Вы находитесь на странице: 1из 264


October 1986

Design Manual

Municipal Wastewater Disinfection

U.S. Environmental Protection Agency

Office of Research and Development
Water Engineering Research Laboratory
Genter for Environmental Researc:h Information
Cincinnati, OH 45268

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 oftrade names or commercial products does not constitute
endorsement or recommendation for use.


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. Ozclne 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
6.6i Safety ...................................... 151
6.7 References ...................................... 153

7. Ultraviolet Radiation ...................................... 157

7.1 Introduction ............................................ 157

7.2: Disinfection of Wastewaters by Ultraviolet Radiation ......... 164
7.::1 Process Design of UV Wastewater Disinfection System ... 184
7..4~ UV Disinfection System Design Examp.le 216
7.Ei System Design and Operational and Maintenance
Considerations for the UV Process ...................... 223
7.Ei References ................................... 245


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 ofthe 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 c'urves 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 Headloss thru spray nozzle diffusers ........................ 66
List of Figures (Cont'd)

Number Page
5-16Nomograph for design of multiple perforated diffusers .......... 67
5-17 Details of a submerged weir mixing structure 69
5-18 DI~tails ofa hydraulic jump mixing structure ..................... 69
5-19 SGaie diagram of jump as designed 71
5-20 RE~sidence time distribution functions for contact basins .......... 72
5-21 Types of baffled contact chambers ............................ 73
5-22 V~med serpentine contactor design ......................... 73
5-23 Dissipation of chlorine residual and point of sampling for control ... 75
5-24 DC3finition sketch for rectangular contactor .................... 79
5-25 Tcm 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
inltermediates, 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 O,mne disinfection process gas and liquid flow diagram ......... 108

6-7 Diagrams showing feed-gas flow of typical ozone disinfection

processes .......................................... 108

6-8 O)<ygen requirement for ozone disinfection compared to oxygen

requirement for activated sludge ....................... 109

6-9 E>l:ample treatment schemes using ozone disinfection ........... 110

6-10 Fe!cal coliform survival for rotating biological contactor effluent,

screened effluent, anaerobic lagoon effluent, and strong
Wi3stewater ................................................. 111

6-11 Effect of water quality and performance criteria on ozone

dClsage requirement ....................................... 112

6-12 TCltal 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 discharj~e

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 I,evel 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 , ................... 1.17

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 controHed, water cooled

ozone generator ....................... , ...... , .... 122

6-24 Details of a vertical tube, voltage controlled, water cooled

ozone gt;lnerator .................................... 123

6-25 Details ofa vertical tube, frequency controlled, double cooled

ozone generator ........................................... , . 123

6-26 Details of an air-cooled, Lowther pla~e 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 ............................................... 1'31

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

Figures (continued)

Number' Page
6-36 Schematic of a turbine mixer ozone contactor ... 0 0 0 00000 000137
6-37 Example diagram of a thermal destruct unit with a heat
exchanger 0 0 0 0 0 0 0 00000 0 0138

6-38 Specific energy consumption versus off-gas temperature rise

through the thermal destruct unit ... 0 0 00 00 0 0000000 00 138

6-39 Example diagram of a thermal/catalyst ozone destruct unit. 000000139

6-40 Dose rElsponse curve for nitrified effluent at Marlborough 00 0 0 142

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. ... 000 0 0 0 0 0000 0143

6-43 Specific energy consumption for a typical air-fed ozone

generation system 0 '. 0 0 00 0 0 0 00 0 00 00 145

6-44 Design example projected ozone production rate for various

operating conditions .. 0 000 0 00.0 00'.0 0 0000 00 149

6-45 Human tolerance for ozone ... 0 0 0 0 00 0 0 0000 000152

7-1 Generell description of UV design. 00.... 0" 0 0.0.158

7-2 Example of closed vessel UV reactor with flow parallel to lamps. 0 160

7-3 Schematic of quartz UV unit in Vinton, IA 0 0.0.000. 161

7-4 Schematic of quartz UV unit in Suffern, NY . 0 161

7-5 Example of open channel unit at Pella, Iowa with flow directed
perpendicular to lamps ... 0 0 00 0 0 0 00 000 0162

7-6 Schematic of quartz UV unit in Albert Lea, MN 00000.0.00 0.0 162

7-7 Schematic of open channel, modular UV system .. 0 0.00 0.00.0 163

7-8 Example of UV system Teflon tubes 0.00. 0 0.0000 163

7-9 Electromagnetic spectrum ..... 0.0 0.0 0 0.0.0 174

7-10 Relative germicidal effectiveness as a function of wavelength .0 174

7-11 Relative abiotic effect of UV on Eo coli, compared to relative

absorption of ribose nucleic acid .. 0 0 00 0 000 0 00 000175

7-12 Example of DNA and UV damage to DNA. 0 0. 0.... 0.00.0 00 175

7-13 Schematic representation ofthe effects of photoreactivation. 000o. 176

7-14 Effect of particulates on UV disinfection efficiency ... 0 000 0 0 185

7-15 The rate K increases with increasing intensity for a given

residence time .................. 0 0 00 0 0 0 00 0 00 0 0186
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 arra\'s ... ~ ......... 204

7-27 Effect of Teflon system sizing on the power requirement

efficiency 205

7-28 Uniform lamp array intensity as a function ofthe reactor UV

density and.UV~absorbancecoefficient 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 densit\( for different

lamp array configurations ........................ 206

7-34 Example for deriving an estimate of the residuell 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 ~ f.~Jnction of the calculated elverage intensity ... 210
. '.
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 severa'l

wastewater treatment plants 214
7-39 Estimation of Np from several plants 2'15
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 techriiquefor estimating the total
UV output of a lamp 22p
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 ota 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

Number Page
2-1 Typical Composition of Domestic Wastewaters 6
2-2 Typical Influent Concentration Ranges for Pathogehic 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 M.odeand
Container .......................................... '.' ..... 16
3-7 Accident Rates per Metric Ton-Km ............., 17
5-1 Early Geographic Distribution of ChJorination 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 Dio~ide ........................ 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

5-9 Summary of Kinetics of HOCI-and OCI-Reduction by

Miscellaneous Reducing Agents 46
5-10 Physical Properties o'f Sulfur Dioxide 48

List of Tables (continued)

Number Page
5-11 Chick-Watson Parameters for Microbial Inactivation by Chlorine ... 53
5-12 Parameters in the Collins et a!. Model Describing Wastewater
Coliform Inactivation by Chlorine .................. q5
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 Trouble Shooting Guide 85
5-18 O&M Schedule, Sacramento Regional Wastewater Treatment
Plant. ..........................................., 88
6-1 U.S. Municipal Wastewater Treatment Plants Using Ozone ....... 98
6-2 Propc:lrties of Pure Ozone ...................... 99

6-3 Solubility of Ozone in Water 100

6-4 Terminology for Measured Ozone Parameters ........ 104
6-5 English Unit Equivalentsfor Ozone Concentration ........... 105
6-6 Terminology for Calculated Ozone Parameters .............. 105
6-7 Moisture Content of Air for Air Temperatures from
-80 0 to 40C ............................... 106
6-8 Atmclspheric 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
Variclus 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 Munf:cipalities 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 UtilizingUV
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 Delsign Phase 173

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 SevE~ral 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 Actinometry ...................... 232


Many individuals contributed to the preparation and rElview of this manual.

Contract administration was provided by the U.S. Environmental Protection
Agency, Water Engineering Research Laboratory (WERL), Cincinnati, Ohio.


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
O. Karl Scheible, 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

Sidney Ellner, Ultraviolet Purification Systems, Inc.,
Bedford Hills, New Yorl<
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.

Chapter 1

1.1 General nowto assess m,ore accurately the impact of present-

Chlorination has long been the accepted and pre- day disinfection practices.
ferred method of disinfection for both water and
wastewater in the United States. Major factors in the That a specific disinfection process has been reported
implementation of chlorination were efficiency and as troublesome at a particular facility could very well
low cost when compared to other means of disinfec- be the result of many factors, some or all of which
tion. New considerations that are now being exam- may not have been considered prior to engineering
ined very closely are the environmental and biota and installation. Critical factors such as disinfection
type, tank configuration, contact time, conditioning
impacts of chlorination versus attainment of public
health. The U.S. Environmental Protection Agency and pre-conditioning criteria, power costs, mainte-
recognized the adverse effects of chlorination of nance requirements, and others may not have been
fully understood or evaluated. Recent information
wastewaters and reported the following conclusions
releasEld by the EPA and the General Accounting
in its Task Force Report (1):
Office (GAO) has indicated that manyofthe municipal
disinfection requirements (Le., the need) should be wastewater treatment plants in the United States
evaluated on a case-by-case basis with considera- have not been meeting their effluent discharge
tion of beneficial uses and criteria. limitations (2). An estimated 18,000 municipal waste-
chlorine and subsequent residuals are extremely water treatment plants were in operation or under
toxic to aquatic wildlife. construction by the end of 1979. More than half of
chlorine and organic compounds form chloro- these plants were not functioning as designed with
organics and are potentially toxic to man. respect to 800, TSS, and fecal coliform removal.
In effect, the policy and summary conclusions have Reasons for plant problems were site-specific and
not changed since 1976. The agency policy essen- complicated with long-term violations determined to
tially states: be due to a combination of problems, as follows:

disinfection should not be required in those operation and main~enance (O&M) deficiencies.
instances where significant benefits are not dem- equipment problems.
onstrated. infiltration and inflow.
prospective disinfection benefits must be weighed industrial waste overloads.
against the environmental risks and costs. design deficiencies.
chlorine should be considered only when there are
public health hazards to control. With respect to design deficiencies, four basic types
alternative disinfection methods and/or dechlor- of design problems that have been implicated fnclude
ination must be considered when and where public the following:
and aquatic health and/or life impacts co-exist.
limited state-of-the-art duri!,g the design phase.
lack of expertise.
With the enactment of the Clean Water Act, its sampling errors prior to plant design.
amendments and new assertions on site specific time and funding constraints.
criteria, there is no longer a "generic" disinfection
procedure. The realization that chlorine is toxic to All these problems have been responsible for design
aquatic life and reacts with precursors of trihalo- deficiencies associated with wastewater disinfection
methanes (THMs) and other chlorine oxidizable and facilities. Limited state-of-the-art during the design
substitutable compounds in wastewater effluents phase has been a real problem associated with the
has caused considerable concern. As a result, the design of alternative disinfection processes, such as
existing practices of disinfection have fallen under ozone and ultraviolet light. With the increasing need
close scrutiny. Disinfection studies, chemical char- for wastewater treatment plant upgrading modifica-
acterizations, and other evaluations are ongoing even tions, including more stringent disinfection criteria,
elimination of design deficiencies and expanding the Information and results developed from the EPA
state-of-the-art for process design become very internal efforts and externally funded research
important considerations. programs.
Communications with investigators and equipment
The importance and need for a comprehensive design manufacturers.
document for municipal wastewater disinfection is Plant site visits and discussions with operating
apparent. A single source information document does personnel.
not exist today that will assist and direct the design
engineer, and other conc:erned parties, in an educated
The disinfection alternatives that are the subject of
and informed decision, selection, and design proce-
this manual and of which sufficient information
dure for the best disinfection alternative process
exists for design purposes are halogenation/dehalo-
suited to site specific constraints. A major tool that is
genation (including chlorine, bromine chloride, and
lacking in the area of disinfection is a concise and
chlorine dioxide), ozonation, and ultraviolet irradia-
unbiased design manual dealing with the disinfection
tion. In the future the acceptable risks associated with
alternatives, and the design of the most effective
the various disinfection alternatives and the levels of
alternative for the application desired. The purpose of
disinfection required may be refined and thus create
this manual is to bridge this gap and present the
disinfection criteria other than those in present use.
engineer with comprehensive design guidance for
The design approaches presented in this manual are
the implementation of the most efficient and cost
flexible and applicable to the various levels of
effective disinfection mothod needed for a particular
disinfection presently required as well as those that
site. may be required in the future by covering the range of
1.2 Purpose and Objectives no detectable coliforms up to any desired level in
wastewater effluents.
Over the past 14 years the EPA has actively pursued
an extensive program to investigate alternatives to 1.3 Scope
chlorination of wastewaters through internal and
The first part of this manual presents an overview of
externally funded research and development. This
the disinfection process, the types of disinfecting
research has produced a wealth of information on
agents, and the advantages and disadvantages of
disinfection alternativEls from literature surveys,
each. The manual then discusses in separate chapters
laboratory studies, pilot plant studies, and full-scale
how to design each of the primary disinfection
alternatives. Although many more alternative dis-
The EPA contractors, grantees, and project officers infection methods have been identified, only those
involved in this extensive effort have developed considered in this manual are cost effective and
results and experience in design and operation of the presently being implemented. The following criteria
most applicable disinfection alternatives. The EPA were used to select the disinfection processes
disinfection program has advanced to the point where discussed: technical feasibility, flexibility, reliability,
the wealth of information and expertise developed on complexity, safety, costs, environmental impacts, and
disinfection alternatives can be compiled into a single hazardous material impact or formation.
document for evaluation of process alternatives and
design of the selected disinfection process. The The primary thrust of this manual is to present
primary objective of this document is to provide a thorough design guidelines on ozone, ultraviolet
comprehensive process design manual for waste- light, chlorination, and chlorination/dechlorination
water disinfection to bEl used by design engineers, facilities. Bromine chloride and chlorine dioxide are
regulatory and review agencies, and owners and discussed in the manual but not to the level of detail
operators of disinfection processes. The development as the previous alternatives. Bromine chloride ap-
of this document has been based on both the pears to be as flexible and effective as chlorine;
theoretical and practical application of process design however, some questions still remain regarding the
criteria for the implementation of disinfection tech- use of bromine chloride, such as equipment reliability
nology in municipal treattmentfacilities. The informa- and future chemical cost. Full-scale operating data
tion in this process design manualfordisinfection has and a proven track record for bromine chloride
been developed from thl~ following sources: disinfection are still lacking. Chlorine dioxide is a
proven bactericide and virucide and has certain
Literature features making it attractive for drinking water
Extensive experience of the individuals preparing treatment. However, for wastewater disinfection,
the manual from participation in disinfection chlorine dioxide is not so attractive, principally
research and development. because of its high cost. Also, its persisting residual is
Active involvement and participation by the EPA toxic and may require removal through chemical
project officers involved in the disinfection pro- reduction prior to discharge of the disinfected ef-
gram. fluent.
Each of the respective chapters presents a brief Figure , ..,. Sequence of Manual's use.
historical background of the development and use of
the particular 'disinfectant of interest. The chapters
then review process chemistry and disinfection
Need for Disinfection
kinetics. Analytical measurement methodology is Technologies
discussed relative to both wet chemistry and instru- (Chapter 2)
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. I Disinfection Alternatives
and Options
(Chapter 3)
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 Kinetics and Hydraulic
municipal wastewater disinfection. A qualitative Considerations
screening procedure is presented in Chapter 3 for {Chapter 4)
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- Halogen
Ozone Ultraviolet
inant advantages and disadvantages of each ofthese (Dehalogenation)
Disinfection Disinfection
disinfection alternatives are also discussed. (Chapter 6) (Chapter 7)
(Chapter 5)
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 procedures are then presented, detailing the mechan-
Chapter 5, ozone disinfection is discussed in Chapter ics of how to size and specify equipment based on
6, and UV disinfection is discussed in Chapter 7. dose r~lquirements, disinfectant demand, NPDES
However, before the user proceeds to the appropriate permit limitations, andcost minimization. The latter
design chapter, he may want to review Chapter 4 on section of each chapter is the heart of the manual, but
disinfection kinetics and hydraulic considerations. the previous sections provide essential technical
Chapter 4 presents a general overview of kinetic back-up for making the proper design decisions.
considerations, mixing requirements, and contacting
requirements that apply to each of the disinfection 1 .5 R.~ferences
alternatives. The suggested sequence to follow for 1. Disinfection of Wastewater Task Force Report.
use ofthis disinfection design manual is presented in EPA-430/9-75-012, U.S. Environmental Pro-
Figure 1.1. tection Agency, Washington, DC, 1976.
After the appropriate disinfection technology has 2. Mc)rrison, A. GAO Finds Massive Failure of
been selected for design, the user can proceed Wi!lstewater Treatment Plants. Civil Engineer-
through the specific design chapter indicated in in~751(4):474, 1981.
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

Chapter 2
Need for Disinfection Technologies

2.1 Need for Disinfection disease in human beings. Waterborne transmission

of these disease-causing organisms can occur via
Population increases and much greater demands for four pathways(1): .
water supply and water recreational uses within the
past 15 to 20 years have significantly increased the direct ingestion of untreated water.
opportunity for human exposure to wastewaters direct ingestion of treated drinking water.
being discharged into the environment. Natural ingestion of aquatic food species infected with
safeguards, such as dilution and distance or time pathogens absorbed from contaminated waters.
before contact or use, have been reduced due to the invasion resulting from skin contact with contam-
large volumes of wastewater being discharged and inated watE!r.
the number of discharge locations. Domestic waste-
waters carry human pathogens excreted in the fecal The first three pathways are sometimes classified as
discharges of infected individuals. Even treated the fecal-oral route. The second pathway described
effluents can affect sources of domestic water supply, above occurs when a drinking water treatment
recreational waters, and shellfish growing areas. system fails or the integrity of the water distribution
Disinfection is necessary to reduce transmission of system is violated. The fourth pathway is likely to
infectious diseases when human contact is probable. result in skin, mucous membrane, or urinary tract
Chlorine in the past has been used almost universally infections but is seldom implicated in gastrointestinal
as the disi nfectant for wastewaters. However, studies illness in the United States. The risk of disease by
have shown that chlorine and its by-products can be exposure to wastewater effluent in recreational
toxic to aquatic life, repel and deny spawning grounds water, especially non-disinfected effluent, is not well
to anadromous fish, and decimate fish larvae and established on epidemiological grounds; however,
other forms of life. Out of these concerns, questions recent work has demonstrated a cause-effect rela-
have arisen relative to both disinfection needs in tionship via this pathway (2).
general and disinfection with chlorine. Although the
need for disinfection is site specific, in general, Wastewater treatment plants historically discharge
disinfection is considered to be a beneficial unit their effluents to natural receiving streams that are
process and required for most discharge applications. often tributaries of larger recreational bodies of water
or that are used as water supply sources by down-
The organisms of greatest concern in human expo- stream communities. City potable water supplies are
sure to wastewater-contaminated environments are often extracted from these tributaries or lakes,
the enteric bacteria and viruses and the intestinal physically and chemically treated and distributed to
parasites. Diseases that are spread via water con- customers. The only protection that the recreational
sumption and/or contact can be severe and some- users receive is the hope that the wastewater was
times crippling. Bacterial diseases such as salmo- adequateiy disinfected prior to discharge. So long as
nellosis (including typhoid and paratyphoid fevers), disinfection guidelines and standards are being met
cholera, gastroenteritis from enteropathogenic at the sources, public safety and water quality will be
Escherichia coli, shigellosis (bacillary dysentery) and protected. Factors that influence and potentially bias
viral diseases caused by infectious hepatitis virus, this type of rational thinking are equipment design,
poliovirus, coxsackieviruses A and B, echoviruses, operator training, equipment dependability, operator
reoviruses, and adenoviruses may be contracted by attention, and others.
contact with or by consumption of wastewater
contaminated water supplies (Le., potable and/or It is significant that infectious hepatitis has main-
recreational). tained a level of 50,000 to 60,000 cases per year in
the United States (3), while typhoid fever dropped
The alternative of discharging wastewater that has from 2000 cases in 1955 to 300 in 1968 (4).There are
not been disinfected allows discharging of pathogenic more than 100 viruses excreted in human feces that
organisms and other resultant hazards posed to have been reportedlyfound in contaminated water, of
humans. Pathogenic organisms, by definition, cause which anyone could cause a waterborne disease.

A wide variety of enteric pathogens including viruses, characteristics for which he is designing a disinfec-
bacteria, and parasites is known to occur in all tion system. He must have a knowledge of the
community-derived wastewaters. Infective dose composition of the conventional parameters as well
studies with a variety of enteric organisms have been as the concentrations of the pathogenic agents or
, conducted over the pa~;t 30 years in hurna n vol unteers indicator organisms for which he is designing the
(5). The widest dose range required to produce a disinfection system. The influent pathogen or indi-
response was found with the bacterial agents. cator organism concentration is a critical parameter
Salmonella spp. r,equired the largest 'dose with the for design of any of the disinfection technologies.
ingestion of 105 to 10 8 cells needed to produce a 50 Typical compositions of raw wastewater through
percent illness rate. In contrast, three species of various levels of treatment are summarized in Table
Shigella produced illness in a significant percent of 2-1. The total or fecal coliform (indicator organism)
subjects dosed with 10 to 100 cells. Protozoan level is a critical design parameter that should be
infections have been produced with Entamoeba coli determined on a site specific basis where possible.
and Giardia lamblia dosed in gelatin capsules at the
level of 1 to 10 cysts. Enteric viruses have produced The total bacterial population of I)uman feces has
infection at low dosage levels via oral ingestion, been estimated to reach a density of 101~ organisms
inhalation, and conjunctival exposure. These studies per gram (3). The density range of fecal coliforms in
have clearly shown that specific enteric organisms of human feces has been estimated at 10 6 to 109 per
all three classifications, Le., bacteria, vir.uses, and gram with total coliforms estimated at 10 7 to 10 9 per
animal parasites, can produce infections at relatively gram. Hubley et al. (1) presented an assessment of
low exposure levels. Available data ara insufficient to the concentration ranges of certain organisms in
evaluate the actual health hazards that exist for domestic wastewater, reduction through primary and
individuals exposed to wastewater subjected to secondary treatment, and estimated secondary
various degrees of trelatment and dilution; however, treated effluent concentrations, as summarized in
the data indicate that enteric pathogens can cause, Tables 2-2, 2-3, and 2-4. If one assumes that the
infections at exposum levels found in undisinfected pathogenic organisms ar,e removed in proportion to
Table 2-2. Typical Influimt Concentration Ranges for
Sobsey reported the viral content of wastewater Pathogenic and Indicator Organisms (7) ,
(United States origin) ranged from 2 to 3 t9 more than Number/100 ml
1000 infectious units/100 ml sample (6). Peak Organism Minirnum Maximum
periods were observed to occur in late summer and Total Coliforms 1,000,000
early fall. Most enteric virus concentrations have Fecal Coliforms 340,000 49,000,000
been isolated in heavily polluted surface waters, but Fecal Streptococci 64,000 4,500,000
Berg and coworkers were able to detect enteric Virus ' 0.5 10,000
viruses in the Missouri River having fecal coliform
concentrations as low as 60/100 ml (3). Although Table 23. Microorganism !Reductions by Conventional
very little quantitativEt information is available "'lith Treatment Processes (S)(9)
respect to the concen1:ration levels of enteric viruses Primary Secondary
in the United States surface and groundwaters, much Treatment Treatment
evidence leads to the indication that wastewaters are Removal Removal
a primary source. Microorganisms (%) (%)
Total coliforms <10 90-99
It is difficult to accurately identify general municipal Fecal coliforms 35 90-99
Shigella sp. 15 91-99
wastewater characteristics due to differences in Salmonella sp. 15 96-99
locations, water use~;, seasonal variations, diurnal Escherichia coli 15 90-99
variations, etc. HowEwer, it is important that the Virus <10 76-99
design engineer have a knowledge of the wastewater Entamoeba histolytica 10-50 10

Table 21. Typical Composition of Domestic Wastewater

Total Fecal
BOD TSS Total N CoJiforms CoJiforms
Water Quality mg/I mg/I mg/I #/100 ml #/100 ml
Raw Wastewater 200 200 35 '107-108 106-107
Primary Effluent 130 100 30 107-108 106-107
Secondary 20 20 20 105-10 6 104-105
Filtered Secondary 12 5 18 '104-10 5 103-10 5
Nitrified 7 10 18 '104-10 5 103-105
Filtered Nitrified 5 5 18 '104-10 5 103-105

Table 2-4. Secondary Effluent Ranges for Pathogenic and by the United States Public Health Service to protect
Indicator Organisms Prior to Disinfection the quality of coastal shellfish beds. The 1964
, Number/100 ml National Shellfish Sanitation Workshop defined the
Organism Minimum Maximum bacteriological water quality for approved shellfish
Total Coliforms 45,000 2,020,000 areas (those areas where shellfish may be taken for
Fecal Coliforms 11,000 1,590,000 direct marketing) as having a total coliform mean
F~cal Streptococci* 2,000 146,000
Viruses 0.05 1,000 Most Probable Number (MPN) not exceeding 70 per
Salmonella sp. 12 570 100 ml. In1977 a fecal coliform median concentration
*Assuming removal efficiencies for fecal streptococci similar to of 14 MPN per 100 ml with no more than 10 percent
the fecal coliform removal efficiencies of the samples exceeding 43 MPN per 100 ml was
recommended for approved shellfish waters by the
the indicator orga nisms (totaI andlor feca I col iforms), EPA. These fecal coliform levels were based on an
conventional treatment of domestic wastewaters extensive total coliform to fecal coliform correlation
without disinfection cannot be considered sufficient study and were subsequently accepted for waste-
for removal and control of human pathogens where water effluent discharges into shellfish waters.
benefiCial uses and body contact occur.
The two effluent guidelines or standards most
2.2 Disinfection Criteria commonly used by the various states and territories
In 1972 Congress enacted the Federal Water Pollu- in the United States have been a total coliform value
tion Control Act (PL 92-500)(amended in 1977, Clean of 1000 per 100 ml and a fecal coliform limit of 200
Water Act, PL 9,5-217), to "restore and maintain the per 100 ml. The microbial guideline for primary
chemical, physical and biological integrity of the contact recreational waters as adopted by many
nation's waters." Effluent standards were to be states requires that the fecal coliform content shall
imposed on industrial and municipal wastewater not exceed a geometric mean of 200 per 100 ml and
dischargers based on the limits of current technology. that no more than 10 percent of the total number of
Today, secondary treatment is the minimum level samples taken during any 30 day period shall exceed
required for municipal wastewater treatment plants. 400 fecal coliforms per 100 ml.
The Secondary Treatment Information Regulation
was promulgated by the EPA in 1973 as Part 133, Some states have adopted even more stringent
Title 40 of the Code of Federal Regulations (40 CFR disinfection requirements than 14 MPNper 100 ml
133). Included in the definition of secondary treat- for mean effluent fecal coliform levels. Since No-
ment was a fecal coliform limitation,' but this was vember 1976, the Maryland Department of Health
omitted from the CFR inJuly 1976, thereby delegating and Mental Hygiene has required all facilities plans to
to the individual states, the responsibility to establish provide for disinfection to total coliform levels of 3
site specific water quality criteria. MPN per'100 ml unless a special exemption is
granted by the department (10). This high level
Municipal wastewater treatment plants often dis- disinfection is applicable to both coastal and inland
charge into potable water sources and recreational waters but is geared toward protection of the state's
bodies of water. The disinfection practices for potable intensive shellfish industry. The California State
water have helped control waterborne disease out- Department of Health "Uniform Guidelines for
break from such sources; however, wastewater Sewage Disinfection" incorporates consideration of
disinfection through the destruction of pathogenic dilution, receiving water quality, and beneficial uses
agents, provides a barrier to possible waterborne in disinfection requirements, which results in varying
disease before the wastewater is released into the coliform standards for different discharge situations
environment. Public health and protection of the (11). The standard for nonrestricted recreational uses
aquatic and human environments should be the of wastewater and shallow ocean discharges in close
overriding considerations affecting disinfection re- proximity to shellfish areas specifies a 7 -day median
quirements. With the site specific constraints and total coliform value 'of 2.2 per 100 ml or less at some
considerations, such as water quality, seasonal point in the treatment process. The 2.2 total coliform
versus year round disinfection requirements, etc., per 100 ml standard applies to those discharge
each individual state has established its own waste- situations where reclaimed wastewater (Le., 100
water disinfection policies. A review of the types of ' percent wastewater effluent) is impounded for body
disinfection criteria of various, states follows. contact recreation activities, where wastewater is
discharged to ephemeral streams with little or no
Asa result of the 1924-1925 typhoid epidemic dilution, and where body contact recreation has been
resulting from consumption of contaminated shell- designated as a peneficial use of the stream.
fish, and because of several other less severe
incidents related to shellfish contamination, a na- The various states hav.e fecal coliform standards
tional shellfish sanitation program was established ranging from less than 2.2/100 ml up to 5,000/1 00
ml and total coliform standards from less than unit process in the wastewater treatment system.
2.2/100 ml up to 10,000/100 ml. Some states have Disease transmission via the aquatic route including
seasonal disinfection rElquirements, and Illinois, at recreational water, drinkin!~ water; and seafood from
the time of writing this manual, has no bacterial polluted water has been and continues to be a
standards for certain, designated non-contact waters. problem. To the extent that the wastewater disinfec-
These disinfection critl3ria have been established tion process mitigates that problem asserts for the
relative to discharge stl'eam water quality with the continued use of that practice in this country.
most com man sta ndard~l bei ng fecal coliform limits of
200/100 ml to 1,000/1i 00 ml. Over 45 states have 2.3 Projected Applications of Disinfection
multi-level standards fo,r disinfection relative to the Technologies
discharge stream water quality criteria, with 200
A summary of the 1984 EPA wastewater treatment
fecal coliforms per 100 ml as the most common plant needs survey has been prepared (12). A
standard. At least 15 sUites have both fecal and total breakdown of the number of municipal wastewater
coliform disinfection criteria. At least 15 states also treatment plants by flow capacity in 1984 and those
require effluent disinfel::tion levels of 14 fecal coli-
projected for the year 2000 is presented in Table 2-5.
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- Table 2-5. Number of Wastewater Treatment Plants by Flow
charges. Capacity (12)
Percent of Total
As indicated by Cabelli (2), the development of water Number of Wastewater
Flow Capacity Facilities Flow
quality guidelines, standards, and associated health
m 3/d (mgd) 1984 2000 1984 2000
effects for recreational waters, such as disinfection
380 5,032 8,416 0.7 0.9
requirements, has historically followed a pattern (O - 0.1)
characteristic of many such efforts to control pollu-
tion; the establishment of health and ecological 380 - 3,800 6,962 8,313 7.4 6.9
effects. The first step has been the development of (0.1 - 1.0)
guidelines and standards dictated largely by attain- 3,800 - 38,000 2,833 3,255 25.7 24.6
ment with the Best CI:mtrol Technology Available (1.0 - 10.0)
(BAT). The guidelines have been based upon limited
epidemiological and ec~ological evidence in many >38,000 551 687 66.2 67.6
cases, and little, if any. data quantifying the risk in (>10.0)
relation to the level of the pollutant in the environ- Total 15,378 20,671
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 The total number of facilities broken down by type of
is then the development of guidelines based upon treatment system in 1984 and projected for the year
acceptable risk, which rl3quires an epidemiological or 2000 is shown in Table 2-6. An evaluation of the
ecological data base broad enough to mathematically number of facilities and quantity of flow treated by
model the relationship of some measure of water municipal wastewater treatment facilities with pri-
quality to the risk. As more studies are conducted and mary treatment through advanced wastewater treat-
more data become available, the effluent disinfection ment shows that approximately 41 percent of the
guidelines may change. This design manual has been total wastewater flow presently receives secondary
prepared with the realij~ation that different disinfec- treatment, with similar requirements in the year
tion criteria are presently required and still others 2000. Approximately 39 percent of the total flow
may be required in the future. Therefore, the design of presently treated receives greater than secondary
the various disinfection alternatives has been
presented with this flex:ibility in mind, as well as the Summary of Wastewater Treatment Processes
Table 2-6.
considerations of different water qualities. For ex- in the United Sta1tes (12)
ample, this manual can !be used to design disinfection Number of Processes
facilities for primary treatment through advanced Type of Process 1984 2000
wastewater treatment. It can be used for effluents Lagoons 7,500 12,210
from lagoons, activat,ed sludge, trickling filters, Activated Sludge 5,690 8,275
oxidation ditches, and rotating biological contactors, Trickling Filter 2,463 2,570
Land Treatment 926 1,454
as well as other unit processes. Oxidation Ditch 741 1,215
Rotating Biological Contaetor 347 291
Attainment of the disin1l ection guidelines can only be
achieved by the disinfection process, which, from a Total Design Flow, m 3/d (mgd) 1.3 X 108 1.6 X 108
disease prevention standpoint, i,s the most important (35.9 X 109 ) (43.2 X 109 )

treatment, with around 51 percent expected to require 2.4 Re'ferences
this level of treatment in the year 2000.
When an NTIS number is cited in a reference, that
In 1984 at least 80 percent of the total wastewater reference is available from:
flow treated required secondary or more stringent
National Technical Information Service
levels of treatment. This value is expected to increase
5285 Port Royal Road
to over 90 percent of the total flow treated in the year
Springfield, VA 22161
2000. Most of this increase will be in requirements
(703) 487-4650
for advanced secondary treatment with the number of
facilities increasing from around 3,000 in 1982 to 1. Hubley, D., et aI. Risk Assessment of Waste-
around 6,400 facilities in the year 2000. The per- water Disinfection. EPA-600/2-85/037, NTIS
centage of flow treated that requires secondary No. PB85-188845, U.S. Environmental Protec-
treatment will remain about the same; however, an tion Agency, Cincinnati, OH, 1985.
estimated increase of around 5,000 new facilities has
been projected. These more stringent requirements 2. Cabelli, V.J. Health Effects Quality Criteria for
for treated effluent quality are related to concerns for Marine Recreational Waters. EPA-600/1-80-
protection of public health and the aquatic environ- 031, NTIS No. PB83-259994, U.S. Environmen-
ment. These concerns will most certainly place tal Protection Agency, Cincinnati, OH, 1980.-
greater emphasis on disinfection requirements and
alternatives to chlorine disinfection. Advanced 3. Berg, G.lndicatorsofVirusesin WaterandFood.
wastewater treatment facilities are excellent candi- Ann Arbor Science, Ann Arbor, Michigan, 1978.
dates for alternative disinfection with ozone and
ultraviolet irradiation. Due to the health hazards and 4. White, G.C. Disinfection: The Last Time of
environmental concerns with chlorination, there Defense for Potable Water, JA WWA 67(8):410,
appears to be an excellent market potential for 1975.
alternative disinfectants, especially ozone and ultra-
violet irradiation, over the next 15 years. 5. Akin, E.W. Infective Dose of Waterborne Path-
ogens. In: Municipal Wastewater Disinfection,
Many of the existing treatment facilities are required Proceedings of Second National Symposium.
to disinfect their effluents, although the degree of EPA-600/9-83-009, NTIS No. PB83-263848,
bacterial reduction varies from plant to plant. How- Cincinnati, OH, 1983.
ever, there are many facilities that presently do not
disinfect their effluents prior to discharge. Due to the 6. Sobsey, M.D. Enteric Viruses and Drinking
present concerns and future requirements for more Water Supplies. JAWWA 67(8):414, 1975.
stringent effluent qualities, more stringent disinfec-
tion criteria relative to both microorganism control, 7. Health Risks Associated with Wastewater Treat-
disinfectant residual, and biotoxicity or bioassay ment and Disposal Systems, State of the Art
testing requirements can be expected. Many of the Review. EPA-600/1 -79-016a, NTIS No. PB-
existing facilities may be required to upgrade their 300852, U.S. Environmental Protection Agency,
present disinfection systems to include dechlorina- Cincinnati, OH, 1979.
tion or install alternative disinfection processes. New 8. Okun, D. and G. Ponghis. Community Waste-
facilities being built can be expected to require proper water Collection and Disposal. World Health
disinfectant residual controls or alternative disinfec- Organization, Geneva, 1975.
tion processes.
Over 400 facilities have been designed for chlorina- 9. Craun, G.F., et al. Waterborne Disease Out-
tion/dechlorination, and approximately 50 percent of breaks in the United States. JAWWA 68:420,
these facilities are presently in use. At the time of 1976.
writing this manual, there were 24 wastewater 10. Maryland Water Quality Standards Policy, Oc-
treatment plants i'n the United States confirmed to I;>e tober 1978.
using ozone disinfection, 20 water treatment plants
using ozone either for disinfection or chemical oxida-~ 11. Crook, J. Wastewater Disinfection in California.
tion, and 30 drinking water plants in Canada using In: WastewaterDisinfection-The Pros and Cons.
ozone (R.G. Rice, Personal Communication, 1985). Proceedings WPCF Pre-conference Workshop,
An estimated 125 wastewater treatment facilities New Orleans, LA, Water Pollution Control
were reported to be using ultraviolet light disinfection Federation, 1984.
in 1984. As of March, 1985, 33 municipalities with
ultraviolet light disinfection had been granted funds 12. 1984 Needs Survey: Report to Congress. EPA
by the EPA under the Innovative and Alternative (II A) 430/9-84-011, U. S. Environmental Protection
Technology Program. Agency, Washington, DC, 1985:

Chapter 3
Disinfection Alternatives and Options

3.1 General Considerations tion/dechlorination with sulfur dioxide, chlorine

Since municipal effluents are an identifiable source dioxide, bromine chloride, ozone, and ultraviolet light.
of pathogenic contamination and disinfection pro-
cesses themselves can create hazards to human To properly evaluate and select alternative disinfec-
health and the aquatic environment, the decision to tion systems two levels of review are required. In the
disinfect or not disinfect is a very complicated one first level of review, a number of non-monetary
that must be made on a site-specific basis. It is factors are considered. This qualitative assessment is
therefore, impossible to establish universal policies comprised of three primary components, including
on wastewater disinfection requirements. Resolution the previously described technical factors, environ-
ofthe need for municipal wastewater disinfection at a mental impacts, and safety. In order to assess the
particular site involves the investigation of receiving disinfection alternatives with respect to their non-
water uses and the associated risks to human health, monetary factors, a qualitative matrix approach, as
an assessment of the options that are available for , shown 011 Table 3-2, can be used. A relative ranking of
control of fecally-contaminated discharges, and an the alternatives based on this qualitative assessment
evaluation of the environmental effects that control can also be made, as shown in Table 3-3. The ranking
measures may create. Figure 3-1 presents an ap- scale is based on a scale of one to five, with one
proach for the type of rationalization that can be indicating the least impact or best degree of con-
involved in assessing the need for, and consequences fidence. From these types of analyses, the number of
of, disinfecting municipal wastewaters (1). appropriate alternatives can be narrowed, and some
alternatives may be completely eliminated.
In general, Figure 3-1 indicates that human health The remaining acceptable alternatives can then be
risks are the initial concern, and upon establishing evaluated in a second more detailed level of review. In
the level of risk involved and the potential for reducing this second level of review, a preliminary design can.
or eliminating the risk, the environmental considera- be developed, cost estimates performed, and an'
tions determine the applicability of the proposed economic analysis comparing the alternatives 011 an
control measures. Development of an option, chlori- equitable basis can then be evaluated. Detailed
nation or an alternative disinfectant, that satisfies capital and operation and maintenance costs can be
both the human health and environmental concerns developed for each alternative disinfection system.
at a specific site is the next step. Capital costs include structures, process equipment,
major auxiliary equipment, special foundation re-
3.2 Selecting a Disinfection Alternative quirements, electrical and instrumentation, site
As many as 25 disinfection alternatives could be work, miscellaneous process and piping, construc-
considered and have been previously identified from tion contingencies, engineering, project administra-
the literature without regard for physical or opera- tion, and interest during the estimated period of
tional constraints (2). The major factors that must be construct.ion. The operation and maintenance costs
considered when evaluating disinfection alternatives are annual costs and include labor, electrical power,
are summarized in Table 3-1 (3). The first four factors, chemicals, routine equipment maintenance, and
effectiveness, use cost, practicality, and pilot study materials and supplies. The specific details required
requirements relate to the disinfection process itself. for performing the second level of review are ad-
The fifth factor, potential adverse effects, relates to dressed in the respective design chapters for each of
effects ofthe disinfectant on the receiving water and the disinfection alternatives.
other environmental concerns and considerations.
Evaluation and thorough consideration of all the The predominant advantages and disadvantages of
criteria listed in Table 3-1 relative to practical, these disinfection alternatives are well known and
physical, and operational constraints of municipal commonly cited in the literature. Some of the more
wastewater disinfection reduces the available alter- obscure elements have not been emphasized or have
natives to chlorination, hypochlorination, chlorina- 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)
. ._ ....N_o
r 4
.!__....;v,.;,e..s_ _ ...
1 .
1'5 receiver used for primary Does the discharge impact on
c:ontac recreation. Shellfish the quality of intake water?
propagation/Harvesting. (Dilution. Distance. Background
Agriculture. Industry? quality. Other sources)

I... ,
~I_O _v_es--.
..... No Ves

Does the discharge impact on Evaluate the feasibility of

the quality of water at the disinfecting the discharge with
point of use (potential use)? chlorine/chlorine compounds

.! Is there any other reason

(Dilution. Distance. Background.

Ves ..
-Target levels
-Effluent quality
. to disinfect? 4 1 -Water treatment

Evaluate the feasibility of I
distinfecting the discharge Does the use of chlorine for

r------ Yes on a seasonal basis

-Water use
wastewater disinfection pose
a hazard to human health?

