(2.5 percent wt). It should
be noted that operation at a lower ozone concentra-
tion tends to reduce ozone generation specific energy,
but will require a higher feed-gas flow rate and higher
air treatment cost. A balance between these two
opposing factors may be evaluated. The maximum
ozone concentration is typically 1.5 percent wt, where
the average specific energy is 17.6 Wh/g (8 kWh/lb).
In Figure 6-20 a range of specific energy consump-
tions is shown to account for different cooling
179
-------
capabilities of various; ozone generators and different
feed-gas temperatures that may be encountered. It
should be noted that the data shown is for a high
quality feed-gas through clean ozone generators;
thus. Figure 6-20 can be used to estimate energy
consumption for typical ozone generation systems.
The manufacturer should be contacted for more
precise information on performance characteristics
of individual ozone generators.
The information from Figure 6-20 may be rearranged
to show an ozone "generator mapping" curve by
selecting different gas flow rates. The mapping curve
may be used to evaluate specific energy requirement
relative to ozone production and assist in the selection
of feed-gas flow rate. An example generator mapping
curve based on actual test results is shown in Figure
6-21 (17). The approach to developing the generator
mapping curve is as ifollows: (1) Select a desired gas
flow rate, (2) Calculate the ozone production at a
selected ozone concentration, (3) Plot the data point
(i.e., calculated ozone production at specific energy
requirement for corresponding ozone concentration),
(4) Repeat steps 1 -3 for various ozone concentrations,
and (5) Select another gas flow rate and repeat steps
1-4.
The design ozone concentration for an air-fed ozone
generator typically ranges from 12to24g/m3(1 to 2
percent wt). A concentration of 18 g/m3 (1.5 percent
wt) is most common. While individual ozone gener-
ators may achieve concentrations greater than de-
sign, it is recommended that the discharge ozone
concentration from all ozone generators be main-
tained at or below design conditions in order to
maximize generator cooling effectiveness. Also, it is
recommended that individual ozone generators not
be operated at power settings greater than 75 percent
of maximum, unless necessary. For example, rather
than operate one ozone generator at 100 percent of
capacity, 2 ozone generators should be operated at 50
percent capacity. These operating practices will cause
less stress on generator dielectrics, will decrease
generator maintenance problems, and will usually
minimize electrical consumption.
The typical range of ozone concentration and specific
energy consumption for an oxygen-fed ozone gener-
ator is shown in Figure 6-22. The typical (40 g/m3)
design ozone concentration for an oxygen-fed ozone
generator at 3 percent wt, is twice the concentration
for an air-feed system. Also, the specific energy is
about half of an air-feed system.
It should be noted that the energy consumption
information shown in Figure 6-20 and Figure 6-22 is
for the ozone generator only. These data must be
coupled with the energy consumption of the auxiliary
equipment in order to determine the overall energy
Figure 6-20.
Specific energy consumption versus ozone
concentration for an air-fed ozone generator.
30
Ozone Concentration (% by Weight)
0.5 1.0 1.5 2.0
2.5
o>
v.
25
I 20
Q.
3
10
111
o
a. 5
CO
12
11
10
9
8
7
6
5
4
3
2
1
Q.
(O
5 10 15 20 25
Ozone Concentration (g/m3)
30
Figure 6-21. Example ozone generator mapping curve
using air feed-gas (17).
600
— 500
jz
CO
400
300
200
100
8.3 l/s
6.9 l/s
5.6 l/s
4.2 l/s
- 2.8 l/s
2.1 l/s
12 14 16 18 20 22 24 26 28 30 32 34 36
Specific Energy (Wh/g)
consumption of the ozone disinfection system. More
information on energy consumption on auxiliary
equipment is discussed in Section 6.5.2.
Ozone Generator Monitoring and A/arms. The follow-
ing are required to adequately monitor ozone genera-
tion equipment:
a. Inlet feed-gas flow rate—Monitor system load-
ing.
b. Inlet feed-gas temperature—Monitor system
operation. Also provide alarm and shut-down
device.
720
-------
Figure 6-22. Specific energy consumption versus ozone
concentration for an oxygen fed ozone gen-
erator.
15
x 12.5
I 10
Q.
E
Ozone Concentration (% by Weight)
1.0 2.0 3.0 4.0
01
c
o
O
c
o
.
o>
10 20 30 40 50
Ozone Concentration (g/m3)
60
c. Inlet pressure—Monitor system operation.
d. Discharge ozonized gas temperature—Monitor
system operation. Also provide alarm and
shut-down.
e. Discharge ozone concentration with in-line
ozone meter—Monitor system performance.
Also provide a recorder. Monitor may be con-
sidered optional for smaller installations.
f. Discharge ozone concentration using wet-chem-
istry procedures (40)—Monitor system concen-
tration.
g. Inlet voltage—Monitor system loading.
h. Inlet amperage—Monitor system loading.
i. Inlet frequency—Monitor system loading.
j. System watt-hour meter—Monitor system
energy. May be considered optional for smaller
installations.
The operator should know the various aspects of
system loading, performance and operation to
be able to determine production levels and
operating characteristics. The alarm and shut-
down on the inlet and discharge gas temperature
is used to protect the generator from damage.
The recorder on the discharge ozone concentra-
tion meter is used to monitor generator per-
formance. This meter and recorder may be
considered optional equipment on very small
systems, but the operators should at least be
provided with the equipment and procedures for
determining ozone concentration using wet-
chemistry methods.
k. The cooling water system should be designed to
achieve adequate cooling for the worst case
conditions.
I. Cooling water system monitoring equipment
should provide detailed information regarding
system operation:
« Water temperature—Monitor system opera-
tion. Also provide alarm and shut-down.
9 Inlet water flow rate—Monitor system opera-
tion. Also provide alarm and shut-down.
e Outlet water temperature—Monitor system
operation.
o Heat exchanger cooling water inlet and outlet
temperature and inlet water flow rate—
Monitor system operation.
The cooling water system is vital to the
operation of the water-cooled ozone genera-
tors. The operators should be provided with
the instrumentation to be able to make control
adjustments, as necessary, to maximize
generator cooling. The alarm and shut-down
on the water flow rate and temperature are
provided to protect the generator dielectrics
from breakage due to heat stress.
m. Cooling air system monitoring equipment should
provide detailed information regarding system
operation.
« Ambient air temperature—Monitor system
operation.
The air cooling system is primarily dependent
upon the room temperature. The "cooling"
fans typically operate at a constant speed.
The cleanliness of the cooling air is important
to proper cooling, although no direct measur-
ing is typically done to monitor this parameter.
The air should be as cool as possible to
maximize generator cooling, and as clean as
possible to prevent dust accumulation on the
cooling surface. Also, no oil should be used,
because an oil film will develop on the cooling
surface.
n. Inlet feed-gas temperature with alarm and shut-
down—Monitor system operation and protect
dielectrics.
727
-------
o. Inlet feed-gas dew point with alarm and shut-
down—Monitor system operation and protect
dielectrics.
p. Safety inter-lock to prevent power energizing
during generator cleaning or repair.
6.4.1.3 Types of Ozone Generators
Ozone generators are typically classified by their
control mechanism, cooling mechanism and physical
arrangement of the dielectrics. Another method of
describing an ozone generator is by name of the
inventor. The control mechanism may be a voltage or
frequency unit. The cooling medium may be water,
water plus oil or freon, and air. The physical ar-
rangement of the dielectrics is typically either vertical
or horizontal.
Figure 6-23. Details of a horizontal tuba, voltage controlled,
water cooled ozone generator (1).
Horizontal Tube, Voltage Controlled, Water Cooled.
The horizontal tube, voltage-controlled, water-cooled
ozone generator is the most commonly used (1).
Figure 6-23 shows details of the ozone generation
equipment. The feed-gas enters one end of the ozone
generator and the ozonized gas exits the opposite
end. The stainless steel jacket acts as the low tension
electrode, and contains multiple, cylindrical tubes
where glass dielectrics are inserted. The internal side
of the glass dielectric is coated with a metal lie coating
which acts as the high tension electrode. The feed-
gas passes between the external side of the glass
dielectric and the stainless steel jacket.
The equipment is normally designed to operate at
pressures up to 103 kPa (15 psig). The majority of
these generators operate at an electrical frequency of
60 hertz, although operation at frequencies from 600
to 800 hertz has been practiced (1). The horizontal
tube, voltage-controlled, water-cooled ozone gener-
ators are installed at a number of wastewater
treatment plants including: Frankfort, Kentucky; Vail,
Colorado; Pensacola, Florida; Murfreesboro, Tennes-
see; Brookings, South Dakota; and Olympia, Wash-
ington.
Vertical Tube, Voltage Controlled, Water-Cooled. The
vertical tube, voltage controlled, water-cooled ozone
generator utilizes the cooling water both as the
grounding electrode and the coolant. Details of the
generation system are shown in Figure 6-24 (1). Input
gas is "pulled" through the dielectrics by means of a
vacuum created in the ozone contacting system.
The ozone generator consists of three compartments.
The feed-gas is drawn into the upper compartment
where it enters hollow metal tubes which are the high
tension electrodes. The gas is drawn downward
within the tubular metal electrode to emerge into the
closed end of a glass dielectric tube. The feed-gas
then passes upward through the corona discharge,
G J F E
Legend
A-Air Inlet
B-Ozonized Air Outlet
C-Coolant Inlet
D-Coolant Outlet
E-Dielectric Tube
F-Discharge Zone
G-Tube Support
H-H.V. Terminal
l-Port
J-Metallic Coating
K-Contact
which is created between the tubular stainless steel,
high tension electrode and the glass dielectric, low
tension electrode. The ozonized gas is discharged into
the middle compartment, from where it is drawn to
the ozone contactor.
The vertical tube, voltage-controlled, water-cooled
ozone generator is coupled with a proprietary,
aspirating turbine mixer contacting system. The
process was installed at the first wastewater treat-
ment plant to use ozone disinfection in the United
States (Indiantown, Florida), and was evaluated in ah
EPA research study (17).
722
-------
Figure 6-24. Details of vertical tube, voltage controlled.
water cooled ozone generator (1).
Figure 6-25. Details of a vertical-tube, frequency control-
led, double-cooled, ozone generator (1).
1
y/M4W/
o
>- E N.
LL.
"
i ^. uoonng wate
1 Discharge
Air or
•c^
)zone Out
Lowther Plate, Frequency-Controlled, Air-Cooled.
The basic elements of the Lowther plate, frequency-
controlled, air-cooled ozone generator are shown in
Figure 6-26 (1). The dielectric consists of an alum-
inum heat dissipator, a steel electrode coated with a
ceramic material, a silicone rubber spacer to establish
the discharge gap, a second ceramic coated steel
electrode with gas inlet and ozonized gas outlet, and a
second aluminum heat dissipator. Thirty to forty
dielectrics are manifolded to form a module. Atypical
Lowther plate module is also shown in Figure 6-26.
The unit uses ambient air for cooling and operates at
an upper frequency range of 2,000 hertz, at a
potential of 9,000 volts. The maximum operating
pressure is 103 kPa (15 psig).
A Lowther plate ozone generator was installed at
Springfield, Missouri; Holland, Michigan; Moorhead,
Minnesota; and Concord, North Carolina. Another
type of air-cooled generator was installed at the
Mission Viejo plant near Denver, Colorado (41).
723
-------
Figure 6-26. Details of an air-cooled, Lowther plate type
ozone generator (1).
Exhaust
Cooling Air
8'-7
Recessed
Control'
Panel
Cooling
Air Intake
Ground
High Voltage __ r* A Steel
Steel Electrode \ ,5*,-Electrode
Aluminum
Heat Dissipator -
P)
Ozone Generator
Cells
Solid State
-Electronics
Behind Lowered
Panels
Cooling Air Intake
Silicone
Rubber
Separator
Ceramic
Dielectric
Coated
Steel
Electrode
Ceramic J5|
Dielectric
Silicone
Rubber
Separator
Section A-A
6.4.1.4 Specific Manufacturers' Ozone Generators
Ozone generators manufactured by different com-
panies have unique characteristics, but also have
some common requirements. For example, all ozone
generators should have a similar quality feed-gas,
and most ozone generators can operate independent
of the type of contact basin used. A listing of ozone
generator manufacturers can be obtained from the
International Ozone Association (IOA).
6.4.2 Feed-Gas Supply and Treatment Equipment
Feed-gas to an ozone generator may be air, once-
through oxygen, or recycled oxygen, as discussed in
Section 6.3.1. Typically, either air or once-through
oxygen is used. Once-through oxygen systems typ-
ically do not require further moisture removal.
However, filters to remove particulates should be
provided. Air and oxygen recycle feed-gas ozone
generation systems must provide treatment to re-
move both excess moisture and particulates.
Inadequate air or oxygen recycle feed-gas treatment,
especially inadequate moisture removal, has caused
several ozone system problems and failures of ozone
treatment processes (4,42). As such, air treatment is
a critical aspect in ozone system design. The major
equipment design considerations are the quality of
the treated air and the reliability of the treatment
equipment.
6.4.2.1 Quality of Feed-Gas
The quality of the feed-gas for ozone generation is
described by its moisture content and particulate
content. Moisture content is mass of water per unit
volume of gas (lb/1,000 ft3), but the design spec-
ification for moisture content is typically expressed as
dew point temperature. The relationship between
moisture content and dew point temperature is
discussed in Section 6.2.3.3.
Excessive moisture content in the feed-gas to the
ozone generator can decrease ozone production, can
increase generator maintenance, and can cause
damage to internal components of the generator. To
maximize ozone production and minimize mainte-
nance problems, it is recommended that the dew
point temperature of the feed-gas to all corona
discharge ozone generators not exceed -60°C (-76°F),
unless absolutely necessary. Operation at a dew
point temperature greater than -50°C (-58°F) should
be avoided altogether.
The primary function of the air treatment system is to
remove moisture from the air. Moisture can be
removed by increasing pressure, by cooling, or by
adsorption techniques (21). Equipment used to re-
move moisture is further described, The procedure for
calculating moisture content at various temperature
and pressure conditions is discussed in Section
6.2.3.3.
The particulate content of the feed-gas should also be
minimized in order to optimize generator production
efficiency and reduce maintenance. Prior to the ozone
generators, filters which remove particles greater
than 1 micron in diameter followed by filters which
remove particles greater than 0.3 to 0.4 micron in
diameter are recommended (18,35).
6.4.2.2 Air Feed-Gas Treatment Systems
Air feed-gas systems are typically classified by their
operating pressure. The most common is a low
pressure system, which operates at a pressure
ranging from 69 to 103 kPa (10 to 15 psig); although
pressures up to 275 kPa (40 psig) have been reported
when the pressure is reduced prior to the ozone
generator. High pressure systems operate at a
pressure ranging from 480 to 690 kPa (70 to 100
psig), reduce the pressure prior to the ozone gen-
erator, and are typically used in "small to medium"
sized applications. Either system may be used in
conjunction with most of the ozone generation
equipment discussed in Section 6.4.1.3, and all of the
ozone contacting systems discussed in Section 6.4.3.
It should be noted that the decision to use a high or
724
-------
low pressure system is often based on a qualitative
evaluation of potential maintenance requirements, in
addition to the quantitative capital cost evaluation.
Some of the issues to consider are listed below:
a. The high pressure air pre-treatment equipment
generally has a higher maintenance require-
ment for the air compressors.
b. The high pressure air pre-treatment equipment
generally has a lower maintenance requirement
for the desiccant dryers.
c. The high pressure air pre-treatment system
generally has a lower capital cost. At small to
medium sized installations this lower capital
cost may offset the additional maintenance
required for the air compressors and associated
equipment, such as filters for the oil type
compressors. The engineer should investigate
the potential maintenance associated with the
high and low pressure systems rather than
evaluating the design on capital cost alone.
Another type of air feed-gas treatment system is the
"nominal pressure" system, which typically operates
at a negative or in some cases a slightly positive
pressure. The nominal pressure process is a propri-
etary process used in conjunction with the Kerag
ozone generator and aspirating turbine mixer ozone
contactor.
Pressure System Flow Schematic. A schematic
diagram of a low pressure air treatment system is
shown in Figure 6-27. The diagram illustrates a dual
component process, and shows desired flexibility for
the equipment provided. More information on process
flexibility design considerations is discussed in
Section 6.5.1.3.
The pre-compressor filters are provided to protect the
air compressors from damage due to large particles.
The air compressors are typically positive displace-
ment, oil-less units. Positive displacement compres-
sors are used in order to obtain constant air flow at
variable operating pressures. Variable pressures are
often encountered due to variable pressure losses in
downstream equipment and processes such as filters
and ozone contact basins. Oil-less compressors are
used to eliminate oil contamination of the down-
stream desiccant dryer medium and ozone generator
dielectrics. Liquid-seal and rotary lobe compressors
have been used most frequently.
The compressors may be followed by an after-cooler
or a refrigerant dryer. These components are depicted
by dotted lines in Figure 6-27, which indicate that
they are optional. Typically, either one or the other
option is provided. These cooling mechanisms are
used to remove moisture in the air at minimal
operating expense.
Figure 6-27. Example low pressure air feed-gas treatment
schematic.
Pre-Compressor
Filter
Pre-Compressor
Filter
Compressor
Compressor
.1 , f"]-Optional | __] ^
-Cooler ! j After-Cooler !
« ^_ __l I— — —_ ^_._ _ _ ^_aJ
! After-Cooler !
u --- .
r
| Refrigerant
i Dryer —
L_
~l
1
1
1
J
[ Refrigerant
i Dryer
1
I
1
Pre-Desiccant
Filter
I
I
Pre-Desiccant
Filter
Tower A
1
Tower B
Heat-Reactivated
Desiccant Dryers
^ ^
Tower A
1
Tower B
1
Post-Desiccant
Filter
I
L
Post-Desiccant
Filter
J
To Ozone Generator
To Ozone Generator
The compressed, cooled air is directed to a pre-
desiccant filter, which is used to remove dust and dirt
particles greater than 3 to 5 micron in diameter.
Particulate removal prior to the desiccant dryers
reduces plugging in the desiccant medium.
Probably the most important component of the air
treatment prodess is the desiccant dryer. The desic-
cant dryer consists of two towers containing moisture
adsorbing media. One tower operates while the other
tower is regenerating. The low pressure system
desiccant dryer uses heat for reactivation of the
desiccant.
The post-desiccant filters are installed to remove
particulates less than 0.3 to 0.4 micron in diameter.
Two-stage filtration is preferred. The first stage filter
removes particulates greater than 1 micron and the
second stage removes particulates less than 0.3 to
0.4 micron in diameter (18,33,35).
725
-------
High Pressure System Flow Schematic. A schematic
of a high pressure air treatment system is shown in
Figure 6-28. The pre-compressor filters are used to
remove larger particulates and protect the air com-
pressors. The compressors are typically oil-less units;
however, oil-seal compressors followed by extensive
oil removal equipment have been utilized.
Figure 6-28. Example high pressure air feed-gas treatment
schematic.
Pre-Compressor
Filter
Pre-Compressor
Filter
Compressor
Compressor
After-Cooler
After-Cooler
Pre-Deslccant
Filter
Pre-Desiccant
Filter
Tower A
Tower B
Heat-Less
Desiccant Dryers
Tower A
Tower B
Post-Desiccant
Filter
Post-Desiccant
Filter
PRV I-*- Pressure Relief Valve
To Ozone Generator
The after-coolers following the high pressure com-
pressors are essential, as they are used to remove the
heat of compression. The f ilter(s) before the desiccant
dryer are used to remove particulates less than 3 to 5
micron in diameter when oil-less compressors are
used. When oil-seal compressors are used, filtration
to remove oil droplets less than 0.03 micron is
provided (33).
The high pressure system desiccant dryer consists of
two towers with moisture adsorbing media. One
tower operates while the other tower is regenerating.
Regeneration is accomplished without additional
heat; thus, the high pressure desiccant dryers are
called heat-less units. The post-desiccant filters
remove particulate matter less than 0.3 to 0.4 micron
in diameter. The high pressure system also has a
pressure reducing valve to regulate operating pres-
sures in the ozone generator.
Nominal Pressure System Flow Schematic. A sche-
matic diagram of the nominal pressure air treatment
system is shown in Figure 6-29. The nominal
pressure system typically operates at a pressure
slightly below ambient pressure conditions. In cases
where a chiller is used to cool the air, a blower maybe
used to overcome the pressure drop within the chiller
and the system may then operate under a slightly
positive pressure. The nominal pressure system is a
proprietary process used in conjunction with an
aspirating turbine mixer contacting unit.
Figure 6-29. Example nominal pressure air feed-gas treat-
ment schematic.
Pre-Filter
'L-_~i:n-—J
^ i i Chille" !
:T
£ J- Optional
! Chiller ! j
Tower A
Tower B
Heat Reactivated
Dessicant Dryers
Tower A
Tower B
Post-Desiccant
Filter
Post-Desiccant
Filter
To Ozone Generator
To Ozone Destruct
To Ozone Destruct
The primary difference between the nominal pres-
sure system and the other air treatment process
schematics is the method of moving the air. With the
nominal pressure system the air is drawn through the
air treatment process and through the ozone gener-
126
-------
ators with the vacuum created by the operation of the
turbine mixer. The pre-desiccant filter is used to
remove particulates greater than 5 microns in
diameter, and the post-filter removes particles greater
than 1 micron in diameter(33). The nominal pressure
system uses a heat-reactivated desiccant dryer.
The energy input and air supply to the turbine mixing
contacting unit can be controlled by a variable speed
drive on the mixer to adjust its pumping action, and by
adjusting the orifice size of the water inlet to the
mixer. The specific details of achieving this flexibility
can be arranged with the manufacturer.
/
6.4.2.3 Air Compressor Design Considerations
Air compressors are used for the low and high
pressure air treatment processes. Typically, oil-less
compressors are used. However, for high pressure
applications oil-seal compressors have been used
when they are followed with extensive oil removal
equipment.
Low Pressure Air Compressors. Low pressure air
compressors operate at a pressure ranging from 69 to
103 kPa (10 to 15 psig), although operation up to 276
kPa (40 psig) has been reported when the pressure to
the ozone generator is reduced. The following design
considerations are applicable for low pressure air
compressors:
a. Positive displacement compressors have been
used most frequently because good control over
air flow rate is achieved at variable operating
pressures encountered.
b. The air compressors should be located in a
remote area to avoid noise disruption, or should
be located in a sound-proofed housing.
c. Liquid-seal compressors often have been used
for ozone, despite a higher capital cost and in
some cases a slightly higher energy require-
ment. These disadvantages were acceptable
considering an overall design objective of reli-
able, continuous performance. It should be
noted that when liquid-seal compressors are
used the size of the downstream moisture
removal equipment does not change because
that equipment is designed to treat saturated air
no matter what type of compressor is used.
d. Air compressors should be followed by an after-
cooler or by refrigerant dryers. Water-to-air
after-coolers are generally used, rather than air-
to-air after-coolers.
e. Instrumentation that should be provided to
monitor and control air compressor equipment
operation is:
• Discharge pressure—Monitor system opera-
tion.
• Discharge temperature—Monitor system per-
formance. Also include system alarm and
shut-down.
• Discharge flow—Monitor system perform-
ance.
• Seal water temperature, when liquid-seal
compressors are used—Monitor system oper-
ation.
• Seal water flow—Monitor water availability.
Also include system alarm and shut-down.
f. Instrumentation that should be provided to
monitor and control after-cooler operation is:
• Feed-gas temperature—Monitor system load-
ing.
• Discharge temperature—Monitor system per-
formance.
• Cooling water inlet and outlet temperature
—Monitor system operation.
• Cooling water flow rate—Monitor system
operation.
High Pressure Compressors. The following design
considerations are applicable for high pressure air
compressors:
a. Non oil-lubricated, high pressure compressors
(piston type compressors that use a Teflon seal)
should be designed with a duty cycle less than
60 percent and should be coupled with pre-
planned methods of replacing the Teflon seal on
a periodic interval. A replacement such as once
per year may occur.
b. High pressure compressors may be oil-lubri-
cated if followed by a sophisticated oil removal
process. Successful performance has occurred
with a cyclonic moisture/oil separation device,
followed by an impingement filter, followed by a
coalescing filter. The final filter should remove
oil particles greater than 0.03 micron in diam-
eter. At the same time, it should be anticipated
that the medium in the desiccant dryers may
need to be replaced quite frequently (e.g., every
12 months).
c. Instrumentation should be provided for process
monitoring and control, as discussed previously
in the Low Pressure Air Compressor section.
6.4.2.4 Refrigerant Dryer Design Considerations
The value of refrigerant dryers has been questioned
at some operating installations. Some problems that
have occurred with their use include sensitivity to
icing because operating temperatures are quite close
to the freezing point of water and non-availability of
maintenance expertise required to repair the refrig-
727
-------
eration equipment (42). If these problems can be
overcome, refrigerant dryers can be very effective
because they can remove excess moisture in the air
for a small amount of energy consumption. The
decision to use a refrigerant dryer is typically based
on cost savings that can result in reducing the size of
the desiccant dryer because of the moisture removed
by the refrigerant dryer (21).
The following design considerations are applicable
for refrigerant dryers;;
a. If the desiccant dryers are reduced in size such
that operation of the refrigerant dryer is essen-
tial to the overall treatment process, back-up
equipment and spare parts must be provided to
insure continued operation of this important,
sensitive-to-operate piece of equipment.
b. Plant maintenance personnel must be trained
on refrigeration equipment, or refrigeration
equipment service and repair must be available
on a responsive contract basis.
c. Instrumentation and control equipment that
must be provided to prevent overload of the
downstream desiccant dryers are:
• Inlet temperature—Monitor unit loading.
• Outlet temperature—Monitor unit perform-
ance. Also provide an alarm and shut-down
for high temperature.
• Feed-gas flow rate—Monitor system loading.
6.4.2.5 Desiccant Dryer Design Considerations
The desiccant dryer is the most important unit in the
air treatment system. Poor performance of the
desiccant dryer will reduce ozone production and
damage internal components of the generator.. De.sic-
cant dryers used for ozone systems are special
application units, since "off-the-shelf"dryerstypical-
ly are capable of achieving dew points in the range of
only -40°C (-40°F) or higher. Dew points of -60°C
(-76°F) and lower are required for ozone generation. It
should be noted that dew point temperature is a
function of pressure, and recorded dew point temper-
atures always should be referenced to pressure. In
this manual all dew point temperatures are based on
standard pressure conditions, unless otherwise
noted. Refer to Section 6.2.4.3 for the procedure to
adjust dew point temperature to standard conditions.
Heat-Reactivated Desiccant Dryers. Heat-reactivated
desiccant dryers are used in nominal pressure and
low pressure air treatment systems. A schematic of a
heat-reactivated desiccant dryer is shown in Figure
6-30. Wet-air is directed to the operating tower
where moisture in the air is adsorbed onto the
desiccant. The desiccant is typically activated alumina
and molecular sieves (21). In some instances silica
gel is used. After several hours of operation the
desiccant becomes saturated with moisture and is
unable to maintain the desired dew point. At that time
the desiccant must be regenerated. The cycle valve(s)
switch the operating tower so the "used" medium
can be regenerated. It should be noted that the switch
valve is an important component in the dessicant
dryer. Stainless steel valves have been reported to
work best. Also, routine maintenance is necessary to
make sure the valves switch every time.
Figure 6-30. Diagram of a heat-reactivated desiccant dryor
with internal heating coils.
Pressure
Relief Valve
"Wet-Gas"
Pressure
Relief Valve
Drying
Tower
Internal
Heating'
Coil
Shroud
x *"
f
Des
M(
*,
W- iniei _
X I ^
« Nfl T t^fi
N.^ *^1 w.
JJ 1
^
CCc
;di
in
3
t
i
Check
Valve y
(
\
\
-2
Purge
i Air
Filter
1 Rotometer
!l Flow
rControl
i Valve
| 1
ft
D
esic
Me
r
i
U. I
\
cant
dia
111
Regenerating
Tower
Control
Valve
Check
Valve
Process
Air
Filter
To Ozone Generator
Regeneration involves heating the desiccant to a
temperature between 90 and 260°C (200 and 500°F)
and purging with dry air (21). The typical design is for
heating to a temperature between 120 and 170°C
(250 and 350°F) for 1 to 2 hours, and cooling for a
minimum of 6 hours. Heating enhances moisture
evaporation. The evaporated moisture is removed by
purge air. After heating is completed the purge air will
continue to flow through the tower and cool the
desiccant before the regenerated tower is placed back
into service. It is extremely important that the
desiccant be cooled before being placed back into
service so that a "temperature spike" does not occur.
An aftercooler may be considered after the dessicant
dryer and before the dew point analyzer, with alarm,
in order to insure that a "temperature spike" does not
728
-------
occur. The dew point analyzer is used to prevent
possible moisture contamination due to a leak in the
aftercooler or excessive moisture from the dryer.
From 5 to 20 percent of the dry air is recycled (called
purge air) to the regenerating tower. If the purge air
flow is inadequate, hot air and excess moisture will
remain in the regenerated tower. This moisture laden
air will be discharged to the ozone generator when
the regenerated tower begins its operating cycle, and
will cause damage to generator components. To
prevent unexpected or unnoticed loss of purge air
flow, it is essential that the purge air flow be
monitored and controlled. It is important to note that
this is typically an option that must be specified in
detail.
The purge air is filtered to protect the equipment,
particularly the air supply control or to weir switching
valves. The schematic diagram in Figure 6-30 shows
that the source of the purge air is subsequent to the
process air filters, which remove particles greater
than 0.3 to 0.4 micron in diameter. This piping
scheme provides initial filtration of the purge air and
decreases the frequency of plugging of the smaller
filters on the purge air line. Eliminating plugging is a
very important operational consideration, given the
importance of constant availability of purge air flow.
After months of operation the desiccant will lose its
moisture adsorbing capability. A life expectancy of 36
months is considered normal but can range from 12
to 60 months. The factors affecting the desiccant life
are frequency of heating and cooling, plugging, initial
capacity(i.e., pounds of desiccant per Ib/d of moisture
loading), and method of heating.
An internal heating mechanism is shown in Figure
6-30. Internal heating is typically used in smaller
installations because of capital cost considerations.
However, internal heating coils often do not result in
even heating of the desiccant. The desiccant closest
to the heating elements will be over-heated and will
deteriorate faster. In the desiccant dryer system
shown in Figure 6-31, external heating capability is
illustrated. The purge air flow is heated and directed
to the regenerating tower. External heating is pre-
ferred because better distribution of heat throughout
the desiccant is provided (37), but this system is more
complicated because of the additional equipment
requirements.
The amount of desiccant used and the regeneration
cycle time are the most important design considera-
tions. The cycle time must be long enough to obtain
sufficient cooling of the regenerating tower, and
short enough to maintain the desired dew point by the
operating tower. The recommended minimum cycle
time, because of cooling considerations, is eight
hours (i.e., time from beginning of regeneration mode
to end of operating mode). However, the design cycle
time should be longer (e.g., preferred 16-hr and
minimum 12-hr) because the desiccant loses its
moisture adsorbing capability in a few months of
operation. At initial start-up the cycle time should be
longer in order to give the operator f lexibi lity to reduce
the cycle time as the desiccant loses its moisture
adsorbing capacity. When the minimum, 8-hr cycle
time is reached, the operator must replace the
desiccant and can then re-set the cycle time to the
original design settings.
Figure 6-31. Schematic of a heat-reactivated desiccant
dryer with external heating equipment.
Relief Valve "Wet-Gas" Relief Valve
v
i
i
rr Inlet v
1
^ • Check i •
S N r, , -P> Valve / — ^\
f \ External ~ / \
Desiccant
Media
v
Heater = 3 Purge
|j
^
2
/ *
. — ''Purge JT
Air ,
i
V •
Check
Valve
Filter
-J
Rotometer
. Flow
Control
Valve _ „ .
To Dram
1 \
Desiccant
Media
A \ J
Control 1 Control N^_^^X
Valve I Valve 1
' l^i^ ' f^rd <^
Purge Air Outlet
) k i 'I
i ?
' w \ r *
Check
r-L-i Valve
Process
Air
1 Filter
To Ozone Generator
The regeneration cycle may be controlled on a timed
or on a demand basis. The timed basis uses a pre-
established time to initiate the regeneration cycle,
whether or not the operating (drying) tower has
exhausted its capability to remove moisture and
achieve the desired dew point. The timed cycle
control is adjusted by changing the timer settings.
A regeneration cycle based on demand is more
expensive to install because of the controls involved.
However, the system has potential O&M cost savings.
729
-------
Using the demand regeneration system the cycle is
initiated when a pre-set maximum value for the dew
point of the discharge air from the dryer is reached.
Dew point temperature is monitored on a continuous
basis. Energy savings occur because the cycle time,
and corresponding heating time would be activated
only when the tower has exhausted its drying
capability. Further, the life expectancy of the desic-
cant is prolonged because it would not be subjected to
the stress of reactivation (i.e., heating and cooling) as
often.
Demand regeneration control has some definite
advantages, but has important operational considera-
tions that must be addressed, namely the reliability of
the dew point monitor and sensitivity of the control
logic. Ozone equipment suppliers, desiccant dryer
manufacturers, and existing ozone plant operating
personnel may be contacted to determine which
meters have been successfully used and how sensi-
tive the control logic must be. Successful perform-
ance has been obtained from various systems, but
problems have been encountered with others. All in-
line dew point monitors should be routinely checked
by performing a "dew point cup" test (32,42).
A word of caution is noted in operation of all dew point
probes. When the measuring probe gets wet due to a
moisture spike, it takes several hours of operation
under dry conditions before the probe will accurately
reflect the correct dew point temperature reading.
The amount of desiccant and the operating pressure
(i.e., absolute pressure) controls the moisture removal
capability of the process. Once the type of air
treatment process is selected (i.e., low or high
pressure), operating pressure is not considered
further in design. This leaves amount of desiccant as
a most important factor. Inadequate desiccant can
cause operating problems in a short period of time. At
one plant with insufficient desiccant, dew point break-
through occurred within 5.5 hours of an 8-hr drying
time and the desiccant was only 3 months old. The
performance objective was a dew point temperature
of -50°C (-58°F). At this plant the desiccant moisture
loading was 13 Ib desiccant/lb of moisture loading in
an 8-hr period.
Apparently, 13 Ib desiccant/lb of moisture loading is
insufficient. Varas suggests that at least 18 Ib
desiccant be provided per Ib of moisture received
during one cycle of operation (33). Damez suggests
that at least 5.5 Ib of desiccant be provided per Ib/d of
moisture received, which corresponds to 16.5 Ib
desiccant/lb of moisture received during an 8-hr
regenerating cycle (37). These suggested desiccant
amounts may be used as guidelines for the design of
desiccant dryers, but should not be considered the
final design criteria because the type and quality of
desiccant are also very important.
A recommended design approach is to specify that a
minimum dew point of -60°C (-76°F) (corrected to
standard pressure conditions) will be achieved during
a minimum cycle time of 1 6 hr while being operated
at the maximum expected moisture loading condi-
tions(i.e., maximum temperature, maximumflowand
minimum pressure). In addition, a minimum amount
of desiccant for each operating tower should be
specified.
A summary of the design considerations for heat-
reactivated desiccant dryers is listed below:
a. The tower switching mechanisms should be
designed for long life and must be maintained
per the manufacturer's recommendations to
insure continued, reliable performance.
b. The desiccant dryers should be designed with
sufficient robm around the units to allow for
inspection or repair of the heating elements, for
maintenance and repair of the tower switching
mechanism, and for ease of replacement of the
desiccant.
c. The heating coils used in internal heating, heat-
reactivated desiccant dryers should not be
contacted directly with the desiccant. A protec-
tive shroud should be provided.
d. The decision to use a timed or demand reactiva-
tion cycle control system should be based upon
economic considerations. Demand reactivation
cycle systems are generally desireable, espe-
cially for larger systems. For either control
system the operators must routinely check and
calibrate the in-line dew point monitor to assure
reliable performance from this critically impor-
tant instrument.
e. The minimum pperating cycle time of a heat-
reactivated desiccant dryer should be 8 hr, in
order for the regenerated bed to cool down
before being placed back into operation. The
design cycle time should be 16 hr (minimum 12
hr) to allow for decreasing this time as the
desiccant gradually deteriorates.
f. The process air compressors must be sized to
also handle the desiccant dryer purge air flow
rate, which may be as much as 20 percent of the
ozone generation process air flow rate. Reliable
process monitoring and control instrumentation
that should be provided are:
• Feed-gas flow rate—Monitor system loading.
• Inlet temperature—Monitor system loading.
• Outlet temperature—Monitor system per-
130
-------
formance. Also include alarm and system
shut-down capability.
• Inlet and outlet pressure—Monitor system
operation.
• Purge air flow rate—Monitor system opera-
tion.
• Discharge feed-gas dew point, in-line meter—
Monitor system performance. Also include
recorder and system alarm and shut-down.
• Discharge feed-gas dew point, dew point cup
measurement—Measure dew point by man-
ual methods to check on the reliability of the
in-line meter.
Heat-Less Desiccant Dryers. The heat-less desiccant
dryers typically operate at a pressure ranging from
480 to 690 kPa (70 to 100 psig) (21). Another name for
a heat-less desiccant dryer is a pressure swing
desiccant dryer because the unit operates with
varying pressures ranging from high pressure to low
pressure. Heat-less desiccant dryers are considered a
viable alternative for moisture removal for small to
medium sized ozone disinfection applications.
A schematic diagram of the heat-less desiccant dryer
is shown in Figure 6-32. The desiccant used is
typically activated alumina and molecular sieves. The
principle of operation of the heat-less desiccant dryer
is adsorption of the moisture onto the desiccant under
a high pressure. After a period of time ranging from 1
to 5 minutes the drying towers are switched. The
tower to be regenerated is reduced to atmospheric
pressure conditions and purged with "dry" air from
the operating tower. The moisture that has adsorbed
onto the desiccant is evaporated into and carried
away by the "dry" purge air because of the lower
pressure air's capacity to hold much more moisture
(21).
The purge air flow rate for a heat-less desiccant dryer
is normally 15 to 25 percent of the process air flow
rate. It is important to note that the compressors must
be sized to accommodate this additional flow require-
ment. The amount of moisture that is removed and
the corresponding dew point is lowered when the
amount of desiccant is increased and/or the cycle
time is decreased.
The heat-less desiccant dryers have a history of
reliable performance, if properly sized and main-
tained. The tower switching mechanism must be
maintained per the manufacturer's instructions, and
the amount of desiccant must be adequate. Over a
period of time (1 to 5 years) the desiccant will lose its
moisture adsorbing capability and will have to be
replaced. This process will occur gradually and must
be monitored closely. The design considerations for
heat-less desiccant dryers are similar to those for
Figure 6-32. Pressure swing (heat-less) high pressure desi-
ccant dryer in purging mode.
Moist Air
Intake
Operating
Tower
Regenerating
Tower
Dry Air
Outlet
Wet Air Inlet (7) Adsorption Desiccant
^ Media
Regenerated Air
Outlet Valve
Silencer
Orifice Plate
Check Valve
heat-reactivated desiccant dryers discussed in the
previous section.
6.4.3 Ozone Contacting
The capability of the ozone contacting unit is critical to
the successful performance of the ozone disinfection
system. It is also important to the economical opera-
tion of the process. Two important ozone contact
basin design considerations are the capability to
achieve good disinfection and capability to achieve
good ozone transfer. In a well designed ozone contact
basin, good ozone transfer will exist if good disinfec-
tion occurs. However, the opposite may also occur if
the contactor is poorly designed; good ozone transfer
can exist when disinfection is less than satisfactory.
At some of the plants visited during development of
this design manual, poor disinfection was noted
despite good ozone transfer. Excessive short-circuit-
ing was considered the primary cause of the poor
disinfection performance. Design considerations for
both ozone transfer and ozone disinfection are
discussed in this section of the manual.
131
-------
6.4.3.1 Ozone Transfer Design Considerations
The bubble diffuser contactor is the most common
type of contacting system used. Other ozone con-
tactors, such as the aspirating turbine mixer con-
tactor, are usually coupled with a proprietary ozone
generator (Kerag), and ozone transfer efficiency (TE)
information can be obtained from the manufacturer.
The characteristics of the bubble diffuser ozone
contactor are presented herein.
The transfer of ozone into wastewater has been
evaluated by several researchers (43-45). All con-
clude that ozone transfer into wastewater can be
described by the two-film theory. In this theory the
mass transfer of ozone per unit time is a function of
the two-film exchange area, the exchange potential,
and a transfer coefficient. The exchange area for the
bubble diffuser contactor is the surface area of the
bubbles, which is discussed later. The exchange
potential is called the "driving force" and is depend-
ent upon the difference between saturation ozone
concentration minus residual ozone concentration.
Ozone transfer can be described by the two-film
theory, but in practice contactor basins have not been
designed utilizing this theory. The theoretical basis
has been avoided because the design coefficients
have not been well documented, not because the
theory is unsound. When the design coefficients are
well documented, contactor design indeed may be
established using the two-film transfer model. In this
section of the manual the current state-of-the-art of
ozone contactor design is used. The most important
factors affecting design are discussed, and the design
rationale is presented.
Ozone TE is primarily influenced by the physical
characteristics of the contactor and the quality of the
wastewater. At a given applied ozone dosage, a
wastewater of poor quality will have a high ozone
demand and the contactor will exhibit a high TEE. The
high TE is due to the disappearance of ozone in
oxidation reactions (i.e., ozone demand reactions). An
example of the effect of water quality on transfer
efficiency is illustrated in Figure 6-33 (17). The TE of
the same contactor was higher when treating sec-
ondary quality wastewater than when treating ter-
tiary quality effluent. The differences in TE were more
pronounced as applied ozone dosage increased.
The chemical quality of the wastewater also affects
ozone TE, especially pH and alkalinity. A high pH
and/or a low alkalinity will cause a lower ozone
residual (i.e., other factors being constant) because
the hydroxyl radicals will be maximized (refer to
Section 6.2.1.4). The lower residual will increase the
exchange potential, or driving force, and will increase
TE.
Figure 6-33. Ozone transfer efficiency decreases as applied
ozone dosage increases and as ozone demand
of the wastewater decreases (17).
100
95
" 90
85
°80
75
Example Relationship
Do Not Use as Design Values
468
Applied Ozone Dose, mg/l
10
Wastewater quality will affect ozone TE, and is
important to keep in mind when evaluating the
performance of existing contactors, pilot-scale con-
tactors, and newly installed contactors. Wastewater
quality is typically not used as a basis to modify the
physical characteristics of the contactor. A summary
of the important water quality considerations on
ozone TE design is listed below:
a. Ozone TE will decrease as applied ozone dosage
increases. A specified minimum design TE
should be coupled with a specified applied
ozone dosage.
b. Ozone TE will increase as wastewater quality
deteriorates (i.e., ozone demand increases). A
specified minimum design TE should be coupled
with a specified description of the wastewater
quality.
c. Ozone TE will increase as wastewater chemical
quality favors the presence of hydroxyl radicals
such as a high pH or low alkalinity. A compar-
ison of TE of existing full-scale and pilot-scale
results should consider differences in waste-
water chemical quality.
The physical characteristics of the ozone contactor
are the most important considerations for the design
engineer because the engineer controls this element
of the process. The most important physical char-
732
-------
acteristics for optimizing ozone-TE are depth of the
contactor and type and location of diffusers. The
contactor physical characteristics for optimum dis-
infection performance are discussed in Section
6.4.3.2.
Ozone transfer theory states that the mass transfer of
ozone is dependent upon the exchange area of the
gas/liquid film. The exchange area for the bubble
diffuser contactor is primarily a function of the
diameter of the bubble. The bubble diameter can be
controlled through proper selection of diffuser mate-
rial. BestTE results are obtained when the bubble size
is between 2 and 5 mm in diameter, and preferrably
between 2 and 3 mm in diameter (30). The primary
factors affecting bubble size include diffuser perme-
ability and air flow rate. A lower permeability and
lower air flow rate will decrease the size of the
bubble.
Another consideration to optimize ozone TE is to
optimize the ozone gas to wastewater liquid ratio
(Gi/l_i) and ozone gas to contactor volume ratio
(G/V). Stover et al. (17) observed a decrease in TE as
the GI/LI ratio increased, as depicted in Figure 6-34.
A minor decrease in TE occurred as the Gi/I_i ratio
increased from 0.2 to 0.5. The decrease in TE was
more pronounced as the Gi/l_i ratio approached 1.0.
Grasso (46) and Given and Smith (24) reported a
decrease in TE as the G/V ratio increased from 0.005
to O.O5. They reported that for best results the G/V
ratio should not exceed 0.03.
Figure 6-34. An increase in the gas to liquid ratio causes a
decrease in ozone transfer efficiency.
100
90
I 80
60
50
Applied Dosage =10 mg/L
0.2
0.5
1.0
Gas/Liquid Ratio
2.0
The depth of the diffusers in the contactor is probably
the most important design issue because it repre-
sents a cost consideration as well as a TE considera-
tion. Deep diffusers will increase TE because the
saturation ozone concentration and associated driv-
ing force will increase. However, a deep contactor will
also increase capital cost. A shallow contactor will
reduce capital cost, but will decrease TE. A high
elevation of the contactor will also decrease TE. Hegg
et al. (42) reported that a 3.6-m (12-ft) deep contactor
at an elevation of 2,290 m (7,500 ft) above sea level
achieved a TE of 50 percent while disinfecting a
tertiary effluent.
Most of the bubble diffuser ozone contactors at plants
visited during development of this manual had a
diffuser depth between 4.9 and 6.1 m (16 and 20 ft)
deep. These contactors achieved a TE between 80
and 95 percent when treating a high quality sec-
ondary effluent at an applied ozone dosage equal to or
less than 6 mg/l.
A summary of important design considerations for
the physical characteristics of the ozone contactor are
listed below:
a. The diffusers in the bubble diffuser ozone
contactor should be at Ieast4.9 m (16 ft) deep for
a plant located near sea level treating a high
quality secondary effluent at an applied ozone
dosage less than 6 mg/l. The contactor should
be deeper if the wastewater is of higher quality,
if the applied ozone dosage is higher, or if the
plant is located at a higher elevation.
b. Both rod- and disc-shaped porous stone dif-
fusers have been used in bubble diffuser ozone
contactors.
c. The air flow rate to the diffuser must be within
the range recommended by the manufacturer.
* For the 6.3-cm (2.5-in) diameter by 61-cm
(24-in) long rod-shaped diffuser the air flow
rate per diffuser should not exceed 1.9 l/s (4
cfm).
• For the approximately 260-cm2 (40-in2) disc
diffusers the air flow rate per diffuser should
not exceed 0.6 l/s (1.25 cfm).
• For the approximately 390-cm2 (60 in2) disc
diffusers the air flow rate per diffuser should
not exceed 0.8 l/s (1.8 cfm).
d. Each porous stone diffuser should be secured
with stainless steel holders and sealed with
ozone-resistant gaskets such as Hypalon, Viton,
Teflon or Silicon.
e. The diffusers should be able to achieve a
estimated bubble size of 2 to 3 mm in diameter.
133
-------
f. Porous stone diffusers that have the following
criteria have typically been used in bubble
diffuser ozone contactors.
• Permeability of the porous media typically
has ranged from 12 to 20 cfm/ftVin at 2
inches water column. Note: Permeability is
defined by a test of a porous stone plate that is
12 in by 12 in by 1 in thick, tested under 2 in
water column pressure.
• Porosity of the porous stone diffuser has been
between 35 and 45 percent.
g. Diffusers have typically been located from 15 to
30 cm (6 to 12 in) from the bottom of the
contactor.
h. All diffusers should be installed at the same
elevation in order to evenly distribute the gas
flow to each diffuser.
i. Diffusers have typically been installed in the
liquid's downflow stages of the contactor in
order to maximize ozone TE. The first stage of
the contactor typically has more diffusers in
order to satisfy the initial ozone demand of the
wastewater. For example, a three-stage ozone
contactor has contained 50 percent of the
diffusers in the first stage.
6.4.3.2 Disinfection Performance
The primary factors affecting disinfection efficiency
assuming ozone demand is not excessive (see Section
6.5.1.2 for discussion of zone demand), are: transfer
efficiency, short-circuiting, mixing, and contact time.
Transfer efficiency should be optimized as described
in the previous section, in order to allow effective use
of the ozone produced as a disinfectant and to
minimize costs.
Short-circuiting is also a most important considera-
tion. The effect of short-circuiting on disinfection
performance is illustrated in Table 6-9. A small
amount of short-circuiting significantly increases the
effluent coliform concentration because the short-
circuited wastewater contains a very high concentra-
tion of organisms relative to the concentration
desired in the wastewater effluent.
Table 6-9. Effect of Short-Circuiting on Disinfection F'er-
formance
Influent
Coliform
#/100 ml
100,000
100,000
100,000
100,000
Effluent
Coliform
Target
#/100 ml
200
200
200
200
Percent
Short-circuiting
(%)
0.00
0.10
1.00
2.00
Resulting
Effluent
Coliform
#/100 ml
200
300
1,200
2,200
Mixing is necessary to bring the residual oxidants into
contact with the microorganisms, but back-mixing
can increase the potential for short-circuiting. For
example, in a bubble diffuser ozone contact basin
back-mixing in each stage of the contactor is suffi-
cient to change the liquid flow characteristics from a
plug flow to a near complete mix pattern in that stage.
Therefore, to minimize the effect of short-circuiting,
multiple stages that are positively isolated from each
other should be provided. The optimum number of
stages has not been documented.
Wastewater detention time in the contact basin is
also a factor affecting disinfection performance with
a wide range of contact times reported to achieve
acceptable disinfection. Gan reported a 6-log reduc-
tion in coliform organisms for a high quality activated
carbon effluent with 2 minutes liquid contact time
and an applied ozone dosage of 12 mg/l. However,
only a 1 -log reduction was achieved with the same
conditions when secondary effluent was treated. Gan
concluded that to achieve better disinfection of poorer
quality wastewater additional contact time would be
necessary. He further recommended that multiple
stages be used to provide the additional contact time
(25).
Farooq et al. (16) reported a 3-log reduction in
Escherichia coli after a detention time of only 6
seconds when a residual ozone concentration was
present. This suggests that the disinfection action of
ozone is very rapid. However, Farooq further sug-
gested that the contact time be on the order of several
minutes rather than several seconds since the
microorganism must come into contact with the
ozone.
Bollyky and Siegel (47) reported that disinfection
action continued when the ozonized wastewater was
held for a period up to 10 minutes, with minimal
improvement after another 10 minutes of holding
time. Perrich et al. (48) reported that the disinfection
action was a function of the product of residual ozone
concentration times liquid contact time. Venosa, etal.
(20) also evaluated the effect of contact time on
disinfection performance, and reported that disinfec-
tion efficiency correlated with the product of off-gas
ozone concentration times liquid contact time. In all of
these studies contact time was reported to be an
important factor in achieving desired disinfection
performance.
Stover et al. (17) evaluated the application of ozone to
secondary and tertiary wastewater to achieve high
levels of disinfection. His conclusion was that high
levels of disinfection could be achieved at contact
basin liquid detention times ranging from 1 to 10
minutes, but extremely high applied ozone dosages
and a high level of residual oxidants (including high
734
-------
residual ozone concentration) were required. Stover
was not able to evaluate the effect of a longer contact
time on achieving equal disinfection at lower applied
ozone dosages. However, he suggested that a sub-
stantial improvement in disinfection may be realized
after initial ozone contacting is complete when the
residual oxidants concentration is relatively high.
Legeron (49) and Miller et al. (1) discuss contact time
relative to chemical reaction and mass transfer
limited ozone processes. Both propose that the
required contact basin liquid detention time for
chemical reaction rate limited processes is dependent
upon the rate of the chemical reaction. For example,
the contact time can be very short (0.5 to 1 minute) for
removal of free iron. However, for a mass transfer
limited process, such as disinfection, they suggest a
longer contact time. Legeron recommends a min-
imum contact time of 6 minutes. Miller reported that
the contact time used for disinfection in European
water treatment plants ranged from 6 to 10 minutes.
For purposes of this manual it is concluded that data
are not available to define an optimum liquid contact
time in the ozone contact basin, or to define an
optimum configuration to prevent short-circuiting.
However, general guidelines are considered appli-
cable to address disinfection efficiency from ozone
contact basins.
a. Multiple staged ozone contactors should be
provided to minimize the effect of short-circuit-
ing. A minimum of 3, and preferably more
stages should be provided. Each stage should be
positively isolated from the other to simulate
plug flow characteristics and minimize the
potential for short-circuiting.
b. The liquid contact time should be adjusted
based on the desired disinfection target.
• To achieve the former EPA standard of 200
fecal coliforms per 100 ml, the minimum
contact time should be 6 minutes and prefer-
ably at least 10 minutes at design flow rates
(See Section 6.5.1 for discussion of design
flow).
• To achieve more stringent standards the
contact time should be lengthened to obtain
the benefit of the disinfection potential of the
residual oxidants produced. Pilot testing
should be completed to determine the opti-
mum contact time.
6.4.3.3 Types of Ozone Contactors
Various types of ozone contact basins have been
proposed or used for wastewater disinfection includ-
ing: positive pressure injectors, packed columns,
spray towers, turbine mixers and bubble diffusers
(22,28,50). The spray tower contactor is generally not
used because ozone transfer efficiency is quite poor
(56).
The packed column reactor, positive pressure injector
and bubble diffuser contactor were evaluated by
Venosa et al. (22). Results indicated that all con-
tactors achieved equal disinfection performance at
equivalent levels of transferred ozone dosage; how-
ever, their ozone transfer efficiency was significantly
different. The bubble diffuser achieved better ozone
transfer, especially as the ozone dosage increased.
Nebel reported similar results (50).
The bubble diffuser ozone contactor is the most
common reactor for ozone disinfection. Many dif-
ferent types of bubble diffuser configurations are
available. A schematic of one type is shown in Figure
6-35. Several important design considerations that
maximize ozone transfer and disinfection perform-
ance are illustrated in this schematic. These con-
siderations are discussed below.
Figure 6-35. Schematicof a 3-stage, bubble diffuser ozone
contact basin.
Wastewater Fine Bubble
•T Influent Diffusers (Typ.
• —
A
Lh
Sump
(Typ.)
.Slidegate'
/Tun \ 1
(iyp-)
n 1
u
Train A
Train B
n
Mo'oo
o o o o
o o o o
0 O 0 O
n
B"l
—
Plan: Ozone Contact Basin
a
A
t
Ozone Gas
2
' [
i
Note
J
0
T!
'-Control Valve c
LFIowMeter_4
: Pressu
D iff ere
Stage 1
Fine
Bubble
3 iff users
f^f^f^ f^
re
ntial
f
/
J
i c
! [
•=-
\
tage 2
r
Wastewater
Off-Gas v, Effluent
Stage 3
s
-Jj |LJj |U(— —
5-
Section A-A
To Outfall-
Train B
Train A
s
Section B-B
135
-------
a. The contact basin should be as deep as practical,
preferably greater than 5 m (16 ft) at sea level
and deeper at a higher elevation, such as 6 m
(20ft) at 2,440 m (8,000ft).The maximum depth
maybe limited by the maximum pressure in the
ozone generator, which is usually 103 kPa (15
psig).
b. The bubbles formed by the porous stone dif-
fusers should range between 2 and 3 mm in
diameter. (See text for detail on permeability
and porosity.)
c. The contactor should have at least two inde-
pendent trains with isolated off-gas compart-
ments to allow for continuous operation during
inspection and cleaning.
d. The contactor should have features that simu-
late plug flow and reduce short-circuiting.
• A minimum of 3, and preferably more,
separate stages should be provided.
• Each stage should be positively separated
from the other stages. No chance for short-
circuiting should exist: for example, through
drain holes at the bottom of the walls
separating the stages.
• Each stage should be provided with a sep-
arate drain pit to aid in cleaning on a routine
basis (e.g., once or twice per year).
e. The contactor should have from 1.2 to 1.8 m (4
to 6 ft) of "head" space to allow for foaming.
f. Each set of diffusers should have a flow control
valve on the ozonized air piping and separate
flow measurement. More diffusers should be
located in the first stage to meet the higher
demand for ozone in that stage, and thus provide
capability to maintain a uniform residual oxi-
dants concentration throughout all stages ofthe
contact basin. The rest of the diffusers can be
equally spaced in the remaining stages.
g. The wastewater flow should be counter-current
to the ozonized air flow to maximize ozone
transfer efficiency.
h. The contact basins should be made of typical
construction grade concrete, with ozone resist-
ant (e.g., Hypalon) water stops.
i. The contact basins should be covered and sealed
as much as possible. Sealing with Sika-flex 1 -A
compound covered with coal tar epoxy or teflon
sheeting has been used in some cases (18,51).
However, basin sealing is difficult to maintain,
and periodic leaks through the ceiling of the
contact basin may occur (51). It is suggested that
the ozone contact basin be placed in a location
where the entire roof of the basin is exposed to
the open atmosphere. Also, the basin should
have the capability to operate under negative
pressure.
j. Stainless steel piping for ozonized gas flow must
be provided for positive pressure ozone systems.
• Tungsten Inert Gas (TIG) welding is recom-
mended.
• Schedule 10 or better and type 304L or 316L
stainless steel is recommended.
• Flange-to-flange fittings, rather than thread-
ed fittings, should be used in applications
where welded connections are not made.
k. Ozonized feed-gas and contact basin off-gas
sample lines should be stainless steel tubing.
Teflon tubing may be considered for short runs.
The turbine reactor contacting system was evaluated
by Stover et al. (17) and Venosa et al. (22). In both
studies the ozone transfer efficiency and disinfection
performance were shown to be comparable to that of
the bubble diff user contactor. A schematic diagram of
the aspirating turbine mixer ozone contactor is shown
in Figure 6-36. The turbine draws ozonized gas into
the unit, where it is mixed with the wastewater and
pumped outward through the impeller tips. The speed
of the impeller and the size of an orifice controls the
amount of water "pumped." The pumping rate
controls the amount of mixing within the contactor,
the amount of ozone dissolution or transfer, and the
amount of ozonized-gas that is received from the
ozone generator. At the same time these controls
affect the amount of energy consumed by the process.
The turbine mixer contactor is typically used in
conjunction with the nominal pressure ozone genera-
tion system and is capable of operating at shorter
wastewater detention times because of the intensive
mixing provided. The detailed design of size of the
turbine mixer, amount of water pumped and feed gas
flow rate should be determined in conjunction with
the manufacturer of the nominal pressure system. In
addition, the flexibility for controlling the energy
consumption to match varying operating conditions
should be incorporated into the design of the process.
For example, flexibility should be provided to adjust
the energy consumed by the turbine mixer at different
wastewater flow rates and different applied ozone
dosages.
6.4.4 Ozone Destruction Equipment and Unit
Sizing
Ozone destruction is used to remove excess ozone in
the contact basin off-gas prior to venting, or prior to
recycle or reuse of the off-gas. Safety is the major
136
-------
consideration. The maximum allowable ambient
ozone concentration for an 8-hr working day is
0.0002 g/m3(0.1 ppm by volume). (See Section 6.1).
This concentration is significantly less than the ozone
concentration in the off-gas, which is normally
greater than 1.0 g/m3 (500 ppm by volume).
The primary methods for treating excess ozone in the
off-gas are: thermal destruction, thermal/catalyst
destruction, and catalyst destruction (52-54). Acti-
vated carbon destruction has also been used, but the
reaction with activated carbon causes the formation
of powdery activated carbon which may be explosive
(53). The use of activated carbon for ozone destruc-
tion is not recommended.
Figure 6-36. Schematic of a turbine mixer ozone contactor.
Drive Motor
Contact
Chamber
Off-Gas
Ozonated
Water
6.4.4.1 Thermal Destruction
Thermal destruction is typically not used with oxygen
feed-gas systems because of the high oxygen con-
centration and potential for uncontrollable fires. The
thermal destruct method of reducing ozone in the
off-gas involves heating the off-gas to a high temper-
ature and maintaining this temperature for a period of
time. From 50 to 100 percent ozone destruction has
been reported at operating temperatures between
250 and 350°C (480 and 660°F) that are maintained
from 1 to 3 seconds (52-56). A temperature between
300 and 350°C (570 and 660°F) for 3 seconds is
required to achieve greater than 99 percent ozone
destruction.
Because of the high temperatures involved, heat
recovery units are typically provided on thermal ozone
destruct systems. The outlet gas temperatures for a
heat recovery thermal destruct unit range from 70 to
110°C (160 to 230°F), and typically between 90 and
100°C (195 and 210°F) (52,54-56).
A schematic of a thermal destructor with heat
recovery is shown in Figure 6-37. The ozone contact
basin off-gas passes through a pressure/vacuum
relief valve and demister prior to entering the heat
exchanger. The pressure/vacuum relief valve is
provided to protect the contact basin from structural
damage due to excessive pressure or vacuum build-
up within the basin. The demister (i.e., stainless steel
wire mesh) is provided to reduce foam accumulation
within the heat exchanger and heating elements.
Different types of heat exchangers can be used,
including cross-flow, shell-and-tube, or plate-type
(52). Afan is shown as an optional piece of equipment,
depending on whether or not the contact basin is to be
operated under a pressurized or vacuum condition.
The option to operate under a vacuum condition is
highly desirable.
The energy requirement of the thermal destruct
system and the size of the thermal destruct heating
element can be estimated by knowing the off-gas flow
rate, the off-gas temperature rise and the heat loss of
the equipment (i.e., energy efficiency of the unit).
Assuming no neat loss, the amount of energy
required to raise 1.0 m3 of gas 1.0°C is 0.37 Wh (i.e.,
assume a specific heat of 0.2454 kcal/kg/°C, a gas
density of 1.293 kg/m3, and a conversion factor of
861.29 kcal/kWh (52,53). The heat loss of a thermal
destructor is reported to be about 30 percent (53,55).
Combining these parameters, the specific energy
requirement (Wh/m3) for a thermal destruct unit was
determined for various operating conditions of tem-
perature rise, as shown in Figure 6-38. This specific
energy value can be coupled with the off-gas flow rate
to determine the energy consumption and power
requirement of the thermal destruct unit.
The following design considerations are summarized
for a thermal ozone destruction process:
a. The thermal destruct equipment and piping
should be well insulated to minimize heat loss.
b. A stand-by unit should be provided to allow for
continuous operation during repair and main-
tenance.
c. A pressure/vacuum relief valve should be
located on the contact basin to protect the
contact basin from structural damage due to
excessive pressure or vacuum.
d. A demister (i.e., stainless steel wire mesh)
should be located prior to the ozone destruct
unit to reduce foam accumulation on the heating
elements.
737
-------
Figure 6-37. Example diagram of a thermal destruct unit with a heat-exchanger.
Pressure/Vacuum
Relief
Ozone Contact
Basin
Heat ^
.
— ».
'
i
i
i
'
•
-*•
•
1
' Exchanger ;
I
-ti
1
-»
I
-»
i
-»
I
->
I
->
1
-»
I
«i
1
>, /
* (
i
^_ Heating
p Coils
^\
)
Fan
(Optional)
e. The operating temperature should be 300 to
350°C (570 to 660°F), and the contact time at
least 3 seconds.
f. A heat recovery system should be considered to
reduce operating cost.
g. Instrumentation for process monitoring and
control that must be provided are:
• Inlet and outlet gas temperature—Monitor
system performance.
• Inlet gas flow rate—Monitor system loading.
• Inlet ozone concentration—Monitor system
loading.
• Outlet ozone concentration using meter-
Monitor system performance.
Figure 6-38. Specific energy consumption versus off-gas temperature rise through the thermal destruct unit.
240
|200
|l60
1120
i 80
CO
40
0
I I
200 400
Off-Gas Temperature Rise (°F)
600
140
•£120
|100
I 80
ID
5 60
j^
O
o 40
(A
20
0
40 80 120 160 200 240 280
Off-Gas Temperature Rise (°C)
138
-------
6.4.4.2 Thermal/Catalyst and Catalyst
Destruction
The use of catalysts for ozone destruction is fairly
recent (52). Specific information about the type and
quantity of material in the catalyst is not available
since it is proprietary information. Several general
classifications of catalysts are known to destroy
ozone including: metal catalysts, metal oxides, hydrox-
ides and peroxides (53).
A schematic of a metal catalyst ozone destruct unit is
shown in Figure 6-39. The contact basin off-gas
passes through a demister prior to entering the unit. It
is very important that contaminants such as foam be
kept away from the catalyst. The size of the metal
catalyst should be obtained from the manufacturer,
due to the proprietary nature of the catalyst. A flow
rate per volume of catalyst of 20 scfm/ft has been
successfully used (18). Flow rates as high as 50
scfm/ft3 have been reported (52). A fan is shown as
an optional piece of equipment, depending on
whether or not the contact basin is to be operated
under a pressure or vacuum condition.
A metal oxide catalyst, such as aluminum oxide that
contains palladium, operates at a temperature rang-
ing from 50 to 70°C (120 to 160°F) (53). A dis-
advantage of the metal oxide catalyst is that this is
more sensitive to chemical reactions with nitrogen
oxides, chlorine and its derivatives, and sulfides,
which destroy the catalyst (54). The hydroxide and
peroxide catalysts have not been used on field-scale
plant equipment.
The following design considerations for the catalyst
ozone destruction process should be considered in
addition to the design considerations for the thermal
destruct unit discussed in the previous section.
a. The catalyst and heating elements should be
easily reached for maintenance.
b. The gas pressure differential across the catalyst
should be monitored.
c. The size of the metal catalyst must be obtained
from the manufacturer, given the proprietary
nature of the material. An accurate projection of
inlet and desired outlet ozone concentration
must be presented to the manufacturer.
Figure 6-39. Example diagram of a thermal/catalyst ozone destruct unit.
Pressure/Vacuum
Relief
SL.
Ozone Contact
Basin
Demister
Heating
Chamber
Catalyst
Chamber
Fan (Optional)
Metal catalysts are now being used most often,
because they are more active than metal oxide
catalysts (54). Metal catalysts, such as finely divided
platinum or palladium, can operate at temperatures
as low as 29°C (85°F) (54). The use of metal catalysts
is advantageous, primarily because of lower operating
cost. The required temperature rise is much lower
than for thermal destruction and the catalyst life
expectancy is about 5 years (53). In some instances
the catalyst is operated without any temperature
increase. However, moisture condensation on the
catalyst can blind the catalyst and render it ineffec-
tive; therefore, it is generally advisable to increase the
gas temperature to prevent moisture condensation
on the catalyst.
6.5 Ozone Disinfection Process Design
Considerations
Ozone disinfection process design involves sizing of
ozone generation equipment and basins to meet
disinfection objectives economically over the antic-
ipated range of operating conditions. Aspects of the
ozone system that affect disinfection performance
and cost-effectiveness include: process energy effi-
ciency, process flexibility, effluent disinfection cri-
teria, wastewater quality, wastewater flow variations,
plus others. Guidelines for ozone disinfection process
design are presented in this section of the manual.
139
-------
6.5.1 Ozone Production Requirements
Proper sizing of the ozone generation equipment is
important for meeting desired effluent criteria with-
out excessive capacity that results in high capital
costs. To properly establish the ozone production
capacity the wastewater flow rate and applied ozone
dosage must be properly selected.
6.5.1.1 Wastewater Flow Design Considerations
The design wastewater flow rate for the disinfection
process may be dictated by State or other regulatory
design criteria, such as peak hourly flow rate (10-
State Standards, 1973 ed.), or may be determined
based on an analysis of local conditions. An analysis
of local conditions may involve an evaluation of the
frequency of the flow exceeding a specified value that
may be determined by developing a probability curve
using existing flow data. The design flow rate can be
based on a selected probability of occurrence. Flow
equalization should be evaluated to optimize the size
of ozone generation equipment needed to achieve
disinfection.
At plants surveyed during development of the design
manual, the ozone disinfection system design flow
rate was typically two to three times the average daily
flow rate. Ozone generation capacity was provided to
achieve a specified applied ozone dosage at the peak
design flow rate. Stand-by ozone generation capacity
was not provided at peak conditions. It was antic-
ipated that ozone equipment maintenance that would
require removal of equipment from service could be
completed during expected low flow conditions.
Where peak flows are frequent and unpredictable,
back-up or stand-by equipment during peak flow
conditions should be provided. As such, stand-by
equipment must be evaluated on a case-by-case
basis. In high growth iareas or for a long-term plant
design life (i.e., 20 years), the most economical
alternative may be to provide the space for additional
ozone generation equipment in the original design
and purchase the equipment when needed.
6.5.1.2 Ozone Dosage Design Considerations
Both applied (D) and transferred (T) ozone dosage are
important in ozone process design. Transferred ozone .
dosage is typically used for establishing the relation-
ship between ozone dosage and disinfection per-
formance. Once T and TE are defined, D can be
established. A determination of applied ozone dosage
is required to determine ozone production capacity. In
this section of the manual an approach for determina-
tion of T and D is presented. Refer to Section 6.4.3.1.
for a discussion of ozone transfer efficiency.
Determination of Transferred Ozone Dosage. The
transferred ozone dosage (T) required to achieve
disinfection is dependent upon the quality of the
wastewater (i.e., potential for chemical reaction with
ozone), the plant discharge criteria, and the disinfec-
tion performance capability of the ozone contact
basin. Because of the variables involved, selection of
transferred ozone dosage is probably the most
difficult process design consideration. The preferred
approach to establishing a design-transferred ozone
dosage is to conduct a pilot plant evaluation on the
treated wastewater to be disinfected. The type of
pilot-scale ozone generator used is not critical to
overall results; however, the type of pilot-scale ozone
contact basin must duplicate the proposed full-scale
basin for the results to be applicable to full-scale
design.
In practice, pilot testing has not been routinely
accomplished. Dosage requirements have often been
based on published pilot plant or existing full-scale
plant operating data. However, these data are site
specific and may not be directly applicable to other
installations. In this section both reported data from
existing plants and a rational approach for determina-
tion of transferred ozone dosage are discussed.
The transferred ozone dosage (T) requirement to
achieve various levels of disinfection performance
were evaluated by several investigators and were
discussed in detail in Section 6.3.2. Typically, trans-
ferred ozone dosages between 4 and 10 mg/l met the
former EPA fecal coliform standard of 200 per 100 ml
(27) when the total COD concentration of the treated
wastewater was less than 40 mg/l. Transferred
ozone dosages greater than 10 mg/l were projected
when the wastewater had a large industrial contribu-
tion and a COD concentration greater than 70 mg/l.
To meet a stringent standard of 2.2 total coliforms per
100 ml, a transferred ozone dosage between 36 and
42 mg/l was required when secondary treatment
plant effluent was disinfected (17). A transferred
ozone dosage between 15 and 20 mg/l was required
when nitrified wastewater was disinfected.
The design transferred ozone dosages were not
available for existing plants using ozone disinfection,
but the design applied ozone dosage for seven plants
ranged from 3 to 14 mg/l, as shown in Table 6-10 (4).
Most of these plants were required to achieve a
concentration of 200 fecal coliforms per 100 ml.
Operating data were not reported at these plants;
however, applied ozone dosage operating data ob-
tained from site visits conducted during the develop-
ment of this design manual and are shown in Table
6-11. The reported data for three plants appear to be
in line with the design criteria for the other plants.
However, at two plants visited the disinfection
standard of 200 fecal coliforms per 100 ml could not
be met even though the applied ozone dosage was
significantly higher(i.e.,greaterthan 10mg/l)thanat
the other plants (i.e., 3 to 6 mg/l) and the treated
wastewater was of similar quality. The poor per-
740
-------
formance at these plants was believed to be due to
poor contact basin disinfection capability, especially
excessive short-circuiting.
22 mg/l. The slope of the line was 2.51, intercept,
0.76; and correlation coefficient, 0.76. Other inves-
tigators have reported a similar relationship between
coliform reduction and transferred ozone dosage,
Table 6-10. Reported Design Applied Ozone Dosages for Various Wastewater Treatment Plants (4)
Name of Plant
Rocky River Regional
Upper Thompson
Sanitation District
Frankfort
Southwest
Brookings
Murphreesboro
Madisonville
Location
Concord, NC
Estes Park, CO
Frankfort, KY
Springfield, MO
Brookings, SD
Murphreesboro, TN
Madisonville, KY
Design
Applied
Dosage
mg/l
14
6
3
4
3
6
6
Permit Limitations
BOD/TSS/NHg
mg/l
20/30/13
30/30/20
10/10/1
N/R
20/30/2
N/R
10/30/1
Fecal
Coliform
#100/ml
N/R
200
200
N/R
1,500
N/R
200
N/R = Not Reported
Table 6-11. Reported Operating Applied Ozone Dosages for Various Wastewater Treatment Plants
Name of Plant
Location
Operating
Applied
Dosage
mg/l
Effluent Quality
BOD/TSS/NH3
mg/l
Fecal
Coliform
#7100 ml
Southport
Southwest
Vail
Indianapolis, IN
Springfield, MO
Vail, CO
5.0
5.4
2.5
5/5/NR
5/5/NR
5/5/NR
17
10
1,000
Note: At two other plants visited an applied ozone dosage greater than 10 mg/l reportedly does not allow consistent achievement of
the 200 fecal coliform per 100 ml limit, despite the fact that the wastewater quality was similar.
Data obtained during site visits to these plants.
N/R = Not Reported.
The transferred ozone dosage that is required to
achieve a desired concentration of coliform organ-
isms in the effluent is dependent upon the disinfec-
tion performance capability of the ozone contact
basin, the demand for ozone in reactions not asso-
ciated with disinfection, the influent coliform con-
centration, and the discharge coliform requirement.
A change in any of these parameters can cause a
significant change in the discharge coliform con-
centration. The approach to design presented in the
remainder of this chapter of the manual allows foran
independent evaluation of the effect of each param-
eter on transferred ozone dosage requirement. Data
to support recommended design criteria are pre-
sented. These criteria should be modified as data
become available to justify an adjustment.
The rational approach to design uses the relationship
- between coliform removal and transferred ozone
dosage reported by several investigators. An example
relationship was presented in Section 6.3.2, Figure
6-13 (17). In Figure 6-13 total coliform removal (log
(No/N)) increased as the transferred ozone dosage
(log T) increased. A linear-log relationship was
indicated for the approximately 100 data points over a
range of transferred ozone dosage from 1.5 mg/l to
although the slope and intercept of the individual
lines are quite variable (18,22,24,27). The regression
line of best fit of Figure 6-13 was rearranged, as
shown in Figure 6-40, in order to depict the dose/
response curve in a form that can be more readily
used for design. The equation of the line in this form
becomes:
Log (N/IM0) = n * LogfT/q)
(6-4)
where:
T = transferred ozone dosage (mg/l)
N = effluent coliform concentration (#/100 ml)
No = influent coliform concentration (#/100 ml)
n = slope of dose/response curve
q = X-axis intercept of dose/response curve,
which is the amount of ozone transferred
before measurable kill is observed.
By mathmatical rearrangement the slope (n) of the
dose/response curve for Stover's results (see Figure
6-13) was calculated as -2.51, and the X-axis
intercept (q) was 0.50 mg/l. These results are
presented in Table 6-12, along with results obtained
by other investigators. A wide range of slope and
intercept data are indicated, but individual results
747
-------
Figure 6-40. Dose response curve for nitrified effluent at
Marlborough (17).
-3
z
I
1-2
'I
w
-4
-5
| q = 0.50
I n = -2.51
>100 Data Points
r = 0.76
Log (N/No) = n Log (T/q)
90
99 I
1
DC
99.9 |
100
25
25
25
25
25
13
30
331
329
16
>100
>140
12
Effluent
COD
mg/l
29
30
26
39
39
38
74
13*
95*
92*
95*
— —
21
40
25
Reference
27
22
27
27
27
27
27
24
24
24
24
24
17
17
18
•Indicates BOD concentration
It should be noted that the X-axis intercept (i.e.,
transferred ozone dosage at 100 percent coliform
survival) of the dose/response curve is calculated
from the data; it is not a measured value. It is
improbable that a straight-line relationship occurs
near the X-axis intercept because some degree of
coliform reduction would be expected to occur
immediately as ozone is transferred to the waste-
to 40 mg/l) between 1.0 and 2.0 mg/l; and with a
high COD (74 mg/l) about 5 mg/l. These data maybe
used to estimate an X-axis intercept for ozone process
design, but conservative estimates may be appro-
priate considering the limited data base that is
available. It is recommended that pilot or bench-scale
testing be completed to better define the X-axis
intercept (i.e., initial ozone demand).
742
-------
The selection of the X-axis intercept will affect the
transferred ozone dosage requirement, as shown in
Figure 6-41. For a high quality wastewater with an
initial ozone demand of 0.5 mg/l a projected trans-
ferred ozone dosage of 5 mg/l would be required to
achieve a 3-log reduction in coliform organisms when
the slope of the dose/response curve is -3.0. For a
wastewater with an initial ozone demand four times
greater (2.0 mg/l), the projected transferred ozone
dosage is four times greater (20 mg/l) to meet the
same level of disinfection. Ozone disinfection effec-
tiveness is highly dependent upon the initial demand
for ozone. Wastewaters with a potential high initial
ozone demand may not be good candidates for ozone
disinfection systems.
Figure 6-41. Example curve showing the effect of different
X-axis intercepts on transferred ozone dosage
requirement.
0.1 0.5 1 5 10
Transferred Ozone Dose (T), mg/L
50 100
The slope of the dose/response curve represents the
change in coliform survival per mg/l transferred
ozone dosage. The effect of slope on transferred
ozone dosage required to achieve a 3-log reduction in
coliform organisms is shown in Figure 6-42, as-
suming the X-axis intercept isO.5 mg/l. At a relatively
steep slope of -5.0, only 2 mg/l transferred ozone
dosage is required. At a flatter slope of -3.0, a
projected transferred dosage of 5 mg/l is required
and at a slope of -2.0 a dosage of 16 mg/l would be
necessary.
The slope of the dose/response curve will become
flatter when the disinfection performance capability
of the ozone contact basin is poorer, or when on-
going chemical reactions with ozone reduce the
effectiveness of the disinfectant. For wastewaters
that have similar water quality, characteristics and a
similar initial ozone demand, it is anticipated that the
long-term ozone reactions would be similar. There-
fore, for these conditions and in the rational design
approach, the different slopes are considered to be
primarily a function of the disinfection performance
capability of the contactor.
Figure 6-42. Example curve showing the effect of different
slopes on transferred ozone dosage require-
ment.
O>
3
I'2
w
o>
3
£
-4
O
q = 0.50
Log (N/No) = n Log (T/q)
0.1 0.5 1 510 50 100
Transferred Ozone Dose (T), mg/L
The slope of the dose/response curves reported in
Table 6-12 vary considerably, ranging from -2.51 to
-6.65. A wide variation also exists when the waste-
water quality is similar (i.e., initial ozone demand less
than 1.1 mg/l and COD concentration less than 40
mg/l), ranging from -2.51 to -5.5. However, in all
cases the pilot-scale contact basin disinfection per-
formance capability was much better than the
capability of the field-scale units, as evidenced by the
consistently steeper slopes of the pilot-scale units.
Better plug flow capability in the small scale reactors
and corresponding reduced potential for short-circuit-
ing are believed to account for the improved disinfec-
tion performance of the pilot-scale systems. The
potential superior performance of pilot-scale contact
basins should be considered when scaling-up pilot
plant results to full-scale design.
An overall review of the dose/response data pre-
sented indicates general design criteria that may be
used for a rational approach to the determination of
the design transferred ozone dosage. For a good
quality secondary treatment plant effluent (COD less
than 40 mg/l), an initial ozone demand of 1.0 mg/l
appears reasonable. If a poorer quality wastewater is
anticipated, a higher initial ozone demand should be
selected. Conversely, a lower initial ozone demand
can be selected if a high quality wastewater is to be
disinfected. Pilot plant results may be used to obtain a
reasonably good estimation of the initial ozone
demand.
743
-------
The slope of the dose/response curve Is more difficult
to establish. Pilot plant results were generally better
than full-scale performance capability; however, the
pilot plants had three stages, while the full-scale
plants had only one and two. If the field-scale ozone
contact basins are designed to match the perform-
ance capability of the pilot-scale units (i.e., multiple
stages), then the steeper slopes, -4.0 to -5.0, may be
used in design. Disinfection efficiency design con-
siderations for the ozone contact basin are discussed
in Section 6.4.3.2. Otherwise, a flatter slope of -3.0
appears justified.
A summary of applicable guidelines for determining
the transferred ozone dosage requirement is pre-
sented below:
a. The approach for determination of transferred
ozone dosage may be used for ozone process
design.
• The initial ozone demand can be estimated
based on the quality of the wastewater
treated. For a good quality secondary treat-
ment plant effluent (COD less than 40 mg/l
and negligible nitrite nitrogen), an initial
ozone demand of 1.0 mg/l appears reason-
able.
• The slope of the dose/response curve can be
based on design features that enhance
contact basin disinfection capability. For a
contact basin v/ith good design features that
emulate reported pilot scale performance, a
slope of -4.0 to -5.0 may be used. Otherwise,
a flatter slope should be used.
• Influent coliform concentration should be
determined based on existing data, if avail-
able, or on reported concentrations for similar
plants.
• Effluent coliform concentration should be
based on the most stringent design limita-
tions.
b. To properly establish the transferred ozone
dosage requirement, pilot testing should be
conducted for all wastewaters and especially for
unique ozone disinfection applications such as:
• Disinfection of "strong" or highly industrial
wastewaters.
• Disinfection to achieve permit standards
more stringent than the former EPA standard
of 200 fecal coliforms per 100 ml.
• Disinfection using a type of ozone contact
basin that does not have a proven record of
performance.
c. Literature-reported ozone dosages may be used
for conventional applications of ozone disinfec-
tion.
e A transferred ozone dosage between 4 and
10 mg/l appears satisfactory to meet the
former EPA standard of 200 fecal coliforms
per 100 ml, when disinfecting a good quality
secondary or tertiary treatment plant effluent
in a properly designed ozone contact basin.
• A transferred ozone dosage between 15 and
20 mg/l reportedly meets the stringent
standard of 2.2 total coliforms per 100 ml,
when disinfecting good quality tertiary plant
effluent in a properly designed ozone contact
basin.
• A transferred ozone dosage between 36 and
42 mg/l reportedly meets the stringent
standard of 2.2 total coliforms per 100 ml,
when disinfecting highly polished secondary
treatment plant effluent in a properly de-
signed ozone contact basin.
Determination of Applied Ozone Dosage. The applied
ozone dosage is the mass of ozone from the generator
that is directed to a unit volume of the wastewater to
be disinfected. Design considerations for determining
transferred ozone dosage (T) were discussed in the
previous section, and for determining ozone transfer
efficiency (TE) in Section 6.4.3.1. The following
equation can be used to determine applied ozone
dosage.
D = T* 100/TE
(6-5)
By determining the applied ozone dosage and the
design wastewater flow rate, the design ozone
generation system production capacity can be estab-
lished. (See Section 6.2.3.2 for the procedure to
calculate ozone production).
6.5.2 Feed-Gas Supply and Process Flexibility
The design feed-gas flow rate to the ozone generator
is dependent upon the design ozone production rate
and design ozone concentration. Determination of
ozone production rate was discussed in Section 6.5.1.
Determination of ozone concentration was discussed
in Section 6.4.1.2. From these data the design feed-
gas flow rate can be determined using the equations
discussed in Section 6.4.1.2.
The peak design feed-gas flow rate is a fairly
straightforward calculation. However, most ozone
disinfection systems seldom operate at the peak
ozone production capacity. Requirements at start-up
are usually less than design conditions, and require-
ments at average design conditions are typically less
than at peak conditions. Process flexibility must be
installed to provide economical operation at variable
ozone production rates. Flexibility in ozone generation
equipment and feed-gas supply equipment are most
important. Also, flexibility must be provided to obtain
reliable, continuous operation. Design considerations
to address process flexibility are further discussed.
744
-------
Ozone production by most ozone generators can be
reduced to at least 20 percent of the peak production
rate, and often lower, by reducing the power supply to
the generator. Therefore, flexibility is available for the
ozone generator to reduce power consumption while
operating between 20 and 100 percent of peak
capacity.
Typically, at least two ozone generators of similar
capacity are provided to allow for continuous opera-
tion during routine maintenance, such as cleaning
the generator. Under these conditions the minimum
ozone production rate is 10 percent of the peak
production rate. If the projected minimum ozone
production requirement is less than the minimum
production rate, additional flexibility in ozone gen-
eration equipment must be provided. This flexibility
may be obtained with additional, smaller generators
or by providing capability to remove some of the
dielectrics from service (18).
Flexibility in the feed-gas supply and treatment
process is necessary for several reasons, but it is
especially important for energy conservation and
process reliability considerations. However, providing
flexibility in this auxiliary equipment is typically more
involved and more difficult to obtain than providing
flexibility in the ozone generator's air pre-treatment
process. Flexibility is more involved because it
includes design considerations for piping as well as
equipment size and type. It is more difficult to achieve
because of equipment capabilities, especially the
capability to vary the air flow rate from the air
compressors and at the same time reduce power
consumption.
When oxygen is the feed-gas, the oxygen requirement
of the ozone disinfection system must be coordinated
with the oxygen requirement of the biological treat-
ment process. Balancing these oxygen needs-was
discussed in Section 6.3.2. Generally, fine-tuning the
system to achieve a precise oxygen supply balance is
difficult in full-scale applications because of the
difficulty in rapidly adjusting the oxygen supply from
the oxygen production facilities. Before an oxygen
supply control system is designed, it is recommended
that existing oxygen-fed ozone disinfection plants be
visited to obtain information on problems encount-
ered with the oxygen supply control approaches that
have been attempted.
When air is the feed-gas, energy efficient operation
may be improved by providing flexibility in the air
supply equipment. The air compressor(s) are typically
the second largest consumer of energy in the ozone
disinfection system, and may even exceed the
consumption of the ozone generator in some in-
stances. The compressor(s) are usually sized to meet
the design ozone concentration at the ozone gener-
ator's peak production rate. At these peak design
conditions the total or overall system-specific energy
consumption is best, as shown in Figure 6-43. The
total system-specific energy consumption averaged
about 23 Wh/g (10.5 kWh/lb) at the typical design
ozone concentration of 18 g/m3 (1.5 percent wt).
However, when the ozone generator power level was
at its lowest setting and the feed-gas flow rate was
not adjusted (i.e., ozone concentration was 5 g/m3
(0.4 percent wt), the total system-specific energy
consumption was significantly higher at 33 Wh/g (15
kWh/lb).
In Figure 6-43 a range of specific energy consumption
data is shown. The variability is due to type of
equipment selected and operating conditions, such
as dew point, dielectric cleanliness, temperature, etc.
The range shown is typical of most air-fed ozone
generation systems and may be used to estimate
energy consumption for proposed ozone disinfection
processes. A summary of applicable design con-
siderations to achieve energy efficient operation and
process reliability is listed below.
Figure 6-43. Specific energy consumption for a typical air-
fed ozone generation system.
Ozone Concentration, Yi (% by Weight)
0.5 1.0 1.5
a.
b.
c.
0 2.5 5.0 7.5 10.0 12.5 15.0 17.5
Ozone Concentration, Yi (g/m3)
Specific energy consumption should be lowest
at expected operating conditions, rather than
only at peak ozone production rates.
Multiple air compressors and/or variable speed
compressors should be considered to reduce the
air flow rate and reduce energy consumption.
The ozone concentration should not be greater
than the design concentration in order to avoid
745
-------
excess heat build-up in the ozone generator and
damage to generator dielectrics.
d. Independent skid-mounted generation systems
should be avoided. If skid-mounted units are
provided, then the equipment should be inter-
connected with appropriate piping so that each
component can be operated independent of
other components on the skid.
6.5.3 Ozone Process Control and Automation
The amount of ozone system control and automation
has ranged from almost negligible capability to very
extensive provisions. The purpose of addressing
control and automation has been to reduce energy
consumption rather than to balance disinfection
capability with water quality effects, as is the case
with chlorine disinfection systems. Overdosing with
ozone does not create a water quality problem
because the ozone will simply "degrade" back to
oxygen. Factors to consider in evaluation of process
control and automation requirements are discussed
in this section.
The primary consideration in an evaluation of process
control and automation needs is the potential for pay-
back of the capital investment. The issues to consider
are the complexity of the system for the size of the
plant and the potential for savings as affected by
anticipated variations in ozone production (i.e., varia-
tions in wastewatsr flow and quality). Generally,
operating flexibility to minimize energy consumption
should be provided, as discussed in the previous
section. Operating flexibility not only provides energy
savings potential, but also improves process reli-
ability and enhances system maintenance. Process
automation, on the other hand, should be evaluated
on a case-by-case basis.
Automation of the auxiliary equipment, especially air
supply and treatment equipment, is typically more
complex than automation of the ozone generator.
Unless the plant is quite large or variation in ozone
production extensive, it is generally cost effective to
control the auxiliary equipment manually. Manual
adjustments might include adding or deleting an air
compressor or increasing or decreasing the compres-
sor speed at certain times of the day, week, season, or
year, in order to operate in an acceptable range of
ozone concentration from the ozone generator. If
automated control is provided, the number or speed of
the air compressor(s) and number of ozone generators
on-line is directly controlled by the ozone concentra-
tion.
Typically, the ozone production requirement to
achieve disinfection is going to vary on a routine
basis. This production requirement can be adjusted
by changing the power setting of the ozone gener-
ator(s). Three approaches to analyzing and controlling
the generator power settings have been used,
including applied ozone dosage control, residual
ozone concentration control, and off-gas ozone
concentration control. The approach, advantages,
and disadvantages of each method are further
discussed.
6.5.3.1 Applied Ozone Dosage Control
The simplest method of ozone system process control
is to make adjustments to the ozone production rate
so that the applied ozone dosage is maintained at a
constant value. Once the applied ozone dosage has
been selected the only other variable involved is the
wastewater flow rate. The adjustments to ozone
production to maintain a specified applied ozone
dosage may be completed by manual means, or can
be completed automatically. The primary advantage
of this method of process control is the simplicity of
the system. The primary disadvantage is that the
control method is completely unresponsive to changes
in water quality.
Manual control using the applied ozone dosage
method requires that the operator know the waste-
water flow rate being treated and the ozone produc-
tion rate of the ozone generator. The operator then
adjusts the generator power setting to maintain the
desired applied dosage as the wastewater flow rate
varies.
The applied ozone dosage method may be fairly
simple or can be quite complex. A simple method of
automating the applied dosage control technique
involves varying the power supply to the ozone
generator in proportion to the wastewater flow rate.
This method is best used when the ozone production
rate is linear to the generator power setting. A more
common method is to use a microprocessor control
system that receives information for wastewater flow
rate, feed-gas flow rate and ozone concentration;
calculates the applied ozone dosage; and compares
this figure with the set-point value. The output signal
increases or decreases the ozone generator power
supply so that the calculated applied ozone dosage is
equal to set-point. This will cause some changes in
the ozone concentration since the feed-gas flow rate
remains the same. Large variations in ozone demand
will require adjustments to the air supply rate.
The microprocessor method of automated process
control is fairly reliable, if reliable instrumentation is
provided. The applied ozone dosage may be main-
tained at the desired set-point over a fairly wide range
of wastewater flow rates. The set-point may be
changed as disinfection performance indicates, al-
though the operator is usually not aware that a
change is needed until the bacteriological test results
are available. The method works best when the
wastewater quality and disinfection performance do
not vary considerably.
746
-------
6.5.3.2 Wastewater Ozone Residual Control
The ozone residual control method is effective
because disinfection performance is closely related to
ozone residual. However, the method is generally
unreliable because it is difficult to maintain calibra-
tion of the residual monitors as liquid characteristics
change (57). Also, the probes tend to foul-up in the
wastewater environment, although improvements in
probes are continually being made. Attempts to use
the residual ozone control method must be coupled
with a commitment to keep the monitors calibrated.
According to Grunwell and Gordon, the Indigo method
or the arsenic (III) method of chemical analyses may
be used as a laboratory check of in-line ozone residual
meters (8,57).
The ozone residual control method is fairly straight-
forward from an instrumentation standpoint. The
signal from the residual monitor is sent to a micropro-
cessor, which compares the signal received to the
value set by the operator. The output signal is sent to
the ozone generator power supply which changes
ozone production until the residual ozone concentra-
tion is equal to the set-point value. A disadvantage of
this control scheme is the time delay between the
residual monitor signal to the microprocessor due to
the wastewater contact time. Also, improperly de-
signed systems could produce excessive "hunting" of
the equipment.
6.5.3.3 Off-Gas Ozone Concentration Control
The off-gas ozone concentration control method was
recently presented as a possible control approach for
wastewater disinfection (15,19,58). It was proposed
based upon an evaluation of pilot-scale disinfection
performance data. A trend of improved disinfection
occurred as the off-gas ozone concentration in-
creased. Initially, the method was proposed as
applicable when the ozone feed-gas to wastewater
liquid flow ratio(Gi/l_i) was constant. However, more
recent testing has indicated that strict control over
the Gi/Li ratio is not required, if the Gi/L-i ratio is
kept between 0.1 and 0.5 (15,18,58).
The off-gas control method has been used at the Vail,
Colorado wastewater treatment plant (18). However,
the system was not automated. The plant operators
simply check the monitor reading on a routine basis,
and make adjustments to the ozone generator power
setting to maintain a pre-established range of off-gas
ozone concentration.
Currently, an automated off-gas control system has
not been installed or tested. However, the method
shows promise of being an effective process control
device. The biggest advantage is that instrumentation
seems to be quite reliable, while at the same time
responsive to water quality and wastewater flow rate
changes.
6.5.4 Ozone Disinfect/on Design Example
The example presented in this section of the manual
is intended to illustrate the design considerations
involved with development of an ozone disinfection
system. The wastewater treatment design informa-
tion for the example design problem is shown in Table
6-13. The plant is an air activated sludge plant located
at an elevation of 3,500 ft (1,067 m) above sea level.
The influent to the ozone disinfection system is the
effluent from the secondary clarifiers.
Step 1—Determine the Transferred Ozone Dosage,
Applied Ozone Dosage, and Ozone Production Design
Values.
Method A. Based on a literature search of existing
facilities (See Section 6.5.2.2) the design applied
ozone dosage ranged from 4 to 10 mg/l at plants
required to meet the 200/400 fecal coliforms per 100
ml standard. The design transfer efficiency ranged
from 85 to 90 percent. Compare these data with the
rational approach to design presented in Method B.
Method B. Refer to Section 6.5.1.2 and develop the
design ozone dosage based on a rational approach to
design. In developing the rational design values, two
important assumptions are required including the
initial ozone demand and the slope of the dose/
Table 6-13. Ozone Disinfection System Criteria for Design Example Problem
Average Daily Wastewater Flow 7.5 mgd
Peak Daily Wastewater Flow '. .15.0 mgd
NOTE: The daily peak flow rate will not exceed 15.0 mgd because of storm flow equalization facilities.
Start-up Daily Average Wastewater Flow 3.5 mgd
Start-up Peak Daily Wastewater Flow • 7.5 mgd
Average Effluent BOD/TSS 15/15 mg/l
Maximum Daily Effluent BOD/TSS 30/30 mg/l
Design Required Effluent Fecal Coliform
Weekly Maximum Limitation 400 per 100 ml
Geometric Mean Monthly Limitation 200 per 100 ml
Disinfection System Influent Fecal Coliform
Geometric Mean Concentration 500,000/100 ml
Maximum Concentration 200,000/100 ml
Chlorine Residual =£0.05 mg/l
747
-------
response curve. According to the discussion in
Section 6.5.1.2, the initial ozone demand at most
secondary treatment plants was about 1.0 mg/l.
Therefore, in this design example the initial ozone
demand is assumed to be 1.0 mg/l.
l
The discussion in Section 6.5.1.2 indicated that the
slope of the dose/response curve was more difficult
to establish. Evaluations using pilot-scale contactors
reported a slope between -4.0 to -5.0, while reported
field-scale data indicated a flatter slope of -3.0. The
transferred dosage for a slope of -3.0, initial ozone
demand of 1.0 mg/l, and coliform log survival of -4.0
is calculated below using Equation 6-6, which is a
rearrangement of equation 6-4.
T = q * 1 o[LoB(N/N-Wn
(6-6)
where
q = 1.0 mg/l
n = -3.0
N = 200 per 100 ml
No = 2,000,000 per 100 ml
then
T = 21.5 mg/l
The transferred ozone dosage for a slope of -4.0 and
-5.0 was also calculated for an initial ozone demand
of 1.0 mg/l and log survival of-4.0, and was compared
with the dosage determined for the slope of -3.0.
Slope of Dose/
Response Curve
-3.0
-4.0
-5.0
Transferred Dosage
21.5 mg/l
10.0 mg/l
6.3 mg/l
As shown, the slope of the dose/response curve has
a significant impact on the resulting transferred
ozone dosage. The flatter slope reported for field-
scale evaluations may be due to "scale-up" factors, or
may be due to the shape of the contactors. The two
field-scale contactors had only one and two stages of
contacting, while the pilot-scale contactors had three
stages of contacting. Contactor staging has a signif-
icant impact on disinfection performance, as dis-
cussed in Section 6.4.3.2.
It should be noted that two alternative design
considerations exist relative to obtaining desired
disinfection performance at reasonable ozone dos-
ages. One alternative is to provide a tertiary filter,
which reportedly will reduce the influent coliform
concentration by a factor of 10 (i.e., 1 log reduction)
(17). If filters were installed and the maximum
influent coliform concentration were reduced from
2,000,000 to 200,000 per 100 ml, then the log
survival is -3.0 instead of -4.0. The resulting trans-
ferred ozone dosage for a dose/response curve slope
of -3.0 is 10.0 mg/l, which is less than half of the
dosage required at a log survival of -4.0. Further, the
10 mg/l dosage is equal to the dosage required when
the log survival was -4.0 and the slope was -4.0.
Therefore, the effect of reducing the influent coliform
concentration by one log is similar to the effect of
improving the slope of the dose/response curve from
-3.0 to -4.0.
The second alternative to obtaining desired disinfec-
tion performance at reasonable ozone dosages is to
provide contact basin design features that improve
disinfection performance and achieve a steeper
dose/response slope of -4.0 to -5.0 instead of -3.0.
For this design example it is assumed that a tertiary
filter is not an available option. Therefore, a coliform
log survival of -4.0 will be used to establish the
required transferred ozone dosage.
Currently, data a're not available to document the
optimum number of stages for an ozone contact
basin. Pilot-scale data indicate that three stages
consistently achieved a log survival between -4.0 and
-5.0. However, the stages must be designed to
simulate plug flow conditions (i.e., eliminate short-
circuiting), which apparently was the case with the
pilot-scale contactors. For this design example it is
proposed that a 4-stage contactor will be used and
that the slope of the dose/response curve will be -4.0.
Also, it is assumed that the contactor diffuser depth
will be 5.5 m (18 ft), the ozone transfer efficiency will
be a minimum of 85% at the plant elevation of 1,070
m (3,500 ft), and the detention time will be 15
minutes at the average flow rate. It should be noted
that the conditions for applied ozone dosage obtained
from these criteria range from 3 to 11.6 mg/l, and
therefore are consistent with the literature reported
dosages of 4 to 10 mg/l presented in Method A above.
Based on the discussion above the following design
basis is used for the example design problem.
Initial Ozone Demand
Slope of Dose/Response Curve
Ozone Transfer Efficiency
Type of Contactor
Number of Stages
Number of Basins
Contact Basin Detention Time at
Average Flow
Diffuser Depth
1.0 mg/l.
-4.0
. 85%
Bubble Diffuser
4
2
15 min
18 ft (5.5 m)
Step 2—Determine the Design Ozone Production
Rate.
Calculate the transferred ozone dosage, applied
ozone dosage, and ozone production rate for the
148
-------
various operating conditions that may occur, as
shown in Table 6-14.
Table 6-1 4, Transferred Ozone Dosage Calculations for Design Example
Effluent Influent
Fecal Fecal Log Trans
Trial Wastewater Coliform Coliform Coliform Dose
Number Flow (mgd) (#/100 mL) (#/100 mL) Survival (mg/L)
1 15 200 10,000 -1.70 2.66
2 15 200 500,000 -3.40 7.07
3 15 200 2,000,000 -4.00 10.00
4 7.5 200 10,000 -1.70 2.66
5 7.5 200 500,000 -3.40 7.07
6 7.5 200 2,000,000 -4.00 10.00
7 3.5 200 10,000 -1.70 2.66
8 3.5 200 500,000 -3.40 7.07
9 3.5 200 2,000,000 -4.00 10.00
Applied
Dose
(mg/L)
3.13
8.32
11.76
3.13
8.32
11.76
3.13
8.32
11.76
Applied
Mass
(Ib/d)
391
1,041
1,472
196
520
736
91
243
343
Shown below is an example calculation for Condition Figure 6-44. Design example projected ozone production
1, and shown in Figure 6-44 is a graph of the rate for various operating conditions.
production results for all conditions.
T n * 1 otL°s(N/N-)1/n ^ c
I q lu 15"
1.4-
where ,
— •»» •«"
q = 1.0 mg/l 1 1.2-
n = -4.0 ' £ 1.1-
N = 200 per 100ml ^10
N0 = 10,000 per 100 ml f ' '
(D "jgy 0.9'
then §• g 0.8-
T = 2,66mg/l ||°£
D = TMOO/TE | 0,5-
£ 0,4-
where <= o 3-
TE = 85% 0 o.2-
0.1-
tnen 0.0.
I
77
ti.
h
i
y
\
T77
//,
'//,
//,
//,
'/
tt
'//.
i
\
i
/
f
I
i
I
\
h m/,
D = 313 mg/l 1 2 3 4 5 6 7 8
Condition Number
///
//
-600
-500
~4QQ_
- . 1"
"300g
-200
-100
9
P = D * L * 8.34
where
L = 15 mgd
then
P = 391 lb/d(123g/min)
From Figure 6-44 and Table 6-14, the ozone produc-
tion rate that would satisfy all assumed design
conditions is 463 g/min (1,472 Ib/d). Therefore, for
this design example the design ozone production rate
is assumed to be 473 g/min (1,500 Ib/d). ,
Step 3—Select the Number of Ozone Generators
The considerations for selection of the number of
ozone generators is to address both the maximum
and the minimum expected ozone production rates. If
three ozone generators each with a capacity of 158
g/min (500 Ib/d) were provided, the maximum
production rate would be met. However, if the
maximum production rate were expected to occur on
a frequent basis, then a fourth generator or additional
capacity of the three generators should be provided so
that the generators are not required to operate at peak
production rates for extended periods of time (See
Section 6.4.1.2).
749
-------
The minimum production rate from an ozone gener-
ator is 10 to 20 percent of its maximum production
capability. From Figure 6-44 and Table 6-14 the
minimum expected O2:one production rate at start-up
conditions is 29 g/min (91 Ib/d). If the capacity of
each generator were 158 g/min (500 Ib/d), then the
minimum ozone production rate would be 16 to 32
g/min (50 to 100 Ib/d) and would match the projected
requirements at start-up conditions.
For this design example three ozone generators each
with a capacity of 158 g/min (500 Ib/d) are proposed.
Also, room for a fourth generator will be provided in
case the design peak ozone production rates are
required for extensive periods of time. Since air
activated sludge is used for secondary treatment (i.e.,
oxygen activated sludge is not used; thus a readily
available source of oxygen is not present), an air-fed
ozone system is proposed.
Step 4—Determine the Size and Number of Air
Compressors
Proposed Design Basis:
1. The maximum ozone concentration is 18
g/m3{1.5 percent wt).
2. A low pressure iair-fed system is used.
3. The desiccantdryer purge airflow is 20 percent.
Calculate the generator air flow rate and total system
airflow rate at both peak design and average start-up
conditions.
At Peak Design Conditions
Generator air flow rate
= 1,500 lb/d/(0.015 Ib O3/lb air)/(0.0753 Ib
air/ft3)/1,440i
-------
Operating absolute pressure is:
Pressure at altitude of 1,070 m (3,500 ft) is 12.93
psi
Operating gauge pressure is 13.5 psi
Operating absolute pressure is 13.5 + 12.93 =
26.43 psia (89 kPa)
Calculate moisture content at operating pressure
= (0.394 lb/1,000 ft3) * (14.7 psia/26.43 psia)
= 0.219lb/1,000ft3
Calculate moisture loading
= (0.219 lb/1,000 ft3) * (1,200 ftVmin * 60 min/hr)
= 15.8 Ib/hr (7.2 kg/hr)
Calculate amount of desiccant for all operating
towers
Desiccant/tower = 18 Ib/lb * 15.8 Ib/hr * 16 hr
= 4,550 Ib/towers that are drying
(2,065 kg/towers)
Consider using three desiccant dryers, each with two
towers. Determine the desiccant amount per tower.
NOTE: The amount for the dryer is twice the amount
per tower, since there are two towers per dryer.
Desiccant/tower/dryer = 4,550 lb/towers/3
dryers
= 1,517 Ib/tower/dryer
(690 kg/tower/dryer)
The analysis of the amount of desiccant required for
the case with and without a refrigerant dryer indicates
that significantly less desiccant is required (i.e.,
desiccant dryers would be smaller) when a refrigerant
dryer is provided. An option exists as to whether or
not to provide the refrigerant dryer in conjunction
with the desiccant dryer, or to eliminate the re-
frigerant dryer from the process because of its
sensitive operation (See Section 6.4.2.4). In some
instances, the desiccant dryer has been sized to
operate without the refrigerant dryer, yet a refrigerant
dryer was installed to provide flexibility for a case
when problems may occur with the desiccant dryer
(18). If one desiccant dryer were out of service, and
normally three would be required, the refrigerant
dryer could be operated to reduce the moisture
loading to the other two operating desiccant dryers. In
this regard, the refrigerant dryer could be considered
a back-up unit for the desiccant dryers. A most
important consideration is to absolutely never mini-
mize the importance and impact of the desiccant
dryer. The dryer should never be undersized.
The design calculations shown above are a portion of
the overall design considerations of an ozone dis-
infection process. Other considerations include in-
strumentation, layout, flexibility, access for main-
tenance, type of materials, alarm systems, type of
equipment, control features, automation, etc. The
engineer should include all this design related
information in the contract documents so that all
equipment manufacturers can bid the ozone equip-
ment on an equal basis. The various sections within
this manual are intended to assist the engineer in
making these design decisions.
6.6 Safety
Ozone is a toxic gas, and like chlorine can cause
severe illness and death if jnhaled in sufficient
quantity. However, ozone systems have safety advan-
tages not available with the chlorine disinfection
process. Ozone is generated on-site, thus eliminating
transportation hazards. Also, the generation system
can be shut down if an ozone leak develops. Another
safety advantage is the physical characteristic of
ozone that allows it to be detected (smelled) at
concentrations much lower than harmful levels.
In addition to safety precautions against exposure to
ozone, protection against noise and electrical hazards
should be incorporated into the design and operation
of an ozone disinfection system.
6.6,1 Recommended Exposure Limit to Ozone
A study of the health effects of ozone exposure was
conducted by the United States Air Force (59). The
results were summarized in the graph shown in
Figure 6-45. Another summary of the health effects
of ozone was compiled by the American Society for
Testing and Materials (ASTM) in support of their
recommended standard for limiting human exposure
to ozone. The reported biological effects range from
dryness of mouth and throat, coughing, headache,
and chest restrictions at concentrations near the
recommended limit, to more acute problems at higher
concentrations.
The recommended ambient ozone exposure levels
have been proposed by the Occupational Safety and
Health Administration (OSHA), the American National
Standards Institute/American Society for Testing
and Materials (ANSI/ASTM), the American Confer-
ence of Government Industrial (ACGI), and the
American Industrial Hygiene Association (AIHA) as
follows (1):
Control occupational exposure such that
workers will not be exposed to ozone con-
centrations in excess of a time weighted
average of 0.2 mg/m3(0.1 ppmbyvolume)for
eight hours or more per workday, and that no
worker be exposed to a ceiling concentration
of-ozone in excess of 0.6 mg/m3(0.3 ppm by
volume) for more than 10 minutes.