-Target levels
No -Effluent quality
Is there a potential for significant
chlorine induced toxicity to aquatic
life? (TRC limits. Chlorinated
No Eliminate Chlorine

Evaluate alternate disinfection
technologies: Dechlorination
techniques. Alternate methods of
disposal (Tables 3-2 and 3-3)
. I
Select method

,,-_-.-L J
of protection
for regulatory
hearings etc.
monitoring and
Annual reviews
and cost

review of the pertinent factors associated with each Major factors include size, flexibility, local construc-
of the alternatives is presented in the following tion costs, energy costs, energy efficiency and power
sections of this chapter. consumption, ozone dose requirements, and site'
specific constraints. The cost for producing ozone is
From Tables 3-2 and 3-3, it may be possible to select
normally higher than the alternative disinfection
an alternative disinfElction system, as demonstrated
methods; however, the other advantages associated
in the following situations. Examples follow where
with ozone disinfection (Tables 3-2 and 3~3), may
ozone, ultraviolet light, and chlorination/dechlorina-
outweigh cost considerations in some cases. The
tion alternatives were selected for different specific
following example is cited where ozone would be
selected as the disinfectant of choice based on site
The cost of installing and operating an ozone dis- specific constraints; however, this situation may be
infection process is dependent on many variables. unique since ozone would be cost competitive.
Table 3-1. Major Factors in Evaluating Disinfectant jng chlorine through the recreational community
Alternatives were also resolved with the selection of an ozone
Effectiveness Ability to achieve target levels of disinfection process..
selected indicator organisms
- Broad spectrum disinfecting ability In a second community, ultraviolet disinfection was
- Reliability
chosen over chlorination/dechlorination and ozona-
Use-Cost Capital cost tion when the following factors were given high
- Amortization cost priority:
Operating and maintenance cost
- Cost of special wastewater pre- effectivElness in destroying bacterial and viral
treatment .pathogens.
Practicality - Ease of transport and storage, or
safety precautions for transport, storage, and
ease of on-site generation application.
- Ease of application and control environmental impacts, especially as relates to
- Flexibility halogenated organics formation and chlorine
'Complexity residual toxicity.
- Ability to predict results
- Safety considera~ions equipmenneliability and ease of maintenance.
process reliability and simplicity.
Pilot Studies Required - Dose requirements process control flexibility.
- Refine design details overall construction, operation, and maintenance
Potential Adverse Effects - Toxicity to aquatic life
- Formation and transmission of un-
dersirable bio-accumulating sub- The treatment process consists of activated sludge,
stances final clarification, chemical addition for phosphorus
- Formation and transmission of removal, and filtration pr:ior to disinfection. The
toxic, mutagenic, or carcinogenic following design conditions applied to this specific
Average Flow = 7,600 m /d (2 mgd)
A large tourist community has selected the oxygen Peak Daily Flow = 13,200 m /d (3.5 mgd)
activated sludge process as the secondary treatment Effluent BOD = 10 mg/I
process of choice. The treatment plant is to be located Effluent TSS = 10 mg/I
. next to the municipal utility district power plant with Effluent NH 3 -N :: 1.0 mg/I
electrical costs of $0.025/kWh. The following design Effluent P0 4 -P = 1.0 mg/I
conditions were established due to the stringent 30-Day Geometric Mean Fecal Coliform = 200/1 00
effluent discharge criteria: ml
3 Maximum Daily Fecal Coliforms = 400/100 ml
Average Flow= 378,000 m /d (100 mgd) Maximum Chlorine Residual = 0.1 mg/I
Peak Daily Flow = 568,000 m /d (150 mgd)
Effluent BOD = 5 mg/I Ultraviolet light was considered to be the most
Effluent TSS = 5 mg/I desirable of the three disinfection.alternatives con-
Effluent NH 3 -N = 1 mg/I sidered with respect to non-monetary factors. Ultra-
30-Day Geometric Mean Fecal Coliforms = violet light systems were considered to be simple to
200/100 ml operate and maintain relative to ozone systems, yet
Maximum Daily Fecal Coliforms = 400/100 ml provide the same advantages for overall disinfection
Maximum Chlorine Residual = 0.02 mg/I capacity relative to bacteria, spores, and virus inacti-
Dissolved Oxygen = 6.0 mg/I vation. Ultraviolet light was the safest alternative
since it is a physical disinfection process, and its use
Due to the recreational nature ofthe community and eliminated the handling, transportation, and storing
the use of the downstream water, major considera- of to><:ic, hazardous, and corrosive chemicals. Due to
tions included effective virus destruction, Giardia cyst the size of the plant and assu.rance of good effluent
destruction, and color removal. Maintaining the quality especially with respect to the suspended
pristine image of the service area and protection of solids levels, ultraviolet light was chosen as the
the quality of the downstream water supplies were disinfectant .of choice over both ozone and chlorina-
the primary reasons for the choice of ozone disinfec- tion/dechlorination. Ultraviolet light disinfection was
tion. The total cost for ozone disinfection represented cheaper than ozonation and cost competitive with
a small part ofthe total plant costs, and in this unique chlorination/dechlorination in this particular applica-
situation, at the design ozone transferred dose of 5.0 tion.
mgl', ozone disinfection was cost competitive with
chlorination/dechlorination and ultraviolet disinfec- The third specific case cited here involved a com- .
tion.The safety concerns for handling and transport- munity in the northern part of the United States
Tabla 32. Applicability of Alternative Disinfection Techniques
Consideration CI 2 CI/deCI 2 BrCI UV
Size of plant all sizes all sizes all sizes small to medium to small to
medium large medium

Applicable level of all levels all levels secondary secondary secondary secondary
treatment prior to
disinfection ...

Equipment good fair to ? ? fair to fair to

Reliability good good good

Process Control well fairly well problematic no developing developing

developed developed experience

Relative Complexity of simple to moderate moderate moderate complex simple to

Technology moderate moderate

Safety Concerns yes yes yes yes no no

Transportation substantial substantial substantial substantial moderate minimal
on site

Bactericidal good good good good good good

Virucidal poor poor fair to good good good good

Fish Toxicity toxic non-toxic slight to toxic none non-toxic

moderate expected

Hazardous By-products yes yes yes yes none no


Persistent Residual long none short moderate none none

Contact Time long long moderate moderate moderate short

to long

Contributes Dissolved no no no no yes no


Reacts with Ammonia yes yes yes no yes (high no

pH only)

Color Removal moderate moderate ? yes yes no

Increased Dissolved yes yes yes yes no no


pH Dependent yes yes yes no slight no

(high pH)

O&M Sensitive minimal moderate moderate ? high moderate

Corrosive yes yes yes yes yes no

Tabla 3-3. Technical Factors and Feasibility Consider- and associated fish toxicity; however, the discharge
atfons stream was not used as a water supply downstream.
Considerations CI 2 CI/deCI 2 BrCI CI0 2 03 UV The following design conditions apply to this plant:
Flexibility 2 2 2 2 2 2
Reliability 1 2 3 2 3 2
Complexity 2 2 2 3 3 2 3
Average Flow = 3,800 m /d (10 mgd)
Effectiveness 2 2 1 1 1 2 3
Pilot Studies 1 1 4 4 3 3 Peak Daily Flow = 5,700 m /d (15 rngcl)
Rating based on scale of 'I to 5, with 1 indicating best degree of Effluent BOD = 30 mg/l
confidence. Effluent TSS = 30 mgll
30-DayGeometric Mean Fecal Coliforms= 200/1 00
where seasonal disinfection was required during the ml .
early spring and summer months. In this particular Maximum Daily Fecal Coliforms = 400/100 ml
application there was a concern for chlorine residuals =
Chlorine Residual 0.1 mg/l
Due to the seasonal disinfection requirements, capital Today, chlorination is the most used disinfectant at
costs, and operation and maintenance costs, chlori- water and wastewater treatment plants in the United
nation/dechlorination was chosen as the design States. Chlorine reacts very rapidly when mixed with
alternative. The non-monetary advantages typically water, and both hydrolysis and ionization occur.
associated with ozone and ultraviolet light over Environmental factors such as temperature, pH,
chlorination/dechlorination did not outweigh the alkalinity, suspended solids, chemical oxygen demand
advantage of chlorine disinfection in this particular (COD), and nitrogen containing compounds influence
application. the effectiveness of chlorine disinfection. Chlorine
reacts rapidly with ammonia and certain organic
As indicated in the previous examples, site specific compounds to form chloramines and chlorinated
constraints and criteria will set guidelines for selec- organic compounds. The combined chloramines are
tion of appropriate disinfection alternatives. Important lower in germicidal value compared to the free
considerations include factors such as the size of the chlorine residuals, with organic chloramines in
plant and effluent discharge quality requirements. wastewater offering significantly lower germicidal
The more stringent the effluent requirements, the value than inorganic chloramines.
more feasible alternative disinfectants such as ozone
and ultraviolet light become. Capital and operation The use of chlorine disinfection of wastewater can
costs are obviously important factors; however, the result in several adverse environmental impacts,
cost of any of the disinfection options is a very small especialfyduetotoxic levels of total residual chlorine
part of the total system costs. in the receiving water and formation of potentially
,toxic halogenated organic compounds. Chlorine
residuals have been found to be acutely toxic to some
3.3 Chlorination species of fish at very low levels. The chlorine
Chlorine has not been required by EPA for inclusion in residuals are stable and can persist for many hours at
States' nonconventional pollutant standards, but toxic levels. Other toxic or carcinogenic chlorinated
approximately 15 states have taken steps to develop compounds can bioaccumulate in aquatic life and
specific criteria for determining the impact and contaminate public drinking water supplies.
adverse effects of chlorine on aquatic life. Most states
have not adopted site-specific criteria to determine Chlorine is normally handled in steel containers of 68
the need(s) and adverse effects of disinfection using kg (150 Ib) cylinders up to 82 metric ton (90 ton)
chlorine. As a result, there has been limited success railroad cars. Chlorine is an extremely volatile and
in ensuring that public health and aquatic wildlife are hazardolls chemical, and proper safety precautions
adequately protected. must be exercised during all phases of chlorine
shipment, storage, and use.
Current developments consist of revised water quality
standards regulation and guidance documents to Hubley et al. conducted an extensive analysis of the
ascertain the appropriateness of existing water transportation risks of chlorine (4). A summary of
quality standards and assess use attainability. EPA's some, of their findings is presented. A compilation of
Office of Research and Development has completed a the Department of Transportation (DOT) accident
draft criteria ,document for chlorine with directions on reports for chlorine transportation from 1971 to 1980
how to apply the criteria and the chemistry and fate of is summarized in Table 3-4. These reports have
chlorine in natural receiving bodies of water. It is included information on number of deaths or injuries
anticipated that guidance for establishing chlorine and amount of chlorine released. The data in Table
effluent limitations for NPDES permits will be pub- 3-4 are summarized by railroad, truck, and barge
lished as well as other documents to help encourage shipment. The major accident atYoungstown, Ohio in
states to consider public and wildlife health issues. February 1978, has been broken out separately.
The Advanced Technology (AT) review policy requires Chlorine movements for 1972, obtained from the U.s,
the evaluation of chlorine toxicity and the construc- Bureau of Census, in metnc ton-kilometers by
tion and operation of dechlorination facilities where transportation mode and by shipm~nt weight are
chlorine will exceed EPA criteria. presented in Table 3-5 (5).
, However, due to the toxicity of chlorine residuals at A breakdown of the information in Table 3-5 for
extremely low concentrations (11 to 19 pg/I) and the chlorine shipments by transportation mode and
relatively high limit of detection of chlorine residual container is presented in Table'3-6 (4). The informa-
test procedures (50 to 100 pg/I), it is difficult to tion in Table 3-4 was averaged to obtain estimates of
control chlorine-induced toxicity in the receiving deaths, injuries, and property damage per year.
stream. Therefore, use of alternative disinfection
processes should be considered where aquatic The percentage breakdown by transportation mode
toxicity is the overriding concern. and container in Table 3-6was then used to calculate
Table 3-4. Compilation of [)epartment of Transportation Accident Data (January 1971 to December 1980) (4)
Number of Property Amount
Accident Damage Released
Reports Deaths Injuries ($) (kg)
Railroad 72 8 247 1.'1 x 106 1.4 X 105
Excl. 71 87 22,500 99,300

to 114 kg 14 o 60 8,000 574
0.911 metric
ton containers 4 o 15 23,500 245
trucks 2 o 71 15,000 23
Barge 2 o 3 o *
Information not available prk)r to 1976.

Table 3-5. Percent Distribution of Chlorine Shipped by cost is often a great advantage for chlorine. Even
Transportation Mode and by Shipment Weight when dechlorination is required, it is still normally the
(metric ton-kilometers) (4) lowest cost alternative in most cases.
A. Transportation Mode Hypochlorination refers tothe use of solid (calcium) or
Rail 84.9 liquid (sodium) hypochlorite compounds as the dis-
Truck (combines DOT motor carrier
and private truck di3ta) 15.0 infecting agent. The active co'mpounds are the same
Water 0.3 as with chlorination, primarily monochloramine in
lOQ.2 wastewater effluents. The mechanism of bacterial kill
is the same as with liquid or gaseous chlorine.
B. Shipment Weight Adverse environmental impacts associated with the
Under 454 kg 0.1
454 to 4,500 kg 5.0 use of chlorine are also applicable with sodium
4,500 to 13,600 kg 6.3 hypochlorite. Concentrated solutions from 10 to 15
13,600 to 27,200 kg 6.1 percent chlorine must be stored in rubber-lined steel
27,200 to 40,900 kg 2.5 or fiberglass storage tanks. Sodium hypochlorite is a
40,900 kg and over 80.2 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
Table 3-6. Breakdown of Chlorine Shipments 'by operate and maintain. The primary advantage of
Transportation Mode and Container (4) hypochlorination over chlorination is the increased
Annual Average safety in transporting, storing, and handling of
Percentage metric ton-km chemicals; however, chemical costs per Linit of free
(for 1972) (for 1971 to 1980)
chlorine are generally much higher.
Rail 84.9 1858
Cylinders to 114 kg 0.1 2.2 Dechlorination, normally with sulfur dioxide, has
0.91 metric ton cylinders 5.0 109 been used to reduce the environmental impacts and
Tank Truck 9.9 219 concerns associated with chlorination or hypochlori-
Barge 0.3 6.4 nation. Other dechlorinating agents such as other
lOQ.2 m4.6 sulfite reducing compounds or activated carbon have
been used; however, costs normally make these
options prohibitive. The levels of total chlorine
a corresponding breakd.own by ton-km, assuming a residuals can normally be reduced to below levels
yearly total of 2191 mimon ton-km. that are toxic to aquatic life. The potential for
formation of halogenated organics may be reduced;
The results ofthis analysis, shown in Table 3-6, were
however, it appears that many halogenated organics
then used to normalize the accident data per ton-km
are formed rapidly upon chlorine addition, and the
as shown in Table 3-7 .It can be observed in Table 3-7
application of sulfur dioxidl3 will probably not affect
that the accident rate fOIl' truck-transported cylinders
these compounds.
was consistently higher than the other categories
Dechlorination facilities are similar in most respects
Chlorination systems am reliable and flexible, and the to chlorination systems. Sulfonators meter gaseous
equipment is not complElx. It is relatively easy to apply sulfur dioxide and are similar to chlorinators. The
and control in wastewater treatment, and low use amount of sulfur dioxide applied is normally based on
Table 37. Accident Rates Per Metric Ton-KM (4)
Property Chlorine
Damage Released
Deaths Injuries ($) (kg)
Railroad 4.3 x 10- 10 1.4 X 10-8 6.0 X 10- 5 3.3 X 10- 5

excluding Youngstown 0 4.7 x 10-9 1.2 X 10- 6 2.3 X 10- 5

Cylinders to 114 kg 0 2.7 X 10- 6 3.6 X 10-4 1.2 X 10-4
0.91 metric ton
containers 0 1.4 x 10- 8 2.15 X 10-5 1.0 X 10-6

Tanker Truck 0 3.2 x 10-8 0.7 X 10-5 4.6 X 10-8

Barge 0 4.7 X 10-8 0 *

*lnformation not available.

flow and/or chlorine residual signals because con- applications of chlorine dioxide disinfection of mu-
tinuous sulfur dioxide monitoring equipment is not nicipal wastewaters are known at the present time.
presently available. Reaction with chlorine residual is
very rapid and separate contact chambers are nor- 3.5 Bromine Chloride
mally not used. Chlorination/dechlorination is more Bromine chloride hydrolyzes to hyprobromous acid
complex to operate and maintain than chlorination or very rapidly and forms monobromamines and di-
hypochlorination alone. The total cost of chlorination bromamil1es when added to water containing am-
will be increased by approximately 30 to 50 percent monia-nitrogen. Bromamines have been shown to be
with the addition of the dechlorination step. very effective as a disinfectant with shorter lived
residuals compared to chloramines. Because brom-
3.4 Chlorine Dioxide amines are more unstable than chloramines, shorter
Chlorine dioxide has been used to disinfect potable contact times are needed. Environmental impacts
waters, especially waters containing phenols or other associated with bromine chloride disinfection are
taste and odor producing compounds. It is a proven less adverse than those associated with chlorine.
bactericide equal to or greater than chlorine in Contrasting results relative to the toxic effects of.
disinfecting power, and has a higher oxidation bromine residuals have been cited in the literature,
potential. Chlorine dioxide has been indicated to by a possibly due to the stability of the bromine residual
much more effective virucide than chlorine. The and their rapid decay to chloride and bromide. Other
environmental impacts associated with chlorine brominated organic compounds can also be formed
dioxide use as a wastewater disinfectant have not during the disinfection process. These compounds
been well established. Advantages of chlorine dioxide may not be produced in appreciable quantities;
include its lackofreaction with ammonia and reduced however, they may be toxic at high enough concen-
halogenated organic compound formation. trations and may bioaccumulate.
Chlorine dioxide is an extremely unstable and ex- Bromine Ghloride is manufactured from bromine and
plosive gas and any means of transport is potentially chlorine. It is a hazardous and corrosive chemical and
very hazardous. Therefore, it must be generated on- requires special transportation, handling, storage,
site. Gaseous chlorine dioxide is normally generated and use precautions. Bromine chloride is shipped in
using a process that involves a reaction between modified chlorine cylinders, or tank cars, and es-
sodium chlorite and chlorine. Sodium chlorite is sentially the same safety and handling procedures
much more expensive than gaseous chlorine per unit are required as required for chlorine.
weight expressed as chlorine. Chlorine dioxide has
not received a great deal of attention as a wastewater Bromine chloride disinfection facilities are very
disinfectant due to its on-site generation require- similar to those used for chlorination. It is applied in
ments and high chemical costs. The overall system is much the same manner as chlorine, using a vacuum
relatively complex to operate and maintain compared chloJ:obrominator to prepare a concentrated solution,
to chlorination. Safety hazards include the handling which is then mixed with the wastewater. Shorter
of two dangerous chemicals. Chlorine dioxide has contact times are possible, resulting in a construction
been shown to be an effective wastewater disin- cost saving due to a smaller contactor. The use cost is
fectant; however, few pilot studies have been done heavily dependenton the cost ofthe bromine chloride
and experience is quite limited in this country. No itself. Secondary treated yvastewaterl'l r13quire no
special pretreatment prior to disinfection; however, any remaining ozone would be extremely irritating
bromine chloride use, for wastewater disinfection is and possibly toxic, the off-gases from the contactor
relatively new and a ~Jood data base and track record must be treated to destroy the remaining ozone.
are presently not available. Bromine chloride disinfec- Ozone destruction is normally accomplished by
tion has been proven effective in field-scale applica- thermal or thermal-catalytic means.
tions for wastewater, but problems with the liquid
feed equipment have occurred. Brominated organics An ozonation system can be considered to be
such as bromoform and mixtures of chlorinated and relatively complex to operate and maintain compared
brominated organics can be expecte~ to be formed. to chlorination. The process becomes still more
complex if pure oxygen is generated on-site for ozone
3.6 Ozone production. Ozonation system process control can be
Ozone is an unstable gas that is produced when accomplished by setting an applied dose responsive
oxygen molecules arEI dissociated into atomic oxygen to wastewater flow rate (flow proportional control), by
and subsequently collide with another oxygen mole- residual control, or by off-gas control strategies.
cule to produce ozone. The ozone molecule can co- Ozone disinfection is relatively expensive, with the
exist as a gas with other gases such as air or oxygen, cost of the ozone generation equipment being the
or it can dissolve in a liquid such as water. Ozone is an primary capital cost item, especially since the equip-
extremely reactive oxidant and a very effective ment should be sized for the peak hourly flow rate as
bactericide and virucide. Over 40 full-scale facilities with all disinfectant technologies. Operating costs
have been constructc:ld in the United States, and an can also be very high depending on power costs, since
extensive amount of research and development has ozonation is a power intensive system. The important
been .expended to develop ozone disinfection of criteria for design include maximum transfer effi-
wastewaters. ciency in the contactor to maximize ozone utilization
and minimize applied dose and power consumption
Unlike some ofthe other chemical disinfecting agents requirements along with efficient ozone generation
previously discussed, ozone can exert beneficial equipment design. Equipment design is presently
impacts on the envimnment. Since ozone decompos- provided by the equipment manufacturers, and they
es rapidly to oxygen after application, the dissolved are improving and updating their equipment to
oxygen levels in the treated effluent can be eleyated improve production efficiencies and reduce the
significantly, often to saturation levels. In most, if not associated operating costs.
all cases, the need for effluent reaeration to meet
required dissolved oxygen water quality standards 3.7 Ultraviolet light
can be eliminated. Ozone residuals can be acutely The effectiveness of ultraviolet light as a bactericide
toxic to aquatic life; however, since ozone dissipates and virucide has been well established. It is a physical
rapidly,' residuals al'e normally not found in the disinfecting agent compared to the previously dis-
effluent by the time it reaches the receiving water. cussed chemical agents. Radiation at a wavelength of
Ozonation has been. shown in some instances to 254 nm penetrates the cell wall and is absorbed by
produce toxic mutagenic and/or carcinogenic com- the celluli3r nucleic acids. This can prevent replication
pounds, but little is presently known about these and cause death of the cell. Since ultraviolet light is
organic byproducts (15). Ozone is believed to present not a chemical agent, no toxic residuals are produced.
fewer potential environmental and health hazards Although certain chemical compounds may be altered
than chlorine. by the radiation, the energy levels used for disinfec-
tion are too low for this to be a significant cause for
Due to the instability of ozone, it must be generated concern. .
on-site from air or oxygen carrier gas. The most
efficient method of producing ozone today is by the Major advantages of ultraviolet light are its simplicity,
electric discharge tec:hnique, which involves passing lack of impact on the environment and aquatic life,
the air or oxygen c~arrier gas across the gap of and minimal space requirements. There is a negli-
narrowly spaced electrodes under a high voltage. Due gible likelihood of producing harmful chemicals in the
to this expensive method of producing ozone, it is wastewater. Required contact times are very short,
extremely important that the ozone is efficiently' on the order of seconds rather than minutes. The
transferred from the gas phase to the liquid phase. equipment is simple to operate and maintain, but
The two most often used contacting devices are fouling of the quartz sleeves or Teflon tubes must be
bubble diffusers and turbine contactors. With the dealt with on a regular basis. Fouling is normally
bubble diffusers, deep contact tanks are required. . handled by mechanical, sonic, or chemical cleaning.
Ozone transfer efficiloocies of 85 percent and greater . High suspended solids concentrations, color, turbid-
can be obtained in most applications when the ity, and'soluble organic matter in the water can react
contactor is designed properly. The contactors must with or absorb the ultraviolet radiation reducing the
be C?overed to contre,1 the off-gas discharges. Since disinfection performance. High levels of wastewater
disinfection (e.g., 2.2 total coliform's 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-

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 ofthe 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.

Kinetics and Hydraulic Considerations

4.1 Disinfection Kinetics coliform decay rate coefficient was 0.03 per day.
Based on these results and using Equation 4-1, a
4.1.1 Natural Die-Off hydraulic residence time of at least 23 days in the
The natural processes of dilution, physical removal, summer months would be required to reduce an
and die-off or inactivation can reduce pathogen influent coliform level of 10 7 organisms/1 00 ml to an
levels. The impacts of dilution on the concentration of effluent coliform level of 102 organisms/100 ml.
organisms in a wastewater discharged into a stream, Redesign of lagoon systemsto meet stringent effluent
river, or lake can be estimated. Several mathematical coliform disinfection criteria appears to be econom- '
models have been proposed for estimating bacterial ically impractical, especially because of the low die-
die-off, and these models are mostly based on first away rates during winter months. The only remaining
order kinetics. For example, the first order model for alternative consists of adding disinfection processes
die-off in streams is (1): to the final lagoon effluents. .
N ::: No e (-ktl
4.1.2 Die..Off In the Presence of a Disinfectant
where: In cases where 'natural die-off is not sufficient to
prevent the potentia! for humans to ingest pathogenic
No ::: the initial concentration of microbes dis- organisms, disinfection should be required prior to
charged into the stream discharge. Disinfection is a time-dependent process.
N ::: the concentration of the microbes t time The ultimate outcome of bacteria and virus destruc-
units after discharge into the stream tion is the result of a series of physical, chemical, and
k ::: rate constant bioohemical actions that can be approximated by
simple kinetic expressions. It must be pointed out that
The first order model for die-off in standing water although the kinetic descriptions are simple, applica-
bodies (i.e., lagoons, lakes, etc.) is~Jollows: tion ofthese 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
No::: the concentration of microbes in the water The information needed for the design of a disinfec-
body's inflow tion system includes knowledge of the rate of
N::: the concentration of microbes in the water inactivation of the target organism(s) by the disin-
body's discharge fectant. In particular, the effect of disinfectant
td::: the hydraulic detention time in the water concentration on the rate of the process will deter-
body based on the water body's discharge mine the most efficient combination of contact time
k ::: rate constant and disinfectant dose to use. The major precepts of
disinfection kinetics were first enunciated by Chick
The rate constant k can be determined from a die- (4), who recognized the close similarity of microbial
away study of typical lakes or streams in a planning inactivation by chemical disinfectants to chemical
area. Rate constants may also be found in the reactions.
literature (2).
Chick stated that "disinfection is a gradual process,
Johnson, et al. (3) found that coliform die-away or without any sudden effects, and if the disinfectant is
removal rates can be as much as 16 times higher in sufficiently dilute to admit a reasonable time being
summer months when compared to winter months. taken for the process, the reaction velocity can be
Results of a study on the Logan City, Utah lagoon studied by enumerating the surviving organisms at
system indicated that the summer coliform decay rate successive intervals of time." Therefore, for a given
coefficient was 0.5 per day, and that the winter number oJ organisms and chemical disinfectants, the
rate ofdisinfection (number of organisms inactivatedl efficiency of ozone mass transfer and greater ozone
volume-time) can be desicribed by: decay.
dN The observation has frequently been made that
- Cit = kN (4-3)
inactivation of organisms in batch experiments, even
when disinfectant concentration is kept constant,
where: does not follow the exponential decay pattern pre-
dicted by Equation 4-3. Two common types of devia-
dN = rat.e 0 f c h ange In
- Cit . organism
. popu- tion are noted, as shown in Figure 4-1. In addition to
lation the linear Chick's Law decay, the presence of
"shoulders" or time lags until the onset of disinfec-
k = organism die-off rate constant tion, is often observed (Curve 2 in Figure 4-1); Also,
N = number of surviving organisms .per some organisms and disinfectants exhibit a "tailing"
unit volume at any given time wherein the rate of inactivatipn progressively de-
creases (Curve 3 in Figure 4-1). In some cases, a
Equation 4-3 expresses the rate of die-off of micro- combination of both of these behaviors can even be
organisms as an empirical first order kinetic model observed.
and is commonly referred to as Chick's Law. Chick's
,Law is presented graphically in Figure 4-1. Diver- Figure 4-1. Chick's Law and deviations.
gences from exponentiul decay are commonly ob-
served, and it is recognized that many factors can
cause these deviations, such as changes in disin- o
fectant concentration with time, differences in resist-
ance between individual organsims of various ages in
the same culture, existence of clumps of organisms, Curve 2
occlusion of organisms by suspended solids, etc. -Presence of "Shoulders"

Watson analyzed data with varying concentrations of

disinfectant and demonsitrated a definite logarithmic
relationship between ooncentration of disinfeotant
and the mean reaction velocity (5). He proposed the -2
following equation to relate the rate constant of
inactivation to the disinflectant concentration:
k ;;: k'C n (4-4)


C =disinfectant conc(mtration
n =coefficient of dilution -4
k' = correcteddie-off rate constant, presumably
independent of disiinfectant concentration and Curve 3
organism concentration -Presence of "Tailing Off"
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:
o 2 3 4 5 6
kf = k20 BIT-201 (4-5) Time

where: The presence of "shoulders" can be accounted for if it

kf = rate constant at some temperature T (0C) is assumed that multiple targets within each organ-
k&o = rate constant at 20C ism must each be damaged by independent dis-
8 = empirical constant infectant molecules prior to kill (6). An alternative
explanation is that diffusion through microbial outer
Little is known about disinfection efficiency at layers, or binding of disinfectant to the cell is a
elevated temperatures, bllJt for agents such as ozone, necessary precursor to inactivation (7). The "tailing"
a significant reduction would occur due to the lower survivor curve may be explained (in the absence of
changes in disinfectant concentration with time) by Figure 4-2. Presentation of the residence time distribution
either inherent differences in sensitivities of organ- (RTD).
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 RTD Curve IE-Curve)
resistance in survivors over the time course of Residence
exposure to disinfectant (8). The tailing phenomenon . Time
has been explained by Poduska and Hershey (9), who Distribution
assumed different subpopulations with differing E
sensitivities, and by Hom (10), who developed a
flexible, but highly empirical kinetic formulation
based on modifying Equations 4-3 and 4-4 to the
following form:

dN One should note that in these discussions, it is

(4-6) assumed that the system is at steady-state, that there
are no reactions taking place (conservative tracer),
Depending upon the value of m, both "shoulders" and and thatthe density of the liquid is uniform through-
"tailing" may be depicted by Equation 4-6. out the reactor.

4.1.3. Summary of Kinetic Considerations Given an E curve, the system mean residence time
In reality, the rate of bacterial kill generally increases may be determined using:
or decreases with time. This can be attributed to 00
several factors such as oxidation, complexation, poor
mixing, cell resistance, dispersion, clumping, and
(J = f tEdt (4-8
others. Deviations from Chick's Law can generally be
described by modifying the first-order expressions. It It is noted that this definition is similar to that for a
must be emphasized that all the reported disinfection moment of inertia. However; knowing the value for (J,
kinetic equations are empirical in nature. Unfor- an infinite number of E curves may still exist, differing
tunately, Chick's Law does not accurately predict in the spread of actual residence times. It can readily
coliform numbers as a function of dose in real world, be shown that (J is identical to V/O in systems such as
continuous flow systems, and therefore, the kinetics those considered here (11). The determination of the
of disinfection, as with any process, must be deter- system RTD provides information on the numerical
mined experimentally. As a result, pilot studies are spread of actual residence times, otherwise referred
required for effective design of any disinfection to as the dispersion. .
system. But pilot studies are expensive and many
times not conducted. Therefore, the need exists for Thus, the evaluation of a specific reactor relies on the
compiling all the available data and information on experimental determination of RTD for that reactor.
disinfection to address this problem. This can be accomplished by a number of procedures;
4.2 Mixing and Contactor Hydraulics subsequent analysis of the RTD curves may be used
In dealing with the hydraulic behavior of any con- to characterize mixing behavior of the unit.
tinuous flow reactor, the ideal situation is one .in
which the velocity profile is known at any point within 4.2.1 Experimental Determination of the
the reactor. This most often is not possible; the flow Residence Time Distribution Curve
characteristics are instead partially described by the The means to develop the RTD of a reactor is generally
residence time distribution (RTD).lf one considers the referred to as a stimulus-response technique. The
time each element resides in the reactor, the fre- state of the reactor is pe'rturbed in some fashion and
quency distribution of these times when plotted the resulting response is observed. In this case, the
against the time forms the residence time distribution stimulus is the input of a conservative (non-reactive)
curve. This is presented graphically in Figure 4-2. By tracer to the fluid entering the vessel and the
normalizing the distribution, the RTD can be repre- response is the time record of the tracer leaving the
sented in such a way that the area under the curve is vessel. The method by which the tracer is introduced
equal to unity: to the fluid will influence the type of response seen at
the outlet ofthe vessel. Typically, direct experimental
00 tracer analyses entail two types of inputs. The.first is
f . Edt = 1 the pulse input in which the tracer is injected in a
o relatively short period of time to the influent. This is
shown on Figure 4-3(a); the response is referred to as
where E = the residence time distribution. os C-curve.
Figure 4-3. Representation of pulse and step inputs and ative in the carrying fluid is injected into the influent
resulting outputs. 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
(a) the effluent by rapidly taking aliquots for subsequent
Pulse Input
analysis or by direct instream measurements (e.g.,
Tracer //~ conductivity). For RTD determinations in chlorine or
ozone .systems, batch sampling techniques suffice
C Output because of the long system mean residence times.
The nature of typical UV system designs presents
I----&;~-------"'""- - - - - -- _ logistical problems to the ~xperimental procedures
,I for the tracer studies. These arise from the very short
o Time
average residence times typical of most UV systems,
Tr,llcer Step input (b) and in some cases to thl~ difficulty of injecting the
(elevate concentration)
tracer at an appropriate point in the approach to the
reactor. The average residence times generally
Output encountered in these systems are between 1 and 50
(F-curve) 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-
I--!,-------=- - - - .-:.- _. - - - __ tration. High frequency automatic samplers and a
o Time very rapid injection of the tracer at the upstream
"F----.----.:- lct portion are generally required. Johnson and Qualls
(12,13) described such a system for a pulse injection
and sampling method. The method is generally

r Output
(F-curve) appropriate where there are distinct inlet and outlet
pipes on the UV reactor (see Figure 4-4(a. The tracer
Step Input
,r (~;hut-off 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 ofthe tracer and wastewater prior to
Time entry into the reactor. Samples should be drawn
immediately after exiting the reactor; alternatively,
The second method is the step input. In this case, the conductivity can be measured directly if salt is used as
tracer is introduced as a constant input, Le., the fluid the tracer.
is adjusted from one st,eady-state concentration level
to a new steady-state concentration level. The A step injection method was demonstrated in the Port
response is the time re(:ord of the concentration in the Richmond study (14). The reactors in this evaluation
effluent reaching this adjusted level, as shown on were open vessels, in effect simulating open channel
Figure4-3(b) and (c). This response is referred to as an flow. Thus, there were 110 distinct inlet or outlet
F-curve. The step input can be applied in either constrictions in which to inject or measure the tracer.
direction: on Figure 4-:3(b), it is shown as an input to Additionally, the times were very short, such that
increase the concentration; the tracer can also be direct sampling of the effluent would be difficult. The
applied as a steady input and then shut-off, with the technique that was developed (Figure 4-4(b incorpo-
response as illustrated by Figure 4-3(c). The E-curve rated the use of the F-curve; the step input was
(Figure 4-2) and the C-curve are equivalent: generated by first injecting a steady-state stream of
tracer into the unitand then shutting off the injection.
E =C (4-9) A conductivity probe was used to search fol' the point
of maximum concentration on the exit side of the
The F-curve can be tnmsformed to the C-curve by
lamp battery. Note that the probe can also be used to
taking its derivative:
scan the entire exit plane in order to define the
dF/dt = C (4-10) conductivity profile and location of the plume as it
exits the lamp battery.
The utility of this equation will be demonstrated in a
later discussion. Once the probe is situated and fixed at the center of
the plume, a steady-state condition is allowed to
Experimentally, the development.ofthe RTD curve is develop at fixed wastewater and salt solution flow
fairly straight-forward. A substance that is conserv': rates. The high frequencyou~putfrom the probe is
Figure 44(a). Techniques for experimentally determining Figure 44Ib). Techniques for experimentally determining
the RTD curves of UV reactors. (closed vessel the RTD curves of UV reactors. lope" vessel
analysis) tracer analysis)

Event Signal
Pulse Injection
// Wheatstone
Tracer Bridge
Injection Conductivity Probe



Sample Port o
Rotating I/--U--n
Sampling Tray \ ')
""",- - --Ll/'

Cone. dC/dt Trace after shutoff
.r- .
of step input

Time After Shutoff of Tracer

Time After Injection characteristics of a system and indicate if it conforms

with proper design for a disinfection process.
amplified and recorded by an oscillograph in this
case; it is necessary to have a high speed recorder in Consider the RTD curves on Figure 4-5, for three
order to obtain a readable and undistorted trace in the different flow regimes. The responses shown are all
short time frames. Once steady-state is indicated by to a pulse input. The worst case for a disinfection
the recorder, the salt solution pump is shut-off. This reactor is shown as a complete mixed flow reactor
event should be automatically signalled and recorded. (Figure 4-5(a)). A pulse input enters the reactor and is
The die-away of the salt tracer is then monitored by completely mixed in the total volume of the reactor
the fixed probe and continuously recorded. Readings instantaneously. This implies that a fraction of the
(conductivity) can then be taken of the tracer, input will be immediately discharged without any real
converted to salt concentration (mg/I), and trans- time of contact with the disinfectant. This is a wholly
. posed to a plot of concentration against time. For a unacceptable condition fora disinfection process.
particular type of system this can also be repeated at Recall that the inactivation of microorganisms re~
several points across the plane of the lamp battery. quires the time element; thus, in a completely mixed
reactor, a significant fraction of the microorganisms
The step input method is generally applicable to any in the wastewater will exit with little chance of being
type of reactor system. It is particularly useful in inactivated.
situations where there is a short average residence
time and where direct in-line monitoring of the tracer The ideal case is the pure plug flow (Figure 4-5(b)).
is possible. The response to the pulse input is a spike of
essentially zero Width. This implies that every element
4.2.2 Analysis of Residence Time Distribution of the pulse input resides in the reactor for an equal
Curves amount ohime. This ideal situation is not achieved in
There are a number of uses for the RTD as a tool for actual applications; there will be some degree of
design and as a diagnostic to determine the effect of dispersion such that the width of the plug flow C-
the hydraulics on the system's performance. The curve will expand. The objective of the hydraulic
shape of the curve and the distribution of the area design aspect of a system will be to minimize the
under the curve will describe much of the hydraulic degree of this spread in the RTD curve. .