157
-------
Figure 6-46. Human tolorance for ozone (59).
102
Exposure Time, minutes
103
10"
These recommended limits for ozone concentration
are much higher than the concentrations at which
ozone can typically be smelled. Generally, an indi-
vidual can detect ozone at concentrations ranging
from0.02to0.1 mg/nn3{0.01 to 0.05 ppm by volume)
(1). The more often a person is exposed to ozone the
higher the required concentration for detection.
6.6.2 Ambient Ozone Concentration Monitors
The subject of safety in the design and operation of an
ozone system should receive a high priority. All ozone
systems should be provided with an ambient ozone
monitor or monitors which are set up to measure the
ozone concentration at potential ozone-contaminated
locations within the plant (e.g., at various places in
the room housing the ozone generators). A single
monitor may be installed, and the air from different
locations pumped to the monitor for detection of
ozone concentration. The monitors should be set up
to sound an alarm v/hen the ozone concentration
reaches 0.2 mg/m3 (0.1 ppm by volume), and should
be set up to shut down the ozone system when the
concentration exceeds 0.6 mg/m3 (0.3 pprn by
volume). However, if the ozone equipment is located
in an area where "smoggy" days due to ozone levels
in the atmosphere are common, higher values may be
necessary to be able to detect ozone leakage by the
ozone generation system.
A listing of reliable ozone monitors used for ambient
monitoring purposes may be obtained from the ozone
generation equipment manufacturers and from oper-
ating plants. These monitors may be checked and
calibrated using extensive calibration procedures
(60), but loss of calibration typically has not been a
problem. Under normal operating conditions the
monitor does not "alarm" and the operators cannot
smell ozone. Therefore, on a periodic basis the
operators should check the operation of the monitor
by directing a small volume of ozone from the
generator to the monitor to test the meter's respon-
siveness. In this manner the operators can be assured
that the monitor will respond in case of an ozone leak.
6.6.3 Miscellaneous Safety Considerations
Some ozone generation systems have been labeled
as "noisy" installations. However, the source of the
noise is usually thefeed-gas compressors and notthe
ozone generators. The ozone generators themselves
have a slight "hissing" sound that is typically not
objectionable. High frequency generators may give
off a high frequency pitch.
Ideally, the feed-gas compressors should be isolated
in a room that has some degree of sound proofing.
However, in smaller installations this alternative may
be economically unattractive. In these situations the
operators should wear ear protection equipment
when operating or maintaining the ozone equipment.
As with any toxic chemical, the operators should be
trained concerning the potential hazards involved and
the emergency operating procedures required if a
problem occurs. Equipment should be provided to
assist the operator. Applicable equipment is listed
below:
a. A self-contained breathing apparatus should be
provided, and should be located at a place where
access is not restricted by ozone in case an
ozone leak occurs.
b. An eye-washing sink should be provided to
enable the operator to rinse ozone from the
eyes, if needed.
c. Safety manuals on performing artificial respi-
ration should be provided.
d. Separate ladders should be provided to enable
the operator to enter the ozone contact chamber.
Fixed steps in the contact basin should not be
relied on.
752
-------
6.7 References
When an NTIS number is cited in a reference, that
reference is available from: .
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
(703) 487-4650
1. Miller, G.W., etal. An Assessment of Ozone and
Chlorine Dioxide Technologies for Treatment of
Municipal Water Supplies. EPA-600/2-78-147,
NTIS No. PB-285972, U.S. Environmental Pro-
tection Agency, Cincinnati, OH, 1978.
2. Rice, R.G., et al. Ozone for Drinking Water
Treatment —Current State-of-the-Art. Proceed-
ings of the Seminar on The Design and Opera-
tion of Drinking Water Facilities Using Ozone or
Chlorine Dioxide, New England Water Works
Association, Volume 1, June 1979.
3. Novak, F. Two Years of Ozone Disinfection of
Wastewater at Indiantown, Florida. Presented
at the IOI-EPA Seminar on the Current Status of
Wastewater Treatment and Disinfection With
Ozone, September 1977.
4. Weston, R.F., Inc. Factors Affecting the Opera-
tion and Maintenance of Selected Ozone and
Ultraviolet Disinfection Systems. Draft for
MERL, U.S. EPA Contract No. 68-83-3019,
February 1983.
5. Rice, R.G. Personal Communication, 1985.
6. Klein, M.J., et al. Generation of Ozone. First
International Symposium on Ozone for Water
and Wastewater Treatment, International
Ozone Institute, 1975.
7. Manley, T.C. and S.J. Niegowski. Kirk-Othmer.
Encyclopedia of Chemical Technology. John
Wiley & Sons. Second Edition, Volume 14,410-
432, 1967.
8. Grunwell, J. et al. A Detailed Comparison of
Analytical Methods for Residual Ozone Meas-
urement. IOA OZONE Science & Engineering,
5(4), 1983.
9. Hoigne, J. and H. Bader. Identification and
Kinetic Properties of the Oxidizing Decomposi-
tion Products of Ozone in Water and its Impact
on Water Purification. Second International
Symposium on Ozone Technology, International
Ozone Institute, 1976.
10. Perry, R.H., et al. Perry's Chemical Engineers
Handbook. McGraw-Hill Book Company, New
York, NY, 1963.
11. Jolley, R.L., etal. Effects of Chlorine, Ozone, and
Ultraviolet Light on Nonvolatile Organics in
Wastewater Effluent. In: Progress in Waste-
water Disinfection Technology—Proceedings of
the National Symposium, Cincinnati, Ohio.
EPA-600/9-79-018, NTIS No. PB-299338, U.S.
Environmental Protection Agency, Cincinnati,
OH, 1979.
12. Fochtman, E.G. and J.E. Huff. Ozone-Ultraviolet
Light Treatment of TNT Wastewater. Second
International Symposium on Ozone Technology,
International Ozone Institute, 1976.
13. Singer, P.C. and W.B. Zilli. Ozonation of Am-
monia in Municipal Wastewater. First Interna-
tional Symposium on Ozone for Water and
Wastewater Treatment, International Ozone
Institute, 1975.
14. Narkis, N., et al. Ozone Effect on Nitrogenous
Matter in Effluents. Journal of the Environ-
mental Engineering Div., ASCE, 103(EE5):877-
891, 1977.
15. Venosa, A.D. Current State-of-the-Art of Waste-
water Disinfection. JWPCF, 55(4):457-466,
1983.
16. Farooq, S., et al. Criteria of Design of Ozone
Disinfection Plants. Forum on Ozone Disinfec-
tion, International Ozone Institute, 1976.
17. Stover, E.L., et ai. High Level Ozone Disinfection
of Municipal Wastewater Effluents. EPA Grant
No. R804946, 1980.
18. Rakness, K.L, et al. Design, Start-Up, and
Operation of an Ozone Disinfection Unit.
JWPCF, 56(11 ):1152-1159, 1984.
19. Stover, E.L. Optimizing Operational Control of
Ozone Disinfection. In: Municipal Wastewater
Disinfection—Proceedings of Second National
Symposium, Orlando, Florida. EPA-600/9-83-
009, NTIS No. PB83-263848, U.S. Environ-
mental Protection Agency, Cincinnati, OH,
1983.
20. Venosa, A.D., et al. Disinfection of Secondary
Effluent With Ozone/UV. JWPCF, 56(2):137-
142,1984.
21. American Society of Heating, Refrigerating and
Air-Conditioning Engineers, Inc., Handbook,
Equipment Volume, 1983, Chapter 7; Funda-
mentals Volume, 1981, Chapters 5, 6, 19;
Systems Volume, 1984, Chapters 16,28.
753
-------
22. Venosa, A.D., et al. Comparative Efficiencies of
Ozone Utilization and Microorganism Reduction
in Different Ozone Contactors. In: Progress in
WastewaterDisinfectionTechnology—Proceed-
ings of the National Symposium, Cincinnati,
Ohio. EPA-600/9-79-018, NTIS No. PB-
299338, U.S. Environmental Protection Agency,
Cincinnati, OH, 1979.
23. Rosen, H.M. and C. Scaccia. Ozone Contacting
for Wastewater Disinfection. Presented at the
Third International Symposium and World Con-
gress of the International Ozone Institute, May
4-6,1977.
24. Given, P.W. and D.W. Smith. Pilot Studies on
Ozone Disinfection andTransfer in Wastewater.
In: Municipal Wastewater Disinfection—Pro-
ceedings of Second National Symposium, Or-
lando, Florida. EPA-600/9-83-009, NTIS No.
PB83-263848, U.S. Environmental Protection
Agency, Cincinnati, OH, 1983.
25. Gan, H.B., et al. The Significance of Water
Quality on Wastewater Disinfection With
Ozone. Forum on Ozone Disinfection, Interna-
tional Ozone Institute, 1976.
26. Venosa, A.D., et al. Disinfection of Filtered and
Unfiltered Secondary Effluent in Two Ozone
Contactors. Environment International. 4:299-
311, 1980.
27. Meckes, M.C., et al. Application of an Ozone
Disinfection Model for Municipal Wastewater
Effluents. JWPCF, 55{9):1158-1162, 1983.
28. Masschelein, W.J. Contact Columns and Bub-
ble-Dispersing Systems. Ozone Manual for
Water and Wastewater Treatment. John Wiley
& Sons, New York, NY, 1982.
29. Mostin, R. Principles in Supplying Electrical
Energy to Ozonators. Ozone Manual for Water
and WastewaterTreatment. John Wiley & Sons,
New York, NY, 1982.
30. Masschelein, W.J. The Direct Action of Dis-
persed Ozonized Gas Bubbles. Ozone Manual
for Water and Wastewater Treatment. John
Wiley & Sons, New York, NY, 1982.
31. Diaper, E.W.J. Gas Preparation for Ozone
Generation. Proceedings of Seminar on The
Design and Operation of Drinking Water Facil-
ities Using Ozone or Chlorine Dioxide, New
England Water Works Association, Volume 1,
June 1979.
32. Rakness, K.L, et al. Case History: Ozone Dis-
infection of Wastewater with an Air/Ozone
System. Proceedings of Wastewater Disinfec-
tion. State-of-the-Art Workshop, 1979.
33. Varas, A.J. New York City's Ozone Demonstra-
tion Plant Design. Presented at International
Ozone Association Conference, Montreal, Sep-
tember 11, 1984.
34. Robson, C.M. Engineering Aspects of Ozona-
tion. Proceedings of the Seminar on The Design
and Operation of Drinking Water Facilities
Using Ozone or Chlorine Dioxide, New England
Water Works Association, Volume 2, June
1979.
35. Gerval, R. Specifications and Performance
Control for Ozone Generators. Ozone Manual
for Water and Wastewater Treatment. John
Wiley & Sons, New York, NY, 1982.
36. Chapsal, P. A Practical Type of Thermal Residual
Ozone Destructor. Ozone Manual for Water and
Wastewater Treatment. John Wiley & Sons,
New York, NY, 1982.
37. Damez, F. Materials Resistant to Corrosion and
Degradation in Contact with Ozone. Ozone
Manual for Water and Wastewater Treatment.
John Wiley & Sons, New York, NY, 1982.
38. Damez, F. Safety Measures to Protect the
Equipment. Ozone Manual for Water and Waste-
water Treatment. John Wiley & Sons, New
York, NY, 1982.
39. Masschelein, W.J. Practical Aspects of the
Recycling of Effluent Gas Into Generation
Systems. Ozone Manual for Water and Waste-
water Treatment. John Wiley & Sons, New
York, NY, 1982.
40. Birdsall, A.C., et al. lodometric Determination of
Ozone. Analytical Chemistry, 24(4):662-664,
1952.
41. Lorenz, W., et al. Operations Histories of Two
Ozone Systems for Wastewater Disinfection.
Presented at the WPCF 57th Annual Confer-
ence, New Orleans, Louisiana, October 1984.
42. Hegg, B.A. et al. Evaluation of Pollution Control
Processes: Upper Thompson Sanitation District.
EPA-600/2-80-016, NTIS No. PB80-212855,
U.S. Environmental Protection Agency, Cincin-
nati, OH, 1980.
43. Hill, A.G. and H.T. Spencer. Mass Transfer in a
Gas Sparged Ozone Reactor. First International
754
-------
Symposium on Ozone for Water & Wastewater
Treatment, International Ozone Institute, 1975.
44. Opatken, E.J. Economic Evaluation of Ozone
Contactors. In: Progress in Wastewater Disinfec-
tion Technology—Proceedings of the National
Symposium, Cincinnati, Ohio. EPA-600/9-79-
018, NTIS No. PB-299338, U.S. Environmental
Protection Agency, Cincinnati, OH, 1979.
45. Houston, M., et al. Mass Transfer of Ozone to
Water: A Fundamental Study. Ozone: Science
and Engineering, 2:337-344, 1981.
46. Grasso, N. The Effect of the Gas Flow Rate to
Static Liquid Volume Ratio on Disinfection in a
Diffused Bubble Ozone Contactor. Master's
Thesis, Purdue University, May 1979.
47. Bollyky, LJ. and B. Siegel. Ozone Disinfection of
Secondary Effluent. Water & Sewage Works,
124(4):90-92, 1977.
48. Perrich, J., et al. Ozone Disinfection and Oxida-
tion in a Model Ozone Contacting Reactor.
AlChE Symposium Series, Volume 73, Number
166,1976.
49. Legeron, J.P. ContactTime of Ozonation. Ozone
Manual for Water and Wastewater Treatment.
John Wiley & Sons, New York, NY, 1982.
50. Nebel, C. Ozone Water Treatment Systems.
Water Engineering and Management, 1981.
51. LePage, W.L. Case Histories of Mishaps Involv-
ing the Use of Ozone. Presented at International
Ozone Association Conference, Montreal, Sep-
tember 1984.
52. Coste, C. Excess Ozone Disposal. Ozone Manual
for Water and Wastewater Treatment. John
Wiley & Sons, New York, NY, 1982.
53. Horst, M. Removal of the Residual Ozone in the
Air After the Application of Ozone. Ozone
Manual for Water and Wastewater Treatment.
John Wiley & Sons, New York, NY, 1982.
54. Orgler, K. Methods and Operating Costs of
Ozone Destruction in Off-Gas. Ozone Manual
for Water and Wastewater Treatment. John
Wiley & Sons, New York, NY, 1982.
55. Chapsal, P. Trailigaz Ozone Generator Tech-
nology. Ozone Manual for Water and Waste-
water Treatment. John Wiley & Sons, New
York, NY, 1982.
56. Masschelein, W.J. Thermodynamic Aspects of
the Formation of Ozone and Secondary Products
of Electrical Discharge. Ozone Manual for Water
and WastewaterTreatment. John Wiley & Sons,
New York, NY, 1982.
57. Gordon, G. and J. Grunwell. Comparison of
Analytical Methods for Residual Ozone. In:
Municipal Wastewater Disinfection—Proceed-
ings of Second National Symposium, Orlando,
Florida. EPA-600/9-83-009, NTIS No. PB83-
263848, U.S. Environmental Protection Agency,
Cincinnati, OH, 1983.
58. Venosa, A.D., et al. Reliable Ozone Disinfection
Using Off Gas Control. JWPCF, 57(9):929-934,
1985.
59. Langerwerf, J.J. Prolonged Ozone Inhalation
and its Effects on Visual Parameters. Aerospace
Medicine, #36 (June 1963)
60. Paur, R.J. and F.F. McElroy. Technical Assist-
ance Document for the Calibration of Ambient
Ozone Monitors. U.S. EPA-600/4-79-057, NTIS
No. PB80-149552, U.S. Environmental Protec-
tion Agency, Cincinnati, OH, 1979.
Safety Information:
Occupational Safetyand Health Administration.
title 29, Chapter XVII, 1910.1000.
Occupational Health Guideline for Ozone. U.S.
Department of Labor, 1978.
Safety and Health Requirements Relating to
Occupational Exposure to Ozone. American
Society for Testing and Materials, 1977.
Standard Practices for Safety and Health Re-
quirements Relating to Occupational Exposure
to Ozone. American Society for Testing and
Materials, 1977.
Chemical Hazards Bulletin—regarding ozone.
American Insurance Association, 1969.
Standard Practices for Safety and Health Re-
quirements Relating to Occupational Exposure
to Ozone. American Society for Testing and
Materials, 1977.
755
-------
-------
Chapter 7
Ultraviolet Radiation
7.1 Introduction
The use of ultraviolet (UV) radiation for the disinfec-
tion of wastewaters, relative to the established
technologies of chlorination and ozonation, is an
emerging process application which has been devel-
oped over the past 10 years. As a perspective, three
major committee reports were issued in the mid to
late seventies (1 -3). All effectively described the use
of ultraviolet radiation as "potentially" advantageous
for the disinfection of relatively high quality treated
wastewater. The U.S. Environmental Protection
Agency (USEPA) Task Force Report on the Disinfec-
tion of Wastewater (1) concluded that: "although
ultraviolet light has not been widely used to disinfect
wastewater, there is limited information that indi-
cates it may become a potentially desirable alterna-
tive. It is the only physical process whereas all the
disinfectants are chemical processes. On-going
research will provide answers as to its applicability to
adequately disinfect wastewater." The Task Force
went on to recommend that "the use of alternate
disinfectants should be further pursued because of
recent findings of the potentially hazardous halo-
genated organics in drinking water."
This chapter presents current state-of-the-art know-
ledge on the design of the UV disinfection process,
much of which represents information developed
over the past decade through research and demon-
stration of large scale applications of UV to the
disinfection of wastewater. As a new application of
the technology, the process design procedures are
still formative, and in-field experience in the operation
and maintenance of UV facilities is limited, but
growing as new plants come on line. The underlying
conclusion which should be stated at the beginning of
this chapter is that the potential which had been
foreseen earlier has been confirmed; the recom-
mended investigations into UV and the demonstration
of its application on a full scale basis have shown the
process to be viable, feasible for application to a wide
range of wastewater qualities, effective in the
inactivation of pathogens, capable of complying with
disinfection goals, and cost-effective. Its advantages
lie in its relative simplicity and in the absence of both
a residual and any chemical intermediates.
7.7.7 Chapter Description
It is reasonable to state that the UV disinfection
process has reached a state of development where
the mechanisms are understood and the critical
design parameters have been identified and generally
demonstrated. Field experience is limited, but gain-
ing. There is not, however, a clear and concise
compilation of this information, including the pro-
cedures by which a UV system can be designed or
evaluated. This manual attempts to provide this, in
addition to O&M considerations, which have been
identified and demonstrated by direct field exper-
ience.
The objective, then, is to bring together the knowledge
and experiences with UV as it is applied to the
disinfection of municipal wastewaters. It is not the
intent of these discussions to give an exhaustive
teaching on the technical aspects of the components
which make up the UV hardware, e.g., lamps, ballasts,
etc.; rather the approach will be to discuss the basic
concept and status of these components. References
are provided if the reader wishes to pursue these
aspects in more detail. The chapter will focus its
attention .on primary considerations for the design of
a system such that it will meet both its performance
requirements and will be amenable to efficient
operations and maintenance (O&M). •
Introductory Section (7.1). This gives an overview of
the technology and its current status relative to
wastewater disinfection application. Of particular
interest are descriptions of UV reactors; these will
give the reader a visual perspective of the system,
which will be helpful in the subsequent sections on
design. The chapter also provides listings of UV plants
in Tables 7-2, 7-3, and 7-4, which will be helpful to
the designer or operator who wishes to learn about
UV disinfection experience at other installations.
Background Discussions (7.2). This section is not
critical to the designer. It is useful, however, if one
wishes to gain a perspective on the mechanisms of
UV inactivation and the evolution of the process as it
is applied to wastewater disinfection.
Process Design Considerations (7.3). This is the
757
-------
section most important to the designer. It details the
process design elements critical to effective design
and offers guidance on defining specific design
parameters. These particularly address:
a. Hydraulics (7.3,2); of interest are discussions on
dispersion, reactor layout, headloss, and resi-
dence time distribution.
b. Intensity (7.3.3); a calculation technique is used,
solutions are presented to give the intensity of
any practical lamp reactor configuration, as a
function of the UV density and the wastewater
UV absorbance coefficient.
c. Wastewater characteristics (7.3.4); this gives
the designer guidance on the important waste-
water parameters. These include the flow, initial
density, suspended solids, UV absorbance co-
efficient, and tthe inactivation rates. Existing
data are compiled for these parameters.
Process Design Example (7.4). A design example is
given to demonstrate the design protocol, incorpo-
rating the considerations discussed in 7.3. The
designer can use this as a stepwise outline for
developing the design of a UV reactor.
O&M and Facilities Design Considerations (7.5). This
last section should be used by both the designer and
operator. It provides guidance related to effective
O&M. Of particular interest are the cleaning aspects
of the reactors. This is the single most important
element for effective reactor performance.
7.7.2 General Description of the UV Process
Disinfection by ultraviolet radiation is a physical
process relying on the transference of electromag-
netic energy from a source (lamp) to an organism's
cellular material (specifically, the cell's genetic
material). The lethal effects of this energy result
primarily from the cell's inability to replicate. The
effectiveness of the radiation is a direct function of
the quantity of energy, or dose, which was absorbed
by the organism. This dose is described by the product
of the rate at which the energy is delivered, or
intensity, and the time to which the organism is
exposed to this intensity.
The basic kinetics of disinfection have been discussed
as part of Chapter 4. The ideal UV disinfection model
follows first order kinetics, whereby:
N = N0e-klt
where:
N = bacterial density remaining after exposure to
UV
No = initial bacterial density
k = rate constant
I = intensity of UV radiation
t = time of exposure
The product It is the UV dose. Thus the response,
noted by the log of the survival ratio, N/N0, can be
plotted against dose; the slope is the rate coefficient,
k. This is shown on Figure 7-1 (a). Deviation from this
model is generally manifested by "shoulders," where-
by minimal response is noted below a "threshold"
dose; and by tailing effects, often attributed to
occlusion (shadowing) of bacteria by particulate
matter.
Figure 7-1. General description of UV design.
(a)
z
o
a
Shoulder
" (Threshold dose)
Tailing
UV Dose (Dose = IT)
Ballast
(b)
(c)
Intensity
_ Power
Supply
Wastewater
Flowpath
Actual
UV "Demand"
Ideal (No Absorption)
Distance from Source
(d)
i /— Ideal Plug Flow
(
/filNs r- High Dispersion
X II ^/
The primary artificial source of UV energy, at present,
is the low pressure mercury arc lamp. It is almost
universally accepted as the most efficient and ef-
fective source for disinfection systems application.
The primary reason for its acceptance is that approx-
imately 85 percent of its energy output is nearly
monochromatic at the wavelength of 253.7 nan-
ometers (nm), which is within the optimum wave-
length range of 250 to 270 nm for germicidal effects.
The lamps are long (standard lengths are typically
158
-------
0.75 and 1.5 m (2.5 and 4.9 ft) arc lengths) thin tubes
(typically 1.5 to 2 cm (0.6 to 0.8 in) in diameter). The
radiation is generated by striking an electric arc
through mercury vapor; discharge of the energy
generated by excitation of the mercury results in the
emission of the UV light.
These lamps can be suspended outside the liquid to
be treated or submerged in the liquid; the intent is to
get the energy into the liquid as efficiently as possible.
Typically, if the lamp is to be submerged into the
liquid, it is inserted into a quartz sleeve to minimize
the cooling effects of the water. Figure 7-1 (b) is
presented to schematically represent the principal
concerns when considering UV disinfection. In this
example, the lamp is placed in the liquid, with the
lamp perpendicular to the direction of flow. Other
configurations may have the lamp parallel to flow, or
the lamp may be suspended above the flowing liquid.
Referring to Figure 7-1 (c), as the lamp emits radiation,
the intensity will attenuate as the distance from the
lamp increases; this is due simply to the dissipation or
dilution of the energy as the volume it occupies
increases. A second attenuation mechanism involves
the actual absorption of the energy by chemical
constituents contained in the wastewater. This,
analogous to the chlorine demand, is the "UV
demand" of the wastewater.
The UV demand of a wastewater is quantified by a
spectrophotometric measurement at the key wave-
length of 253.7 nm; this expresses the absorption (or
transmittance) of energy per unit depth. The output is
absorbance units/cm, or a. u./cm. The percent trans-
mittance can be determined from this unit by the
expression:
% Transmittance = 100 x 10"(a'u-/cm)
The term most often used for design purposes is the
UV absorbance coefficient, a, expressed in base e:
UV absorbance coefficient, a = 2.3 (a.u./cm)
The unit for a is cm"1.
Although wastewater characteristics will be different
site to site, ra nges of the UV demand can be described
for different levels of treatment:
UV Absorbance
Coefficient Percent Absorbance
g(cm"1) Transmittance (a.u./cm)
-Primary
\ Treatment
Secondary
Treatment
Tertiary
Treatment
0.4 to 0.8
0.3 to 0.5
0.2 to 0.4
67 to 45 0.174 to 0.3 5
74 to 60 0.13 to 0.22
82 to 67 0.087 to 0.174
A second major concern is the provision of adequate
exposure time to the microorganisms in order to meet
the dose requirement at a given intensity. This was
also generally discussed in Chapter 4; the key is to
have plug flow through the system (see Figure 7-1 (d))
such that each flow element resides in the reactor for
the same amount of time. Perfect plug flow is not
going to be achieved, of course; some dispersion will
exist, such that there will be a distribution of exposure
times about the ideal, theoretical exposure time. A
design objective will be to minimize this distribution.
7.1.3 Current System Designs
In all, the design of a UV system must accommodate a
few simple considerations: satisfy the UV demand of
the wastewater; maximize the use of the UV energy
being delivered by the lamps; and provide the
conditions which encourage plug flow. Before pro-
ceeding with the detailed discussions of various
technical aspects of the UV process, it is appropriate
to first gain a perspective of UV system configura-
tions. This is best done by reviewing design con-
figurations which are currently being used at full-
scale plants. This is done to enable the reader to
better "visualize" the subsequent discussions. The
use of these figures does not suggest that the
configurations represent optimal designs; in fact,
certain design configurations have been demon-
strated to be inefficient.
Two basic generic reactors encompass current de-
signs. The first is a contact reactor in which the lamps
are submerged at all times in the wastewater; the
submerged systems have the lamps encased inquartz
sleeves which are only slightly larger in diameter
than the lamp itself. The second reactor design does
not allow contact of the water with the lamp (i.e., the
quartz sleeve), but rather suspends the lamp above
the liquid or surrounding conduits carrying the liquid.
These conduits are transparent to the UV radiation.
Let us first consider the so-called submerged quartz
systems. These can take on any number of configura-
tions, generally described by the arrangement of the
lamps relative to the direction of flow and to the
hydraulic design of the lamp reactor. A common
approach is the encasement of the lamp battery in a
sealed reactor shell, as shown by the schematic on
Figure 7-2. Flowenters the unit through an inlet pipe,
typically perpendicular to the lamps, redirects and
flows parallel to the lamps, finally exiting the reactor
through the outlet pipe. A modification of this
arrangement was provided at the Vinton Water
Pollution Control Plant, Vinton, Iowa, as shown on
Figure 7-3 (4). A steel plate was installed to split the
cylinder in half lengthwise. Flow is directed down the
unit on one side, then turns and flows down the
second half before discharge. This encourages a plug
flow condition by increasing the length of travel
159
-------
Figure 7-2. Example of closed vessel UV reactor, with flow parallel to lamps (Courtesy of Ultraviolet Purification Systems, Inc.,
Bedford Hills, NY).
Illuminated Lamp
Monitoring Panel
Quartz Jackets
Enclosing UV Lamps
Optional Ultrasonic
Cleaning Transducer
UV Intensity
Measuring Cell
160
-------
Figure 7-3. Schematic of quartz UV unit in Vinton, Iowa (4).
Cut-Out View of
Lamp Battery
Inlet
Flow Pattern
End Panel
Top View
Outlet
Internal Baffle Wall
End Panel
Concentric Circles of Lamps,
160 Lamps Total
Internal
Front View Baffle Wall
Figure 7-4. / Schematic of quartz UV unit in Suffern, NY.
relative to the unit's hydraulic radius. Subsequent
testing of this unit indicated that shortcircuiting
occurred within the reactor and that its effective
volume was significantly reduced. This is discussed
further in Section 7.3.2.
The lamps can also be arranged perpendicular to the
direction of flow in the same type of cylindrical reactor
shell. Baffle plates distribute the wastewater along
the length of the lamp battery; the flow is then
directed upward through the lamp battery and over an
internal overflow weir which runs the length of the
reactor.
The sealed reactors can also be arranged to simulate
channel flow. An example is provided on Figure 7-4,
which is a schematic cross section of the UV units
installed at the Suffern Water Pollution Control Plant,
Suffern, New York. The lamps in this case are
arranged in a symmetrical array, perpendicular to the
direction of flow. The wastewater is pumped to the
inlet chamber; a perforated baffle plate separates the
chamber from the lamp battery to distribute the flow
across the inlet plane of the lamp battery. A second
plate is installed on the outlet side of the lamp section
Baffle Plate
Outlet Pipe
Top View
Ultrasonic
Transducers
Flow
Pattern
\ UV Lamp
Bankl
Motor-Operated
Inlet Valva
\Bank2 \Bank3
UV Lamp Inside Quartz
/ Sleeve (Typ.)
ii
I
Bank 4
Manually-Operated
Outlet Valve
nnn
Side View
161
-------
before the liquid enters the outlet chamber for
discharge through the effluent pipe. The lamps are
staggered to encourage turbulance and the system is
arranged with a long path length to influence a plug
flow condition.
Figure 7-6.
Schematic of quartz UV unit in Albert Lea,
Minn.
Figure 7-5.
Example o'f open channel unit at Pella, Iowa,
with flow directed perpendicular to lamps (5).
3/8"Mesh
Screen
Guide Arms for
. Mechanical Wipers
Power Controls and
Ballasts for UV Lamps
Plan View
Weirs
Support Frame
Drive Cy'inder
for Wiper
Symmetrical
Lamp Array
Wastewater
Mechanical Flowpath
Wiper Frame
Quartz-Sheathed Lamp
The submerged quartz systems are also arranged as
open channel systems operating under gravity flow.
An example is provided in Figure 7-5, which schemat-
ically presents the UV system installed at the Sents
Creek Water Pollution Control Plant, Pella, Iowa (5). In
this case the lamps are arranged in a symmetrical
array, in the fashion of an open rectangular box as
shown on the lower panel of Figure 7-5. The lamp
battery is inserted into an open channel with the
lamps perpendicular to the direction of flow (upper
panel). Downstream of the lamp battery, the waste-
water is collected in effluent launders for final
discharge. This open-channel effect is also simulated
by encasing the same type of lamp battery between
open influent and effluent tanks. An example of this is
the Albert Lea Water Pollution Control Plant, Albert
Lea, Minnesota; a schematic of the units installed at
this plant is presented on Figure 7-6.
The UV system for the Tillsonburg, Ontario (Canada)
Water Pollution Control Plant is also installed as an
open channel system. In this case, however, the
Flovv
/
Wiper Pulley
|]-" Mechanism
Inl
Ch
1
3t
amber
3 Banks
of UV Lamps
'erpendiculai
to Flow.
1 Bank=
.108 Lamps
Outle
Cham
je r
7^
Cut-Out View of_y
Lamp Battery
Top View
Baffle
Wall (Typ.)
Filtered I /
Water*"|/
N
Header "'
Flow Pattern '
Screen (Typ.)-
4
\
^
^
1 8 Rows of
Lamps
8 Columns o
Lamps
Col. 1-6, Ban
Col. 7-13, Ba
Col. 13-18, B
UV
.amps
uv
.amps
k1
nk2
ank3
'
/
'/I
Floor
-------
Figure 7-7. Schematic of open-channel, modular UV system
(Courtesy of Trojan Technologies, Inc., London,
Ontario, Canada).
Figure 7-8. Example of UV system utilizing Teflon tubes
(Courtesy of Ultraviolet Technology, Inc.,
Rancho Cordova, California).
System UV 2000
UV Modules in
Effluent Channel
Removable Lamp Support Rack
with Internal Wiring
Teflon Tubes to
Carry Water
System UV 2000
UV Module Lifted
from Effluent
Channel.
held in enclosures above the lamp battery (Figure
7-8), or are remote from the reactors in separate
power panels (Figures 7-2 through 7-7). Instrumenta-
tion generally entails UV intensity monitors and
individual lamp operations circuitry. Control of the
system can be on a manual basis, often involving the
selective operation of modules, or banks of lamps
within a module, as a function of the hydraulic load to
the system. Automatic controls generally slave the
lamp bank (or in some cases the lamp voltage)
operations to the flow rate and/or the water quality.
A major element in the operation of UV systems is the
cleaning of the surfaces which must be kept trans-
parent to the UV radiation for efficient performance.
These include the quartz sheaths and the Teflon
tubes. Most commercial systems include accessory
equipment to assist in this cleaning task; these
include the use of mechanical wipers, ultrasonic
transducers, and the provision to chemically restore
the surfaces.
7.7.4 Current Technology Status
As suggested by the foregoing discussions, the UV
process is relatively simple. Not unlike chlorination,
an agent is added to the wastewater in sufficient
quantity to effect the inactivation of bacteria. Time in
this case is not provided to allow for a specific
reaction to take place, but rather to accomplish the
necessary dose. The effect on the microorganism is
not in itself lethal to the microorganism; the main
effect is to impose sufficient damage such that the
organism is unable to replicate. The process, like
ozonation, requires on-site generation of the germ-
icidal agent; the generator (the UV lamp), however, is
far simpler in concept and operation than that
required to produce ozone. Like both chemical
processes, UV must also satisfy a "demand" of
energy exerted by the wastewater itself.
163
-------
Other than the simplicity of the process, UV also
offers the advantages of system flexibility and a
capability of responding quickly to changes in de-
mand. There is relatively little complexity to the
hardware, and maintenance generally requires low
skill levels. The haxards of the process are low,
principally related to the high electrical loads and the
personal exposure to the UV radiation; these are
conditions which are easily safeguarded. A major
advantage of the process is the absence of a residual
in the wastewater and any subsequent impact on the
receiving water. A corollary to this is the ability to
"overdose" with UV land still not affect the receiving
water. This allows for a less rigorous control require-
ment than associated with the use of chlorine. The
absence of a residual can also be viewed as a
disadvantage when considering the operational con-
trol of the process. There is no immediate monitor of
performance analogous to the chlorine residual.
Since the energy levels are not high enough to affect
chemical reactions, there are no significant inter-
mediates formed by the process, even at overdose
levels. This is clearly an advantage over the chemical
addition processes.
Given its advantages, UV was still not seriously
considered as an alternative to chlorination for
wastewater disinfection until the mid-seventies. At
that point the process was considered more in
response to the negative aspects of chlorination; its
potential was acknowledged if the perceived disad-
vantages with the process were overcome. These
related to the lack of information on UV application to
low grade waters and the impact of various water
quality parameters (particularly suspended solids) on
design, the lack of any clearly defined design pro-
cedures, and the previous history of system fouling.
Subsequent investigations focused on these and
other aspects of UV disinfection. A number of full
scale plants were installed, encouraged by the
support of the USEF'A through its Innovative and
Alternative Technologies funding, under the Con-
struction Grants program. A total of 14 plants were
funded under I/A by the USEPA (6). These and other
facilities are providing much needed information on
the operation and maintenance requirements of the
systems and on refinements necessary for current and
future installation designs. Plants funded under the
I/A program are listed on Table 7-1.
Table 7-1. MunicipalitiesThat Have Received I/A Fundsfor
Designing and/or Constructing UV Disinfection
Facilities (October 1978 to June 1981)
Village of Suffern, New York
Woodstock, New York
Crawford, New York
Rhlneback, New York
Smithburg, Maryland
Clear Spring, Maryland
Evanston, Wyoming
Northfield, Minnesota
Albert Lea, Minnesota
Pel la, Iowa
Cassville, Missouri
Dexter, Maine
Kennebunk, Maine
Heston, Kansas
There are a significant number of plants now installed
throughout the United States for the disinfection of
treated municipal wastewaters. A list of UV installa-
tions which are operational, or in the design, bid, or
construct stage is presented in Tables 7-2, 7-3, and
7-4 (5). This information is further reduced in Table
7-5 to reflect the distribution of plants with regard to
size; as indicated, the existing plants in operation are
predominantly small (less than 3800 mVd or 1 mgd),
while plants in the planning or construction stage
tend to be larger.
7.2 Disinfection of Wastewaters by Ultra-
violet Radiation
7.2.1 Ultraviolet Light
As presented in Figure 7-9, the ultraviolet region of
the electromagnetic spectrum is generally defined as
those radiations with wavelengths greater than the
longest X-ray and less than the shortest wavelength
visible to man; these wavelengths typically are set
between40 and400 nanometers(nm).The ultraviolet
region itself is divided. Near ultraviolet radiation is
between 300 and 400 nanometers; far ultraviolet is
between 200 to 300 nm. These two bands of
ultraviolet are observed in solar radiation. Extreme
ultraviolet radiation describes that energy between
wavelengths 40 and 200 nm; energy at these
wavelengths is strongly absorbed by air and its
observation requires working in a vacuum or in a gas
which does not absorb the energy.
Qualitatively, light is almost universally described by
its wavelength. For the sake of convenience, a single
unit of wavelength will be used throughout this
chapter. This is the nanometer (nm), or 10~9 m. There
are 10 Angstroms per nanometer.
A more fundamental quantity describing electromag-
netic radiation is its frequency of vibration. The
frequency and wavelength of radiation are related by:
c =
(7-1!
where:
c = the velocity of light (3x1010 cm per second in
free space)
v = frequency of vibration (vibrations per second)
A = wavelength (cm)
A number of terms are used to express the quantity of
radiation. These physical units relate to work or
energy. The ones most commonly used are the erg,
calorie, and watt-second (joule); all are measures of
total quantity of energy or work. The time rates at
which this energy is delivered in the corresponding
164
-------
Table 7-2. Summary List of Facilities in the U.S.A. or Canada Utilizing Ultraviolet Light (UV) Disinfection Which are in Design
Size, mgd Other Treatment Equipment and/or
Facility Name Design Firm Design Start-up Processes Comments
Bristol
(CONNECTICUT)
Ridgefield
(CONNECTICUT)
Salmon
(IDAHO)
Ucon
(IDAHO)
Camp Point
(ILLINOIS)
Iowa City
(IOWA)
Kennebunk
(MAINE)
Limestone AFB
(MAINE)
Pittsville
(MARYLAND)
Poolesville
(MARYLAND)
Morton
(MINNESOTA)
Calhoun City
(MISSISSIPPI)
Deer Lodge
(MONTANA)
Dillon
(MONTANA)
Lewiston
(MONTANA)
Keys Associates
Jim Geremis
(401)861-2900
Alfaertson, Sharp & Ewing
Mike Pastore
(203) 846-4356
Ellsworth Engineering, Inc.
Gary Marshall
(208) 523-1662
Forsgren-Perkins Engineering
Dick Dyer
(208) 356-9201
W.H. Klinger & Associates
Dan Oliver
(217) 223-3670
Keenstra & Kimm, Inc.
Jim Kimm
(515) 225-5000
E.C. Jordan
A. Peter Krauss
(207) 775-5401
Dufresne & Henry
Barry Bastian
(207) 797-2010
Harrington & Associates
William Harrington
(301)768-5400
Kamber Engineers
Dennis Kamber
(301)840-1030
Ayres Associates
Dale Philstrom
(612) 644-0604
Willis Engineering Co.
Joe Sutherland
(601)226-1081
Christian, Spring, Sielbach
& Associates
John Connel
(406)656-6000
S & A Engineers
Robert Scruton
(406) 442-1532
HKM & Associates
Jim Kaercher
10.75 7.0
0.120 0.040
Not estab- Not estab-
lished lished
0.115 0.07-0.115
(80 gpm) (50-80 gpm)
0.16 0.11
13.0 9.0
1.3 0.7
2.5 (ADWF) 1.5
(peak 5.2)
0.125 0.06
0.6 0.35
0.132 0.067 (dry
weather)
0.342 0.175
1.5 1.3
(peak 3.3)
0.8 0.8 "
(winter-0.65,
summer-1.10)
2.9 2.3
(current 9.0)
Activated sludge
RBC System
Aerated lagoon/
Facultative lagoon
(secondary cell)
Facultative lagoon
(secondary cell); UV
is followed by 4-
acre storage pond
& land disposal
Two-cell stabiliza-
tion pond (30-day
Det. time) w/inter-
mittent sand filter
Activated sludge
RBC system
RBC system dry
weather
In wet weather, RBC
(2.5 mgd) & up to
4.1 mgd primary ef-
fluent
Oxidation ditch, ter-
tiary filtration
Sequencing Batch
Reactor, pressure
filter
Aerated stabilization
basin system
Hydrograph-control
release facultative
lagoon. Store
wastewater during
dry weather;
discharge 1 to 10
times design flow
during rainy season
Aerated lagoon sys-
tem
Aerated lagoon sys-
tem, storage basin
(180AF)
RBC system
(Equipment not yet
purchased)
N.C.
N.C.
N.C.
N.C.
N.C.
(Equipment not
purchased)
N.C.
N.C.
Flow through UV
unit will be under
slight pressure
N.C.
UV selected due to
periods of no flow,
not because of ef-
fluent require-
ments.
N.C.
N.C.
N.C.
(406) 245-6354
765
-------
Table 7-2. (Continued)
Size, mgd Other Treatment
Facility Name Design Firm Design Start-up Processes
Stratford #1
(NEW HAMPSHIRE)
Stratford #2
(NEW HAMPSHIRE)
Whitefisld
(NEW HAMPSHIRE)
Woodstock
(NEW YORK)
Harrimon
(NEW YORK)
Loomis
(NEW YORK)
Ironton
(OHIO)
Northridge Subdivision
Piedmont
(SOUTH DAKOTA)
Beckley
(WEST VIRGINIA)
Buckhannon
(WEST VIRGINIA)
Green Valley-
Glenwood PSD
(WEST VIRGINIA)
Marshall Co. PSD #1
(WEST VIRGINIA)
Opequen Hcdgesville
(WEST VIRGINIA)
PAX
(WEST VIRGINIA)
Harper Eccles WWTP
Raleigh Co. PSD
Hoyle, Tanner & Associates 0.056
Gene Forbes
(603) 669-5420
Hoyle, Tanner & Associates 0.024
Gene Forbes
(603) 669-5420
Phillips & Emberly 0.185
William Emberly
(1302) 434-2142
Phillip J. Clarke 0.230
David Wright
(716) 454-4570
Phillip J. Clarke 4.0
John Tarolli
(914) 294-8818
Phillip J. Clarke 0.080
David Wright
(716) 454-4570
Brundage, Baker & Stouffer, 1 .7
Ltd
George Haggard
(614) 888-3100
Hoskins, Western & Son- 0.023
deregger
Al Foster
(605) 342-4105
Greenhorns & O'Mara 3.5
Gary Beech
(301)982-2837
Kelley, Gridley, Blair Wolfe 2.5
Jim Downey
(304) 345-0470
Pentree 1 .5
Will Smith
Bob Hazelwood
(304) 425-9581
Green International 0.17
Norman Katz
(412) 471-5348
HNTB 0.80
Dave Wright
(4'I4) 463-2310
G.A. Tice 0.06
Geiorge Tice
(304) 255-5400
Greenhorne & O'Mara 0.100
Turgay Ertugal
0.053 Slow rate (0.5 gpd/
ft2) sand filters
0.021 Slow rate sand fil-
ters (0.5 gpd/ft2)
0.133 Aerated lagoon sys-
tem
0.180 Oxidation ditch,
sand filters in sum-
mer months only
2.0 Draft tube oxidation
(existing) ditch, tertiary sand
filters, parallel train:
existing activated
sludge, aerated pol-
ishing lagoon
0.054 Overland flow, aer-
ated polishing la-
goon
1.7 Trickling filter
0.011 Extended aeration
2.5-3.0 Activated sludge ex-
tended aeration
(post-aeration fol-
lows UV)
1.5 Oxidation ditch
(post-aeration fol-
lows UV)
1.0 Oxidation ditch
(post-aeration fol-
lows UV)
0.10 Extended aeration
(package plant)
0.55 Oxidation canal,
post-aeration fol-
lows UV
0.045 Facultative
lagoons
0.100 Oxidation ditch
(package unit)
Equipment and/or
Comments
N.C.
N.C.
N.C.
N.C.
Lead design firm
is: Erickson Schmitt,
Al Schmitt
(914)294-8838
N.C.
N.C.
State wants a ter-
tiary filter installed
before they will ap-
prove design
N.C.
Design around
ENERCO
N.C.
N.C.
N.C.
N.C.
N.C.
(WEST VIRGINIA)
(301)982-2800
166
-------
Table 7-2. (Continued)
Facility Name
Design Firm
Design
Size, mgd
Start-up
Other Treatment
Processes
Equipment and/or
Comments
Salt Rock PSD
(WEST VIRGINIA)
Riverton
(WYOMING)
Athens
(WISCONSIN)
Dunn Engineers 0.260 0.246
Dave Schultz
(304) 342-3436
Little Black
(WISCONSIN)
Collingwood
(ONTARIO, CANADA)
Airex Engineers
Harry La Bonde
(307) 856-6505
Becker Hoppe Engineering
Gerald Bizjak
(715) 359-6147
Carl C. Crane, Inc.
Victor Marz
(608) 238-4761
4.95 2.3
0.225 0.070
0.012 0.008 to
0.010
Ainley & Associates, Ltd. 1.2 0.5 low
(Owen Sound, Ont.)
Colin Kent
(705) 445-3451
Oxidation ditch
Oxidation ditch
Aerated lagoon (3
cells: primary, sec-
ondary, storage).
Plant has controlled
discharge of efflu-
ent by using the
storage lagoon
Recirculating sand
filters
Activated sludge,
tertiary filtration
N.C.
U.V. Technologies
Anticipating the
use of ENERCO
N.C.
Anticipating Trojan
Industries
N.C. - No comments.
units are ergs per second, calories per second, and
the watt. These units and their cgs equivalents are as
follows:
Watt-second (or joule) w-sec 107 ergs
Calorie cal 4.2 x107 ergs
Erg - -
The intensity or energy density of the radiation is
expressed in terms of energy incident upon a unit
area. The unit used in the context of this report is the
micro-watt per square centimeter (/uwatt/cm2).
Quantum theory states that radiant energy occurs in
discrete units, or quanta. The energy of these
fundamental units is related to its frequency:
E = hv = he/A
(7-2)
where:
E = energy of a single quantum (ergs)
h = Planck's constant (6.62 x 10~27 erg-sec)
c = velocity of light (3 x 1010 cm per sec)
v = frequency (vibrations per second)
/\ = wavelength (cm)
The quantum is a very small energy unit, equivalent to
(19.86 x 10~17)/wave length, in cm-ergs. From this
expression it is shown that the energy content of a
quantum is identical for a given wavelength of light.
7.2.1.1 Source of UV Radiation
Practical application of UV for purposes of disinfection
required a high intensity source at the desired
wavelengths. This can be traced by the evolution of
the mercury vapor lamp. In 1835, Wheatstone
described the intense light emitted when mercury is
vaporized in an electric arc. The first true mercury
vapor lamp was constructed by Downing and Keating
in 1896 by passing an electric discharge through
mercury in a partially evacuated tube. The problem
was that the arcs would eventually go out because of
the increase in the pressure of the vapor. Cooper-
Hewitt resolved this problem in 1901 by devising a
lamp in which the mercury was condensed at the
same rate at which it was vaporized. This, along with
the development of fused quartz and ultraviolet
transmitting glass, initiated the successful commer-
cial development of mercury vapor lamps (7).
The discharge type lamps were relatively inefficient,
however, due to their low selectivity in the use of the
energy, or electrical input. The generation of heat,
excitation of several different spectral lines, and
inefficient electrodes resulted in a distribution of
energy to many outputs. The key development came
in the 1920s, when it was determined that a dis-
charge through a mixture of mercury vapor at a
precisely optimum pressure and a rare gas (typically
argon) at a somewhat higher pressure was extremely
efficient in converting the electrical energy to ultra-
violet light. Fully 60 percent of the energy input could
be converted to monochromatic radiation at 253.7 nm
(8).
The commercial development of the mercury-rare gas
discharge lamps was directed to its use as a light
source. The development of a suitable fluorescent
167
-------
Table 7-3. Summary List of Facilities in the U
Construction.
Facility Name Design Firm
Galney Ranch
(Arizona)
Payson
(Arizona)
Augusta
(Arkansas)
Hebor Springs
(Arizona)
Presque Isle
(Maine)
Clear Springs
(Maryland)
Milford
(Massachusetts)
Bomidji
(Minnesota)
North Koochiching
Area San. District
(Minnesota)
Bonne Terre
(Missouri)
Emminence
(Missouri)
Frederick Town
(Missouri)
Mineral Belt Area
WWTP
(Flat River, Mo.)
(Missouri)
Noel
(Missouri)
Greely & Hanson
Elizabeth Zureick
(602) 992-5000
Moore, Knickerbocker
& Assoc.
Terry Moore
(602) 265-3776
Mehburger, Tanner,
Robinson & Assoc.
Daryl Laws
(501)375-5331
Boulder Engineers
Jim Little
(501)362-3118
Wright, Pierce, Barnes &
Wyman
Dave Fuller
(207) 725-8721
Fellows, Reed & Assoc.
Ed Renn
(301) 739-5660
Haley & Ward, Inc.
Ben Bugbee
(617) 890-3980
Rieke, Carroll, Muller &
Assoc.
Warren Kerstan
(612) 935-6901
Widseth, Smith, Nolting &
Assoc.
Don Anderson
(218)829-5117
Metropolitan Engineering
Robert Vogler
(314) 948-3860
Missouri Engineering Co.
Corky Stack
(314) 364-4003
Crane & Fleming
Greg Boettener
(314) 221-4048
Metropolitan Engineering
Robert Vogler
(314) 467-3860
Allgeier, Martin, & Assoc.
Jan Tupper
.S.A. or Canada Utilizing Ultraviolet (UV) Disinfection Which are Under
Size, mgd Other Treatment Equipment and/or
Design Start-up Processes Comments
1.7 (8 mo.)
1.0 (4 mo.
winter)
1.7
0.6
(current 0.4)
1.8
(1250 gpm)
2.3 dry
weather
(5.4 wet
weather)
0.20
4.3
2.7
2.3
0.6
2.9
0.85
2.0
0.2
1.7
0.5
0.4-0.5
1.8
(intermittent
pumped
flow; ulti-
mately; flow
will be con-
tinuous)
0.7
(1.9 with I/I)
0.12
2.0
1.6 (current)
1.1 (current)
0.4
0.5
0.6
1.6
0.15
Extended aeration,
sand filters
Bardenpho (& Clari-
fiers), backwash fil-
ters
Orbal treatment
system, Aeration
disks
Three-cell aerated
lagoon (facultative)
of 19 acres rapid
sand filters
Oxidation ditch sys-
tem
Oxidation ditch
RBC & tertiary fil-
ters
Activated sludge &
tertiary filtration
Trickling Filter
Oxidation ditch
Oxidation ditch,
sand filter
Oxidation ditch
Oxidation ditch
Oxidation ditch
ENERCO (formerly
UV Technologies)
UV Technologies
(Now ENERCO)
ENERCO #L1 000
UV Purification
Systems, Inc. (70
lamps; 9-sec.
retention)
Pure Water Sys-
tems, Inc. (Two
units-1 standby)
Not yet purchased
Not yet purchased
UV Technologies
UV Purification
Systems, Inc.
Not yet purchased
UV Purification
UV Purification
Not yet purchased
UV Technology
(417) 624-5703
168
-------
Table 7-3. (Continued)
Facility Name
Summerset Plant Div.,
South Jefferson Co.
(Missouri)
Winona
(Missouri)
Chinook
(Montana)
Bennington
(Nebraska)
Chatham Township
(New Jersey)
Rhineback
(New York)
Thompson
(New York)
Beech Mountain
(ski resort)
(N. Carolina)
Waynesburg
(Ohio)
Mt. Pleasant
(S. Carolina)
Coalville
(Utah)
Baker Heights
Berkley Co. PSSDa
(W. Virginia)
Moorefield
(W. Virginia)
Evanston
(Wyoming)
Worland
(Wyoming)
Madison Met
(Wisconsin)
Design Firm
Horner & Shifrin
(314) 531-4321
C.B. Simmons
C.B. Simmons
(417) 732-2092
Robert Peccia & Assoc.
Alden Beard
(406) 442-8160
Johnson, Erickson,
O'Brien & Assoc.
Terry O'Brien
(402) 443-4661
Keller, Kirkpatrick
Bob Kirkpatrick
(201)377-8500
Brinnier & Larios
Dennis Larios
(914) 338-7622
Phillip J. Clark & Assoc.
David Wright
(716)454-4570
Davis, Martin, Powell
Ed Powell
(919) 883-0032
Hammontree & Assoc.
Richard Hunsinger
(216)499-8817.
E.M. Seabrook, Inc.
Louis Couthen or Brian
Wright
(803) 884-4496
DMJM
Reed Fisher
(801) 262-2951
HNTB
David Wright
(414) 463-2310
Kelley, Gidley, Blair, Wolfe
Dick Kline
(304) 345-0470
Eckoff, Watson, Preater
John McNeil
(801) 486-5621
Airex Engineers
Harry LaBonde
(307) 856-6505
Consour Townsend
Ron Reising
(312) 938-0300
Size, mgd
Design Start-up
0.117
0.175
0.50
(current 1.10)
0.186
0.120
0.130
1.0
0.400
(peak @
2 1/2X)
4.0
3.2
3.0
0.34
0.477
2.9
(current 2.5)
1.12
50
(peak 11 5)
largest UV fa-
cility in world
0.1
0.100
0.35
0.065-0.070
very low
0.080
0.80
0.04-0.20
seasonal
0.23
1.0
(current)
0.25
0.34
0.400
2.0
0.8
(current)
35-40
Other Treatment Equipment and/or
Processes Comments
Lagoon facility (30-
day detention time)
Oxidation ditch
Oxidation ditch
Extended-aeration
activated sludge
RBC, multi-media
filter
Oxidation ditch
Draft tube oxidation
ditch, aerated pol-
ishing lagoon
Contact stabilization
Bio-drum
Conventional acti-
vated sludge
Oxidation ditch
Trickling filter
Aerated lagoon
Oxidation ditch
Aerated lagoons
Activated sludge
w/nitrification
UV Purification
Systems Inc.
ENERCO
ENERCO
Pure Water
Systems, Inc. and
U.V. Purification
Systems, Inc.
UV Technologies,
Inc.
ENERCO
ENERCO
ENERCO
UV Purification
Systems, Inc.
UV Purification
Systems, Inc.
Not yet purchased
ENERCO
ENERCO
UV Purification
Systems, Inc.
UV Purification
Systems, Inc.
UV Purification
Systems, Inc.
°PSSD = Public Service Sanitation District.
169
-------
Table 7-4. Summary Lint of Facilities in the U.S. A. or Canada Utilizing Ultraviolet Light (UV) Disinfection Which are in Operation
S!ze,mgd n.hpr Tr.»tmont
Facility Namo
Lake Croason, Cow
Hide Cove oroa
(AR)
Uko Ouachita, Lit-
tle Fir area
(AH)
Umar (AR)
Tillonsburg
(Ontario, Canada)
Edon
(Wisconsin)
Conttor Canter
(Colorado)
Eila (Colorado)
Cypron-Thompson
Creek, Challls
(Idaho)
Design Firm Design
U.S. Army Corps of En- 0.015
gineors
"Mac" Montgomery
(601) 634-5301
U.S. Army Corps of En- 0.018
gineors
"Mac" Montgomery
(601) 634-5301
Burrough, Uerling & 0.106
Brasuoll
David Uerling
(501) 646-5S59
Anderson Assoc. 2.4
Peter Laugh :on
(416) 497-8630
Arthur Technology 0.16
John Masters
(414) 922-6973
ADG Engineering, Inc. 0.015
Roger N. Venables
(303) 761-5142
Keith Bell & Assoc. 0.30
Keith Bell
Hamilton & Voeleur 0.720
(No longer in (500 gpm)
business) Contact:
Current Processes
0.015 Extended
aeration-activated
sludge, rapid sand
filters
0.018 Extended aeration-
activated sludge rapid
sand filters
0.100 Overland flow
1.3 Extended aeration
0.10 1° clarifier, roughing
filter extended activated
sludge (for nitrification)
0.005 Package-activated
sludge extended
0.18 Aerated lagoons (20-
day det. time)
0 3-stage lagoon
Equipment
U.V. Technologies
(Teflon Tubes)
U.V. Technologies
(Teflon tubes)
Now ENERCO
1 unit, 8-lamps
rated for 7500
hrs. use
U.V. Technologies
Trojan Industries
UV Technology
Ultradynamics
(Santa Monica,
CA)
ENERCO
ENERCO
Comments on Performance
In operation 1-1/2 years. Good perform-
ance.
In operation 1-1/2 years. Good perform-
ance. Operates March-Oct.
Disinfection requirements met. Frequent
bulb replacement. Failure to operate auto-
matically. Weir causes UV tanks to be filled
with silt. (Identical to Hatfield facility).
In operation for 2 years.
Achieving 1/2 100 MPN coliform/100 ml.
Good performance.
No performance problems to date, only 1
month of operation.
Tube fouling. Operational since Dec. 1 983.
Operates 30 days/year in September. Too
early to judge performance; flows too early
to judge performance; flows too low. Fecal
Nowdilo (Idaho)
former employee of
H & V Ranee Bane
Ellsworth Engrg.
(208)523-1662
Forsgren-Perkins Engrg. 0.045
Dick Dyer
(208) 356-9101
Earl Keemp
(8011364-4735 (Salt
Lake City Of:.)
Rod Top Meadows J.U.B. Engineers
(Kolchum) James Colernan
(Idaho) '(208)733-2414
Polla (Iowa)
Huston (Kansas)
0.180
(dry 0.06)
3.4
1.3
Sabbalus (Maine)
Togus VA Hosp.
(ME)
Old Towno
(Maryland)
Smithiburoh
(MD)
Voonstra, Kimm
Engineers
Jim Kimm, Mike
Foreman
(515)225-8000
Wilson & Co. Engineers
& Architects
Jim Dowell
(913) 827-0433
Contact: City of Heston,
Maurice Bowersox or
Bill Nitzsche (plant) @
(316) 327-4412 or 327-
2535
Woodward it Curran
Frank Woodward
(207) 839-6751
Hunter Balleau
Barrio Patrie
(296) 671-4721 or V.A.
Center,
Bob White
(207) 623-8411 X338
Allogany County Sani- 0.04
tation Commission
Kevin Beachy
Fellows, Reed & Assoc. 0.20
Ed Renn
(301) 739-56GO
0.022
0.180
1.5
Facultative lagoon
(Land disposal)
Extended aeration oxi-
dation ditch
Activated sludge
0.25-0.30 Orbal Activated Sludge
(wet weather (Effluent goes to golf
-0.40) course)
0.25
0.2
0.1
0.15
0.03
0.12
Imhoff tank & intermit-
tent sand filter system
Oxidation, ditch system U.V. Technologies
coliform die-off in ponds is 100%, prior to
UV disinfection.
ENERCO Start-up: Nov. 1983-99.9% bacterial die-
(Model G-30) off. Design engineer is Dick Dyer.
U.V.Technologies Good performance. Fecal coliform count =
(Now ENERCO) 0 for 100% fecal coliform kill. Bulbs
replaced annually. Operating since 10/82.
Pure Water Sys- Mechanical wiper system not functioning.
terns. Inc. Using chemical cleaning with weak acid so-
(Quartz jacketed lution for lamp jackets (outside) approxi-
UV lamps) mated every 2 weeks vs. 6 mos to 1 year.
New end seals were provided by the
manuf. Operational since April 1982 (Oct.
'81 for entire facility).
U.V. Purification Ultrasonic cleaning not working up expec-
Systems, Inc. tations. Consequently, light intensity is not
as expected. Must chemical clean the
lamps.
U.V. Technologies Meeting coliform count requirements. Op-
(G-500) erating 1 year. Still in shake-down. I/I prob-
lems; must bypass due to poor filtration.
Operating 1 year. Fecal coliform levels met.
Initially, wiring and bulbs were faulty. Cur-
rently, ballast that runs UV bulbs weakens,
diminishes intensity of bulbs. Must change
ballast often; life is less than 2,000 hours.
Extended Aeration acti-
vated sludge
Extended aeration
Pure Water Sys-
tems, Inc.
U.V. Purification
Systems, Inc.
High maintenance. Initial ballast problem-
now corrected. (Auto wiper system shuts
down UV unit when tubes get too dirty).
System in start-up.
770
-------
Table 7-4. (Continued)
Facility Name Design Firm
Thurmont (MD)
Albert Lea
(Minnesota)
Northfield (MN)
Cossville
(Missouri)
Clinton (MO)
Ozark (MO)
Briarwood (MO)
Yellowtail Power
Plant
(Montana)
Environmental
Disposal Corp.
(Pluckerrnan) (New
Jersey)
Educational
Testing (N.J.)
Crawford (NY)
Pennyann
(NY)
Suffern (NY)
McPherson
(Kansas)
Marietta
(Oklahoma)
Harrington & Assoc.
William Harrington
(301) 768-5400
Tolz, King, Duvall &
Anderson
Dave Kirkwold
(61 2) 292-4400
Bonestroo, Rosene,
Anderlick & Assoc.
Dick Turner
(612) 636-4600
Allgeier, Martin &
Associates
Jan Tupper
(417) 513-5703
Bucher, Willis J.
Ratliffe
Jim Swanson
(913) 827-3603
Anderson Engineering
Steven Brady
(417) 866-2741
Sanders, Stewart,
Gaston
Paul Kinshella
(406) 245-6366
Bureau of Reclamation
Craig Peterson
(408) 657-6141 or
Mr, Hergenraidar
(406) 666-2443
Environmental Design
Inc. (out of business)
Contact: Ray Ferrara
Princeton University
(609) 452-4653 or Neil
Callahan (Operator)
(201) 234-0667
CUHZA
Manny Dios
(609) 452-1212
Phillip J. Clarke &
Associates
John Tarolli
(914) 294-8818
Hershuy, Malone, &
Associates
Greg Barbour
(716) 381-9250
Thomas Riddick
Norman Lindsay
(914) 365-0446
Wilson & Co., &
Architects
Jim Dowell
(913) 827-0433
Contact: Plant operator,
Walt Hundley
(316) 241-3940
Bob McCoy
(Now retired) Contact:
Mark Daniels
State Environmental
Agency
(405) 276-5493
Size, mgd
Design
1.0
(4.0 weather)
12.53
2.5
0.5
2.0
0.72
0.180 (peak
0.72) (Health
Dept.
approved
0.123 to
date)
0.0006
0.85 (590
gpm)
0.080
0.15
1.8
1.9 (peak 4.0)
0.29 (200
gpm) (entire
plant @ 2.0
mgd)
0.231
Current
0.3
(2.9 weather)
3.91
2.1
0.7
(exceeding
design
capacity)
1.3
0.20
0.0005 (500
gpm)
0.001
0.43-0.57
(300-400
SP<")
0.030-0.038
0.08-0.085
0.6
1.2
0-0.14 (0-100
gpm) (entire
plant® 1.7
mgd)
0.19
Other Treatment
Processes
Oxidation ditch with
tertiary filtration
2-Stage activated
sludge, tertiary filters
2 Stage secondary
system (trickling filters
&RBC)
Oxidation ditch
Oxidation ditch
Oxidation ditch
Oxidation ditch
Extended aeration
package plant, tertiary
filtration
Bardenplho and
multi-media filter
Extended aeration,
filtration (package,
multi-media, high rate)
Oxidation ditch (septic
tank effluent)
RBC
"Lighting Complete
Mix" aeration (similar
to activated sludge.)
Trickling filter and
Contact Stabilization
basin in parallel.
Combined discharge
bypasses UV to creek
or a portion (100 gpm)
is discharged to lake.
when lake level is low.
Oxidation ditch
Equipment
ENERCO
Pure Water
Systems, Inc.
Pure Water
Systems, Inc.
Aquafine (No
longer in market)
U.V. Purification
Systems, Inc.
U.V. Purification
Systems, Inc.
ENERCO
U.V.
Technologies, Inc.
U.V.
Technologies, Inc.
(Now ENERCO)
U.V. Purification
Systems, Inc.
U.V. Purification
Systems, Inc.
U.V. Purification
Systems, Inc.
U.V. Purification
Systems, Inc.
Aqua-fine
U.V. Purification
Systems, Inc.
Comments on Performance
Too soon to tell.
Good performance. Meeting coliform kill
requirements. Coliform count is less than
1 MPN/100 ml. Some minor mechanical
problems, which have been corrected.
Operating since June 1 983.
Poor operation. Have not achieved contract
specification for operation. Still putting in
corrective measures. Mechanical
difficulties; electrical components burned.
Excess heat in UV Bldg. - uncomfortably
warm for operator. Fans alleviated this
problem.
Plant is meeting discharge requirements. In
operation 2-3 years.
Summer use only; in use for 1 yr. Some
ballast problems. Some chemical cleaning
problems.
In operation 1 year. Initial problem with
ultrasonic cleaning system achieving
.desired bacterial kill.
Just starting up (as of 10 am 3/27/84) Using
clear water.
Good performance.
Good performance. Fecal coliform @ 10-50
MPN/100 ml. (Permit 200 MPN per 100 ml)
Feed to UV Is "clean" 4-5 ma/1 SS, 4-S mg/l
SS. Initially seals leaked • now replaced,
Once clay got in stream, colored water, and
reduce effectiveness of UV system, 2-750
gpm units run intermittently batch.
Operational since 9/81. Faulty photo cell.
replaced in 1981.
Not yet in operation — summer
requirements only (Plant on-line since
10/83).
In operation since 1 1/83. High flows blew
out UV Tubes; cause uncertain: freezing or
obstacles in flow.
Not yet in operation. Awaiting stabilization
of activated sludge system.
Poor performance. Coliform count is high.
Difficult to keep quartz sleeve over lamps
clean. Harness (mineral deposits won't
wipe off. Sometimes operators use soap or
chlorox. However, it'sfrequently necessary
to chemically clean lamp sleeves.
Maintenance problems i.e. burned out
lamps.
171
-------
Table 7-4 (Continued)
Facility Nima C«sign Firm
Berkeley County
(South Carolina)
Civilian
Conservation
Center, Nemo
Oanvilto
(Vermont)
Jacksonville (VD
Pawlot (VD
Whitlnflhamn
(VT)
Cumberland
Hospital
Tangier Island (VA)
Lamtor (Wyoming)
Rock Springs (WY)
Brooklyn
(Wisconsin)
Cross Plains
(W1J
Eltrick
(Wl)
Hotomon
(Wl)
Oeertield
(Wl)
Lodi
(Wl)
Lyons
(Wl)
Poynetta
(Wl)
Spring Valley
(Wl)
EM. Seabrook, Inc.
Ryan Wright
(803) 1184-4496
Case, Colter, Inc.
(Denv.jr)
Ralph Olson
(3031:183-1511
Other contacts:
U.S.F.S:
Carl E/ikson
(605) 348-3636
Terry Ambraster
(303) :'34-5223
Dufresne & Henry
Bobbi Trudell
(802) 886-2261
Dufresne & Henry
Bobbi Trudell
(802) £85-2261
Dufresne & Henry
Dufresne & Henry
Greshnm, Smith &
Partners
David Shood
(803) 572-1300
Shore Engineering
Emme'rt Renson
(804) 787-2773
Western Design
Consultants
Mr. Chen
(801) 486-5621
Johnson, permelia, and
Crank
Dale Crank
(307) 877-9093
Carl C. Cranca, Inc,
Dennis Truttman
(600) 238-4761
Mead IJt Hunt, Inc.
Bill Bulh
(608) 233-9706
Davy Engineering
Mike Davy or
Arnie F'inski
(608) 7132-3130
Davy Engineering
Mike Davy
(60S) 782-3130
Carl C. Crane, Inc.
Dennis Truttman
(608) 2:38 4761
Mid Stnte Assoc.
Jim Owen
(608) 3B6-8344
Robbers & Boyd
Larry Boyd
(414) 713-2652
Lakeland Engineers
Mark Koletzke
(608) 274-3898
Davy Engineering
Arnie Pinski or
Mike Davy
(608) 782-3130 '
Size, mgd
Design
5.0 (peak 8.0)
0.024
0.070
0.050
0.040
0.013
0.030
0.1
1.82
2.0
0.160
0.450 (avg.)
0.064
0.8
0.195
0.620
0.100
0.190
(0.470
wet weather)
0.189
Current
3.0
0.017
0.040
0.025
0.023
0.006
0.002 to
0.003
0.032
1.3
2.0
0.025-0.030
(low)
0.180 to
0.200
0.040
0.4
0.100
flowmeter
not working.
0.230 to
0,288
0.020 (low)
0.300
0.100
Other Treatment
Processes
Oxidation ditch
Extended aeration,
tertiary filtration.
Aerated lagoon system
(30-day del. time)
RBC system, which
treats septic tank
effluent
RBC system; which
treats septic tank
effluent.
RBC system, which
treats septic tank
effluent.
Extended aeration
RBC
Aerated lagoon
Oxidation ditch
Oxidation ditch
Extended aeration
oxidation ditch
RBC
Extended aeration
Extended aeration
oxidation ditch
RBC
Oxidation ditch
Oxidation ditch
RBC
Equipment
Pure Water
Systems, Inc.
U.V. Purification
. Systems, Inc.
U.V. Purification
Systems, Inc.
Ultra Dynamics,
Inc.
U.V. Purification
Systems, Inc.
Ultra Dynamics,
Inc.
U.V. Purification
Systems, Inc.
Aquafine
ENERCQ
UV Technologies
U.V. Purification
Systems. Inc,
U.V. Purification
Systems, Inc.
U.V. Purification
Systems, Inc.
U.V. Purification
Systems, Inc.
U.V, Purification
Systems, Inc.
U.V. Purification
Systems, Inc.
Aqua fine Corp.
U.V. Technologies
(now ENERCO)
U.V. Purification
Systems, Inc.
Comments on Performance
In operation for past 9 months. Good
performance.
Discharge requirements met. Some
mechanical difficulties.
Good performance.
Good performance. Meeting discharge
requirements.
Good performance. Meeting discharge
requirements.
Good performance. Meeting discharge
requirements. Operating since 12/83.
Good performance. Meeting discharge
requirements. Operating since 1 2/83.
Discharge requirements not met in 6 of 9
months. In compliance for 2/84.
Difficulty meeting dicharge requirements.
Burning one end of the UV lamps due to
float switch problem.
Test runs of system resulted in overheating
of bulbs. Fans were used to correct this.
150-200 fecal coliform count, (Currently
there is no state requirement.)
Good performance. Requirements met.
Problems-w/turbidity & coli killo.
Problems-start-up only
Operational since 10/83.
In start-up for 2 months. Still in
shake-down.
Operational since 7/83. Poor performance.
Requirements not met.
Meeting discharge requirements.
Operational since late 1981.
Operational since 8/83. Meeting coliform
count requirements.
UV is not to start-up until 4.1.84
172
-------
Table 7-4. (continued)
Facility Name Design Firm
Otay Lowry & Assoc.
(California) Matt Tebbetts
(619) 283-7145
Size, mgd
Design Current
0.65 0.3
Other Treatment
Processes * Equipment
Activated sludge, UV Technology
filtration, pH reduction
prior to UV, the R.O.
unit.
Comments on Performance
Jehovah Witness Radmes Torres
Church, Tujuas (809) 725-5878
(Puerto Rico) (in San Juan, P.R.)
0.22
0.006
Septic tank, sand filter. Unknown
equalization tank.
Table 7-5. Summary of UV Installations in U.S. in Operation, Construct, or Design Phase
Size (Design) In Operation In Construction
In Design
<380mVd«0.1 mgd)
380-1900 (0.1-0.5)
1900-3800 (0.5-1.0)
3800-19000 (1-5)
19000-38000 (5-10)
3800-190000(10-50)
>190000 (>50)
15
17
7
11
-
1
-
-
10
5
14
-
-"
1
7
10
4
11
-
2
.
53
30
34
Note: List compiled Spring of 1984.
phosphor for application to the walls of the tube (to
convert UV light to visible light) and efficient long-
lived electrodes was accomplished in the 1930s. By
the 1940s the fluorescent lamp was a commercial
reality. Although there was no significant demand for
the UV lamp ("germicidal lamps") per se, the suc-
cessful commercial development of the fluorescent
lamp technology resulted in the immediate availability
of a relatively inexpensive, efficient, UV source.
7.2.2 Mechanism of UV Disinfection
One of the earliest reports relating to the germicidal
effects of UV was by Downes and Blount (9). They
described the lethal effects of solar radiation on a
mixed microbial population and assigned the cause of
these effects to shortwave UV radiation.
The early interest in the application of UV for
disinfection centered first on potable water. The
equipment was not reliable, however, and the lamps
were highly inefficient, as discussed earlier. Chlorine
was becoming readily available by the early 1900s
and was inexpensive. Chlorine also exhibited the very
real benefit for potable water applications; this was
the ability to maintain a residual. Interest in the
application of UV subsequently faded, but not in the
effects of UV or its mechanism. Research continued,
at first centering on the effects of UV on different
organisms and the optimum conditions for germicidal
effectiveness. The more recent research, conducted
primarily since the early fifties, was directed to the
actual mechanisms by which radiant energy affects
an organism.
These research efforts have been well documented in
the literature and it is not the intent of this discussion
to give a detailed accounting. Rather a brief descrip-
tion of the basic mechanisms is provided. The reader
should refer elsewhere for greater detail (10-17).
The basic premise to understand is that radiation
must be absorbed before it can have an effect. Visible
light is absorbed by molecules called pigments; color
is observed by reflectance or transmittance. Radia-
tion outside the visible spectrum can also be ab-
sorbed. Proteins and nucleic acids are basically
colorless, but strongly absorb invisible shortwave UV
light.
Recall from the earlier discussion of quantum energy
that it is constant for a given wavelength, and will
change as a function of the wavelength. The longer
the wavelength the lower the energy (see Equation
7-2); conversely, the shorter the wavelength, the
higher the energy. The effect of a quantum when it
interacts with matter is a function of its energy
content. Referring to the electromagnetic spectrum
on Figure 7-9, infrared radiation at wavelengths
greater than 1200 nm has relatively little energy and
is unable to effect any chemical change. The energy is
immediately converted to heat (hence the infrared
heat lamps). At wavelengths from 1200 nm (near
infrared) to about 200 nm (far UV) the energy content
is sufficient to produce photochemical changes.
Radiations with wavelengths less than 200 nm
(extreme UV, X-rays, gamma rays, and cosmic rays)
have energy contents so high that molecules in their
path become ionized.
773
-------
Figure 7-9. Electromagnetic spectrum.
Electromagnetic Spectrum
High
-Quantum Energy-
•Low
icr
i ' rr i ' i ' i ' i ' i ' i
I0'1 101 | 103 105 107 10" 1011 1013
Wavelength '(Nanometers)
i—X-Rays-
Gamma.
ible
-Hertzian Rays-
Radio
I- Waves"
100 200 300 400 500 600 700 800 900 1,000
102
Expanded-Arithmetic Scale
—Ultraviolet 1 Visible Light
X-Raysn
Violet Green Orange
Blue Yellow Red
Living organisms can use parts of solar radiation
advantageously. The obvious examples are photo-
synthesis, phototaxis, and vision. The lethal effects
are related primarily to the photochemical changes
induced by molecular absorption of radiation. Cellular
proteins and nucleic acids are strongly absorptive of
far UV radiation; the photochemical changes caused
by this absorption are very injurious to living cells,
hence the bactericidal properties of UV. The most
effective spectral region lies around 260 nm, which is
the region of maximal absorption by nucleic acids.
Cell death following UV radiation is almost entirely
attributable to the photochemical damage of these
compounds.
7.2.2.1 Photochemical Damage of the DNA
Molecule
Deoxyribonucleic acid (DNA) and ribonucleic acid
(RNA) are chain-like macromolecules that function in
the storage and transfer of a cell's genetic informa-
tion. These compounds generally comprise 5 to 15
percent of a cell's dry weight, and effectively define
the operations of a cell, particularly the type and
quantity of enzyme production. The DNA molecule is
considered to be the principal target of UV photons,
and the primary component where significant bio-
logical effect, or damage, is incurred.
The monomeric units of the DNA (and RNA) are
nucleotides. These all have three characteristic
components: each has a nitrogenous heterocyclic
base which can be either a purine or pyrimidine
derivative; each contains a pentose sugar; and each
has a molecule of phosphoric acid. There are four
different deoxyribonucleotides which comprise the
major components of DNA, differing only in their base
components. Two are the purine derivatives adenine
and guanine; the other two are pyrimidine derivatives
cytosine and thymine. Similarly, four different ribo-
nucleotides comprise the major components of RNA.
As with the DNA, they contain the purine bases
guanine and adenine; the pyrimidine bases are
cytosine and uracil. Thus, thymine is characterist-
ically present only in DNA, while uracil is normally
present only in RNA.
As had been mentioned earlier, the most effective
spectral region for germicidal activity lies about the
260 nm wavelength. This is demonstrated on Figure
7-10 which presents relative germicidal effectiveness
as a function of wavelength (17). The action spectrum
of nucleic acids is very similar to this, as shown by
Figure 7-11. On a relative scale, the extinction
coefficients (a measure of the inhibiting effect on
bacterial colony formation) are plotted as a function of
wavelength. Maximal effect is shown to occur
between the wavelengths of 250 nm and 265 nrn.
Overlaying this is the relative percent absorption for a
solution of RNA. The similarities are striking, sup-
porting the premise that the lethal effects of UV
radiation are induced by the photochemical damage
to the cell's nucleic acids.
Figure 7-10. Relative germicidal effectiveness as a function
of wavelength (17).
g 100
3
1
"5 80
I
ID
Q.
CO
U)
CO
60
20
£
to
•TS
a
£
o5 2000
O
\ I I
2540 A
Emitted
from Low
Pressure
Mercury Lamps
2500
Wavelength (A)
3000
The photochemicalchanges induced by UV radiation
on the DNA of an organism have been thoroughly
studied. Although several mechanisms exist, the
most dominant is the dimerization of two pyrimidine
molecules. To visualize this effect, consider the
174
-------
Figure 7-11. Relative abiotic effect of UV on E. coll
compared to relative absorption of Ribose
nucleic acid (16).
Figure 7-12. Exampleof DNAandUVdamagetoDNA(18).
5!
I
i
°x
<0
"o
0)
a.
100
90
80
70
60
50
40
30
20
10
Relative Percent
Absorption
for 0.5 v, 100g/l
(Absorption
at 2550 A = 26%)
Extinction
Coefficients
I
2300 2400 2500 2600 2700 2800 2900
Wavelength (A)
schematic representation of the DNA molecule on
Figure 7-12. Recall that the DNA is a long polymer
comprised of a double helix chain of simple mono-
meric units called nucleotides. The order of these
nucleotides constitutes the genetic information of the
cell. These are represented on the Figure by the
letters A (adenine), G (guanine), C (cytosine) and T
(thymine).
In the two strands, G is always opposite C and T is
opposite A; if damage occurs in one strand the
information still remains in the second strand. Thus,
to repair the damage, a C is inserted opposite a G and
a T opposite an A, and so on. As long as the
information is retained on one strand, the second
strand damage can be rebuilt. These are enzymatic
processes. Before cell division occurs, a duplicate of
the DNA is prepared by building a complementary
strand to each of the parental strands.
The UV induced dimer between two adjacent pyrim-
idines in a polynucleotide strand has been demon-
strated for all combinations of the pyrimidines
(thymine, cytosine, and uracil). The thymine dimer is
formed with the greatest efficiency, however. This is
shown on Figure 7-12. There are two adjacent
Hypothetical
DNA Double
Strand
i r i n i i i i \
A C G T A A C AC
T G C ATT G T G
I i i i i i i i i
Replicating DNA
1 •
A
A
T
i
A
T
i
C
G
i
A
T
C
G
1
Dimerization
ofThymine
Nucleotides
A
T
C
G
C
T A A
A T= T
C
G
A
T
G
i
thymine monomers on one of the strands; during
exposure to UV light new bonds are formed between
the two such that a double thymine molecule, or
dimer, is formed. Formation of many dimers along a
DNA strand makes replication very difficult.
No comparable interactions of the purines have been
demonstrated. The effect of the pyrimidine dimeriza-
tion is a blocking of normal replication. Total and
permanent inhibition of DNA replication would in
itself be a lethal event.
Alternatively, replication may bypass such_a distor-
tion, producing an error in the copy and a subsequent
mutant daughter cell which is unable to replicate.
7.2.2.2 Recovery from Photochemical Damage
Just as a cell can be lethally affected by photo-
chemical damage, there is a widespread prevalence
in the world of living organisms to repair and reverse
the lethal effects of UV. The mechanism is typically a
photoenzymatic repair, requiring longer wavelength
light in the near UV and visible spectrum. This
phenomenon, uniquetoUV, has been broadly termed
photoreactivation. Jagger and Stafford (19) suggested
an explicit definition: "the reduction in response to
far-ultraviolet irradiation of a biological system
resulting from concomitant or post-treatment with
non-ionizing radiation."
Although not explicitly characterized as such, photo-
reactivation effects were noted during the first half of
this century. These observations related to the
counteracting effects of visible and UV light. The
discovery of the phenomenon is generally attributed
775
-------
to Kelner (20) and Dulbecco (21), working independ-
ently In the late forties. The reader is referred to Harm
et. al. (22) and Harm (23) for detailed discussions and
reviews of research of this phenomenon.
The repair mechanism is not universal and there is no
clearly defined delineation of characteristics which
suggest which species would have the ability to repair
and which would not. Organisms which have been
shown not to have the repair mechanism include
Haemophilus influenzas, Diplococcus pneumonias.
Bacillus subtilis, and Micrococcus radiodurans. Vi-
ruses generally do not have the repair ability except
when in a host cell which can repair. Organisms
shown to photorepair include Streptomyces, Escher-
ichia co//', Saccharomyces, Aerobactor, Micrococcus,
Erwinia, Proteus, Penicillium, and Nuerospora.
The catalyzing, non-ionizing radiation wavelength is
not the same for all. It generally falls between 310 nm
and 490 nm. In some cases photorepair has been
induced by radiations between 230 nm and 240 nm
(although these wavelengths are absorbed in the
atmospheric ozone layer and would not naturally
occur at the earth's surface). It is important to note
thatphotoreactivating light is present in sunlight and,
as such, is universally available. The effects are quick,
occuring within minutes after exposure to the neces-
sary reactivating light.
Observation of the effect has been accomplished by
comparing, after UV radiation, the "dark survival" of
cells with "photoreactivated survival."Quantitatively,
this is described as the dose decrement. Referring to
Figure 7-13, consider a dark survival curve as a
function of UV dose. After a UV dose Dl( the culture is
exposed to photoreactivating light and the survival
increased to the level marked by PR; the resulting
survival can be considered equivalent to the dark
survival accomplished by dose Di'. The difference,
Di-Di', called the dose decrement, can be used as a
measure for the extent of photoreactivation. Since
Figure 7-13. Schematii; representation of the effects of
photoreactivation (23).
1.0
-ID"1
1
to
10-*
1CT
PR
1.0
D,' D,
UV Dose
10'
the number of lethal photoproducts (thymine dimers)
is directly proportional to the UV dose, the dose
decrement can be considered a measure of the
number of repair events.
In the case of maximal photoreactivation, the dark
and PR survival curves differ by a constant displace-
ment factor. The two curves would coincide if the
dose scale for one of the curves was changed by an
appropriate factor. This was described as the "prin-
ciple of constant dose reduction" and the displace-
ment factor was called the dose reduction factor (24):
Di'/Di
this suggests the repair of a constant fraction of
lesions. This fraction,
1 - (Di'/Di)
is called the phbtoreactivable sector, or PRSmax.
There are several mechanisms by which these repairs
can be made. Rupert (67) established that the most
dominant repair mechanism was by a photoreac-
tivating enzyme (PRE). It is similar to other cellular
enzymes, except that it requires light energy to
initiate its activity. The reaction scheme suggested for
this enzymatic repair can be expressed as:
E + S
ES
E + P
(7-3)
This is a conventional Michaelis-Menton expression
for enzymatic reactions, except that the rate ka is
absolutely dependent on light energy. E is the
photoreactivating enzyme, S is the substrate (the
photorepairable lesion), ES is the enzyme-substrate
complex, and P is the repaired UV lesion. The enzyme
binds the pyrimidine dimers, and upon exposure to
the appropriate light energy monomerizes the dimers.
The complex formation and dissociation can occur in
the dark; the value of k3, however, is zero in the dark.
A second mechanism has been demonstrated to
occur without the light requirement, called dark
repair. It is a multi-enzymatic mechanism (termed
excision repair) in which the dimer is recognized by an
enzyme; this enzyme nicks the dimer from the DMA
strand on one side. An exonuclease then releases the
dimer completely from the DMA strand and a replica-
ting DNA enzyme then repairs the gap.
Environmental conditions which tend to inhibit active
cell metabolism and cell division for a time after
exposure will tend to decrease the effects of UV
radiation. These inhibitory effects allow time for the
cell to repair its DNA before it is erronebusly (and
lethally) replicated. Such conditions include low
775
-------
temperature or low nutrient levels. Conversely, the
repair mechanism will be attenuated by conditions
which encourage high growth rates. A population in
its lag or stationary growth stage will have a greater
chance of recovery because, by definition, it is not
replicating its DMA as quickly as an exponentially
growing population. The recovery of irradiated phage
is also dependent on the physiological condition of its
host cell.
\
Photoreactivation is a phenomenon which can impact
the performance and design of a UV system in certain
situations. The conditions which exist in a treated
effluent are conducive to the occurrence of photo-
reactivation; the nutrient levels are low and the
population of organisms would likely be in the
stationary growth stage at the point of the disinfection
process. There are several variables involved in
predicting the recovery effect in systems such as a
wastewater treatment plant. Certainly sunlight, the
source of the photoreactivating light, will differ in
intensity and spectral distribution according to the
season, time of day, and cloud cover. Effluent
characteristics will affect the penetration of the
photoreactivating wavelengths; this will in fact
extend to the receiving water conditions. Shallow,
clear receiving streams will be more conducive to
repair than discharge to deeper, slow-moving, and
turbid receiving waters.
Given the environmental factors which influence the
degree and effect of photoreactivation, it is likely far
more practicable to control (i.e., account for) the
mechanism by increasing the applied UV dose.
Manipulating the growth stage of the microorganisms
would not be a practical operation in typical waste-
water treatment operations. By designing for an
increased UV dose, however, the effect of photo-
reactivation can be concurrently reduced (it cannot be
eliminated). This was demonstrated by the discussion
of Figure 7-13.
7.2.3 Recent Application of UV to Wastewater
Disinfection
The following discussions present a summary of the
major studies which have been conducted to date
with regard to the application of UV for wastewater
disinfection. Several of these studies will be ref-
erenced again in subsequent sections. The intent at
this point is to present an overview of the evolution of
the technology and to identify those efforts which
were and are important to the current state-of-the-
art. The review is limited primarily to direct applica-
tions to wastewater disinfection and, as will be noted,
to work reported after 1970. Before this, three studies
are presented which, although not directed to waste-
water treatment, were important in their evaluation
of the applicability of UV on a large scale, its
effectiveness, and the factors which would influence
its design.
Kelly (25) reported on a study which evaluated the
ability of UV to disinfect seawater used for the
depuration of oysters. Tests had shown that the
activity of the oysters was adversely affected when
the water was treated by chlorination/dechlorination.
Two different designs were set up to apply the UV;
one was at Pensacola, Florida and the other was in
Purdy, Washington. Both were tray designs in which
the lamps were suspended over a shallow tray which
received continuously flowing water. The Pensacola
unit was the deeper of the two—6.4 cm (2.5 in); the
slow moving system tended to accumulate particu-
lates. The lamps and reflectors were cleaned on a
daily basis and the troughs were flushed regularly.
The unit at Purdy was found to be a far more efficient
system operationally, most likely because of the
higher velocities, the thinner film thickness of the
water, the greater agitation, and the improved flow
distribution. The liquid level was kept at 1.9 cm (0.75
in) by the placement of a downstream weir. A
perforated pipe provided for equal lateral distribution
at the inlet to the unit. Six internal baffles were then
installed to provide a rolling motion to the liquid as it
traveled down the length of the unit. Forced draft
ventilation of the unit was also provided in an attempt
to optimize the operating temperature of the lamps.
The studies demonstrated the effectiveness of the
systems in reducing the coliform levels by greater
than three logs from an initial density between 1,000
and 10,000 MPN/100 ml and at turbidities up to 20
JTU. Kelly stated that a dose of 57,600 /uWatt-
sec/cm2 was required to accomplish this. The
intensity at the surface of the liquid was computed by
the author by simply distributing the rated output of
the lamps (unreflected) across the surface of the
water. The pilot scale studies suggested that, with
proper maintenence, the unit could operate with a
high degree of dependability. The maintenance
requirements were of a housekeeping nature; flush-
ing the units on a periodic basis and cleaning the
lamps and reflectors of accumulated debris.
Huff et al. (26) demonstrated that ultraviolet disinfec-
tion would be effective in treating ship-board potable
water supplies. Doses varied between 4,000 and
11,000 /uWatt-sec/cm2; the intensity was measured
by an intensity meter on the side of the UV unit.
Effective kill of E. coli, A. aerogenes, and S. faecalis
was accomplished. The study also studied the atten-
uating effects of turbidity, color, and iron on UV
intensity and consequent effects on the performance
of the UV unit. The apparatus was shown to effectively
inactivate certain enteric viruses when operated at
the recommended intensity and flow rates.
Hill et al. (27) followed up, in a sense, on the work of
Kelly and others in investigating the application of UV
for the disinfection of seawater. Their efforts were
177
-------
directed to the virucidal efficiency of the process,
which had become the treatment of choice for
disinfecting seawater that is to be used for shellfish
depuration systems. The Kelly-Purdy UV Seawater
Treatment Unit was used for the study. Static tests
first demonstrated the effectiveness and rate of
inactivation for eight enteric viruses. The exposure
required to obtain effective disinfection (99.9 percent
reduction) at an applied intensity of 1160 /M/att/cm2
was as follows:
Poliovirus 1
Poliovirus 2
Poliovirus 3
Echovirus 1
Echovirus 11
Coxsackievirus A-9
Coxsackievirus B-'l
Reovirus 1
28 seconds
31
27
28
31
31
40
40
These static bioassays were conducted in shallow,
unstirred petri dishes. The devitalization rate de-
termined from dynamic tests with Poliovirus 1 were
found to be significantly different than for the same
virus under the conditions of the static test. The rate
was, in fact, significantly increased in the flowing
seawater system. This difference was attributed
primarily to UV dose and the mixing effects provided
by the unit. In all, the study concluded that the
continuously flowing UV units would be highly
effective for the inactivation of viruses in contam-
inated seawaters.
At approximately the same time, a federally spon-
sored study was investigating the application of UV to
wastewaters which were being discharged to shell-
fishing waters in St. Michaels, Maryland. Roeber and
Hoot (28) used a system similar in design to the Kelly-
Purdy shallow tray unit to disinfect the effluent from
an activated sludge plant.
First order reductions were observed in total coliforms
and bacteriophage densities, with a tailing effect after
99.99 percent kill of the total coliform. The average
dose required to reduce coliform densities to 70
MPN/100 ml or less was estimated by the investi-
gators to be 25,000 yuWatt-sec/cm2. The intensity
was calculated on the basis of measured estimates of
the absorbance coefficient of the wastewater at
253.7 nm and direct measures of the intensity in the
liquid. The average intensity was then estimated by
integration of the Beer-Lambert equation over the
fluid depth:
Average Intensity = l0[(1-e~Kd)/kd] (7-4)
where:
I0 = initial intensity (yuW/cm2)
d = depth of the fluid (cm)
k = the absorption coefficient (cm 1)
This average intensity would be multiplied by the
average detention of the system in order to estimate
the applied UV dose. The ultraviolet transmittance of
the liquid averaged approximately 65 percent; its
value was related more to the organic makeup of the
water (as COD) than the turbidity. The study indicated
a dependence on the initial coliform density and
suggested that higher turbidity levels would affect the
unit's performance.
Roeber and Hoot also presented the results of a series
of photoreactivation tests. Coliforms and bacterio-
phage were shown to exhibit significant repair upon
exposure to sunlight for one hour. In all, the report
concluded that ultraviolet disinfection would be
practicable for application to well-controlled activated
sludge plant effluents. They suggested that at opti-
mum conditions, an energy consumption of 0.092
kWh/m3 would be required to accomplish a coliform
level less than 70 MPN/ 100 ml.
Singer and Nash (29) reported on a study which
evaluated UV disinfection of a secondary step aeration
plant effluent. The UV system was one normally used
in potable water applications and consisted of a
closed vessel containing nine germicidal lamps. The
lamps were sheathed in quartz sleeves and were
continually submerged. Flow would enter the rec-
tangular reactor at one end perpendicular to the
lamps, turn and flow down the length of the reactor
and exit through a pipe located on the same side as
the inlet pipe. This was one of the first repbrted
applications of the submerged quartz configuration to
wastewater disinfection. A water quality meter was
attached which measured the intensity at a single
point in the reactor. This was found to inversely
correlate with the turbidity, suspended solids, and
BOD of the wastewater.
Singer and Nash reported a tailing of the first order
reduction with increasing dose, such that a base level
of coliforms would be present in the waste. This was
attributed to a degree of short-circuiting and the
suspended solids in the wastewater. Effluent levels
less than 200 MPN/100 ml could be consistently
achieved as long as suspended solids levels were kept
below approximately 22 mg/l. The report concluded
that UV would adequately and cost-effectively dis-
infect a step aeration effluent and recommended
further work to investigate closer spaced lamps to
counteract the poor quality of the effluent and
suggested that greater attention be paid to the
hydraulic characteristics of the reactor. The quartz
surfaces were found to require a periodic chemical
cleaning, although the appropriate frequency was not
determined.
178
-------
Oliver and Carey (30,31) investigated methods to
apply UV to conventional wastewater treatment
systems. I n a sense, the scheme was a modification of
the Kelly-Purdy design; lamp units (double lamp units
with reflector and ballast) would be suspended above
secondary clarifiers, close to the clarifier overflow
weirs. The effluent would be exposed to about the
same dosage as it rises to the surface and passes over
the weir in a thin film. Earlier laboratory studies
indicated that relatively low dose levels were required
to achieve a 2-Iog reduction (Log survival = -2) in total
and fecal coliforms and fecal streptococcus (32). They
also determined that the bacterial inactivation was
independent of light intensity; thus, a relatively low
intensity arrangement could be implemented as long
as sufficient exposure time were provided to achieve
the desired dose. A significant result of these tests
was the demonstration of the bacterial occlusion by
particulates in the water. When sonication was
applied as a pretreatment before UV exposure,
greater inactivation efficiencies were accomplished.
The ultrasonics appeared to disperse the particulate
aggregates, making the bacteria contained in the
particles more susceptible to the UV radiation.
Pilot studies of the overflow weir arrangement were
conducted and provided excellent results. The inves-
tigators also noted that the lamps did not foul and
maintained a constant output over a six-week period.
The absorbance of the effluent ranged between 0.12
and 0.25 a.u./cm with little direct effect on the
system efficiency. A dose of 130 Watt-sec/imperial
gallon was reported to achieve a reduction of 2 logs.
This is relatively inefficient, reflecting the ineffective
use of the lamps' output by having them suspended
above the surface of the wastewater. Greater energy
efficiency would obviously be obtained by submersing
the source in the liquid. Nevertheless, the report
concluded that UV would be highly effective and
would be competitive with the use of chlorine.
A limited study conducted in Syracuse, New York,
evaluated the use of UV for the disinfection of waters
from combined sewer overflows (CSO) (33). One liter
samples were irradiated in a bell-jar vessel in which
the walls were equidistant—5.7 cm (2.25 in)—from
the UV lamp. The intensity at the lamp surface was
computed to be 5,800 /uWatts/cm2. The study sug-
gested that a dose of 500,000 /uWatt-sec/cm2 would
be required to achieve a residual coliform level of
2500 MPN/100 ml and concluded that UV disinfec-
tion of CSO waters was feasible and particularly
attractive because of the absence of a residual. The
scheme studied by Oliver and Carey (30), which was
suggested for the CSO application (the UV lamps
would be suspended over high rate swirl separators)
would be impractical, however, because of the high
dose requirements.
Petrasek, et al. (34), reported on a study conducted in
Dallas, Texas, which investigated the feasibility of
ultraviolet disinfection of treated municipal waste-
water effluents to achieve fecal coliform levels less
than 200 MPN/100 ml. Two system configurations
were evaluated: the first system was the Kelly-Purdy
shallow tray design; and the second system was a
closed vessel design with the lamps (enclosed in
quartz sleeves) submerged in a 53.6-liter (14.1-gal)
chamber. The flow was parallel to the lamps.
The investigators concluded that the more appropriate
system configuration for wastewater disinfection
was the submerged quartz system. The Kelly-Purdy
design, although effective in disinfection, was less
efficient in the utilization of UV energy, would require
greater space, and was susceptible to solids deposi-
tion in the unit. Hydraulic studies indicated that the
unit operated poorly at low flows and at deeper liquid
depths. Actual retention times approached theoretical
when the unit was operated at shallow depths—2.5
cm (1 in)—and higher velocities. Relatively high
dispersion was still observed under these conditions.
The hydraulic analyses of the submerged system also
indipated relatively poor flow characteristics when
compared to the ideal condition of plug flow. The time
distribution curves constructed for a number of flows
indicated a high degree of dispersion. The investi-
gators also demonstrated the importance of the
absorbance coefficient of the wastewater. This was
spectrophotometrically measured at 253.7 nm and
effectively quantified the UV "demand" of the liquid.
The parameter was not found to be affected signif-
icantly by the suspended solids or turbidity of the
water.
The authors recognized the lack of any direct meas-
urement capability for estimating the UV intensity
within a complex multi-lamp system, such as the
submerged quartz unit. An approach suggested by
the report was to mathematically calculate the
intensity at any point in the reactor; they assumed
that the radiation emits perpendicularly from the
lamp and that the lamp is an infinite line source. The
average intensity was then calculated on an areally-
normalized basis. The technique accounted for the
attenuation of intensity due to the absorptive char-
acteristics of the liquid. No attempt was made to
account for the deterioration of the lamp output or the
quartz surfaces. The estimated intensity was found to
correlate well with the total coliform reduction of the
unit at a constant flow to the unit.
The authors found that UV was effective in the
inactivation of Type 1 poliovirus and F2 coliphage.
The coliphage was suggested as a good indicator for
the inactivation of poliovirus. Photoreactivation was
775
-------
also investigated; static tests indicated that sub-
sequent exposure of UV irradiated conforms to
sunlight for 30 minutes induced a 1.1 log increase in
total coliforms, and a 0.6 log increase in fecal
coliforms.
Scheible and Bassell (35,36) reported the results of a
full-scale prototype demonstration study conducted
at the Northwest Bergen County Water Pollution
Control Plant, Waldwick, New Jersey. The unit,
similar to that shown on Figure 7-5, was a submerged
quartz system comprised of 400 lamps. These were
arranged in a symmetrical array, axially parallel to
one another, and equidistant on horizontal and
vertical centerlines. The flow path was perpendicular
to the lamps. The spacing between quartz surfaces
was only 1.25 cm (0.5 in), imposing a thin film of
liquid as the wastewater passed through the lamp
battery. The lamp battery was inserted into two
bulkhead walls constructed in an existing chlorine
contact chamber. The arrangement simulated an
open channel, with the lamp battery inserted into the
' channel and the inlet and outlet planes of the lamp
battery wholly exposed to the wastewater.
The time-distribution characteristics of the unit were
not measured directly, although it was established
that a relatively uniform velocity field existed across
the exit plane of the lamp battery. The time of
exposure was assumed to be the theoretical detention
time of the unit, which ranged between one and four
seconds under normal operating conditions. In order
to estimate the dose under a given set of sampling
conditions, the authors incorporated the use of the
radial light model, as had been desribed by Petrasek,
et al. (34), to calculate the intensity. In this case,
however, the authors calculated the incident intensity
at the surface of the lamp (dividing the rated UV
output by the quartz surface area) and accounted for
some deterioration in UV output with time. The lamps
were presumed to be transparent to UV from a
neighboring lamp, which is not the case, and likely
resulted in an overestimate of the applied dose. The
intensity calculations were demonstrated for any
given symmetrical system and presented as a func-
tion of spacing, UV absorbance coefficient, and lamp
rating. Dose was estimated by multiplying the inten-
sity (which varied as a function of the UV absorbance
coefficient and the output of the lamps) by the
theoretical detention time.
Empirical regressions were developed to describe the
performance of the system as a function of the
applied dose. These reflected more an attempt to
linearize the correlation of the coliform reduction
with the dose; this was effective in determining the
system performance and design requirementsforthe
specific plant application. The UV absorbance coef-
ficient was suggested to be the key wastewater
parameter for the design, control, and monitoring of
the UV disinfection process. It correlated well in this
case with the wastewater COD.
An extensive series of static and dynamic photo--
reactivation tests were conducted during the study.
The static-bottle technique, in which a light and dark
bottle are held in sunlight for a period of one hour,
was suggested as an effective and practical procedure
for evaluating and/or monitoring the effects of
photoreactivation. The results of the tests indicated a
degree of temperature dependency, although this
may have implicitly included the influences of re-
duced sunlight intensity during the winter months. At
10°C (50°F), a 0.3 log increase in coliform density
was observed, while at 20°C (68°F) the increase was;
approximately 1.0 log. A detailed economic evalua-
tion showed the cost of the process to be approx-
imately $0.008/m3($0.012/1,000 gal) for secondary
treatment plants (1979$). When compared to alter-
native processes, UV was found to be less costly than
ozonation, more expensive than chlorination, and
competitive with chlorination/dechlorination.
Severin (37) reported on the application of a com-
mercially available UV system to the disinfection of a
variety of treated municipal effluents. In determining
dose, he estimated the average intensity by the
integrated solution of Beer's law over a fluid depth, as
presented by Roeber and Hoot (Equation 7-4) and by
Petrasek et al. The UV unit was a closed flow-through
vessel containing 10 lamps set longitudinally; flow
was directed parallel to the lamps. A nominal average
depth of liquid was computed for the reactor to use in
the computation of the average intensity. Severin
assumed a single source in the estimate of intensity
and did not account for the additive affects of a multi-
lamp system.
Experiments were conducted at a number of waste-
water treatment plants which provided different
levels of treatment. The quality of the wastewatersi
was very good in most cases, except where the
effluents were artificially adjusted to yield higher
absorbance values. Typically, the effluents had an
absorbance coefficient between 0.2 and 0.3 cm"1
(base e). The inactivation results of the overall
experimental effort were described by a disinfection
model which was linear with time to the one-third
power. Time was assumed to be the theoretical
detention time of the unit. The least squares regres-
sion yielded the expression:
Logio(100N/N0) = -1.73 (Pavg/Po) (t1/3) + 2.598
(7-5)
The expression was explicitly described as an empir-
ical relationship applicable only to similar reactors;
under the same water quality conditions. Pavg/Po is'
180
-------
the ratio of the average intensity to the incident
intensity.
Hydraulictraceranalyses of the system indicated that
the unit did not behave in a plug flow fashion but
exhibited a relatively high degree of dispersion. The
author indicated that this deviation from ideal plug
flow was not enough to account for the empirical
observation that disinfection is a f untion of time to the
one-third power. Overall, the conclusion was that UV
was highly effective for wastewater disinfection.
Further work was recommended in the areas of
intensity estimates and the direct analysis of the
impact of hydraulics and the effect of channelized
flow in a non-uniform intensity field.
Johnson and Quails (38) reported the results of a
substantial pilot and laboratory scale effort which
focused on a number of parameters which were key
to the understanding and design of the UV process. In
the early phases of the work they investigated the
performance of two commercial units on contact
stabilization effluent. The first unit was based on the
close spaced, or thin film concept and contained
fourteen 25 Watt submerged lamps at nominal
spacings of 1.25 cm (0.5 in). The second unit utilized
six40-Watt, widely spaced lamps submerged in an 11
liter exposure chamber. Significant differences were
noted, with the widely spaced unit providing far better
performance. This was attributed to short-circuiting
occurring in the first unit. The study also reported
significant photoreactivation in UV irradiated samples
which had been exposed to sunlight for 45 minutes at
25°C (77°F). Total coliforms were found to increase in
density by 1.4 logs. The degree of photoreactivation
was found to decrease with decreasing temperature.
A point source summation model was applied by the
authors to estimate the intensity in a multi-lamp
system. This calculation method treats a finite line
source (the tubular germicidal lamp) as a series of
point sources radiating in all directions. The attenua-
tion of the radiation was inversely proportional to the
square of the distance from the point source and was
absorbed by the liquid medium according to the Beer-
Lambert law/The intensity at any given point in a
system was assumed to be the sum of the radiation
received from all point sources in the system. The
authors also demonstrated that direct beam, spectro-
photometric measurement of the absorbance of the
liquid overestimated the UV absorbance because
scattered light would be measured as absorbed light,
when in fact it is still available. The true absorbance of
the liquid should be estimated by correcting for this
scattering effect. The authors also demonstrated a
spherical integration method for measuring the
actual output of a lamp at any given time.
Quails et al. (39), reported on the effect of suspended
solids on the performance capability of the disinfec-
tion process. Aside from the effect on the UV
absorbance characteristics of the liquid, the major
impact of the suspended solids normally found in a
biologically treated domestic wastewater is the
harboring or occlusion of the bacteria within the
particle. These are protected from the UV radiation
and will be measured as viable organisms subsequent
to UV disinfection. The authors suggest that these
protected coliforms are the major factor in limiting
disinfection efficiency at -3 to -4 log survival units.
Improved disinfection would be accomplished (if
necessary) by prefiltration or improvements in solids-
liquid separation steps at a typical plant.
A bioassay approach to estimate the actual dose in a
system was proposed by Quails and Johnson (40). A
dose-response relationship would first be developed
for a known, pure culture (spores of Bacillus subtilis
were suggested as an appropriate test organism).
This involved exposing the organism to a known, and
measurable, intensity of collimated light (at 253.7
nm) over several exposure times. Dose would be the
product of this measured intensity and the time of
exposure. The calibrated spore would then be injected
into a system. The response (log survival ratio) would
then be compared to the calibration curve to de-
termine the effective dose delivered by the UV unit.
This technique was further applied to a dynamically
flowing system concurrent with a conservative tracer
to determine the time distribution characteristics of
the unit. The spore survival ratio is measured, in this
case, at several times after injection. By accounting
for the hydraulic distribution in this fashion, the
authors demonstrated that it was possible to implicitly
solve for the average intensity within the unit. The
estimates of intensity by the point source summation
calculation method were found to compare favorably
to the intensity estimates made by the bioassay
technique. The authors proposed the use of this
technique to evaluate, and/or compare UV systems,
and as a method to separately evaluate the intensity
and residence time distribution.
Bellen et al. (41), applied the bioassay technique to
seven commercial UV units to determine the dose
application performance of each unit and to compare
their effectiveness. The application was to be ship-
board potable water supplies. They coupled this with
a separate analysis of the residence time distribution
for each unit. The bioassay method was suggested as
the only available technique to directly compare the
performance of commercial units.
Haas and Sakellaropoulos (42) presented a series of
rational analysis solutions for UV disinfection which
incorporated first-order inactivation kinetics with the
hydraulic characteristics of the UV reactor. These
181
-------
were allowed to range from a completely mixed
reactor to a perfect plug flow reactor. Their analysis
supported the premise that the hydraulic character-
istics of a reactor can strongly influence disinfection
efficiency. The optimum hydraulic regime for dis-
infection involves turbulent flow with minimal axial
mixing. Plug flow would be supported in systems with
high aspect ratios (ratio of length to hydraulic radius).