Figure 46. ExemplelS oir RTD curves for varioull flow variance. Whereas the mean locates the RTD, the
characteristic: variance will represent the spread of the distribution: .
q2 = 1: t~ CIL!.tl _ t m2 (4-14)
Flow Characteristics 1: CIL!.tl
I/t (a) (b) Artibrary (c)
Complete Plug Flow The parameters shown on Figure 4-5(d) are defined

I ~ as follows:
(Age Distribution)

Area = 1
'-/. .
Area = 1 ,

' \ Area = 1
::. '--_~' ....u.._\~.L.....<'---_~
= time at which the tracer first appears
= time at which the peak concentration of the
t- tracer is observed
t10 = time at which 10 percent of the tracer had
passed through the reactor
(d) t50 = time at which 50 percent of the tracer had
passed through thl:! reactor
C B = the meantime (centroid) of the curve
Concentration T = the theoretical mean time (Vv/Q)
Tracer t90 = the time at which 90 percent of the tracer
had passed through the reactor
These parameters describing the RTD curve can be
Time combined to a series of indices that in turn describe
too After-+- the hydraulic behavior of the reactor (15). In the
Input evaluation of reactors specifically designed for dis-
infection, the critical areas of the RTD are the initial
A third flow condition is displayed (Figure 4-5(c)). This portions and their relationship to the theoretical and
is in a sense an arbitrary response, having little the mean residence times. Referring to Figure 4-5(d):
similarity to a uniform flow condition. Such a curve is
indicative of different volume elements ofthe reactor tf/T ... Measures the most severe short-circuiting. In
yielding different flow responses; examples are .an ideal plug flow reactor, the ratio is one, arid
channeling, eddies, or quiescent zones. Such a approaches zero with increased mixing.
response from a contactor would suggest potential
problems. tp/T ... Measures the average degree of short-circuit-
ing and will indicate the presence of significant dead
A number of numerical characteristics can be used to space areas. This in turn provides an estimate ofthe
describe the RTD curve and particularly its spread for effective vol,ume of the reactor. This ratio will
a particular system. These can then be used to approach one in a plug flow reactor and zero with
determine the adequacy of its design. The C-curve on
Figure 4-5(d) is a hypothetical response to a pulse t 90 /t10 ... This is the Morrill Dispersion Index, and is
input for a given reactor. First, define the theoretical the ratio ofthe time for 90 percent ofthe tracer to pass
detention time of the reactor as the void volume V v to the time for 10 percent of the tracer to pass: It is a
divided by the rate of flow, Q: measure of the spread of the curve; a value of 1.0
would indicate the ideal plug flow case, and 21.9for
T = Vv/Q (4-11 ) ideal mixing. Reactor designs should allow for this
index to be less than 2.0 for an effective deSign.
The mean residencetimEI, B, isthe mean value, orthe
centroid of the distribution: BIT '" For any reactor, this should equal 1.0; this
would imply that full use is being made ofthe volume.
co co When this value is significantly less than 1.0, it
8 = J tCdt/ J Cdt (4-12) indicates that the effective volume is much less than
o o the actual volume. By definition, the ratio of BIT
cannot be greater than 1.0.
This can be approximateld by summing the areas of
discrete time intervals: t 5018 ... In an effective plug flow reactor, the RTD
curve is very similar to a normal or Gaussian
tm = (4-13) distribution. The ratio of the median (t50) 'to the
centroid (8), will be a measure of the skew of the
curve. A skew to the left, in which t5018 is less than
The mean of the RTD curve is the first moment; the -1.0, would.be detrimental to an effective reactor
second moment of the c:urve about its mean<is the 'design. -

i Dispersion Models single parameter for the resulting C-curve. The mean
The above discussion centered on indices that could and variance of this curve, C, are:
be readily obtained from the RTDto evaluate or
diagnose the flow regime of a. specific reactor. (fm,e tm,e = 1 (4-18)
However, recalling that the primary design approach = 0
incorporates the use of a disinfection model, it is
necessary to also evaluate the RTD as a means of and
estimating the dispersive characteristics of the 2
system. To accomplish this, one can write a mass OJ. = u = 2(~) = 2d (4-19)
bala nee over the reactor for a conservative substance ~ uX
(at steady-state), where E is defined as the dispersion
coefficient (length 2/time), u is the linear axial veloc- Thus, if the RTD closely approximates a normal
ity, and x is the axial distance from the inlet. distribution, the parameter d =(E/uX) can be obtained
from the C-curve. It can be shown that minimal error
o = .!... (d 2C) _ dC (4-15) 5 percent) will be incurred by the above estimate
uX dx2 dx when:
The dimensionless group, E/uX, is called the dis- d <0.01
persion number (d), and is a measure of the axial
dispersion in a reactor: In cases where a significant degree of dispersion
exists, (Le., d > 0.01), causing the C-curve to skew,
as E
- 0; plug flow 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.
as E
- 00; complete mix In open vessels, flow is undisturbed at the boundaries
of the vessel. For closed vessels the mean and
variance are:
Levenspiel (11) gives the solution to Equation 4-15 for
pulse inputs in relatively low levels of dispersion. First (fm,e::: tm,e = 1 (4-20)
define the dimensionless parameter, 0'; o
0' = t
o} = 2 _E_ -2 ( -!.. )2 (1 _e- UX/ E) (4-21 )
o uX uX
which describes time in units of the mean residence In the case of open vessels with a significant degree
time. In case of small values of d the solution of of axial dislPersion, the mean and variance are:
Equation 4-15 gives the symmetrical C-cufve:

1 exp - (1 - 0')2
8'm,e = _tm,e = 1 + 2 E (4-22)
Co = 2(l1d)1/2 4(d)
(4-17) o uX

As shown on Figure4-6, E/uX (or d), is effectively the

~ = 2 _E_ -8 ( )2.!... (4-23)
uX uX
Figure 4-6. Relationship of C-curve to E/uX for small Note that in situations of small E/uX, the RTD is not
degrees of dispersion. affected by boundary conditions. From Equation 4-22
. it can be implied that at values of E/uX 0.5,
"open" and "closed" vessels are equivalent.
It is suggested that for closed vessels, Le., plug flow
entrance and exit reactors, different values of E/uX
68% Total Area can denote the following degrees of dispersion: (11)

CD No Dispersion E/uX = 0
Low Dispersion E/uX = <0.01
Total Area = 1 Moderate Dispersion =
E/uX 0.01 to 0.10
High Dispersion. E/uX=>O.1

A distinct advantage of the use of the dispersion .

o number for the characterization of mixing behavior is
that several predictive models are available. In the potential for short-circuiting would exist. Also, each
chapter on chlorination, correlation for dispersion in stage of other high mixing type reactors, such as the
pipes and baffled contact chambers will be presented. turbine mixer reactor, would exhibit characteristics of
There are experimental measurements reported for a complete mix reactor. To minimize the effect of
UV and ozone systems; however, these have yet to be short-circuiting, multiple stages that are positively
reduced to simple cOlrrelations. isolated from each other should be provided.
The characterization of mixing in an ozone contact 4.3 References
basin is more compleJ( and is dependent on the type of
contactor used. For e:Kample, with turbine contactors When an NTIS number is cited in a reference, that
a high degree of mi>cing is a necessary part of the reference isavailable from:
process of dissolving ozone into the wastewater. At
the same time, the violent mixing helps bring the National Technical Information Service
organisms into contact with the ozone. On the other 5285 Port Royal Road
hand, back-mixing ca:n increase d, and care must be Springfield, VA 22161
taken' to ensure that good baffling is provided to (703) 487-4650
prevent short-circuitihg.
1. Hubly, D.et al. Risk Assessment of Wastewater
With bubble diffuser ()zone contact basins the mixing Disinfection. EPA/600-2-85-037, NTIS No.
is much less violent than with turbine contactors. PB85-188845, U.S. Environmental Protection
Even here, however, basin back-mixing in each stage Agency, Cincinnati, OH, 1985.
of the contactor due tl::> the flow of gas in that stage is
sufficient to change the liquid flow characteristics 2. Berg, G. Indicators of Viruses in Water and Food.
from a plug flowtowclrd a complete mix pattern. The Ann Arbor Science, Ann Arbor, Michigan, 1978.
tendency toward complete mix operation with in-
creasing gas flow for bubble diffuser ozone contact 3. Johnson, B.A., et al. Waste Stabilization Lagoon
basin was demonstrated byVenosa and Opatken (16) Microorganism Removal Efficiency and Effluent
by performing dye tests with and without gas flow Disinfection with Chlorine. EPA-600/2-79-018,
(Figure 4-7). NTIS No. PB-300631, U.S. Environmental Pro-
tection Agency, Cincinnati, OH, 1979.
Figure 4-7. Dye teat shows effects of g8S flow rate on plug 4. Chick, H. An Investigation of the Laws of
flow ChllJ\llcteristics through a bubble diffuser
ozone conitact basin. Disinfection, J. Hyg., 8, 92, 1908.
5. Watson, H.E. A Note on the Variation ofthe Rate
60....------------------,-----, of Disinfection with Change in the Concentra-
tfon of the Disinfectant. J. Hyg. 8, 536, 1908.
o =
Gas Flow 0 L/min
60 o Gas Flow = 20 Llmin 6. Chang, S.L., Modern Concept of Disinfection. J.
6. Gas Flow = 80 Llmin Sanit. Eng. Div., ASCE (97):689-707, 1971 .
. Liquid Flow =75 Llmin
7. Haas, C.N. Rational Approaches in the Analysis
of Chemical Disinfection Kinetics. Chemistry in
Water Reuse, Volume 1, p. 381-399 (ed. W.
8o Cooper), Ann Arbor Science, Ann Arbor, Michi-
~ 30 ga~1981; .
.~ 8. Cerf, O. Tailing of Survival Curves of Bacterial
.!! 20 Spores. J. Appl. Bacterial. 42(1), 1977.
9. Poduska, R.A. and D. Hershey. Model for Virus
Inactivation by Chlorination, JWPCF 44:738,
10 1972. '

10. Hom, L.W. Kinetics of Chlorine Disinfection in

an Ecosystem. J. Santo Eng. Div., ASCE 90
6 8 10 12 (SA1):1983-1994,1972.
Time, minutes
11. Levenspiel, O. Chemical Reaction Engineering.
Each stage of a bubble diffuser contactor would Second Edition, John Wiley and Sons, New
exhibit similar characteristics; therefore, a similar York, NY, 1972.
12. Johnson, J.D. and R.G. Qualls. 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,

13. Qualls, 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 Overffows, Draft Report submitted to
U.S. Environmental Protection Agency, WERL,
under Cooperative Agreement No. CR807556,

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,

Chapter 5
Halogen Disinfection and Dechlorination

5.1 Coverage 1826 (1). The first public health or environmental

application of chlorine appears to have been its use as
This chapter covers the use of chlorine, chlorine
a prophylactic agent during the European cholera
dioxide, bromine chloride and related compounds for
epidemic of 1831 (1), although Baker(2) asserted that
wastewater disinfection. In addition, design of sys-
Javelle water (chlorine gas dissolved in an alkaline
tems for the dechlorination of wastewaters will be
potassium solution) was used in France for waste
discussed. treatment as early as 1825.
The material in this chapter may be used in several The earliest U.S. reference to chlorine as a dis-
ways. Sections 5.2 to 5.6 present an overview ofthe infectant occurs in 1832, when the known germicidal
history of chlorination and the other halogens, and properties of chlorine for control of disease epidemics
the fundamental chemical and kinetic aspects of were summarized by Averill (3). It is particularly
these materials as disinfectants. The chemical as- noteworthy that these results were obtained approx-
pects of dechlorination are also reviewed. imately 50 years before the advent ofthe germ theory
of disease.
Section 5.7 covers in detail the specification and
design of the various components comprising a The first formal recognition of chlorine or chlorine
halogen disinfection system, and a dechlorination cpmpounds for wastewater treatment occurred in
system. The engineer may proceed directly to this 1854, by the English Water Commission (2), although
section, referring to material in Sections 5.2 to 5.6 as not until 1884 was wastewater in England actually
background. For application to a particular case under chlorinated (4).
consideration, the engineer may also use the design
example presented at the end of Section 5.7 as a During the last years of the nineteenth century,
checklist in the construction of other calculations. electrolytic generation of chlorine and hypochlorites
became sufficiently competitive with the chemical
Section 5.8 treats safety considerations in the use of oxidation synthesis routes to spur the use of chlorine
halogens, to which the design engineer must pay compounds for disinfection. The growth of chlorine
particular attention. Section 5.9 and the following disinfection for both water and wastewater applica-
material summarize important O&M considerations tions occurred simultaneously.
that should be addressed in the design stage, and
relate performance experience at several full-scale The threel;larliest applications of chlorine as a
treatment plants. wastewater disinfectant used patented on-site elec-
trolytic generation. In 1892, operations commenced
5.2 History of Halogen Disinfection at both Hamburg, Germany and Brewster, New York
5.2: 1 Chlorine and Hypochlorites
While chlorides in the form of salts were known by The electrolytic generation of chlorine for wastewater
the ancients, the first preparation of chlorine gas, . treatment was covered by the patenting of the Woolf
then known as "dephlogisticated marine acid air" is Process in 1893 (5), which superseded the earlier'
credited to Scheele who achieved this result from the Powers process for the chemical production of
reaction of manganese dioxide with hydrochloric acid chlorine by a method similar to Scheele's original
in 1774. However, it was not until 1808 that this procedure. Many variants on the electrolytic proqess
material was regarded by Davy as a chemical element, were reported up to and including the work of Rideal
and general acceptance of Davy's hypothesis did not in 1908.
occur until 1815 (1).
Systematic investigations of the efficiency of chlorine
Medical germicidal applications ofthe new compound disinfection of wastewater were carried out in the
soon followed, with hospital disinfection equipment first decade of the twentieth century. Three research
installed as early as 1823, and with the use of groups reported on studies conducted almost con-
chlorine in surgical applications by Semmelweis in temporaneously.
During 1907 and 19051, Phelps conducted laborato,ry While Phelps, Kellerman, and the Lawrence, MA
and field investigations of chlorination using chloride group contributed to knowledge concerning the
of lime. Field chlorination at Boston, Baltimore and application of chlorinated lime to wastewater dis-
Red Bank, NJ led to the conclusion that a dose of infection, perhaps the first serious study of the
several milligrams per liter chlorine and 15 minutes kinetics of wastewater disinfection, per se, was that
contact time resulted in effective microbial reduc- of Avery in 1913 (14). In contrast to prior work, and
tions. Furthermore, Phelps asserted that it was particularly, that of Phelps, Avery concluded that it
Infeasible to attain complete microbial reduction was important to achieve some minimum amount of
using chlorination. In his work, Phelps used the contact time between the points of chlorine addition
coliform organism as an indicator of disinfection and discharge. He suggested that this time be at least
efficiency. As a justification forthis, he demonstrated 20 minutes. Furthermore, Avery's studies were
that these organisms had equivalent sensitivity to the significant in noting the relative difficulty of disinfect-
typhoid bacillus in was;tewater chlorination (6). ing septic, as opposed to fresh, wastewater.

Kellerman et al. (7). conducted two field studies on the The use of solid hypochlorite compounds as dis-
chlorination of (slow) sand filter effluents in Ohio and infectants was superseded following the development
came to conclusions similar to Phelps. In addition, of practical and economic means of generating and
this latter work estimated costs for chlorination at two dispensing liquid chlorine. In 1909, production of
particular plants studied as $7.50 and $16/million liquid chlorine commenced at Niagara Falls, NY, by
gallons ($2.00 and $4.25/1,000 m 3 ) using chlor- Electro Bleaching Gas Co. In 1912, Dr. Georg Ornstein
inated lime. Futhermore, it was concluded that "the of this company developed and patented a device for
quantity of chlorine immediately absorbed cannot be the application and feeding of chlorine in solution
estimated from the determination of the oxygen form. Simultaneous to this, C.F. Wallace and M.F.
consumed factor of the sewage effluent," thus Tiernan developed and installed a direct feed gas
contradicting the hypothesis advanced by Rideal (8). chlorinator at the Boonton, NJ water works, and went
on to develop solution feed chlorinators. The acquisi-
The third set of Initial investigations was conducted tion and abandonment of the Ornstein patent rights
by Clark and Gage (9) at the Lawrence, MA, exper- by W&T in 1917 finalized the adoption of modern
iment station. In addition to corroborating the studies solution feed liquid chlorination treatment for both
at Ohio and by Phelps, these workers appear to have water and wastewater (13).
been the first to investigate the phenomenon of post-
disinfection regrowth. Their cost estimates for the In wastewater treatment, the liquid chlorine direct
use of calcium hypochlorite for disinfection were feed process was used as early as 1918 in Millville
$1.75 to$3.75/million gallons ($0.45 to$1.00/1 ,000 ,and Camp Merritt, NJ (14). The W&T solution feed
m3 ). liquid chlorination process was used at New Haven,
eT commencing in 1919, where costs were estimated
FolloWing these early studies, a rapid introduction of at $4/million gallons ($1.00/1 ,000 m 3 ) (15).
chloride of lime disinfection into wastewater treat-
ment plants occurred. While a commission of the City In light ofthe apparently rapid adoption of chlorination
of Milwaukee, WI rec:ommended, in 1909, that a for wastewater disinfection it is also of interest to
primary treatment/chlorination plant be built, the note that reservations were expressed then, some of
first major city to install chlorination appears to have which are still relevantto practice today. For example,
been Philadelphia, PA in October 1910 (10). In as noted above, the issue of regrowth of bacteria
addition, Baltimore, MID commenced chlorination in subsequent to chlorination and release was noted by
February of 1912 and Providence, RI commenced Clark and Gage at Lawrence (9). In addition, as early
chlorination of a 20 MGD (0.9 m 3 /s) stream that same as 1922 (16) reservations were expressed about the
year (1 0,11). As of 1913, chlorination facilities using relative resistance of coliforms to chlorine vis a vis
chloride of lime were installed at Red Bank, NJ; pathogenic bacteria, and the resulting adequacy of
Rahway, NJ; Shore Harbor, NJ; Ventnor, NJ; Atlantic the coliform test as an indicator of disinfection
City, NJ; Bridgeton, NJ; Keyport, NJ; Margate City, NJ efficiency.
(10); and at Pleasantville, NY (9) and College Park, CA
(12). Thegenesis of modern chlorination practice is seen in
the experiences at the Easterly and Westerly plants of
The rise in chlorine treatment of wastewater was the city of Cleveland, OH (17). At these plants, with
paralleled by the use of chloride of lime in water' average flows of 90 and 30 mgd (3.9 and 1.3 m 3/s),
disinfection. The first applications of this technology respectively, disinfection using liquid chlorine was
occurred in 1908 at Boonton, NJ and Bubbly Creek commenced in 1923, using heated water evapo-
(Chicago), and in 1909 at Poughkeepsie, NY (13). rators. In 1929, at Easterly, one-ton container
handling systems were installed, while at Westerly, Table 5-2. Development of Chlorination Installations for
the 15-ton unit tank car system was placed on line. At Wastewater Treatment after Laubusch (5, 20)
these two plants, over the period 1923 to 1929, costs Wastewater
averaged $4.09 and $5.50/million gallons ($1.08 Treatment % of Population with
and $1.45/1000 m 3 ), respectively. The final stage in Plants Wastewater Chlorination
development of modern wastewater chlorination Of Those
concepts was the demonstration of the validity of With with Waste Of Total U.S.
Year Total Chlorination Treatment Population
operational control based on measurement of the
1910 619 22 2.4 0.12
chlorine residual by Tiedeman (18). 1916 846 55 4.6 0.28
1934 3697 655 N/A N/A
The geographic and temporal spread of wastewater 1940 5580 1127 35.3 10.9
disinfection by chlorine is illustrated in Tables 5-1 1945 5786 1262 34.2 12.1
and 5-2, which indicate the geographic distribution of 1948'6058 1307 37.0 12.3
1957 7518 2216 49.5 N/A
chlorination plants in 1910 and 1916, and the growth
N/A not available.;,,,
in chlorination from 1910 to 1957.

5.2.2 Chlorine Dioxide unpublished study of chlorine dioxide at a full scale

The discovery of chlorine dioxide, produced from the wastewater treatment plant in France (24),
reaction of potassium chlorate and hydrochloric acid,
is attributed to Davy in 1811 (21). However, itwas not 5.2.3 Bromine Chloride and Bromine
until the industrial scale preparation of sodium Due to the relative economics of bromine and
chlorite, from which chlorine dioxide may more chlorine, little interest in the former compound as a
readily be generated, that its widespread use occurred disihfectant was shown until the Second World War,
(22). when shortages of chlorine and increased supplies of
bromine as a byproduct, occurred (24). McCarthy
Chlorine dioxide has supplemented and supplanted demonstrated that, as a potable water disinfectant,
chlorine as a bleaching agent in pulp and paper elemental bromine was of similar effectiveness as
manufacture (22); however, despite early investiga- chlorine (25). In a more detailed study,Wyss and
tions on the use of chlorine dioxide as an oxidant and Stockton demonstrated the relative insensitivity of
disinfectant (23), its ascendancy in both water and bromine disinfection to high ammonia nitrogen
wastewater treatment has been slow. c,oncentrations (26). These results were confirmed by
Johannesson (27), who subsequently attributed this
As of 1977, 84 potable water treatment plants in the finding to chemical properties of bromamines (28).
United States used chlorine dioxide treatment, al-
though only one of these relied upon it as a primary While the existence of bromine chloride has been
disinfectant (21). In Europe, chlorine dioxide is used known since the work of Balard in 1826, final
as either an oxidant or disinfectant in almost 500 equilibrium coefficients were not well defined until
potable water treatment plants (21 ). the work of Mattraw et al. (29). However, little interest
in the industria'i applications of bromine chloride
While there have been numerous laboratory and pilot were evidenced until work by Mills and associates at
plant investigations of chlorine dioxide disinfection of Dow Chemical (30-32).
wastewater, there does not appear to have been any
full-scale operating experience with this disinfectant In the, mid 1970s, several investigations of the
in wastewater treatment. The situation in Europe is effectiveness of bromine chloride in laboratory and
not known in detail; however, White cites one pilot plant disinfection of wastewaters were reported.

Table 5-1. Early Geographic Distribution of Chlorination Facilities (19)

1910 1916
% of Serviced % of Serviced
No. of Population with No. of Population with
State Plants Chlorination Plants Chlorination
California 1 0.7 1 0.7
Connecticut o o 1 5.9
Maryland 2 0.7 2 0.4
New Jersey 6 0.7 16 38.8
New York 1 0.4 1 0.2
North Carolina o o 1 8.2
Ohio 1 0.4 2 1.7
Pennsylvania 4 7.0 5 9.9
Texas 6 3.8 25 15.0

These included pilot or demonstration studies at pressed to pressures in excess of its vapor pressure,
Hampton Roads, VA (:33,34), Freedom District(Sykes- chlorine condenses into a liquid with the release of
ville), MD (35,36), Hatfield, PA (37), Hawaii (38-41), heat and with a reduction in specific volume of
and Michigan (42,43). approximately 450 fold. Hence, commercial ship-
ments of chlorine are made in pressurized tanks to
The number of full scale operating wastewater reduce shipment volume. When chlorine is to be
disinfection systems using bromine chloride is un- dispensed as a gas, it is necessary to supply thermal
known, although a number of bromine chloride energy to vaporize the compressed liquid chlorine.
installations for power plant condenser biofouling
control exist (24). Table 5-3 summarizes the major physical properties
of chlorine (44). Figure 5-1 summarizes the vapor
5.3 Chemistry and Physical pressure of chlorine as a function of temperature.
Characteristics clf Disinfectants These data are obtained from a publication of the
The fundamental mechanisms of action of disinfect- Chlorine Institute (45), from which more extensive
ants and the problems raised in the design of tabulations of the physical properties of gaseous and
wastewater disinfection systems can be related, in liquid chlorine may be obtained.
many cases, to the physical properties of the dis-
infectants themselve;s, and their chemical reactions Commercially, sodium hypochlorite and calcium
with other constituents that may be present. In this hypochlorite are also used as sources of chlorine
section, these properties will be reviewed. compounds for wastewater disinfection. The relative
amount of chlorine present in these alternative
5.3.1 Properties of Disinfectants sources of chlorine is expressed in terms of "available
chlorine." Chlorine and Hypochlorites
Elemental chlorine CI 2 is a gas of density greater than The concentration of hypochlorite (or any other
air at room temperature and pressure. When com- oxidizing disinfectant) may be expressed as available

Table 53. Physical Properties of Chlorine (44) .

Liquid Gas
Affinity for Water Slight Slight

Bolling Point (@ 1 atm) -34.05C (-29.3F)

Color Clear amber Greenish-yellow

Corroslvlty Extremely corrosive to steel in pres- Same as liquid

ence of small amounts of moisture.
See discussion.

Density, Ib/cu ft 88.79 @ 60F 0.2003 @ 32F and

(85.61 psia) 1 atm

Explosive limits (in air) Non-explosive Non-explosive

Flammability Non-flammable Non-flammable

Melting (freezing) Point (@ 1 atm) -100.98C Non-flammable


Odor Penetrating and irritating Same as liquid

Solubility Below 9.6C


Chlorine hydrate CI 2 :8H 2 0, may


Specific gravity (compared to 1.468 @ 32F and 3.617 atm

4 water)

Relative Vapor Density 2.482 @ 32F and

(Alp- 1) 1 atm.
Vapor pressure See Figure 5-1

Viscosity . 0.385 centipoise @ OC 167.9 micropoise

0.729 centipoise @ -76.5"C @ 100

Figure 5-1. Vapor pressure of liquid saturated chlorine gas. (46) (Reprinted by permission of The Chlorine Institute.)





1000 --


200 -

::J 100






-80 -40 -20 o 20 40 60 80 100 120 140 160


chlorine by determining the electrochemical equiv- may be said to contain 70.91 grams of available
alent amount of CI 2 to that compound. By Equation chlorine (identical to the molecular weight of Cb).
5-1, it can be seen that one mole of elemental
chlorine is capable of reacting with two electrons to Since calcium hypochlorite (Ca(OCI)2) and sodium
form inert chloride: hypochlorite (NaOCI) contain two and one moles of
hypochlorite per mole of chemical, respectively, they
(5-1 ) also contain 141.8 g and 70.91 g available chlorine
per mole. The molecular weights of Ca(OCI)2 and
From Equation 5-2, it can also be noted that one mole NaOCI are, respectively 143 and 74.5, so that pure
of hypochlorite (OCn may react with two electrons to preparations of the two compounds contain 99.2 and
form chloride: 95.8 weight percent available chlorine; hence they
are effective means of supplying chlorine for dis-
OCI- + 2e- + 2H+ = CI- + H20 (5-2) infection purposes.
Hence, one mole of hypochlorite is equivalent (elec- Commercially, calcium hypochlorite is available under
trochemically) to one mole of elemental chlorine, and a variety of trade names (HTH, Pittchlor, Perchloron)

as a dry solid. It is relatively stable in the dry form, Hypochlorous acid is a weak acid and may dissociate
subject to a loss in strength of approximately 0.013 according to:
percent/d (46).
Sodium hypochlorite is available in solution form (5-5)
commercially in strengths of 1 to 16 percent by KA = [OCI-] [H+ ]/[HOCI]
weight. It is not practical to provide higher solutions
since chemical stabUity rapidly diminishes with
increasing strength. At ambient temperatures, the The pK A of hypochlorous acid at room temperature is
half-life of sodium hypochlorite solutions varies approximately 7.6 (48).
between 60 and 1700 days, respectively, for solutions
of 18 and 3 percent available chlorine (46,47). The three equilibrium constants describing the free
chlorine system are each a function of system
When either chlorine gas or one of the hypochlorite temperature. Morris has provided a correlating
compounds is added to a water containing insign- equation for KA as a function of temperature (48):
ificant quantities of Kjeldahl nitrogen, organic mate-
rial, and other chlorine demanding substances, a In (KA) = 23.184 - 0.0583 T - 6908/T (5-6)
rapid equilibrium is established among the various
chemical species in sor:ution. The term "free available In Equation 5-6, T is specified in degrees Kelvin (OK =
chlorine" is used to refer to the total of the con- C + 273).
centrations of molecular chlorine (CI 2), hypochlorous
acid (HOCI) and hypclchlorite ion (OCn, each ex-
For engineering purposes the chlorine equilibria in
pressed as "available chlorine" as described prev-
water may be adequately described using the Gibbs
free energy and enthalpy at 298K with the assump-
The dissolution of gasElous chlorine to form dissolved tion of constancy of enthalpy over the practical range
molecular chlorine is expressible as a phase equi- of interest. For each of the important species, Table
librium, and may be described by Henry's Law: 5-4 indicates the relevant thermodynamic quantities.

CI2101 = Cl2laq) : H (mo.JIL-atm) = [CI 2IaQI]/Pcl. (5-3) Using the standard thermodynamic relationships
from the data in Table 5-4, the equilibrium constants
for each of the chemical reactions governing free
In Equation 5-3, quantities within square brackets available chlorine may be calculated. These constants
represent molar conclmtrations, P CI is the gas phase are summarized in Table 5-5.
partial pressure of chlorine in atmospheres, and H is
the Henry's Law constant.
Table 5-4. Thermodynamic Functions of Free Chlorine
The dissolved aqueous chlorine is capable of reacting Species-298K (49)
with water to form hypochlorous acid, chloride ions, Standard Gibbs Free Standard Enthalpy of
and protons as indicated by: Energy of Formation Formation (H)
Species (G) (kJ/moJ) (kJ/moJ)
Cl z (g) o o
CI2IaQ) + H20 = W + HOCI + CI- Cl z (aq) +6.90 -23.43
(5-4) HOCI -79.9 -120.92
OCI- -36.82 -107.11
KH = [H+] [HOCI] [CI- ]/[CI 2IaQ )] CI- -131.30 -166.94

The kinetics of this reaction have been discussed in Table 5-5. Equilibrium Constants for Free Chlorine
detail by Morris and have been found to be extremely (Applicability Approximately 283 - 308K)
rapid (47). The reaction mechanism appears to involve
KA = 7.349 X 10- 6 exp (-1660.89m (molll)
elementary reactions between dissolved molecular KH = 2.581 exp (-2581.93m (molz/IZ)
chlorine and hydroxyl ions. H = 4.805 X 10- 6 exp (2818.48m (mol/I- atm)
Note: T is in degrees Kelvin.
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 From the use of these equilibrium constants it is
reaction decreases with decreasing pH and incr'eas- possible to determine the relationship between gas
ing salinity; hence, the solubility of gaseous chlorine phase partial pressure of chlorine (PCI.) and solubility
maybe increased by the addition of alkali or by the use under two particular conditions of importance. If
of fresh, rather than brackish water. gaseous chlorine is dispersed into pure water with
n~gligible amounts of demand, buffering potential, Step 1-We may use the same values for KHand Has
and chloride ion, then the following equation can be the prior problem, since the temperature is the same.
used to determine solubility: A value for KA is needed, however.

S = PCI,H + (K HPCI ,H)'/3 (5-7) From Table 5-5:

where S is solubility in moll I as C1 2. K A = 7.349 X 10-6 exp (-1660.89/285)

= 2.'16 x 10-6 molll
To determine the solubility as free available chlorine
in mg/I multiply the latter quantity by 71,000. Also, since the pH = 6, [H+] =:= 10-pH = 10-6 mol/I
However, if a solution of chlorine is prepared using Step 2-Apply Equation 5-8:
water with significant buffering potential and/or
chloride, or if a solution is prepared from calcium or (1,000 mg/I free chlorine/71 ,000 = 0.0141 mol/I)
sodium hypochlorite, both of which may contain
chloride, then Equation 5-8 is appropriate. This would 0.0141 = P CI , x 0.0948
also be applicable to determine the partial pressure of x[1 + (3 x 10-4 )/(10-6 x 0.0141) + (3 x 10-4 )
chlorine above a solution containing the free chlorine x (2.16 x 1O-a)/[(1 O-a)2 X 0.0141]]
0.0141 = PCI, x 2061
PCI, = 6.84 x 10-6 atm
S = Pcl,H [1 + (KH/[H+] [CI-]) + (K AK H/[H+]2 [CI- ])]
The second example illustrates an important aspect
The use of these equations is illustrated byway of two to chlorine chemistry, namely that adjustment of pH
examples. upwards (in this case to pH 6) results in a substantial
diminution ofthe equilibrium gas pressure. This is of
Example 1: practical importance in the proper responses to spills
of hypochlorite, or solutions of chlorine. Adjustment
A solution of chlorine is prepared by bubbling chlorine of pH, by the addition of an alkaline material such as
gas at 0.5 atm pressure into pure water until lime or sodium bicarbonate will reduce the volatility
equilibrium is reached. The temperature is 54F of chlorine from such spills and thus minimize danger
(285K). Determine the solution concentration of free to exposed personnel.
chlorine. Chlorine Dioxide
Step 1-Determine constants at the temperature of Chlorine dioxide (CI02) is a neutral compound of
concern (285K) chlorine in the +IV oxidation state. It has a boiling
point of "1 C at atmospheric pressure. The liquid is
From Table 5-5: denser than water and the gas is denser than air (50).
Relevant physical properties of chlorine dioxide are
KH = 2.581 exp (-2581.93/285) = 3 x 10-4 mo1 2/1 2 summarized in Table 5-6 (51).
H = 4.805 X 10-6 exp (2818.48/285)
= 0.0948 mol /I-atm Table 56. Physical Properties of Chlorine Dioxide (50)
Property Value
Step 2-Apply Equation 5-7 to determine solubility: Melting point ':"59C

S = (0.50 x 0.0948) + (0.50 x3 x 10-4 x 0.0948)1/3 Boiling point 11C

= 0.0716 mol/I Density (OC) 1.640 glml (liquid)

2.4 gIl (vapor)
= 5,084 mg/I free available chlorine Heat of Vaporization 27.28 kJ/mol
Example 2: Critical temperature 153C

An effluent is used to make up a chlorine stock Vapor Pressure (OC) 0.626 or 0.673 atm
solution. The effluent is chlorinated until it contains
1,000 mg/I free available chlorine. At this point, it Heat of Solution (DOC) 27.61 kJ/mol
has a pH of 6.0 and a chloride ion concentration of Color
500 g/m 3 (this includes the initial chloride that was Solid Red
present). What would be the equilibrium pressure of Liquid Orange
chlorine in the gas phase?Thetemperature is 285 K. - Gas Orange

Chemically, chlorine d'ioxide is considered to be a ride:oxygen produced was 5:3:1 :.75, and that the
stable free radical. Hence, at high concentrations it addition of chloride enhanced the rate of decomposi-
may react violently with reducing agents. It is tion and resulted in the predicted 1:1 molar ratio of
explosive, with the lower explosive limit in the vapor chlorite:chlorate (55). This suggests that the oxidation
phase variously reportod as 10 percent (49,51 ) or 39 of chloride by chlorate, and the possible formation of
percent (50). As a result, virtually all applications of intermediate free chlorine, may be of significance in
chlorine dioxide requiro the synthesis of the gaseous the decay of chlorine dioxide in demand free systems
compound in a dilute stream (either gaseous or liquid) (53).
on location as needed.
The concentration of chlor'ine dioxide in solution may
The solubility of gasec,us chlorine dioxide in water be expressed in terms of "g/I as chlorine" by
may be described by Henry's Law, and a fit of the multiplying the molarity of chlorine dioxide by the
available solubility data (52) results 'in the following number of electrons transferred per mole of chlorine
relationship for the Henry's Law constant: dioxide reacted and then multiplying this by 35.5
grams CI 2 per electron mole. Conventional/y, the five
H(mol/I-atm) = [CI0 2 IsQI]/Pclo. electron reduction (Equation 5-14) is used to carry out
(5-9) this conversion.
= 3.819 x 10-6 exp (3680/T)
In dilute aqueous solution, the spontaneous dispro-
portionation of chlorinl3 dioxide into chlorate (CI0 3 -) Hence, one mole of chlorine dioxide contains 67.5
and chlorite (CI0 2-) leads to a continuous loss of grams of mass, and is equivalent to 177.5 (= 5 x 35.5)
measurable chlorine dioxide from solution. Under grams C1 2 In other words, one gram of chlorine
alkaline conditions, thEl following disproportionation dioxide contains 2.63 grams "as chlorine." In exam-
occurs (53): ination of any studies on chlorine dioxide, due care
with regard to units of expression of disinfection
2 CI02 + 2 OW = H 20 + CI0 3- + CI0 2- (5-10) concentration is warranted.
This reaction has been reported to be catalyzed by Bromine Chloride
carbonate ions; in thl3 absence of carbonate, the
Bromine chloride (BrCI) is a chemical compound
reaction is governed by parallel first and second order
formed by the combination of an atom of chlorine and
kinetics according to the following rate law (53,54):
an atom of bromine. Below-5C, BrCI exists as a dark
rCIO. = rate of C/0 loss; (moles/liter-second)
red liquid at atmospheric pressure. The vapor pres-
sure above bromine chloride liquid (Figure 5-2) IS
(5-11 )
= k, [CIOil + k [CIOil
2 intermediate between the vapor pressures for the
molecular halogens chlorine and bromine (Table 5-7).
k, = (5 x 105 + 1.7 x 10 11 [OW]) exp (-7348/T) (5-12)
Figure 5-2. Vapor pressure over liquid bromide chloride.
k2 = 2.7 X 109 exp (-7:348/T)
+ 7.3 x 108 [OW] exp (-3875/T) 60 ,..------,-----,,-----,-----,,---,-----,-----,

These kinetics imply that the half-life of aqueous 50

chlorine dioxide sol utions decrease substa ntia Ily with
increasing concentratkm, and that above pH 9, the
til o
'~ 40
further increase in pH decreases the half-life. Even at
neutral pH values, however, in the absence of :::I

carbonate at room tElmperature, the half-life of ~ 30

chlorine dioxide solutions of 0.01, .001, and 0.0001 0:
mol/liter is respectively, 0.5, 4, and 14 hours. Hence, g, 20
the storage of stock solutions of chlorine dioxide for ~
even a few hours is impractical. 10

Recent evidence also suggests that the simple

disproportionation reac:tion to chlorate and chlorite is Ol.-----L--J.----'-:-L----::!:----=~-_!=_--
-20 -10 0 10 20 30 40 50
Insufficient to explain the decay of chlorine dioxide in
water free of extraneous reductants. Equation 5-10 Temperature, C
predicts that the molar ratio of chlorate to chlorite
formed should be 1:1; however MediI' and Giralt Physical properties of BrCI are summarized in Table
found that the molar ratio of chlorate:chlorite:chlo- 5-7 in comparison to properties of molecular bromine

Table 5-7. Comparison of Properties of BrCl and Br2 and Aqueous solutions of bromine chloride in which the
CI 2 (57) concentrations of reducing agents are negligible
Property Br2 BrCI CI 2 undergo a series of reactions similar to solutions of
Freezing Point -7.3C -66C -101C chlorine gas. Bromine chloride may hydrolyze to form
hypobromous acid (HOBr) and chloride according to
Boiling Point -58.8C 5C -34C the following equation, which is analogous to Equa-
Density (gas, tion 5-4 for chlorine:
20C, 1 atm) 3.12 g/I 2.34 g/I 1.40 g/I
BrCI + H 20 = HOBr + H+ + CI- (5-18)
Vapor Pressure
atm, 20C 0.98 3.03 7.81 I<H = [H+] [HOBr] [CI- ]/[BrCI]
Solubility The value of K H at 25C has been reported as 2.94 x
(g/100 g H2 O) 3.3 8.5 0.75 10-5 M 2/1 2, which is approximately fivefold less than
that for molecular chlorine (30,60).
and chlorine. In general, the properties of BrCI are
intermediate between the two halogens, with the Hypobromous acid (HOBr) is formed on hydrolysis in
exception of the solubilityof BrCI, which exceeds that preference to hypochlorous acid, since the latter is
of either of the two molecular halogens. able and will readily act to oxidize bromide ion to
HOBr. Hypobromous acid is a weak acid and will
Commercially, bromine chloride is supplied as a dissociate in accordance with:
containerized liquid under pressure. Since the vapor
pressure over liquid bromine chloride at ambient HOBr = W + OBr- (5-19)
temperatures is substantially in excess of atmos- KA = [H+] [OBr-]l[ HOBr]
pheric (Figure 5-2), the chemical may be withdrawn
under its own pressure. Using thermodynamic data on hypobromous acid and
hypobromite (51), the following relationship for KA
In both liquid and gas phases, BrCI may dissociate may be derived for temperatures in the range of
into its constituent elements according to the follow- normal operation:
KA ::::: 5.24 x 10-6 exp (-2265/T) (5-20)
2 BrCI{g) = Br2(9) + C1 21g ) (5-15)
In general, the equilibrium constant for HOBr dis-
2 BrCl m = Br2m + C1 2111 (5-16) sociation is lower than that for HOCI at a given
temperature (Le. the pK for HOBr is higher than that
The equilibrium constant for the gas phase reaction
for HOCI). Hence, the percentage of bromine present
(Equation 5-15) is given by:
as hypobromous acid at a given pH is greater than
(5-17) would be the relative abundance of hypochlorous acid
in a system containing chlorine at identical concen-
This equilibrium has been well studied, and the value trations.
for Kg appears to vary only slightly with temperature Concentrations of bromine chloride can be expressed
and assumes the value of approximatelyO.12 at 25C in an analogous manner to those for free chlorine.
(29,30,58,59). No equilibrium constant for Equation From Equation 5-21, in comparison with equation (1),
5-16 has been reported; however, it has been it may be seen that one mole of BrCI is equivalent to
asserted on the basis of unpublished results that liquid . one mole of chlorine:
bromine chloride is less than 20 percent dissociated
(30). (5-21 )
Due to the unavoidable presence of molecular Since one mole of BrCI contains 115.4 grams of
bromine and molecular chlorine in both the gas and material, there are 0.62 (71 1115.4) grams of available
liquid phases of a bromine chloride system, the halogen as chlorine per gram of bromine chloride.
thermodynamic significance of the vapor pressure,
boiling and melting point data for BrCI is ambiguous. 5.3.2 Onsite Production Chemistry
In a closed vessel containing BrCI in both liquid and In the case of chlorine and chlorine dioxide, tech-
vapor phases, the vapor phase will be enriched in CI 2 nologies exist that enable a wastewater treatment
and the liquid phase will be enriched in Br2 under plant to produce disinfecting chemical at the plant
. normal conditions. To ensure withdrawal of bromine location itself, rather than relying upon shipment of
chloride containing a 1: 1 stoichiometric ratio of the externally generated materiaL In the case of chlorine,
halogens it is necessary that all withdrawals be made these technologies are alternative to external produc-
from dip tubes below the liquid level. tion, while in the case of chlorine dioxide the onsite
synthesis of disinfectant is mandatory due to de- electrolytic cell ranges from 1.73-2.19 volts. The
composition discuSSEld above. The chemistry asso- minimum theoretical energy required is the product
ciated with the production of these chemicals will of this and the quantity (2.72 C/mg available chlo-
now be discussed. rine). Thus, 4.7-5.96 Watt-seconds are required to
produce 1 mg available chlorine. This may also be Chlorine and Hypochlorites expressed as 1.31-1.66 kWh/kg available chlorine.
'Chlorine and hypochl()rites have been produced from
the electrolysis of brines and saline solutions'since In actual practice, it is necessary to operate electro-
the early days of thl3 20th century (8). The basic lytic chlorine generating units at voltages as high as
principle is the use of Cl direct current electrical field to 3.85 volts in order to increase the rates of the
effect the oxidation clf chloride ion with the simul- generation reaction. At these overvoltages, however,
taneous and physically separated reduction usually of additional oxidations such as formation of chlorate,
water to gaseous hydrogen. ohmic heating, and incomplete separation of hydro-
gen from oxidized products with subsequent dissi-
Atthe anode of such an electrolytic cell, one or both of pative reaction, combine to produce system inef-
the following reactions takes place: ficiencies. For typical electrolytic generating units,
current efficiencies (based on actual to theoretical
CI- = 1/2 CI 2 lg1 + e- EO = 1.36 volts (5-22) C/mg) of 97 percent may be obtained along with
(5-23) energy efficiencies (based on kw-hr/kg) of 58 percent
CI- + 20W = OCI- + H 0 2 + 2e- EO = 0.90 volts (49). These efficiencies are related to the physical
configuration of the electrolysis cells, brine concen-
In Equations 5-22 and 5-23, EO is the standard tration, and desired degree of conversion to available
electromotive potential at unit activity of all products chlorine (61,62).
and reactants at 298K, and the actual half-cell Chlorine Dioxide
voltage changes may depend slightly on concentra-
Theoretically, chlorine dioxide may be produced by
tion. This dependence may be determined by standard
either the oxidation of a lower valence compound or
reduction of a more oxidized compound of chlorine.
At the cathode of an electrolytic cell, reduction of Chlorites (CI0 2 -) or chlorous acid (HCI0 2 ) may be
water occurs produc:ing molecular hydrogen gas oxidized by chlorine or persulfate to chlorine dioxide,
according to Equaticm 5-24, written for alkaline or may undergo autooxidation (disproportionation) to
conditions. chlorine dioxide in solutions acidified with either
mineral or organic acids. Chlorates (CI0 3 -) may be
(5-24) reduced by use of chlorides, sulfuric acid, sulfur
H2 0 + e- = OH- ;. 1/2 H2 /g 1 EO = 0.83 volts dioxide, or oxalic acid, or electrochemically to form
chlorine dioxide (51).
Combination of Equations 5-22 and 5-24 or 5-23 and
5-24 yield the two overall possibilities, depending For practical purposes in wastewater treatment,
upon whether the chlorine is withdrawn as gaseous chlorine dioxide is generated exclusively from chlorite
chlorine evolving at the anode or as hypochlorite in inasmuch as the reductive processes using chlorate
spent brine. as a starting material are capital intensive and
competitive only at larger capacities (51).
cr + H2 0 = 1/2 C1 2 /g 1 + 1/2 H 2lg1 + OW
In, the acid-chlorite process, sodium chlorite and
EO :: 2.1 9 volts hydrochloric acid react according to:

CI- + H2 0 = OCI- + H2 0 + 1/2 H 2 lg1 (5-27)

(5-26) 5 NaCI0 2 + 4 HCI = 4 CI0 2 + 5 NaCI + 2 H 20
EO :: 1.73 volts

From these thermodynamic relationships, it may be The resulting chlorine dioxide may be evolved either
concluded that for each mole of Cb (or the equivalent as a gas, or removed in solution. Mechanistically, this
one mole of OCn produced, two moles of electrons process occurs by a series of coupled reactions, some
are required. Since there are 96',493 coulombs (= of which may involve the in situ formation of chlorine,
amfere seconds) per Iquivalent, theoretically 1.93 x the catalysis by chloride, and the oxidation of chlorite
10 coulombs are re,quired per mole of available 'by chlorine (50,51,53). In addition, the yield of the
chlorine, or 2.72 C/mll available chlorine. reaction as well as the rate of the process are
improved by low pH values in which both gaseous
Additionally, from Equations 5-25 and 5-26, the chlorine and chlorous aCid formation are favored.
minimum voltage difference between the plates of an Under these favorable conditions, the reaction pro-
ceeds in the order of minutes; however, to achieve To understand chlorine demand reactions, a further
these conditions, excess hydrochloric acid is required. definition is necessary. As noted above, the term
"free available chlorine" is used to denote the
Alternatively, chlorine dioxide may be produced by concentrations of hypochlorous acid plus hypochlorite
the oxidation of chlorite with chlorine gas according ion expressed on a mass equivalent Cb basis. The
to: term "total available chlorine" is defined as the mass
equivalent CI 2 contained in all materials that contain
2 NaCl0 2 + Cl2 = NaCI + 2 CI0 2 (5-28) chlorine in an oxidized state. The difference between
total available chlorine and free available chlorine is
As in the previous case, low pHaccelerates the rate of defined as ,icombined available chlorine" and repre-
this process, as does excess amounts of chlorine gas. sents the amount of chlorine that is in chemical
However, if chlorine gas is used in stoichiometric association with various compounds (usually amino-
excess, the resultant product may contain a mixture or ammonia-nitrogen) but that is also capable of
of unconsumed chlorine as well as chlorine dioxide. carrying out oxidation. The significance of this
distinction, which will be further discussed in the
5.3.3 Disinfection Demand Reactions section on kinetics, is that the free chlorine forms are
In the presence of such dissolved impurities as exist generally more effective disinfectants than the com-
in wastewater effluents, each of the halogen dis- bined chlorine forms.
infectants may undergo reactions in which they
decompose or transform to less effective chemical One example of this behavior, along with a graphical
forms. In the case of chlorine and bromine chloride, demonstration of typical chlorine behavior in waste-
these principally involve reactions with ammonia and water, is given in Figure 5-3 (64). In this depiction, as
amino-nitrogen compounds, while chlorine dioxide the dose of chlorine is increased, the total available
demand may result from reactions with other organic chlorine residual (Le. remaining in the system after
materials. The nature of these processes will now be 30 minutes) increases until a dose of approximately
. discussed. 50 mg/I, whereupon residual chlorine decreases to a
very low value, and subsequently increases linearly Chlorine and Hypochlorites with dose indefinitely. The "hump and dip" behavior
Available chlorine, whether in the form of chlorine is paralleled by the sensitivity of microorganisms to
gas or hypochlorite, when added to wastewater, can the available chlorine residual (indicated by the time
undergo a series of dissipative reactions that resu It in required for 99 percent inactivation of Bacillus
a loss of disinfectant from the system, or a change in metiens spores). At the three points indicated, the
disinfectant form to a less active chemical species. total available chlorine is approximately identical at
The following reactions of chlorine or hypochlorite 22 to 24 mg/I yet there is a 32-fold difference in
may occur (63): microbial sensitivity.