In similar fashion, Severin et al. (43) concluded that
mixing in the radial direction (perpendicular to the
direction of flow) was beneficial to disinfection
efficiency. Mixing in the longitudinal direction (direc-
tion of flow) was not advantageous, although they
suggested that some degree of axial (longitudinal)
mixing may have to be accepted in order to ensure
adequate radial mixing. Severin et al. (44), suggested
the use of a series event inactivation kinetics model to
describe the inactivation of coliforms in reactors of
differing mixing characteristics. The model effectively
predicted efficiency, and suggested the importance of
discouraging stratified flow conditions (lack of radial
turbulence) in a non-uniform intensity field.
Nehm (45) reported on pilot plant studies conducted
at the Nine Springs Water Pollution Control Plant,
Madison, Wisconsin, which evaluated commercial
units from four different manufacturers. The ob-
jectives of these studies were to assess operational
requirements; no attempts were made to measure
dose or to scale-up to design. The results showed that
consistent performance could be accomplished by all
systems when they were kept clean. Scale formation
occured on both quartz and Teflon surfaces. This was
found to be readily removed by the addition of a citric
acid solution to the reactor. The report recommended
the capability of chemically cleaning on any scale-up,
in addition to a UV intensity sensor to monitor the
status of the quartz surfaces. As a post-script to these
studies, UV disinfection was recommended for the
2,200 L/s (50 mgd]i plant, and the system is currently
under construction. This was also one of the first
major plants to require bioassays of equipment which
could be scaled up so as to compare the performance
capability of commercial units. This was imposed as a
pre-bid qualification specification.
Ho and Bohm (46) and Bohm et al. (47) reported on the
pilot scale application of UV to the disinfection of a
number of tertiary and secondary effluents at Ontario
(Canada) Water Pollution Control plants. The process
was demonstrated to consistently meet effluent
disinfection goals. Total coliform and fecal coliform
densities were targeted at 2000 and 200 MPN/100
ml in the effluent, respectively. Corresponding reduc-
tions were demonstrated for Pseudomonas aerugin-
osa, fecal strepococci, and E. coif. Salmonella spp.
were reduced to less than 4/100 ml in 80 percent of
the samples. The investigators propose that UV
transmission is a good surrogate parameter for
correlating effluent water quality to expected UV
effectiveness. The authors estimated dose by calcu-
lating intensity from the Roeber and Hoot model and
the measured mean contact time.
Bohm et al. (47), also assessed the effects of
photoreactivation. Total coliforms increased by ap-
proximately two logs, fecal coliforms by one to two
logs. Little or no photoreactivation was shown for
Pseudomonas aeruginosa and fecal streptococci.
Some increase was observed for the Salmonella ssp.
Temperature did not affect repair, nor did dilutions
with stream water by a factor of 1 to 10.
Whitby et al. (48) reported the results of a full-scale
evaluation of a commercial UV disinfection process at
the Tillsonburg Water Pollution Control Plant, Till-
sonburg, Ontario(Canada). The system design, in this
case, allowed the use of the plant's existing secondary
effluent channels to retrofit a UV unit (see Figure 7-7).
The units were comprised of a series of four-lamp
(quartz-sheathed) modules. The modules had the
lamps arranged in a vertical row; these were sus-
pended on a support frame which had been inserted
into the effluent channel. The number of modules
was dependent upon the width of the channels. The
wastewater flowed parallel to the lamps. The lamp
battery was kept submerged at all times by a
downstream control gate.
A direct comparison between UV and chlorination in
disinfection efficiency was made in this study. UV
was found to outperform the chlorination system; if
photoreactivation were allowed to proceed in the UV
exposed samples, the performance of the two pro-
cesses was similar. This applied to the total and fecal
coliform analyses. Fecal streptococci do not photo-
reactivate and substantial reductions were accom-
plished by the UV systems. A spore forming bacter-
ium, Clostridium perfringens, which is known for its
resistance to disinfection, was also tested. UV was
found to be nearly twice as effective as chlorination in
the inactivation of this organism. Similarly, the units
accomplished greaterthan 99.97 percent inactivation
of bacteriophages, as compared to the chlorination
process, which averaged 95.1 percent.
The studies at Tillsonburg also included fish (rainbow
trout yearlings) toxicity studies downstream of the
plant's discharge which compared the effects of the
chlorinated effluent and the UV irradiated effluent.
Complete mortality was observed within 24 hours
during the chlorination study; the UV disinfection test
was non-lethal for a 48-hour exposure period. The
report suggests very consistent operation of the
system over an 18 month period. Maintenance was
minimal; the lamps had been manually cleaned once
and the lamp life had extended for greater than
12,000 hours for the reporting period.
182
-------
Kirkwold (49) reported on the installation and per-
formance of a UV disinfection system at the Albert
Lea Water Pollution Control Plant, Albert Lea, Min-
nesota. The systems (see Figure 7-6) were reported to
be operating well, after some startup problems, and
performing very effectively. In fact, due to the high
quality of the plant effluent, only a small fraction of
the overall system is needed on a continuous basis.
The author suggests that the cost of operating the UV
system is $5.3/1000 ms($0.02/1000 gal); this is less
than half the operating costs estimated for a com-
parable chlorination/dechlorination process.
Scheible et al. (50) reported on the early studies
conducted for an EPA project at a New York City
treatment plant. They discussed the development of
software to compute the average intensity of UV in a
system of any lamp configuration. They presented the
computed intensity as a function of UV absorbance
and showed the effect of lamp spacing and lamp
output. A procedure was also presented to measure
and analyze the retention time distribution of a UV
system. A subsequent article reported further results
of the New York City project, particularly with regard
to the maintenance of UV systems and the effect of
quartz fouling and lamp aging (51). The development
of a disinfection model was also suggested which
incorporated the hydraulics of a system and the
inactivation rate of colifornis as a function of the
intensity.
7.2.3.1 Current Evaluations of UV Disinfection
A significant amount of the information used to as-
semble this chapter on the UV process is from studies
which were as yet unreported. Only drafts can be
cited at this time, although these may be formally
reported at the time this manual is printed and
released. The following is a short review of these
studies. Further reference will be made to them as
appropriate in the subsequent discussions.
A major research and demonstration project was
completed recently under joint sponsorship by the
USEPA and the New York City Department of
Environmental Protection (52). Conducted at the Port
Richmond Water Pollution Control Plant, Staten
Island, New York, it investigated the performance of
three large scale UV systems in the disinfection of
secondary effluent and high rate settled raw waste-
water. The systems included two 100-lamp quartz
systems which differed only in the spacing between
quartz surfaces—1.25 cm and 5.0 cm (0.5 in and 2.0
in). The third unit used Teflon tubes to carry the
wastewater, with the lamps suspended outside the
tubes. The effluent characteristics were highly var-
iable; the suspended solids ranged between 5 and 50
mg/l, with a UV absorbance coefficient (base e)
between 0.25 cm"1 and 0.5 cm"1. Primary effluent
was also treated to determine the application of UV to
combined sewer overflow wastewaters. These were
characterized by high suspended solids levels and UV
absorbance coefficients between 0.5 cm"1 and 1.0
cm"1.
The UV process was found to be very effective in the
disinfection of secondary effluent. Log survival ratios
between -3 and -4 could be achieved under practical
loading conditions. Similarly, it was shown that a log
survival ratio up to -3 could be accomplished with
primary effluent. The studies also indicated that the
quartz systems were more energy efficient than the
Teflon system. An empirical system loading rate was
suggested to monitor and compare systems; this was
the ratio of the flow (Q) to the actual output of the UV
system (W), in watts at 253.7 nm. This output would
account for the age of the lamps and the degree of
fouling on the surfaces through which the energy
must be transmitted.
A major element of the Port Richmond study was the
development of a protocol for the design of a UV
disinfection process. The resulting model incorpo-
rates the retention time distribution of the system,
and the inactivation rate of the bacteria described as a
function of the calculated intensity in the reactor. This
intensity was calculated by the point source sum-
mation technique and can be adjusted for the
measured (or assumed for design purposes) average
lamp output and the losses of energy due to absorp-
tion by the liquid and the fouling of the quartz (and
Teflon) surfaces. The model was found to correctly
respond to the variables associated with the UV
process when applied to the Port Richmond exper-
imental data. The study demonstrated the importance
of suspended solids in the application of UV. Conforms
occluded by suspended particles will not be affected
by UV and will, in effect, set the limiting density which
can be achieved by UV radiation.
Photoreactivation effects were demonstrated in this
study for both total and fecal conforms. The results
showed a constant fraction increase over densities
measured immediately after UV exposure, regardless
of the initial UV dose. The study also provided
suggestions with regard to the maintenance and
monitoring of the system which can enhance the
efficiency and cost-effectiveness of the process. A
cost analysis of the process shows it to be competitive
with a comparable chlorination system.
As a follow-up to the funding of UV installations
under the I/A program, the USEPA contracted for a
number of post-construction evaluations (PCEs) to
assess the status of these plants. These were
summarized by White et al. (53). Six plants were
visited: Pella, Iowa; Suffern, New York; Northfield,
Minnesota; Togus, Maine; Eden, Wisconsin; and Lodi,
Wisconsin. Overall, the report was favorable to the
183
-------
application of the UV process. It cited the problems
relating to equipment fabrication (ballasts, wiring),
and O&M requirements which were more than
originally anticipated. These will be discussed further
in later sections. An earlier survey of a number of UV
facilities in 1983 was limited due to the lack of
operating experience at the plants.
A series of special studies were conducted at four
operating UV plants to assess plant operations, plant
performance, and to determine the applicability of the
process evaluation techniques developed by the Port
Richmond project (4,54). A one-month study was
conducted atthe Suffern, New York facility. The units
(see Figure 7-4) were evaluated for their hydraulic
characteristics, and performance in the disinfection
of fecal coliforms and fecal strep. Tracer studies
indicated some dispersion and a reduced effective
volume. Analysis of data showed that the systems
were adequately sized and would be able to meet
performance requirements at design loads. The tests
conducted at Suffern showed that sodium hydro-
sulfite is a very effective chemical cleaning agent.
Side by side testing of the system's ultrasonics
cleaning device indicated that it was of limited
benefit, and because of its energy costs, not cost-
effective.
The Teflon system at Eden, Wisconsin was also
evaluated. As expected, the Teflon tube arrangement
was good hydraulically. Tracer analyses showed
relatively low dispersion. Analyses indicated, how-
ever, that the system would not be able to meet
requirements under design loading conditions. Al-
though not directly tested at Eden, significantly
reduced levels of transmissivity by the Teflon was
suggested by the analysis as a major cause of the poor
performance. Special studies were conducted to
evaluate a method to directly measure the Teflon
transmittance, utilizing chemical actinometry tech-
niques. These showed that virgin Teflon transmitted
75 to 85 percent of the UV(the range being a function
of thickness). Samples of Teflon from the Port
Richmond facility, which were heavily fouled, were
found to transmit only 5 to 30 percent of the UV.
When the surfaces were thoroughly cleaned, the
transmittance increased to 65 to 70 percent. Limited
field studies were also conducted at the Vinton, Iowa
facility (see Figure 7-3K The hydraulic tracer analyses
indicated a large degree of dispersion and a signif-
icant reduction in effective volume.
A series of studies were recently completed which
used the same pilot plant to test wastewaters at
several wastewater treatment plants (55). The tests
were directedto investigating the inactivation rate as
a function of the intensity, and the level of bacterial
occlusion by the suspended solids.
7.3 Process Design of the UV Wastewater
Disinfection System
This section presents the design protocol for the UV
process. The mathematical expressions which are
presented are based on the analysis of the system as
a chemical reactor, incorporating the retention'time
distribution of the system, and the inactivation rate as
a function of the intensity of UV radiation within the
reactor. The discussions in this section will be in the
following format; first, the process design model is
presented. This expression is the framework about
- which the process is designed and the system sized.
The data requirements for the system design are then
presented in general form. As will be shown, the
design information falls into three major categories:
hydraulics; UV radiation intensity; and wastewater
characteristics (including bacterial sensitivity to UV).
Second, the hydraulic design considerations, as they
relate to the UV reactor, will be discussed. These will
encompass the flow and dispersive characteristic of
the reactor, and the considerations of head loss,
turbulence, and effective volume.
Third, the intensity of UV in the reactor will be
discussed. The procedure for estimating the average
intensity will be presented, including solutions for
almost all practical UV lamp configurations.
Fourth, the relevant wastewater quality parameters
are reviewed. Aside from the normal water quality
parameters we will need to know, such as coliform
densities, UV absorbance coefficients, etc., the
discussions will also present procedures to determine
the coefficients which describe the sensitivity of the
coliforms to UV, and the densities associated with
particulates normally found in treated municipal
wastewaters.
Finally, the design protocol is presented in Section
7.4, incorporating the discussions of this section. The
protocol will be demonstrated by using the design
example which has been carried throughout this
manual.
7.3.1 Process Model to Describe UV Reactor
Performance
7.3.1.1 First Order Kinetics for the UV Process
Recalling the discussions presented in Chapter 4, the
inactivation of bacteria by UV radiation can" be
approximated by the first order expression:
N = No e-k"
(7-6)
where:
N = bacterial density after exposure to UV
(organisms/L3)
184
-------
No = the initial bacterial density (organisms/L3)
k = inactivation rate constant (L2Watts~1T~1)
I = the intensity of the germicidal UV energy
(Watts/L2)
t = time of exposure (T)
The intensity is the rate at which the energy is being
delivered to the liquid; in the context of this report,
intensity has the unit microwatts per square centi-
meter (//Watts/cm2). When multiplied by the time to
which an entity is exposed to this rate, the quantity of
energy, or dose, is determined:
Dose (/uWatt-sec/cm2) = Intensity (/uWatt/cm2)
x Time (seconds) (7-7)
The rate constant, k, is the slope of the relationship of
ln(N/N0) as a function of the dose. It is generally held
that the intensity and time are reciprocal in their
effect on dose.
Figure 7-14. Effect of particulates on UV disinfection
efficiency.
Ideal
CO
cr
£
3
in
Dose
7.3.1.2 Incorporation of Particulate Coliform
Densities
Although the first order expression is generally a
good first approximation of the response to a given
dose, direct testing on mixed cultures will often show
a reduced efficiency with increasing dose. In the
disinfection of treated wastewaters by ultraviolet
radiation, in particular, this is attributed to the
aggregation or the occlusion of bacteria in particulate
matter (28,31,33,36). Ultraviolet light is unable to
penetrate this material and effect inactivation of the
bacteria. Quails et al. (39) presented data which
demonstrated that removal of particles (by filtration)
large enough to harbor coliforms exerted dramatic
effects on the dose-survival relationships. They
concluded that these protected coliforms were the
major factor limiting improved disinfection at -3 or -4
log survival units. Thus, as the dispersed or singlet
bacterial organisms are inactivated, continued eleva-
tion of the dose will show a diminishing response as
the residual active bacteria are protected in the
particulates. This is schematically presented on
Figure 7-14. In light of this. Equation (7-6) can be
more accurately written as:
N = Ni, exp (-kit) + Np
(7-8)
where N0' is the initial, non-aggregated density, and
Np is the density associated with the particulates and
unaffected by the UV radiation. When considering
treated domestic wastewater, N0' » Np, such that
the total initial density, N0, can be considered equal to
N0' + Np. The expression can then be written:
7.3.1.3 UV Process Design Model
Considering Equation 7-9, the use of a single
exposure time presumes the ideal case of perfect plug
flow in the reactor, with no axial dispersion. Under
actual conditions, this ideal plug flow does not exist.
Axial dispersion and velocity gradients will cause a
distribution of residence times; this will be a function
of the dispersion characteristics of the reactor, which
can be quantified by defining the spread (or variance)
of the time distribution relationship for a specific
reactor (see Chapter 4).
A disinfection model was developed and reported by
Scheible et al. (56). Reference is made to that report
for a detailed development of the model. It presumes
the first order expression given as Equation 7-9, but
also incorporates the dispersive properties of the
reactor, in effect describing the residence time
distribution of the reactor under steady-state condi-
tions. This model forms the basis for the design
protocol presented herein for the UV process. The
general expression is written:
N = N0exp[—{1-(1
2E
where:
N
(7-10)
N = No exp (-kit) + Np
(7-9)
N = the bacterial density remaining after exposure
to UV (organisms/100 ml).
NO = the initial bacterial density, measured imme-
diately before entry into the UV reactor
(organisms/100 ml)
755
-------
x = the characteristic length of the reactor, defined
as the average distance traveled by an element
of water white under direct exposure to UV
(centimeters)
u = the velocity of the wastewater as it travels
through the reactor (cm/sec). This is calculated
as:
u = x/(Vv/Q)
where Vv is the void, or liquid volume in the
reactor (liters) and Q is the total flow (liters/
second). In cases where significant "dead
volume" is indicated within a reactor, Vv is
adjusted to more closely approximate the
"effective" liquid volume.
the dispersion coefficient (cmVsecond). E
quantifies the spread of the residence time
distribution of a particular reactor.
the rate of bacterial inactivation (seconds"1)
the bacterial density associated with the
particulates and unaffected by exposure to UV.
Figure 7-15.
E =
K
Np
The rate of inactivation, K, is expressed as a function
of the UV intensity. Thus, for a given time of exposure
(or distribution of exposure times), the rate of
inactivation will increase (or decrease) with an
increase (or decrease) in the intensity. This is shown
graphically on Figure 7-15. In this fashion, an
expression can be developed in which K is estimated
as a function of intensity:
K = f(lntensity)
(7-11)
As will be discussed in a later subsection, this
correlation will be developed by relating the log K to
the log Intensity, where the intensity is the average
reactor intensity, lB,va. This yields the expression
(when transformed).
K =
a l(avg)b
(7-12)
where a and b are the slope and intercept of the linear
regression.
With regard to the paniculate bacterial density, Np,
this is generally described as a function of some
measureable index of paniculate density in a waste-
water, such as suspended solids or turbidity. Sus-
pended solids is used in the context of this manual
since it is the parameter most commonly measured
and most relevant to wastewater treatment applica-
tions. The value of Np is described as a function of the
suspended solids by the correlation of the log effluent
coliform density to the log SS. When transformed, the
expression is in the form:
Np = cSS"
(7-13)
where SS is in mg/F. As will be discussed in a later
section, the effluent densities for this analysis must
The rate K increases with increasing intensity
for a given residence time.
I = Intensity
N.
(a)
Time
(b)
attn
I (yuWatts/cm2)
be generated under very high dose conditions. In this
fashion, it is appropriate to assume that the remaining
bacteria are those which were occluded in the
paniculate matter and were unaffected by the UV
radiation.
By incorporating Equations (7-12) and (7-13), into
Equation (7-10), the UV design model can be ex-
pressed as follows:
N-=
b 1/2
2E
cSSm(7-14)
7.3.1.4 Data Requirements to Use the UV Design
Model
The information that would be required to effectively
use the UV design model relate to the characteristics
of the wastewater to be disinfected, and to the
physical characteristics of the reactor itself. Consider
first the wastewater application; the data which must
be generated either by direct testing, or by estimates
based on experience include:
Initialbacterialdensity, N0. Explicit in the expression,
this should be determined under the average and
maximum conditions anticipated for the plant.
186
-------
Flow, Q. The flows to be handled by the disinfection
process. This is implicitly required to determine
velocities and loadings to the system.
UV absorbance coefficient, a. This will affect the
intensity of radiation in the reactor, and is a direct
measure of the energy "demand" of the wastewater.
Suspended solids (SS). This will be defined by the
permit limitations the plant is designed to meet. As
discussed later, the paniculate bacterial density will
be related to the suspended solids concentration.
Paniculate bacterial density, Np. This density asso-
ciated with the particulate forms the minimum
density level which can be achieved by the UV
process. It is typically determined as a function of the
suspended solids concentration.
Coefficients, c and m. These are determined from
Equation (7-13) and describe the particulate coliforrn
density associated with the suspended solids.
Rate of inactivation, K. This rate is a measure of the
sensitivity of the bacteria to UV radiation, and will be
site specific. As discussed, the value of K is estimated
as a function of the intensity of UV radiation which a
particular reactor can deliver. An estimation of this
rate will, therefore, require knowledge of the actual
intensity levels within the UV reactor.
Coefficients, a and b. From Equation (7-12), these
describe the rate of inactivation as a function of the
average intensity.
The remaining parameters which are to be addressed
relate to the physical design of the reactor:
Velocity. The velocity of the liquid, as described above,
is set by the rate of flow, Q, and by the physical
dimensions of the reactor. Specifically, these are the
characteristic length, x, and the liquid volume, Vv, of
the reactor.
Length. The characteristic length of the reactor is the
distance traveled by the liquid while under direct
exposure to UV light.
Dispersion coefficient, E. This parameter accounts for
the deviation of the reactor's hydraulic behavior from
that of perfect plug flow; in effect, the distribution of
residence times at steady-state is forced by the
dispersion coefficient.
Average intensity, /ava. The average intensity in a
reactor is a function of the lamp (i.e., UV energy)
density in the reactor and the UV absorbance charac-
teristics of the liquid.
In the situation when one is evaluating an existing
system, the physical dimensions are fixed. The task is
to properly calibrate the model, which can then be used
to assess the system's capacity and to optimize its
operations. When a new system is to be designed, the
approach is to establish the wastewater parameters,
ideally by direct bench or pilot scale testing, and then to
determine the optimum hardware configuration and
sizing.
7.3.2 Characterization of the Hydraulic
Behavior of a UV Reactor
Recall the objectives of an effective hydraulic design,
as it would apply to the design of the UV disinfection
reactor. (The reader should refer to the discussions in
Chapter 4 regarding hydraulic considerations in disin-
fection reactor design.) First, the unit should be a plug
flow reactor (PFR) in which each element of fluid
passing through the reactor resides in the reactor for
the same period of time. Second, the flow motion
should be turbulent radially from the direction of flow.
This is to allow for each element to receive the same
overall average intensity of radiation in the non-
uniform intensity field which exists in the reactor. The
tradeoff in this requirement is that some axial dis-
persion will be introduced, yielding a dispersive or non-
ideal flow reactor.
Third, maximum use must be made of the entire
volume of the reactor; conversely, dead spaces must be
minimized, such that the effective vol ume is very close
to the actual volume available.
When evaluating an existing reactor, the hydraulic
evaluation should entail direct testing of the unit to
establish the residence time distribution (RTD). Sub-
sequent analysis, as described in Chapter 4, can serve
as an excellent diagnostic tool in examining non-
performance or to determine system capacity.
New systems design requires the engineer to specify
equipment configurations which will be hydraulically
efficient. The indices and dispersion characteristics
discussed in Chapter 4 can serve as design specifica-
tions. Evaluation of commercial reactors can rely on
the development and evaluation of the necessary
hydraulic information from scaleable pilot units or full
scale modules.
The following discussions present the major elements
of effective hydraulic design for UV systems:
• residence time distribution (RTD)
o dispersion
• turbulence
• effective volume
7.3.2.1 Residence Time Distribution
The evaluation of a specific reactor relies on the
construction of the RTD appropriate for that reactor
configuration. This can be accomplished by a number
of experimental procedures; subsequent analysis of
the residence time distribution curves determines the
187
-------
hydraulic characteristics of the unit. Experimental
procedures and analysis techniques were presented in
Chapter 4. Particular attention should be paid to the
discussions regarding reactors with short residence
times.
Consider the analysis of a specific RTD to demonstrate
the appropriate calculations and interpretation. The
example is taken from the Port Richmond project
described by Scheible et al. (55). Unit 2 in this study
was a submerged quartz system configured in a
fashion similar to that shown on Figure 7-6. It
contained 100 lamps, parallel to one another, each
held in quartzsleeves with an outer diameter of 2.3 cm.
The method by which the RTD was developed was
described in Chapter 4 and presented schematically on
Figure 4-4(b). The relevant unit characteristics are:
ti
3.8
4
4.6
5.1
5.4
6.3
7.0
7.2
8.3
10.0
11.2
13.0
13.9
14.1
16.2
Ci
0
0.003
0.011
0.02
0.028
0.039
0.049
0.061
0.044
0.029
0.017
0.01
0.008
0.005
0.001
tid
0
0.012
0.0506
0.102
0.1512
0.246
0.343
0.439
0.409
0.29
0.190
0.13
0.111
0.071
0.016
ti2Ci
0
0.048
0.233
0.520
0.816
1.548
2.401
3.162
3.806
2.9
2.13
1.69
1.55
0.994
0.259
Cumulative
(%)*
0.47
2.46
6.44
12.3
21.9
35.3
53.5
68.4
79.8
87.0
92.2
96.5
99.2
100.0
x = distance between tracer input and output,
which is approximately equivalent to the lamp
battery dimension in the direction of flow.
= 47cm
Vtf = liquid volume of reactor
= 100 liters
T = theoretical mean residence time, VV/Q
= 7.0 seconds
At a flow of 890 Lpm, the velocity is computed to be 7
cm/sec. Figure 7-16 presents the tracer curve and
resultant RTD developed for one run. The upper panel
Is the F-curve developed by the so-called step input
tracer analysis described in Chapter 4 (see Figure 4-
3{c)). A salt tracer is continuously injected upstream of
the lamp battery until a steady-state concentration is
read by the conductivity probe positioned immediately
downstream of the battery (all at t < 0).
Att = 0, the salt injection is discontinued. The trace on
the upper panel is a record of the die-away at the
downstream probe until a new steady-state condition
is reached, which in this case, is the background level.
Recalling that the derivative of the F-curve is the C-
curve (see Figure 4-3), the C-curve can be constructed
by plotting the slopes of tangents (dc/dt) drawn at
points along the curve against time. This is shown on
the middle panel of Figure 7-16.
The analysis of the RTD curve can be accomplished
graphically by breaking the curve into discrete areas at
discrete time intervals. The calculations are demon-
strated by the following, where Ci is dc/dt and tj is the
corresponding time:
- 0.325 2.561
24.39
*Cumulative percent tracer at observation j =
J
1
n
I tiCi
x100
where j < n
n = total number of observations
The mean residence time, ft is the centroid, or first
moment of the distribution (from Equation 4-13):
0
ItiCi
XCl
2.561
0.325
= 7_g seconds
The last column in the above calculation is the
cumulative area as a percent of the total. By plotting
this against time, as shown on the lower panel of
Figure 7-16, one can display the cumulative tracer with
time. This then allows one to evaluate any number of
the indices defined by Rebhun and Argaman (57) and
discussed in Chapter 4. For this particular tracer
analysis, the following parameters are determined:
ratio of initial to
theoretical time
ratio of peak to
theoretical time
Morrill Dispersion
Index
ratio of mean
residence to
theoretical time
ratio of median
to mean residence
time
tf/T .= 3.9/7 = 0.56
tp/T = 7.2/7 = 1.03
t90/tio = 12.4/5.3 = 2.3
0/T = 7.9/7.0 =-1.13
teo/6 = 7.2/7.0 = 1.03
188
-------
Figure 7-16. Example of RTD curve developed for Unit 2 at
Port Richmond by the step input method (55).
0.8
c 0.6
o
E
a. 0.4
ra
CO
0.2
0
0.08
(D "o"
O
-------
Fiouro 7-17.
100000
10000
1000
100 £
Relationships of velocity, length, and disper-
sion.
Characteristic Length, x (cm)
100
1000
10000
0.1
0.001
0.01 0.1
d (E/ux)
The dimensionless variance is computed (Equation
4-19):
The value of d can be estimated by (Equation 4-21 ):
-2
ux
ux
or
= 2d-2d2(1 -
Ignoring the second term on the right, the first
approximation of disO.102. Adjusting by trial and error
forthe second term, the value of dbecomesO.104. This
suggests that the first Gaussian approximation was
adequate, i.e.:
d~
Note also that d is « 0.5; thus, it is reasonable to
consider it as a closed vessel. The value of d (0.104)
also confirms the moderate to highly dispersive nature
of the reactor.
The tubular reactor variance is 0.624 sec2. The
dimensionless variance is 0.026; from this, the dis-
persion number, d, is estimated to be 0.013, reflecting
the low dispersion, plug flow nature of this particular
reactor design.
The dispersion number and the dispersion coefficient
are utilized in the design equation for the UV process
(Equation 7-10). Correlations can be developed to
estimate the dispersion number as a function of
reactor characteristics relating to friction losses,
hydraulic radius, velocity, etc. Such predictive models
exist for pipe systems. Such models have not been
developed for UV reactors. There is limited data
available to attempt this; certainly further work is
needed in this effort.
The dispersion coefficient should be expected to vary
with velocity. In current practice, values of E are
selected to represent conditions under high flow; a
design goal is then set with the dispersion number
(e.g., d = 0.02 to 0.05) which will force limits on uxfor
the given E. This is discussed further with the design
example presented in Section 7.4, particularly as it
relates to the impact on head loss.
7.3.2.3 Turbulence and Head Loss
An important consideration in the hydraulic design of a
UV reactor is the turbulence of the fluid. By having
turbulent flow, any particle has an equal probability of
being at any point in the cross-section of the conduit,
as it travels in the direction of flow. The importance of
turbulence lies in the fact that the intensity field in the
reactor, regardless of the way the lamps are con-
figured, is non-uniform. Thus, if a particle is forced to
move erratically by th? turbulent conditions, it will
likely see all intensity levels in the non-uniform field. In
this case, then, it is acceptable to use the average
intensity in the reactor to evaluate dose levels micro-
organisms receive as they move through the reactor. If
true laminar conditions existed, streamlines may move
through areas of low-intensity and receive little dose
relative to the streamlines moving close to the lamps.
Flowthrougha reactor can be characterized by eddies,
swirls, and irregular movements of large fractions of
the fluid; these do not constitute turbulence. They may
more correctly be described as "disturbed flow." Thus,
a reactor flow may be laminar, but have disturbances;
the lamps themselves can be the source of these
disturbances. Turbulence is generally induced by high
friction losses and high velocities.
Turbulence indicatedby headloss. If the log of the head
lossfor a given length of uniform pipe is plotted against
the log of the velocity, it will be found that effectively
two regions exist. Where the velocity is low enough to
190
-------
assure that laminar flow exists, the head loss, hu, due
to friction, will be directly proportional to the velocity, u:
hi. — un, where n = 1 (laminar) (7-15)
As the velocity increases, at some point turbulence is
induced and the hu is found to increase at a higher rate
than the increase in velocity. In uniform pipe this value
of n is greater than 1.0;
hi. = u", where n = 1.75 to 2.0 (turbulent)
Such measurements were taken during the Port
Richmond study for the quartz systems. The data for
one of the quartz units are presented on Figure 7-18;
these demonstrate the transition from the laminar to
the turbulent flow condition. At velocities less than 10
cm/s(logu = 1.0), the value of n is approximately 1.0;at
velocities greater than 10 cm/s, the n is estimated to
be 2.0.
Figure 7-18. Log-log plot of head loss against velocity for
Unit 2 at Port Richmond indicating transition
from laminar to turbulent flow regime (55).
10 100
Velocity (cm/sec)
This type of information can be developed by direct
measurement on full scale modules or hydraulically
scaleable pilot units. Care should be taken that the
head loss due only to the lamp reactor itself is being
measured. Losses due to reactor entrance and exit
conditions should be separated from the analysis.
Turbulence indicated by Reynolds Number. Velocity is
not the only factor that determines if a flow is laminar
or turbulent. The criterion is the Reynold's number.
This dimensionless number, NR, is the ratio of inertia
forces to friction forces in a completely filled conduit.
Thus:
Lu
V
(7-16)
where:
u = velocity
p = density of fluid
fj = viscosity
L = linear dimension significant to pattern of flow
v = kinematic viscosity
The linear dimension, L, for pipes is generally taken as
the pipe diameter. A straightforward example of the
Reynolds number analysis is the tubular flow array.
The linear dimension is taken as the tube diameter. The
Reynolds Number is plotted on Figure 7-19 against the
velocity and the flow rate per 3-meter longTef Ion tube.
Figure 7-19.
Estimates of Reynold's Number for 8.9 cm
diameter Teflon tubes.
100000
80000
60000
40000
20000
-S10000
5. 8000
i 6000
=> 4000
,w
1 2000
-------
Hydraulic Radius, RH = -r-^
Aw
(7-17)
where:
Vv = liquid, or void volume (cm3)
Aw = total wetted surface area (cm2)
The void volume is the total reactor volume minus the
volume occupied by the lamps and quartz sheaths.
The wetted surface area is the sum of the surface
areas of the quartz sleeves and the internal wall area
of the reactor.
For a circular conduit flowing full, the hydraulic
radius is equal to one-fourth the diameter of the
conduit. Thus, for Equation 7-16:
L = 4RH
(7-18)
The Reynolds Number can then be estimated for the
quartz, submerged reactors:
NF
(7-19)
Consider Unit 2 from the Port Richmond study as an
example of this calculation:
systems designs. It should be understood that al-
though the difference in performance has not been
demonstrated, it is implicit in the physical mechanism
of UV disinfection that the liquid be in turbulent
motion. The Reynold's Number calculation offers a
method to qualitatively evaluate this criterion.
Estimating Head Loss. There is little information on
head losses caused by the UV lamp batteries. This will
vary according to the size of the reactor, the velocity,
and the placement of the lamps. The tubular reactor
head loss can be estimated from pipe flow equations,
assuming smooth wall friction coefficients. With
regard to the quartz systems (submerged), one would
expect higher losses in the unit with the lamps placed
perpendicular to the flowpath, than those configured
with the lamps parallel to the flowpath.
Scheible (56) reported an empirical relationship
developed for the quartz units at Port Richmond. This
would be representative of only that type of configu-
ration: open channel structure, uniform lamp array,
and the flowpath perpendicular to the lamps. Diresct
testing would have to be done on alternative con-
figurations in order to determine head losses, al-
though a similar approach can be taken in doing so.
The head loss expression is based on Darcy's equation
for pipe flow:
RH
_ Vv _ 10.2x104cm3
Aw 7.2x10" cm2
fxu*
'
At a kinematic viscosity of 0.0098 cmVsec (water at
20°C), the Reynolds number is estimated:
(7-20)
where the velocity, u,, is in cm/s. This estimate of NR is
plotted on Figure 7-1!8 for Unit 2. From this analysis,
the breakpoint appears at an NR slightly less than
6,000. This is somewhat higher than the NR of 4,000
normally considered as minimal for turbulent flow.
The linear dimension estimate (4RH) may be a factor
in this. The fact is, that as long as the method of
estimating the Reynolds number is kept consistent, it
is possible to qualitatively evaluate a unit design for
turbulence.
Table 7-6 presents a summary of the Reynolds
Numbers estimated for several lamp array configura-
tions. These are based on the estimated hydraulic
radius and the design flow range for each unit. As
shown, all systems would typically operate at veloc-
ities high enough to yield turbulent flow. If a Reynolds
Number of 6,000 were set as a minimum, the
minimum velocity cam also be shown. Such a criterion
can be used in establishing specifications for new
(7-21)
where x and u are the length (cm) and velocity (cm/s),
respectively, dr is the diameter (cm), and f is the
coefficient of friction. In this case, dr is set as the
approximate hydraulic radius of the system from
Equation 7-17. Finally, a new coefficient of friction is
defined:
Cf (cmVsec ) =
RH2g
such that:
hL = c,(x)(u)2
(7-22)
(7-2:3)
At Port Richmond, direct hi. measurements of the
lamp batteries in the two quartz systems yielded a ct
between 0.0001 73 and 0.00023 secVcm2.
/
The head loss can become a factor in the design of a
reactor. Earlier discussions cited the dispersion
number as a good design guideline; e.g., designing a
reactor to yield a low d value. Recalling that:
ux
192
-------
Table 7-6. Summary of Reynolds Number Estimates for Different Lamp Configurations
System
Port Richmond 1*
Port Richmond 2"
Vinton'
Suffern*
Port Richmond 3"
Configuration"1
uniform array
uniform array
concentric array
staggered array
tubular array
RH
(cm)
4.7
1.4
4.74
3.65
2.23a
Hydraulic"
Flow Range
(Ipm)
1,000-2,600
1,000 - 2,600
2,000 - 4,500
3,000 - 10,000
40 - 100°
Equivalent
Velocity
(cm/sec)
3.5 - 9.1
9.3 - 24.0
6.8-15.4
7.3 - 24.3
10.8 - 27.0
Reynolds
Number NR
6,720 - 17>500
5,400 - 14,000
13,000 - 30,000
10,900 - 36,200
9,800 - 24,600
Velocity at
NR = 6000
(cm/sec)
3.1
10.3
3.1
4.0
6.6
"Diameter/4.0.
bDesign flows.
cLiters/min/tube.
dSee section on UV intensity for definition of array configurations.
"Reference 56; also see Figures 7-6 and 7-8.
'Reference 4; also see Figures 7-3 and 7.4, respectively.
it can be seen that for a given E, the d is forced by ux.
Thus, for an increasing E, the ux must be propor-
tionately increased to maintain a fixed value of d. We
see by Equation 7-23, however, that increasing
velocity can have dramatic effects on the head loss,
since it will increase by the square of the velocity.
Increasing x can be done to some degree; this must
stay within practical limits, however. One should
understand that the head loss we are considering is
that incurred through the lamp battery itself; addi-
tional head losses will be incurred at the inlet and
outlet structures.
In all, d will need to be reconciled with the head loss.
Increasing the acceptable d will allow one to reduce
and keep the hi. within design limits. Generally, for
gravity flow systems, one should design about a d
between 0.02 and 0.05. One'should understand that
these are still excellent dispersion numbers for a
disinfection reactor.
7.3.2.4 Effective Volume
The lamp battery volume is that portion of the total
system occupied by the UV lamps..By this fact, it is
very important that the reactor is designed such that
full use be made of the entire volume. Dead zones or
short-circuited areas mean ineffective use of lamps
and power, the two components which comprise a
major portion of the capital and operating costs. This
is primarily a design consideration for the submerged
quartz systems. The tubular array configuration
should inherently provide for maximal use of the
Teflon tube volume.
Maximum use of the reactor volume is directly related
to the approach and exit conditions of the reactor. The
design intent must be to first enter the front plane of
the lamp battery with equal fluid velocity at all points.
This same condition must exist at the exit plane. The
approaches taken to encourage this include:
• open channel flow before and after the lamp
battery,
• overflow weirs placed the full length of the reactor.
• perforated stilling walls before and after the lamp
battery, and
• unidirectional flowpath throughout the approach,
battery, and exit sectors.
Figure 7-20 schematically displays these considera-
tions. The upper panel shows a plan-view of an open
channel type configuration. The inlet and outlet
chambers should be independent of the lamp battery.
Weirs can be placed across the width of the reactor to
distribute the flow evenly across the unit. The
perforated baffle plates then serve to distribute the
flow evenly both horizontally and vertically. The weir
Figure 7-20.
Inlet and outlet considerations for submerged
quartz systems.
Weir Plate Perforated Plate
Weir
- Z i; Lamp Battery£&t _ _
Effluent
Chamber'
Influent Chamber z- Lamps are Parallel or
Perpendicular to Flowpath
(a) Open Channel-Type Configuration
Lamp Battery (Generally)
Parallel to Flowpath)
I
-Potential
Dead Zones
Perforated Baffles
/
u
\
(b) Sealed Cylindrical Reactor Configuration
793
-------
plates may not be needed before the perforated baffle
if there are no extreme velocity gradients coming into
the unit. By having the same arrangement on the exit
side, the flow paths are kept stable and unidirectional
through the lamp battery.
In the middle panel of Figure 7-20, a sealed cylindrical
shell type reactor is shown. The lamps are generally
parallel to the flow piath and the wastewater enters
and exits the reactor perpendicular to the lamps. As
shown, there is the tendency to induce flow chan-
nelling with this arrangement, causing dead zones in
the reactor. A solution is shown on the lower panel, in
which perforated baffle plates are installed at both
ends of the unit. These serve to distribute the flow
over the entire cross-sectional plane of the reactor.
To evaluate a system for effective volume, the indices
derived from the RTD curve are useful. The ratio of the
mean residence time to the theoretical residence
time (0/T) should be approximately 1.0. The actual
fraction is reflective of the actual volume being
effectively utilized.
It is suggested that the &/J ratios be used in specifying
the hydraulic design of a UV reactor. A value greater
than 0.9 is an appropriate requirement.
7.3.2.5 Summary Considerations for Effective
Hydraulic Design
In summary, the key points to address when evalu-
ating or specifying the design of a UV reactor are as
follows:
Residence Time Distribution. This should be con-
structed at a number of flow conditions for an existing
system; it should also be required when specifying
commercial systems. The RTD provides key informa-
tion on the actual or anticipated hydraulic behavior of
a reactor.
Plug Flow. This can. be quantitatively described by
indices derived from the RTD analysis. Appropriate
guidelines for specifications can be:
tf/T > 0.5
tgo/tio < 1.0
tp/T > 0.9
tso/0 = 0.9 to 1.1
Additionally, the dispersion coefficient, E, should be
relatively low (<500 cmVs); the dispersion number
should be less than 0-.1 (preferably less than 0.05 if
the head loss is acceptable).
Reactor designs which are conducive to plug flow
have high aspect ratios, x/L Thus, the length, x (i.e.,
the distance in the direction of flow), should be
significantly higher than the appropriate cross-sec-
tional dimension, L. In tubular reactors, such as the
Teflon tube units, this is the diameter. In submerged
quartz units, L is 4Rn, where RH is the hydraulic
radius. As a guideline, an aspect ratio greater than 15
should be incorporated into a reactor design.
Maintenance of plug flow within a reactor will be
influenced by the approach and exit conditions. The
design should have minimal disturbances at the inlet
and exit planes of the lamp battery; directional
changes in the flowpath would best be made outside
of the lamp battery.
Dispersion Number. A key goal is to minimize the
dispersion number, d. A design goal should be to have
a d between 0.02 and 0.05. Levenspiel (58) suggests
that this would be representative of a plug flow
reactor with low to moderate dispersion. This can be
accomplished by increasing the product of ux, even in
a system with a relatively high dispersion coefficient.
The designer should be aware, however, that ex-
tended lengths and higher velocities will cause higher
head losses. In certain situations some adjustment of
the dispersion number may be necessary in order to
meet specific head loss requirements.
Turbulence. Radial turbulence isnecessaryduetothe
non-uniform intensity field. The reactor design should
induce an estimated (by the procedure discussed
earlier) Reynold's Number greater than 6,000 at
minimum flow. If possible, it would be beneficial to
confirm the laminar/turbulentflowtransition velocity
by direct head loss measurements on the lamp
battery.
Head Loss. Direct measurements should be required
for full-scale modules or scaleable pilot units as part
of commercial equipment specifications. These
should be determined over a wide velocity range and
should exclude entrance and exit losses.
Effective Volume. Maximal use of the reactor lamp
battery is essential to keep the process cost-effective.
This will be related directly to the reactor's inlet and
outlet design. The goal must be to have equivalent
velocities at all points upon entering and upon exiting
the lamp battery. Stilling walls (perforated baffles),
and weirs should be incorporated into reactor designs
to assure this. Guidelines for specifying commercial
equipment should require the ratio 0/T to be greater
than 0.9 and/or the tp/T to be greater than 0.9.
7.3.3 Estimation of the A verage Intensity in a UV
Reactor
The second element of dose, after time, is the
intensity of energy during the exposure time. Recall
that the intensity is the rate, or flux, of delivery of
photons to the target. In the UV process design model,
Equation (7-10), the rate of bacterial inactivation is
described as a function of the intensity. By this fact it
becomes important to be able to quantify the intensity
in a given system. The intensity in a reactor is a
194
-------
function of the UV source (output), the physical
arrangement of the source relative to the wastewater
(the arrangement of the lamps and their placement in
or out of the liquid), and the energy sinks present
which will attenuate the source output before it can
be utilized for disinfection purposes.
The UV source, as discussed earlier, is typically the
low pressure mercury arc lamp. Table 7-7 presents
lamp specifications for a series of germicidal lamp
models. The lamps generally used in UV reactor
systems are equivalent to the G64 and G36 units. The
overall lamp length is approximately 0.9 m for the
G36and1.6 m for the G64. The arc length defines the
active, light emitting portion of the lamp (0.75 m and
1.5 m, respectively.) The diameter of the lamp is
small, typically 1.5 to 1.9 cm. The lamp envelope is
made of fused quartz or other highly transparent (to
the 253.7 nm wavelength) glass, such as Vycor.
In the quartz systems, the individual lamps are
sheathed in quartz sleeves only slightly larger in
diameter (2.3 cm) than the lamp and the entire
lamp/quartz bundle is submerged in the flowing
liquid. In systems where the wastewater does not
contact the quartz or lamp surface, separate conduits
carry the wastewaters. The conduits are translucent
to the UV light, with the lamps placed near the outside
conduit wall.
Determining the intensity at any point in these
complex lamp reactors is not straightforward. At
present, there is no commercially available detector
which can measure the true intensity in such a
system. The problem lies in the fact that the detectors
are planar receptors; only energy striking a flat
surface will be measured. Such detectors wilf inter-
cept fractions of light striking the surface at an angle.
Only light which is normal to the surface, i.e.,
collimated light, however, will be wholly measured.
"Cosine-corrected" detectors attempt to compensate
for this by adjusting for the angular light. These still
measure, however, only the planar intensity. Where
light is not collimated, as is the case with a multi-lamp
UV reactor, the flux of energy is three-dimensional.
This same concept is enforced when the target of the
radiation is considered. In turbulent motion, all
particles can be expected to receive equal exposure.
Several approaches have been proposed to estimate
light intensity, including chemical actinometry, bio-
logical assays, and direct calculation. The two pro-
cedures which have received the greater attention
are the bioassay and direct calculation methods. The
bioassay procedure has been applied in a limited
fashion for a number of design specifications, pri-
marily as a technique for quantifying the dose
delivered by a specific piece of UV equipment. It can
also be used to implicitly derive the intensity within a
system.
The second method which is used and which is
generally emphasized within the context of this
manual, is the direct calculation of intensity. This is
accomplished by the point source summation method.
The discussions briefly describe the calculation
framework which yields the average intensity as a
function of the UV absorbance coefficient of the
wastewater. A series of solutions are then presented
which the designer can use to estimate intensity in a
number of lamp configurations. Finally, discussions
are presented regarding the factors of lamp deter-
ioration and enclosure fouling which will absorb the
UV energy, and thereby reduce the intensity.
7.3.3.1 Bioassay Procedure to Estimate Intensity
and UV Dose
The assay procedure has been proposed as an
effective method for estimating delivered dose and
system intensity (40). This technique is shown
schematically on Figure 7-21.
UV sensitive pure culture is calibrated to the UV dose
using the collimated light device shown on Figure
7-21 (a). The collimating device allows one to accur-
ately measure the intensity directly with a com-
mercial radiometer. Aliquots of the bacterial sus-
pension are then exposed to this given intensity for a
series of fixed time intervajs, yielding known doses.
The response is then plotted against the dose. This
dose-response relationship serves as the calibration
for the subsequent reactor assays (Figure 7-21 (b)).
The unit to be tested is set to the desired flow and
operating conditions and the culture is injected into
the influent. The effluent is then sampled with time
and assayed for the known bacterium. This same
procedure is repeated without the lamps in operation.
The resulting densities are as shown on Figure 7-
21 (c). For each time interval, the log survival rate is
determined. The equivalent dose can then be esti-
mated from the dose response calibration curve.
When the dose is plotted against time (Figure 7-
21 (d)), the slope of the correlation is the dose-rate, or
intensity.
This method for determining intensity in a system can
require a fair sized laboratory effort. A quality analysis
requires very frequent sampling and analysis and
should be replicated to assure precision. This can be
costly and is not cost-effective when compared to the
alternative calculation method. It should be noted
that Quails and Johnson (40) used this assay
procedure to independently verify the point source
summation calculation method.
A possible simplification of the procedure for esti-
mating the intensity involves injecting the bacterium
stock at a steady rate. The effluent is assayed and the
dose is determined from the log survival ratio/This is
195
-------
Tabla 7-7. Examples of Low Pressure Mercury Arc Lamp Specifications (Courtesy Voltarc Tubes, Inc., Fair-field)
Lamp Watts
Lamp Current, mA
Ultraviolet Output, watts
G36T6L
G36T6H
39
425
13.8
G37T6VH
40
425
14.3
G36T6
36
425
12.7
G64T5L
65
425
26.7
G64T6L
62
425
25.5
(at 100 hrs, 253.7 nm)
Microwatts/cm2 @ 1 meter
Ozone Generation
(approximate gm/hr)
Nominal Length,
Inches
cm
Arc Length
inches
cm
Tube Diameter, mm
Tube Material
120
H .5
L 0
36
91.4
30
76.2
15
H: Vycor
7912
L: Vycor
7910
124
15
37
94.0
31
78.7
15
Quartz
110
0
36
91.4
30
76.2
19
Vycor
7910
190
0
64
162.6
58
147.3
15
Vycor
7910
180
0
64
162.6
58
147.3
19
Vycor
7910
Rated Life, hours
(average, at 8 hrs/start)
7500
7500
7500
7500
7500
Figure 7-21. Schematic of bioassay procedure for estima-
ting dose and intensity.
UV Lamp
7
^-Shield
_ Collimating
Tube
i—Sample
I—-k Mixer
(a) Dosing Apparatus
Dose (Ixt)
(b) Dose-Response Calibration
8
-------
As such, it can be used effectively as a post-construc-
tion performance test or to compare the performance
of competing commercial units during design and/or
bid phases of a facility installation.
The following is an outline of the procedure which
can be used to develop the dose performance curve
for a commercial unit. Example results are given from
an actual test series conducted in response to
specifications for the Bristol, Connecticut plant
expansion; these demonstrate the data generated
from the bioassay analysis.
1. Selection and Culturing of Bacterial Culture
The species selected for the assay should be one
which is relatively easy to culture, identify, and
harvest, and which has a dose-response which
is reproducible and consistent. Bacillus subtilis
spores, which are used in the example on Figure
7-22, have been used on several recent equip-
ment assays. Originally used by Johnson and
Qualls(38), the spores are relatively resistant to
UV, and have been shown to be very consistent
and reproducible within a specific harvest. M.
lutea has also been used successfully and
shows a response similar to that of the coliform
group.
The culture should be harvested in sufficient
quantity such that all necessary dose calibration
and assay work can be accomplished with a
single harvest. The B. subtilis are particularly
suited to this since they can be stored for long
periods and have been shown to retain their
dose-response behavior through this period.
In situations where units are being compared
(such as in a pre-qualification procedure for
several manufacturers), a single organism
should be specified. Additionally, the use of a
single laboratory should be encouraged. It has
not been established that reactors will yield the
same effective dose for organisms with different
dose-response relationships.
It has also been shown that different dose-
response curves are developed lab to lab for the
same organism. The better approach, then, in
order to assure consistency and the ability to
validly compare different units, is to use one
organism generated from a single mixed batch
and to have one laboratory conduct the calibra-
tion and equipment assays.
2. Dose-Response Calibration
A dose-response calibration curve can be de-
veloped using the laboratory collimated beam
apparatus shown schematically on Figure 7-
21 (a). As an example, the apparatus used for the
example bioassay had a G8T5 lamp as the UV
source. All but two inches of the lamp were
shielded. The exposed portion of the lamp was
suspended above a 25 cm long, 5 cm diameter
non-reflective tube. The sample to be exposed
was placed in a petri dish below the tube. The
sample size (20 ml) was sufficient to give a liquid
depth of 1 cm.
The purpose of the tube is to collimate the light,
such that the light reaching the liquid is
perpendicular to the surface. In this manner, the
light can be accurately measured by a radi-
ometer detector.
Prior to exposing the sample, the intensity of the
ultraviolet radiation is adjusted (by movement of
the la mp position) to 100 //W/cm2 at the surface
of the sample. A narrow band, calibrated
detector should be used for these measure-
ments.
During exposure, the sample should be gently
stirred continuously, using an insulated mag-
netic mixer with a micro spinbar. The organism
density should be adjusted to approximately
10s; a buffered water should be used for dilution.
After exposure, samples should be plated im-
mediately, using a culturing medium appropriate
to the organism being assayed. The same
medium should be used for both the dose
calibration and the equipment assay tasks. In
situations where a clean, potable water can be
used for the carrying liquid in the equipment
assay, a non-selective nutrient medium can
generally be used for microorganism enumera-
tion. This was the case for the example bioassay;
total plate count agar (pour plates) were used to
grow the B. subtilis spores.
Five to seven exposure times should be run to
develop the dose-response curve. Thus, at a set
intensity of 100 /uW/cm2, running exposure
times between 50 and 400 seconds would yield
a dose (I x t) range between 500 and 4000
juW-sec/cm2.
The dose runs (series of exposure times) should
be conducted in triplicate, each from a separate
dilution of the stock suspension. In all three
dose runs the controls and exposed samples
should be sampled in triplicate; three dilutions
should then be plated in triplicate. A minimum
of two controls (unexposed) should be run with
each dose run, representing time zero and the
longest exposure period.
197
-------
Figure 7-22. Example of bioassay analysis of commercial UV system to determine dose (By permission of Ultraviolet Purification
Systems, Bedford Hills, New York).
Flow
— Drain
Tank
(b) t p 3" 0 Outlet
12"#
UV
Chamber
3" Inlet-i I
IF%
r
Static
Mixer
Spore
Feed
Thlo
Feed
• 4" Pipe
O
Hydrant
Schematic of Pilot Plant
(a)
0
-0.2
-0.4
-0.6
2-0.8
"5 -1.4
1-1.6
w -1.8
§"-2.0
-2.2
-2.4
-2.6
-2.8
-f- Run 1
• Run 2
Ruh3
100
10
3
W
1.0
10 20 30 40 50 60
Dose (u Watt-second/cm2)
Dose-Response Calibration of B. subtilis
70
80
(c)
100
d~ 80
60
40
Q 201-
50 100 150 200 250 300 350 400
Flow(gpm)
Assay Results of Test Unit
A dose-response calibration curve developed for
the example assay is given on Figure 7-22(a).
The log of the survival ratio (Log N/N0) for the B.
subtilis is plotted against the delivered dose.
3. Test Unit and Experimental Setup
The test unit to be evaluated by the bioassay
should closely simulate the design of the full-
scale system proposed for the treatment facility.
Since the UV equipment is generally comprised
of modules, the test unit need simulate only one
module. In some cases, it may be practical to test
the full scale module itself.
Particular attention should be paid to scaling the
hydraulic design of the full-scale unit. The
parameters of dispersion and the indices de-
veloped from the RTD of a unit can be con-
sidered in specifying the test unit. These were
discussed in Section 7.3.2. Additionally, the test
unit should have a similar aspect ratio (ratio of
198
-------
length to diameter, or cross sectional dimen-
sion), and inlet and outlet designs. In particular,
the inlet and outlet velocities should be equiv-
alent. Although not always specified, it is
recommended that residence time distribution
curves be developed for the test module.
Generally, specifications should require that the
test unit, once its similitude is established, be
tested at the hydraulic loads to be encountered
for the full scale system. This is determined on
the basis of flow per unit lamp, e.g., Ipm/lamp.
The range of flows to be tested should en-
compass the peak design flow anticipated for
the plant. The performance requirement gen-
erally specifies that the system sizing would
meet a desired dose level under peak design
conditions. Other requirements imposed on the
test are that the lamp output be reduced to
simulate end of life conditions; this is generally
considered at 70 percent of the lamps' nominal
output. The lamp output can be reduced by using
a rheostat to adjust the voltage, or by using
lamps which have reached 70 percent (this
should be confirmed by direct measurement,
see Section 7.5.1) of their original output.
The transmittance of the carrying water should
also be adjusted to yield an absorbance coef-
ficient, or percent transmittance, anticipated
under design conditions. This is accomplished
by adding a chemical which will absorb energy
at the 253.7 nm wavelength, but will not
interfere with the test An appropriate coin-
pound is sodium thiosulfate.
Figure 7-22(b) presents, schematically, the
experimental setup to test the UV module for the
example bioassay. The water source was pot-
able water from a hydrant at a wastewater
treatment plant (with appropriate backflow
protection devices). Sodium thiosulfate is in-
jected at a rate needed to yield (by direct
measurement) the desired transmittance.
The microorganism suspension is injected in
similar fashion to yield a desired density level.
Both the thiosulfate and spore suspension are
injected upstream of an in-line static mixer to
assure a homogenous solution before entering
the UV chamber. Flow rates are set and meas-
ured by determining the rate of fill in a large
(1000 L) tank. The tank drains to the primary
clarifiers in this particular case.
4. Experimental Field Test Procedure
Three to four flow rates should be tested in
triplicate; these should, at minimum, bracket
the peak design flow. Once the appropriate
water flow rate is set through the unit, a near
saturated solution of sodium thiosulfate is
metered directly into the water line. The feed
rate is adjusted until the desired transmittance
level is reached in the effluent.
Once the water is adjusted to the desired
transmittance with sodium thiosulfate, the B.
subtilis spore suspension (or other test orga-
nism) which is continuously mixed, is metered
into the line with a second metering pump. The
feed rate is adjusted, in this case, at each flow
setting to yield an influent density of approxi-
mately 104 spores/ml.
The flow with thiosulfate and spore suspension
is continued long enough to allow a minimum of
seven volume changes in the unit before
sampling. The influent and effluent are then
sampled in triplicate using sterile sampling
containers. The influent and effluent sampling
lines should be kept flowing continuously to
assure that the samples taken are representa-
tive of the run being done. An additional sample
istaken of the influent for percent transmittance
analysis.
Samples should undergo immediate (within four
hours) analysis, A minimum of three dilutions
should be plated in triplicate, using the appro-
priate medium. The percent ultraviolet transmit-
tance at 253.7 nm should be measured by
standard spectrophotometric procedures.
The results of the example bioassay are pre-
sented on Figure 7-22(c). The log of the survival
ratio (Log N/N0) are first determined from the
experimental data. These are then used to
determine the effective dose delivered by the
test unit by reference to the dose-response
calibration curve (Figure 7-22(a». This effective
dose is then plotted against the corresponding
flow rate. The example unit was operated at 70
percent lamp output (set by adjusting the
voltage); the water transmittance was 70 per-
cent (at 253.7 nm).
The relationship presented on Figure 7-22(c)
allows one to then determine the flow (or flow
per unit lamp) which corresponds to the mini-
mum desired dose. As discussed earlier, this
has often been used in equipment specification
as a pre-bid or bid qualification requirement.
The bioassay is also used to set specific per-
formance requirements for equipment supplied
to a facility.
199
-------
7.3.3.2 Calculation of the Average Intensity by the
Point Source Summation Method
The calculation approach is suggested as the method
of choice because of its versatility and flexible
application to varying configurations. The technique
used to calculate intensity is the point source
summation method. A brief description is presented
herein; the reader is referred elsewhere for a detailed
discussion of the calculation framework (52).
The point source summation technique was evalu-
ated by Jacob and Dranoff (66) for light intensity
profiles in a perfectly mixed photoreactor and was
first applied to UV disinfection reactors by Johnson
and Quails (38). It presumes that the lamp is a finite
series of point sources that emit energy radially in all
directions. The intensity at a given point in a reactor
would be the sum of intensities from each of these
point sources.
Intensity Attenuation. UV intensity will attenuate as
the distance from the source increases. This occurs
by two basic mechanisms: dissipation and absorption.
Dissipation is simply the dilution of the energy as it
moves away from the source. The area upon which
the energy is being projected is increasing; thus the
energy per unit area is decreasing. This dissipation
can be calculated by surrounding the point source by
a sphere of radius R:
S/(47rR2)
(7-24)
where 1 is the intensity at a distance R in/uWatts/cm ,
R is the distance in centimeters, and S is power
available from the UV source in /M/atts. Thus,
dissipation is seen to attenuate the intensity as the
inverse of the radius squared.
The second attenuation mechanism relates to the
absorptive properties of the medium through which
the energy is transmitted. This is best described by
Beer's Law:
!0exp[-orR]
(7-25)
where I0 is the intensity at a given surface on the
source (/M/atts/cm2), or is the absorbance coefficient
of the medium through which the energy is passing
(cm"1), and R is the distance at which I is measured
relative to the point represented by I0. The absorbance
coefficient reflects the absorbance at the specific
wavelength being emitted; in the case of the low
pressure mercury arc lamps, the wavelength is 253.7
nm.
Combining Equations 7-24 and 7-25 yields an
expression which describes the intensity at a given
distance from a single point source of energy:
This equation serves as the basis for the point source
summation calculation technique. A basic assump-
tion is that a receiver (i.e., a microorganism) passing
through the reactor is infinitely small and is spherical;
by this it can then be presumed that the energy
emitted from any point source element of the lamp
will strike the receiver normal to its surface.
The model analysis also neglects the phenomena of
reflection, refraction, diffusion, and diffraction of light
and assumes that the absorptive properties of the
liquid are independent of the light intensity. The
intensity at a receiver is then the summation of the
intensities from each of the point source elements of
a lamp, (or lamps in a multilamp system). Figure 7-23
is a schematic representation of this calculation. As
shown, the intensity at the receiver location (r, z0) is
the summation of the intensities from each of the
lamp elements:
n=N
I(r,z0) = I
n—1
S/N
exp [-ct(r2+zn2)1/2] (7-27)
where N is the number of point source elements in
the lamp. The value of zn is:
zn = z0 - L(n/N)
(7-28)
Figure 7-23.
Lamp-\ |
NR
•4-
Division
of Lamp
into
Point Sources
Lamp goometry for point source summation
approximation of intensity.
Receiver Location
^Sample Lamp Element
zn = z0-L(n/N)
z = O
I = [S/(477-R2)] exp (-cR)
(7-26)
A practical analysis of intensity in a submerged or
Teflon tube lamp battery system requires that the
calculations be made at numerous receiver locations
within the lamp battery. This is accomplished by
dividing the cross-sectional space between lamps
into an equal-area grid system. The average of the
200
-------
receivers located in the center of equal area grid
elements would then be equivalent to the average
intensity within the total grid area. Within a system,
this grid is expanded to encompass all or a section of
the unit and can be moved about to evaluate boundary
effects and other configurations which may affect the
overall unit average intensity.
The model takes into account the geometry of a
system, the characteristics of the lamps and en-
closures (e.g., quartz or Teflon), and the given UV
absorption properties of the fluid. Since the low
pressure mercury arc lamps are excellent absorbers
of light at the 253.7 nm wavelength, the model
calculations presume that any energy at this wave-
length entering a lamp from a neighboring lamp will
be completely absorbed by that lamp.
Figure 7-24 is presented to illustrate the intensity
field calculated by the point source summation
method. This example shows four lamps, spaced 5 cm
apart. The important note to this is the non-uniformity
of the intensity field and thus, the need to have
turbulent flow, as discussed earlier, such that a
particle will have the opportunity to be exposed to all
intensity levels. I n this fashion, it is appropriate to use
the average intensity computed for the reactor.
Computationally, the point source summation method
is not convenient and is best handled by computer
with the appropriate software. In lieu of this, a series
of solutions have been developed, and are presented
in this manual, which describe the average intensity
for almost any practical lamp configuration which
would be considered by the designer.
7.3.3.3 Nominal Average Intensity Estimates for
Alternative Lamp Configurations
Different lamp configurations will yield different
nominal intensities in the reactor. Calculations have
been performed for a number of designs, and
subsequently reduced to show the nominal intensity
as a function of the UV density of the reactor, and the
wastewater absorbance coefficient.
UV Density. The UV density, D, is defined as total
nominal UV power (at 253.7 nm) available within a
reactor divided by the liquid volume of the reactor:
D = total UV output/liquid volume (7-29)
= UV watts/liter
As an example. Unit 1 in the Port Richmond study
contained a total of 100 lamps. The liquid volume
(internal reactor volume minus the volume occupied
by the quartz sleeves) was approximately 375 liters.
The lamps were the G37T6VH lamps described in
Table 7-7. The nominal output is shown as 14.3
W/lamp; thus, the total UV output is 100 x 14.3 =
1430 W. The UV density of the reactor is then 1430
W/375 liters, or 3.8 W/L.
Obviously, the density will be directly related to the
spacing of the lamps. The closer the spacing, the
higher the UV density of the reactor.
Lamp Array Configurations. Four lamp "arrays" are
considered; these are in common use today and, in
effect, cover almost all practical configurations one
would consider. The only assumptions which are
made are that the lamps are always parallel to one
another and that the single array pattern is continu-
ous and symmetrical throughout the reactor. Both are
appropriate and would be expected from a practical
design. The four arrays are: (1) uniform array; (2)
uniform staggered array; (3) concentric array; and (4)
tubular array.
Uniform Array—
A cross-section of a uniform array is given on Figure
7-25(a). The lamps (with quartz sleeves) are arranged
in even horizontal and vertical rows, with the
centerline spacings equal in both directions.
Staggered Uniform Array—
This is similar to the uniform array, except that the
alternating vertical rows are offset by one-half the
vertical spacing, Sv, as shown on Figure 7-25(b). The
flowpath is typically perpendicular to the lamps; the
staggered effect is designed to influence turbulence.
Concentric Array—
In this configuration, the lamps are arranged in
concentric circles. This is illustrated on Figure 7-
26(a), which is the cross-section of the unit at the
Vinton, Iowa wastewater treatment plant. Typically,
the array is designed to shut off banks of lamps. The
banks are distributed throughout the reactor; this in
effect, alters the UV density of the reactor as lamp
banks are turned on or off. As more and more lamps
are turned off, this can possibly cause very non-
uniform intensity fields. This is not the case with the
uniform arrays.
Tubular Array—
The tubular array describes the Teflon tube systems
in which the lamps are suspended outside and
parallel to a Teflon conduit. This is illustrated on
Figure 7-26(b). The lamps and tubes are stacked
vertically in alternating rows, with equivalent vertical
and horizontal centerline spacing.
When considering the tubular array configuration, it
is important to understand the effect which the
number of vertical rows of tubes in a system have on
the lamp requirement (and the consequent UV
density). The meters of arc required for each meter of
Teflon tube will vary with the number of vertical tube
rows in the system. This is shown on Figure 7-27
207
-------
Figure 7-24. Illustration of the intensity field calculated by the point source summation method.
5.0 cm Spacing
(7.5 cm *E Spacing)
11,000
15,000
Isointensity Lines l/j Watts/cm2)
Absorbance Coefficient = 0.4 cm"1
where n is the number of vertical rows of Teflon tubes
In the unit. The number of lamp rows would be n+1.
Thus, if the system has only one row of tubes, two
rows of lamps would be required, yielding the ratio of
2 meters arc/meter tube. As n becomes greater, the
ratio approaches 1.0. From Figure 7-27, the greater
efficiency is achieved in systems with greater than 10
Teflon tube rows. In computing the UV density for the
UV intensity solutions in this manual (Figures 7-31
and 7-32), the ratio of meters arc to meters tube is
assumed to be 1.1.
Nominal UV Intensity Estimates. Figure 7-28 pre-
sents the estimated nominal UV intensity as a
function of the UV density for the uniform lamp array
configuration. These are shown for a range of UV
absorbance coefficient values between 0.2 and 0.9
crrf1. In similar fashion, solutions are presented in
Figure 7-29 and 7-30 for the staggered uniform array,
and the concentric array, respectively.
Figure 7-31 presents the solutions for the tubular
array, in which the lamp/tube/lamp centerlines are
202
-------
Figure 7-25. Schematic of uniform and staggered uniform
lamp arrays.
Centerline Spacing
(Horizontal)
XTX
c' Centerline Spacing
(Vertical) '
I ~ri-j/Z = Arc Length
r- SH-rr
~\~
Liquid Volume/Lamp
-4-4—
/ Quartz
Sleeve
Lamp (Typical)
(a) Uniform Array
(b) Staggered Uniform Array
15 cm, as shown on Figure 7-26(b). The density for
the system is adjusted by changing the diameter of
the Teflon tube. Current designs use the 15 cm
centerline configurations with a Teflon tube diameter
of 8.9 cm.
An additional analysis is presented on Figure 7-32 for
the tubular array configuration. This presents a
comparison of the solutions for the 15 cm centerline
spacing to the more compact(but equivalent density),
10 cm centerline spacing configuration. The tube
diameter is 6 cm (D = 1.95 W/L); the nominal UV
intensity is shown as a function of the UV absorbance
coefficient. By drawing the lamps in closer to the
Teflon tube, greater energy efficiency is achieved.
The compactness of the tubular array, however, will
be influenced and limited by practical fabrication
considerations.
As shown on Figure 7-33, the greater efficiency is
demonstrated by the quartz arrays. The tubular arrays
are shown to be less efficient in accomplishing an
intensity for a given UV density. This is an artifact of
the physical constraints on the unit fabrication. In
order to arrange the lamps and tubes and still
maintain the system for easy assembly/disassembly
and for access to these components, there will be
limits as to how closely spaced the lamps and tubes
can be.
7.3.3.4 Energy Loss Factors to Adjust the Nominal
Average Intensity Estimate
An important note applies to the solutions presented
in Figures 7-28 through 7-33; the intensity is
calculated at the nominal output of the lamp and
assumes that the quartz sheath or the Teflon tube will
transmit 100 percent of the energy emitted by the
lamp. Thus, the term "nominal" average intensity.
Under actual operation, and for design purposes, this
nominal average intensity must be adjusted to
account for the aging of the lamps, and the conse-
quent reduction in UV output, and for the losses of
energy as it passes through the quartz sleeve or
Teflon tube wall. These losses are due to the quartz or
Teflon wall itself and to fouling of the inside and
outside surf aces. Thus, in order to estimate the actual
intensity under a given set of conditions, it is
necessary to adjust the nominal intensity:
= (Nominal lavg) x (Fp) x (Ft) (7-30)
where:
Fp = theratiooftheactualotitputofthelampstothe
nominal output of the lamps
Ft = the ratio of the actual transmittance of the
quartz sleeves or Teflon tubes to the nominal
transmittance of the enclosures; the nominal
transmittance is presumed to be 100 percent
in the intensity calculation
Procedures are described in a later section to directly
monitor the average values of Fp and Ft for a gjven
system, and are strongly recommended as control
procedures in a plant's O&M program.
When designing a new unit, it is suggested that the
system be designed at an average Fp of 0.7, which is
representative of a lamp inventory output at approxi-
mately one half its operating life. In a sense,
economics come into play here. The UV lamps are
expensive. Their rated life of 7,500 hours can
sometimes be greatly exceeded, but at a cost of
reduced output. The inefficiency may be balanced by
not having to buy new lamps.
The Ft should reflect the anticipated maintenance
input; if the system will be well attended, a reasonable
203
-------
Figure 7-26. Schematic of concentric and tubular lamp arrays.
Outlet
O - Bankl, Lamps 1-40
• - Bank 2, Lamps 41-80
<8 - Bank3, Lamps 81-120
8 - Bank 4, Lamps 121-160
(a) Concentric Array
Lamps
— 15 cm
Lamp
15 cm
Teflon Tubes
[6 cm or 8.9 cm ^ 5 cm
diameter)
15 crn
(b) Tubular Array
204
-------
Figure 7-27. Effect of Teflon system sizing on the power Figure 7-29.
requirement efficiency.
Staggered uniform array intensity as a function
of UV density and UV absorbance coefficient.
•S 2.0
\ "
> 1-4
h
<
j
u
a
i
C 1-2
= 1.0
"a 0.8
•§0.6
*OA
£0.2
ro
cB
^
? 10,000
•s
^J
to
o
Absorbance Coefficient
(cm-')
100% Lamp Output
100% Quartz Transmittanee
8 12 16 20
UV Density (Watts/Liter)
Figure 7-28. Uniform lamp array intensity as a function of
the reactor UV density and UV absorbance
coefficient.
Figure 7-30. Concentric lamp array intensity as a function
of UV density and absorbance coefficient.
Absorbance
Coefficient
(crrf1
Assumes 100% Lamp Output
100% Transmittance
Quartz O.D. = 2.3 cm
20,000
0.25-
3
>.
•ffl
I
-------
Figure 7-31. Tubulararray(15cm
-------
The following discussions will present the key
elements of the wastewater application as they apply
to the design,(and subsequent operation) of a UV
disinfection system:
• wastewater quality parameters. These are the flow
(Q), initial bacterial density (N0), suspended solids
(SS), and UV absorbance coefficient (a);
• estimated bacterial density associated with the
suspended solids (Np). This includes a determina-
tion of the coefficients c and m in Equation (7-13);
• estimated inactivation rate (K). Recall that this is
set as a function of the average intensity, requiring
a determination of the coefficients a and b in
Equation (7-12).
Relevant wastewater characterization data (including
estimates of the coefficients, a, b, c, and m) from
existing wastewater treatment plants are summar-
ized to demonstrate the range of values to be expected
under typical wastewater applications.
Photoreactivation is then discussed, with a presenta-
tion of field data to demonstrate levels of repair one
can expect to occur. Finally, some discussion is given
to sampling considerations and to suggested monitor-
ing programs.
7.3.4.1 Key Wastewater Quality Parameters
The four wastewater parameters which most affect
the design or performance of a UV system are the
flow, initial bacterial density, suspended solids (or
some measure of the particulates in the wastewater),
and the UV absorbance of the wastewater.
Flow Rate. The flow rate is set by design of the main
plant and projections of the hydraulic load to the
plant. In evaluating the design requirements for the
disinfection process, some consideration should be
given to the equalization effects of the treatment
processes before disinfection. This can have an effect
on the sizing of the UV system.
Flow estimates should be for the design year of the
plant. There should also be some knowledge of the
progressive increase in the flows through the design
life of the plant in orderto determine if the system can
be phased in by the addition of modules as the
demand increases. Some consideration should also
be given to the hydraulic load to the unit. The flow
rates important to the design and evaluation of the
system are those typically considered for wastewater
treatment systems:
Annual average daily flow
Maximum 7-day average flow
Maximum 30-day average flow
Peak daily flow
Peak hourly flow
For disinfection, average flows are not critical to the
design sizing; rather they are important to estimating
average utilization of the system for operation and
maintenance needs. Peak flows should be used for
sizing, particularly reflecting diurnal variations.
Initial Coliform Density. The performance of a UV
disinfection system is directly related to the initial
density of the indicator organisms. This is not a
parameter which is generally monitored at a treat-
ment plant, particularly where the disinfection is
accomplished by chlorination. In the case of disinfec-
tion by UV, however, it is critical. Performance is
given by the log of the survival ratio, IM/N0, or by the
number of "logs" the density is reduced.
Expected initial densities cannot be predicted solely
from the type of treatment process preceding the
disinfection process. Order of magnitude levels are
given in Chapter 2 as guidelines. Examples drawn
from several plants (presented in a later discussion)
vary widely and do not correlate well with the types of
systems or plant residence time. It is recommended
that these data be generated before design; effluents
can be analyzed from similar plants in the area, or at
the existing facility if an upgrade or retrofit is being
considered.
Suspended Solids. From the development of the
disinfection model, it is clear that the occlusion of
bacteria in the particulates will have a significant
effect on the design of a UV system. It is recommended
that the suspended solids measurement be used as
the primary indicator to quantify these particulates.
The level of suspended solids in the effluent of a
wastewater treatment facility is, in effect, set by the
design of the plant. This then will limit trie range of
suspended solids concentrations to be considered in
the design of the UV process. A further consideration
is to understand the variability associated with the
effluent suspended solids. As an example, if a plant is
designed not to exceed 30 mg/l on average for any
consecutive 30-day period, the suspended solids
levels it must meet on an annualized basis will likely
be between 10 and 20 mg/l. This can affect the sizing
of a UV facility and determination of its average
operational requirement.
UV Absorbance. The one parameter which is solely
in the venue of UV disinfection is the UV "demand" of
the wastewater. Specific organic and inorganic
compounds in the wastewater will absorb energy at
the 253.7 nm wavelength. This absorbance will affect
the intensity of the radiation within the reactor; in
specific design situations, the level of absorbance will
affect the sizing of a system and possibly the
configuration (spacing) of the lamps. Recall from the
discussions of intensity, and its calculation in a
207
-------
complex reactor, that the final product of these
calculations is the average nominal intensity as a
function of the UV absorbance coefficient.
There are a n umber of ways to express the absorba nee
of a wastewater. First, consider the manner in which
it is measured. The wastewater sample is placed in a
quartz cell (transparent to the 253.7 nm wavelength)
of a given width. A spesctrophotometric measurement
of the absorbance is made of a direct beam of light (at
253.7 nm) which is passed through the quartz cell
containing the liquid. A detector determines the
amount of light which passes through, and by
inference, the amount of light "absorbed" by the
liquid sample can be determined. The output of this
measurement is absorbance units per centimeter, or
a.u./cm. The pathlength is set by the quartz cell;
typically this is 1 centimeter.
The transmittance of the wastewater is a common
parameter used to describe the "demand" of the
wastewater. This can be determined from the ab-
sorbance measurement, and is most often expressed
on a percent basis:
% Transmittance = 100 x 10~'a'u-/cml
(7-31]
Conversely, the percent absorbed is simply 100
percent minus the percent transmittance. The pa-
rameter which is most often used for design purposes,
and is the parameter used within the context of this
manual, is the UV absorbance coefficient, expressed
in base e:
UV absorbance coefficient, a = 2.3{a.u./cm) (7-32)
The unit of the UV absorbance coefficient is cm~1. The
reader is referred to Section 7.1.2 which presents
typical absorbance levels for varying degrees of
treatment.
The single beam, spectrophotometric method for
measuring the UV absorbance of the liquid is the
simplest procedure, requiring minimal effort and
instrumentation. It is important to note however, that
this "direct" UV absorbance measurement assumes
that light which does not pass through the cell and is
not seen by the detector has been absorbed by the
liquid. This is not necessarily the case, especially in
samples which have suspended or colloidal particles
in the liquid. These will cause a portion of the light to
be scattered; the light is still available, but it will not
be seen by the detector since it has been deflected
from its direct path through the quartz cell. Thus, the
direct method tends to overestimate the "true"
absorbance of the liquid.
Johnson and Quails (38) and Scheible et al. (52),
demonstrated that suspended or colloidal particles
will not absorb any significant amount of light energy
and will in fact scatter the light back to the liquid. It
becomes important, therefore, that the absorbance
measurement must in some fashion account for the
scattering effect and give a value representative of
the true absorbance of the liquid. Note that the
procedures to calculate the intensity in a reactor
inherently presume that the UV absorbance coef-
ficient reflects the true absorbance of the liquid.
The Port Richmond study incorporated the use of a
standard accessory to the UV/Visible spectropho-
tometer which would correct the absorbance meas-
urement for the effect of scattering. A sphere, in
effect, surrounds the quartz cell; any scattered light is
absorbed on the surface of the sphere, which
integrates the quantity of light collected and corrects
the absorbance measured by the direct beam de-
tector. This absorbance, referred to by the Port
Richmond report as the "spherical" absorbance
coefficient, is felt to more closely represent the true
absorbance of the liquid, and is more appropriate for
use in estimating the intensity in a reactor.
In the case where the capability to measure the
corrected UV absorbance coefficient is not available,
the UV absorbance coefficient should, at minimum,
be determined on filtered samples by the direct
method. In most cases this will give an approximation
of the true absorbance. The results would be further
improved if this is accomplished by membrane
filtration to remove particles greater than 1 micron in
size. Care should be taken to prewash the filters; in
some instances the filter material itself can contribute
UV absorbing materials.
At Port Richmond, limited testing was conducted on
an in-line continuous monitor of the UV absorbance.
This was a prototype instrument which would con-
tinuously sample the influent of a UV system and
determine the UV absorbance at 253.7 nm; the
instrument would also correct, to some degree, for
scattering. The monitor was found to respond well to
the absorbance of the wastewater. Such a direct
monitor would be useful, in conjunction with the flow
rate meter, in controlling the operations of a UV
system.
The daily average, maximum7-day, and the maximum
30-day average UV absorbance coefficient would be
important to design. Unlike the suspended solids,
which is limited by permit and by the treatment
process, the UV absorbance coefficient is not a
parameter describing treatment goals. It is an artifact
of the wastewater and the treatment of that waste-
water. Thus, an estimate will have to be made of the
UV absorbance coefficient, and its variability, either
by direct measure of the treated wastewater (as in the
case of an existing plant) or a similar wastewater
undergoing the same degree of treatment.
208
-------
In summary, the parameters of primary concern
regarding the characteristics of the wastewater are
the flow rate, UV absorbance coefficient, suspended
solids, and initial coliform density. The peak design
condition for a plant should not necessarily consider
the concurrent occurrence of these parameters as the
worst case condition. If sufficient data are available,
running averages of combinations of these param-
eters should be constructed to determine the maxi-
mum 7-day and 30-day average of the combined
parameter set.
The maximum 7-day coliform density may occur
under average flow conditions, as an example.
Analysis of the data in this fashion may allow a more
realistic system sizing to achieve the desired per-
formance under the anticipated worst case condition.
7.3.4.2 Estimating NP, the Bacterial Density
Associated with the Suspended Solids
The estimate of Np requires generating data under
high dose levels. Recalling earlier discussions, the
premise is that by determining the residual density
after high doses, the residual can be attributed to
'those bacteria which were occluded from the radi-
ation. These are then correlated to the suspended
solids concentration which were present in the same
sample.
In effect, one is generating information which is well
out on the "tail" of a typical dose-response curve (see
Figure 7-14). An example is given on Figure 7-34,
which shows the log effluent fecal coliform (after
exposure) plotted against the log effluent suspended
solids. These data are from the Port Richmond study.
A linear regression analysis yields the expression
(when transformed):
Figure 7-34. Example of deriving an estimate of the residual
fecal coliform density associated with partic-
ulates as a function of suspended solids (55). •
No = 0.26 SS1-96
(7-33)
where Np is in the units colonies/100 ml and SS is in
mg/l. The coefficients 0.26 and 1.96 are the values of
c and m, respectively, from Equation 7-13. The
intercept of the regression on Figure 7-34 is c; m is
the slope.
It is best to determine c and m by direct testing. Values
determined at several plants are presented in a later
section. These had been developed from flow-through
tests. Although batch tests have not been conducted,
one should expect that these would provide accept-
able data under the proper conditions.
7.3.4.3 Estimating the Inactivation Rate, K
It is necessary, again, to generate a specific set of data
to estimate the inactivation rate as a function of the
average intensity. The required samplings are those
in which the operating conditions would not allowfor
maximum kill at some point in the reactor. In other
0.5 1 2 3 4 6 8 10 20 30 4060 100 200 300
Effluent Suspended Solids (mg/l)
words, the apparent dose is low enough such that a
significant coliform density would still be evident in
the exposed effluent. This allows a valid estimate of
the rate of inactivation in a specific sampling utilizing
the initial and final coliform densities. One is, in
effect, operating in the portion of the dose-response
curve (see Figure 7-14), where the relationship is
linear, with a constant slope.
The data are generated by piloting a system on the
subject wastewater. For proper analysis of the data,
the hydraulic characteristics of the unit need to be
determined; these are the dispersion number and the
dispersion coefficient. Once the data subset is
developed, the rate coefficient can be estimated for
each sampling by manipulation of the model equation
(Equation 7-10) to solve for K. Estimates are first
made of Np from the suspended solids data and
Equation 7-13, as described above. This is then
subtracted from the densities measured after ex-
posure:
N' = N-Np
Solving Equation (7-10) for K yields:
K = [u2(P2-1)']/4E
where:
P = 1 - [2E In (N'/No)]/ux
(7-34)
(7-35)
(7-36)
The rate K has the units second"1. Equation (7-35) can
be used to solve for K for each sampling; the inputs
are the observed initial and final coliform densities,
203
-------
the velocity based on the observed flow rate, the x
dimension based on the operating condition of the
unit, and the dispersion coefficient determined for the
reactor. These values of K are then correlated to the
estimated Uvg in the reactor corresponding to the
conditions for each sampling.
The correlation of log K as a function of log lavg for
fecal coliform data generated at Port Richmond are
presented on Figure 7-35 as an example of the rate
analysis. Linear regression analysis yielded the
expression (when transformed):
K = 0.0000145 (lava!
,1.3
(7-37)
where Uvg is the average intensity in /tM/atts/cm2. The
coefficients 0.0000145 and 1.3 are the values of the
intercept a and the slope b, respectively, of Equation
7-12. Values of a and b derived at a number of plants
are presented in a later section.
Figure 7-36. An example for deriving an estimate of the
Inactivtitlon rate for fecal coliformu as a
function of the calculated average intensity
(56).
2
3
OC
20.0
10.0
8.0
6.0
5.0
3.0
2.0
1.0
0.8
0.6
0.5
0.4
0.3
0.2
Unitl
Unit 2
600 1000 2000 4000 10000 20000
Average Intensity, Uvo (//W/cm2)
60000
Batch testing, although not applied or demonstrated
at the time of this publication, can also be used to
derive the coefficients a and b. A clear advantage to
the batch technique would be its independence of
hydraulic considerations.
7.3.4.4 Checking the Coefficients Determined for
Use in the Model
The coefficients a, b, c, and m are specific to a given
wastewater application and reflect the site-specific
sensitivity of the microorganisms to UV(a and b) and
the level to which microorganisms are occluded in the
effluent suspended solids. When these are deter-
mined by direct piloting, it is appropriate to verify their
values by checking against data generated inde-
pendent of the data set used to derive the coefficients.
Influent and effluent data should be collected over a
range of conditions. Using Equation 7-14, calibrated
to the proper coefficients and operating conditions,
the performance (Log N/N0) can be predicted for a
given sample. This can then be compared to the
observed performance. Again, an example of this type
of analysis is presented on Figure 7-36, which shows
the observed versus predicted fecal coliform densities
from the Port Richmond study. An analysis of the data
indicated that the regression line for the observed
versus predicted correlation was not significantly
different from the ideal line in which the slope is 1.0
and the intercept is 0.0. In all, the analysis suggests
that the model correctly responds to the varying
operating parameters of the UV system, and when
properly calibrated, will successfully predict perform-
ance under any matrix of operating and hardware
configurations.
Figure 7-36. An example of the comparison of disinfection
model estimates to Observed effluent fecal
coliform densities (56).
6 -
.a
O
"3
c
• Secondary Effluent
° Primary Effluent
Regression Line
2345
Log Effluent Coliform (Calculated)
7.3.4.5 Summary of Wastewater Data from
Existing Plants and Recent Field Studies
Wastewater quality data were compiled from several
wastewater treatment plants. This was done to give
the reader a perspective on water quality character-
istics, particularly those parameters relevant to the
210
-------
UV disinfection process. The list of plants is sum-
marized on Table 7-8, including a brief description of
the treatment process at the plant and the level of
treatment the plant is designed to achieve. The plants
were selected because full or pilot scale UV evalua-
tions had been conducted at each yielding a consis-
tent set of data relevant to UV.
The Port Richmond plant is listed first; data are given
for both the primary and secondary treatment levels
(52). The Suffern, Vinton, and Eden data are from
special one-month studies (54). The Northfield plant
(59) is an existing full scale facility and is one of the
few facilities that routinely monitors on a frequent
basis. Data from the 1984 disinfection season are
presented. The Tillsonburg plant is a full-scale UV
facility in Ontario; data are presented from a long-
term demonstration study (48). The Northwest Bergen
study entailed a one-year full-scale pilot evaluation
(36).
Five plants are listed from Ontario, Canada: Toronto-
Main, Georgetown, Milton (secondary and tertiary),
Hamilton, and Toronto-Lakeview. They were all sites
of UV pilot plant studies (47). Finally, the last four
plant sites. New Windsor, Newburgh, Suffern, and
Monticello (secondary and tertiary), are from special
pilot studies conducted in 1985 (52). These studies
also investigated the inactivation rate, K, and the
paniculate coliform density, Np, as will be presented
in the following discussions.
Initial Bacterial Density. Average bacterial densities
are summarized on Table 7-9. These are all geometric
means. The fecal conforms are listed in all cases; this
remains the indicator of choice in almost all permit-
ting activities. Other indicators are listed when
available.
As can be seen, there is little consistency in the
density levels, nor is there any obvious correlation to
the level of treatment. The fecal conforms range from
104 to 10s; for the purpose of preliminary sizing
and/or system designs, a reasonably conservative
initial fecal coliform density, N0, would be 2 to 5 x 105
for secondary treatment levels, and 1 to 2 x 105 for
tertiary levels.
Treated Effluent Quality. Table 7-10 presents aver-
age UV absorbance coefficient data and suspended
solids information for each of the plant sites. Where
available, turbidity, COD, and TKN data are also
provided. Of particular interest is the UV absorbance
coefficient; in all cases, the direct measurement
(unfiltered) is provided. In several cases, both the
direct and spherical measurements (unfiltered and
filtered) are presented. Recalling the discussions in
Section 7.3.4.1, the preferred method is the spherical
(unfiltered) since it corrects for scattering and is the
most representative of the actual UV absorbance
characteristics of the wastewater. An alternative is to
use the direct method on a filtered sample; except for
a few cases, this is shown to give a reasonable
approximation of the spherical UV absorbance coef-
ficient.
The direct analysis of an unfiltered sample is certainly
the easier procedure, requiring no sample preparation
or special accessory to the UV spectrophotometer.
Figure 7-37 indicates an excellent correlation be-
tweenthedirect(unfiltered) and spherical (unfiltered)
data from nine plants. Thus, for preliminary design
purposes, it is reasonable to estimate the spherical
absorbance coefficient from the direct (unfiltered)
analysis:
as = 0.6 (aD)
0.64
(7-38)
where as and ao are the spherical and direct
(unfiltered) absorbance coefficients (base e), respec-
tively, with the unit (cm~1).
Inactivation Rate, K. Several plant sites were evalu-
ated to directly determine the inactivation rate, K, as a
function of the UV intensity (52). These used 2 pilot
units; each had 12 lamps and differed only in spacing,
and, therefore, in intensity. The results are presented
on Figure 7-38, including the coefficients a and b
(Equation 7-12). The Port Richmond regression (from
Figure 7-34) is also presented (this also used two
units, each differing significantly in intensity).
There is significant variability among the regression
coefficients presented on Figure 7-38, although the
actual values of K are not as varied. At the lower
intensities (—3000) the average K ranges between
0.2 and 0.6 sec"1, a factor of 3. The K ranges between
1.6 and 3.8 at an lavg of 10,000, a factor of 2.4.
The data from the four plants were combined; the
resulting regression line is also given on Figure 7-38.
The Port Richmond plant is shown to be nearly
equivalent to this combined regression.
At this point, sufficient data are not available to
clearly demonstrate a uniformity in the K rate as a
function of the lavB. For this reason, it is strongly
suggested that these data be generated by direct
testing for specific plant applications. Preliminary
design calculations can use an a = 1.4 x 10~5 and a b =
1.3, based on the results of the Port Richmond and
combined plants regression analysis. The reader
should understand that the lavg is estimated on the
basis of the spherical absorbance coefficient.
Paniculate Denisty, /Vp. The same plant studies
which had evaluated the inactivation rate, also
evaluated, on a limited basis, the coliform density
associated with the suspended solids. These data are
shown on Figure 7-39. As would be expected, there is
211
-------
Table 7-8. Wastewatsr Treatment Plants Which are Sources of Wastewater Characterization Data
Treatment Plant Location
Treatment
Level
Description
Port Richmond WPCF'
Staten Island, NY
Port Richmond
Staten Island, NY
Suffern, NY
Vinton, IA
Eden, Wl
Northfteld, MN
Tillsonburg, Ontario
NW Bergen, NJ
Toronto-Main, On-
tario
Hamilton, Ontario
Georgetown, Ontario
Milton, Ontario
Milton, Ontario
Toronto-Lakeview
Ontario
New Windsor, NY
Newburgh, NY
Suffern, NY
Monticello, NY
Monticello, NY
Secondary
Primary
Advanced
Secondary
Secondary
Secondary
Secondary
Advanced
Secondary
Secondary
Secondary
Secondary
Tertiary
Secondary
Secondary
Secondary
Advanced
Secondary
Tertiary
Step aeration activated sludge, sec-
ondary clarification
High rate primary clarification
Trickling filter; single stage (nitrifica-
tion) activated sludge, clarification
Extended aeration activated sludge,
final clarification
Activated sludge, secondary clarifi-
cation
RBC and secondary clarification
Activated sludge; clarification
Single stage (nitrification) activated
sludge; clarification
Conventional activated sludge
Conventional activated sludge
Conventional activated sludge
Conventional activated sludge
Conventional activated sludge; rapid
sand filtration
Conventional activated sludge
Trickling filters; clarification
Activated sludge; clarification
Trickling filters; single stage (nitrifi-
cation) activated sludge; clarification
Oxidation ditch; secondary clarifica-
tion
Oxidation ditch; secondary clarifica-
tion; sand filtration
considerable scatter, not unlike the variability shown
on Figure 7-34 for the Port Richmond study. The data
tend to fall about the regression line developed for
Port Richmond. This poor correlation can be due to
several factors, including analytical precision at low
levels; site differences; differing particle size distri-
butions; etc. For preliminary design purposes, it is
suggested that the coefficients c a nd m be set to 0.25
and 2.0, respectively.
7.3.4.6 Photoreactivation
The phenomenon of photoreactivation had been
described in Chapter 7.2.2.2. Unique to ultraviolet
radiation, the mechanism involves the repair of
damage caused by exposure to UV, allowing for
subsequent replication of the organism. The enzy-
matic mechanism generally involved requires subse-
quent (or concurrent) exposure to light at wavelengths
between 300 and 500 nm; such light is available in
sunlight and in most incandescent and fluorescent
light sources.
The procedure to quantify the effects of photoreacti-
vation is by the so-called static bottle technique. The
method involves splitting an exposed sample to three
aliquots: the first is set immediately for coliform
enumeration; the second is placed in a bottle opaque
to visible light; and the third is placed in a bottle
272
-------
Table 7-9. Initial Bacterial Density Before Disinfection3 (organisms/100 ml)
Plant
Port Richmond
Port Richmond
Suffern
Vinton
Eden
Northfield
Tillsonburg
NW Bergen
Toronto-Main
Hamilton
Georgetown
Milton
Milton
Toronto-Lakeview
New Windsor
Newburgh
Suffern
Monticello
Monticello
Treatment
Level
Secondary
Primary
Advanced
Secondary
Secondary
Secondary
Secondary
Advanced
Secondary
Secondary
Secondary
Secondary
Tertiary
Secondary
Secondary
Secondary
Advanced
Secondary
Tertiary
Total
Coliforms
1,020,000
31,700,000
—
-
-
-
53,000
190,000
1,300,000
1,300,000
77,000
140,000
77,000
160,000
-
-
-
-
-
Fecal
Coliforms
361,000
12,500,000
95,500
89,100
44,700
75,500
15,600
48,000
120,000
180,000
9,200
23,000
13,000
14,000
1,900,000
77,000
60,800
1,124,000
736,000
Fecal' .
Streptococci
-
-
25,120
-
21,400
-
1,240
-
33,000
15,000
3,500
1,900
820
2,500
•
-
-
'
-
Escherichia
coli
-
-
. —
.
-
-
-
-
110,000
30,000
8,300
22,000
6,600
12,000
.
-
-
-
-
Pseudomonas
aeruginosa
-
-
—
-
-
-
-
-
2,500
1,300
29
140
40
80
-
-
-
-
-
"Geometric Means; data from analysis of grab samples.
Table 7-10. Treated Effluent Characteristics From Several Wastewater/Treatment Plants"
Average UV Absorbance Coefficient
(base e) (crrr1 at 253.7 nrtj)
Plant
Port Richmond
Port Richmond
Suffern
Vinton
Eden
Northfield
Tillsonburg
NW Bergen
Toronto-Main
Hamilton
Georgetown
Milton
Milton
Toronto-Lakeview
New Windsor
Newburgh
Suffern
Monticello
Monticello
Treatment
Level
Secondary
Primary
Advanced
Secondary
Secondary
Secondary
Secondary
Advanced
Secondary
Secondary
Secondary
Secondary
Tertiary
Secondary
Secondary
Secondary
Advanced
Secondary
Tertiary
Direct
(Total)
0.466
0.865
0.290
0.331
0.391
0.378
0.250
0.390
1.07
0.565
0.348
0.366
0.271
0.657
0.894
0.739
0.454
1.627
0.687
(Filtered)
0.404
0.747
0.273
0.275
0.354
-
-
-
-
-
-
-
-
-
0.705
0.495
0.379
0.589
0.440
Sphetfical
(Total)
0.372
0.593
0.282
0.296
-
-
- -.
-
- •
-
-
-
-
-
0.578
0.518
0.374
0.766
0.416
[Filtered)
0.358
0.533
0.271
0.260
-
-
-
• . - •
-
-
.. -
-
-
-
0.500
0.444
0.362
0.352
0.328
Suspended
Solids
(mg/l)
14.3
80.9
8.3
8.3
33.2
12.2
6.1
6.4
26.8
12.2
5.8
9.1
2.6
6.4
31.6
26.7
7.3
72.4
13.3
Turbidity
(NTU)
4.0
25.7
4.4
-
8.6
-
1.9
3.7
6.3
3.2
2.0
2.0
1.2
3.2
-
-
—
-
-
COD
(mg/l)
44.5
134.0
34.4
28.9
39.0
-
14.6
28.0
92.0
39.1
34.8
38.8
23.2
49.1
-
-
—
-
-
TKN
(mg/l)
6.9
-
—
-
-
-
1.2
14.0
28.6
26.5
14.5
9.2
4.3
3.9
-
-
—
-
-
"Arithmetic means; data from analysis of grab samples.
213
-------
Figure 7-37. Correlation to estimate the spherical absorb-
ance coefficient from direct unfiltered absorb-
ance coefficient.
o
1
2
"E
£
0>
8
3
2.0
1.5
1.0
0.8
0.6
0.5
0.4
S 0.3
8 0.2
Q.
W
0.15
as = 0.6
Plant No.
17
Averages from 9 Wastewater
Treatment Plants (see Table 7-10)'
0.2 0.3 0.4 0.6 0.8 1.0 1.5 2.0
Direct Absorbance Coefficient, aD; Unfiltered (cm"1)
transparent to visible light. The two bottles are then
held for sixty minutes (this time is not standard, it can
vary from one half hour to three hours), at 20°C,
exposed to sunlight. The samples would then be set
for coliform enumeration; the opaque bottle is the
"dark" sample, and the transparent bottle the "light"
sample. Holding the visible light exposure at constant
temperature (20°C) is not critical. In situations where
it is desired to monitor photoreactivation seasonally,
it is more appropriate to suspend the two bottles just
belowthe surface of the plant effluent or the receiving
water. In this fashion, the degree of photoreactivation
is being monitored under current temperature condi-
tions.
The repair mechanism has a dependency on tempera-
ture (36). The analysis suggests that at 10°C, a two-
fold increase in fecal coliform density will be caused
by photoreactivation. If the temperature is approxi-
mately 20°C, a ten-fold (one log) increase in the
effluent density is observed.
A series of tests were also conducted as part of the
Port Richmond study to evaluate the impact of
photoreactivation. Again, the tests centered on the
indicators total and fecal coliforms, and utilized the
static bottle technique. The resultant data are pre-
sented on Figure 7-40, which is taken from the Port
Richmond report. Total and fecal coliforms at time 0
and 60 minutes (light) are plotted as a function of the
loading parameter, Q/W, which is the ratio of flow
(Lpm) to the UV output of the system (Watts at 253.7
nm). The suspended solids averaged 11.8 mg/l during
Figure 7-38. Comparison of inactivation rate estimates
from several wastewater treatment plants.
a(x1Q-6)
Newburgh 1.08
1.29
New Windsor 0.00061
Suffern 0.004
Monticello 0.69
(Unfiltered)
Monticello 10.0
(Filtered)
2.2
1.92
1.4
1.09
Combined
Port
Richmond
1.387
1.45
1.28
1.3
lavg Based on Spherical
Absorbance Coefficient
6.0
4.0
3.0
2.0
1.0
0.8
0.6
0.4
0.3
0.2
0.15
1500 2000 3000 4000 6000 10000
Average Intensity, Ug (^ W/cm2)
214
-------
Figure 7-39. Estimation of Np from several plants.
E
§3
E>
o_
"E
8
.75 2
c
<0
a 1
<»
jo
3
o
CD
Q.
C
<1>
.= 0
CD
3
-1
I I
T Newburgh
& New Windsor
I I I
ND = 0.26 SS1-96
I 1
From Port Richmond
(see Figure 7-34)
o Suffern
• Monticello(Unfiltered)
x Monticello (Filtered)
4 +
x x
J J_
I
I
I
I I
10 20 30 40
Effluent Suspended Solids (mg/l)
60 80 100
200
this period and the average spherical absorbance coef-
ficient was 0.448 cm"1. The temperature was approx-
imately 23°C during the test period. As shown on the
figure, the photoreactivation mechanism causes an
increase of approximately 1.3 logs in either the total
or fecal coliform. The increase is relatively constant,
regardless of the applied dose. This is expected, since
the degree of photoreactivation is independent of
dose.
At present, the effects of photoreactivation are not
directly addressed in most state permitted activities.
Thus, it is appropriate to minimize the effect in the
sampling and analysis of the exposed effluent. The
degree to which this phenomenon exists among the
pathogenic organisms is not fully understood; as an
example streptococci do not photorepair, while
Shigella do exhibit the ability. Viruses generally
cannot photorepair, except in cases where the host
cell can photoreactivate.
If the permitting agencies require that the photorepair
phenomenon must be addressed in assessing the
performance of a UV disinfection process, the design
of the process can accommodate such a requirement.
Based on the results of previous studies, it is also
appropriate to assume that total and/or fecal coli-
forms would still serve as adequate indicators of
pathogen activity. The design of the system must now
address the inactivation of an additional fraction of
the initial coliforms in order to meet guidelines after
photorepair takes place. The critical design period is
during the warmer temperature, higher sunlight
intensity summer months. The system should gener-
ally be designed to.accomplish approximately one
additional log reduction. Thus, if it is determined that
a 3-log reduction will be required to meet the permit
levels, the system should be designed to accomplish a
4-log reduction to account for the effect of photo-
reactivation.
7.3.4.7 Wastewater Sampling Considerations for
Design
The design and performance monitoring of a disin-
fection process is based on the measure of bacterial
density. These are typically coliforms or fecal strep.
Sampling for these can be accomplished only on a
grab basis. Furthermore, the sampling for coliforms
must typically be accomplished during daylight hours,
often set between 10 a.m. and 3 p.m. This time period
215
-------
Figure 7-40. Photoreactivation effects for total and fecal
coliform at Port Richmond (55).
-1
I-3
-5
-6
0
-1
-2
-5
-6
-7
too (Light)
to
Total Coliform
Unit 1
o to
• teo
& to
I
I
Unit 2
I
12345
Flow/UV Output (LPM/Watts)
12345
Flow/UV Output (LPM/Watts)
is considered the maximum loading period for a
treatment plant, and reflects the maximum density
levels with regard to the disinfection process. With
this in mind, it is appropriate that the data that are
generated to characterize the effluent for suspended
solids, UV absorbance, initial coliform density, and
flow should also be collected as grab (or short-term)
analyses, and should correspond to the time of day in
which the system is to be monitored for disinfection
performance.
Twenty-four hour composites, which are typically
collected at a plant, will not directly reflect conditions
under which the system should be designed. At
minimum, sufficient data should be generated to
understand the variability of these parameters during
a diurnal period. There) is an additional benefit which
can be gained from this information. The data will
likely indicate significant improvement in absorbance
and lower initial densities and suspended solids
during the off-peak, early morning hours. The system
can be adjusted to account for this, potentially
resulting in energy savings.
The data base developed for design should reflect
analyses of grab samples taken during the peak load
hours. The parameters which must be monitored
include suspended solids, coliform density, and the
UV absorbance coefficient. Secondary parameters
useful to characterizing a wastewater for design
purposes are grease/oil, iron, and hardness. These
will be important in considering the cleaning re-
quirements of a system.
7.4 UV Disinfection System Design
Example
The preceding sections presented the various ele-
ments important to the design of the UV process.
These considered the hydraulic design of a reactor,
the intensity of radiation, and the wastewater quality
for the specific application. All were related to a
design protocol based on a disinfection model (Equa-
tion 7-10).
This section presents a design example to demon-
strate the use of the design protocol. It does not:
attempt to provide comprehensive solutions, but
rather the procedures and calculations which the
designer can use for specific applications. This can
encompass several situations: to design a new
reactor; to determine the adequacy of a proposed
reactor design (e.g., by equipment manufacturer);
and/or to evaluate the capacity and design adequacy
of an existing system.
7.4.1 UV Disinfection Design Example
The example presented in this section of the manual
is intended to illustrate the design considerations
involved with development of a UV disinfection
system. The wastewater treatment design informa-
tion for the example design problem is shown inTable
7-11. The plant is an air-activated sludge plant
located at an elevation of 1,070 m (3,500 ft) above sea
level. The influent to the UV disinfection system is the
effluent from the secondary clarifiers.
For this example, we are assuming that the waste-
water data (fecal coliform and UV absorbance)
represent grabs taken during the peak diurnal period
(see discussion in Section 7.5.4). The disinfection
portion of the plant will operate under gravity flow.
There are no area constraints and the plant treats
domestic wastewaters only. The plant's permit calls
for year-round disinfection.
7.4.2 Setting Design Conditions and
Parameters for Equipment Sizing
From the disinfection model (Equation 7-10), the
information needs, aside from wastewater character-
istics described earlier, are an estimate of the
dispersion properties of the proposed reactor con-
figuration (i.e., dispersion coefficient and the disper-
sion number), the inactivation rate, K, and the
coliform density associated with the particulates, Np.
216
-------
Table 7-11. Example UV Disinfection System Design
Average Daily Wastewater Flow 28.4 mL/d (7.5 mgd)
Peak Daily Wastewater Flow 56.8 mL/d (15.0 mgd)
NOTE: The daily peak flow rate will not exceed 15.0 mgd because
of storm flow equalization facilities.
Start-up Daily Average Wastewater Flow .. 13.2 mL/d (3.5 mgd)
Start-up Peak Daily Wastewater Flow 28.4 mL/d (7.5 mgd)
Average Effluent BOD/TSS 15/15 mg/l
Maximum Daily Effluent BOD/TSS 30/30 mg/l
Design Required Effluent Fecal Coliform
Weekly Maximum Limitation
(Geometric Mean) 400 per 100 ml
Average Monthly Limitation
(Geometric Mean) 200 per 100 ml
Disinfection System Influent Fecal Coliform
Geometric Mean Concentration (Daily) 500,000/100 ml
Maximum Concentration (7-day Mean) 2,000,000/100 ml
Maximum Concentration (30-day Mean)... 1,000,000/100 ml
UV Transmittance (% at 253.7 nm)
Daily Average 70%
Minimum 30-day Average 65%
Minimum 7-day Average 60%
7.4.2.1 Model Coefficients
The hydraulic characteristics are a direct function of
the reactor configuration, particularly the lamp
placement and spacing, and the reactor's inlet and
outlet design. This can best be characterized by direct
testing of full-scale modules, or hydraulically scale-
able pilot modules. Alternatively, these data can be
required from the equipment manufacturer as a bid
and/or warranty specification. The procedures for
these analyses have been described earlier. Specifi-
cally, the information should encompass the follow-
ing:
a. Residence time distribution curves developed at
several flow rates. These should encompass, at
least, the minimum, average, and maximum
design flows for the system.
b. Head loss measurements, again over a range of
velocities (i.e., flow rates).
Calibration of the model requires direct determination
of the inactivation rate as a function of the intensity
and an estimate of the residual coliform density
associated with the suspended solids. These are the
coefficients a, b, c, and m in Equation 7-14. They will
be site specific and will need to be determined
experimentally. As more field experience is gained
with the application of UV, these coefficients may be
found to cluster about certain levels relative to the
type of wastewater and the degree of treatment.
Estimates of these coefficients for several plants
were summarized in Section 7.3.4.5. These may be
used as approximations for first-cut design estimates;
it is recommended at this point, however, that direct
testing be conducted to verify and/or refine these
estimates.