cyanide may react with chlorine to form chlorides, Fi~ure 5-3. Effect of increased chlorine dosage on residual
bicarbonate, .and a variety of oxidized nitrogen chlorine and germicidal efficiency: pH 7.0,
forms; 20,oC NH 3 1 0 mg/1. (64) (Reprinted by permis-
sion of the American WaterWorksAssociation.)
ammonia or amino-nitrogen groups may react with
chlorine to form chloramine compounds by. the
replacement of a proton by a chlorine atom; 120 r - - r - - , - - , - - r - - r - - , - - , - - , - - - - ,

organic molecules containing unsaturated (double I 100 ~~

or triple) bonds may react to form chlorinated -; 2l 83 min 88 min ;,<:-0
.2 '"
organic molecules; .~ 280 ~o 80 ca ...Ql
V'<:- .~ 0
.2~ 0~C:>' 0 Co
a variety of inorganic reducing agents, including t5 ~60 - ~,~ 60:g ~
hydrogen peroxide and reduced iron and sulfur Q) ... ~o - c
:E2l CJ'<:- "*.~
compounds~ may consume chlorine by redox ~<i: 40 - 24.2 . o~~ 40 ~ Gi
~ /ppm~0C:>L1" 2.7 min ~ .E
20 24.9 Residual =22.5 ppm - 20 ~ '0
Of these processes, probably the most significant ppm Available Chlorine i=
process in determining the fate of chlorine added to O,LJJ-L_JlJ-.;J_ I I I 0

wastewater is the reaction with ammonia or amino- W ~ 00 00 l00lWl~lOOlooWO

nitrogen groups. However, the reactions with organic Available Chlorine Added (ppm)
materials are also of importance in the production of
chlorinated organic byproducts during wastewater The explanation for this behavior is founded in the
chlorination. "breakpoint" reaction between free chlorine and
ammonia, which is pre-sent in non-nitrified effluents. Figure 5-4. Effect of chlorine or hypochlorite dose on pH of
At doses below the "hump" in the chlorine residual settled wastewater.(70) Reproduced from
curve, only combined clvailable chlorine is detectable Journal of the Water Pollution Control Federa-
tion, by permission.
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 g,-----,----...---------r----,--"::?""-..
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 8

\ ',l
completely oxidized, the residual remaining consists
almost exclusively of free chlorine. The minimum in
the chlorine residual \(s dose curve (in this case 70
Hypochlorite Solution
mg/l) is called the "breakpoint" and denotes the of pH = 11.1
amount of chlorine that must be added to a waste-
water before a stable free residual can be obtained.
The relative inefficien(~y of combined chlorine forms
as disinfectants vis a vis free chlorine forms has been 6
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 ____ Chlorine Water of pH =2.0
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:chlorin1a) of 1:10;
2 L.. --'- 1
..J- -'--_ _---'1'-- -'
in practice, mass dose ratios of 1:15 may be a 5 10 15 20 25
needed to reach breakpoint. Chlorine Dose, mg/I

In later work with breakpoint chlorination of waste-

waters, Griffin and Chamberlin found that measure- the total chlorine residual at a given contact time, is
ment ofthe 30 minute and 2 hour residuals failed to due to ammonia nitrogen, as well as organic and
produce a "hump and dip" curve, and that only at 18 inorganic compounds. In early work, Symons et al.
hours could a classical breakpoint pattern be observed determined that domestic wastewater contains an
(69). Furthermore, it was found that even above the average chlorine demand of 45 g/capita-d, including
breakpoint, the concentrations of organic nitrogen all sources (this is demand in the raw wastewater and
were unaffected. However, this work did confirm that fails to account for possible changes in quality
a mass dose ratio of 10 mg chlorine to 1 mg ammonia through prior unit processes) (71,72). In addition,
nitrogen results in brElakpoint at 18 hours in raw, superimposed upon this average were significant
primary, and secondary (activated sludge) waste- fluctuations of various frequencies.
It was found that chlorine demand fluctuated sea-
The breakpoint reaction may also affect the pH of a sonally approximately with temperature. Weekly and
wastewater. Figure 5-4 indicates that, if sodium diurnal fluctuations also existed, and the latter were
hypochlorite is used as the source of active chlorine, found to be more variable than flow. During storms,
as breakpoint occurs, the pH decreases due to an the chlorine demand was found to increase shortly
apparent release of protons during the breakpoint after the onset of the storm flow, and then to rapidly
process (70). If gaseous; chlorine is used, this effect is return to baseline levels.
obscured by the release of protons by hydrolysis of
gaseous chlorine according to Equations 5-4 and 5-5. The rate of exertion of chlorine demand in both
potable water and wastewater chlorination has been
In wastewaters, the chlorine demand, representing the subject of numerous studies. The most systematic
the difference between the applied chlorine dose and work has been that of Taras (73), who chlorinated

pure solutions of various organic compounds and These three compounds, monochloramine (NH 2CI),
found that chlorine demand exertion follows: dichloramine (NHCI2), and trichloramine (NCb), each
contribute to the total (or combined) available chlorine
D = kIt)" (5-29)
residual in a wastewater. Each chlorine atom asso-
where t is the time in hours, D is the chlorine demand, ciated with a chloramine molecule is capable of
and k and nare empirical constants. It was found that undergoing reduction to chloride, and in the process
n for various substances ranged from 0.02 to 0.29. accepting 2 electrons; hence, each mole of mono-
chloramine contains 71 grams available chlorine;
In subsequent work, Feben and Taras chlorinated each mole of dichloramine contains 2 x 71 or 142
potable water, and potable water blended with up to grams; and each mole of trichloramine contains 3 x
1.5 percent wastewater to a free residual. and found 71 or 213 grams of available chlorine. Inasmuch as
that the results could be correlated to Equation 5-29 the molecular weights of mono-, di-, and trichlor-
with the value of n determined by the one hour amine are 51.6, 86, and 110.5, this means that the
chlorine demand as follows (74): chloramines contain respectively 1.38, 1.65, and
2.02 grams available chlorine per gram. However, the
n = 0.18 - 0.017 log (D1) (5-30)
efficiency of the various combined chlorine forms as
where D1 is the one hour chlorine demand. disinfectants differs, and thus the concentration of
available chlorine also is insufficient to characterize
Lin and Evans extended the significance of Taras' process performance.
work by chlorinating secondary effluents to combined
residuals (75). They found that Equation 5-29 could As Equation 5-32 indicates, the formation of mono-
be used to describe the kinetics of chlorine demand chloramine is accompanied by the loss of a proton,
exertion except that two sets of k and n values were which corresponds to the experimental findings of
required. At times from 1 to 12 minutes, the rates of McKee at al. (70). The loss of a proton is due to the fact
chlorine demand development were significantly that chlorination reduces the affinity of the nitrogen
greater than between 12 and 60 minutes. It was also moiety for protons; this was verified experimentally
found that when calcium hypochlorite was used as a for ammonium as well as a variety of amines by Weil
chlorinating agent, a lower chlorine demand was and Morris (78).
obtained as compared with chlorination using dis-
solved chlorine gas. Under conditions where ammonia nitrogen is in
excess of the chlorine dose (Le. below the "break-
A more recent kinetic model for chlorine demand poinC) and where the pH is below 9.0 (so that the
exertion has been developed by Haas and Karra (76): dissociation of ammonium ion is negligible), the
amount of combined chlorine in dichloramine relative
D = Co (1 - X exp (-k,t) - (1 - X) exp (-k2t)) (5-31 ) to monochloramine after Equations 5-32 and 5-33
have attained equilibrium may be given by the
In Equation 5-31, X is an empirical constant typically
relationship described by McKee et al. (70):
0.4-0.6, and k, and k 2are rate constants, typically 1.0
min-' and 0.003 min-1, respectively, and Co is the A = BZ/(1- (1-Z(2-Z)B)) -1 (5-35)
chlorine dose in mg/1.
In Equation 5-35, A is the ratio of available chlorine in
While it has been generally observed that nitrified the form of dichloramine to available chlorine in the
effluents that are low in ammonia-nitrogen have form of monochloramine, Z is the ratio of moles of
lower chlorine demands in comparison to non nitrified chlorine (as C1 2) added per mole of ammonia nitrogen
effluents that have higher chlorine demands, recent present, and B is defined by:
work suggests some anomalies. White et al.(77)
observed that the addition of low levels of ammonia to B = 1 - 4 K eq [W ] (5-36)
nitrified effluents may reduce chlorine demand. This
The equilibrium constant in Equation 5-36 refers to
phenomenon may be due to competitive formation of
the direct inter-conversion between dichloramine
inorganic versus organic chloramine compounds,
and monochloramine as follows:
and will be discussed in greater detail below.

In the presence of ammonium ion, free chlorine may (5-37)

react in a step-wise manner to form chloramines. Keq = [NH/] [NHCI 2 ]/[W] [NH 2CI]2
This process is depicted by the following equations:
At 25C, K eq has a value of 6.7 x 10 5 liters/mole
NH a + HOCI = NH 2CI + H 20 + H+ (5-32) (70,79).
NH 2CI + HOCI = NHCb + H 20 (5-33)
Using these relationships, it is possible to determine
NHCb + HOCI = NCb + H 20 (5-34) the equilibrium ratio of dichloramine to monochlor-
amine as a function o'f pH and applied chlorine dose As noted above, hypochlorous acid dissociates into
ratio (assuming no dissipative reactions other than hypochlorite with a pK o'f approximately 7.5; analo-
those involving the ino,rganic chloramines). Table 5-8 gously, ammonia is able to associate with a proton to
summarizes these call:ulations. As pH decreases and the ammonium cation, with the pK for the latter of
the CI:N dose ratio inoreases, the relative amount of approximately 9.3. For a constant chlorine:nitrogen
dichloramine also increases. dose ratio, the maximum rate of monochloramine
formation occurs at a pH where the product HOCI x
Tablo 58. Ratio of DichloramIne Combined Chlorine to NH s is maximized, which is at the midpoint ofthe two
Monochloramlne Combined Chlorine as a pK values or 8.3. At this optimum pH and the usual
Function of pH and Applied Molar Dose Ratio
(Equlllbrlumi Assumed)
temperatures encountered in practice, the formation
of monochloramine attains equilibrium in less than a
Molar second; however, at either a higher or lower pH, the
CI2:N pH
Ratio 6 7 8 9 speed of the reaction slows.
.1 .13 .014 1E-03 o
.3 .389 .053 5E-03 o The formation of dichloramine from reactions be-
.5 .668 .114 .013 1E-03 tween hypochlorous acid and monochloramine obeys
.7 .992 .213 .029 3E-03 the following rate law (81):
.9 1.392 .386 .082 .011
1.1 1.924 .694 .323 .236 (5-39)
1.3 2.7 1.254 .911 .862
1.5 4.006 2.343 2.039 2.004 I' (mol/I-s) = 3.0 x 10 7 exp (201 OfT) [NH 2 CI] [HOCI]
1.7 6.875 4.972 4.698 4.669
1.9 20.485 18.278 18.028 18.002 While this reaction has been reported to be catalyzed
by protons and acetic acicl, the effectiveness of these
As the CI:N molar dOSl3 ratio increases beyond unity, catalysts is such that they would not be of importance
the amount of dichloramine relative to monochlor- at the usual concentrations encountered.
amine rapidly increases as well. For the conversion
from dichloramine to trichloramine, the equilibrium Given the situation where free chlorine is contacted
constant given at O.Ei M ionic strength and 25C with ammonia, the initial velocity of the monochlor-
indicatesthatthe amount oftrichloramine to be found amine formation process given by Equation 5-38 is
in equilibrium with di-and mono-chloramine at molar substantially greater than the velocity of the sub-
dose ratios of up to 2.0 is negligible (79). This is in line sequent formation of dichloramine given by Equation
with experimental ml3asurement of the individual 5-39. Hence, relative to equilibrium levels, there will
combined chlorine spelcies as a function of approach be an initial accumulation of monochloramine if large
to breakpoint, which indicates that detectable tri- dose ratios are used until the dichloramine formation
chloramine is not formed until shortly before, or after process can be driven. Thus in practice, the ratio of
the break point itself, and then only in relatively small dichloramine to monochloramine found may be less
amounts (80). than that' given by equilibrium analysis as in Table
These findings, coupled with the routine observation
of the breakpoint at molar doses below 2:1 (weight Additional formation reactions ofthe chloramines are
ratios CI 2:N below 10:11) indicate that trichloramine is disproportionation processes, such as Equation 5-37.
not an important species in the breakpoint reaction. However, both Morris (82), and Gray et al. (79) provide
Rather, the breakpoint reaction leading to oxidation of data to indicate that the formation of dichloramine via
ammonia nitrogen and reduction of combined chlo- monochloramine disproportionation is relatively slow
rine is initiated with the formation of di-chloramine. compared to Equation 5-38 under the usual condi-
tions encountered in wastewater disinfection.
The kinetics of formation of chloramine species have
been investigated byvarious researchers. It has been Morris has also studied the reaction of organic
found that the formation of monochloramine is a first amines to form organic monochloramines (82). The
order process in each of hypochlorous acid and rate laws for these reactions follow a similar pattern
uncharged ammonia. Solely through kinetic argu- as those for the inorganic monochloramine formation
ments, however, it is not possible to determine process, except that the rate constants are generally
whether this, or a proCl3SS involving hypochlorite ions less. In addition, the rate constants for this process
reacting with ammonium cations, is the actual correlate to the relative basicity ofthe amine reactant.
mechanism of reacticm. If the neutral species are
selected as the reactants, then the rate of formation The mechanism of the breakpoint reaction has been
of monochloramine may be given by (81): extensively studied by Morris, and more recently by
Saunier and Selleck (83). The oxidation of ammonia
(5-38) nitrogen by chlorine to gaseous nitrogen would
I' (mol/I-s) = 6.6x10 8
exp(-1510/T)[HOCI][NH s] theoretically yield a stoichiometric ratio of 1.5 moles
of chlorine (CI 2 ) consumed per mole of nitrogen several compounds. For example, Trofe et al. deter-
oxidized according to: mined that monochloramine may react with bromide
ions to form monobromamine, the chemistry of which
(5-40) will be discussed below (84). If trichloramine is
formed, as would be the case for applied chlorine
As noted previously, the observed minimum stoich- doses in excess of that required for breakpoint, it may
iometric ratio between chlorine added and ammonia decompose either directly to form nitrogen gas and
nitrogen consumed at breakpoint is 2:1, suggesting hyp,ochlorous acid or by reaction with ammonia to
that more oxidized nitrogen compounds are produced form monochloramine and dichloramine. Saguinsin
at breakpoint. Evidence (83) suggests that the prin- and Morris determined the kinetics of these process-
cipal additional oxidized product may be nitrate es, the relative importance of which are dependent
formed via: upon solution pH, trichloramine, and ammonia con-
centrations (85).
(5-41 )
NH/ + 4HOCI = N0 3 - + 4 CI- + 6 H+ + H2 0 Organic chloramines may be formed by processes
similar to inorganic chloramine formation, although
As indicated, the production of nitrate from ammonia usually at lower rates. Pure solutions of amino acids
resu Its in the consumption of four moles of CI 2 (or four and some proteins may display breakpoint curves of
moles of HOCI) per mole nitrate formed. Hence, identical shape to those of ammonia solutions (86,87).
depending upon the relative amount of nitrate formed Organicchloramines may also be formed by the direct
in comparison to nitrogen at breakpoint, between 1.5 reaction between monochloramine and the organic
and 4.0 moles of available chlorine may be required, amine, and this is apparently the most significant
which is consistent with the available data. mechanism of organic N-chloramine formation at
higher concentrations such as might exist at the point
The breakpoint reaction consists of a complex series of application of chlorine to a wastewater (88).
of elementary reactions, of which Equations 5-40 and
5-41 are the net results. On the basis of extensive Free chlorine residuals may react with other inorganic
kinetic investigations, Saunier and Selleck proposed compounds likely to be present in wastewaters. Table
that hydroxylamine (NH 20H) and NOH may be inter- 5-9 summarizes available data on rates of these
mediates in this reaction (83). A complete kinetic processes (63). These reactions are generally first
scheme for the breakpoint process as proposed by order in each of the oxidizing agent (hypochlorous
these authors is given in Figure 5-5; however, there acid or hypochlorite anion) and the reducing agent.
does not yet appear to be sufficient experience with Notable among these possible reactions are the
the use ofthis reaction mechanismto justify its use in reductions in the presence of nitrites, and sulfites.
wastewater applications. Nitrites might be present in incompletely nitrified
effluents and react via a complex, pH-dependent
At chlorine doses below breakpoint, the inorganic mechanism (89). The reaction with sulfite is used in
chloramines can decompose by direct reactions with dechlorination and will be discussed subsequently.

Figure 5-5. Proposed kinetic mechanism forthe breakpoint reaction. (83) (Reproduced from American WaterWorks Association
by permission.)

k S3 ka ,
N02 NO;;

If ,Kb <Xl 0
.:.! ::c
NH 3

k S1
T' K S2
NH 20H -_.-NOH
,)0- N2

~ If ~

11 ,Km

II ,K' NCb NH 3 0W

0. N 2


Table 6-9. Summary of Kinetics of HOCI and OCI- Reduc- ials during wastewater chlorination to residuals on
tion by Miscellaneous Reducing Agents (63) the order of 12 mg/I (92). However the kinetics ofthis
process were sufficiently slow in that the formation of
organic chlorine compounds was found to increase
Oxidizing Reducing Oxidation with time up to at least4 hours following chlorination.,
Agent Agent Product
Similar findings were noted by Chow and Roberts,
OCI- -5.04
OCI- -7.63 who determined that the production of both organic
OCI- -5.48 halogen compounds, and trihalomethanes (chloro-
OCI- 3.93 form and its analogs) specifically was greater in
HOCI 0.82 nitrified than in non-nitrified effluents (93)~ In break-
HOCI -1.38
point chlorination of effluents, a large spectrum of
HOCI 3.47
HOCI -0.55 individual organic compounds may be isolated from
HOCI 1.20 wastewater effluents at levels of 10 to 500 micro-
HOCI 8.52 grams/I (94). Bromine Demand

Free chlorine can react with organic constituents to The reaction of free bromine (hypobromous acid and
produce chlorinated organic by-products. Murphy et hypobromite ion) follows many ofthe same pathways
al. surveyed the reactivity of many classes of organic as the reactions involving hypochlorous acid and
materials, and indicated that phenols, amines, alde- hypochlorites. The key differences in chemistry
hydes, ketones and pyrrole groups are all readily between free chlorine and free bromine, as outlined
susceptible to chlorination (90). The first rigorous by LaPointe et al are as follows: (95)
study of this point Wi3S that of Granstrom and Lee,
who noted that phem:>1 could be chlorinated by free formation of bromamines is rapid in comparison
chlorine to form chlorophenols of various degrees of to chloramines;
substitution (54). The, kinetics of this process were
dependent upon both phenolate ions and hypochlo-
except for the irreversible oxidation of ammonia
rous acid. However, i1' excess ammonia was present in the breakpoint reaction, the bromamine
the formation of chlorophenols was substantially system is adequately described purely by equil-
inhibited. ibrium relationships;
In more recent work, De Laat et al. determined the
rate of reaction of a variety of organic compounds at normal pH and Br:N dose ratios, the major
with hypochlorous aciid to form chloroform (91). This form of bromamiml is dibromamine (NHBr2)
reaction, the prototype of which is the reaction of versus monochloramine; and
acetone with hypochlc1rous acid to form chloroform as
in the following, is of major interest in potable water at low pH values and relatively high Br:N dose
chlorination. ratios, tribromamine may be of significance,
whereas trichloramine almost always is insig-
(5-42) nificant.
+ 2 H2 0 . The comparative rate of formation of monobrom-
amine versus monochloramine has been studied by
Polyhydric phenols are substantially more reactive Wajon and Morris (96). These results indicate that the
than simple ketones in the production of chloroform, monobromamine is formed 14 to 20 times more
and thatthe rates ofth ese processes are typically first rapidly than monochloramine at pH 7 and equivalent
order in each ohhe phenol concentration and the free molar concentrations of free halogen and ammonia.
chlorine concentration (91). More significantly, it was
observed that the reactivity of these compounds, On this basis, it is' satisfactory to describe the
many of which are analogous to materials such as predominance of various bromamine compounds by
tannins and humic alcids that could be present in the use of an equilibrIum diagram in which pH and
effluents, is greater than the reactivity of ammonia N:Br molar dose ratio are the two master variables.
with hypochlorous acid. Therefore, even if sub-break- Such a diagram is reproduced in Figure 5-6 (97).
point chlorination is practiced, some chloroform may
be formed rapidly prior to the conversion of free to The oxidation of ammonia nitrogen by free bromine
combined chlorine. residuals during the break point process follows the
analogous reaction as in the case of chlorine. As
Experimentally, this hi3S been verified by Jolley who described by Equation 5-43,1.5 moles of hypobromous
noted that more than 1 percent of the applied chlorine acid are consumed by the reaction with one mole of
dose may be converted to chlorinated organic mater- ammonia. This corresponds to the equivalent of 1.5
Figure 6-6. Distribution of bromamine species as a function Comparatively little information exists on the forma-
of pH and N:Br molar dose ratio (97). (Repro- tion of organobromine compounds during wastewater
duced by permission, American Society of Civil
bromination. However, it is now well known from
water treatment practice that transient free bromine
may react to form bromoform and mixtures of
-2 - chlorinated and brominated compounds, and itwould
o ...
.-~:J HOBr
be prudent to anticipate that these reactions would
a: 0.-1 -
_E also occur during wastewater bromination.
o ..
E 00 0 _
... to Chlorine Dioxide Demand
:g, J:
+1 The reaction of chlorine dioxide with material present
...J in effluents resulting in chlorine dioxide demand
+2 - 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
2 3 4 5 6 7 8 9 10 direct reactions with water and interconversions to
pH chlorite and chloride, as outlined above.
moles of molecular bromine per mole ammonia
In' one of the only studies directly applicable to
nitrogen (97).
wastewater, Roberts et aL (100) and Chow and
(5-43) Roberts (93) determined the chlorine dioxide demand
2NH/ + 3HOBr = N 2 + 3Br- + 3H 20 + 5H+ of wastewater effluents and the formation of halo-
,genated organics and trihalomethanes. It was found
There do not appear to have been any detai led studies that at equal mass doses (Le. 10 mg/I chlorine versus
on the mechanism of this reaction in real waste- 10 mg/I chlorine dioxide), the chlorine dioxide demand
'waters, or the production of more oxidized nitrogen of a wastewater (expressed as electron equivalents)
forms, such as nitrate, which the reports noted above was less than the chlorine demand of that waste-
indicate may be produced during breakpoint chlo- water. In addition, notrihalomethanes were formed
rination. It is reasonable to suspect that the amount of from the application of chlorine-free chlorine dioxide
bromine consumed in the breakpoint reaction would (prepared via the acid-chlorite process, rather than
be somewhat greater than that indicated by Equation the chlorine-chlorite process), and the amount of
5-43 due to the possible existence of these other halogenated organics produced was 10 to 20 times
processes. less than those produced by equal mass doses of,
One factor that must be mentioned is that combined
bromine residuals (inorganic bromamines) possess
disinfection efficiencies comparable to free bromine It has been established that, at usual concentrations,
residuals. Thus, while the ammonia-bromine reaa- ammonia nitrogen, peptone, urea, and glucose have
tions do exist and do occur in wastewater disinfection, insignificant chlorine dioxide demand in 1 hour
their significance in governing process performance (101,102).
is somewhat less than the comparable ammonia-
chlorine reactions. Masschelein reviewed the reactions of various
classes of organic materials with chlorine dioxide,
The kinetics of decomposition of bromamine residuals and only the following would appear to be of
have been studied in some detail in pure laboratory significance to wastewater applications (51):
systems (98,99). Both of these studies suggest a
mechanism of the bromine breakpoint process that 1. CI02 may oxidize tertiary amines to secondary
may involve the elementary reaction between di- amines and aldehydes, with the formation of
bromamine and tribromamine acting via a series of chlorite. For trimethylamine, this reaction is first
complex intermediates. One consequence of this order in chlorine dioxide and the amine, and has
process appears to be that high ammonia concentra- a rate constant of approximately 100,000 Iiters/
tions may reduce the rate of the breakpoint reaction mol-s.
due to a reduction in the amount of tribromamine
present (98). In addition, the decomposition of 2. Ketones, aldehydes, and alcohols may be oxi-
dibromamine may lead to the reappearance of free dized to acids. While Masschelein cites this as a
bromine (which could, in turn, form additional significant reaction, Somsen indicates that the
amountsof combined bromine) via thefollowing (98): rate of ethanol reaction with chlorine dioxide is
(5-44) insignificant at pH 7, while carbonyl compounds
NHBR 2 + NBr3 + 2H 20 = N2 + 3H+ + 3Br- + 2HOBr do exhibit reactivity (103).

3. Phenols and phenol derivatives may react with air of 2.26 (106). Sulfur dioxide gas is supplied in
chlorine dioxide to form oxidized and chlori- cylinders similar to those used for the supply of
nated products. gaseous chlorine.

4. Sulfhydryl amino acids, particularly cystine, Table 5-10 summarizes physical properties of sulfur
may be oxidized to cysteic acid. dioxide at various temperatures.

Wajon et al. have nacently studied the reaction of

phenols with chlorinc3 dioxide under dilute conditions Table 5-10. Physical Properties of Sulfur Dioxide (106)
typical of water and wastewater disinfection applica- Liquid Vapor
tions (104). The reaction's stoichiometry is 2 moles of Temperature Density Pressure Solubility
chlorine dioxide consumed per mole of phenol (or (OC) (g/mll (atm) (gil)
hydroquinone) consumed. Products formed included -40 1.530
chlorophenols, aliphatic organic acids, benzonqui- -29 1.510
none, and (in the case of phenol) hydroquinone. The -18 1.480
-6 1.450 0.15
mechanism appeared to include the possible forma- 4 1.420 0.81 199.8
tion of hypochlorous acid as an intermediate which 16 1.390 1.74 130.5
would thus chlorinate, and the rate of this process 27 1.360 3.83 87.6
was found to obey the following law:

(5-45) Upon .addition to water, sulfur dioxide forms sulfur-

r = 2(kb + kaKa/[H~) [Phenol] [CIOil (mol/I-s) ous acid (H 2 SO s), which may lose, successively, two
protons, according to the following reactions:
Where Ka is the acidity constant for the phenol, and kb
and k a are the rate constants. For phenol, these were
0.24 and 2.4 x 107 I/mol-s, respectively, while for (5-46)
hydroquinone, these rate constants were respectively
39,000 and 6.5 x 109 I/mol-s. (5-47)

5.3.4 Dechlorination Chemistry The pK for the second dissociation (Equation 5-47) is
In certain circumstances, it is desirable to reduce the 7.2 at room temperature (107), and hence sulfur (IV)-
chlorine residual in a disinfected wastewater prior to is present in most wastewater effluents as a mixture
discharge. In order to dothis, it is necessary to contact of both sulfite and bisulfite ions. In some circum-
the chlorinated wastewater with a substance that stances, sulfur (IV) may be applied as sodium sulfite
reacts with, or accelerates the rate of decomposition or bisulfite, and the resulting equilibrium mixture is
of, the residual chlorine. While many compounds may established via the above equations; however, the
perform this function, including thiosulfate, hydrogen form of application of sulfur (IV) does not, in and of
peroxide, ammonia, sulfite/bisulfite/sulfur dioxide, itself, affect the nature of the subsequent dechlori-
and activated carbon, only the latter two materials nation reactions.
have been widely used for this purpose in either
water or wastewater treatment (1 05). In this section, The reaction between sulfur dioxide and free chlorine
the chemistry of the reaction between residual is such that one mole of sulfur dioxide reacts with one
chlorine and the latter two substances will be mole of either free chlorine or monochloramine via
discussed. the following stoichiometric equations (105): 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 combint3d chlorine residuals to form
oxidized sulfur products, principally. sulfate, and
hence convert available chlorine to chloride. The rate of reaction between free chlorine and sulfur
(IV) compounds appears to be quite rapid. This rate
Commercially, S(lV) used for dechlorination is sup- has been found to be first order in each of sulfite and
plied principally as sulfur dioxide under pressure as a free chlorine concentration. However, Lister and
gas in equilibrium with its liquid. Sulfur dioxide is an Rosenblum (108) have indicated that the kinetics are.
irritanf gas, with a waiter solubility 20 times greater dependent on OCr- concentration, while Srivastava et
than chlorine gas, and with a gas density relative to af. (109) proposed that HOCI is the reactive speCies.
The rate laws given by these investigations are, to which catalysis of Equation 5-52 might occur in
respectively: dechlorinated effluents has not been thoroughly
r = 2.7 x 10 9 exp (-3773/T) (5-50) Activated Carbon Dechlorination
x [OCI-] [S03- 2 ] (M/I-s)
Activated carbonis a material prepared by controlled
r = 1.2 x 10 15 exp(-7851/T) (5-51 ) combustion and oxidation of an organic material. This
section will describe the chemical properties of
x [HOCI ] [S03- 2 ] (M/I-s) activated carbon peculiar to dechlorination applica-
tions. However, it should be noted that the use of
At 25C and 0.2 mM sulfite concentrations (16 mg/l activated carbon in wastewater treatment may sim ul-
as sulfite), these rate laws give half-lives of 0.4 and taneously achieve other objectives, such as the
0.8 s, respectively. Thus, it appears, for free chlorine, reduction of total organic carbon to low levels, or the
that the rate of reaction is sufficiently rapid so as to be specific removal of biologically refractory trace pol-
regarded as instantaneous (80). lutants.

Surprisingly, there does not appear to have been any The reduction of free chlorine residuals by activated
direct measurements made of the rate of reaction of carbon at low loadings has been found to be due to an
sulfur (IV) species with inorganic combined chlori~e. actual reaction of the chlorine with the carbon
However, based on the results reported for organic surface, rather than merely by a catalytic decomposi-
chloramines, it may be expected that the dechlorina- tion (113). Products of this reaction include chloride
tion of combined chlorine residuals is slightly slower and surface-bound oxides of carbon that evolve as
than free chlorine. Stanbro and Lenkevich reported carbon monoxide and carbon dioxide, as well as
that the sulfite dechlorination of organic chloramine water, when the spent carbon is thermally regener-
residuals follow kinetics first order in both total ated. Hence, the dechlorination process using acti-
reduced sulfur and chloramil:te concentrations (11 0). vated carbon represents a consumptive use of the
For the monochloramine derivatives or methylamine, activated carbon itself.
N-alpha-acetylysine, alanine, leucine, and alanyl-
At higher loadings, once the surface of the carbon has
alanylalanine(N-chloro derivative), at pH 7 and 25C,
accumulated surface oxides (denoted by CO*), reduc-
the measured second order rate constants were
tion of free chlorine residuals may continue to occur.
4,333,2,167,1,167,2,333, and 83 liters/mol-s. For
This has been postulated to be due to the catalysis of
the peptide chloro derivative, in particular, this is suf-
free chlorine decomposition to chloride and chloric
ficiently slow to suggest that small amounts of some
acid according to (114):
organochloramines may not be dechlorinated com-
pletely in some sulfur (IV) processes, particularly 3HOCI = 2HCI + HCI03 (5-53)
where extremely short contact times are used.
The reaction of combined chlorine residuals with the
One side reaction that may occur during sulfur carbon surface is more complex, and differs between
dioxide dechlorination, and that may be of concern for mono and dichloramine. For fresh carbon, before a
the receiving water quality, is the deoxygenation of large amount of surface oxides have formed, mono-
the effluent. This occurs by the following reaction chloramine is reduced by the carbon surface (denoted
stoichiometry (105): by C*) to form surface oxides (which are converted to
carbon monoxide and dioxide when thermally regen-
02 + 2S0 2 + 2H 20 = 4H+ + 2S0 4 - 2 (5-52)
erated) according to (115):
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,
NH 2CI + H 20 + C* = NH/ + CI- + CO*
and manganese, and is inhibited by ethanol, glycerol After some surface ~oxides have formed, they may:
and mannitol. In uncatalyzed systems, at sulfur (IV) react with monochloramine by the additional reaction
concentrations below 0.02 M it has been found to be (115 ):
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, . (5-55)
and increasing rates as pH increases (105). Thus, 2NH 2 CI + CO* = N2 + 2W + 2CI- + H20 + C*
unless catalyzed, the rate of deoxygenation is much
less than the rate of sulfur dioxide dechlorination. In Reaction 5-55 serves to reduce the concentration of
pilot studies, it has been noted that no significant ammonia nitrogen present in a wastewater, and this
oxygen depletion occurred until sul"fur dioxide over- process has been suggested as a means for the
doses exceeded 50 mg/l (112). However, the degree removal of ammonia nitrogen from effluents.

Dichloramine residuals appear to react primarily by chlorine. Each of the latter two fractions may be
the formation of nitro~len gas and surface oxides by further subdivided; free chlorine into HOCI and OCI-,
(115): and combined chlorine into mono-, di-, and tri-chlor-
amine (as well as organic chloramine compounds). In
(5-56) most cases, however, the concentration of total
2NHCI2 + H 20 + C = N 2 + 4H+ + 4CI- + CO residual chlorine is sufficient to allow adequate
process monitoring.
The kinetics of monochloramine-carbon dechlorina-
tion occurring by Equations 5-54 and 5-55 have been All currently available methods for the analysis of
studied in detail (116). Since the rate of monochlor- total residual chlorine rely upon the oxidation of
amine decomposition ii, accelerated by the formation iodide to form molecular iodine, which then can be
of surface oxides, the overall process is autocatalytic, analyzed. The chemistry 'of this process, which is
and for low effluent monochloramine residuals, favored at low pH, is as follows:
prolonged contact times are necessary-for example,
for a 2 mg/I monochloramine residual, 16 minutes HOCI + H+ + 21- = b + CI- + H 2 0 (5-58)
contact in a packed beel carbon contactor are neces-
sary to reduce the r~sielual below 0.3 mg/1. OCI- + 2W + 21- = 12 + CI- + H2 0 (5-59)

The reaction between dichloramine and carbon is NH 2 CI + 2H+ + 21- = NH/ + b + CI- (5-60)
faster than that between free chlorine and carbon,
NHCI 2 +3W + 41- = NH 4 + 21 2 + 2CI- (5-61 )
and much faster than that between monochloramine
and carbon. In kinetic: investigations, Kim et aI., NCb + 4H+ + 61- = NH/ + 31 2 + 3cr- (5-62)
(117,118) determined that the reduction of dichlor-
amine occurred by Equation 5-56 in parallel with a At a pH of4.0, potassium iodide will react with all of
catalytic breakpoint reaction, given by: the above forms of chlorine residual to produce
NH/ + 3NHCI 2 = 2N 2 + 6CI- + 7W (5-57)
At this point, the analysis can be conducted in two
Hence, dechlorination of predominantly dichloramine general manners. First, in the so-called forward
residuals will also produce significant degrees of methods, the released iodine may be directly meas-
ammonia nitrogen removal. ured by titration using either a starch colorimetric
Since, as noted above, in sub-breakpoint chlorination, indicator or amperometricmeasurements for end-
monochloramine predominates, it would appear that point determination. In the reverse methods, simul-
the design of activated ci3rbon beds for dechlorination taneously with the production of iodine in Equations
would be limited by the reduction of monochloramine 5-58 to 5-62, reductant is added. By measuring the
residuals, which are thl3 more slowly reacting form, amount of this reductant that has been oxidized
particularly when final I::hlorine residuals below 0.1 during the analysis, the amount of total residual
mg/I were desired. chlorine may be determined.

Only a small number of wastewater treatment plants The two most common forward titration methods
practice GAC dechlorination. Other than the models involve the use of starch indicator in conjunction with
of Kim et al. (118) and Kim and Snoeyink (116), little sodium thiosulfate (Na 2S203) titrant and the use of
guidance for design is available. amperometric titration equipment in conjunction with
phenylarsine oxide titrant.
5.4 Analysis of Disinfectant Residuals In the first case, sodium thiosulfate is added to the
The monitoring and control of halogenation process- sample until the reddish iodine color is almost
es requires that a means be available for the analysis discharged. At this point, starch indicator is added,
of residual disinfectant, which, as previously noted, is which forms a strong bluish complex with iodine, and
one major variable governing microbial inactivation. sufficient additional titrant is added until the color
Ideally, the analytical method should be simple, disappears. The amount of sodium thiosulfate that
rapidly performed, and not subject to analytical has been added is equivalent to the amount of iodine
interferences by other constituents likely to be that had been released, and this in turn is equivalent
present in the wastewatl3r. to the total residual chlorine.

5.4.1 Chlorine The second procedure substitutes the amperometric

Chlorine may be analyzed as total residual, including titrator for the starch indicator solution (119). This
all chemical forms capable offunctioning as oxidizing instrument measures the current developed across
agents (free chlorine plus inorganic chloramines), or electrodes maintained at a constant potential suf-
it may be differentiated as free chlorine and combined ficient to reduce free halogen residuals. In the forward

titration, phenylarsine oxide (PAO) titrant is added partof the wastewater. In the reverse titrations, any
until there is no further change in the indicated iodine produced is rapidly removed by reaction with
current; this is the point at which all halogen residual the reducing reagent (thiosulfate, PAO, LCV or DPD),
has been reduced, and the amount of PAO added is .and thus there is little opportunity for demand
equivalent to the total residual chlorine. exertion.