7.4.2.2 Testing Requirements
Testing need be directed only at the data which are
necessary for the model calibration. Demonstration
of long term performance at a given loading or an
evaluation of O&M needs over time are not issues
which can be effectively resolved by limited piloting or
lab tests. These elements of a system evaluation are
best answered by observation of existing full scale
facilities and the experiences of engineers and
operators directly involved in their design and opera-
tion.
The tests should incorporate the ability to obtain data
at two significantly different intensity levels. If direct
piloting is conducted, this may best be accomplished
by using two units which differ significantly in lamp
spacing, such that the UV density, and consequently
the intensity, differ significantly. The sizing of the
systems does not need to be large, although it is
recommended that the units should have 10 lamps
(0.75 arc lamps would be sufficient) at a minimum.
There are no restrictions on the configuration of the
lamps relative to flow (e.g., parallel or perpendicular),
or in the arrangement of the lamp array (the lamps
should be parallel to one another, however).
The wastewater effluent should be piped to the unit
and there should be the capability to vary the flow
rate. Flow rates must be accurately measured. The
sampling and analyses will center on measurement
of the influent and effluent bacterial density (typically,
this will be total and/or fecal coliforms), the UV
absorbance coefficient, and the suspended solids.
The pilot unit lamp configuration would first require
evaluation by the point source summation method to
calculate the intensity in the unit as a function of the
UV absorbance coefficient. This would be similar to
the relationships presented in Figures 7-28 through
7-32, depending on the array configuration. During
the term of the pilot study, direct measurements
should be made to determine the~output of the units'
lamps and the transmittancy of the quartz and/or
Teflon enclosures. These tasks can be accomplished
by methods which are described in Section 7.5.1. The
information on lamp output and quartz transmittance
would be used to adjust the calculated intensities.
The adjusted intensity would also be affected by the
UV absorbance coefficient measured at the time of
sampling.
The hydraulic characteristics of the pilot units would
then have to be defined directly by running tracer
analyses. The procedures were described in Section
7.3.2, including the analysis of the resulting data. The
required information would be an estimate of the
dispersion coefficient, E, and an estimate of the
actual detention time (mean) versus the theoretical
detention time. If the mean detention time is signifi-
cantly different from the theoretical detention time,
2/7
-------
then this measured value should be accounted for in
the analysis of the system by adjusting the effective
volume.
The procedure for determining Np was discussed
earlier. Samplings should be conducted under very
high dose levels in the pilot units (high exposure
times); the premise is that the residual density
measured after such a high apparent dose is attribu-
table to the bacteria being protected by occlusion in
the suspended solids. Linear regression analyses of
the log effluent residual coliform density as a function
of the log of the effluent suspended solids will yield
the coefficients c (intercept) and m (slope). It is
important to evaluate the units in this fashion over a
significant range of suspended solids concentrations.
It may be necessary to artificially adjust the waste-
water to accommodate this requirement.
Influent and effluent data should be collected at flow
rates which are high, yielding low "apparent" doses
in order to determine the rate of inactivation; this
requires that the exposure time be sufficiently low to
allow significant bacterial density levels in the
effluent. The rate, K, is then determined by solution of
the model expression. A linear regression analysis of
the log K as a function of the log intensity will then
yield the intercept and slope, which are the coef-
ficients a and b, respectively.
Testing can also be conducted on a bench-scale basis
to determine the UV coefficients. Although proce-
dures have not been reported in the literature, simple
batch test methods can be used. This, in turn, would
greatly simplify the testing requirements described
above. Care should still be taken, however, in devising
the experimental apparatus. In order to simulate the
high intensity levels; experienced in the multi-lamps
full-scale units, the bench-scale batch units should
use several lamps to attain higher intensities. The
collimated beam apparatus in Figure 7-21 would not
be adequate as shown; intensities generally do not
exceed 1,000 /M//cm2, even with modifications of
the arrangement.
7.4.3 Assumptions for the Design Example
Given the protocol presented in the earlier discus-
sions, the disinfection model can be used to determine
the optimum design for a given application. The
model approach allows the testing of several design
scenarios and any number of unit configurations. At
minimum, it allows the designer to evaluate directly
the systems proposed by manufacturers.
The primary design objective (and operating goal) is to
maximizethe loading to the system while still meeting
performance goals. For the UV disinfection process,
this UV loading is defined as the ratio of the flow, Q, to
the nominal UV wattage (at 253.7 nm) of the reactor,
Wn:
UV Loading = Q/Wn = Lpm/UV Watt, nominal
Additionally, we will define the performance of a
reactor as the log of the survival ratio, Log N/N0.
Thus, our goal is to design a system which can handle
the maximum loading, Q/Wn, and meet the desired
Log N/No.
At the start, let us make the following assumptions.
The model coefficients, based on direct testing, are:
a = 1.45x10"5
b = 1.3
c = 0.25
m = 2.0
Further, the assumptions we will make regarding the
reactor are:
a. Quartz system with a uniform lamp array. This
was described in section 7.3.5.3 and schemat-
ically presented on Figure 7-25(a).
b. The centerline spacing will be 6.0 cm. The
average nominal intensity is presented on Figure
7-28 as a function of the UV absorbance
coefficient.
c. The lamps will be G64T5, or equivalent (see
Table 7-7). The lamps will have insignificant
transmission at 185 nm, in order to minimize
the production of ozone. This ozone is generated
in the air gap between the lamp and the quartz
sleeve; the ozone absorbs energy at the 253.7
nm wavelength, resulting in attenuation of the
UV energy before it can reach the liquid.
d. The lamps will he..] .6-m long with an effective
arc length of 1.47 m; the nominal UV output is
approximately 18.2 W/m arc.
e. Each lamp is sheathed in a quartz enclosure
with an outer diameter of 2.3 cm.
f. The lamps will be configured axially parallel to
one another; the flow path will be perpendicular
to the lamps.
g. The values of the energy loss factors, Fp and Ft
(see Sections 7.3.3.4 and 7.5.1) are set at 0.8
and 0.7, respectively.
h. The maximum allowable headloss through the
lamp battery is set at 40 cm. This is exclusive of
the entrance and exit losses for the reactor.
Regarding the wastewater characteristics, several
adjustments are made to the Table 7-11 parameters
218
-------
to reflect diurnal and maximum average design
conditions.
For convenience, the problem is restated. The treat-
ment plant is a conventional activated sludge facility
with the following effluent:
Maximum 30-day average BOD5 =15 mg/l
Maximum 30-day average SS = 15 mg/l
Maximum 7-day average BOD5 = 30 mg/l
Maximum 7-day average SS = 30 mg/l
Maximum 30-day average
fecal coliform = 200 org/100 ml (GM)
Maximum 7-day average
fecal coliform = 400 org/100 ml (GM)
The anticipated design hydraulic capacity of the plant,
will be:
Average daily flow (dry weather) =
7.5 mgd (28,000 Lpm)
Peak daily flow = 15.0 mgd (56,000 Lpm)
These flows are anticipated at 5 years; the average
daily flow is expected to be 3.5 mgd at startup. The
relevant wastewater characteristics are:
Average daily fecal coliform = 5 x 10s org/100 ml
UV transmittance (% at 253.7 nm)
Daily average = 70 percent
Minimum 30-day average = 65 percent
Minimum 7-day average = 60 percent
Other wastewater quality characteristics relate to the
variability of the parameters; these can be established
by the analysis of data collected over an extended
period of time:
Ratio of maximum 7-day average flow/average
daily flow = 1.25
Ratio of maximum 30-day average flow/average
daily flow = 1.1
Ratio of maximum hourly flow/average 24-hour
flow = 1.3
7.4.4 Design Sequence
The following steps will comprise the sequence of
calculations for the design example:
1. determine UV density, D,
2. establish intensity as function of D; adjusted for
loss factors,
3. establish inactivation rate, K,
4. set hydraulic parameters to accommodate dis-
persion and headless limitations,
5. establish UV loading-performance relationship,
6. establish performance goals, and
7. reactor sizing.
7.4.5 Design Example
The following calculations demonstrate the procedure
for sizing the UV system. The wastewater character-
istics have been given on Table 7-11; the design
criteria and assumptions are discussed in Section
7.4.3 and 7.4.4.
Step 1—Reactor UV Density
The liquid volume per lamp (see Figure 7.25(a)) is
computed:
Vv/Lamp = (S2z) - [(7rd2/4)z]
(7-38)
where S is the centerline spacing (cm), z is the lamp
arc length (cm), and dq is the diameter of the quartz
sleeve (cm). For the uniform array, with
S = 6.0 cm
z = 147 cm
dq = 2.3 cm
Vv/lamp = [(6.0)2(147)]-[^2.3)2 147/4]
Vv/lamp = 4700 cm3 (4.7 liters)
The UV density, D, is calculated from Equation (7-29):
D = (1.47 m arc x 18.2 W/m arc)/4.7 liters
D = 5,7 W/liter
Step 2—Intensity
The nominal average intensity can then be estimated
from Figure 7-28 (uniform array) for this density and
the anticipated wastewater conditions, i.e., absorb-
ance coefficients. These are summarized on Table
7-12. The percent transmittance is first converted to
the UV absorbance coefficient, a(base e). The nominal
average intensity is then determined from Figure 7-
28. Note that the absorbance coefficient information
has been assumed to be derived from measurements
corrected for scattering (ors-Total).
This nominal average intensity must then be adjusted
to account for the anticipated average lamp output in
the reactor and the minimum average transmittance
of the quartz sleeves. Recalling Fp = 0.8 and Ft = 0.7:
Uvg = Nominal Iav8 x 0.8 x 0.7
The adjusted lavg values are given in Table 7-12 to
reflect these conditions.
Step 3—Inactivation Rates
Once the adjusted intensity values are determined,
the inactivation rates can be estimated from Equation
7-12, with the coefficients a and b determined earlier.
The rates are given on Table 7-12 for each of the
design conditions.
219
-------
Table 7-12. Estimate of Intensity and Rate, K for Design Example
Design Condition
Average Daily
Maximum 30-Day Average
Maximum 7-day Average
UV
Transmittance
at 253.7 nm
(%)
70
65
60
UV
Absorbance
Coefficient
(cm-1)
0.35
0.43
0.51
Nominal"
Average
Intensity
(juW/cm2)
17300
15100
13000
Adjusted"
Average
Intensity
(/M//cm2)
9700
8450
7300
Inactivation0
Rate (sec'1)
K
2.21
1.85
1.53
"From Figure 7-28 for a density of 5.7 W/L
"Assumes an Fp = 0.8 and an Ft = 0.7
ca = 0.0000145; b = 1.3, Equation 7-12.
Step 4—Set Hydraulic Rates
Recalling the discussions on hydraulics (Section
7.3.2), the maximum performance would ideally be
accomplished in a perfect plug flow reactor. Since
there will always be the non-ideal case, the goal is to
design a plug flow reactor with a low dispersion
number, d. As was also discussed, this dispersion
number must be reconciled with the headloss in-
curred by forcing the d to be low.
Scheible et al. (54) had estimated the c( in Equation
7-23 for headloss to be between 0.00017 and
0.00023 secVcm2, based on direct head loss meas-
urements for quartz units with a flowpath perpen-
dicular to the lamps. This estimate would apply only to
this type of configuration. There is little direct data
available to determine Cf for a variety of configura-
tions. It is very important that these data be generated;
the designer should at least specify this in the
equipment specs.
It is suggested that a conservative value of ct =
0.00025 sec2/cm2can be used in estimating headloss
in most quartz reactors where the flow path is
perpendicular to the lamps. One should understand
that this applies only to the lamp battery itself and
does not take into account losses from pipe inlets,
stilling walls, etc.
The example plant is to operate by gravity with the
maximum allowable headloss duetothe lamp battery
set at 40 cm. We can use this to set a practical design
goal with regard to the dispersion number. From
Equation 7-23:
lu= c,(x)(u)2
40 cm = 0.00025 (x) (u)2
this implies that:
x a 160,000 (u)~2
Consider the product, ux. We can estimate u and x
over a range of ux values which will keep the hL below
40 cm. As an example, at:
ux= 10,000 cmVsec
substitute Equation 7-42:
10000
(up =
160000
= 0.0625
or u = 16 cm/sec
and x ss 10,000/16 = 625 cm
Figure 7-41 presents similar solutions; these are all
to yield an hi. < 40 cm for the example system.
Figure 7-41. Example of calculating the limiting U and X on
the basis of head loss (flowpath perpendicular
to the lamps).
(7-42)
S1
4.5
4
x
3
o>3.5
3
3
2.5
2
Log x
1.5 2 2.5 3 3.6 4
N
_ hu = Cf
Maxir
X
\
x)(u)2an
num ht =
C) =
X
^x
d x = 1 60,(
40cm
0.00025 c
\
K
joo (ur2
mVsec2
^
X
\
^^
\
0 0.5 1 1.5 2 2.5 3
Log u
We have defined d as:
ux
A dispersion number should then be approximated for
design purposes. From earlier discussions, a practical
range of d is 0.03 to 0.05. If we consider d = 0.03,1:his
implies a plug flow reactor with low to moderate
220
-------
dispersion. The components of d are the dispersion
coefficient E, the velocity u, and the characteristic
length, x.
Little direct testing information is available to select a
dispersion coefficient as a function of the selected
reactor configuration. Procedures have been given to
develop the RTD curve for a given unit; the d and E can
be estimated from these data. For this reason, it is
strongly recommended that it be developed directly,
or that the equipment manufacturer supply direct,
certified, test data from hydraulically scaleable units.
In the long term, as more information is developed, it
may be possible to develop empirical relationships
which will give reasonable approximations of E. To
date, direct testing on quartz units has yielded an
estimated E between 10 and 500, with values
typically between 50 and 200 cmVsec. Little data is
available on the Teflon tube configuration; these
indicate a value between 10 and 50 cmVsec.
Although Ewill likelyvarywithflowrate(i.e., velocity),
it is sufficient to consider one E; this should be
representative of the maximum flow conditions. For
purposes of design, one should check that the sizing,
based on a selected, d, u andx, implies an E less than
300 cmVsec.
Let us select a characteristic length, x, of 200 cm for
the UV unit (this can be varied in evaluating alterna-
tive design configurations). From Figure 7-41, this
implies a ux of 5,600 cmVsec and a u of 5,600/200 =
28 cm/sec. Maintaining u below 28 cm/sec will
assure an hi. below 40 cm at peak flow.
Recall that we set a d of 0.03 as a design guideline:
d = 0.03 =
ux 5600
This implies an E of approximately 170 cmVsec,
which falls within acceptable limits.
In summary, as the first approximation of the system
design, we are setting:
d = 0.03
x = 200 cm
u = 28 cm/sec
E - 170 cmVsec
Step 5—UV Loading
Recall the disinfection model (Equation 7-10):
N
ux
4EK
The next step in the design procedure is to develop the
relationship of the reactor performance (Log N/No) as
a function of the maximum UV loading (Q/Wn), based
on the above relationship, and the assumptions
discussed earlier.
Before this, one should understand that for a given
loading, there is an equivalent nominal exposure
time, tn:
At the 6-cm spacing for the uniform array the liquid
volume, Vv, associated with the lamp and quartz
sleeve is 4. 7 liters. Thus, for the reactor configuration
we are considering for this example:
Vv
Wn
and
4.7 L
1.47 m arc x 18.2 Wn/m arc
t _ (0.176)
In -
•= 0.176
(Q/Wn)
Calculations are summarized on Table 7-13 for the
design conditions. The log (N'/N0) values are plotted
on Figure 7-42 as a function of the Q/Wn. Understand
that the performance is based on the non-particulate
effluent fecal coliform density, N'. The particulate
fecal coliform density, Np, is additive. The velocity is
also plotted as a function of the loading; the limiting
velocity is based on the maximum headless.
Figure 7-42. Predicted performance as a function of loading
for design example.
-1
-2
^
o> -5
1-6
10
0
1.0 2.0 3.0 4.0 5.0
Maximum Loading, Q/Wn (Lpm/W)
Uniform Array, 6.0 cm *£ Spacing
Solutions at:
d = 0.3
x = 200 cm
E = 170 cmVsec
D = 5.7 W/L
Fp = 0.8
F, = 0.7
6.0
221
-------
Table 7-13. Calculations of Performance on the Basis of Loading for the Design Example
Log N'/No
UV
Loading,
Q/Wn
(1 pm/Wn)
0.5
1.0
1.5
2.0
3.0
4.0
Nominal"
Exposure
Time, tn
(seconds)
21.1
10.6
7.04
5.38
3.52
2.64
Characteristic
Length, x
(cm)
200
200
200
200
200
200
Velocity"
u
(cm/sec)
9.5
18.9
28.4
37.2
56.8
75.8
Daily
Average
k = 2.21 sec"1
-7.8
-6.2
,-5.0
-4.2
-3.1
-2.4
Maximum
30-Day
Average
K = 1 .85 sec"1
-7.0
-5.4
•-4.4
-3.6
-2.6
-2.0
Maximum
7-Day
Average
K = 1 .53 sec"1
-6.2
-4.8
-3.7
-3.1
-2.2
-1.7
X = [(0.176)/(O/Wn)]x60 sec/min
"u = x/tn
No
, = exp[.
2E
{1 - (1
where E = 170 cmVsec, x = 200 cm, and N' = N-NP
(see Equation 7-34)
Step 6—Establish Performance Goals
Returning to the plant design, it is necessary to
determine the performance goal for the system design
conditions. These Cpalculations are summarized on
Table 7-14. The design values for N0 are set to reflect
the variability of the data base, as given on Table
7-11. Regarding the suspended solids, in order to
meet the maximum 30-day permit level, it is neces-
sary to achieve an overall average no more than 50 to
70 percent of the maximum 30-day average. For this
example, then, the daily average SS is set at 10 mg/l.
Step 7—Reactor Sizing
Table 7-15 presents a summary of the reactor sizing
calculations for the design examples. The lava and
performance goals are restated. The maximum
allowable loadings are then determined from the
performance curves on Figure 7-42.
The lamp requirement is estimated on the basis of
using 1.47 m arc lamps with a UV output of 26.7
W/Lamp:
Number of Lamps = [(Q)/(Q/Wn)]/26.7 Wn/Lamp
Table 7-14. Estimation of Reactor Performance
Requirements for the Design Examples
Maximum Maximum
Daily 7-Day 30-Day
Table 7-15. Sizing Calculations for the Design Examplei
Initial Fecal Coliforms
Density, N0(org/100 ml)
Suspended Solids (mg/l)
Paniculate Coliforms
Density, Np (org/100 ml)
Permit Requirement, N
(org/100ml)
Performance Goal, N'
(org/100ml)
Log (NVN0)
500,000 2,000,000 1,000,000
25
200
175
-3.45
30
225
400
175
-4.05
15
56
200
144
-3.84
'Assumed value; acknowledges that the verage daily is generally
50 to 70 percent of the maximum 30-day average
The values of Np are calculated on the basis of
Equation 7-13, with the coefficients c and m equal to
0.25 and 2.0, respectively. These are then subtracted
from the permitted effluent fecal coliform densities to
yield the design performance goal. This can then be
used to compute the design Log (N'/N0).
Daily Maximum Maximum
Average 7-Day 30-Day
Adjusted Uvg (/uW/cm2)
Performance, Log (N'/N0)
Maxmium Q/Wn (Ipm/W)a
Flow, ( Ipm )
Adjusted Design Flow ( Ipm )b
Peak Dry Weather Flow ( Ipm )
Peak Wet Weather Flow ( Ipm )
Nominal Exposure Time, tn
(seconds)0
Characteristic Length, x" (cm)
Nominal Velocity, u (cm/sec)d
Lamp Requirement
at peak dry weather
9700
-3.45
2.63
28000
36400
57000
76000
4.0
200
50.0
518
811
7300
4.05
1.35
35000
45500
7.8
200
25.6
1262
8450
-3.84
1.90
30800
40000
5.6
200
3Ei.7
788
"From Figure 7-42
bSet flows to peak diurnal conditions; adjusted daily = 1.3 x daily;
adjusted 7-day = 1.25 x adjusted daily; adjusted 30-day
= 1.1 x adjusted daily. ~
°tn = (0.176)/(Q/Wn) x 60 sec/min
du = x/tn
From Table 7-15, the maximum design requirements
are the 7-day maximum and the peak wet weather
conditions. The minimum number of lamps required
is appproximately 1,300.
222
-------
Table 7-16. Reactor Sizing Requirement for the Design
Example
Daily Average
7-Day Maximum
30-Day Maximum
Required Number of Lamps
(see Table 7-15)
Length (x) cm
Height (y) cm
Width (z) cm
Total Lamps
518
1262
788
Reactor Sizing Requirement— Perpendicular Flow Path
200 (34 lamps)
96(16 lamps)
150(1 lamp)
544
200 (34 lamps)
222 (37 lamps)
150(1 lamp)
200 (34 lamps)
138 (23 lamps)
150(1 lamp)
1258
782
Once the number of lamps is determined, the actual
modular configuration of the UV system can be
considered. Table 7-16 summarizes the combined
reactor dimensions which will satisfy the critical
design elements of the Q/Wn (i.e., number of lamps)
and the characteristic length, x. This is to form a
uniform array reactor in which the flowpath is
perpendicular to the lamps.
Recall that, the x dimension is set by dividing the
required x by the centerline spacing S, to determine
the number of lamps in the x direction. Thus, for the
design example, we can determine the number of
lamps in x:
As shown on Table 7-1 6, this is set at 34 for a total x
dimension of 204 cm. The width of the reactor is set
by the length (i.e., effective arc length) of the lamp.
This is approximately 1 .5 m. The height then, is set to
satisfy the total lamp requirement: For the average
daily condition:
No. of Lamps in y = 51 8/34 = 1 5.2 or 1 6 lamps
This yields a total y dimension of 96 cm.
The maximum requirement is set by the 7-day
condition. A possible arrangement of UV modules
would be as follows:
4 modules; 408 lamps/module(Total 1500 lamps)
Each 34 lamps long (2.0 m)
12 lamps high (0.72 m)
1 lamp length wide (1 .5 m)
The average condition can be met by 2 units; three
modules would be required under the peak condition.
The fourth unit would be solely for standby.
The reactor modules can be installed in any number
of configurations. The critical consideration is the
design of the approach and exit portions of the
system. This was discussed in Section 7.3.2.4. In both
parallel and perpendicular flow path cases, the key is
to simulate open channel flow with a constant velocity
profile across the cross-sectional plane on both the
inlet and outlet sector of the lamp battery (see Figure
7-20):
Weirs and stilling walls should be used on both ends
of the reactor to effect this even flow distribution.
Other design considerations regarding system layout
and facilities requirements are discussed in Section
7.5.
7.4.4 Summary
, The preceding calculations are given to demonstrate
the design protocol. The numbers used should not be
used for an actual design application. The procedure,
once the model is calibrated, can be very effective.
Several alternative configurations can be evaluated
and, most importantly, the sensitivity to the key
design parameters can be assessed. This applies
particularly to the hydraulic parameters of velocity,
dispersion, and head loss.
The calculations also demonstrate the importance of
having direct test information on several aspects of
the design:
Wastewater:
Reactor Characteristics:
flow and flow variability
UV absorbance
initial coliforms
suspended solids
coefficients a, b, c, and m
RTD curve (range of flows)
head loss (lamp battery
only)
dispersion coefficient
The reader is referred to Section 7.5 for discussions of
other design aspects relating to O&M and to facilities
requirements.
7.5 System Design and Operational and
Maintenance Considerations for the UV
Process
Sections 7.3 and 7.4 present the protocols by which a
UV disinfection system can be designed and evalu-
ated. The discussions centered on the design basis
and the process elements which are key to the design.
These were the hydraulic behavior of the unit, the
223
-------
calculation of the intensity in the reactor, and the
generation of the appropriate wastewater character-
ization data. Finally, an example was presented to
illustrate the design procedure.
This section presents other peripheral topics which
the designer (and operator) must consider. These
address system design elements which will affect the
operation and maintenance of the system, and the
overall economics. Specifically, the following items
are presented:
First, the factors thai: affect the reactor intensity will
be addressed. These relate primarily to the lamp
output and lamp aging, and the attenuation of
intensity due to fouling of surfaces in the reactor.
Second, methods are presented to monitor a system
directly for lamp aging and unit fouling.
Third, design considerations are presented which will
encourage effective maintenance, and assure the
disinfection performance of the unit.
Fourth, a brief discussion will be given in regard to the
major reactor components and system controls.
Fifth, the safety aspects of the system will be
discussed.
Sixth, and finally, design considerations will be
presented as they relate to ancillary facilities, mate-
rials, labor requirements, and system layout.
7.5. / Factors Affecting UV Intensity in a Reactor
A critical element in the evaluation of a system's
performance or in the design of a system is the actual
energy available in the germicidal range. The key is to
understand how efficiently the 253.7 nm energy is
being utilized and, conversely, how it is being lost.
This information is necessary to make the required
adjustments to the calculated nominal average
intensity solutions presented in Figures 7-28 through
7-32, and quantified in Equation 7-30 as the factors
Fp and Ft.
The reduction in available energy can be divided into
two major areas: the loss of lamp output (Fp) and the
change in transmittance of the enclosures separating
the lamp from the liquid (Ft). These enclosures are
typically the quartz sleeves or the Teflon tubes.
7.5.1.1 UV Lamp Output
Electrical discharge lamps generate light by trans-
forming electrical energy into the kinetic energy of
moving electrons, which is then converted to radi-
ation by some kind of collision process. Mercury vapor
kept at an optimum pressure in the presence of a
rare-gas (generally this is argon—which is the reason
for the blue-green glow seen with germicidal lamps).
is a remarkably efficient emitter of light at 253.7 nrn,
when an electrical potential difference is applied
across the device. The basic process as it takes place
in the discharge lamp can be described in three steps:
1. Free electrons are accelerated by a potential
difference applied across the lamp. This voltage
drop is maintained by an external source of
power, the current being the motion of the
electrons.
2. The kinetic energy of the accelerated electrons
is released as they collide with atoms in the
plasma.
3. The energy of the atoms is dissipated as
radiation as the atoms relax back to their lowest
energy levels.
The lower the vapor pressure of mercury in an electric
discharge, the greater the intensity of the mercury
resonance line at 253.7 nm. Exploiting this fact,
construction of the low-pressure mercury arc lamp
yields a lamp which is nearly monochromatic in its
radiation. The greatest output occurs under condi-
tions which favor having the highest numbers of
"excited" atoms close to the wall of the lamp. In this
way, the radiation they emit as they relax will have a
higher probability of passing through the wall, and
not being re-absorbed by atoms in the low energy
state. This is a reason for the thin tubular design of
the lamps, the smaller the pathlength (diameter of the
tube) the less likely the radiant energy will be re-
absorbed. In the same fashion, the lower the pressure,
the greater the chance of the radiation reaching the
wall before making a collision. Under higher pres-
sures the concentration of the atoms in the excited
resonance state is high, but because of this higher
density, the chances of the radiation reaching the
bulb wall is" small. Conversely, at extremely low
pressures the number of available atoms becomes
too small.
The resonance output of the lamp depends to a large
extent on maintaining the optimum conditions of the
discharge requirements in the vapor, and on the
operating conditions during the life of the lamp. In
large part, many of the factors which establish the
output are independent of the designer or operator
and will simply be set by the environmental conditions
the unit must operate under. Still it is important to
understand the major factors which come in to play.
Output at any given time will be influenced by lamp
temperature and by the voltage potential applied
across the lamp. Additionally, output at the resonant
frequency will always degrade with time of operation
due to any number of "aging" factors.
224
-------
Temperature. As was discussed, there is a maxi-
mum resonance output which is dependent on the
vapor pressure in the lamp. For a given lamp, this
pressure will be influenced by the temperature of the
lamp wall. The optimum wall temperature for maxi-
mum efficiency is generally between 35 and 50°C (95
and 122°F). Figure 7-43 presents data relating the
relative UV output (at 253.7 nm) to the bulb wall
temperature. In the quartz systems which are sub-
merged in the flowing liquid, the lamps are insulated
by inserting them in the quartz sheaths. The air layer
between the quartz and the lamp wall serves as a
buffer and prevents the lamps from being cooled by
the wastewater. There is little information available
regarding actual lamp skin temperatures during
normal wastewater disinfection operations. Scheible
and Bassell (36) reported that cold water tempera-
tures had little effect on the measured bulb wall
temperature. In the study at Northwest Bergen, the
lamp temperature averaged 43°C (110°F) at an
average water temperature of 21.3°C (70.3°F). At
water temperatures averaging 10.5°C (50.9°F), the
lamp wall temperature decreased to an average of
40°C.
Figure 7-43.
Effect of bulb wall temperature on the UV
output of a low pressure mercury arc lamp
(7,68).
100 -
20 40 60 80 100 120
Bulb Wall Temperature (°C)
In the submerged systems it is not practical under
most design conditions to control the lamp tempera-
ture. In the non-contact systems, such as the tubular
arrays, it is possible to maintain the lamps at their
optimum wall temperature by controlling the temper-
ature of the ambient air surrounding the lamps. This
is currently being practiced in commercial applica-
tions. Heat given off by the lamp ballasts is circulated
into the lamp reactor in cases where heat is required;
otherwise fans vent the reactor with cooler outside
air. These operations are thermostatically controlled.
Voltage. A characteristic of electric discharge arcs
is that they have a negative volt-ampere relationship.
This means that the voltage decreases with an
increase in the current. Such devices are inherently
unstable. This instability is counteracted by putting
the arc in series with a resistance. This resistance (or
reactance in the case of a-c circuits), is called the
ballast. This ohmic resistance has a positive volt-
ampere characteristic.
Radiance will be a function of the arc current. This
fact can be exploited by adjusting the voltage, in order
to vary the output of the lamp. Decreasing the voltage
will result in a decrease in the current. Such a control
mechanism has been installed at full scale facilities
as an energy conservation measure. During periods
of low UV demand, the lamps are "dimmed" by
turning the lamp supply voltage down. This results in
a reduction in the power draw of the lamp. Generally,
the lamp intensity can be reduced to levels no less
than 50 percent before the lamp current becomes too
low and the lamp will begin to flicker and eventually
turn off.
Lamp Aging. A number of factors combine to effec-
tively age a lamp and limit its useful life. These
include failure of the electrodes, plating of the
mercury to the interior lamp wall (blackening), and
solarization of the lamp enclosure material (reducing
its transmissibility). These all cause a steady deteri-
oration in the lamp's output at the 253.7 nm
wavelength, such that its output at the end of the
lamp's life can be 40 to 60 percent of its nominal
output.
The output of the lamp through its life is affected
primariJy by the extent of blackening and solarization
of the glass tube; the actual life of the lamp is
governed by the condition of the electrodes. The
germicidal lamps are typically of the hot cathode type.
These will progressively deteriorate with increasing
number of starts. Thus, the lamp life expectancy is
generally rated according to the number of times the
lamp is started, or the burning cycle. The lamp life
normally cited by most manufacturers is 7,500 hours,
based on a burning cycle of eight hours. The average
UV output at this point is estimated to be 70 percent of
the lamps output at 100 hours (note that the nominal
output of the low-pressure mercury arc lamps
represents its output after a 100 hour "burn-in"
period).
An unusuaMy long lamp life has been demonstrated
by the units atTillsohbgrg, Ontario. Lamps have been
in operation for a documented period of greater than
13,000 hours. This may be due to the fact that the
entire lamp, including the electrode connections, are
225
-------
submerged and thus cooled by the flowing water (see
Figure 7-7). The cooling of the electrodes may be a
factor in extending the life of the electrode. Con-
versely, excessive blackening and deterioration of the
end of the lamps has been noted for systems in which
the lamps are inserted through a metallic bulkhead.
There is no cooling in this case and, in fact, this may
cause a buildup of heat.
Measurement of the UV Output of the Lamp. Table
7-7 presented the specifications for a number of
slimline type germicidal lamps. Each is specified with
a rated UV output at 253.7 nm, expressed in Watts,
and the arc length, which is the length of the radiating
portion of the lamp column. Figure 7-44 presents the
output of the 1.5-cm diameter lamps as a function of
the arc length. The slope of this is 18.2 UV W/m of
arc. This will be slightly lower for lamps with a
diameter of 1.9 cm. One should understand that this
is the nominal output of the lamp; as discussed, this
will decrease with time.
Figure 7-44. Nominal lamp output as a function of arc
length.
Figure 7-46. Measurement and analysis technique 1For
estimating the total UV output of a lamp.
Lamp
IL 570 Detector
• Arc of Radius r
Measurement at r = 1 22 cm
BulbN-1-12
Angle IL570
B (//W/cm2)
90°
75°
60°
45°
30°
15°
87.0
85.5
82.0
69.0
55.5
24.0
W C
iS
I.4
2
0
Total
P=2x6.5 = 13 W-
at 253.7 nm
Area <=* 6.5
80
60 40 20
8 (Degrees)
40
§ 30
It
CC
20
10
^18.2
Watts/m Arc
(Lamp Diameter = 1.5 cm)
I I i
0.4
0.8 1.2
Arc Length (m)
1.6
2.0
compute the power, P, at any point at distance r from
the center of the lamp, at angle 6, is written:
where:
P
6
I =
r =
a =
0°
/ 2n a(0) l(0) rdff
90°
(7-46)
the total power at 253.7 nm (Watts)
angle of the detector from the longitudinal
centerline of the lamp (0 to 90 degrees)
intensity reading at radius r at a point (T)
along the arc
radius of arc (distance from center of lamp
to detectors) (cm)
distance from detector to axial line of lamp
(cm)
A test procedure for determining a lamp's UV output
was described by Johnson and Quails (38). The
procedure, in effect, treats the lamp as a single point
source of light. A UV detector measures the intensity
along an arc scribed at a fixed radius from the center
of the lamp. These intensity readings are integrated
over the surface of a sphere with the same radius,
resulting in an estimate of the total output from the
lamp.
Figure 7-45 is a schematic representation of the
experimental set-up to measure the total output and a
sample analysis of a set of data. The expression to
This can be solved graphically by plotting dP/d0
against 0, as shown by the example on Figure 7-45.
The area under the curve is power output at the
wavelength 253.7 nm, and represents one-half of the
sphere about the lamp. Doubling this power estimate
yields the total output of the lamp.
In the Port Richmond study, a total of 9 lamps were
measured by this method after 100 to 300 hours
operation. The measured output ranged from 12.2 to
15.9 W, with an average output of 13.6 W; this is
close to the rated output of 14.3 W given by the
manufacturer.
226
-------
A far simpler technique can be used to implicitly
monitor the average lamp output with time. The setup
with the lamp and the detector is the same as shown
on Figure 7-44. In this case, however, a single
intensity reading can be taken at 0 equal to 90
degrees and compared to the same intensity reading
for a new lamp. Figure 7-46 is a sketch of a simple
table-mounted unit which can be used at a full-scale
facility.
Figure 7-46. Sketch of lamp monitoring set-up.
4
Barrier
Lamp^_
Ballast
Leads
!flfc
^
Tfl
\
IT \ '
I
and
>5 Feet t
6 Inches /
Detec
W'
/
\ /
Lamp Being Detector
Measured Bracket
-Table
' — '
Section View
Meter
To 110V Supply
Barrier
^ Lamp Being
Lamp /"I
(One of Two) illi
^--Brackets (Fixed) /
Measured
Table
i
Midpoint of Lamp Detector
Bracket
Meter
Detector
(Fixed)
2-Lamp Ballast
w Leads
No Less Than 5'
Plan View
Two lamps are operated off a single ballast. The lamp
to be measured would be placed on brackets which
are in a fixed position from the detector. The other
lamp should be operating but placed behind a barrier
to prevent it from interfering with the lamp being
measured. Between three and five minutes warm-up
time should be allowed before a measurement is
taken. The leads from the ballast should be wired to
end caps which can then be quick fitted onto the
lamps. The lamps should not be repositioned by
disconnecting them; they should be moved with their
leads in-place.
As shown, the setup can be placed on a table-top; the
lamp brackets should be thin and should not shield
light. The lamp which is not being measured can be
shielded by fitting a length of cardboard tubing (or thin
opaque plastic tubing) over it. The tube can be slit
lengthwise and should fit loosely (to prevent over-
heating the lamp). Alternatively, as shown on the
sketch, a barrier can be set up between the lamps.
The detector is set up at the other end of the table. The
mounting bracket should be rigid and fixed. The
bracket should allow for removing and inserting the
probe without changing the position of the detector in
any direction. The objective is to always have the
lamp and detector in a fixed position; these positions
must also be reproducible from day to day. The
detector should be on the same horizontal plane of
the lamp centerline and perpendicular to the midpoint
of the lamp. The distance between the lamp and the
detector should be no less than five feet.
The procedure for monitoring the lamp intensity, once
the setup is in place, is rather simple. The idea is to
measure a lamp's intensity relative to that of a new
lamp. The first step is to measure the intensity, at the
fixed distance, of 3 to 5 new lamps which had been
burned for approximately 100 hours. The average of
the five then becomes the benchmark to determine
the relative output of the lamps in use with the system
(percent of new lamp average). Each lamp should be
tagged and given an I.D. number; this allows direct
monitoring of individual lamps and allows the oper-
ator to keep an appropriate mix of lamps in a system
and to know when to discard a lamp.
The same procedure is used to monitor the trans-
mittance of a quartz sleeve. In this case a single lamp
is used; first the intensity is measured with and
without a new, clean quartz sleeve in place over the
lamp. Similar measurements are then taken with the
unit's quartz sleeves and compared to the transmit-
tance of the new quartz. This can be done before and
after the quartz is cleaned.
7.5.1.2 Losses of UV Energy through the Quartz
and Teflon Enclosures
The UV output of the lamps themselves can be
monitored with time, as discussed above. This is
generally a non-controllable parameter, although
optimum conditions can be maintained (such as
voltage and temperature) to keep the output at its
maximum. The lamps will age, however, and lose
output. A key consideration regarding the UV source
in both system design and subsequent operation and
maintenance relate to maximizing the utilization of
the source output and understanding the known
energy sinks within a given system.
Figure 7-47 schematically presents two typical UV
lamp configurations and the known energy sinks. The
first is the tubular array; the second is the submerged
quartz array. In either system there are several ways
the UV energy is lost before it reaches the liquid and
can be utilized for its primary germicidal role. First,
the lamp wall itself can become dirty. In the non-
contact tubular array systems the lamps are in an
open air environment within the reactor. The air is
often circulated to keep the ballast cool and/or to
control the lamp temperature. This can introduce dust
227
-------
Figure 7-47. Energy links In UV reactor.
Wastewater
Carrier (Teflon)
UV Source
100%
Non-Contact System
(Teflon Carrying Tubes)
Lamp
Quartz
Envelope
'
X
Wastewatei
^
_
: j100%
UV Output
1
Contact System
(Submerged)
which settles on the lamp surface and becomes an
energy absorber. This same problem will also cause
the outer surface of the Teflon tubes to become dusty
and reduce the Teflon's transmittance. Filters are
now installed in such units to minimize these
problems.
In the quartz systems, some units are installed which
either seal the quartz ends or leave them open. In the
open arrangement, convective air currents can carry
air (often humid) through the quartz sleeve, causing
some deposition on tthe lamp surface. Additionally,
the same air convection will cause the inner surface
of the quartz sleeve to become dirty. This may also
occur to some degree in sealed systems due to
condensation effects, although there is no current
information regarding these effects.
Passage through the quartz sleeve or the Teflon tube
wall will itself cause a loss of energy. The fused quartz
sleeves are highly transmissible of UV at 253.7 nm.
The transmittance of the Teflon will vary with the
thickness of the tube wall and is typically less then
than that of the quartz. Lastly, the surfaces of the
quartz sleeves or the Teflon tubes which contact the
wastewater will foul and cause the transmissibility of
either to be reduced. The O&M tasks will naturally be
directed to keeping these surfaces clean and main-
taining the maximum transmittance of the quartz or
Teflon.
Quartz Systems. The quartz sleeves are typically
high quality fused quartz, with a transmittance
greater than 90 percent when in a new, clean
condition. In the Port Richmond study, relative output
readings of a lamp were taken with and without the
quartz sleeve in place by the procedure discussed
earlier. The average (of 12 measurements) reduction
in intensity measured when the quartz was placed
over the lamp was approximately 25 percent, with a
range of 15 to 35 percent. Obviously, this is signifi-
cantly different from the 90 percent for a new and
clean quartz; the reasons relate to the dirtiness of the
surfaces and the loss due to ozone absorption.
The lamps used at Port Richmond were made of
quartz, which is transparent to energy at the 185 nm
wavelength, a spectral line characteristic of the low
pressure mercury arc. Energy at this wavelength will
ionize free oxygen to ozone which, in turn, is an
excellent absorber of energy at the 253.7 nrn
wavelength. Thus, with any production of ozone in
the gap between the lamp and the quartz sleeve, it is
likely that there would be a consequent absorption of
UV energy. Direct testing in the Port Richmond
project confirmed this effect.
Not all lamps will transmit this energy. In fact, the
majority of lamp designs utilize a lamp envelope
which has a low transmittance at the 185 nm
wavelength. This is shown on Table 7-7. It is
recommended that these types of lamps be used for
UV disinfection systems."
Table 7-17 is excerpted from the Port Richmond
report (54), and shows the effects of fouling on the
quartz transmittance. Intensity readings taken im-
mediately after the quartz were removed from the
unit averaged 61.8/uW/cm2, which was 81.6 percent
of the reading obtained with the new quartz sheath.
Cleaning the inside surface of the quartz improved
the output to an average 70.1 /uW/cm2, or 92.8
percent of the new quartz reading. Finally, the output
increased to 75.3 /M//cm2 when the outer surfaces
were cleaned, which is essentially equivalent to the
new quartz reading.
The inside of the quartz had last been cleaned five
months before these tests; the report then ascribed a
15 percent output decrease per five months due to the
inside surface fouling. The outside surfaces were
cleaned with an acid/detergent solution on a frequent
basis, and immediately before these readings were
taken. The results indicated that although this
procedure was effective, there is apparently a film
layer which stays on the surface; this was presumed
to cause a loss of approximately eight percent of the
UV output at any time.
The Port Richmond report presented a summary
analysis of the lamp output and quartz transmittance
monitoring conducted during the term of the project.
This is repeated herein as Figure 7-48, and provides
an excellent example of the importance of both
monitoring these conditions and accounting for them
228
-------
in the design of a system. The procedures are rela-
tively straightforward and require little expense in
terms of monitoring equipment. The primary com-
mitment is the labor requirement for taking the
necessary readings.
The manufacturer's rating for these lamps was
confirmed by direct measurement of their UV output;
this was shown to degrade to approximately 60
percent of this output after 8300 hours. A 25 percent
reduction is taken to account for the quartz sheath
absorbance, and losses attributable to ozone ab-
sorbance within the air gap between the lamp and the
quartz sheath. If low or zero ozone producing lamps
are used, this reduction will be approximately 10
percent. A constant eight percent loss is taken to
account for the film layer on the outer surface of the
quartz sheaths. This was considered the base line
loss as discussed earlier; with increasing time
between chemical cleaning cycles, this loss will
increase. If not attended to (cleaned) this surface
fouling can cause significant deterioration in the
system performance. The internal fouling, although
not as significant as the outside surface, will still have
an effect. It should be a requirement, particularly in
systems where the quartz ends are open, to clean
these surfaces at least once to twice per year.
The lower line on Figure 7-48, therefore, is an
estimate of the actual average UV output for the
lamps at any time during the operating period for the
Port Richmond study. It is significant to note that
although a system may start with a nominal output
from a lamp source, this is immediately reduced to
approximately 70 percent of nominal simply by its
placement in the quartz sheath and by the develop-
ment of a film layer with time. Over the operating age
of the lamp, this output, even with good maintenance,
will deteriorate to approximately 35 to 45 percent of
this nominal output after one'year. Thus, it is
important to maintain the system and keep this
output at its maximum.
Teflon Tubular Systems. A number of tests were
also conducted during the Port Richmond study On a
sample of the Teflon tubes used in Unit 3 of the
project. Lamp intensity readings were taken with and
without clean Teflon in place; these readings also
evaluated the transmittance with the lamp on both
the convex and concave side of the Teflon. The resu Its
are presented on Figure 7-49, excerpted from the Port
Richmond report. They are not wholly conclusive;
there appears to be some apparent effect due to the
curvature of theTeflon, with the highertransmittance
measured when the lamp is on the inside of theTeflon
arc.
All of the used samples were vigorously cleaned in
the lab with hot soapy water and a soft brush after the
transmittance measurements were made. The UV
transmittances after this cleaning are also shown in
Table 7-17. As can be seen, considerable improve-
ment was obtained. However, none of the cleaned
samples demonstrated UV transmittances at the level
one would expect for a virgin Teflon sample of the
same wall thickness, suggesting that Teflon may
undergo a transformation and lose some of its ability
to transmit UV light over time. This may be caused by
continued exposure to the UV lamps.
The Teflon used in the Port Richmond tests had been
in use for a period of time. The inside surface was
observed to be dirty, but was cleaned before the
readings were taken. These results compare poorly
with the 75 percent transmittance cited by the
manufacturer as characteristic of Teflon. In their
report, the authors suggested a transmittance level
for the Teflon to range between a maximum of 75
percent when new to as low as 30 percent under
significant fouling conditions.
A special series of tests were conducted using an
alternative and possibly more accurate method to
determine the transmittance of the Teflon. These
tests (4) involve the exposure of a chemical acti-
nometer to UV light. The compound was 0.006 M
potassium ferrioxalate; upon exposure, the ferric ions
will be reduced to ferrous ions in proportion to the
amount of UV light received by the actinometer. The
ferrous concentration is measured spectrophoto-
metrically at 510 nm, using phenanthroline as the
color reagent. The chemistry procedures follow those
by Baxendale and Bridges (60) and Parker (61).
A sketch of the bench-scale setup is given in Figure
7-50. Note that, although these tests addressed
Teflon, the same procedures are applicable also to
quartz. Work should be conducted in a darkened
room. The ferrioxalate is placed in a 1.3-cm diameter
by 8-cm long fused quartz test tube (volume = 6 ml)
capped with a non-reactive stopper, and covered with
an aluminum foil sheath. The UV lamp is allowed to
stabilize for a minimum of five minutes before any
test. The test tube containing the actinometer would
be clamped in a fixed position from the lamp; exposure
would be accomplished by slipping the aluminum foil
off the test tube for a preset period of time. The
ferrous concentration would then be measured.
Exposures are conducted with and without the Teflon
tube in place.
The tests are conducted to encompass a range of
reaction times. The reaction rate is determined by
plotting the ferrous concentration against exposure
time. The transmittance of the Teflon is estimated by
determining the reduction in the reaction rate of the
actinometer. Examples of these results are given in
Figure 7-51. Results are given without theTeflon and
223
-------
Figure 7-48. Approximation of average lamp UV output at 253.7 nm with time for quartz systems, accounting for lamp aging arid •
surface fouling (54).
2
-------
Table 7-17. Effects of Fouling on the UV Transmittance of Quartz (54)
Intensity Readings
Tube
Number
4
34
64
74
84
94
After Removal
from Unitb
(/uWatts/cm2)
50.5
60.5
63.5
68.5
66.5
60.0
Relative
to New
Quartz
66.9%
80.1%
84.1%
90.7%
88.1 %
79.5%
Cleaned
Inside of
Quartz
(/uWatts/cm2)
66.5
67.5
70.5
73.5
73.5
69.0
Relative
to New
Quartz
88.1 %
89.4%
93.4%
97.3%
97.3%
91.4$
Cleaned
Outside
of Quartz
(AM/atts/cm2)
77.5
73.5
74.0
73.5
72.5
81.0
Relative
to New
Quartz
102.6%
97.3%
98.1%
97.3%
96.0%
107.3%
Average
61.5
81.6%
70.1
92.8%
75.3
99.8%
"Intensity of bulb without quartz is 90.5 /nW/cm at 1.2 m. The intensity with a new quartz in place is 75.5 //W/cm2. This is used as
the reference intensity.
""Cleaned by acid/detergent solution while in the unit.
Figure 7-49. Estimate of Teflon transmittance by use of a
UV detector (54).
100
90
r- 80
C
j§ 70
1 60
o> 50
c
'•5
(n
, S 40
oc
20
10
0
u> t a la i ivd
I I I I I II
]. asiis /gl J-
Detector _>^ IT
Teflon _
^^
^ . ____^--^^^^
_X
b ^^^^
^^~~— -— ^^
-
-
I I I I I I I I
Figure 7-50. Test set-up to conduct actinometry experi-
ments (4).
To Ballast
To Ballast
0 10 20 30 40 50 60 70 80 90 100
Distance from Bulb to Detector (cm)
Quartz
Test Tube
Containing
Teflon Tube Actinometer
(Can be Moved Back
& Forth on Ring Stand ^,^-
Holder to Cover Test ^~-
Tube if Desired)
—25 cm
Ring Stand
231
-------
Figure 7-51. Example of chemical actinometry tests to Figure 7-52.
determine Teflon UV transmission (4).
Effect of wall thickness on Teflon transmit-
tance, as determined by chemical actinometry
(4).
f
S4
3
§
r
D=6 cm
D W/O Teflon
A Avg. T = 0.81 mm
• Avg. T = 0.94 mm
I
100
50 100 150
Exposure Time (sec)
200
Tibia 7-18.
UV Transmittances of New and Used Teflon as
Determined by Chemical Actionmetry
Teflon Tube I.D.
%UV
% UV Transmittance
Transmittance After
As Received Lab Cleaning
New. D=6 cm, T=0.81 mm
Now. D=6 cm, T=0.94 mm
New, D=8.9 cm, T=0.46 mm
New, D=8.9 cm, T=0.84 mm
Port Richmond #1"
Port Richmond #2
Port Richmond #3
Port Richmond #4
Port Richmond US
Chinook—3800 hours"
Chinook—3800 hours
(Cleaned in field with
water and vinegar)
Chinook—100 hours
Chinook—100 hours
(cleaned in field
with water and vinegar)
Baech Mountain #1°
Beech Mountain #2
85%
72%
92%
85%
7
9
5
18
30
30
66
55
70
33
47
68
71
72
64
76
75
78
80
85
74
78
•Port Richmond tubes had D=3.5" and T=0.030".
"Chinook tubes had D=3.5" andT=0.032". One half of the tube had
been exposed to UV light for 3800 hours, the other half for 100
hours. Each of those halves was cut into sections; one section was
cleaned with a high pressure nozzle washer with water and
vinegar, the other not.
"Beech Mountain tubes had D=2.375" and T=0.033".
Note: D * inside diameters, T = wall thickness
90
c
I 80
70
60
I
0.2 0.4 0.6
Teflon Wall Thickness (mm)
0.8
1.0
ranged from 5 to 30 percent. The Chinook samples
were not as obviously fouled, but they did have
noticeable whitish precipitate deposits and some
scum and grease attached to the Teflon. Their UV
transmittance were in the range of 50 to 60 percent.
Beech Mountain samples also demonstrated UV
transmittances in the 50 to 60 percent range.
7.5.1.3 Summary—Adjustments to the Estimated
UV Intensity
The lamp output will decrease with operating time. It
is recommended that the system be designed (and
subsequently operated) on the basis that the average
output is approximately 80 percent of the nominal
output. This is equivalent to an Fp of 0.8 in Equation
7-30. With time, it will be necessary to mix newer
lamps(Fp>0.85) with older lamps(Fp<0.75) in order
to maintain the desired output level. This will require
monitoring by the procedures suggested earlier.
The quartz sleeves and Teflon tubes will require
effective maintenance to keep their transmittances at
reasonable levels. For design purposes, one needs to
consider the minimum transmittance to be expected.
With the quartz systems this is suggested to be 60 to
70 percent of nominal. This is equivalent to an Ft of
0.6to0.7 in Equation 7-30. For the Teflon systems, an
FtbetweenO.5 andO.6 is recommended. In situations
where the disinfection units are not to be frequently
attended (remote, smaller plants), the values of Fp and
Ft should be reduced further. Values of 0.7 and 0.4 are
suggested, respectively.
232
-------
7.5.2 System Design Considerations for Effective
Maintenance
An overriding concern in the proper maintenance of
the UV reactor for effective performance is to keep all
surfaces through which the radiation must pass as
clean as possible. The effects of surface fouling on
energy utilization efficiency were discussed in detail
earlier in this section. It is critical, and can very often
be pointed to as the primary reason for the non-
performance of a particular system. Other concerns
relate primarily to the accessibility to UV reactors and
to keeping adequate records to control replacement
cycles and maintenance schedules.
7.5.2.1 Reactor Maintenance
The most reliable method to determine if a reactor is
becoming dirty and requires cleaning is by visual
inspection. The unit should be drained and the
surfaces observed for fouling. In open systems this
can be done rather conveniently and quickly. Reactors
which are sealed vessels can be difficult; these
designs should accommodate such visual inspections
by incorporating large portholes or manways in the
reactor shell.
Generally the surfaces of submerged quartz systems
contacted by the wastewater will become coated by
inorganic scale, very much like boiler scale. This will
especially be the case in areas where there is hard
water. Additionally, the inside surface of the quartz
and the outer surfaces of the Teflon tubes will
eventually develop a grimy dust layer, primarily from
airborne dirt and water vapor.
Fouling of the reactor's internal surfaces will be
signaled by reduced performance efficiency, or by
reductions in the intensity measured by in-line
probes. While these may provide some signal of
fouling, it is still necessary to be able to physically
inspect the surfaces.
Subsequent discussions will present procedures and
equipmentfor routine cleaning of the reactor surfaces
which contact the wastewater. First, however, it is
appropriate to discuss a maintenance task which
should be conducted at least once per year, or once
per disinfection season. This is to completely overhaul
the reactor, cleaning all interior surfaces, and deter-
mining the lamp outputs and quartz (or Teflon)
transmittances.
These procedures were demonstrated for quartz
systems at Port Richmond (54). Each lamp is removed
from the reactor and washed with a mild soap
solution, rinsed, and swabbed with an alcohol
(isopropyl) soaked rag (cheesecloth). Then the interior
surfaces of the quartz sleeves are cleaned by the
same procedure by using a gun-barrel type cleaning
rod to swab the interior surfaces. At the same time,
each lamp, which is tagged with an ID number, is
measured for relative output. Those which are below
a specified level are discarded and replaced with new
lamps. These new lamps are also tagged with a
number. In this manner each lamp can be traced on
the basis of operating time and output. A reactor lamp
inventory can then be mixed and controlled to
maintain a minimum average output level.
In similar fashion, the quartz should be monitored for
transmittance. It may be cumbersome, however, to
remove all the quartz from a system. It is recom-
mended, instead, that a representative fraction of the
quartz sleeves be monitored. Ten to fifteen percent of
the quartz inventory would be sufficient. The same
quartz should always be monitored; these would be
considered as representative of all quartz in the
system. If the tagged quartz begin to show marked
deterioration due to aging and wear, it may then be
appropriate to broaden the monitoring and to begin
replacing the quartz sleeves. This replacement can be
accomplished gradually. As with the lamps, there will
eventually be a mix of old and new quartz in the
system. There is little experience in determining the
effective life cycle of the quartz sleeves; certainly it
will vary by site, but generally should be between four
and seven years.
In Teflon systems, the lamps are removeable on
racks; they should be cleaned and monitored in the
same manner as the quartz systems. The Teflon tubes
should be cleaned on occasion; this can be done by
swabbing the tubes with soapy water/alcohol. A non-
abrasive material should be used. Each tube should
also be monitored for transmittance, just as with the
quartz sleeves. This may not be as straightforward,
however, because of their limited accessibility and
problems in getting direct measurements with a UV
radiometer/detector. The actinometry procedure
described earlier may be the more appropriate
method for this application.
This system overhaul, as mentioned, should be
accomplished at least once per year. In the case of
plants with seasonal disinfection requirements, the
most appropriate time would be before the start of the
disinfection season.
7.5.2.2 Routine In-Place Cleaning
Regardless of a particular system's accessory clean-
ing equipment, it is likely that periodic chemical
and/or detergent cleaning will be required to main-
tain the outer quartz, or inner Teflon surfaces. This is
particularly the case where the wastewater is rela-
tively dirty (secondary or primary effluents), has a
relatively high grease and oil content, or has a high
hardness content. A major cause in fouling the quartz
surfaces (and to a lesser extent, the Teflon surfaces)
has been found to be inorganic magnesium and
calcium carbonates. The inorganic deposition is also
233
-------
easiest to control; simple acidification of the reactor
water will generally dissolve the material and restore
the surface. In the case of organic fouling (usually
from a high grease content) it is necessary to use a
detergent or some combination of cleaning agents.
This is typically determined by trial and error for a
particular situation. The frequency with which this
cleaning task has to be accomplished will also be site
specific, and will be determined with experience.
A number of systems commercially available offer
accessory equipment which are purported to maintain
the surfaces of the Teflon or quartz. Currently, these
are the mechanical wiper, ultrasonic transducer, and
a high pressure spray nozzle. The mechanical wiper
and ultrasonic devices are applicable to the quartz
systems, while the high pressure spray wash is
applicable to both the Teflon and quartz systems.
Although these components may be effective in
cleaning the appropriate surfaces, intermittent clean-
ing with chemicals is generally required, and it is
strongly recommended that provision be made in the
overall system design to allowfor chemically cleaning
the system. In certain cases, the accessory cleaning
devices can be used to assist and increase the
efficiency of the chemical cleaning task.
Chemical Cleaning. The task of cleaning the UV reactor
on a routine basis is generally a very straightforward
and simple task. The procedures and equipment re-
quirements will generally be a function of the type of
reactor and the system size. There is no standard pro-
cedure or equipment design, nor should there be. A
system and procedure should be developed which best
suits the application. This will be influenced by the size
of the plant, the level of operator attention, the type
and characteristics of the wastewater, and the UV
system configuration.
The simplist procedure is applicable to the exposed
open channel units. Examples are the systems at Pella,
Iowa; Northfield, Minnesota; and Tillsonburg, Ontario.
These are all open channel systems where the lamp
batter can be drained, is accessible, and can be easily
inspected visually. The unit at Port Richmond were also
similar to these types, of systems.
At Port Richmond, a combination of acid and an
industrial detergent was used. The unit to be cleaned
was isolated by diverting flowto the second unit; acid
was then added to adjust the reactor water pH to
approximately three. The acid was sulfuric in this
case; a pH meter would be used as the entire reactor
liquid volume was acidified by direct addition. Once
acidified, the detergent would be added in similar
fashion. The wiper stroke would usually be increased
to assist the cleaning operation. This procedure was
found to be very effective and typically required no
more than one hour of a single operator's time.
An operational concern regarding this aspect of
system maintenance is determining the frequency
and/or need to chemically clean. Visual inspection
has already been mentioned as the most effective
procedure. An in-place intensity monitor would be
effective, although there are concerns for this type of
fixed place detector. The window for the probe must
itself be kept clean in some fashion; additionally, the
probe sees only the surfaces in the near vicinity, and
the probe cannot directly account for the absorbance
characteristics of the wastewater itself. At Port
Richmond, a portable radiometer was found to be
more effective, when used in conjunction with the UV
absorbance measurements taken on each sampling.
Intensity measurements were taken by placing the
detector at selected (and reproducible) positions
along the influent and effluent plane of the lamp
battery. Figure 7-53, taken from the Port Richmond
report, shows detector intensity readings versus the
absorbance readings of the wastewater. These were
taken after the unit was chemically cleaned. The
relationship is important, not from the absolute
intensity readings as a function of absorbance, since
these will change with lamp age, but from the relative
change as a function of absorbance. In this particular
example, the intensity is shown relative to the
intensity at an absorbance coefficient of 0.2 cm~1.
Thus, by knowing the relative effects of the waste-
water absorbance, the operator can make reasonable
judgements from the radiometer readings as to the
condition of the quartz surfaces. This same procedure
can be accomplished with the Teflon system by
inserting the detector (always to the same fixed
position) into the Teflon tubes.
Figure 7-53. Example of radiometer intensity readings as a
function of UV absorbance at Port Richmond
(64).
? 100
1? 9°
B 53 80
££ 60-
II 50
12
-------
At Pella, Iowa, direct provision was not made for
chemical cleaning (62). The system has mechanical
wipers which are generally effective, but the units do
foul. The operators isolate one of the two units, drain
the channel and inspect the lamp battery. If dirty, the
unit is rinsed, then sprayed with an acidic detergent.
The wiper is stroked across the unit several times and
the lamp battery is rinsed once again. The surfaces
are inspected and the procedure is repeated, if
necessary. The unit is then put back into service. This
entire procedure is efficient, requiring one-half to one
hour, and is typically done once every one to two
weeks.
At Northfield, Minnesota, the quartz units are con-
structed very much like the units at Port Richmond,
with inlet and outlet tanks. The lamp battery is open on
both the inlet and outlet planes. Again, this plant had
been designed with mechanical wipers and no provision
to chemically clean the quartz. The area is marked by
very hard water and the quartz were found to foul fairly
quickly with an inorganic scale. The operators had
cleaned the unit by adding citric acid to the reactor
water (having first isolated it from the system) and
allowing it to soak for several hours. The wipers would
also be kept in operation during this period. This was
generally effective, but relatively expensive. Each ap-
plication involved adding 50 to 100 Ib of citric acid,
costing over $ 100 per sequence.
This procedure has subsequently been modified. The
citric was replaced by a cheaper mineral acid, (sulfuric
acid). Less than a nine pound bottle is needed to
reduce the pH to approximately three. It is recom-
mended that the acid input be controlled by a portable
pH meter to prevent addition of excess acid. A small
metering pump with a small bulk acid drum (stored
outside the building), is also recommended to further
reduce the cost of the acid (low grade or waste
mineral acid is sufficient). A recirculation pump is
also added to rapidly mix the tanks and provide
agitation. The water level is typically lowered to the
top of the lamp battery to further reduce acid
requirements. Once acidified, the unit is allowed to
stand for a period of time, with the recirculation pump
in operation and the wiper moving. The tank is then
fully drained (drainage to the head of the plant) and
the quartz is inspected. It is rinsed, sprayed with an
acidic detergent, and then rinsed again with a high
pressure hose. A wand sprayer is used to be sure the
internal quartz are reached. The unit is then brought
back into service. This procedure is effective and
requires only one to two hours to accomplish. The
frequency is generally once every week.
The Tillsonburg, Ontario unit is set in the plant
secondary effluent channel and does not have any
accessory cleaning equipment (63). This too can be
easily cleaned because of its accessibility. The unit is
isolated, the channel drained, and the quartz cleaned
by the procedures described above. Alternatively, this
particular design allows for individual lamp modules
(each containing four lamps) to be pulled from the
unit. The unit does not have to be shut down, and, if
necessary, a spare lamp rack can be inserted tempo-
rarily. The lamp racks are then inspected and
thoroughly washed and rinsed. The unit design also
incorporates the application of a proprietary poly-
meric coating to the quartz surface which is designed
to retard fouling.
In sealed quartz systems, it is not possible to easily
inspect the quartz or to access the internal quartz
with sprayers or high pressure wash lines. At Vinton,
Iowa the units are designed with ultrasonic devices,
and a supplemental ability to chemically clean the
reactor (64). The reactor is first drained, then it is filled
with clean water and a chemical cleaning agent.
These are added through piped inlets to the reactor.
The system is also designed with inlet ports to inject
either high pressure air or water to agitate the
solution inside the unit.The unit isthendrained(back
to the head end of the plant), and returned to service.
At Suffern, New York, the system is also sealed.
Ultrasonics are incorporated as the primary cleaning
device, with a supplementary chemical cleaning
system (4). The chemical cleaning system includes a
solution mix tank, and recircujation pumps. A sche-
matic is provided on Figure 7-54. A solution of the
cleaning agent is prepared in the tank with warm
water, and then recirculated for a period of time. Food
grade citric acid or sodium hydrosulfite is used.
During startup at Suffern, the quartz became heavily
fouled (the units had been allowed to sit for a long
time filled with wastewater); the citric acid was not
effective in this case. The sodium hydrosulfite was
very effective; however, it is highly reactive and a
strong oxidant. This material would likely be the most
effective in sealed systems where the cleaning relies
solely on contacting the surface under agitated
conditions. It must be handled with great care,
however, and special precautions would be required
to properly store and handle the material. The
chemical manufacturer should be consulted on these
aspects.
A special note is also made under this topic of routine
chemical cleaning. In several cases, it has been
observed that a luxuriant growth (believed to be a
fungus) will develop on the wetted metallic surfaces
of the UV reactor. Additionally, particularly in reactors
which have quiscent zones, sludge accumulations
can develop. When the systems are drained and/or
cleaned, an effort should be made to remove these
accumulations.
235
-------
Figure 7-54. Schematic of in-place chemical cleaning system at Suffern, New York (4).
Solution
Return Lines
Dry Polymer Eductor
Overflow
Chemical Make-Up Tank
No. 2
Cleaning Solution
Feed Pumps
No. 1
Mechanical Wipers. A number of full-scale systems
incorporate the use of a mechanical wiper. Recall
from Figure 7-5 the schematic of the wiper blade on a
submerged quartz system. These entail a machined
frame in which the wipers are fixed; these then fit
over the quartz sleeves. A single frame usually
services the entire lamp battery. The wiper is driven
by cable (pnuematically pulled) or by a piston. This can
then be stroked across the reactor at a preset
frequency. Examples of systems which have wipers
are Pella, Iowa; Northfield, Minnesota; and Albert
Lea, Minnesota.
Generally, the wipers are looked upon favorably by
plant operators. At Pella, the wiperblades are a
rubber-base ring. These tend to wear and will typically
require replacement every one to two years. The
wiper does not accomplish its original intent, which
was to keep the surfaces clean, precluding the need
for chemical cleaning. The wiper is felt to serve a
useful purpose, however, by continually removing
small debris particles, including grit, plastic fibres,
and strings of algae (from the secondary clarif iers a nd
channels). When the unit is taken down for chemical
cleaning, the wipers are used to provide a degree of
scrubbing.
Similar observations are made at the Northfield plant.
The rings, in this case, are made of Teflon. Although
they should not wear as quickly, it is found that the
Teflon ring becomes distorted. It has no memory;
thus, if there is a variation in the surface of the quartz,
the Teflon will respond but will not recover its original
shape. The wiper is still considered a benefit,
however, because of its ability to keep the surface
free of debris, and the ability to use it during the
chemical cleaning task.
At Albert Lea, the wiper also uses Teflon rings but of a
different design (and at present the most current
design). These rings are split by cutting the ring on a
bias at one point on its circumference. A spring then
surrounds the ring, its compression causing the ring
to always try to close. In this fashion, if the wiper
passes over a section of the quartz which is larger in
diameter, the ring will expand. As the quartz diameter
decreases, the Teflon will also close down on the
surface because of the spring action.
236
-------
A limited series of tests were performed at the Albert
Lea plant to assess the effectiveness of the wipers.
These tests involved monitoring quartz clarity from
modules which were operated with and without the
wiper. The results were not wholly conclusive. Kreft
et al. (4), reported that the wipers were effective in
reducing the buildup of scale and biological growth on
the surfaces of the quartz, when compared to the unit
with no wiper in operation. The report also noted
problems which must be addressed in the design and
operation of system with the wiper devices. The cable
drive is sheathed in Teflon; this was found to crack
and cause water leaks onto the outside electrical
connections (this same problem was reported in the
Port Richmond study). This resulted in electrical
hazards which on one occasion caused a small
electrical fire. The second problem arose due to the
misalignment of the wiper frame. This will cause, at
minimum, incomplete wiping of the quartz; at worst,
the misalignment will cause breakage of the quartz
sleeves.
Other considerations which should be taken into
account are the time requirements for the disas-
sembly and repair of these wiper mechanisms.
Experience up to now has involved the equipment
manufacturer accomplishing this task; this may be
appropriate since it requires extensive handling of the
quartz sleeves and precise alignment of the frame
when reassembled. The design of the UV system
should address these tasks, particularly their costs,
including the costs of replacing the individual wipers
(e.g., the spring loaded split Teflon rings).
Ultrasonics. Ultrasonic devices rely on the surface
cavitation caused by high frequency sound waves.
When properly applied, debris which coats a surface
will simply fall off. This concept has been applied at a
number of full-scale plants, including the Suffern,
New York and Vinton, Iowa installations. The ultra-
sound transducers are typically inserted across the
length or width of the reactor, parallel to the quartz
sleeves. They are operated on an on/off cycle, the
on-time determined by the specific site requirements.
The units at Vinton each have two transducers which
have 1.8 kW ultrasound input; thus, each unit is
equipped with a total of 3.6 kW input. The units are
rated to have an effective radius of activity of
approximately 0.75 m over a 180° arc.
The total lamp power in each of the units at Vinton is
approximately 12.8 kW; this yields a ratio of ultra-
sonic power to lamp power of 0.28. With the
ultrasonics operated 25 percent of the time, the
energy utilization by the ultrasonics comprises ap-
proximately seven percent of the total when all lamps
are being operated and 28 percent when one bank of
lamps (out of four banks) is in operation. These would
increase if the ultrasonics are required for greater
periods of time.
The system at Vinton has been in operation for
approxi mately two years. At first the u Itrasonics were
operated with approximately 25 percent on-time. This
was not felt to be effective and was gradually
increased to nearly full time. This is excessive; its cost
of operation negates its utility. Hypochlorite had been
used to chemically clean the Vinton reactors. This is
not effective, and may have contributed to the poor
performance of the ultrasonics.
The Suffern units each have one 1.5 kW transducer
per two banks of lamps, or two per unit. The ratio of
ultrasonic to lamp power is approximately 0.14, or
half that at Vinton. A limited series of tests were
recently performed at this plant to evaluate the
effectiveness of the ultrasonics (4). The tests were
conducted over a two week period, using unit 1
without ultrasonics and unit 2 with ultrasonics. The
ultrasonics were operated on a cycle of 30 minutes on
and 30 minutes off (50 percent on-time). Flow was
generally split evenly between the two units. As
discussed earlier, the quartz will accumulate buildup
differently, depending on whether the lamps remain
on or off for extended periods of time. With the lamps
on, an inorganic scale tends to develop; with the
lamps off, the quartz simply provide a surface for a
biological film to develop. The ultrasonics were
evaluated as to the ability to retard the buildup of
either type of material. In either unit a select number
of quartz sleeves were monitored in a bank where the
lamps were kept off and in a bank where the lamps
were continuously operated.
The results of the evaluation, taken from the study
report, are presented on Figure 7-55. The dashed
lines are from Unit 1, which operated without the
ultrasonics; there appears to be no significant dif-
ference between the banks with the lights on or off. It
may be said that the bank with the lights on tended to
degrade at a faster rate than the bank with no lamps
on. By the fourteenth day, however, the transmittance
of the quartz sleeves in either bank was between 20
and 30 percent of the initial transmittance.
The results of the test in Unit 2, which operated with
the ultrasonics, show a significant difference be-
tween the banks with or without the lamps on. With
the lamps off, the ultrasonics appear to have been
very effective; after 14 days, the quartz transmittance
was still approximately 70 percent of the initial
transmittance. In the bank which had the lamps on,
however, the ultrasonics was ineffective; by day
eight, the transmittance was between 35 and 40
percent of the initial transmittance. It appears that the
ultrasonics are unable to retard the softening effect in
which the inorganic carbonates plate out on the
237
-------
Figure 7-65. Comparison of ultrasonic cleaning performance at Suffern, New York (4).
100
W/Ultrasonics, UV On
Unit 2, Bank B
7 8
Time (days)
12
13
14
quartz surface. This phenomenon will occur only
when the lamps are on.
In all, current experience has not confirmed the utility
of ultrasonics as a cleaning device. This would
especially be the case in areas with hard waters.
Additionally, consideration should be given to the
cost of the accessory device on the basis of both
additional capital cost and added operating cost due
to its energy requirements.
High Pressure Wash. This is closely related to the
earlier discussions of routine chemical cleaning.
Commercially available industrial cleaning units are
available which use pressurized water to clean
surfaces. The type of hose nozzle or spray wand used
to discharge the water will vary and will depend on
the application. In the Teflon units, a nozzle is fitted to
the end of a flexible hose; spray is directed radially
from the nozzle as the hose is snaked down the Teflon
tubes. The systems have the capability of educting a
detergent or acid solution into the wash stream. The
same type of system would be applicable to quartz
systems. In this case, however, a wand type sprayer
would be more appropriate to reach the insides of the
lamp battery. The system is inexpensive, easy to use,
and recommended for most UV reactor applications.
Equipment manufacturers for the Teflon tube systems
have been supplying a high-pressure nozzle spray
cleaning system with several recent installations.
Kreft et al. (4) reported on visits to wastewater
treatment plants at Chinook, Montana and Rock
Springs, Wyoming to observe performance of the
cleaning systems. The Rock Springs UV disinfection
unit had moderate to heavily-fouled Teflon tubes
(black and brown coatings) caused by heavy solids
and foam and grease carry-over into the tubes.
During cleaning, it was noted that the high-pressure
nozzle system was able to remove, in some cases.
235
-------
significant portions of the fouled material from the
inside surfaces of the Teflon, but was not consistent
for all tubes. In some tubes, very little material was
removed and, therefore, the UV transmittance would
be assumed to still be minimal. A series of repeated
cleanings may be necessary to remove the material
from all tubes. This is a critical point; if a few tubes
remain dirty, their consequent poor performance can
have a dramatic effect on the overall performance of
the system.
At Chinook, the amount of fouling on the tubes was
not as heavy as at Rock Springs and the high pressure
cleaning system appeared to remove most of the
material attached to the inside of the Teflon tubes. A
sample of a cleaned tube was measured by the
chemical actinometry method along with the used
samples that were taken from the Chinook plant. It
was verified, as shown in Table 7-17, that the in-.
place cleaning did return the Teflon tube to a UV
transmittance in the range of 65 to 70 percent from
an uncleaned transmittance of 30 to 55 percent.
The authors concluded that the high-pressure nozzle
washing systems have some benefit in helping to
clean the interior surfaces of Teflon tubes. Since the
Teflon tubes are themselves usually difficult to
access, the nozzle cleaning system offers some
advantages and provides the operator with a simpler
cleaning task. They did indicate, however, that
internal swabbing of the Teflon tubes with a soft rag,
and possibly with a detergent, will be necessary on an
occasional basis to ensure that the Teflon tubes
remain in a fairly clean state. It was also noted that
many of the plants visited that had Teflon UV units
had noticeable amounts of dust on the outside
surfaces of the tubes. It is highly recommended that
in all installations, a routine maintenance task must
be to clean the outside with a rag and, possibly
isopropyl alcohol or water; this will improve the
transmission of the UV light through the Teflon into
the wastewater.
7.5.2.3 Other Design Elements for Effective
Maintenance
The following observations are made on the basis of
current full-scale and pilot scale operating experi-
ences. These are directed to considerations for
design, fabrication, and installation which will ease
maintenance tasks or provide for more effective
maintenance:
a. The reactors and related tankage should be
equipped with drains which will allow for
complete and rapid dewatering. Drainage should
be to the main plant drainage system.
b. A clean water supply should be permanently
available, in addition to all requirements for
chemical cleaning.
c. The systems should be designed modularly with
the ability to readily isolate a module from the
plant flow.
d. A bypass should be constructed around the
entire UV disinfection system, particularly in
plants which require only seasonal disinfection.
This would allow for greater convenience for
maintenance tasks during the non-disinfection
season.
e. The accessibility to the lamps, quartz sleeves,
and Teflon tubes is critical to the ease of
maintenance. Manways should be provided on
larger scale systems.
f. Strict inventories should be kept of the lamps in
use, their relative output, and their estimated
cumulative operating life. This should also apply
to a more limited extent to the quartz sheaths,
Teflon tubes, and ballasts.
g. The reactors and all other accessory equipment
should be installed in an area that is adequate to
accomplish all the required maintenance tasks.
The systems should not be so cramped that it is
virtually impossible to work on the units.
h. If reactors are taken out of service, they should
always be drained; a clean water rinse would
also be appropriate. The units should then be
held in a drained, dry condition.
7.5.3 System Components
The major components of the UV systems are the
lamps, enclosures, and the ballasts. Discussion of
these have been interspersed throughout this text.
The following observations are made to highlight the
major points to consider when evaluating or designing
a new system, or in the operations of an existing
system.
a. The low pressure mercury arc lamps are cur-
rently the most efficient source of UV radiation.
Costwise, the longer arc length lamps are more
efficient. Because of the negative effects of the
ozone produced by the 185 nm light, lamps with
fused quartz envelopes are not recommended.
The lamps of vycor or other high transmission
glass are appropriate.
b. Care should be taken to minimize temperature
effects. In quartz systems, O-ring spacers should
, be slipped over the lamps to prevent direct
contact with the cooler quartz sleeve. These are
generally provided in the newer systems.
c. Fittings holding the quartz sleeve should be tight
and leakproof. A number of plants have had
problems with leaks at these points, causing
239
-------
electrical hazards and corrosion. Additionally,
some designs are difficult to disassemble and
reassemble, causing excessive labor, quartz
breakage, and continuing leaks.
d. The quartz sheaths are fairly standard in com-
position. Variations have been noted in the wall
thickness; attention should be paid to the
structural strength of the sleeve. Another
variation is the single end quartz in which only
one end of the quartz is open; the other end of
the quartz is fused close. A plant which will use
this type of quartz is the Nine Springs plant in
Madison Wl. The wiring from the lamp electrode
at the fused end of the quartz will be snaked
back to the open end of the quartz.
e. Frequent on-off cycles for the lamps will shorten
their life. More ejffective lamp control may be
accomplished by voltage dimming, in conjunc-
tion with on-off control of banks of lamps.
f. Control panels should be remote from the UV
reactor.
g. The ballasts must be properly mated with the
lamps being used. It is strongly recommended
that both the lamp and ballast manufacturers be
consulted on this aspect. It would be appropriate
to require certification that the ballast is correct
for the UV lamp. The ballasts should be thermally
protected; this forces the ballast to shutdown if
it overheats.
h. The power panel containing the ballasts must
have adequate ventilation to discharge the heat
generated by the ballasts. This has been a
recurring problem at full-scale installations. The
life of the ballast is greatly shortened and in
several cases excessive heat build-up caused
rapid failure of a number of ballasts. High
volume ventilation fans should be installed in
the power panels to cool the ballasts. During
warm temperature months, this should be
vented out of the building.
i. Careful attention should be paid to the electrical
wiring of the UV systems, at the points of both
fabrication and installation. Improper wiring at
several plants resulted in electrical hazards,
component failures, and in some cases, small
electrical fires. The wiring should be properly
sized and the wire covering should be resistant
to UV radiation effects. Typically, Teflon coated
wiring is specified.
j. The Teflon tubes are generally standard. The
variables are the wall thickness and the tube
diameter. The transmittance will decrease with
increasing wall thickness; very thin walls will
limit the structural integrity of the tube, how-
ever, causing them to collapse. Greater inten-
sities can be achieved with smaller diameter
tubes; this will have to be weighed against
increased head losses.
k. Air bleeds should be considered on certain
Teflon systems (in particular the pressure units)
to minimize air binding in the Teflon tubes.
I. Removeable screens should be placed upstream
of the lamp reactor to prevent large debris from
entering the system. This is especially important
for quartz systems.
7.5.4 Monitoring and Control
System Controls. The sophistication of the monitor-
ing and control systems for the UV process can vary
from minimal to fully automated. This is no different
from any other unit operation in a treatment facility. It
is recommended that the minimum should always be
provided; any increased capability should then be
considered on a cost-benefit basis. The minimal
requirements suggested for the UV disinfection
process are flow metering per unit, individual lamp
operating monitors, a portable radiometer, power
panel temperature (with alarm), and the ability to turn
portions of the system on and off on the basis of time.
The following observations are made, again as points
which should be considered when evaluating or
designing a UV system:
a. The units should be arranged such that banks of
lamps can be shut off or on. In the simplest mode
this can be controlled by timers. Modest adjust-
ments can then be made on a diurnal basis to
reflect the normal variation in the plant's flow.
This can be further advanced by automatically
slaving the lamp bank operations to the plant
flow; some systems also will adjust the unit
voltage, using bank shutoffs as a gross adjust-
ment.
b. Concurrent with the plant flow, the control of
the system can be coupled to the water quality.
This is done in some systems by use of an
intensity monitor fixed to the reactor; this may
cause problems if it cannot be effectively
maintained. An alternative method is to utilize a
continuous monitor of the wastewater UV
absorbance.
c. As had been discussed earlier, it is important
that a continuing record be kept of the average
output of the lamps, and the transmittance of
the quartz and Teflon. Procedures have been
given by which to monitor these parameters. It is
strongly recommended that these become part
of the system's routine O&M, and that the
necessary equipment be available to accomplish
these tasks.
240
-------
d. In-line pilot light monitors should be installed in
the control panels to indicate each lamp opera-
tion. This is normally available on most UV
systems. Alarms should be installed to alert the
operator if a preset number of lamps fail.
e. The flow to each module should be metered.
f. Elapsed time monitors (non-resetable) should
be installed for each bank of lamps. This will
allow an accurate accounting of cumulative
operating time and will allow the operator to
balance the use of the various lamp banks in a
system.
g. The temperature in the ballast power panel
should be monitored. An alarm system should
be available to alert the operator if the panel
temperatures exceed an acceptable level.
h. A watt-meter would be useful, particularly in
larger systems, to monitor on a continuous basis
the power requirements relative to the remain-
der of the plant.
7.5.4.2 Wastewater Monitoring for an Existing
System
The UV disinfection process does, not have the
monitoring advantage of a measurable residual (as is
the case with chlorine). As such, greater care is
required in controlling operations efficiently and still
maintaining performance. It is strongly recommended
that this entail frequent sampling and analysis. A
suggested protocol is to sample the system a mini-
mum of three times per week (alternating days)
between the hours of 10 a.m. and 3 p.m. The influent
to the UV system should be analyzed for suspended
solids, UV absorbance, and cpliform density (i.e., the
bacterial monitor prescribed by the plant's permit).
Additionally, the flow rate should be recorded at the
time of sampling, as well as the operating conditions
of the reactor (number of lamps in operation, etc.).
The effluent sample should be analyzed for coliform
density.
An important note should be made with regard to the
subsequent handling of the exposed UV effluent
sample. All precautions should be taken to protect the
sample from exposure to visible light (sunlight and
normal fluorescent and incandescent light), before
the sample has been set and put into incubation. This
will prevent the occurrence of photoreactivation.
Normal precautions include sampling in an opaque
(and covered in foil) sample bottle, and keeping the
sample covered during the procedure to set the
sample (filtration and plating for the MF procedure,
and dilutions and inoculation for the MPN procedure).
Alternatively, if photoreactivation is to be accounted
for by the permit requirements, the effluent should be
taken with the transparent glass bottle and left in
direct sunlight for approximately ond hour. Enumera-
tion would then be accomplished by routine pro-
cedures. Either the direct Membrane Filter (MF) or the
Most Probable Number (MPN) procedures have been
demonstrated to yield equivalent recoveries (65).
Collection of such a relatively comprehensive data set
is felt to be important in controlling the operations of a
UV system. It allows direct evaluation of the system
performance under current wastewater conditions
and provides a data base from which the disinfection
model can be calibrated and/or refined. The model
itself then becomes an excellent tool in controlling
the system and optimizing operations for maximum
use of lamps and minimal use of energy. The
continuous collection of the appropriate data also
allows a rational approach to troubleshooting the
non-performance of a system.
7.5.5 Safety Considerations
Ultraviolet disinfection is basically a safe process; the
activity is generated on-site; thus/there are no
transport concerns to or from the site, or concerns
regarding storage of reactive material. Normal plant
safety precautions apply relative to physical layout
(railings, etc.) and to electrical hazards. Power
supplies are high voltage, requiring the adherence to
normal electrical safety codes. Electrical interlocks
should be provided to shut off systems when opened
(reactor end panels); particular attention should be
paid to electrical wiring, groundings, and water-
proofing.
The storage, handling, and disposal of the expendable
components should also be considered from the
standpoint of safety. Storage of lamps, quartz sleeves,
and ballasts should be in a separate dry area.
Adequate shelving should be designed to store the
materials such that they are protected from breakage,
and are easily and safely accessed. Used lamps,
quartz sleeves, and ballasts which are to be discarded
should be repackaged and overpacked for safe
disposal.
Personnel safety training should address and require
strict adherence to personal protection from excessive
UV radiation. A lamp battery would not present a
hazard while submerged and operational; the water
absorbance will sufficiently attenuate the radiation.
These lamp batteries should not be operated while in
a dewatered and dry state. Similarly, "dry lamp"
systems such as the Teflon unit, should have all
covers in-place during operation. Plastic (e.g., plexi-
glas) will not transmit the 253.7 nm wavelength; this
material can be used for the end plates (windows or
end plates) or unit shields to protect against exposure
but still allow visual inspection of the lamp ends. If it is
necessary to engage a system without shields .in
241
-------
place (or with the lamp battery exposed) it is absolutely
necessary that the proper protective gear be worn by
all personnel in the area and that adequate warning
signals be active during these operations to warn
anyone entering the area. This will also apply during
the routine lamp monitoring tasks discussed earlier.
The skin and eyes readily absorb UV radiation and are
particularly vulnerable to injury. Sunburn (erythema)
is a common example, although this effect is most
pronounced with UV light between wavelengths 285
and 300 nm. Absorption by the mucous membranes
of the eye and eyelids can cause conjunctivitis
{commonly referred to as "welders flash"). The injury
becomes apparent 6 to 12 hours after exposure;
although painful and incapacitating, the damage is
usually temporary.
Personal protection must include plastic goggles
(wrap around), or face-shields. These must be rated to
absorb the UV spectral lines. Protective clothing
should be worn to prevent exposure to the hands,
arms, and face.
7.5.5 Facilities Requirements for Full Scale
Installations
The facilities needs are divided to two specific areas:
equipment and the physical plant. The equipment
elements include the hardware requirements directly
associated with the installation of the UV process.
The installation at a given plant site must then
address the hookups and physical plant needs to
install the system.
7.5.6.1 Equipment
The equipment generally supplied through the vendor
include the UV reactor itself and the ancillary
equipment used to control and monitor the system:
a. UV lamp battery (UV reactor)
b. Power supply and power panels (with single
point hookup to 'the plant power)
c. Instrumentation for the control and monitoring
of the system; this generally includes a UV
intensity monitor per module, pilot monitors for
each lamp, controller to direct the number of
lamp banks on as a function of flow (and/or
water quality), and alarms to signal deficiencies
in lamp operation and/or performance.
d. Accessory cleaning equipment; this is generally
in the form of chemical cleaning, mechanical
wipers, or ultrasonics.
e. Manufacturer's engineering and startup ser-
vices are generally at the option of the buyer.
f. Replacement parts supplied with purchase; this
should include no more than 100 percent lamps,
50 percent ballasts, and 20 percent quartz or
Teflon tubes. At minimum, this should be 20
percent, 10 percent and 10 percent, respec-
tively.
When specifying a UV system, the design should
analyze the cost of the.ancillary monitoring and
cleaning devices to be provided with the reactor
against this need and benefit. These can account for
15 to 25 percent of the total equipment costs.
7.5.6.2 Physical Plant
The structural and installation requirements are
sensitive to the specifics of the site and the equipment
to be installed. It is difficult to give a detailed
assessment of these needs; by way of guidance,
however, a generic installation is considered. .
There is a basic space requirement for the UV system
at a plant, based on the number of lamp modules to be
installed. The smallest plants would generally require
only one module; the reactor itself is rather small. A
minimum space of 10 m2 (108 sq ft) should be
allowed, however, for the unit. It is further suggested
that this space allocation be increased to 25 ma(270
sq ft) for large modules which may contain several
hundred lamps. Thus, a 500 kW system may contain
approximately 6,000 lamps in 10 modules; the total
space allocation would be approximately 250 m2
(2,700 sq ft). In most cases, the system should be
housed in a standard building. Certain configurationis
proposed by engineers/manufacturers (including
existing full-scale systems), do not require such
housing. The designs generally call for open channel
installations of the UV equipment. The power supplies
and instrumentation in this case require more
stringent specifications with regard to weatherproof-
ing and protection against water/electrical hazards.
Reactors which are housed are generally character-
ized by piped inlet and outlet structures. The reactors
are typically sealed units or have integral influent and
effluent tanks attached to the lamp battery. The
housing itself may be shared with other unit opera-
tions in the plant. Power supplies and control systems
are contained in the building, typically remote and
elevated from the reactors. Storage for spare parts
(lamps, ballasts, etc.) can also be accomplished within
the building. The entire area should be adequately
ventilated, particularly with regard to humidity
control and venting from the power panels.
7.5.7 Estimating O&M Requirements
UV is a capital intensive process, with the equipment
requirements directly proportional to the design peak
hydraulic and performance needs. The operational
and maintenance needs, however, are reflected more
by the average utilization of the system. In fact, a key
operational consideration is to use only that portion of
the system necessary to meet current performances
242
-------
demands. Over-utilization of the system, in an
attempt to simplify operations, will have a significant
impact on the costs to operate the process. The
following discussions focus on the three major
elements which comprise the costs associated with
the operation and maintenance of a UV system.
7.5.7.1 Labor Requirements
Estimating labor requirements is a subjective task,
relying on current experiences and being selective in
defining the tasks which should be assigned to the UV
process. In estimating labor needs, Scheible et al. (54)
assessed the experience in the operation of the Port
Richmond plant, as well as previous studies, and the
current experience at full scale facilities. The esti-
mates are summarized in graphical form on Figure
7-56. As shown, the labor is divided to three major
categories: direct UV operation and maintenance
tasks; general maintenance; and system overhaul. It
is important to note that the labor estimates are based
on the O&M requirements for the entire installed
system.
Direct UV Operation and Maintenance. The tasks
which are considered in this category may be
described as follows:
1. Operations and Monitoring
• daily systems checks for proper operation;
• appropriate recording of data (lamps in opera-
tion, meter readings, power readings, flow
rates, water quality readings, temperatures,
etc.);
e sampling and analysis for SS, bacterial
density, UV absorbance;
• direct manual control of the systems, or the
monitoring and control of automatic opera-
tional instrumentation.
2. Maintenance
• checking and maintaining system compo-
nents (lubrication, etc.);
• storage and maintenance of appropriate parts
inventory;
• routine systems cleaning, the labor associ-
ated with this task will include monitoring of
the quartz/Teflon surfaces, switching sys-
tems during special cleaning cycles, and
maintenance of the chemical feed systems;
• replacement of worn or broken components
in the system.
The labor needs assigned to these direct O&M tasks
are estimated to range from 2-3 hr/wk for small
systems (less than 100 lamps) to 15-30 hr/wk for
larger plants (greater than 1,500 lamps).
General Maintenance. As had been discussed,
there are space, building, and ventilation require-
ments associated with the installation of a UV
process. The general maintenance of these physical
facilities will be required. For purposes of this
discussion, the labor is assigned to the labor require-
ments for the disinfection process at a plant. It is
suggested that approximately one-half the labor
required for the direct O&M tasks discussed above be
assumed for the general maintenance tasks.
System Overhaul. The reader is referred to the
earlier discussions which dealt with the cleaning and
direct measurement of the lamps and the quartz/
Teflon which comprise the UV reactor. It is strongly
recommended that the entire system be broken down
on a yearly basis to accomplish the following tasks:
a. clean the outside surface of each lamp;
b. clean the inside surfaces of each quartz sleeve,
and the outside surfaces of the Teflon tubes;
c. measure each lamp for relative UV output;
replace those which fall below a specified level;
d. measure a representative sampling of the
quartz/Teflon enclosures for transmittance,
replace those which are worn excessively; and
e. check internal components for wear and replace
if necessary.
These tasks are suggested to serve as an efficient
means to control the system's output and energy
efficiency at acceptable levels. By having direct
measurements of the unit's average output, the
lamps can_bejJtilized to their maximum life. Keeping
the surface clean will allow for efficient use of the UV
energy.
Based on the experiences of Port Richmond, the labor
required to accomplish the system overhaul each
year is estimated to be approximately 16 hr/100
lamps. Relative to the total labor requirements for the
UV process, the system overhaul is small, but can
yield significant overall O&M cost savings.
Total Labor Estimates. The total yearly estimated
labor requirement is presented on Figure 7-56. Note
that these are based on year-round disinfection. In
cases where seasonal disinfection is allowed, the
labor estimates for the direct O&M, and the general
maintenance tasks would be reduced; the system
overhaul is still recommended on a yearly basis.
Overall, the labor needs for the UV process are
relatively low, ranging from approximately 40 man-
days/yr for a small 10 kW (120 -lamps) system to
approximately 400 mandays/yr for a 400 kW system
(5,000 lamps).
243
-------
Figure 7-56. Estimate of labor requirements for the operation and maintenance of UV systems (54).
1000
30
60
Approximate Number of Lamps (1.5m Arc)
120 300 600 1200
3000
6000
Estimated Total Yearly Labor
Note: Labor Based on
Total System KW.
I I
6 8 10
7.6.7.2 Materials
The major materials cost associated with the UV
system are the lamps, the ballasts, and the quartz or
Teflon enclosures. Note that these requirements are
considered as a function of the annual average
system utilization (kW). Thus, although the system
may be sized to meet the peak power demand, the
need to replace the mayor expendable components of
the system will depend on their actual use; this can be
represented by the estimated annual average utiliza-
tion of the system. To estimate the annual average
requirement for material, the following suggestions
are offered:
20 40 60'
System Size, Total KW
100
200
400 600 1000
Lamps. Low pressure mercury arc (1.5 m/arc) are
standard. This replacement cycle should be assessed
at one year (8,700 hours). This is conservative; there
are cases where considerably longer life cycles have
been demonstrated.
Ballasts. A single ballast serves two lamps. The
average life cycle is five years; this can be considered
conservative if the ballast is properly mated and the
power panel is properly ventilated to prevent over-
heating.
Quartz/Teflon Enclosures. The estimate should
assume one quartz sleeve per lamp, and one Teflon
244
-------
tube (3 m) per two lamps. An average life cycle of five
years is suggested for both, although there is little
demonstrated experience in this regard.
Miscellaneous. To account for miscellaneous parts
replacement, a cost equivalent to five percent of the
annual lamps, ballasts, and enclosures costs are
suggested.
7.5.7.3 Power Requirements
The third element, power, should also be estimated
on the basis of the annual average system utilization.
This requirement can be accounted for by addressing
the lamps only. Ancillary power use is relatively
insignificant, except in cases where ultrasonic de-
vices are used for cleaning. The total power per lamp
is 80 Watts (for 1.5 m arc lamps), including the
ballast.
7.5.7.4 Estimating Average Annual Utilization
As discussed, the materials and power requirements
(and to a lesser extent, the labor needs) should be
based on the annual average utilization. An example
of this analysis was provided in Section 7.4 for the
design example. It is based on the design needs for
average wastewater conditions (f low, UVabsorbance
coefficient, initial density, and suspended solids).
This can be significantly less than the peak system
requirement (20 to 30 percent of peak), particularly
with plants which are not at capacity and in cases
when only seasonal disinfection is required.
/
7.6 Referenced
1. Disinfection of Wastewater—Task Force Report,
EPA-430/9-75-013, U.S. Environmental Pro-
tection Agency, Washington, DC, 1976.
2. Environment Canada. Wastewater Disinfection
in Canada. A Background Paper Prepared by the
WPC Directorate, Environmental Protection
Service, Report No. EPS 3-WP-78-4, 1978.
3. International Joint Commission. IJC Chlorine
Objective Task Force Final Report, 1976.
4. Kreft, P., Scheible, O.K. and A.D. Venosa.
Hydraulic Studies and Cleaning Evaluations of
Ultraviolet Disinfection Units. Presented at the
58th Annual Conference, Water Pollution Con-
trol Federation, Kansas City, MO, 1985.
5. J. M. Montgomery Engineers, Inc. Ultraviolet
Disinfection; I/A Technology Assessment.
Summary List of Facilities in the USA or Canada
Utilizing UV Light Disinfection. Contract No. 68-
03-1821, U.S. Environmental Protection
Agency, Cincinnati, OH, 1984.
6. Venosa, Albert D. Current State-of-the-Art of
Wastewater Disinfection. JWPCF 55(5), 1983.,
7. Kohler, Lewis R. Ultraviolet Radiation, Second
Edition. John Wiley & Sons, Inc., New York, NY,
1965.
8. Waymouth, J.F. Electric Discharge Lamps. The
M.I.T. Press, Cambridge, MA, 1971.
9. Downes, A., and Blount, T. Research on the
Effect of Light Upon Bacteria and Other Orga-
nisms. Proc. Roy. Soc., London, 26,488,1877.
10. Jagger, J. Introduction to Research in Ultraviolet
Photobiology. Prentice-Hall, Inc., Englewood
Cliffs, NJ, 1967.
11. Setlow, R.B. and Setlow, J.K. Effect of Radiation
on Polynucleotides. Annual Reviews of Bio-
physics and Bioengineering 1:293-346, 1972.
12. Smith, K.C. and Hanawalt, P.C. Molecular
Photobiology, Inactivation and Recovery. Aca-
demic Press, New York, NY, 1969.
13. Harm, W. Biological Effects of Ultraviolet Radi-
ation. Cambridge University Press, Cambridge,
England, 1980.
14. Stanier, R., Doudoroff, M., and Adelburg, E. The
Microbial World. Prentice-Hall, Englewood
Cliffs, NJ, 1970.
15. Kirby-Smith, J.S. and Craig, D.L. Genetics
42:176-187,1957.
16. Loofbourow,J.R. Effects of Ultraviolet Radiation
on Cell Growth. 12:75-149, 1948.
1/7. Oda,A. Ultraviolet Disinfection of Potable Water
Supplies. Ontario Water Resources Commis-
sion, Division of Research, Paper 2012, 1969.
18. Bernstein, I.A. Biological Influences on Envi-
ronmental Toxicity. Deeds, and Data, Water
Pollution Control Federation, 18(5), 1981.
19. Jagger, J. and R.S. Stafford. Journal Biophysi-
ology5:75, 1965.
20. Kelner, A. Proceedings of National Academy
of Science, U.S. 35, 73, 1949.
21. Dulbecco, R. J. Bacteriology 59, 329, 1950.
22. Harm, W., Rupert C.S. and H. Harm. Photo-
physiology, 6:279, 1971.
23. Harm, H. Contributor to Photochemistry and
Photobiology of Nucleic Acids, Volume II. Aca-
demic Press, Inc., New York, NY, 1976.
245
-------
24. Novick, A. and L Szilard. Proceedings National
Academy of Science, U.S. 35, 591,19,49.
25. Kelly, C.B. Disinfection of Sea Water by Ultra-
violet Radiation, American Journal of Public
Health 51(11), 1981.
26. Huff, C.B., Smith, H.F., Boring, W.D. and N.A.
Clarke. Study of Ultraviolet Disinfection of
Water and Factors in Treatment Efficiency.
Public Health Report 80(8), 1965.
27. Hill, W.F., Akin, IE.W., Benton, W.H. and F.E.
Hamblet. Viral Disinfection of Estuarine Water
by UV. Journal Sanitary Engineering Division,
ASCE, SA5, 1971.
28. Roeber, J.A. and F.M. Hoot. Ultraviolet Disin-
fection of Activated Sludge Effluent Discharging
to Shellfish Waters. EPA-600/2-75-060, NTIS
No. PB-249460, U.S. Environmental Protection
Agency, Cincinnati, OH, 1975.
29. Singer, M. and N. Nash. Ultraviolet Disinfection
of a Step Aeration Effluent. Presented at Winter
Meeting of the New York Water Pollution
Control Association, 1977.
30. Oliver, B.G. and J.H. Carey. A Scale-up Investi-
gation of Ultraviolet Disinfection as an Alterna-
tive to Chlorination for Sewage Effluents. The
Canadian Journal of Chemical Engineering,
Volume 53, 1975.
31. Oliver, B.G. and J.H. Carey. Ultraviolet Disin-
fection: An Alternative of Chlorination. JWPCF
48(11):2619, 1976.
32. Oliver, B.G. and E.G. Cosgrove. The Disinfection
of Sewage Treatment Plant Effluents Using
Ultraviolet Light. The Canadian Journal of
Chemical Engineering, 53, 1975.
33. Drehwing, F.J., Oliver, A.J., MacArthur, D.A.
and P.E. Moffa. Disinfection/Treatment of
Combined Sewer Overflows, Syracuse, New
York. Report of Project Number S802400, U.S.
Environmental Protection Agency, Cincinnati,
OH, 1978.
34. Petrasek, A.C., Jr., Wolf, H.W., Elsmond, S.E.
and D.C. Andrews. Ultraviolet Disinfection of
Municipal Wastewater Effluents. EPA-600/2-
80-102, NTIS No. PB81 -111049, U.S. Environ-
mental Protection Agency, Cincinnati, OH,
1980.
35. Scheible, O.K. and C.D. Bassell. Ultraviolet
Disinfection of a Secondary Wastewater Treat-
ment Plant Effluent. In: Proceedings of Waste-
water Disinfection Alternatives—State-of-the-
Art Workshop, 52nd Annual WPCF Conference,
Houston, TX, 1979.
36. Scheible, O.K., and C.D. Bassell. Ultraviolet
Disinfection of a Secondary Wastewater Treat-
ment Plant Effluent. EPA-600/2-81-152, NTIS
No. PB81-242125, USEPA, Cincinnati, OH,
1981.
37. Severin, B.F. Disinfection of Municipal Waste-
water Effluents with Ultraviolet Light. JWPCF
52(7), 1980.
38. Johnson, J.D. and R.G. Quails. Ultraviolet
Disinfection of a Secondary Effluent: Meas-
urement of Dose and Effects of Filtration. EPA-
600/2-84-160, NTIS No. PB85-114023, U.S.
Environmental Protection Agency, Cincinnati,
OH, 1984.
39. Quails, R.G., Flynn, M.P. and J.D. Johnson. The
Role of Suspended Solids in Ultraviolet Disin-
fection. JWPCF 55(10), 1983.
40. Quails, R.G. and J.D. Johnson. Bioassay and
Dose Measurement in Ultraviolet Disinfection.
Applied and Environmental Microbiology 45:
872,1982.
\ ,
41. Bellen, Gordon E., Randy A. Cottier, Ronald
Dormand-Herra. Survey and Evaluation of Cur-
rently Available Water Disinfection Technology
Suitable for Passenger Cruise Vessel Use.
National Sanitation Foundation report to Center
for Prevention Services, Center for Disease
Control, Atlanta, GA, Contract No. 200-80-
0535, 1981.
42. Haas, C.N. and G.P. Sakeltaropoulos. Rational
Analysis of UV Disinfection Reactors. Presented
at ASCE National Conference of Environmental
Engineers, San Francisco, CA, 1979.
43. Severin, B.F., Suidan, M.T., Englebrecht,.R.S.
Mixing Effects in UV Disinfection. JWPCF 56(7),
1984.
44. Severin, B.F., Suidan, M.T., Rittmann, B.E.,
Englebrecht, R.S. Inactivation Kinetics in a
Flow-Through UV Reactor. JWPCF 56(2), 1984.
45. Nehm, P.H. Operating Experience Disinfection
Secondary Effluent with Pilot Scale Ultraviolet
Units. In: Municipal Wastewater Disinfection
Proceedings of Second National Symposium,
246
-------
Orlando, FL EPA-600/9-83-009, NTIS No.
PB83-263848, U.S. Environmental Protection
Agency, Cincinnati, OH, 1983.
46. Ho, L.W.A. and P. Bohm. UV Disinfection of
Tertiary and Secondary Effluents. Water Pollu-
tion Research Journal, Canada, 16, 33, 1981.
47. Bohm, P., Ho, K.W.A. and J.E. Pagel. Application
of UV Disinfection Technology in Ontario Water
Pollution Control Plant Effluents. Ontario Min-
istry of the Environment, Toronto, 1981.
48. Whitby, G.E., Palmateer, G., Cook,. W.G.,
Maarschalkerweerd, J., Huber, D. and K. Flood.
Ultraviolet Disinfection of Secondary Effluent.
JWPCF 56(7), 1984.
49. Kirkwold, D. Disinfecting with Ultraviolet Radi-
ation. Civil Engineering Magazine, ASCE, pg.
62, December, -1984.
50. Scheible, O.K., Forndran, A. and W.M. Leo. Pilot
Investigation of Ultraviolet Wastewater Disin-
fection at the New York City Port Richmond
Plant. In: Municipal Wastewater Disinfection—
Proceedings of Second National Symposium,
Orlando, FL. EPA-600/9-83-009, NTIS. No.
PB83-263848, U.S. Environmental Protection
Agency, Cincinnati, OH, 1983.
\
51.. Scheible, O.K. 'Design and Operation of UV
Disinfection Systems. In: Proceedings of the
Pre-Conference Workshop on Wastewater Dis-
infection, 56th Annual WPCF Conference,
Atlanta, GA, 1983.
52. Scheible, O.K., Casey, M.C. and A. Forndran.
Ultraviolet Disinfection of Wastewaters from
Secondary Effluent and Combined Sewer Over-
flows. EPA-600/2-86/005, NTIS No. PB86-
1 451 82, U.S. Environmental Protection Agency,
Cincinnati, OH, 1986.
53. White, S.C., Jernigan, E.B. and A.D. Venosa. A
Study of Operational Ultraviolet Disinfection
Equipment at Secondary Treatment Plants.
Presented at the 58th Annual Conference,
Water Pollution Control Federation, Kansas
City, MO, 1985. '
54. Scheible, O.K., Kreft, P. and A.D. Venosa.
Demonstration of Process Design and Evalua-
tion Procedures for Ultraviolet Disinfection.
Summary prepared for U.S. Environmental
Protection Agency, WERL, Cincinnati, OH,
55. Scheible, O.K. Pilot Studies of UV Disinfection
at Several Wastewater Treatment Plants.
Summary report, U.S. Environmental Protection
Agency, WERL, Cincinnati, OH, Cooperative
Agreement CR 808978, 1986.
56. Scheible, O.K. Development of a Rationally
Based Design Protocol for "the Ultraviolet Light
Disinfection Process. Presented at the 58th
Annual Conference, Water Pollution Control
Federation, Kansas City, MO, 1985.
57. Rebhun, M. and Y. Argaman. Evaluation of
Hydraulic Efficiency of Sedimentation Basins.
Journal SED, ASCE 91 (SA5):37, ] 965.
58. Levenspiel, 0. Chemical Reaction Engineering,
2nd Edition. John Wiley and Sons, New York,
NY, 1972.
59. HydroQual, Inc. Evaluation of the UV Disinfec-
tion Process at the Northfield Water Pollution
Control Plant. Report to the City of Northfield,
MN, 1984.
60. Baxendale, J.H. and N.R. Bridges. The Photo-
reduction of Some Ferric Compounds in
Aqueous Solutions. Journal Physical Chemistry
59:783, 1955.
61. Parker, C.A< A New Sensitive Chemical Acti-
nometer, Parts I and II. Proc. Royal Soc. London
Abs. 235:518, 1956.
62. HydroQual, Inc. Trip report to the USEPA,
Cincinnati, OH, summarizing visit to Sents
Creek WPCP, Pella, IA. Cooperative Agreement
No. CR 808978, 1983.
63. HydroQual, Inc. Trip report to the USEPA,
Cincinnati, OH, summarizing visittoTillsonburg
WPCP, Tillsonburg, Ontario. Cooperative Agree-
ment No. 808978, 1983.
64. HydroQual, Inc. Trtp report to the USEPA,
Cincinnati, OH, summarizing visit to Vinton
WPfeP, Vinton, IA. Cooperative Agreement No.
808978,1983.
65. Quails, R.G., Chang, J.C.H., Ossoff, S.F. and
J.D. Johnson. Comparison of Methods of
Enumerating Coliforms After UV Disinfection.
Applied and Environmental Microbiology 48(4):
699-701,1984.
^66. Jaceb, S.M., and J.S. Dranoff. Light Intensity
Profiles in a Perfectly Mixed Photoreactor.
Journal AlChE 16(3), 1970.
67. Rubert, J.S. Journal Genera! Physiology45:725,
4-9S&
68. Luckieseh, M., Taylor, A., and G. Kerr. Germi-
cidal Energy. General Electric Review, 1944.
247
-------
------- |