In the reverse titration, the amount of consumed Experimental studies comparing these methods,
reductant (either, thiosulfate or PAO) is measured by however, have tended to yield disparate results,
titration with standard iodine solution using either which appear to vary with the quality of the waste-
the starch indicator or amperometric end-point water being chlorinated. Browning and McLaren
detection methods. Alternatively, the reductant used (127) developed a modified starch-iodide forward
may be a compound that changes color with oxidation, titration (MSI), which was found to yield results
and thus the amount of residual chlorine may be closely comparable to the reverse amperometric
determined either colorimetrically or by titrating the method and greater than the forward amperometric
color with reductant until disappearance. This latter method, and reverse starch-iodide methods. In con-
procedure is exemplified by the use of DPD (N,N- trast, Collins and Deaner (128) found that the MSI
diethyl-p-phenylenediamine) or Leuco Crystal Violet procedure consistently produced lower chlorine
(LCV). residual readings than did the reverse starch-iodide
method, and that the extent of this difference
In the DPD procedure, DPD reagent is added prior to
increased with the degree of pollution ofthe effluent
iodine addition in the reverse titration. Oxidized DPD (e.g. BOD).
is reddish in color, whereas DPD itself is colorless;
thus the amount of residual chlorine may be deter- Lin et al. compared the reverse starch iodide titration
mined by a spectrophotometric determination of the (SIB), the DPD method, the LCV method and the MSI
amount of oxidized DPD produced. This method is titration using three wastewaters. These authors
widely available in a field kit form. Alternatively, the concluded that the MSI procedure generally, but not
oxidized DPD may be titrated with ferrous ammonium always, yielded results less than SIB and LCV
sulfate reductant. methods, but that the results were indistinguishable
from the DPD procedure. Furthermore, following
The LCV method is a colorimetric assay based upon chlorine addition, in some cases, the total chlorine
the vivid bluish color exhibited by oxidized LCV. residual measured by the DPD and LCV procedures
increased with time; the MSI procedure did not
For each of the above procedures, detailed methods produce such anomalous results. On this basis, Lin et
may be found in Standard Methods (111). All colori- al. suggested that the MSI procedure was preferable
metric and visual titrimetric methods are subject to to either DPD or LCV methods (129).
interferences in highly colored or turbid wastewaters.
Oxidized manganese and chromium, as well as other
Recent data suggest that all of the methods used for
materials capable of converting iodide to iodine,
total chlorine residual may produce false negative
interfere with all ofthe above methods. Silver or Cu(l)
readings if sulfite or similar reducing agents are
ion interfere with the amperometric methods.
present(110). This situation mayoccurwhendechlor-
In previous years, orthotolidine reagent had also been ination using S(IV) compounds is practiced, partic-
used for total chlorine determinations in wastewater. ularly if organic chloramines are also present. In this
However, a number of studies have indicated that the case, it is possible that total chlorine residuals actually
OT method is substantially less satisfactory than exist in the effluent, but that when the pH is lowered
either the amperometric or starch iodide titration during the initial addition of iodide, the reaction
methods (120-126). In addition, orthotolidine is a between S(IV) and chlorine residuals may be more
suspect carcinogen and therefore, on the basis of rapid than the reaction with iodide. Thus, chlorine
these two reasons, this method is no longer approved may be deemed to be absent when it is in fact present.
(111 ).
Two additional methods for the analysis of total
The amperometric and starch-iodide forward titration residual chlorine in wastewater have been reported;
methods for total chlorine residuals have been found however, neither of them has yet appeared in
to produce comparable results (122). For waste- Standard Methods. Jenkins and Baird (130) utilized a
waters, it is recommended that the total chlorine polarographic iodine electrode to measure the iodine
reverse titrations be used. In the forward methods, released following acid-iodide addition to waste-
free iodine residuals are in contact with the waste- water. This method is essentially a forward titration
water for a period of time during the completion of the procedure using a direct electrode analysis of iodine
titration. During this time, there may be a . loss of production. It was found that this m.ethod was equal
halogen residual due to exertion of demand on the to the SIB method in precision and accuracy at total

chlorine residuals above 1 mg/l; below 1 mg/l, the In addition to the above procedures, Knechtel et al.
iodine electrode procedure was more precise than the (133) determined that acid chrome violet K (ACVK)
SIB procedure, which is not recommended for reagent is suitable for the direct colorimetric deter-
residuals below 1 rng/l (111). The principal inter- mination of chlorine dioxide in wastewaters. ACVK is
ference with the elec:trode iodometric assay occurred not subject to interference by free or combined
in effluents with more than 50 mg/l BOD-in these chlorine, chlorite, or nitrite, and results in waste-
cases, the iodine progressively declined to zero prior water were found to be comparable to determinations
to a stable electrode reading, presumably due to using electron spin resonance., However, if large
exertion of iodine demand. ' amounts of turbidity are present, a pretreatment via
centrifugation is necessary to prevent optical inter-
Liebermann et al. (131) investigated the use of the ference.
syringaldizine reagent for total chlorine determina-
tion following acid-KI addition, with a colorimetric Similarly, Wheeler et al. (134) proposed chlorophenol
assay. This test yielded a linear response to 10 mg/l red (CPR) as a selective indicator for chlorine dioxide;
total chlorine, and was found to be comparable to the this compound is suitable for use in a direct colori~
forward amperometric method. Syringaldizine forms metric or a forward titrimetric assay. No interferences
the basis of the FACTS test for free chlorine which have been reported by chlorite, chlorate, free chlorine,
has been accorded tentative status in the current Cr(VI), Fe(III), or Mn(V).
edition of Standard Methods (111 ).
In certain cases, particularly in breakpoint chlorina- 5.5 Kinetics of Microbial Inactivation
tion for nitrogen rElmoval, or the chlorination of The information needed for the design of a disinfec-
nitrified effluents, it may be desirable to measure free tion system includes knowledge of the rate of
chlorine apart from total chlorine. The iodometric inactivation of the target, or indicator, organism(s) by
starch-iodide, amperometric, DPD, and LCV methods the disinfectant. In particular, the effect of disinfect-
may be modified to be relatively selective for free ant concentration on the rate of this process will
chlorine. Further details on these methods may be determine the most efficient combination of contact
obtained in Standard Methods (111). It should be time (Le. basin volume at a given design flow rate) and
noted that the determination of small concentrations dose to use.
of free chlorine in the presence of large amounts of
total chlorine (as might be found immediately prior to Chick's Law and deviations therefrom are discussed
the breakpoint itself) is an extremely difficult ana- in Section 4-1, and the reader is referred to that
lytical problem (132). section for the historical development of disinfection
theory. The discussion that follows here is a practical
5.4.2 Bromine and logical outgrowth of that theory.
The analysis of bmmine residuals, as might be In the disinfection of a wastewater by halogens the
produced from BrCI treatment, proceeds directly concentration of disinfectant changes with time, and,
analogous to chlorine residuals. Free or combined particularly during the initial moments of contact with
bromine can oxidize iodide to iodine, and any of the either chlorine or bromine, the chemical form of the
methods for total chlorine can be used to measure halogen undergoes a rapid transformation from the
total bromine. If mixtures of chlorine and bromine free to the combined forms. Since C in Equation 4-6
residuals are present, it becomes impossible to would thus not be a constant, typically wastewater
distinguish between the two by iodoll)etric methods. disinfection results obtained in batch systems exhibit
"tailing," the degree of which may depend on the
5.4.3 Chlorine Dioxide chlorine demand and the ammonia concentration of
For the determination of chlorine dioxide in waste- the system (135).lt is more critical to determine what
waters, it is particularly necessary to differentiate this the chlorine residual is, rather than the'chlorine dose
compound from chlorite (which is produced as a in these systems. According to Heukelekian and
reaction product) and free and combined chlorine Smith: " ...the control of chlorination cannot be based
(which may arise if the chlorine-chlorite process is on the dosage of chlorine, because neither for sewage
used to produce chlorine dioxide). Both the DPD and from different sources, nor for a sewage from a given
the amperometric methods may' be modified to source, but at different times, does a constant dosage
determine chlorine dioxide apart from total chlorine of chlorine produce a constant number of coliform
(111); ihhe total chlorine methods are used, they will organisms per ml in effluent." (136)
tend to indicate the sum of total chlorine, chlorine
dioxide, 'and a portion of the chlorite. Roberts et al. As a result, the analysis of the kinetics of microbial
(100) indicated that, 2lS for total chlorine, use of the inactivation in wastewater effluents by disinfectants
reverse titration procedures produces results in has often been accomplished by empirical models
excess of the forward titration methods. deviating substantially from Chick's Law. For example,
McKee et aI., chlorinated primary effluent, and found While it is recognized that the presence of solids in
that his data could be described by (70),: wastewater effluents may harbor and protect en-
meshed microorganisms from the action of disinfect-
log (N) = 1/(a + bC o) (5-63) ants, few methods are available for the quantitative
consideration of these phenomena.
where N is the viable microorganism concentration
However, it has been shown that organic solid
and. Co is the chlorine dose; in this equation, the
materials present in actual wastewaters can render
contact time is held constant. McKee at al. also
enmeshed microorganisms some protection from a
proposed the following equation as a means to
measurable chlorine. residual (139,140).
extrapolate inactivation results from one contact ti me
to another (70): Analysis of data on the inactivation of a variety of
microorganisms by .free and combined chlorine
Nt/N a = (tla)n' suggests that the combined Chick-Watson law(Equa-
. tion 4-4 substituted into Equation 4-3) provides a
In Equation 5-64, Nt and Na are the microorganisms satisfact,ory description of the disinfection process.
remaining at time t and tim~ a, respectively (chlorine Values for k' and n are a function of the micro-
dose fixed), and n' is a constant. organism, the pH, and the temperature of the system,
and also differ depending upon whether free or
The following equation, similar in concept to McKee's combined chlorine species are used as disinfectants.
model, was developed to describe inactivation in Table 5-11 ,summarizes a number of values of the
various chlorinated effluents (137,138): Chick-Watson parameters for microbial inactivation.
In general, microbial inactivation increases with a
NINo = (bCtra (5-65) decrease in pH for either free or combined chlorine
residuals. An.increase in temperature also increases
In Equation 5-65, a and b are constants, and C is the inactivation. At a given temperature and pH, free
chlorine residual remaining at time t. At values ofthe chlorine residuals are more effective than combined
product (Ct)< 1/b, by definition, NINo equals unity. chlorine residuals. These data may underestimate
The use of this equation will be further discussed the resistance of microorganisms in wastewater.
below. However, the degree of underestimation is unknown.

Table 5-11. Chick-Watson Parameters for Microbial Inactivation by Chlorine

Temp. k'
Reference . Organism pH (OC) I nmg- n min- 1 n
141 E. coli 8.5 2-5 10.9 1.2
9.8 2-5 1.18 1.29
10.7 2-5 0.279 1.11
8.5 20-25 30.6 1.46
9.8 20-25 5.91 1.34
10.7 20-20 1.30 0.79
Aerobacter 7 20-25 1.39x104 3.78
Pseudomonas 8.5 20-25 312 2.74
Pyocyanea 9.8 20-25 2.13 1.26
10.7 20-25 0.738 0.711
Salmonella 7 2-5 7.87X10 12 7.26
typhi 9.8 2-5 0.962 0.76
7 20-25 8.15x 106 4.07
8.5 20-25 2.45X10 4 1.78
Shigella 7 20-25 9.07x107 4.92

142 E. coli 7 25 4.02 .801

Micrococcus 7 25 3.32 1.10
pyogenes var.

143 Bacillus metiens 10 20 .00577 .483

6 20 .029 1.24
7 20 .0219 1.18
8 20 .0209 1.12
9 20 .008 0.99

Table 5-11. Continued
Temp. k'
Reference Organism pH lOG) l"mg-"min- 1 n
9.35 20 .0086 1.04
12.86 20 .0015 .58
10 30 .00324 .868
10 35 .0044 1.0
10 50 .0075 1.26

144 Poliovirus 6 2 11.28 .766 .

Type I 6 10 12.78 .818
(Mahoney) 6 20 30.12 .615
6 30 75.12 .608

145 Poliovirus 6 20 12.20 0.69

Type I 10 20 6.5 3.23

146 Naegleria 7.2 25 .171 .786


Mycobacterium 7 20 .226 .545
phlei 7 20 6.21 x10- 4 3.51
M. fortuitum 7 5 5.68x1Q-4 2.68
Candida 7 20 6.85X10- 3 2.38
parapsilosis 7 5 .656 1.16
E. coli 7 20 1.01 0

148 E. coli 7 35 .084 1.39

8.5 35 ,0109 1.52
9.5 35 2.48x1O- 5 13.3
6.5 20-25 .483 1.07
7.0 20-25 .316 1.04
7.8 20-25 .193 1.18
8.5 20-25 .0854 1.125
9.5 20-25 .049 1.37
10.5 20-25 .0125 2.27
Aerobacter 6.5 20-25 .363 1.19
Aerogenes 7 20-25 .241 1.35
7.8 20-25 .095 1.18
8.5 20-25 .0715 .917
9.5 20-25 .0358 1.16
10.5 20-25 .00809 1.7
Shigella 6.5 20-25 .821 1.3
dysenteriae 7.0 20-25 .55 1.15
7.8 20-25 .341 1.32
8.5 20-25 .151 1.02
9.5 20-25 .064 .995
10.5 20-25 .0301 1.52

Salmonella 7 2-6 .0902 1.32

typhi 8.5 2-6 .0182 1.67
9.5 2-6 6.8X1Q-4 6.26
6.5 20-25 .491 1.13
7.0 20-25 .290 1.84
7.8 20-25 .211 1.07
8.5 20-25 .113 1.16
9.5 20-25 .0417 .878

Pseudomonas 6.5 2.0-25 .44 1.27

pyocyanus 7.0 20-25 .301 1.44
7.8 20-25 .174 1.55
8.5 20-25 .102 1.01
9.5 20-25 .0483 1.05

The use of the parameters in Table 5-11 for the entails the assumption that the microorganisms for
description of waste!water microbial inactivation which k' and n have been measured are identical (in

terms of sensitivity towards disinfectant) to those 5.6 Proc:ess Options
likely to be present in wastewater to be disinfected. Disinfection systems using chlorine, chlorine dioxide,
Unfortunately, all of the data summarized in Table or bromine chloride, with or without dechlorination
5-11 are based upon the use of pure, laboratory share certain common elements. For each of these
strains of microorganisms. These may have different elements, several possible options exist for the
sensitivities as compared with indigenous waste- system designer to consider. By specifying the
water organisms. potential options, the many available halogenation
disinfection systems may be enumerated.
As an alternative approach, the model of Collins et al.
(Equation 5-65) has been used to describe inactiva- In this section, these elements and options will be
tion of coliforms indigenous to wastewater. Table 5- summarized. In the following section, the method-
12 summarizes values for the a and b parameters in ology for detailed sizing and specification of each
Equation 5-65 which have been determined for the option element will be discussed.
inactivation of coliforms indigenous to wastewaters.
The general block diagram of a halogen disinfection
Table 5-12. Parameters in the Collins et al. Model Describing system, with optional dechlorination, is shown in
Wastewater Coliform Inactivation by Chlorine Figure 5-7. In general, the disinfection system
(177) contains four elements and the dechlorination system
b contains three elements. Flows of information are
System I/mg-min a depicted by dashed lines and flows of material are
Palo Alto-1978 .193 3.15 depicted by solid lines.
Palo Alto-1979-Unfiltered .452 2.22
Palo Alto-1979-Lab Filtered 1.190 2.10
Dublin-San Ramon Lab 1.54 1.62
The disinfection subsystem elements are chemical
Dublin-San Ramon Field 0.598 1.79 generation and/or supply, mixing, contacting and
San Jose 0.246 2.82 control. In the generation/supply element the dis-
Note - Palo Alto and San Jose are non-nitrified secondary efflu- infecting agent is maintained and fed continuously. In
ents. Dublin is nitrified activated sludge. the mixing element, the incoming wastewater is
blended with the disinfectant. In the contacting
Use of the Chick-Watson and the Collins et al. models element, the blended wastewater is held for a
for the estimation of microbial inactivation in waste- sufficient period until the desired microbial inactiva-
water chlorination are described in Section 5.7.3. tion has been attained. The control element uses

Figure 5-7. Elements of halogen disinfection systems. with optional dechlorination.

,------------, ~~~~~--~----~~
.... - -

Control '::

Dech lorination
~__ BI'! Control

1 !
i: /// !-j
Influent ! II. II I Effluent
Mixing Contacting
Mixing I
1 '

..--... Hydraulic Flow
- -... Information Flow
L Dechlorination Subsystem
J. I

information about incoming wastewater flow and economics of chemical supply. In these systems, the
residual disinfectant to adjust the operating param- chemical may be fed directly as a solid (proprietary
eters of the generation and supply element so as to tablet feed chlorinators) or may be prepared as a
maintain consistent disinfectant performance. slurry/solution and fed in a similar manner as sodium
hypochlorite. Solution feed of calcium hypochlorite
In the dechlorination subsystem. the dechlorination has inherent disadvantages in that the addition of
chemical supply element functions to provide a feed. water to calcium hypochlorite will produce a calcium
The mixing elements and the control elements have carbonate sludge that will foul surfaces and storage
identical functions as in the case of the disinfection tanks.
subsystem. In dechlorination, there is generally no
substantial contacting element (other than, perhaps, The choice among gaseous chlorine/sodium hypo-
the final outfall pipe). chlorite/calcium hypochlorite is governed primarily
by economic considerations in conjunction with
For the halogen generation/supply subsystem, a inherent safety and handling hazards with the use of
number of choices exist as to the option to be gaseous chlorine. In wastewater treatment plants
selected. These may be t3numerated as follows: serving large municipalities, it may be advantageous
to use sodium hypochlorite, despite higher unit costs,
I. Choice of chemical to minimize risks associated with the transport of
A. Chlorine Gas liquid chlorine through populated urban areas.
1. Method of Withdrawal
(a) Gas In chlorine dioxide systems, the primary design choice
(b) Liquid/oxternal evaporation in the generation/ supply subsystem is the type ofthe
2. Method of feeding generation process. As discussed above, the utiliza-
(a) Solution tion of chlorite is greater in the chlorine/chlorite
(b) Direct Gas process than in the acid-chlorite process. However, in
B. Sodium Hypochli:>rite the former process, the possibility for the production
1. Commercial supply of chlorine dioxide containing chlorine at low levels
2. Onsite generation exists, and thus some of the deleterious aspects of
C. Calcium Hypochlorite chlorine residuals and byproduct reactions may exist.
1 Solid Feed
2. Solution Feed For BrCI systems, liquid withdrawal with external
D. Chlorine Dioxide evaporation is necessary to" prevent chemical dissoci-
1. Method of G,eneration ation and enrichment of gaseous chlorine, which
(a) ChlorinEt-chlorite process would occur in gas phase withdrawal. Solution feed
(b) Acid-chlorite process systems for BrCI plants are universally used.
E. Bromine Chloridl3
In the mixing subsystem, the primary design options
are the use of devices (static mixers or hydraulic
For chlorine gas, the choic:e of gas or liquid withdrawal jumps) that dissipate hydraulic head and the use of
is dictated primarily by the size of the disinfection mechanical mixers (jet mixers, impeller mixers, etc.)
system to be designed. It is impractical to use gas that require external motive power such as pumps or
withdrawal in very large systems due to limitations on motors. In plants where there is little available head
safe withdrawal of galS from ton containers of between the influent and effluent (such as may exist
chlorine. This decision between solution and direct in wastewater treatment plants serving municipal-
gas feed processes is alsl:> based upon the size of the ities with relatively flat topography), the use of the
plant; in all but very small wastewater treatment former processes may be infeasible. In general,
plants (less than 1 MGD), or if chlorine is only to be however, the former processes offer an advantage'in
used on an occasional basis, the solution feed process that their performance is insensitive to any power
Is generally used. fluctuations or outages that may occur, and thus they
are inherently more reliable.
When sodium hypochlorite is to be used as the
disinfecting agent. the chemical may be purchased in The contacting subsystem may consist of a separate
bulk as a liquid, or may be produced continuously contact basin or the outfall pipe itself. The latter
onsite by electrolysis of brine. The latter process option reduces capital costs; however, it necessitates
becomes feasible at remote locations. or at waste- the existence of a sufficiently long outfall pipe, which
water treatment plants where inexpensive sources of may not exist at all plants.
brine are available (i.e., plants on ocean coast lines).
The control subsystem serves to produce an effluent
The use of calcium hypochlorite is confined primarily in which a consistent effluent quality is attained at a
to smaller wastewater treatment plants due to minimum chemical dose. Three options exist: manual,
or "open-loop" control (relying upon operator inter- effluent and the receiving stream, with undesirable
vention); simple flow-proportional feed-forward con- results. Hence, in some plants, even where simpler
trol, and true closed-loop feedback control (relying control systems for chlorination are used, flow-
upon continuous sensing of halogen residual, usually proportional, or feedback control for dechlorination
in conjunction with feed-forward flow proportional may be necessary to minimize the possibility for
control). The last option provides minimum chemical deoxygenation in the effluent and the receiving
utilization, but is more capital intensive. In very small stream.
plants, where dechlorination is not practiced, manual,
or feed-forward control is sufficient. However, in 5.7 Design Coordination
plants subject to effluent constraints on chlorine As previously indicated, an overall halogen disin-
residual, or in very large plants where small decreases fection system may be subdivided into several
in chemical dosage result in large economic savings, components. In this section, the procedures for sizing
the last option may be most desirable. Also, in and specifying each of these components are dis-
dechlorination plants, the consumption of dechlori- cussed.
nating chemical will be a function of the chlorine
residual leaving the chlorine contact system, and 5.7.1 Disinfectant Generation and Handling
thus the use of true feedback control will help to In the case of chlorine or bromine chloride this
minimize costs associated with the dechlorination subelement includes. possible on-site generation (in
process. the case of hypochlorites), storage, metering, and
mixing of the disinfectant for delivery to the point of
For the dechlorination sub-system, the options for
application. In the case of chlorine dioxide, the items
chemical supply and generation are as follows:
included are identical to the above, except on-site
generation is mandatory.
I. Choice of chemicals
A. Sulfur Dioxide Chlorine and Hypochlorites
1. Method of Withdrawal The first decision to be made by the designer is
(a) Gas whether gaseous chlorine or liquid sodium hypo-
(b) Liquid/External evaporation chlorite is to be the chemica I used for disi nfection. For
B. Sulfites or Bisulfites small wastewater treatment plants, generally up to
100,000 gpd (378 m 3 /d), proprietary solid hypochlo-
rite systems are available. However, these are
The choice of sulfur dioxide versus sulfites/ bisulfites generally not used in larger plants.
is dictated by considerations similar to those that
govern the selection of gaseous chlorine versus The basis for deciding between chlorine gas and
calcium or sodium hypochlorites. If sulfur dioxide is sodium hypochlorite as a disinfecting chemical is
used, selection of gas or liquid withdrawal is governed primarily one focusing upon risk of chlorine transport
by considerations identical to those that pertain to the versus additional costs of sodium hypochlorite. Only
selection between these options for the use of in rare instances will the delivered cost of NaOCI
gaseous chlorine. It is virtually universal to use solutions per unit available chlorine be less than the
solution feed processes for sulfur dioxide dechlo- delivered cost of gaseous chlorine.
rination processes.
If sodium hypochlorite is used as the chlorinating
The types of mixing processes that may be used for agent, it may be purchased in strengths of 5 to 15
dechlorination are identical to those used for chlo- percent, or it may be generated onsite. If purchased,
rination.Due to the rapidity of the dechlorination storage tanks for sodium hypochlorite solution are
reaction, and the absence of deleterious side re- necessary. It is recommended that the size of these
actions, it is considered unnecessary to use separate storage tanks should equal the amount of disinfectant
contacting systems for dechlorination, since the short required for a time equal to the shipping time from the
residence time in the mixing and outfall structures vendor plus a 15 day emergency reserve against
generally suffices for the reduction of halogen strikes and transportation problems (149). At smaller
residuals to desired values. wastewater treatment plants, the reserve will be
higher due to the necessity of purchasing minimum
The choice of process control options for dechlor- loads of solution to achieve reasonable chemical
ination is identical tothat for halogenation processes, costs. The concentration of available chlorine in
with one exception. In wastewater treatment plants NaOCI solutions diminishes with time, and allowance
that discharge effluents into low dilution receiving must be made for this in computation of reserve
streams, the use of manual control will result in requirements. A loss of 0.031 percent/d from 10
overdosing of dechlorination chemicals into the percent NaOCI solutions, and 0.075 percent/d from
wastewater. This may produce deoxygenation of the 15 percent NaOCI solutions have been reported (46).

Example: A wastewatc3r treatment plant has an Figure 5-8. Schematic of Chloromat (lonics,lnc.) electro-
average daily flow of 0.116 m 3 /s (2.6 MGD) and an lytic hypochlorite cell (Reproduced by permis-
sion of the Water Pollution Control Federation).
average chlorine dOSEI of 5 mg/I is necessary.

Chlorine is to be supplied in the form of 1 percent
NaDCI solution. The shilPping time for the vendor is 3 Spent Brine
Chlorine Gas _,......;..--.... Caustic
days. Determine the required storage capacity assum- r-+HydrOgen
ing a 15-day emergency reserve.
i II
Step 1. Hourly chlonine requirement = (0.116
m 3 /s) (5 g/m 3 ) (3600 s/hr) (24 hr/d) =
50 kg/d
I -H-
-L_ ----.!L
Step 2. Daily NaDCI volume = (50 kg /d)/ (100/ Anode , .--yr
1 Cathode
kg/m 3 ) = 0.5 m 3 /d Na+--l.+
I, +
Step 3. Storage volume = (3 d + 15 d) (0.5
m3/d) = 9 m3

--11- 'I
Membrane I :I Diaphragm
Step 4. Correct for hypochlorite decay. (18 d)(0.031 -'-l-
percent/d) = 0.56 percent loss in 18 d.
(9 m 3 ) (10 percent)/(1 0-0.56 percent) = 9.5 m 3
I ~r
Therefore a storage tank of at least 9.5 m 3 (2,500 gal) Brine-
would be necessary.

Onsite Generation. For onsite generation of sodium ing strength (Ionics, Inc.); sodium chloride require-
hypochlorite, several systems are available. All re- ments are 2 kg/kg available chlorine without brine
quire a source of DC ele(:tricity (rectification facilities recycle, and 1.75 kg/kg with brine recycle. Doan and
may be included as part ofthe generation package). In Haimes presented a case study of the application of
addition, a source of brine is necessary, and partic- this system to a 24 MGD (1 m 3 /s) wastewater
ularly if seawater or rocksalt is used, pretreatment of treatment plant at Amherst, NY (151). In that case,
brine for the removal of calcium carbonate and iron brine pretreat ment with sodium carbonate and
(which precipitates in the alkaline side of the elec- caustic (some of which may be recovered from the
trolysis cell) may be required. electrolysis operation) to less than 10 mg/I -(as
calcium carbonate) hardness and less than 1 mg/I
In one such system (Chloromat-TM, lonics Inc.), iron was needed since brine was produced from
electrodes are separated by two membranes. Dne rocksalt. Capital costs for the entire system (electro-
membrane is a cation e>l:changer permitting only the lytic apparatus, power rectification, and brine pre-
passage of cations (sUl::h as sodium), and in this treatment) were estimated at $1,400,000 (1978
manner, chlorine gas is evolved, which is then dollars).
immediately dissolved to form hypochlorite solution.
In addition, caustic (NaDH) and hydrogen gas are Electrolytic hypochlorite generating equipment suit-
produced as by-products. The hydrogen gas must be able for on-site use is also manufactured by Engelhard
flared offto minimize fire hazards. A schematic ofthis Corporation (152) and Diamond Shamrock (61).
electrolytic system is shown in Figure 5-8. If onsite hypochlorite generation is used, short term
NaDCI storage (several days) may be desirable to
Individual cells are manifolded together to produce
permit maintenance of the electrolytic units, and to
hypochlorite generation capacities of 9-150 kg/hr
eliminate the need for continuous operation. In
(500-8,000 Ib NaDCI/d). Capital costs have been
wastewater treatment plants serviced by electric
estimatedforthe electrolytic units, exclusive of power
utilities with off-peak billing schedules, it may also be
rectification and brine pretreatment, and fit the
desirable to operate the electrolytic generation equ ip-
following equation (150):
ment only during offpeak hours and to store NaDCI
(5-66) solution for use during peak periods.
Capital Cost ($, 1974) = 10386 (kg NaDCI/hr)o.66
Gaseous Chlorine Supply. When chlorine gas sup-
Replacement of the cell anode and membranes has plied in cylinders or containers is to be used as the
been estimated as a contribution to operating costs to disinfecting chemical, it is important to decide upon
be $4.40-$13.20/1000 kg NaDCI (150). Power the necessary inventory in the same manner as
requirements vary with the desired product strength outlined for hypochlorite. In addition, the working
of hypochlorite and arEI 3.5-5.5 kWh/kg (1.6-2.5 inventory should be increased by an amount equal to
kWh/lb) available chlorine, increasing with increas- the number of chlorine containers simultaneously

under service. However, an inventory above that container). In addition, treatment plants using bulk
minimum calculated may be desirable since the rail or truck shipments generally practice liquid
shipping cost of bulk chlorine in the form of cylinders withdrawal. Benas presents a detailed case study of
or containers is a function of the number of cylinders bulk chlorine storage utilized at a 2.8 m 3 /s (65 mgd)
or containers shipped, and hence a larger working and a 1.3 m 3 /s (30 mgd) treatment plant in San
inventory may enable the treatment plant to take Francisco, CA (153).
advantage of quantity price discounts.
If liquid withdrawal is practiced, the contents of a
Chlorine is generally supplied either in 150 Ib single container are piped to an evaporator. Multiple
cylinders, ton containers, or larger rail or truck evaporators and containers may be used; however, it
tankloads.ln such containers, liquid chlorine exists in is undesirable to manifold several cylinders or
a pool at the bottom of the container and is in containers each with liquid withdrawal due to the
equilibrium with gaseous chlorine in the container possibility of liquid transfer between containers.
vapor space.
In piping systems where liquid chlorine is transported,
The chlorine may be withdrawn for use from ton particular attention must be given to preventing the
containers either as a gas or as a liquid. If it is possibility of heating of the pipeline. If this occurs
withdrawn as a gas, it may be directly piped to the (such as by exposure to sunlight), the liquid flowing
ejector, which produces a chlorine solution in water. full in the pipeline will expand and could hydro-
If it is withdrawn as a liquid, it is necessary to allow statically rupture the line. To guard against this
the liquid chlorine to vaporize in an external evapor- possibility, an emergency expansion chamber (or
ator prior to its dispersal in water. 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
There is a maximum rate of safe withdrawal of rated at less than the maximum bursting pressure of
gaseous chlorine from cylinders or containers. This the supply line. Figure 5-9 details this provision. In
limit exists because, as gaseous chlorine is with- addition, pipelines should be insulated to minimize
drawn, liquid chlorine must vaporize to maintain heating. Further information on the design of chlorine
equilibrium within the container. As this vaporization pipelines may befound in Chlorine Institute Pamphlet
occurs, heat must be withdrawn from the surround- #60 (154).
ings.lfthe rate of gas withdrawal is too high, the rate
of heat abstraction from the surroundings will not be While it is desirable to minimize the length of liquid
sufficient to prevent cooling of the chlorine to. low chlorine pipe, to reduce the risk of leakage in the
temperatures. At a sufficiently low temperature, the event of pipeline breaks, instances exist wherein this
vapor pressure of chlorine will be below the exit may be difficult. Cameron reported on the use of a
pressure of chlorine gas from the cylinder, and thus 490 m (1 ,600 ft) liquid chlorine pipe handling a flow of
no further flow will occur. To preventthis, a maximum 1,365 kg (3,000 Ib)/d of chlorine supplied to the City
gas withdrawal rate is permissible. This is dependent Island Plant at Atlantic City, NJ (155).
upon surrounding temperature and exit gas pressure.
For a 21C (70F) external temperature and 2.4 atm Chlorine evaporators.are usually supplied as integral
exit pressure, the maximum gas withdrawal rate from units by various suppliers and consist of a pressure
a 68 kg (150 Ib) cylinder is 0.8 kg (1.75 Ib)/hr, and vessel surrounded by a heating bath (hot water or
from a 910 kg (1 ton) container is 6.8 kg (15 Ib)/hr thermostated electrical) that serves to convert the
(44). For liquid withdrawal, no such cooling occur's incoming liquid chlorine into a gaseous chlorine
(since the heat of vaporization is supplied by an product stream. Evaporators are inherently high in
external evaporator), and the maximum withdrawal energy requirements, and manufacturers' literature
rate is set primarily by the hydraulics of chlorine flow should be consulted for such details. Design and
through the fittings; it is 90 kg (200 Ib)/hr and 180 kg safety information for evaporators are available from
(400 Ib)/hr, respectively, for the two types of con- the Chlorine Institute (156). In particular, pressure
tainers. relief devices leading to gas absorption tanks and
super heating of effluent gas are necessary. It is
In general, a liquid withdrawal system will be used recommended that an absorption tank (of caustic or
when the number of parallel containers necessary to lime solution) capable of neutralizing all liquid
. maintain the average daily chlorine flow while chlorine in the supply line plus 70 kg (150 Ib)/
obeying the maximum rate of safe withdrawal per evaporator be provided, and also that the exit gas be
container becomes unwieldy. For example, if 90 kg 11C (2()OF) higher in temperature than the boiling
(200 Ib)/hr of chlorine are to be supplied from ton point at the exit pressure (157). This latter require-
containers, at least 14 ton containers must be ment will prevent reliquification in the subsequent
simultaneously manifolded (90 kg/hr+6.8 l<g/hr/ gas piping as the pressure of the gas is reduced.
Flguro 69. Chlorine expansion chambers (reproduced from Chlorine Institute Pamphlet #60, 1982 by permission).

Alternative "A"
Alternative "B"


1/4V /'
~" Sch.aO

- 1/4V
@---/ 3/8V
1)- 1/4V
1" Sch.8iJ 1" Sch.80
(If Required) "-===r:J==::::J
Tee or
=t...l:==::J- Product
Welded Outlet
(Supports may be (As Applicable)
Necessary but
are Not Shown)

Item Name of Part

1 Primary Expansion Chamber Notes:

(Note 1)

2 SElcondary Expansion Chamber

1. Capacity-20% of line volume
(Note 2) 2. Capacity-10% of line volume
3 Rupture Disc (400 psi) (Note 3) 3. 4QO psi setting suitable for many
systems, setting must not exceed
4 Pressure Indicator or Alarm system design pressure
Switch, 1/2" Conn. (Note 4)
4. Liquid-filled protective diaphragm
5 Tee (Note 5)
6 Re'CIucing Tee (Note 5)
5. Fittings shall be forged carbon steel,
7 Elbow (Note 5) 3000 Ib cwp

a Reducing Elbow (Note 5)

9 Union (Note 5)

10 Valve (Note 5)

11 Plllg (Note 5)

For gas withdrawal systems, two major design options chlorinator at which point the pressure is reduced
exist. These have been referred to as pressure/ below atmospheric until the chlorine flows to the
vacuum/pressure (pVP) and all vacuum systems ejectors. In the all vacuum system, the chlorine gas is
(168).ln the former system, chlorine gas is withdrawn immediately expanded to a pressure below atmos-
from the supply containers into a pipe at greater than pheric, eliminating the initial pressure IinEi..ln a PVP
atmospheric pressure, whereupon it flows to a system, the initial pressure line introduces some
potential risk of gas loss in the event of line rupture; Figure 5-10. Chlorine manifold and switchover system
however, it is easier to manifold multiple cylinders (after White, 1972 reproduced by permission
than in the all vacuum case. of Van Nostrand Reinhold).

In gas withdrawal systems, two major concerns are

temperature in cylinders and supply lines, and
impu.rities in the gas. Filters should be placed inline Manifold A
CI 2
Ga g j1 =40 PSI

as close as possible to the chlorine cylinder or ~

manifold to remove liquid droplets and solid impur- ~
ities (80). It is desirable that the chlorine gas lines and Chlorine
Supply CI 2 o c:
subsequent connections be warmer than the chlorine t- .~
System G;age ..Q
cylinders or containers to prevent liquification of (Can be Any .c:
chlorine in the downstream lines. In an all vacuum Size Container) ..,.I.,.._r.-----.:J+--, ~+b.tl""""':c=J!...,t+..4-,.-.r
system (as opposed to a PVP system) this possibility is
minimized since the gas pressure is reduced con-
Manifold B ~
Chlorine Pressure
P2 =20psi

siderably below the vapor pressure of saturated Reducing Valves

chlorine gas at the existing temperature. In PVP
systems, it is also desirable to have duplicate piping
rator, flows to a chlorinator. The function of this
and manifold systems to facilitate cleaning.
chlorinator is to regulate the flowofchlorine gas, and
Gas cylinders manifolded together should be at to couple with the control elements so that this
similar temperatures. This may deserve special chlorine gas flow rate (and ultimately the chlorine
consideration in wastewater treatment plants located dosage to the wastewater) varies with the chlorine
in areas of extreme temperature variability. Other-' demand.
wise, there will be a transfer of contents between
containers, which could lead to over-filling and Chlorinator Design and Hydraulics. Chlorinators are
container rupture. Nevertheless, multiple ton con- sized at a variety of gas flow rates. Typical upper limit
tainers (or gas cylinders) may be readily manifolded to gas flow rates are 100, 200, 500, 1,000, 2,000,
provide easy swit?hover of cylinders upon emptying. 4,000,6,000 and 8,000 Ib chlorineld (1.9,3.8,9.4,
One such configliration is indicated in Figure 5-10 18.8,37.5, '112.5 and 150 kg/hr). The lower limit to
(80). Pressure reducing valve 1 (PRV 1) is initially set flow is 1/20 of this upper limit. If there are frequent
at a higher pressure, e.g. 375 kPa (40 psig), than PRV conditions necessitating a chlorine dose below
2, e.g. 238 kPa (20 psig). Initially, gas will flow only 1120th ofthe maximum, it may be necessary to have
from the cylinders in header A. When the pressure in two chlorinators in parallel, with a second used to
these cylinders approaches 20 psig, flow will com- regulate gas flow under low flow conditions.
mence from header B. The valves connecting the
individual cylinders to header A can now be shut, and In PVP systems, the lines,between gas supply and the
the cylinders replaced (since 20 psig is below the chlorinator should be on an uphill grade, to help
vapor pressure of chlorine at usual ambient temper- minimize liquid chlorine carryover. Typically the exit
atures, relatively little chlorine will be remaining in pressure reducing valve at the gas source is set at 172
these cylinders). Upon replacement, valves a, b, c, and kPa(25 psia). In ton containers, it is customary to
d can now be switched such that the header install a blind drip leg prior to the chlorinator as a final
containing the "old" cylinders is connected to the trap against liquid chlorine and debris passage, which
375-kPa (40-psig) pressure reducing valve. could damage chlorinator components.

In both liquid and gas withdrawal systems, it is The output from the chlorinator in a PVP system is a
desirable to have weighing scales for each gas regulated chlorine gas flow stream to the ejector
cylinder under service. Since chlorine liquid and gas where an aqueous solution is produced (however,
are at equilibrium at a constant pressure given a fixed often the ejector may be physically housed close to or
temperature, pressure will only drop ,from a cylinder in the same mounting as the chlorinator). The chlorine
upon exhaustion of all ofthe contained liquid and will ejector (also variously called the eductor or injector) is
thus provide little warning to the operator of imminent a device that conveys the chlorine gas into the
emptying. Thus, the only reliable means of inventory converging region of a jet of high velocity water. The
control, as well as verification of chlorine consump- diverging water downstream of the point of gas
tion, is weight. A variety of weighing devices exist; injection produces a vacuum, which draws the
however, these should permit the weighing of chlorine gas into solution and serves to convey the
cylinders or containers without lifting. gas stream from the pressure reducing valve at either
the gas manifold or the evaporator. Figure 5-11
The chlorine gas from the cylinder manifold, or in the illustrates the typical configuration of ejectors with
case of a liquid withdrawal system, from the evapo- fixed water flow rates, and with variable water flow
rates set by a variable orifice. Additional configura- be at least 1.2 m (4 ft) of water (1.7 psi) to
tions are available wherein the gas flow rate may be prevent chlorine release in the solution line.
controlled by a variable orifice on the gas inlet pipe.
3. The ejector. should be installed as close as
The inlet water for the ejector may be either city possibletothe mixing pointto minimize the time
water, or, more commonly, chlorinated secondary lag in the solution line and the back pressure at
effluent. As will be indicated below, the inlet water the ejector.
supply to the ejector must be at relatively high
pressure, and there1fore if secondary effluent is used, 4. The vacuum line carrying gas between 'the
a booster pump is required. In all but the smallest chlorinator and the ejector should have a
wastewater treatmElnt facilities it will be less expen- pressure drop of less than 3.8 cm (1.5 in) of
sive to provide this booster pump for secondary mercury (0.7 psi).
effluent ejector watc3r supply rather than to rely upon
purchased city water. Specific information needed to size and specify the
ejector is the following (Fischer & Porter):
Important considerutions for ejector sizing include
(157): maximum capacity of the chlorinator;

inlet water supply pressure (this can be assumed,

1. The ejector water supply should be no less than with several values, for the purpose of determining
that required t,o produce a chlorine solution of ejector water booster pump capacity);
less than 3,500 mg/I to minimize the possibility
of chlorine gas; volatilizing from the feed solu- back pressure at the ejector outlet;
elevations of the ejector and the diffuser; and
2. The ejector back pressure (sum of static and
dynamic pressure losses between the outlet distance between ejector and diffuser, and any
side ofthe ejector and the mixing device) should piping bends or constrictions.

Figure 6-11. Schematic of fixed and variable orifice ejectors. (Reproduced courtesy of Fischer and Porter.)

Diaphragm *Diaphragm
Back-Flow Back-Flow
Check Valve Check Valve

\ -.J-+-- Throat \ 4-\-- Th roat

G.H~ Orifice

Back-Flow Plug
Ball Check
Either One
or Both.
as Required *Either One
or Both,
as Required

Fil<ed Orifice Ejector Variable Orifice Ejector

Ejector SIZing procedures will be illustrated by pressure is 273 kPa (25 psig) and the acceptable
example. Assume a maximum chlorine application waterflowrate is 2.8 lis (45 gpm). Now it is necessary
rate of 18.8 kg/hr(1 ,000 Ibid) has been calculated. It to verify that the system backpressure is below the
will also be assumedthatthe booster pump supplying maximum acceptable (273 kPa). The backpressure is
feedwater to the ejector can provide a pressure of 515 the sum of the following elements:
kPa (60 psig)). At the point of mixing (the outlet side of
the diffuser nozzles) the pressure is 1.8 m (6 ft) of elevation of diffuser with respect to ejector (A);
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 back pressure on diffuser outlet (8);
m) of equivalent pipe (pipe length plus length
friction loss through diffuser (C); and
equivalents offittings obtained from standard hydrau~
lic handbooks) between the ejector outlet and the friction loss in piping (D).
diffuser. It is assumed that the headloss through the
diffuser itself is 13.8 kPa (2 psi) or 1.4 m (4.6 ft). From the assumptions, A + 8 + C = 5 ft + 6 ft + 4.6 ft =
15.6 ft or 6.8 psig. In order to calculate D, the pipe.
For one possible ejector (Fischer & Porter, Figure 5~ diameter needs to be specified. Assumethat 1-1.4 in ID
12), manufacturer's data are available for acceptable Schedule 80 PVC pipe is used. Frictional losses may
combinations of inlet and back pressure and water be estimated using the graph in Figure 5-13. At a
flow rate. Since the inlet pressure is assumed at 515 water flow of 45 gpm in1 -1.4 in PVC Schedule 80 pipe,
kPa (60. psig), from Figure 5-11, the acceptable back the frictiona I loss is 37 ft/1 00 ft. Therefore the tota I

Figure 6-12. Ejector sizing curve (Reproduced by permission of Fischer and Porter).


Chlorine Gas-Feed Rate

180 10001b/d



0> 120
til 100
co 80




0 140 160 180 300

Supply Pressure, psig

34 40 45 50 55 60 65 70 75 80 85 90 93
Water Flow, gpm

Figure 6-13. Frictionalll:lsses in solution piping (Reproduced courtesy of Fischer & Porter Co.).

44 100


u- 20
a 5 .8. 10
CD 5

1 2 3 5
Water Flow Rate, gpm
Rubber Hose - - - - - - - - _
PVC Pipe Sch. 80 - - - - -
PVC Pipe Sch. 40 - - ~ - -

system backpressure (A+B+C+D) is 53 ft or 260 kPa pressure is more variable than in Figure 5-12. For the
(23 psig). Since this is below the maximum acceptable above example, at an ejector inlet pressure of 515 kPa
back pressure, the design is satisfactory. At this point, (60 psig) and 273 kPa (25 psig) backpressure, a water
s different inlet assumEld pressure may be tried to flow of 15.5 I/s (245 gpm) is required for this
investigate the trade-ofts inherent in various booster particular ejector. Losses A, 8 and C will remain as
pumps. It is also necessary to determine the transit above (15.6 ft). If schedule 80 PVC pipe is used, from
time in the solution piping, which is necessary, as will Figure 5-12, to keep the total backpressure below 25
be noted later, to ensure that unreasonably long psig (273 kPa) (and, in other words, to keep the
control lags are not present in the system. frictional losses in the solution line below 231 kPa (19
psig), 3 inch (76 mm) ID pipe must be used. In this
The flow rate in the solution pipe is 2.8 I/s (45 gpm). case, the residence time in the solution line is 9 sec~
Assuming that the entirE~ pipe is straight, the volume or nearly identical to the lag obtained with the
of a full 100 ft pipe of 1-% in ID is 24.5 I (0.85 cu ft). preceding ejector. In this case, however, the neces-
Therefore, the residenc:e time in the line is 0.14 sary ejector water flow rate is substantially higher,
minutes (8.5 sec) which is reasonably short. A larger which may necessitate the use of more costly
diameter line, while serving to reduce back pressure, pumping equipment and larger operational pumping
would also have increasl~dthissystem delay time. As costs. It should be noted that for both chlorine
will be noted, the total ohhe lag time in the solution soJutionpiping itself and the ejector, maximum
line, the time for analysis, and the residence time pressure limits exist above which material failure
between mixing and sample collection for analysis in becomes possible. Generally these pressures are in
the contactor should be under several minutes for excess of 100 psig (790 kPa); however, they decrease
best performance when residual control or compound at above-ambient temperature. Prudent practice
loop control is practiced. dictates that the inlet water supply pressure be
maintained belowthat specified by the ejector or pipe
As a more complex example, a second manufacturer manufacturer.
(Wallace and Tiernan Division of Pennwalt, Figure
5-14) produces ejector specifications in which the At this point, knowing the flow rate of chlorine
relationship between water flow rate and injector solution and the desired maximum headloss through
Figure 6-14. Ejector sizing curve (Reproduced by permis- a nomograph for the estimation of head losses
sion. Wallace & Tiernan Division. Pennwalt through perforated pipe diffusers. In the present
Corp.). example. given a solution flow of 2.8 lis (45 gpm). a
diffuser with a headloss of less than 13.7 kPa (4.6 ft)
is desired.
300 r------.---r---.,-----,-=:--..,..,....----,
A vertical line at 1.4 m (4.6 ft) headloss is constructed
to the point of intersection with the loss of head curve.
275 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
225 number of orifices needed is obtained by dividing 2.8
lis (45 gpm) (the total solution flow) by the flow per
orifice and rounding uptothe nearest whole number.
200 Thus, 6 YHn orifices would be needed; 5 9/16-in
orifices. etc.

175 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
150 known. For example. to feed 18.8 kglhr (1,000 Ibid)
of chlorine using a NaOCI solution containing 10 percent
available chlorine:
1000 Ibid
x 454 glIb =4.540 lid
100 gil =1.200 gpd
would be needed. This flow rate is equal toO.05 lis
75 (O.83gpm). Hence. from Figure 5-15. if a h-in hose
spray nozzle is used. a headloss of approximately 8.1
kPa(2.7 ft) would result.

2.5 Materials Compatibility. In the design of piping and

25 ---~-------_=:..._---- handling systems for both chlorine and hypochlorite.
3500 mgll Satubility 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-
o 25 50 75 100 125 150 cussed. For dry chlorine gas. copper. iron, and steel
Injector Pressure, psi pipe are satisfactory (44). For hypochlorite solutions
(including the solution piping from the chlorine
Note: Extrapolations and interpolations
should not be used.
ejector). rubber. ceramic. glass, Tyril. saran, PVC,
vinyl and Hypalon are all suitable materials (159).
the diffuser structure. one can design the diffuser
itself. For small solution flows. a single spray nozzle Bromine Chloride Supply
on the end of the solution pipe or hose may be Systems for the supply and handling of bromine
adequate. Figure 5-15 can be used to determine the chloride are similar to those for the supply and
headloss for this design as a function of hose handling of gaseous chlorine. BrCI is supplied in
diameter. cylinders of 68 kg (150 Ib) capacity and containers of
1.365 kg (=3.000 Ib) capacity. Since the vapor phase
For the solution flow rates used in the example case. above liquid arCI is relatively enriched in molecular
however. the use of a diffuser with multiple perfora- CI 2 due togas phase dissociation. it is necessary to
tions is more likely. Such a system consists of a pipe withdraw BrCI from either type of container from the
lateral of diameter equal to the solution pipe itself. liquid phase.
with multiple outlet holes. This diffuser pipe may be
suspended horizontally or vertically in a pipe or in a The vapor pressure of BrCI is less than that of gaseous
rapid mixing chamber or device. Figure 5-16 provides chlorine at identical temperatures. This means that.

Figure 6-16. Headloss thru spray nozzle diffusers. (Reproduced courtesy of Wallace & Tiernan Division, Pennwalt Corp.)

10 .---r---r-,.-.-r-_r_-----r----r----'--r---r-r--~-"'TT__..,r__....__r_.,..---r_--....,

Flow in gpm (Under Water Discharge)

while similar liquid evaporators may be used for BrCI Similarly, design of ejectors and diffuser structures
service, it is necessary to attain a higher degree of for the resulting bromine chloride solution follows
superheating of the BrCI product gas to prevent principles noted above for gaseous chlorination
reliquifaction. White recommends that 30F super- systems.
heat be provided to thl:l product gas (24). Chlorine Dioxide Supply
Bromochlorinators (analogous to chlorinators) and It is necessary to generate chlorine dioxide on a
solution ejectors may be designed on a similar basis continuous basis for use as a disinfectant. Although a
tothose used for chlorine service. The only substantial few European potable water treatment plants use the
differences are that materials specified for chlorine acid-chlorite generation process (21 ), the most com-
service may not necessarily be suited for BrCI service. mon synthesis route for disinfectant CI02 generation
It is recommended that piping should be black iron or is the chlorine-chlorite process.
carbon steel, but not cast iron (56). Other suitable
materials for BrCI servicl;l are tantalum, nickel-molyb- In the chlorine-chlorite process, sodium chlorite is
denum alloy, lead, silvElr, platinum, glass ceramic and supplied as either a solid powder or a concentrated
some plastics (56). Early pilot plant studies. noted solution. A solution of chlorine gas in water is
frequent clogging of BrCI evaporators (42). This produced by a chlorinator-ejector system, of design
problem persists in BrCI evaporators, due, apparently, similar to that used in chlorination. The chlorine-
to substantially greater impurities existing in BrCI as water solution and a solution of sodium chlorite are
compared to C1 2 As a result, at least one manufac- simultaneously fed into a reactor vessel, typically
turer has withdrawn its designs for BrCI evaporation PVC, 36-42 inches (91-'107 em) in height and 8
systems, and is currently testing direct liquid BrCI inches (20 em) in diameter, packed with Raschig rings
feeding equipment (1'. Zeh, Capital Controls Co., to promote mixing (21). From Equation 5-28. one
personal communicatil:>n). mole of chlorine is required for two moles of sodium

Figure 5-16. Nomograph for design of multiple perforated diffusers (Reproduced courtest of Wallace & Tiernan Division,
Pennwalt Corp.)

Head Loss thru Orifice, feet

2 3 4 5 6 7 8 9 10 11 12
Orifice Diameters, inches
- 315
26 16 4 16





.::= 16

0 14
0 12


Chlorine Solution Flow, gpm

chlorite-or 0.78 part CI 2 /part NaCI0 2 by weight. It To provide greater yields; several options exist. First,
has been found, however, that for this reaction to it is possible to produce chlorine-water solutions in
proceed to completion it is necessary to reduce the pH excess of 3,500 mg/l using pressurized injection of
below that provided by the typical chlorine-water gas. In this case, however, there will be an excess of
solution produced by an ejector. At 1:1 feed ratios by unreacted chlorine in the product solution, and the
weight, only 60 percent ofthe chlorite typically reacts resultant disinfectant will consist of a mixture of
(21 ). chlorine and chlorine dioxide. The second option
(160), consists of ac~d addition to the chlorine and ian and Day (126). In their studies, the vertical
chlorite solutions; a D.1 M HCIIM chloride addition gradient in distribution of chlorine residual in a
enabled the production of a disinfectant solution of 95 contactor downstream of a simple dropped hose
percent purity in terms of chlorine dioxide, and diffuser was measured and was found to be sub-
achieved a 90 percent conversion of chlorite to stantial unless adequate turbulence was present at
chlorine dioxide. A third process, developed by CIFEC the point of application of the chlorine solution.
(Paris, France), involves recirculation of the chlor- Unless mixing is present, there will exist zones of
inator ejector dischan~ewater back to the ejector inlet high and low disinfectant concentration, and the
to produce a strong chlorine solution (5-6 kg/m 3 ) surviving microorganisms present in the zones of low
(0.04-0.05 Ib/gal), typically at pH below 3.0, and in concentration will act to diminish the process effi-
this manner to increaise the efficiency of chlorite con- ciency.
version (21,24). It h"ls been reported that this last
option is capable of pwducing 95 to 99 percent pure Longley determined that the degree of mixing at the
solutions of chlorine dioxide (24). point of disinfectant application has a pronounced
effect upon the initial rate of inactivation of. a variety
In any of these generation processes, the additional of microorganisms over the range of G (root mean
equipment needed (over and above a chlorinator- square velocity gradient) values of 100-10 5 S-1 (161).
ejector system) includes the reactor vessel, metering White recommends rapid mixing with a G value of
pumps for the chlorite (and, if used, acid) solutions. 500-1,000 S-1 and a residence time of 5-15 seconds
and ancillary piping. (24).

In the acid-chlorite method of generation, sodium In-Line Diffuser

chlorite and hydrochloric acid are used as reagents. Four major options for mixing exist. In the first case,
There appears to bo substantially less operating the diffuser may be placed in the center of a pipe
experience with this process in water or wastewater running full at a turbulent Reynolds number. For this
treatment (21,24); hl)WeVer, it appears to be con- system, the mixing intensity may be computed from
ceptually simpler than the chlorine-chlorite process. hydraulic relationships as follows:
In one plant in Europe, (21), a batch reactor is used to
mix the reagents in amounts necessary to produce a Computation of G value in pipes:
20 percent final chlorine dioxide concentration; 15
minute reaction time has been found to be sufficient. The metric form of Manning's equation may be
There are also reportod to be continuous generation written as:
processes available for use with the acid-chlorite
method (24).
where S is the hydraulic gradient (dimensionless), n
Due to the instability of strong chlorine dioxide
is Manning's coefficient, v is the flow velocity in mis,
solutions, it is not practical. to store disinfectant
and .R is the inside pipe diameter assuming a full-
solution for any significant period oftime. Hence, the
flowing pipe (or hydraulic radius) in m.
rate at which the chle'rine dioxide is generated must
be capable of being cCiupled in some manner directly Camp and Stein's relationshipforthe G value is given
to process sensors (C3. g., flow meters or residual as:
analyzers). This is generally done by varying reagent
flow rates. G = (P/uV)1/2 (5-68)
In addition, with both of the generation processes, a where P is the power dissipation (Watts), V is the
chlorine dioxide solution is produced directly, rather volume over which such dissipation occurs and u is
than a chlorine dioxide gas stream. Since chlorine the viscosity.
dioxide gas at high concentrations may produce an
explosive mixture, its handling as a solution is For a closed conduit in which power dissipation
preferable. This chlorine dioxide solution may be occurs by internal fluid friction, the power dissipation
pumped directly to a diffuser or may be diluted in a is given in terms of the headloss (h 1) by:
solution ejector (in which the strong solution is sent
to a vacuum ejector similar to that used for gaseous (5-69)
chlorine application for the purpose of dilution) and
then piped to a diffusElr. where Sw is the density of water, g is the gravitational
acceleration, and Q is ,the volumetric flow rate. By
5.7.2 Mixing Systems multiplying S obtained from Equation 5-67 by the
The necessity for rapid mixing of the disinfectant length of pipe (l), the headloss (he) may be obtained.
solution exiting the diffuser ports with the bulk The power dissipation then obtained from Equation
wastewaterflowwas uraphically shown by Heukelek- 5-69 is substituted into Equation 5-68, and the

volume is obtained via the geometric relationship for shooting flow (Le. flow at a sub-critical depth) occurs
a volume of a cylinder. The resulting relationship for at the bottom end of the chute, and that a hydraulic
the G value in a pipe flowing full is given by: . jump is located downstream ofthe diffuser but within
the downstream channel.
G = (SwgSV/U)1/2
Figure 5-18. Details of a hydraulic jump mixing structure.
Finally, to ensure turbulent flow, it is necessary to
check the Reynolds number from:

Re = DvSw/u (5-71 ) a':'yc


If Re is in excess of 2,000, turbulent flow may be


Using these relationships, given the design flow, the

velocity may be computed forvarious 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 !3
satisfactory. The headloss may be computed directly
from Equation 5-67.
Sc=tan a Sd=tan!3
Figure 5-17. Details of a submerged weir mi>[ing structure.
The various computations for designing a hydraulic
jump mi)(ing device are outlined below. For more

detailed discussions, standard hydraulic texts should

--- JI('~~ ;;;--,.

. ..;.:"4I.U
__"n (' - _
be consulted (162).

Step 1-Decide upon the entering jump Froude

Cb Diffuser
The Froude Number (F) is defined by Equation 5-72, Submerged Hydraulic Structures where g is the gravitational acceleration (9.8 m/s 2 ), a
The second mixing option involves the use of a is the volumetric flow rate, Y4 is the liquid depth and b
hydraulic structure at which turbulence is induced, as is the channel width.
the point of application of chlorine solution. Two such (5-72)
possible structures are the submerged weir and the
hydraulic jump (24). F must be greater than 1 for a jump to occur
(supercritical flow), and Chow indicates best per-
A submerged weir consists of an open channel in formance in terms of stability and insensitivity to
which a rectangular projection arises from the downstream conditions for F between 4.5 and 9.0
channel base in such a manner that both the free (162).
liquid surfaces above and downstream from the weir
are at a greater elevation than the weir itself. The To illustrate these calculations, a hydraulic jump will
diffuser inlet should be located at a point of maximum be designed to serve as a mixing device for a
turbulence downstream from this structure. Figure wastewater of average flow 26.5 MGD (1.16 m 3 /s).
5-17 details the design of a submerged weir mixing F()r a Froude number of 6.0, and assuming a channel
structure. width (w) of 1 m, the value of Y4 from Equation 5-72 is
0.156 m. From the depth of the water immediately
The major variables necessary to design the sub- upstream of the jump, the final depth immediately
merged weir mixing structure are the distance downstream (y4) may be determined by:
between the weir and the point of maximum turbu-
lence, at which the diffuser should be located, and the (y"/Y4) = 0.5 (-1 + (1 + 8F 2 )1/2) (5-73)
anticipated headloss across the weir.
Step 2-Determine chute slope (Sc) and Y3.
The hydraulic jump following a chute and a sharp
change in channel slope may also be used as a mixing For the hydraulic jump to form downstream of the
device. Figure 5-18 details the general characteristics termination ofthe chute, it is necessary that the depth
of a hydraulic jump used as a mixing structure. To at the chute outlet, Y3, be some value less than the
design such a structure it is necessary to'ensure that depth immediately prior to the hydraulic jump (Y4).

One can assume a chute slope and calculate y from bottom to the initiation of the jump (lj) of 6.12 m is
Manning's equation (Equation 5-67) or vice-versa. obtained.
Both the hydraulic mdius and the flow velocity are '
Table 5-13. Computation of Length to Jump (LJ)' (Assumed
functions of Y3. HE = 0.01)
Assuming a value ofY3 of 0.12 m (0.4 ft), the hydraulic (1 ) (2) (3) (4) (5) (6)
d (m) RH (m) v (m/s) S Savg L12 (m)
radius equals 0.0968 m (WY3/(W + 2Y3)), and the flow
velocity equals 9.67 m/s (Q/WY3). From Equation 5- 0.12 0.0968 9.67 0.473 0.421 1.71
0.13 0.1032 8.92 0.369 0.333 1.68
67, assuming a Manning's coefficient of 0.015 (typical 0.296 0.267 1.74
0.14 0.1094 8.29
of finished concretel), the frictional headloss (or 0.15 0.2154 7.73 0.239 0.226 0.99
required channel slope for uniform flow at the Total Lj if.'l2
indicated depth) is 0.473 (Le., the channel must have
a rise of 0.473 m/m horizontal). Hence, 5 e must be
Therefore, in the example case, the headloss dis-
0.473, and thus the angle of inclination is 28.
sipated in the jump is 1.68 m. By application of
Step 3-Determine le,ngth until jump (lJ)' Equation 5-69, the power dissipation equals 19.1 kW.
It is not readilyfeasibleto estimate the G value, since
The water depth increases from the base ohhe chute the volume of the liquid contained within the jump is
(Y3) until the onset of the jump (Y4) due to frictional not easy to obtain.
losses that reduce thl3 energy (and since the flow at
Step 4-Determine length of jump (lH).
these points is subcritical, increase the depth). To
determine the length oftravel between the chute exit The ratio of jump length to height after the jump
and the jump onset, numerical integration of the (lH/y) has been observed to be a function of the
energy conservation equation in conjunction with Froude number alone. For Froude numbers between
Manning's equation must be performed. 4.5 and 9, this ratio is relatively constant at 6.15
(162). In the example problem, since y has been
This process will be illustrated in tabular form. The
determined to be 1.25 m, the value for lH is 7.69 m.
basic relationships arc" Manning's equation (Equation
5-67) and the following rearrangement of Bernoulli's Step 5-Determine minimum chute length le.
equation accounting for changes in channel elevation
and frictional losses: A sufficient length of chute is, needed so that the
water velocity attained at the base of the chute is its
l12 = (c1 2 - d1) + (v~ - V~) , (5-74) normal value as predicted by Manning's equation. At
2g(h e12 - 5d the onset of the chute, the depth of flow may be
assumed to be equal tothe critical depth (Yd, predicted
In Equation 5-74, l12 is the channel length between from:
two points (1 and 2) having indicated water depths (d)
and velocities (v), and between which there is a (5-75)
channel floor slope of h e12 (dimensionless rise/run)
and a frictional headloss of 5 12 (dimensionless). Thus, for the example under discussion, the initial
depth at the top of the chute, or Ye, equals 0.52 m.
The computation ofl, is executed in Table5-13. First,
the range of depths between Y3 and Y4 is subdivided It is necessary that the length of this chute be
into a number of intervals arbitrarily (however, the sufficient so that the depth changes from 0.52 m to
number of intervals will influence the precision ofthe the design value of 0.12 m at the chute base. The
calculation) as indicatl3d in column (1). In columns (2) slope of the chute bottom (h e12) is 0.473. Equations
and (3), the hydraulic radius and the velocity, respec- 5-68 and 5-73 may be used in the manner in which
tively, associated with each depth (at the known flow they were applied in Table 5-13 to estimate the'
of 1.16 m 3 /s and wcdth of 1.0 m) are computed. minimum length of this transition. By this procedure,
Column (4) represen1ts the frictional slope (5) as- it can be calculated that the length of the required
sociated with each depth, and is obtained by the use chute is 6.99 m.
of Equation 5-67 at eal~h point. Column (5) represents Step 6-Comput13 energy dissipation in jump.
the arithmetic average of the frictional slope in the
range defined by an upstream and a downstream The amount of energy transferred to the fluid in the
depth. Finally, column (6) tabulates the calculated hydraulic jump, and thus available for mixing, may be
length increment from Equation 5-73 associated with given by the difference in the energy of the fluid
the change in depth-a channel bottom slope of 0.01 immediately prior to and following the jump. Chow
has been assumed; however, this has only a very gives the headloss in the jump as (162):
minor effect upon the result. By addition of each of the
length increments, a total length from the chute (5-76)

Step 7-Repeat computations at low and high flows. metries for mixer design exist, and the necessary
motor speed will be a function ofthe specific propeller
Befor~ the above design can be accepted, it is
or turbine utilized.
necessary to determine that an acceptable hydraulic
jump exists in the given physical system under Jet Mixer
anticipated low and high flow conditions, and fur- One additional mixing configuration exists in which
thermore that the location of the diffuser is upstream the chlorine solution (or in some cases, the gaseous
of the jump in all likely flow situations. By using the chlorine) coming from the ejector is introduced into a
design specifications (chute slope and length, chan- large jet, of similar overall design to a chlorine ejector.
nel width, downstream slope and length) with the This jet mixer carries a substantial volumetric fraction
expected low and high flows, the values for position of of the influent wastewater through a high velocity
the jump under all circumstances can be calculated nozzle and the disinfectant solution (or the disin-
by application of the foregoing equations. fectant gas stream) is injected into the vacuum
Once these computations are completed, the hydrau- produced.
lic profile of the design may be sketched, and the
This jet mixing system is of proprietary design;
static headloss computed. A scale diagram of the
however, it has been claimed that a G value in a jet
jump illustrating diffuser location is given in Figure
tube of 10 S-1 or more, and a dimensionless tube Gt
5-19. product of 1.5-15 produces satisfactory mixing (163). Mechanical Mixer The jet mixer may be mounted directly in the entrance
The third mixing option involves the use of a region of the contact chamber, and a booster pump
mechanical mixer, such as a propeller or turbine used to provide the motive power. In one such field
mixer, in conjunction with a small residence time test, it was shown that the required dosage of
mixing chamber containing the disinfectant diffuser. chlorine was less with jet mixing than with poor
If a G value is specified (for example, 500-1 ,OOO/sec mixing (164); however more recent tests suggest that
as recommended by White (24, then Equation 5-68 such reductions in chlorine dose, and in any neces-
can be used to determine the mixing power to be sarydose of sulfur dioxide required for dechlorination
imparted to the fluid per unit volume. may be insufficient to compensate for the additional
energy costs of jet mixing (165).
For example, for the 26.5 mgd (1.16 m 3 /s) waste-
water flow described above, if a G value of 750 S-1 is Ofthe mixing systems, the most common is the use of
specified, and the viscosity is assumed to be 0.001 an in-line diffuser, followed by a rapid mixer, then
kg/m-s, then from Equation 5-68, a power per unit hydraulic jumps and finally jet mixers. In-line dif-
volume of 866 W/m 3 is required. If the rapid mixing fusers are particularly applicable to small plants.
chamber has a residence time of 10 seconds, and However at flows greater than 10 MGD it becomes
hence a volume of 11.6 m, then the required mixing difficult to design a fUllyflowing turbulent section of
power to be imparted to the fluid equals 1,385 W. To pipe. Both in-line diffusers and hydraulic jumps share
determine the actual electric energy consumption for the advantage of-requiring no direct external power
the drive motor for the necessary mixing device, it is input. However, hydraulic jumps are inapplicable
necessary to consult manufacturer's specifications unless upstream equalization of flows is practiced,
as to the efficiency of conversion of electrical to since the horizontal location of the jump is extremely
mechanical mixing energy. Similarly, differing ge'o- sensitive to flow variations.

Figure 5-19. Scale diagram of jump as designed.

Scale: f--l = 1 meter

~ Diffuser Static

6.12 7
--~.+l..---,--~. 7.69 . --.--..l~1

Chute L Increasing
Depth + Development
of Jump
1.75m Below
Base of
Chute Inlet

5.7.3 Contacting Systems long distance is available, the outfall pipe itself may
The role of the chlorine contact system is to permit be used to provide contact. In other cases, it is
sufficient time to elapse for adequate disinfection to necessary to design a separate tank for contacting. In
occur. A characteristic measurement of time is the either case, the volume to be contained in the
mean hydraulic residl3nce time, However, it is also contactor may be determined from the equation V =
necessary to ensure that the bulk of the wastewater e
Qe, where Q is the design flow, is the design contact
has had the opportunity to remain in the contact basin time, and V is the volume of liquid to be contained in
prior to release. Figure 5-20 depicts the frequency the contactor.
distribution of fluid residence times in three basins of
identical e.
The spread of residence times can be For pipe contactors, diameter influences the headloss
characterized by the dimensionless dispersion index, in the pipe, which may be determined via Manning's
d, which decreases as the range of individual equation, the flow velocity in the pipe, and the
residence times decreases. See Section 4.2 for a dispersion. The velocity in the pipe must be kept
theoretical developml3nt of residence time distribu- above the scour velocity for particles to avoid sedi-
tion functions. mentation. Generally a 1 ftls (0.3 m/s) flow velocity
wili suffice, or the designer may use the scour
relationship of Camp to estimate the minimum
Figure 620. Residenl:e time distribution functions for
contact basins. velocity. The dispersion in a pipe may be estimated by
the use of the following equation (166):

d, (J= Mean Residence Time d = 89,500 f3.6 (D/L)o.B59 (5-77)

In the above equation, f equals the Darcy Weisbach .
friction factor for pipe flow, which is obtained as a
function ofthe pipe Reynolds number and the relative
roughness from a Moody diagram (170), D is the pipe
diameter, and L is the length ofthe pipe. This equation
is valid for the case of a straight pipe alone, without
bends or other flow disturbances.

If it is necessaryto'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

o L..- --":.._I-_----O. ------l In a baffled contact chamber, the flow velocity

o 1D (defined by the length of the flow path divided by the
tiD mean residence time) should be greater than the
scour velocity to minimize particle deposition. The
dispersion in rectangular baffled contact chambers
It has been shown that the performance of a chlorine (either end-around or over and under baffles) may be
contact chamber declines substantially (Le., the given by either Equation 5-78 (166) or Equation 5-79
effluent microbial concentrations increase) when the (157):
dispersion in the contactor increases (166), and this
finding has been verified experimentally (167-169). d = 0.14 K/(L/W) (5-78)
There appears to be a point of diminishing return, and
Trussell and Chao (1136) propose that d =0.01 is a d = 1.15 (L/Wr 1 .13 (5-79)
practical lower limit to be attained to optimize the
hydraulic performancl~ of chlorine contact chambers In Equation 5-78, K is a coefficient of nonideality,
in wastewater disinfel:tion. found to vary between 2.3 and 15.8 for real con-
tactors. In both cases, L is the length of the flow path
Once the average residence time for the contactor and W is the width between parallel baffles.
has been determined, from the kinetics of microbial
inactivation discussed above, several options exist for The headloss in baffled contactors is due to frictional
the configuration ofthe contactor. When disinfection losses with the bordering surfaces and dissipative
is the final treatment process (in other words, when losses associated with changes in flow direction.
no dechlorination is to be practiced), and a sufficiently Frictional losses may be evaluated as in the case of

Figure 5-21. Types of baffled contact chambers (174). Figure 5-22. Vaned serpentine contactor design of Louie
and Fohrman(172).

Cross Baffled Tank
(May be Horizontal
or Vertical)

I j

Baffled Tank

Annular Ring Surrounding Vanes
Secondary Clarifier
basins with full sharpcrested weirs performed better
than partial length Cipolleti (Le. 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
any open channel flow system. Losses due to velocity chambers should be designed to introduce flow over
changes may be estimated using the following (171): as large a fraction of the width as possible, and with
minimum velocity. Alternatively, a single pipe inlet
h 1 = ((n + 1) v/ + n vl)/2g (5-80) impinging upon a momentum-absorbing baffle may
be used.
In this equation, V1 and V2 are, respectively, the flow
velocities between parallel baffles and through the Contact chambers, like many wastewater treatment
baffle slots, and n is the number of baffles. It is unit processes, will occasionally foul with a microbial
possible to reduce the headloss associated with these slime. Therefore, at least two parallel contact cham-
velocity changes by adding flow redirectors. Louie bers should be used so that one can be drained and
and Fohrman evaluated a number of such configura- cleaned. The contact chamber should have provision
tions, and concluded that the vaned serpentine design for draining, preferably by gravity, and be in proximity
in Figure 5-22 provided minimum headloss and to a high pressure water hose for cleaning.
satisfactory reduction in dispersion (172).
The reqUired size of a contact chamber is a function of
It has also been suggested that the performance of the chlorine dose to be used, the nature of the
baffled rectangular contact chambers may be im- wastewater to be disinfected, and the required
proved by. the use of air scour. This has been amount of disinfection to be achieved. For a first
demonstrated in scale models (173); however, little approximation, manufacturer's data may be consulted
full-scale quantitative data exist to evaluate this for the doses of chlorine to be used in typical
effect. In addition, such scouring would increase situations. As a more exact estimate of the effect of
dispersion, possibly to a deleterious degree. contact time on required chlorine dose, information
presented earlier on the kinetics of microbial inacti-
The effluent weir from a contact chamber should vation and the decay of chlorine residual may be used.
extend along the full length of the exit channel from This will be illustrated in the design example to be
the flow channel. Marske and Boyle (174) found that presented below.

5.7.4 Process Control a gas chlorinator vacuum line (with the flow being
The function of the process control element is to linear in orifice area) orthe metering pump speed on a
ensure that there is; consistent disinfection perform- hypochlorinator to provide a relatively constant
ance despite fluctuations in the quality and quantity chlorine dose. One drawback of this strategy is the
of wastewater influent to the disinfection process. slight nonlinear behavior of flow rate versus pump
Due to the scarcity of experience with chlorine horsepower due to intervening friction losses in the
dioxide and bromine chloride as wastewater disin- line between the pumps and the chlorine contact
fectants, only procE!sS control of chlorination will be chamber. In addition, there will be a lag time in
discussed. process response due to the hydraulic residence time
in the system between the pump station and the point
The simplest control strategy involves eitner the use of chlorine application. During this lag, there will be
of flow equalization basins prior to chlorination or under or over application of chlorine as flow de-
extremely conservative design factors. If a sufficiently creases or increases, respectively.
large equalization basin exists, e.g., a lagoon with
mUltiple day residence time, upstream of disinfection, A second form of flow proportional control uses a flow
the wastewater quality and quantity entering the meter (e.g., magnetic in-line, or venturi) to sense the.
chlorination proces~; will be relatively uniform. Thus, flow entering the chlorine contact chamber. Due to
the performance of the chlorina'tion process will be the problem of lags, the plant raw wastewater
reasonably constant. However, very few treatment flowmeter should not be used to supply this control
plants are able to pmvide this degree of equalization. signal. The output ofthe flowmeter may be a 4-20 rna
Alternatively, the dose of chlorine or hypochlorite DC signal, or a vacuum or pressure pneumatic signal.
applied to the wastelwater may be made sufficiently This signal is then applied to either the orifice on the
high to ensure that the high flows with the greatest gas chlorinator or the metering pump on a hypo-
chlorine demands are adequately disinfected. This chlorinator to provide a constant dose.
will ensure that low flows, or flows of lesser chlorine
demand, will be at IElast adequately treated. While flow proportional control of either type ac-
counts for a major source of variation in chlorine dose
For high wastewater flows, it will be impractical to requirement, it fails to provide adjustment in chlorine
prOVide equalization or uneconomical to overdose dose with variation in wastewater chlorine demand.
with chlorine. In this case, the amount of chlorine For example, ifthe chlorine demand varies by a factor
must be varied in some manner so that the dose of 2 over the course of a day it will be necessary to
applied is just sufficient to provide the required overdose during times of low chlorine demand, even
degree of disinfection. This may be done manually, by if flow proportional control is used.
the operator, one or more times per day. The feed rate
(kg chlorine per hr) 'from the gas chlorinator (or the To circumvent this difficulty, it is, in principle, possible
metering pump rate, from the hypochlorite storage to measure the chlorine residual after the initial high .
tank) is adjusted in accordance with the flow and the chlorine demand has been exerted (2 to 5 minutes .
chlorine demand, usually to attain a target chlorine following rapid mix) and use an automatic measure-
residual. Typically, this adjustment may be made ment of this residual to control the chlorine applica-
once per shift. tion rate. This strategy, feedback residual control, is
generally not practiced alone, except when flow
For larger treatment plants, the savings on chlorine variations are negligible, dueto problems of hydraulic
utilization will justify some degree of automatic lags in the system.
control. In addition, for treatment plants using
dechlorination, auto,matic control may further be Instead the flow is most commonly measured using
justified by savings o.n dechlorination chemicals. an automatic flow transducer, and the residual after 2
to 5 minutes is measured using an automatic chlorine
The simplest form of automatic control that can be residual analyzer. These signals are then combined to
used is flow proportional control. In this case, the provide a constant residual. In this manner, the
Instantaneous chlorine application rate is maintained variation in flow is used as the major signal for dose
directly proportional to the instantaneous flow rate adjustment, and the variation in residual is used to
coming into the disinfection process, resulting in a fine tune the dose. This option is designated as
constant chlorine dosage. If the disinfection influent compound loop control.
flow Is controlled by cl pumping station upstream, the
chlorine application rate may be set by a signal from Two major types of compound loop control exist. One
the wastewater pumps to the chlorinator. This type, typified by many gas chlorinators such as those
modification is referred to as additive rate control. manufactured by Wallace and Tiernan Division of
Generally the signal from the pumps is a standard PennWalt Corporation, uses the flow signal to vary
4-20 ma DC signal (the typical electrical control the ,size of the gas orifice, and the residual signal
signal) and may be used to vary the. orifice opening on (modified by a square root module) to vary the
pressure differential across the orifice. The square Figure 5-23. Dissipation of chlorine residual and point of
root module takes the incoming signal and converts it sampling for control.
to its square root-this is done because the flow rate
is linear in the square root of pressure drop, and not in 1.0.-------------------,
press4re drop itself. The second type of compound Co =Dose
loop control, implemented in many chlorinators 1\
manufactured by Fischer and Porter and Capital Most Rapid Decay
Controls, electronically combines the two signals Completed in Co 1-5 min.
from the flow meter and the analyzer into a single
_~ Control Sample Should
signal to a chlorine valve, which may regulate either ~~ be Withdrawn Here.
the orifice area or the pressure differential across a Cleo
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 OL-----------------~--'
to produce the vacuum or pressure that drives the
control signal lines. In addition, direct digital com-
puter control, using either dedicated distributed or
system (175). For situations in which the rapid phase
centralized computers, is possible.
of chlorine demand is prolonged (5 to 15 minutes), a
Critical elements in the design of compound loop second chlorine analyzer is recommended at the far
control are the time lags inherent in the residual point to trim the signal from the first sampling
measurement system, and delay times between location.
application of tne chlorine residual signal to the
control device and the attainment of the control The analyzers used for residual may use either the
action. These time lags include those due to the amperometric, or the colorimetric or titrimetric DPD
sample transit time in the sampling lines, the lag time methods, as described above. In either case, manu-
between the ejector and the point of application, and facturers' designs of continuous analyzers incorpo-
the residence time in the contactor prior to sampling. rate in-line filters to minimize solids deposition. For
continuous analyzers, special reagent solutions are
The calculation ofthe lag time in the line between the used that differ somewhat from manual analytic
ejector and the point of application may be calculated techniques in that proprietary additions are incorpo-
as described above once the ejector is sized. The lag rated that are said to reduce fouling inherent in
time in analysis is ofthe order of 1-2 minutes. The lag continuous systems. The maintenance requirements
time in the sampling lines is set by the velocity andthe for continuous analyzers are higher than manual
length of the line. This velocity should be maintained analyzers, and at least daily standardization is
above 3 mls (1 0 ft/sec) to minimize fouling (157). By required to prevent drifting in the automated analyzer.
maintaining the analyzer as close as possible to the
chlorine contacf basin, this lag may be minimized. 5.7.5 Dechlorination
The lag time between the point of application and the From Equations 5-48 and 5-49, one mole of sulfur
sampling line should be set so that the rapid initial dioxide can dechlorinate one mole of either free
decay has been completed. This usually occurs 3 to 5 chlorine or combined chlorine. On a mass basis,
minutes following chlorine application (Figure 5-23). therefore, 64/71 grams of S02 are required to
dechlorinate one gram of chlorine, either in the free
Due to the existence of lags in the chlorine residual or the combined form. On a practical basis, about 1
sampling lines, it may be necessary to incorporate a gram of S02 is required per gram of chlorine.
delay timer on the residual control loop. This insures
that no additional adjustment in chlorine dose occurs Sulfur dioxide used for dechlorination may be sup-
.following a preceding adjustment until sufficient time plied in either 68 kg (150 Ib) cylinders or 1,365 kg
has elapsed for the wastewater receiving the new (3,000 Ib) containers. Railroad tank cars for bulk
chlorine dose to be analyzed for its response. This sulfur dioxide shipment are also available. The
timer should be adjustable and is ordinarily "tuned" discharge from sulfur dioxide containers may be
during start up of a treatment plant to minimize taken from the liquid or gaseous phases (106).
control instability.
For gaseous withdrawal, the maximum safe rate of
Roop presented a comprehensive review of waste- withdrawal on a continuous basis at 21C (70F) is
water chlorine control using the compound loop 0.9 kg (2Ib)/hr for a 68-kg (150-lb) container, or 11.4
kg (25 Ib)/hr for a "ton" container (106). To attain discharge of the effluents (1 09). If a contact chamber'
higher rates of gaseous withdrawal, it is acceptable to is used, the absence of a chlorine residual may allow
immerse the containers or cylinders in a liquid bath, for the growth of a microbial slime layer containing.
or to surround the containers with strip heaters; in coliforms, and possibly other organisms. While.
both cases, only the lower 50% of the container prevention of this slime growth does not appear to be
should be subject tlJ heating, and heating should be feasible, some regulatory a,uthorities (e.g., the State
limited to 325K (1 (6). As in the case of chlorine, the of California) permit the coliform standard to be
pressure at which sulfur dioxide is withdrawn should attained at the chlorine contactor effluent, despite an
be sufficiently below the saturation vapor pressure to increase in bacterial numbers during dechlorination
provide superheating and prevent reliquifaction (157). (112).

For liquid withdrawal, the maximum rate of removal Control of sulfur dioxide dechlorination is generally
is limited by hydraulic considerations to 135 kg (300 more difficult than chlorination. According to Chen
Ib)/hr) from either type of container (1 06), although it and Gan: "simple feedforward sulfur dioxide feed'
is possible to exceed this by pressurization, or control system(s) [Le., systems which rely solely on a
"padding" with dry air or nitrogen up to 515 kPa (60 measurement of influent volumetric flow rate] (are)
psig) (157). With liquid withdrawal systems, external inadequate for most dechlorination installations"
evaporators, of similar overall design to those used (112). This is due to the variation in the chlorine
for chlorine service, are required. residual at the outlet from the chlorine contact
It may be necessary to manifold multiple cylinders of
sulfur dioxide for supply to larger dechlorination Most commonly, control systems for dechlorination
systems. Considemtions for manifold design are consist of either straight flow proportional or feed-
similar to those USEld for gaseous or liquid chlorine forward control based on the product of influent flow
systems. and chlorine residual. In dechlorination installations,
it is necessary to maintain two automatic chlorine
Piping and materials used for chlorine service are residual analyzers-one 3 to 5 minutes from the point
generally satisfactory for sulfur dioxide service (157), of application of chlorine used for chlorine dose
although the physical systems used for chlorine control, and the second located at the contact'
should not be used for sulfur dioxide, or vice versa, chamber outlet, used for control of the sulfonator
prior to thorough Gleaning, to prevent potentially (112). Feedback control using measured chlorine
explosive reactions from occurring. In liquid piping residual following sulfur dioxide addition is not
systems, expansion chambers are required, as for currently feasible due to the instability of currently
liquid chlorine servit::e, to prevent hydrostatic expan- available analyzers measuring low (below 0.1 mg/I)
sion rupture from oc:curring (106). or zero chlorine residuals.
The gaseous sulfur dioxide exiting from the cylinder Two types of feedback control systems may be used to
or the evaporator miElY be dispersed into water using circumvent the problems associated with continuous
vacuum ejector syst,ems of similar design to chlorine measurement of low chlorine residuals. In the first
ejectors. Design curves for chlorine ejectors must be method, which may be described as split dechlorina-
corrected for different flow rates (on a mass basis) for tion, a flow proportional feedback chlorine residual
sulfur dioxide versus chlorine. This is done by system (Le. compound loop control) is used to produce
multiplying the specified flow rate for a chlorinator by an effluent that is dechlorinated to a low residual, but
0.95 (which repres.~nts the ratio of the density of well within the range of commercially available
sulfur dioxide to that of chlorine). For example, if a gas chlorine analyzers-this is usually 90 percent de-
chlorinator is designed to deliver 45 kg (1,000 Ib)/d chlorination of the incoming residual. A subsequent
chlorine at maximum span, it can deliver 43 kg (950 dechlorination step, using simple feedforward con-
Ib)/d of sulfur dioxide. trol, may then be used to produce a chlorine-free
The mixing at the point of application of sulfur dioxide
to the full wastewater flow should be conducted in an The second process uses a chlorine analyzer that is
intensive mannerto allow complete dechlorination of biased withothe addition of a constant flow of a side
all fluid elements. Injection into the center of a stream containing a constant chlorine concentration
turbulent fully-flowing pipe has proven satisfactory, (produced by a chlorinator or hypochlorite feed system
with complete mixing being attained 10 pipe diam- separate from that used for chlorination) to the
eters downstream of the injection point (176). sample being analyzed. The setpoint on this analyzer
is controlled to maintain the chlorine residual sub-
The kinetics of the reaction between sulfur dioxide sequent to dechlorination at the concentration im-
and chlorine are sufficiently rapid to preclude the posed by the biasing stream. In this case, classical
necessity for a sep~lrate contact chamber prior to compound loop systems may be used (112).

One modification of the biased feedback control For the conversion of free chlorine to combined
strategy has recently been devised (177). In this chlorine in wastewaters where the CI:N ratio is far
approach, the chlorine residual immediately prior to below breakpoint, a first order decay relationship may
dechlorination is measured (C pre). Equal flows of the be assumed:
influent to and effluent from the dechlorination
process are blended and the residual measured using (5-81 )
a second analyzer (Cmlxed). The chlorine residual
leaving the dechlorination basin may then be calcu- In Equation 5-81, Cf is the free chlorine concentration,
lated as 2Cmlxed-Cpre, and this value used as a control Co is the chlorine dose, and is the apparent first order
signal. decay constant, which may be given by (81):

This system has been in use in Metro Seattle and is (5-82)

reportedly successful at maintaining a residual below 1
0.067 mg/l. The principal disadvantage of this k (s-1) = 9.7 x10 C N Ko N [W]exp(-1510/T)
approach is the relatively higher complexity of the ([W] + KKA) ([H+] + KON )
control system, and the potential for biological slime
accumulation in the dechlorinated effluent sample In Equation 5-82, CN is the concentration (in Mil) of
lines; however, this system eliminates the need for a ammonia nitrogen, K A is the dissociation constant of
separate chlorinator used to bias the sulfonator HOCI (given by Equation 5-6), and K ON is the
analyzer. dissociation constant of the ammonium cation. T is
the Kelvin temperature. With an influent ammonia
In the absence of such control systems, it will be concentration of 10 mg/l as N (=0.7 mM), a mean
necessary to overdose the wastewater with sulfur temperature of 15C (288K), and an influent pH of
dioxide, which may result in deoxygenation of the 7.5, k 1 equals 193 s-1.
effluent. In this case, reaeration of the dechlorinated
wastewater may be necessary to produce compliance The resultant combined chlorine, once formed,
with an effluent DO requirement. In the presence of decays. It has been found that the kinetics of this
one of the suggested dechlorination control systems, decay may be described by two parallel first order
such' deoxygenation and reaeration requirements reactions (76):
have been found not to occur (112),
Cc = Co [x exp (-k2t) + (1 - x) exp (-k2't)] (5-83)
5.7.6 Design Example
The principles of design of chlorination and dechlor- In Equation 5-83, x, k 2, and k 2' are constants
ination systems will be illustrated by reference to the characteristic of the wastewater. For a variety of
common design example. A chlorination system is to wastewaters, it has been found that values of 0.3,
be designed to treat a non-nitrified activated sludge 1,67 x 10- 2 S-1 and 5 x 10-5 S-1, respectively are typical
effluent. The design flows are 7.5 mgd (mean) and 15 (76). Since the rates of these two decay reactions are
mgd (peak); initial flows are 3.5 mgd (mean) and 7.5 disparate, Equation 5-83 may be simplified to:
mgd (peak). The influent to the disinfection system is
expected to contain a geometric mean of 500,000
Cc = Co (1 - x) + Co x exp (-k2t) for t < 1/k2 (5-84)
fecal coliforms/100 ml, with a peak of 2,000,0001 Cc = Co (1 - x) exp (-k2't) for t > 1/k2 (5-85)
100 ml. The effluent coliform standards are a 30-day
geometric mean of 200 fecal coliforms/1 00 ml, and a With the particular constants assumed, Equations
7 -day geometric mean of 400/100 ml. Other effluent 5-81,5-84, and 5-85 may be written as the following,
parameters are a 30-day mean for both BODs and with t in minutes:
suspended solids of 15 mg/l and a maximum daily
value of 30 mg/l for each of these parameters. It is Cf = Co exp (-11 ,580t) (5-86)
also desired to dechlorinate such that the final
effluent chlorine residual is less than 0.05 mg/l. Cc = ,7C o + .3C o exp (-t) for t < 1 min (5-87)

Step f.-Determine the average desired fractional Cc = .7C o exp (-.003 t) for t > 1 min (5-88)
Step 3-The chlorine decay relationships may now be
From the conditions of this problem, the desired substituted into the kinetic relationships describing
fractional survival under average conditions is 2001 microbial inactivation.
500,000, or 4 x 10-4 This will be used as an initial
basis for design. Method 1: Using Equation 5-65, with parameters
from Aieta and Roberts (178), the following may be
Step 2-Characterize the decay of chlorine residual. written:
This must be done experimentally, or by resort to
information from the literature. NINo = {0.25Ctr2.S2 (5-89)

Using either Equation 5-87 (for times below 1 minute) mined which will satisfy the 30 day coliform standard
or Equation 5-88 (for times greater than 1 minute) to under average conditions. Based on this information,
substitute C as a func:tion oftime into Equation 5-89, a contact chamber will be sized. It is now necessary to
an equation relating contact time, chlorine dose, and determine, for a specific volume contact chamber, the
inactivation may be obtained. In other words, it may performance under various extreme conditions to
be shown that: ascertain the maximum chlorine dose likely to be
necessary. While this sensitivity analysis can be
(5-90) performed using either Method 1 or Method 2 (i.e.,
Equation 5-91 or 5-94), the approach will be illus-
NINo = (.175 Co t + .075 Co t exp (_t))-2.82 t < 1 min trated using Method 2 (Equation 5-94), which, based
NINo = (.175 Co t exp (_.003))-2.82 t > 1 min (5-91) on the computations in Step 3, for this situation,
appears to be more conservative.
Therefore, since N/N c, has been specified, if a value of
t is assumed, Co may be calculated. Hence, for times Based on step 3, it is anticipated that a 60-minute
of 10, 15, 30 and 60 minutes, Co (the chlorine dose) contact tank will be most suitable (obviously, the
equals 9.45, 6.39, 3.~14 and 1.83 mgll, respectively; following can be repeated with different initial
the effluent residual c:hlorine concentrations may be assumed tank sizes to generate economically opti-
determined from EquCltion 5-88 to be 6.42, 4.28,2.14 mized designs). For the desi~n mean flow, this would
and 1.07 mgll, respe,ctively. require a volume of 1,200 m (312,500 gal). Rounding
up, the design volume of 320,000 gallons is selected.
Method 2: From the Chick-Watson equation, assum- At the design mean 'flow and influent coliform
ing inactivation by fl"ee and combined chlorine is concentration, this would require a chlorine dose of
additive, the followin~1 may be written: 2.26 mg/l. However, at higher flow rates and coli-
forms, the chlorine dose must be increased if the
dN =-N(k,c,n, + kcCcnc) effluent is to be in compliance with the coliform
dt standard. Based on the assumed tank volume, at each
combination of chlorine dose, flow rate, and influent
The chlorine demand relationships may be substi- coliform concentration, the anticipated effluent coli-
tuted into this equation, and it can be integrated form and total residual chlorine concentration may be
numerically. computed. For example, at a dose of 3.7 mg/l, for the
given tank, the following are to be expected:
For t> k2, the result may be obtained as an analytical
Integral in the following form: Effluent
-In (NINo) = (k,Con,/k1ntl (1 - e-k,n,t) Effluent Chlorine
+ (kcCon c (1 - x) n e /k 2nc)* (5-93) Assumed Influent Coliforms Residual
(e -k;nJk, _ e -k;n.t) Flow Coliforms (#/100 ml) (mg/l)
For the inactivation data for E. coli at room temper" avg design avg 0.1 2.15
ature and relatively neutral pH, values for kf, k e , nf and max design avg 197.4 2.36
nc of 30, .085, 1.46, clnd 1.25, respectively, may be avg design max 0.6 2.15
assumed (Table 5-11). Using these, and the prior max design max 789.7 2.36
assumed constants, the following equation is ob-
tained: This is in compliance with the standards under all
conditions except for the combination of maximum
-In (NINo) = 2.86 x 10-3Co1.46 design flow and maximum influent coliform concen-
(5-94) trations. For design purposes, it may either be
+ 14.4Co1.25 _ 14.5Co1.25 e-0.00375t assumed that these two factors are unlikely to occur
coincidentally more than once in 7 days (and hence
Equation 5-94 relates NINo, Co, and t, and since the the 7-day standard will be complied with), or the dose
first parameter has beEm specified, given a value for t, necessary to produce compliance at this extreme may
the requisite value of Co may be calculated. For t be computed.
equals 10, 15, 30 and 60 minutes, Co equals 10.03,
6.92,3.85, and 2.26 rng/l, respectively. Using these At this point, the chlorinator(s) that is(are) to be
doses and contact times, from Equation 5-88 the specified must be capable ofa maximum outputof3.7
effluent chlorine residuals may be calculated to be mgll at the maximum design flow of 15 mgd, or 462
6.81, 4.63, 2.46, and 1.32 mg/l, respectively. Ibid. This is close to 500 Ibid, which is a commonly
available chlorinator size; to provide system redun-
Step 4-By two methods, suitable combinations of dancy, two such units would be specified. Alter-
residence time and chlorine dose have been deter- natively, the overall plant design might use two

parallelchlo.rination systems, each sized to half the Figure 6-24. Definition sketch for rectangular contactor.
design capacity; in this situation, two 250 Ibid units,
plus one spare 250 Ibid unit would be specified.
Step 5-Based upon a 320,000 gal volume, the
104m ....
_ 04_m 1
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. ------------------;14m

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 flow Js 389 m/3600 s = 0.11 m/s (21 from Equation 5-80 as 0.41 cm at average design
ft/min). This should be checked against the flow and 1.65 cm at peak design flow.
specific scour velocity of the particles expected to Step 6-Based on conditions at maximum design
be present to ensure minimum sludge deposition. flow, assuming peak chlorine dose of 3.7 mg/l, the
chlorine residual leaving the contact basin is esti-
The dispersion in this pfpe may be estimated
mated as 2.36 mg/l (see table under step 4). Thus, at
using Equation 5-77. The pipe Reynolds number
peak design flow, sulfonation capacity (assuming 1
is given by Re = VOR/u (where R is the density of
mg S02 per mg chlorine residual) of 15 mgd x 2.36
water and u is the viscosity). For turl::iulent flows,
mg/l x 8.34 = 295 Ibid is needed. From this
the friction factor 'f' is given by the following:
information, the sulfonator(s) can be specified, keep-
1/fo. 5 = 210g10(0/e) + 1.14 (5-95) ing in mind the desired redundancy.

In Equation 5-95, e is the absolute pipe roughness Step 7-Annual chemical requirements can be
(0.002 inches is typical for concrete). Thus f = estimated using average design conditions. As noted,
0.0094. Substituting in Equation 5-87 along with at mean design flow and coliform concentrations, a
the pipe diameter and the length, the dispersion chlorine dose of 2.18 mg/l would be satisfactory, and
number is determined as 4.9 x 10-5 , sufficiently would impart a residual exiting the contact chamber
low that the plug flow assumption is justifiable. of 1.27 mg/l (using Equation 5-88). Thus, the average
annual chlorine requirement under design conditions
The headloss may be estimated using Manning's would be 365 x 7 mgd x 2.18 mg/l x 8.34 = 46,300
equation. With an 'n' value of 0.015, the hydraulic Ib/yr (say 24 tons/year). The average annual re-
grade is calculated as 9.767 x 10-7 at mean quirement for sulfur dioxide under design conditions
design flow, and 2.127 x 10-7 at peak design flow. (assuming 1 mg S02/mg chlorine residual) would be
Therefore, the estimated headloss is 0.038 em 365x7 mgdx 1.27 mg/l x8.34=27,000 Ib/yr(say 14
and 0.151 em, respectively, under these two tons/yr). This information could then be used to
conditions. design the chemical handling facilities.

Option 2-Serpentine contact: If chlorine and sulfur dioxide are obtained in ton
containers, and if gaseous withdrawal is utilized, the
A rectangular baffled contactor with a depth of 2 average chemical consumption is 2.4 kg (5.3 Ib)/hr
m is used. The sketch in Figure 5-24 illustrates and 1.4 kg (3.1 Ib)/hr, respectively, for chlorine and
one configuration providing the required resi- sulfur dioxide. Both of these are substantially less
dence time. than the maximum safe rate of supply from ton
containers. Hence, there is no need to manifold more
The flovy path of the system is 40 x 7 = 280 m. The than two cylinders together (one in service and one
L/W ratio is 280/2.16 = 129. The velocity (at awaiting) for either chlorine or sulfur dioxide. It would
average design flow) between parallel baffles is be reasonable for the treatment plant to anticipate
0.076 mis, and through the baffle slots is 0.082 ordering 2-ton containers of chlorine and 1-ton
m/s. This should be checked against the specific container of sulfur dioxide every four weeks. A safe
scour velocity as in the previous option. From inventory would be 3 containers of chlorine and 2
Equation 5-78, the estimated 'd' value is 0.0047 containers of sulfur dioxide (this would provide for a
(for K = 1), again indicating reasonable approach 28-day reserve for transportation, plus excess for
to plug flow. Finally, the headloss is computed emergencies), and provision should be made for,

storage of 2 empty chlorine containers and 1 empty for plants with two chlorinators and feed rates up
sulfur dioxide container. to 180 kg (400 Ib)/d, 15 m 2 (160 sq ft) of area are
needed for the chlorinators; and
Step 8-A detailed c:ost estimate would be made at
this point, based on ,actual catalog costs, or prior bid for each extra chlorinator above two, an additional
data. For the purpose of estimating chemical con- 15 m 2 (160 sq ft) area should be provided.
sumption and annucll O&M costs during the period
prior to attainment 01: actual design flow, the chlorine In addition to the above areas needed for the
dose to achieve inactivation at various interim flows chlorinator and evaporator modules, space is also
can be computed (using Equation 5-94), and inte- required for the chlorine containers or cylinders being
grated over an estimated flow-duration curve. Addi- used to feed the system (or, in the case of sodium
tionally, other alternative design contactor volumes hypochlorite plants, the hypochlorite storage tanks)
may be assumed, anc! the design computations begun and space for inventory and empty cylinders or
again commencing with step 4. containers. The size of gaseous chlorine containers is
given in Table 5-14, and, in conjunction with esti-
5.7.7 Economics mates of required inventory, working supply, and
The estimation of chlorination, chlorine dioxide or storage of empties, can be used to develop site area
dechlorination costs is highly site specific. Particular estimates for chlorine storage.
concerns include the necessity for separate chlorine
contact basins (versus the use of an effluent channel Table 5-14. Physical Dimensions of Chlorine Gas Containers
as a contactor), the site-specific chemical costs, and Diameter Length
the necessary chlorine dosages. However, prelim- Capacity (inches)8 (inches)
inary rough estimatels of process alternatives may be 150 Ibs (70 kg) 10.25 to 10.75 53 to 56
developed from available literature data based on 2000 Ibs (910 kg) 30 79.75 to 82.5
field experience. 8150 Ib cylinders should be used and stored upright, thus requir-
ing approximately 0.6 ft2 (0.06 m 2) (plus separation area) per
Geisser et al. developed cost equations based on their cylinder, while ton containers are used and stored in the horizon-
studies on the disinfection of combined sewer tal position, requiring approximately 17ft2 (1.6 m 2) (plus separa-
tion area) per container.
overflows using either chlorine or chlorine dioxide
(179). Other useful sources of chlorination and
dechlorination cost delta are Chen and Gan (112), who The chlorine storage room (which mayor may not be
have developed estimates for sulfur dioxide dechlo- separate from the room containing the chlorinators)
rination'systems, and Gumerman et al. (180), who should be isolated from any other process by use of a
provide cost estimates for drinking water chlorination separate building, or by use of an isolated room
systems. bounded by fire resistant walls (20). If the chlorine
cylinders are physically separate from the chlorin-
The design engineer may also use standard estima- ators, and ifthe gaseous chlorine withdrawal method
tion manuals for the costing of contact basins and is used, it is necessary that the temperature of the
physical structures, plus manufacturers' quotations room containing the chlorine gas cylinders be kept
on equipment (chlorinators, mixers, etc.). below the temperature of the room containing the
chlorinator in order to prevent reliquifaction in the
5.8 Safety and Occupational Health . lines between the chlorine supply and the chlorin-
Considerations ators. For chlorine installations practicing liquid
Halogen disinfectants have certain properties that withdrawal from either ton or bulk containers,
must be considered at the design stage for the outdoor storage is acceptable. Climatic considera-
protection of operating personnel from risks that may tions will dictate necessary site specific details, such
arise during the plant life. Much detail is available in as provision for prevention of icing of valves and
regard to necessary salfetyfeatures to be incorporated shielding containers"from direct sunlight.
in wastewater disinfc3ction facilities using gaseous
chlorine or hypochlorite as disinfecting agents. Little The chlorine supply room must have at least 2 means
information is available in regard to adequate pre- of egress, with doors opening outwards from the
caution using other hi::JIogens. room. The structural designer must consider the dead
load of full chlorine containers in his calculations. If
5.8.1 Physical Site Layout ton containers are to be used, floor mounted trunions
For plants using gaseous chlorine, general space or scales (Figure 5-25) are needed. Furthermore, for
requirements may be estimated as follows (157): ton containers, the use of overhead crane of at least
2-ton capacity in conjunctionwith a special lifting bar
single chlorinator facilities of less than 200 Ib/d is needed to facilitate container movement, and
(90 kg/d) capacity require at lea~t 64 ft2 (6 m 2 ) of therefore the physical site layout must carefully
space fpr the chlorinator and ancillary equipment; consider ceiling heights. It is highly desirable that the

Figure 5-25. Ton container mounting trunions. (Courtesy chlorine leakage to occur from a cylinder, a valve, or
Force Flow Equipment.) piping. The threshold limit value (TLV) for worker
exposure to chlorine in air is 1 ppm by volume as an
Storage 8-hr time weighted average (181 ). Other atmospheric
Roller chlorine concentrations of interest are presented in
Trunnions For
V2" Dia. Table 5-15 (20). It is particularly noteworthy that the
Length = 21" Bolt 1+----19" 'minimum chlorine concentration detectable by odor
Capacity = 4"
= ~
4000 l b ' ~.
is greaterthan the above specified TLV. Therefore, in
order to provide for the continuous sensing of
Container " ,
chlorine leakage, it is necessary to rely upon some
Weight '= 15 Ibs each ~ L..--- -,-,- -l chemical or electronic device. These may be of
ASTM A7 Steel 3" Dia. Bushed Beam several types.

Epoxy FiniS~C: 'I:j':~~=~=;====1=3="~====-i=-~l

Table 5-15. Gas Phase Chlorine Concentrations Evoking
Specific Effects (20)
Response Concentration (ppm v/v)
Minimum odor threshold 3.5
Minimum 1-hour no serious effect
level 4.0
Length = 21" For .
Width = 4" V2" Dia'~1=====2~1!..:":===;::=~'1 Throat irritation 15.1
Capacity = 4000 ~b 'Bolt~,
Contamer .
w.-rr Coughing 30.2
Weight = 25 Ibs each . --------'-t- 30-minute, 1-hour danger level 40 to 60
ASTM A7 Steel 19" ---..-I"
Epoxy Finished ,---13"---+1,
L.ljlb rbi l
The older type of continuous chlorine sensors rely
chlorine container(s) in use be placed on a scale or upon a version of the iodometric chlorine detection
. load cell to have a positive record of remaining procedure, using starch/iodide or other chemically
disinfection. impregnated paper and measuring the change in
color that results when vapor containing chlorine
Chlorine cylinders of 68-kg (150-lb) oapacity are passes over this material. A more modern version of
emptied in a vertical position, secured to a wall or a this principle is used in a sensor that measures the
sturdy upright by means of a band or chain clamp to current required to electrochemically reduce the
the upper portion of each cylinder. chlorine present in a gas (Mine Safety Appliances,
Pittsburgh, PAl. A third principle used in continuous
In the chlorine storage room, ventilation must be such chlorine gas analyzers uses the change in electrical
to assure a complete air change in 1-4 minutes. conductivity of a gas that occurs as chlorine concen-
Generally, this is achieved by use of an exhaust fan trations increase (International Sensor Technology,
near the floor of an outside wall (20); such fan must Irvine, CAl. '
have switch outside the chlorine room itself as well
as inside. No ductwork, shafts, or other potential
sources of gas travel should exist between the Once the vapor phase analyzer has signaled the onset
chlorine supply room and any other portion of the of a chlorine leak, it is necessary to determine the
plant. The exhaust fan should vent at a level above cause of such event and proceed towards its repair.
neighboring buildings, trees, etc., to afford high Personnel entering the room in which the ambient
diluti.on of any contained contaminant gases with the chlorine levels have exceeded the alarm limit should
atmosphere. be clothed in protective equipment (gloves; breathing
device, suit). Using portable continuous chlorine
No turpentine, ether, ammonia (except for small sensors, the location of highest chlorine concentra-
amounts associated with leak detection equipment), 'tion may be found, and thus the site of the leak
finely divided metals, or other flammable materials inferred. Alternatively, it is possible to locate the site
are permissible within the chlorine storage room (20). of a gaseous chlorilJe leak by spraying a solution or
the vapors of a solution of strong ammonia (com-
5.8.2 Leak Detection mercial, not household strength). A white cloud of
In gaseous chlorine installations, the major safety NH 4 CI precipitate will form in the region where the
and health concerns involve the possibility for chlorine leak exists (181 ).

5.8.3 Emergency Rl9sponses of 68-kg (150-lb) and 910-kg (2,000-lb) chlorine
AI/ breathing devices, used for personal protection in containers. In the event of fire in association with a
chlorine instal/atiom; must be U.S. Bureau of Mines chlorine leak, no water should be used, since this will
approved, and should be sterilized after use if it is increase the corrosivity of the mixture. Non-water
anticipated that thc3y may be used by another based extinguishers should be used until the chlorine
individual. The "buddy" system should be employed leak is stopped.
in situations where it is necessary to go into a room
containing high vapor concentrations of chlorine.
Table 5-16. Neutralization Requirements for Chlorine
Approved breathing devices may be categorized as Containers
follows (20): Container 100 Percent Hydrated
Capacity Caustic Soda Ash Lime
Industrial Canister Mask ("gas masks"). This is only 150 Ib 1881b 450lb 1881b
suitable for atmospheric chlorine concentrations in 60 gallon in 150 gallon in 188 gallon
below 1 percent and clxygen concentrations above 16
percent, and only for Bhort duration use. It should not 2000lb 2500lb 6000lb 2500lb
in 800 gallon in 2000 gallon in 2500 gallon
be relied upon for use during leaks, and rigid
adherence to canistel replacement time is required.

Self-Contained Breathing Apparatus. These may use Should any persons come in contact, either by
a portable cylinder of oxygen or air, or a chemical inhalation, or eye or skin contact, with chlorine
generation system for continuous production of resulting from a leak, the following first aid measures
oxygen. For the latter type of system, it is necessary to may be taken, prior to consultation with a physician
wait until oxygen production has commenced prior to (156):
entry into a contaminated area. No oxygen system
should be used for entry into closely confined areas General. Remove the person to an uncontaminated
(storage tanks) or whlare the danger of sparks or fire area, and remove contaminated clothing, washing
exists. any parts of the body exposed to chlorine with water.

Positive Pressure Blower Mask. A hose is used to Inhalation. If breathing has ceased, commence
supply breathable air 'from either a remote air tank or artificial respiration. When breathing recommences,
a remote compressor. Use is permissible only if or if breathing has not stopped, administer oxygen.
immediate safe escapl9 from the contaminated area is Keep the person warm and at rest.
possible in the event of a failure in the air supply
system. If a compraSS(lr is used, its intake should be at Eye Contact. The eye shou Id be fl ushed with water for
least 6 ft (1.8 m) above grade to prevent inadvertent 15 minutes, holding the eyelids apart to get complete
contamination with chlorine emanating from the area irrigation. In the design stage, eyewash basins should
under investigation. be provided.

Personnel should be trained in the use of such Skin Contact. Wash the eJ<posed parts with soap and
breathing devices, and, in particular, in the location water. It would be desirable to provide in the facility
and management of chlorine leaks while using such design provision for an emergency shower. Installa-
protective equipment. tions using sulfur dioxide should adhere to safety
precautions regarding handling of a nature similar to
When a chlorine leak is located, if possible, the those used in chlorine handling. At the design stage,
chlorine cylinder or container should be turned so it is particularly useful to specify fittings, valves, etc.,
that the defect is shifted at the upper portion of the for S02 service that are compatible with Chlorine
container. This will assure that gaseous (rather than Institute emergency kits.
liqUid) chlorine leakage occurs, and, eventual/y, the
leak may be self-limited by evaporative cooling
described earlier. All facilities handling gaseous 5.9 Operation and Maintenance (O&M)
chlorine should have ~Ivailable the appropriate Chlo- Requirements
rine Institute emergency kit, and be familiar with its In the start-up of chlorination systems initially, or
use. Devices in this kit may be used to patch or repair after a shut-down, a careful procedure must be
the leaking container, valve, or line. followed to minimize the likelihood of chlorine leaks.
Sepp and White outline the following procedure
If the gaseous or liquid chlorine that leaks can be when gaseous chlorine is used (157):
collected, it is possible to neutralize it using alkaline
solutions. Table 5-16 describes the required chem- 1. Check all joints for proper gasketing and check
Icals and water necessary to neutralize the contents that all supply valves are closed;

2. Check ejector for proper vacuum; 4. Evaporator vessels should be inspected for
sludge accumulation every year or each 200
3. Place automatic chlorinators in manual mode, tons of chlorine. If superheating capability
and set feed rate for about 25 percent of declines, this may be indicative of either sludge
maximum; accumulation or heater failure. Piping and
connections to the evaporator should be in-
4. Open one chlorine cylinder slightly, and check spected every six months. Sacrificial anodes
all joints for leaks. If no leaks are found the used for corrosion control in the evaporator
remaining number of required cylinders can be water bath should be inspected, and, if neces-
opened; sary, replaced, every six months.
5. If a leak occurs, immediately close the cylinder 5. The sludge accumulation in the evaporator may
valve, open all other valves fully and increase be cleaned by flushing with cold water until the
the chlorine feed setting on the chlorinator to its effluent water runs clear. At this time, the
maximum value. When all chlorine has been vessel can also be inspected for pitting, and, if
ejected, repair the leak, and return to step (1); this is severe, replaced. Before placing the
evaporator back into service, it must be dried by
, 6. Check performance at maximum rated capacity.
holding the external water temperature at 82C
The following are likely causes of capacity loss:
(180F) and maintaining a vacuum of 25 inches
a. insufficient ejector vacuum, of mercury (85 kPa) for 24 hours.
b. vacuum line leaks,
6. The chlorine gas filter should be inspected every
c. insufficient feed water pressure,
six months, at which time the filter element
d. solution line friction losses too high due to
should be replaced. The sedimenttrap should be
too small pipe diameter,
washed and dried at this time, and the lead
e. air or gas binding in solution line.
gaskets disposed and replaced.
The following modifications to the above must be
7. The chlorine pressure reducing valves may be
made when the liquid chlorine withdrawal method is
cleaned of any deposits with a soft cloth, or, in
used (157):
more severe cases, with isopropyl alcohol or
1. Dry the entire system by heating water in the trichloroethylene. The valve spring should be
evaporator bath and passing dry air (-40 0 Cdew replaced every two to five years.
point) through the evaporator and all supply
lines from the cylinders to the chlorinators. The chlorinator system itself is subject to the fol-
Several hours may be required for this step. lowing maintenance requirements (157):

2. Bring the evaporator water to the design tem- 1. The ejectors should be disassembled and
perature. Start the ejector water flow. Start up cleaned every six months, and iron and man-'
system with gas phase withdrawal to check ganese deposits removed with muriatic acid.
leaks as above. If no leaks are found, then liquid
withdrawal may be commenced. . 2. Booster pumps are subject to similar main-
tenance requirements as any other pump.
The maintenance requirements for the chlorine
supply system are as follows (157): 3. The chlorinator rota meters and floats should be
removed and cleaned every six months, and the
1. Flexible copper tubing reliability must be per- metering orifice inspected. All valve stems and
iodically checked. This can be done by bending seats should be cleaned and inspected once a
the lines slightly. Any "screeching" is ilidicative year, and all valve springs should be replaced
of corrosion, and the tubing must be replaced. every two years.

2. Minute leakage from lines and fittings, undetect- The chlorine analyzer is probably the single system
able by odor or standard leak detection pro- component requiring most careful attention. The
cedures, may be ascertained by inspecting for following operation and maintenance requirements
signs of moisture accumulation or metal dis- , are essential (157):
coloration, both of which are signs of incipient
leak development. 1. Sample lines must be inspected daily for solids
accumulation and filters and screens cleaned
3. In some cases, a one inch flat file may be used to daily. This is particularly important in lines from
reface chlorine cylinder valves to ensure more dechlorination systems that lack residual, which
precise seating. could reduce line fouiing.

2. In amperometric sy~;tems, electrodes should be of 20 fecal coliform/100 ml (enforced year round).
cleaned weekly or biweekly, or more frequently The bacteriological standard is enforced at the
if erratic or drifting I::alibrations occur. chlorine contact tank effluent. In addition, a minimum
dissolved oxygen requirement of 6.0 mg/l is enforced
3. The effluent line pH should be checked daily at the final effluent.
with pH paper. If it i~; not within the range of4.5
to 5.0, a stronger buffer solution may be The treatment system consists of primary and pre-
required. liminary treatment, activated sludge, separate stage
biological nitrification, multi-media filtration, chlo-
4. Periodically, preferclbly at least daily, the ana- rination, dechlorination with sulfur dioxide, and
lyzer reading should be compared with the reaeration.
results from a manual analysis. A control chart
graphing the difference in results as a function The filtered, nitrified effluent routinely contains less
of day should be maintained. If there is a than 1 mg/l suspended solids, and 1-3 mg/l total
consistent deviation, the span control on the Kjeldahl nitrogen (occasional spikes of 10-20 mg/l
analyzer may be adjusted on a weekly basis. A occur). Nitrites have not been found in this effluent.
trend to increasing deviation may be indicative
of instrument malfunction or sample line or The disinfectant used is gaseous chlorine, supplied in
filter fouling. ton cylinders. Chlorine usage is 6.5 kg/h (350 Ibid),
on average (average dose of 8.4 mg/l) and two
6. The adequacy of KI i3ddition may be checked by cylinders may be manifolded for withdrawal. The
adding slightly more KI than recommended to chlorine supply system was originally designed for
the feed buffer solution. If there is a change in liquid phase withdrawal, and the plant has evapora-
cell response, this is indicative of insufficient KI. tors; however, due to the low hydraulic loading, gas
Continue increasin~1 the KI concentration until withdrawal has been found to be satisfactory, and no
no further change in the analyzer output occurs. problems with reliquifaction have been noted. Ap-
proximately 1 cylinder in 30 is defective, and an
Culp and Heim present a useful guide to trouble- inventory of about 6 cylinders is maintained, with
shooting of chlorination systems (182). This is new orders received when only two full cylinders
reproduced in Table 5-17. remain (this is about a 10 day supply).

5.10 Case Studies The chlorine solution is dispensed through solution

In this section, several treatment plants that use feed chlorinators. The plant was originally designed
chlorination and dechlorination will be described, to operate on a residual control mode; however, the
indicating their design dotails, operating experience, operators noted that the analyzers were difficult to
and features oftheirfacilitythat influence their ability maintain, and that the control system was unstable
to achieve satisfactory porformance. (possibly due to a 200-500 ft length of sampling line,
or to the location of the point of sampling prior to the
5.10.1 Stony Brook Re~1ionalSewerage Authority end ofthe rapid decay phase of chlorine demand). The
Treatment Plant #1, PriJl"Jceton, NJ chlorination process is currently operated in manual
This treatment plant is located in Princeton, New mode, with residual monitored 3 times/shift at the
Jersey and services a metropolitan area consisting of end of the chlorine contact chamber. A residual of
the Borough and Township of Princeton, West 1.5-2 mg/l is generally used as a control point (some
Windsor Township, Hopewell Borough and Township, of this residual is free).
Pennington Borough, and South Brunswick Town-
ship, NJ. Wastewater is collected in separate sewers, The chlorine solution is applied at a separate stilling
and the service area is essentially 100 percent well through submerged nozzles. Two parallel, baf-
domestic. Within the service area is the campus of fled (3 baffles of the end-around type) contact
Princeton University. chambers are used. The contact time at design
average flow is 18 minutes, making the current
The current treatment pl;3nt was completed in 1977. contacttime at average flow about40 minutes. Slime
The average daily design flow is 11 mgd (482 IIs), and accumulation has occurred, and the contact tanks
the current average daily flow is 5 mgd (219 lis). The must be cleaned by high pressure water hoses every
discharge is to the Millstone River, a tributary to the 3-4 months. Some difficulty in cleaning these tanks
Raritan River. due to lack of provision for isolated clean outs for
wasting of the loosened debris has been noted.
Permit requirements include a BOD of 8 mg/l, a 2
mg/I ammonia nitrogen requirement, a suspended The dechlorination system uses gaseous sulfur
solids requirement of 10 mg/l, a maximum chlotine dioxide, also supplied as "ton" containers, with gas
residual of 0.06 mg/I, and a bacteriological standard withdrawal. The S02 utilization rate is 89 kg (197
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.
lii) 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 necessalY.
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.
lii) increase velocity in sample line to analyzer cell.
(iii) move cell closer to sample point.
liv) move sample point closer to point of application.
C. Check that sufficient KI is being added for the anticipated amount of chlorine residual, and increase KI 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
G. Check that the rotameter tube on the chlorinator is in the proper range, and replace, if necessary.

Ib)/d. Based on a 2 mg/I chlorine residual, this is 2.4 5-day median total coliform requirement of 240/1 00
mg S02/mg C12. Thl~ sulfonator was designed for ml (with a geometric mean coliform requirement over
manual control and is controlled by monitoring the 30 days of 200 fecal coliform/1 00 011). The bacterio-
influent chlorine residual 3 times/shift. No particular logical requirements are met at the chlorine contact
problems with the sulfonator system were noted. chamber exit, although problems with high counts
after dechlorination have been noted. The instan-
There is no data available on the dissolved oxygen at taneous maximum chiorine residual is 0 mg/1.
the effluent from the sulfonator mixing point; how-
ever, the plant was designed with reaeration due to Treatment consists of preliminary and primary treat-
the effluent DO permit condition. The operating ment, closed tank oxygen activated sludge (non-
personnel only use the aerators sporadically, and nitrifying), chlorination and dechlorination. Chlorina-
believe that the necessity for reaeration stems from tion is achieved using 14 to 15 percent sodium
the oxygen demand from nitrification rather than hypochlorite stored in fixed storage tanks. An apprOlc-
from any demand from sulfur dioxide. imate six day storage capacity exists. Chlorine dose is
generally12 mg/1.
Both the chlorine and sulfur dioxide containers are
stored in a single room isolated from the other rooms While compound loop control hypochlorinators are
in the plant. Temperclture control and ventilation is installed, they have not been used, since it is believed
provided. Necessary si3fety equipment, including self- (by the operators) that before the secondary treatment
contained breathing systems, and leak detectors are process went on line, the use of primary effluent
on-hand. The two cylinders each of sulfur dioxide and fouled up the analyzers. Additionally, the automatic
chlorine which are in use are mounted on hydraulic control system was originally designed for gaseous
load scales. The operating personnel feel that the chlorine service, and only mid-way through construc-
necessary handling and safety features appropriate tion were modifications made to switch to hypo-
to chlorine are also appropriate to sulfur dioxide. No chlorite service. Due to more pressing operational
corrosion or leak pro1blems, other than occasional difficulties, the shakedown of the automatic disinfec-
defective cylinders, h~lve been noted. tion control system has not yet been accomplished.
Flow-paced manual chlorine control is practiced in
Due to the substantial overcapacity, no significant which the residual at the end of the contact cha"mber
down times have b,~en recorded, except for an is measured every hour, and feed rate adjusted to
Instance of failure of the sulfonator due to water maintain a 6-8 mg/I residual in dry weather and 3-5
intrusion. Spare vacuum regulator parts and valve mg/l in wet weather.
springs are maintained. However, there is no periodic
maintenance program. The hypochlorite is supplied to diffusers, and mixing
at the point of application is supplied by two 30-hp
The operating personnel estimate that 3 hr/d of (22-kW) turbine mixers. The chlorine contact system
operator time are required for the manual sampling consists of an underground channel (2 identical
and dose adjustments for the chlorinators and parallel channels, each 680 01 (2,200 ft) long, 3m (1 0
sulfonators, and aboutO.5 hr/wk for cylinder switch- ft) high and 2.8 01 (9 ft) wide) providing 50 minutes
overs. Some complaints as tothe lack of manufacturer contact at85 mgd(3,700 I/s). No solids accumulation
training for maintenance and troubleshooting of the or sliming problems have been noted.
chlorinators and sulfonators were voiced.
The dechlorination system uses sodium bisulfite
The 1983 annual chemical costs for the chlorination- supplied as a 22 percent (as 80 2 ) solution. Tankage
dechlorination system were $14,000 for the chlorine for 1 week's supply exists. The dose of dechlorinating
($0.24/kg) and $6,50() for sulfur dioxide ($0.20/kg). chemical is also controlled in a flow-paced manual
5.10.2 Southeast Wa;ter Pollution Control Plant, mode after hourly analyses just after the point of
City and County of S.'.ln Francisco, CA sulfite addition to provide about a 1-2 mg/I excess
This treatment plant i~j located in San Francisco, CA dose. Mixing at the point of application is provided
and services a portion ofthe combined sewer system using two 22-kW (30-hp) mixers.
of the City and County of San Francisco. The flow is
predominantly domestic. The plant appears to have substantial problems in
achieving automatic chlorine control. However, these
The current treatment plant was completed in 1981. are traceable to the initial use of the chlorination
Design average flow for secondary treatment is 85 equipment for treatment of primary effluent, and the
mgd (3,700 I/s) with a peak flow of 210 mgd (9,200 need to modify the disinfection/dechlorination sys-
1/5). Current flows are 20-200 mgd (875-8,750 lis). tem from gaseous chlorine and sulfur dioxide to
Discharge is to San Francisco Bay. Permit require- sodium hypochlorite and sodium bisulfite late in the
ments are monthly aVElrage BOD/55 of 30/30 and a design and construction process.
5.10.3 Sacramento Regional Wastewater chlorinators. Chlorinators are controlled using a
Treatment Plant, Sacramento, CA compound loop with residual trim algorithm. The
This treatment plant is located in Sacramento County, residual at the end of the contact system is main-
CA and services a metropolitan area including the tained at 9 mg/l.
City and County of Sacramento and the City of
Folsom. Except for a small portion of the downtown The chlorine solution is applied through diffusers at a
point where two two-speed 36/72 kW (35/70 hpj

Sacramento area, the facility services a separate
sewer collection system. About 1 percent of the flow
is due to industrial sources, primarily canneries,
turbine mixers are located. The initial residual
analyzer sampling point is located a short distance
which contribute a high BOD, low nitrogen waste- (probably less than 1 minute) from this point of
water. mixing. The contact system consists of a 3,000 m
(10,000 ft) pipe 2.6 m (8.5 ft) in diameter, that has
The current treatment plant was completed in 1982. never been inspected (in approximately 1 year of
The design flow is 136 mgd (6,000 I/s) (daily average) service) for slime or solids accumulation; however it
. with a design hourly peak flow of 240 mgd (1 0,500 has not appeared to be a problem. The second
lis). The current daily average flow is 130 mgd (5,700 analyzer, providing the residual trim signal, is located
I/s), with a peak hourly flow of 240 mgd (1 0,500 lis). at the end of this pipe.
The discharge is to the Sacramento River.
The dechlorination system uses gaseous sulfur
Permit requirements include a BOD monthly average dioxide. The S02 is supplied either by truck or by rail
of 30 mg/l. 30 mg/l monthly average suspended and off loaded into a fixed 150 ton storage tank. The
solids, monthly average total coliform MPN of 23/100 fixed storage tank was modified such that its fittings
ml, with a daily maximum of 500/1 00 ml. a maximum are compatible with the Chlorine Institute emergency
chlorine residual of 0.1 mg/l. The bacteriological kit specifications. There is no load scale on the
standard applies at any location at the choice of the storage tank, and inventory is monitored using the
facility, and is- currently achieved at the chlorine control computer system. Withdrawal of liquid S02 is
contact tank effluent. accomplished under compressed air padding, and the
compressor system is separate from that used for the
The treatment system consists of primary and pre-
chlorine withdrawal system. Sulfur dioxide dose is
liminary treatment, closed tank oxygen activated
typically 1 mg/mg chlorine residual, plus an excess of
sludge (non-nitrifying, although an average of 0.4
mg/I nitrite is present in the secondary effluent), 2 mg/l for a daily use of 5,450 kg (12,000 Ib).
chlorination and dechlorination.
The sulfur dioxide is fed to evaporators that supply
The secondary effluent contains an average of 12 sulfonators. Gaseous sulfur dioxide under vacuum is
mg/l suspended solids, 9 mg/l ammonia, and 22 then piped 3,000 m (10,000 ft) to the end of the
mg/l total Kjeldahl nitrogen. chlorine outfall, where it is mixed in ejectors and
then, using four circumferentially-mounted pipe
Gaseous chlorine is supplied in 90 ton tank cars diffusers, injected to the wastewater flow. The control
delivered by rail. The chlorine dose is 13 mg/l for an system is compound loop using a biased chlorine
average daily use of260 kg/hr{14,000 Ib)/d (giving a residual signal using a hypochlorite biasing system.
12-day life for a tank car). Two tank cars are located The set point is at a 2 mg/l sulfur dioxide excess,
on site, with one in service and one full tank car partially due to the requirement for maintenance of
awaiting use, or on order. Standby equipment to use an undetectable chlorine residual, and partially due to
ton containers also exists. The tank car loading the inability to achieve precise sulfonation control in
facilities are located away from the chlorinator the long vacuum line (a lag of 5 minutes in the S02
building, and liquid chlorine is withdrawn into an vacuum supply line was cited). No problems with
intermediate storage tank with a level sensor. The dissolved oxygen depression or pH were cited,
drop in the liquid level sensor in the storage tank is although continuous pH sensors are used.
used as a signal for switching to another tank car. No
load scales exist on the tank cars and this is No problems with slime growth or recontamination
recognized as a design deficiency-the plant com- subsequent to dechlorination were expressed, al-
puter control system maintains a cumulative log of though the total coliforms measured after dechlor-
chlorine withdrawals from the tank car as a means of ination are somewhat erratic.
determining when the tank car is nearly exhausted.
Withdrawal of chlorine is accomplished with a The plant has been fully operational for only about
compressed air padding system. one year, so evaluations of corrosion and machine
reliability are difficult. Some problems noted to date
From immediate storage, chlorine is fed to evap- are the corrosion of the trim chlorine analyzers from
orators, and the gaseous chlorine is then fed to the chlorine vapors emanating at the end of the
contact pipe, the breakdown of PVC solution in 1. Belohlav, L.R. and McBee, E.T. Discovery and
sunlight (painting of PVC is suggested), and the lag Early Work. In: Chlorine: Its Manufacture,
time in the sulfonation control system noted above. Properties and Use. ACS Monograph #154,
Reinhold Pub!. Corp., New York, NY, 1962.
The dechlorination analysis system is cleaned and
calibrated twice daily, and it is estimated that total 2. Baker, J.C. Use of Chlorine in The Treatment of
operator time required for operation of the chlorina- Sewage. Surveyor 69, 241, 1926.
tion/dechlorination system is 4 hr/d. Table 5-18
summarizes the overall O&M schedule for the major 3. Averill, C. Facts Regarding the Disinfecting
components of the chlorination-dechlorination sys- Powers of Chlorine. Letter to Hon. J.1. Degraff,
tem at Sacramento. Mayor of the City of Schenectady. SS Riggs
Printer, Schenectady, NY, 1832.
5.11 References
When an NTIS number is cited in a reference, that 4. Gascoigne, G.B. Chlorination of Sewage and
reference is available from: Sewage Plant Effluents. Sewage Works Jour-
nal 3:38-49, 1931.
National Technical Information Service
6286 Port Royal Road 5. Laubusch, E.J. State Practices in Sewage
Springfield, VA 22161 Disinfection. Sewage and Industrial Wastes
(703) 487-4650 30(10):1233-1240, 1958.

Table 518. O&M Sched:ale, Sacramento Regional Wastewater Treatment Plant

Component Action
CI 2 leak indicator Adjust sample flow rates (M)

CI2 pressure reducing shutc,ff Remove and clean valve seats (SA)

S02 ejector Check and clean ( if necessary) (M).

NaOCI biasing pump Check and adjust V-belts (0). Overhaul (A).

Chlorine Unloading platform 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).

Air padding system Check oil level, air intake, drain condensate trap, separator (W). Tighten bolts (M). Check
pressure reducing valves (M). Change oil frame oil (0). Clean cooling coil (0). Inspect intake
air cleaner and valves (0). Check-motor lubrication (A).

Chlorine room exhaust fan Lubricate fan bearings (M). Check V-belt tension (M). Check fan wheel (0). Check fan motor
and bearings, V-belt alignment (SA). .

Air padding dryer 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 (0). Change air
filters (SA). Inspect exhaust mufflers (SA). Inspect dessicant, check valve seats, and
solenoid valves (A).

Chlorinators and sulfonatow Remove organic residues with wood alcohol and inorganic residues with hydrochloric acid

CI2 and S02 emergency Visual inspection (A).

expansion tanks

Flash mixers Check motors (M).

Sulfonator ejector Check motor condition and lubrication, pumps packings, and pump lubrication and condi-
tipn (M).

Hypochlorite mix tanks Clean (M).

Chlorine analyzers 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).

Chlorine evaporators Clean and inspect vaporization chamber (A).

(0) "" daily; (W) == weekly; (11.1) == monthly; (0) = quarterly; (SA) = semiannually; (A) = annually.
6. Phelps, E.B. Disinfection of Sewage and 21. Miller, G.W., et al. An Assessment of Ozone
Sewage Filter Effluents with a Chapter on the and Chlorine Dioxide for Treatment of Munic-
Putrescibility and Stability of Sewage Efflu- ipal Water Supplies. EPA 600/8-78-018, NTIS
ents. U.S. Geol. Surv. Water Supply Paper, No. PB-288196, U.S. Environmental Protec-
229,1909. tion Agency, Cincinnati, OH, 1978.

7. Kellerman, K.E., et al. The Disinfection of 22. Rapson, W.H. From-Laboratory Curiosity to
Sewage Effluents for the Protection of Public Heavy Chemical. Chemistry Can. 18(1 ):25-31,
Water Supplies. U.S. Dep. Agr., Bur. Plant Ind., 1966.
Bulletin #115,1907.
23. Aston, R.N. and J.F. Synan. Chlorine Dioxide
8. Rideal, S. Application of Electrolytic Chlorine as a Bactericide In Waterworks Operation.
to Sewage Purification and Deodorization in Journal New England Water Works Assoc.
the Dry Chlorine Process. Trans. Faraday 62:80-94, 1948.
Society 4:179-206, 1908.
24. White, G.C. Disinfection of Wastewater and
9. Disinfection of Sewage and Sewage Filter Waterfor Reuse. Van Nostrand Reinhold, New
Effluent. Engineering Record 67:14, 1913. York, NY, 1978.

10. Hooker, AH. Chloride of Lime in Sanitation. 25. McCarthy, J.A Brand CIO as Water Disin-
John Wiley & Sons, New York, NY, 1913. fectants. Journal New England Water Works
Assoc. 58:55-68, 1944.
11. Phelps, E.B. The Chemical Disinfection of 26. Wyss, O. and Stockton, J.R. The Germicidal
Sewage. American Journal Public Health, Action of Bromine. Arch. Biochem.12:267,
2:72-86, 1912. 1947.
12. Nikirk, F.A Disinfection of Sewages and its 27. Johannesson, J.K. Bromamines. Part I. Mono-
Success in a Small City. Municipal Engineer- and Di-bromamine. Journal Chemical Society
ing 46:478,1914. 2998-3001, 1959.
13. Faber, H.A How Modern Chlorination Started 28. Johannesson, J.K. Anomalous Bactericidal
The Story of the Solution Feed Process as it Action of Bromamine. Nature 181 :1799-1800,
Began Forty Years Ago. Water and Sewage 1958.
Works 99:45-50, 1952.
29. Mattraw, H.C., et al. Equilibrium Constant,
14. Wigley, C.G. Disinfection of Sewage. Munic- Infrared Spectrum, and Thermodynamic Prop-
ipal Journal Public Works. 47:292-293, 1919. erties of Bromine Chloride. Journal Chemistry
15. Chlorination of Sewage. Municipal Journal Phys. 22:1117-1119, 1954.
46:266, 1919 30. Mills, J.F. and Schneider, .J.A Bromine Chlo-
16. Editorial. Is Chlorination Effective Against all ride: An Alternative to Bromine. Ind. Eng.
Waterborne Disease? JAMA 78:283, 1922. Chem. Prod.. Res. Devel. 12(3):160-165,1973.

17. Ellms, J.W. and Pond, G.T. Sewage Disinfec- 31. Mills, J.F. and Oakes, B.D. Bromine Chloride,
tion. Municipal Sanitation, 1:266-268, 1930. Less Corrosive Than Bromine. Chemical Eng.
Aug. 6, 1973,pp. 102-106.
18. Tiedeman, W.O. Efficiency of Chlorinating 32.. Mills, J.F. Disinfection of Sewage by Chloro-
Sewage Tanl< Effluent. Engineering News- bromination. ACS DivWater Air Waste Chem-
Record 98, 1927. istry 13(1):106,1973.
19. Porges, R. United States Sewage Treatment 33. LeBlanc, N.E., et al. Disinfection Efficiency and
Practices During the Early Twentieth Century. Relative Toxicity of Chlorine and Bromine
Sewage and Industrial Wastes 39: 13-21, Chlorine in an Estuarine Environment. Special
1957. report in Applied Marine Science and Ocean
Engineering #206, Virginia Inst. of Marine
20. Laubusch, E.J. Safe Handling of Chlorine. In: ScL, 1978.
Chlorine: Its Manufacture, Properties and Use.
ACS Monograph #154, Reinhold Publ. Corp., 34. LeBlanc, N.E. and McEuen, J.M. A Compar-
New York, NY, 1962. isol1 of Bromine Chloride and Chlorine as
Wastewater Disinfectants in Pilot Plant Stud- 47. Morris, J.C. The Mechanism of the Hydrolysis
ies. Water Chlorination, Environ. Impact and of Chlorine. Journal American Chemical Soci-
Health Effects 2:637-650, 1978. ety 68, 1946.

35. Greene, D.J. Wastewater Disinfection with 48. Morris, J.C. The Acid Ionization Constant of
Bromine Chloride. Presented at the Chesa- HOCI From 5 C to 35 C. Journal Phys.
peake Water Pollution Control Association/ Chemistry 70, 1966.
Water and Waste Operators Association of
Maryland, DEllaware and District of Columbia 49. Downs, A.J. and Adams, C.J. The Chemistry of
Joint Annual Meeting, 1979. Chlorine, Bromine, Iodine and Astatine. Per-
gamon Press, Oxford, 1973.
36. Greene, D.J. An Alternative Wastewater Dis- 50. Noack, M.G. and Doerr, R.L. Chlorine Dioxide,
infectant. WClter, July 1981. Chlorous Acid and Chlorites. Kirk-Othmer
Encyclopedia of Chemical Technology, 3rded.,
37. Keswick, B.H. et al. Comparative Disinfection
HF Mark et aI., eds., 5:612-632, 1979.
Efficiency of Bromine Chloride and Bromine
for Poliovirus. JAWWA 70(10):573-577,1978. 51. Masschelein, W.J. Chlorine Dioxide. Ann Arbor
Science. Ann Arbor, MI, 1979.
38. Keswick, B.H., et al. Evaluation of Bromine
Chloride as it Disinfectant of Poliovirus Sus- 52. Equilibrium Solubility of Chlorine Dioxide in .
pended in Water Flowing Thru a Model Contact Water. Tech. Info Sheet #2. Tappi 44(8):160A.
Chamber. Pmsented at 19th Ann. ASM Mtg., 1961.
1979. 53. Gordon, G., et al. The Chemistry of Chlorine
Dioxide. Progress in Inorganic Chemistry
39. Keswick, B.H. et al. Comparative Disinfection 15:201, 1972.
by Bromine Ghloride and Chlorine of Viruses
and Bacteria in Wastewater. JWPCF 52 (10): 54. Granstrom, M.L. and Lee, G.F. Rates and
2581-2518, '1980. Mechanisms of Reactions Involving Oxychloro
Compounds. Public Works 88(12):90-92,
40. Keswick, B.H. et al. Comparative Disinfection 1957.
of Poliovirus by Bromine Chloride and Chlorine
in a Model Ccmtact Chamber. Water Research 55. Medir, M. and Giralt, F. Stability of Chlorine
16(1 ):89-94, 1982. Dioxide in Aqueous Solution. Water Research
41. Ward, R.W., C:lt al. Disinfection Efficiency and
Residual Toxicity of Several Wastewater Dis- 56. Dow Chemical Co. Bromine Chloride Hand-
infectants, Volume I. Grandville, Michigan. book, 1979.
EPA-600/2-76-156, NTIS No. PB-262245,
U.S. Environmental Protection Agency, Cin- 57. Mills, J.F. Chemistry and Disinfectant Proper-
cinnati, OH, 1976. ties of Bromine Chloride. In: A.J. Rubin, ed,
Chemistry of Wastewater Technology. Ann
42. Ward, R.W., C:lt al. Disinfection Efficiency and Arbor Science, Ann Arbor, MI, 1978.
Residual Toxf:city of Several Wastewater Dis-
infectants, Volume II. Wyoming, Michigan. 58. Cole, L.G. and Elverum, G.W. Thermodynamic
EPA-600/2-'n-203, NTIS No. PB-274498, Properties ohhe Diatonic Halogens from Spec-
U.S. Environmental Protection Agency, Cin- troscopic Data. Journal of Chemical Physics
cinnati, OH, 1977. 20,1952.

43. Laubusch, E.J. Chlorine. Public Works 94: 59. Greenwood, N.N. Physicochemical Properties
511-521, 19Ei3. of the Interhalogen Compounds. Revs. Pure
Applied Chem, Australia 1, 1951.
44. Chlorine Institute, Properties of Chlorine in SI
Units. The Chlorine Institute, New York, NY, 60. Mills, J.F. Interhalogens and Halogen Mix-
1981. tures as Disinfectants. In: Johnson, ed., Dis-
infection: Water and Wastewater. pp. 113-
45. Laubusch, E.J. Hypochlorites. Public Works 114, Ann Arbor Science, Ann Arbor, MI, 1975.
61. Bennett, J.E. On-Site Generation of Hypo-
46. Baker, R.J. Characteristics of Chlorine Com- chlorite Solutions by Electrolysis of Seawater.
pounds. JWPGF 41 :482, 1969. ArChE Symp. Ser. 74(178):265-269, 1978.
62. Michalek, S.A and Leitz, F.B. On Site Genera- 76. Haas, C.N., and Karra, S.B. Kinetics of Waste-
tion of Hypochlorite. JWPCF 44(9): 1697-1712, water Chlorine Demand Exertion: JWPCF
1972. 56(2):170-173, 1984.

63. Wojtowicz, J.A Chlorine Monoxide, Hypochlor- 77. White, G.C., et al. Problems of Disinfecting
ous Acid and Hypochlorites. Kirk~Othmer Nitrified Effluents. Presented at the ASCE
Encyclopedia of ChemicalTechnology, 3rd Ed., Nat'l Conf. on Environ. Eng., 1981.
H.F. Mark et aI., eds., 5:580-611, 1979.
78. Weil, I. and Morris, J.C. Equilibrium Studies on
N-Chloro Compounds. Journal American
64. Weber, G.R., et al. Effect of Ammonia on the
Chemical Society 71 :312-313, 1949.
Germicidal Efficiency of Chlorine in Neutral
Solutions. JAWWA 32:1904-1912,1940. 79. Gray, E.T., etal. Chloramine Equilibria and the
Kinetics of Disproportionation in Aqueous
65. Taras, M.J. Effect of Free Residual Chlorina- Solution. ACS Symp. Ser. #82: Organometals
tion on Nitrogen Compounds in Water. and Organometal loids, Occurrence and Fate
JAWWA 45:47-61,1953. in the Environment, 1978.
66. Enslow, L.H. Ammonia Chlorine Reactions 80. White, G.C. Handbook of Chlorination. Van
and Chloramine Production in Water and Nostrand Reinhold Co., New York, NY, 1972.
Sewage Treatment. Contract Record and Eng.
Rev. 45:291-295, 1931. 8.1. Morris, J.C. and Isaac, R.A A Critical Review
of Kinetics and Thermodynamic Constants for
67. Griffin, AE. and Chamberlin, N.S. Some the Aqueous Chlorine-Ammonia Systems.
Chemical Aspects of Breakpoint Chlorination. Water Chlorination, Environmental Impact
Journal New Eng. Wat. Wks. Assn. 55(3), and Health Effects 4(1 ):49, 1983.
82. Morris, J.C. Kinetics of Reactions Between
68. Griffin, AE. and Chamberlin, N.S. Relation of Aqueous Chlorine and Nitrogen Compounds.
Ammonia Nitrogen to Breakpoint Chlorination. In: S.D.Faust, ed, Principles and Application of
American Journal Public Health. 31 :803-808, Water Chemistry, John Wiley, New York, NY,
1941. pp. 22-53, 1967.
69. Griffin, AE. and Chamberlin, N.S. Exploring 83. Saunier, B.M. and Selleck, R.E. The Kinetics of
the Effect of Heavy Doses of Chlorine in Breakpoint Chlorination in Continuous Flow
Sewage. Sewage Works Journal 17:730-742, Systems. JAWWA 71 :164-172,1979.
84. Trofe, T.W., et al. Kinetics of Monochloramine
70. McKee, J.E., et al. Chemical and Colicidal Decomposition in the Presence of Bromine.
Effects of Halogens in Sewage. JWPCF 32: Environ. Sci. and Tech. 14:544, 1980.
795-819, 1960.
85. Saguinsin, J.L.S. and Morris, J.C. The Chem-
71. Symons, G.E., et al. Reduction of Bacteria in istry of Aqueous Nitrogen Trichloride. In: J.D.
Sewage by Chlorination. Water Works and Johnson, ed, Disinfection: Water and Waste-
Sewerage 85:983-988, 1938. water. pp. 277-299, Ann Arbor Science, Ann
Arbor, MI, 1975.
72. Symons, G.E., et al. Variation in the Chlorine 86. Baker, R.W.R. Studies on the Reaction Be-
Demand of Buffalo Sewage. Sewage Works tween Sodium Hypochlorite and Proteins. I.
Journal 13:249-264, 1941. Physicochemical Study of the Course of the
Reaction. Biochemical Journal 41 :337-342,
73. Taras, M.J. Preliminary Studies on the Chlor-
ine Demand of Specific Chemical Compounds.
JAWWA 42:462~474, 1950. 87. Wright, N.C. The Action of Hypochlorites on
AminoAcids and Proteins. The Effect of Acidity
74. Feben, D. and Taras, M.T. Studies on Chlorine & Alkalinity. Biochemical Journal 30:1661-
Demand Constants. JAWWA 43(9):22-31, 1667,1936.
88. Isaac, R.A and Morris, J.C. Rates of Transfer
75. Lin, S. and Evans, R.L. Chlorine Demand Study of Active Chlorine Between Nitrogenous Sub-
of Secondary Sewage Effluents. Water and stances. Water Chlorination: Environ. Impact
Sewage Works 121(1):35,1974. and Health Effects 3, 1980.

89. Cachaza, J.M. Kinetics of Oxidation of Nitrite 102. Sikorowska, C. Influence of Pollutions on
by Hypochlorite in Aqueous Basic Solution. Chlorine Dioxide Demand of Water. Gaz.
Can. Journal Chemistry 54:3401-3406, '1976. Woda. Tech. Sanit., 35, 12,4645, 1961.

90. Murphy, K.L., et al. Effect of Chlorination 103. Somsen, R.A. Oxidation of Some Simple
Practice on Soluble Organics. Water Research Organic Molecules with Aqueous Chlorine
9:389, 1975. Dioxide Molecules. I. Kinetics. Tappi 43:154-
156,1960.11 Reaction Products, Tappi 43:157-
91. De Laat, J., Elt al. Chloration de composes
organizes: Demande on chlore et reactivite vis
a vis de la Furmation des trihalometha Inci- 104. Wajon, J.E., et al. Oxidation of Phenol and
dence de I'azote ammoniacal. Water Research Hydroquinane by Chlorine Dioxide. Environ.
16:143-147,1982. Sci. and Tech. 16:396, 1982.
92. Jolley, R.L. Chlorine Containing Organic Con-
105. Snoeyink, V.L. and Suidan, M.T. Dechlorina-
stituents in Chlorinated Effluents. JWPCF tion by Activated Carbon and Other Dechlor-
47:601-618, 1975.
inating Agents. In: J.D. Johnson, ed., Disinfec-
93. Chow, B.M. and Roberts, P.V. Halogenated tion: Water and Wastewater. Ann Arbor
Byproduct Formation by Chlorine Dioxide and Science, Ann Arbor, MI, 1975.
Chlorine. Journal of the Environmental Engi-
neering Division, ASCE 107(4):609-618, '1981. 106. Laubusch, E.J. Sulfur Dioxide. Public Works
94(8):117-121, 1963.
94. Glaze, W.H. and Henderson, J.E. Formation of
Organochlorine Compound from the Chlorina- 107. Scott, W.D. and Hobbs, P.V. The Formation of
tion of a Municipal Secondary Effluent. JWPCF Sulfate in Water Droplets. Jour. Atmos. Sci.,
47:2511, 197fi. 24, 54, 1967.

95. LaPointe, T.F., et al. Kinetics ofTribromamine 108. Lister, M.W. and Rosenblum, P. Rates of
Decomposition. In: J.D. Johnson, ed, Disinfec- Reaction of Hypochlorite Ions with Sulfite and
tion: Water and Wastewater, pp. 301-338, Iodine Ions. Can. Journal Chemistry 41 (12):
Ann Arbor Scil3nce, Ann Arbor, 1975. 3013-3020, 1963.

96. Wajon, J.E. and Morris, J.C. Bromamination 109. Srivastava; R.D., et al. Kinetics of Reaction of
Chemistry: Rates of Formation of NH Br and Sodium Hypochlorite and Sodium Sulfide by
Some N-Bromamino Acids. Water Chlorina- Flow Thermal Method. Ind. Eng. Chem. Fund
tion: Environ. Impact and Health Effects 19:207-209, 1980.
110.. Stanbro, W.D. and Lenkevich, M.J.Slowly
97. Johnson, J.D. and Overby, R. Bromine and Dechlorinated Organic Chloramines. Science
Bromamine Disinfection Chemistry. Journal 215:967-968, 1982.
of the Sanitary Engineering Div., ASCE97
(SA5):617,1971. 111. Standard Methods for the Examination of
Water and Wastewater, 16th edition. APHA,
98. Cromer, J.L., e!t al. Dibromamine Decomposi- AWWA and WPCF, Washington, DC, 1985.
tion Kinetics. In: AJ. Rubin, ed, Chemistry of
Wastewater TC3chnology, Ann Arbor Science, 112. Chen, C. and Gan, H.B. Wastewater Dechlor-
Ann Arbor, 1978. ination: State of the Art Field Survey and Pilot
Studies. USEPA Contracts 14-12-150 and 68-
99. Inman, G.W., et al. Kinetics of Nitrogen 03-2745, Final report, 1981.
Tribromide Decomposition in Aqueous Solu-
tion. Inorg. Chl3mistry 15:3037-3042, 1976. 113. Puri, B.R., et al. Interaction of Charcoal with
Chlorine Water. Journal Indian Chemical
100. Roberts, P.V., et al. Chlorine Dioxid<e for Society 35:181-184, 1958.
Wastewater Disinfection: A Feasibility Evalua-
tion. EPA-60C>/2-81-092, NTIS No. PB81- 114. Puri, B.R., et al. Interaction of Charcoal with
213357, U.S. E:nvironmental Protection Agen- Chlorine Water II. Journal Indian Chemical
cy, Cincinnati, OH, 1981. Society 37: 171 -176, 1960.

101. Ingols, R.S. and Ridenour, G.M. Chemical 115. Bauer, R.C. and Snoeyink, V.L. Reactions of
Properties of Chlorine Dioxide in Water Treat- Chloramines with Active Carbon. JWPCF
ment. JAWWA 40:1207-1227, 1948. 45(11 ):2290-2301, 1973.
116. Kim, B.R. and Snoeyink, V.L. The Monchlor- 130. Jenkins, R.L. and Baird, R.B. Determination of
amine GAC Reaction in Adsorption Systems. Total Chlorine Residuals in Treated Waste-
JAWWA 72(8):488-490, 1980. waters by Electrode. Analytical Lett~rs 12(A2):
117. Kim, B.R., et al. Removal of Dichloramine and
Ammonia by Granular Activated Carbon. 131. Liebermann, J., et aJ. Development of the
JWPCF 50(1), 1978. FACTS Procedure for Combined Forms of
Chlorine and Ozone in Aqueous Solution.
118. Kim, B.R., et al. Analysis of Models for Environ. Sci. and Technol. 14: 1395-1400,
Dichloramine Removal by Activated Carbon in 1980.
Batch and Packed Bed Reactors Using Quasi-
linearization and Orthogonal Collocation Meth- 132. Cooper, W.J., et al. Comparison of Several
ods. Water Research 12(5):317-326, 1978. Instrumental Methods for Determining Chlo-
rine Residuals in Drinking Water.. JAWWA
119. Marks, H.C., et al. Amperometric Methods in 74( 10):546-552, 1982
the Control of Water Chlorination. Analytical
Chemistry 19:200-204,1947. 133. Knechtel, J.R., et al. Determination of Chlorine
Dioxide in Sewage Effluents. Analytical Chem-
120. Marks, H.C., et al. Amperometric Titration of istry 50(2):202-205, 1978.
Residual Chlorine in Sewage. Water and
Sewage Works, May, 1948. 134. Wheeler, G.L., et al. A Rapid Microdetermina-
tion of Chlorine Dioxide in the Presence of
121. Marks,H.C. and Joiner, R.R. Determination of Active Chlorine Compounds. Microchemical
Residual Chlorine in Sewage. Analytical Journal 23:160-164, 1978.
Chemistry 20:1197, 1948.
135. Olivieri, V.P., et al. Inactivation of Virus in
122. Heukelekian, H., et al. Disinfection of Wastes Sewage. Journal of the Sanitary Engineering
by Chlorination. Ind. Eng. Chem. 45:1004- Div., ASCE97(SA5):661, 1971.
136. Heukelekian, H. and Day, R.V. Disinfection of
123. Day, R.V.. et al. Residual Chlorine Methods Sewage with Chlorine. II. Method and Uni-
and Disinfection of Sewage. Ind. Eng. Chem. formity of Distribution of Chlorine in Sewage
45:1001-1004,1953. Plants. Sewage and Industrial Wastes 23:47-
124. Feng, T.H. Behavior of Organic Chloramines in
Disinfection. JWPCF 38:614-628, 1966. 137. Selleck, R.E., Collins, H. and White, G.C.
Kinetics of Bacterial Deactivation with Chlo-
125. Henderson, W.L. Control of Sewage Chlor- rine. Journal of the Environmental Engineer-
ination by Amperometric Titration. Sewage ing Division, ASCE 104: 1197-1212, 1978
and Industrial Wastes 24:1467-1472, 1952.
138. Collins, H.F., et al. Problems in Obtaining
126. Heukelekian, H. and Day, R.V. Disinfection of Adaquate Sewage Disinfection. Journal ofthe
Sewage with Chlorine. Ill. Factors Affecting Sanitary Engineering Division, ASCE 97(SA5):
Coliform Remaining and Correlation of Ortho- 549-562, 1971 .
. Iidine and Amperometric Chlorine Residuals.
Sewage and Industrial Wastes 23:155-163, 139. Stagg, C.H., et al. Chlorination of Solids
1951. Associated Coliphages. Progr. Water Technol.
1O(12):381-387, 1978.
127. Browning, G.E. and McLaren, F.R. Experiences
with Wastewater Disinfection in <2alifornia. 140. . Hejkal, T.W., et al. Survival of Polioviruses
JWPCF 39(8):1351-1361, 1967. Within Organic Solids During Chlorination.
Appl. Environ. Microbiol. 38:114, 1979.
128. Collins, H.F. and Deaner, D.G. Sewage Chlor-
ination Versus Toxicity, Dilemma. Journal of 141. Butterfield, C.T., et al. Influence of pH and
the Environmental Engineering Division, ASCE Temperature on the Survival of Coliforms and
99(SA6):761-772, 1973. Enteric Pathogens When Exposed to Free
Chlorine. U.S. Pub. Health Rep. 58: 1837-1866,
129. Lin, S., et al. Selection of a Method to 1943.
Determine Residual Chlorine in Sewage Ef-
fluents. Water and Sewage Works 118(11): 142. Ozgumus, N.H. Evaluation of Chemical and
360-364, 1971. Bacteriological Methods of Determining Ger-

micidal Activity of Chlorine. Ph.D. Thesis, 156. Chlorine Institute, Chlorine Vaporizing Equip-
Mich. State CoiL, Lansing, MI, 1950. ment. Pamphlet #9, ed. 2, rev. 2. The Chlorine
Institute, New York, NY, 1979.
143. Rudolph, A.S. and Levine, M. Factors Affecting
the Germicidal Efficiency of Hypochlorite 157. Sepp, E.and White, G.C. Manual for Waste-
Solutions. IoWa State College Eng. Exp. 8ta. water Chlorination and Dechlorination Prac-
Bulletin 150, 1941. tices. California Department of Health Serv-
ices, 1981.
144. Floyd, R., et al. Inactivation by Chlorine of
Single Poliovirus Particles in Water. Environ- 158. Connell, G. F. and Fetch, J.J. Advances in
mental Science and Technology 13:438-442, Handling Gas Chlorine. JWPCF 41:150-155,
1979. 1969.
145. Sharp, D.G. and Leong, J. Inactivation of 159. Chemical Aids Manual for Wastewater Treat-
Poliovirus I, Brunhilde Single Particles by ment. EPA430/9-79~018, U.S. Environmental
Chlorine in Water. Appl. Environ. Microbiol. Protection Agency, Washington, DC, 1979
40:381-385, 1900.
160. Jordan, R.W. Improved Method Generates
146. Chang, S.L. Resi!;tance of Pathogenic Naegler- More Chlorine Dioxide. Water and Sewage
ia to Some Common Physical and Chemical Works 127(10):44, 1980.
Agents. Appl. Environ. Microbiol. 38:36-38,
1978. 161. Longley, K.E. Turbulence Factors in Chlorine
Disinfection of Wastewater. Wat. Res. 12:
147. Severin, B.F.lnac:tivation of Proposed Indicator 813-822, 1978.
Organisms by Inorganic Chloramines. M.S.
Special Problem Report, Dept. of Civil Eng., 162. Chow, V.T. Open-Channel Hydraulics. Mc-
Univ. of Illinois at Urbana, 1975. Graw Hill Book Co., New York, NY, 1959.
148. Butterfield, C.T., et al. Influence of pH and 163; Mandt, M. Jet Disinfection Improves Contact
Temperature on the Survival of Coliforms and Chamber Performance, Cuts Costs. Water and
Enteric Pathogens When Exposed to Chlorine. Sewage Works 125:447-453, 1978.
U.S. Pub. Health Rep. 61 :157-192,1946.
149. Scott, W.J., et al. Recommended ProcedUl'es 164. LaGrosa, J.J., et 811. Wastewater Chlorination
in the Use of Chlorine at Water and Sewage by Jet Diffusion. April 1978.
Plants. JAWWA45:1060-1074, 1953.
165. Huang,J.Y.C., Warriner, R. and Ni, N.S.N. Pilot
150. Fiedler, R.P, On-slite Caustic Chlorine Genera- Tests of Chlorination Facility for Disinfecting
tion for Water Dis;infection. JAWWA 66(1 ):46- Secondary Effluent. JWPCF 57(7):777-784,
49,1974. 1985.
151. Doan, R. and A.S. Haimes, Hypochlorite Gen- 166. Trussell, R.R. and Chao, J. Rational Design of
eration at a Water Pollution Control Plant. Chlorine Contact Facilities. JWPCF 49(4):659-
Public Works 10~1(1 ):48-51, 1978. 667,1977.
162. Baur, F. Contract Granted for Sodium Hypo- 167. Hart, F.L. Modifications for the Chlorine Con-
chlorite Generat()r. Water and Sewage Works tact Chamber. Jour. New Engl. Water Pollut.
119:76, 1972. Control Assoc. 13(2):135-151, 1979.
163. Benas, B. Bulk Storage of Liquid Chlorine at
San Francisco Sewage Plants. Water ~nd 168. Hart, F.L. Improved Hydraulic Performance of
Sewage Works 100:381-385, 1953. Chlorine Contact Chamber. JWPCF 51 (12):
2868-2875, 1979.
154. Chlorine Institutl:l, Chlorine Pipelines. Pam-
phlet #60, 2nd ed. The Chlorine Institute, New 169. Hart, F.L. and Vogiatzis, Z. Performance of
York, NY, 1982. Modified Chlorine Contact Chamber. Journal
of the Environmental Engineering Division,
165. Cameron, G.W. Chlorination-Plant Problems ASCE 108:549, 1982.
at City Island Plant ofthe Atlantic City Sewage
Company. Sewal~e Works Journal 19:603- 170. Metcalf & Eddy. Wastewater Engineering, 2nd
606,1947. ed. McGraw Hill, New York, NY, 1979.
171. Fair, G.M., et al. Elements of Water Supply and
Wastewater Disposal. John Wiley & Sons,
New York, NY, 1971.

172. Louie, D.S. and Fohrman, M.S. Hydraulic

Mpdel 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,

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, AseE
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,
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. Culp, 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.

Chapter 6
Ozone Disinfection

6.1 Introduction In contrast to the widespread use of ozone for water

treatment in Europe, very few European wastewater
6.1. 1 General ozone disinfection systems exist. Currently, there are
The planning, design, construction and operation of a more ozone disinfection systems in use at U.S.
wastewater treatment process involves at least five wastewater plants than at U.S. water plants or
separate parties: owner, design engineer, contractor, European wastewater plants.
equipment manufacturer, and regulatory agency.
The first U.S. wastewater plant to use o~one for
Beca use the use of ozone for wastewater disinfection
disinfection was Indiantown, Florida, which began
is relatively new, the responsibility for development
operation in 1975 (3). By 1980 about 10 wastewater
of process and equipment design has been largely the
treatment plants using ozone for disinfection had
equipment manufacturer's. However, many design
been constructed. Unfortunately, some of these
features are not related to a given manufacturer's
earlier ozone disinfection facilities have chosen to
piece of equipment, like process flexibility considera-
abandon ozone disinfection for one or more of the
tions, contact basin size and configuration, and
following reasons: excessive high cost of operation,
influent water quality characteristics. The informa-
equipment problems, excessive maintenance cost,
tion presented in this manual is intended to provide
and inability to attain performance objectives without
necessary design capability and understanding to all
major modifications (4). Despite these eariy set-
parties so that each party can be properly involved in
backs,. many facilities have proposed to utilize ozone
the development of an ozone disinfection system.
for wastewater. A list of wastewater plants that have
During the development of this design manual, site used or are using ozone disinfection is presented in
visits to seven operating ozone installations were Table 6-1 (5).
conducted in order to obtain information on operating
experiences at existing facilities. The information
obtained from these site visits is incorporated in the
6.2 Ozone Properties, Chemistry and
manual. Terminology
Ozone is a molec.!Jle that can co-exist with air or high
6.1.2 History of Ozone purity oxygen, or can dissolve in water. It is a very
A history of ozone has been documented by others strong o>ddizing agent and a very effective. disinfect-
(1,2). This work was based on surveys of operating ant.
ozone installations in Europe and the United States
and is briefly described. 6.2.1 Ozone Properties
Ozone (CIa) is an unstable gas that is produced when
Experiments conducted in 1886 showed that ozon- oxygen molecules are dissociated into atomic oxygen
ized air "will effect the sterilization of polluted and sllPsequently collide with another oxygen mole-
water." The first drinking water plant to use ozone cule (6). The energy source for dissociating the
was built in 1893 at Oudshoorn, Holland. The French oxygen molecule can be produced commercially or
studied the Oudshoorn plant. and after pilot testing can occur naturally. Some natural sources for ozone
constructed an ozone water plant at Nice, France in production are ultraviolet light from the sun and
1906. Because ozone has been used at Nice since lightning during a thunderstorm.
that time, Nice is often referred to as "the birthplace
of ozonation for drinking water treatment." Ozone may be produced by electrolysis, photochem-
ical reaction, radiochemical reaction, or by "electric
In Europe there is a strong commitment to attain a discharge" in a g~,S that contains oxygen (7). The
water of the highest chemical quality. Currently, electric discharge principle has been used in most
there are over 1,000 European drinking water plants commercial applications and in all known water and
that use ozone at one or more points in the treatment wastewater treatment applications. The electric
process. discharge method is presented in this manual.
Table 61. U.S. Municipal Wastewater Treatment Plants Using Ozone (5)
Primary Average Flow Rate
Feed Purpose of Startup
Location Gas Ozone Date mgd m3 /d
Indiantown, FL air disinfection 1975 0.5 1,890
Woodlands, TX air disinfection 1976 1.5 5,680
Upper Thompson Sani air disinfection 1977 1.5 5,680
tatlon District, CO
Hunter, NY air disinfection 1977 0.1 368
Harriman, NY air disinfection 19773 0.36 1,360
Chino Basin, CA air suspended solids 1978 5 18,930
Palo Alto, CA air organics 1978 4 15,100

Collegeville, MN air disinfection 1978 0.22 830

Mahoning County, OH"' O2 disinfection 1978 4 15,100
Hunter Highlands, NY air disinfection 1978 <1 <3,780
Cotter Gasville, AR air disinfection 19783 1 3,780.
Springfield, MO* O2 disinfection 19783 30 113,600
Bull Shoals, AR air disinfection 0.19 720
Sebring, FL air disinfection 0.52 1,960

Maryland City, MD air disinfection 0.58 2,200

Oak Ridge, NY air disinfection 19803 0.12 450
Norton AFB, CA air phenols 1980 3 0.25 950
Carmel,NY air disinfection 19803 1 3,780
Potomac Heights, MD air disinfection 19803 0.2 760 .
Murphreesboro, TN* O2 disinfection 19803 8 30,280
Pensacola, FL* O2 disinfection 1980 20 75,700

Hercules, CA air disinfection 19803 0.4 1,510

Marion, NY air disinfection 19803 0.125 470
Brookings, SO air disinfection 19803 6 22,700
Concord, NC* O2 disinfection 19803 25 94,600
Delaware County, OH air disinfection 19803 1.5 . 5,680
Frankfort, KY air disinfection 19803 7 26,500
Ocean City, MD* O2 disinfection 19808 12 45,400

Madisonville, KY* O2 disinfection 19808 5 18,900

Little Valley, NY air disinfection 1981 0.28 1,060
Yaphank, NY air disinfection 1981 0.12 450
Granby, CO air disinfection 1981 <1 <3,680
Rocky Mount, NC* O2 disinfection 19823 40 151,400
West Knoxville, TN air disinfection 1982 2 7,570
Vall, CO air disinfection 19828 2.7 10,200

Hagerstown, MD* O2 disinfection 1983 8 30,300

Olympia, WA* O2 disinfection 1983 ' 14 53,000
Indianapolis, IN
Belmont Plant* O2 disinfection 1983 120 454,200
Southport Plant* O2 disinfection 1983 125 473,200
Twining, NM air disinfection 1983 0.095 360
Somero, NY air disinfection 1983 0.1 368
Alburndale, FL air disinfection 1984 2 7,570
'Plants operating per 1982 EPA survey
Plants using oxygen activllted sludge process

Ozone is used in relatively low concentrations in Ozone has a very distinct odor and it owes its name to
water and wastewclter treatment applications. T,he its odor. The word ozone is derived from the Greek
properties of pure ozone are presented for general word 'ozein', which means to smell. Ozone can be
background information in Table 6-2 (7), detected at concentrations of only 2x10-5 to 1x1 0- 4
., g/m 3 (Le.,0.01 toO.05 ppm byvolume)(1). The ability Ozone Color and Odor to smell ozone at small concentrations is considered a
At ordinary temperatures ozone is a blue gas, but at safety feature because an ozone odor can be detected
typical concentrations its color is not noticeable before ozone related health considerations develop.
unless it is viewed through considerable depth (7). The present allowable 8-hour exposure concentra-

Table 6-2. . Properties of Pure Ozone (0 3 ) (7)
melting point, C -192.5 0.4
boiling point, C -111.9 0.3
critical temperature, C -12.1
critical pressure, atm 54.6
critical volume, cm 3/mole 111

density and vapor pressure of liquid

Temperature, C Density, g/cm 3 Vapor Pressure. Torr

-183 1.574 0.11

-180 1.566 0.21
-170 1.535 1.41
-160 1.504 6.73
-150 1.473 24.8 .
-140 1.442 74.2
-130 1.410 190
-120 1.378 427
-110 1.347 865
-100 1.316 1605

density of solid ozone, g/cm 3, at 77.4K 1.728

viscosity of liquid, Cp3, at 77.6K 4.17
at 90.2K 1.56
surface tension, dyn-cm, at 77.2K 43.8
at 90.2K 38.4
parachlor* at 90.2K 75.7
dielectric constant, liquid, at 90.2K 4.70
dipole moment, debye 0.55
magnetic susceptibility, cgs units, gas 0.002 x 10- 6
liquid 0.150
heat capacity of liquid from 90 to 1500 K Cp == 0.425 + 0.0014 (T-90)
heat of vaporization, kcal/mole, at -111.9C 3410
at -183C 3650
heat and free energy of formation
aH f , kcal/mole aG f , kcal/mole

gas at 298.15K 34.15 38.89

liquid at 90.15K 30.Q
hypothetical gas at OaK 34.74
*M-yl/4 (0 - d) where M == molecular weight; r == surface tension; D == liquid density; d== vapor density

tion is 2x1 0-Sg/m 3 (i.e., 0.1 ppm byvolume), whi.ch is Figure 6-1. The direct reaction of ozone with solutes (M"
2to 10 times higher than the concentration at which M2) and a hydroxide ion (or radical) catalyzed
ozone can be smelled. Additional safety considera- decomposition reaction, leading to reactive
intermediates. compete for ozone (9).
tions for ozone are discussed in Section 6.6 Ozone Stability

The stability of ozone is greater in airthan 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
pH 7.0 with sodium hydroxide, the half-life was 8
minutes. Hoigne and Bader (9) found that ozone will
react directly with solutes in the water, and that
hydroxide ions and hydroxyl radicals will proviae a
catalyst for the decomposition of ozone into inter-
mediate compounds that are also reactive, such as
peroxide ions and hydroxide ions. Their findingsare
summarized in Figure 6-1 .
Holgne and Bader's results suggest that ozone concentrations has been developed, as shown in
disinfection Is Influenced by raw water chemistry Table 6-3.
characteristics, in addition to the more well known
influences of wastewater pollutants. The water H =Y/X (6-1 )
chemistry influences are generally not utilized in
ozone system design, except as developed in pilot where:
plant studies of thE! specific wastewater to be
disinfected. However, the water chemistry influences
Y = partial pressure of the gas above the liquid
are important to keep in mind when a comparison is X = molar fraction of the gas in the liquid at
made of ozone disinfelction performance at different equilibrium with the gas above the liquid
facilities. H = Henry's law constant (varies with tempera-
ture), atm/mole fraction
The residual ozone c:oncentration in water is de-
creased rapidly by aeration or agitation of the liquid
(7). A wastewater sample with residual ozone must Table 6-3. Solubility of Ozone in Water
not be agitated or collc3cted after a period of agitation Water Henry's Ozone Ozone
has occurred, or the measurement of residual ozone Temperature Constant Concentration Solubility
will be inaccurate. Also, because ozone may be (OC) Atm/Mole mg/I ppm-vol mg/I
released from the water, a plant effluentthat contains 1,940 12.07 6,044 8.31
a high residual ozone concentration that is used as a 5
source of non-potable water, may release excessive
15 2,880 12.Q7 6,044 5.60
amounts of ozone and contaminate the ambient 20 3,760 12.07 6,044 4.29
environment. 25 4,570 12.07 6,044 3.53
30 5,980 12.07 6,044 2.70

The stability of ozone in air or oxygen is significantly 1,940 18.11 9,069 12.47
affected by the templ3rature of the gas. In a clean 5
vessel at room temperature the half-life of ozone may 15 2,880 18.11 9,069 8.40
range from 20 to 100 hours (7). At 120C (248F) the 20 3,760 18.11 9,069 6.43
half-life is only 11 to 112 minutes and at 250C 25 4,570 18.11 9,069 5.29
(482F) only 0.04 to 0.4 seconds. This characteristic 30 5,980 18.11 9,069 4.04
of ozone is important for design because cooling of 1,940 24.14 12,088 16.62
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. 15 2,880 24.14 12,088 11.19
20 3,760 24.14 12,088 8.57
25 4,570 24.14 12,088 7.05
30 5,980 24.14 12,088 5.39 Ozone Physil:al Characteristics
Gaseous ozone is explc)siveat an ozone concentration 1,940 36.21 18,132 24.92
of 240 g/m 3 (20 percent wt in air) (7). Fortunately, the 5
maximum gaseous ozone concentration typically
15 2,880 36.21 18,132 16.79
found in water or wastewater ozone disinfection 20 3,760 36.21 18,132 12.86
systems does not excEled 50 g/m 3 (4.1 percent wt in 25 4,570 36.21 18,132 10.58
air). If, however, a medium that can adsorb and 30 5,980 36.21 18,132 8.09
concentrate ozone is inappropriately located in the Note: The concentration of the ozone gas is determined at a
system, then explosive ozone concentrations could standard temperature of 68F (20C) and a standard pres-
develop. sure of 1 atmosphere (101 kPa). Ozone Solubility 6.2.2 Ozone Chemical Reactions

Ozone solubility In water is important because ozone Ozone is a very strong oxidizing agent, having al1
disinfection is dependent upon the amount of ozone oxidation potential of 2.07 volts (1). Ozone will react
transferred to the wastewater. Henry's law relative to with many organic and inorganic compounds in the
ozone systems states that the mass of ozone that will wastewater. These reactions are typically called
dissolve in a given volume of water, at constant "ozone demand" reactions. They are important in
temperature, is directly proportional to the partial ozone disinfection system design because the reacted
pressure ofthe ozone nas above the water (10). Using ozone is no longer available for disinfection. Waste-
Equation 6-1 the maximum solubility of ozone in waters that have high concentrations of organics or
water at various temperatures and feed-gas ozone inorganics may require high ozone dosages to achieve

disinfection. It is very important to conduct pilot plant Cyanide. To>dc cyanide ions are readily oxidized by
studies on these wastewaters during ozone disinfec- ozone to the much less toxic cyanate ion. At low pH,
tion system design in order to determine the ozone cyanate ion hydrolyzes to produce carbon dioxide and
reaction kinetics for the level of treatment prior to nitrogen.
ozone disinfection. Reactions with Organic Compounds
In most instances, the oxidation reactions produce an An in-depth analysis of the reactions with ozone and
end product that is Jess toxic than the original various organic compounds were developed by
compound (1,2). Numerous studies have been com- several investigators.These reactions were described
pleted describing the reactions with ozone and by Miller et al. (1), and are summarized below.
various inorganic and organic compounds (11,12). A
brief summary of these reactions is presented. Aromatic Compounds. Phenol reacts readily with
ozone in aqueous solution. Oxalic and acetic acids are Reactions with Inorganic Compounds relatively stable to ozonation in the absence of a
The inorganic compounds that most commonly react catalyst such as ultraviolet light or hydrogen peroxide.
with ozone in a wastewater treatment plant are Cresols and xylenols undergo oxidation with ozone at
sulfide, nitrite, ferrous, manganous, and ammonium faster rates than does phenol. Pyrene, phenanthrene,
ions. Other reactions may also occur if the waste- and naphthalene oxidize by ring rupture. Chloro-
. water characteristics are affected by an industrial benzene reacts with ozone slower than does phenol.
contribution or by in-plant recycle loads. Ozone
reactions with various inorganic compounds have Aliphatic Compounds. There is no evidence that
been analyzed by several researchers (2). These ozone reacts with saturated aliphatic hydrocarbons
reactions are summarized below: under water or wastewater treatment conditions.
There is no evidence that ozone oxidizes trihalo-
Sulfide. The degree of oxidation of sulfide depends methanes. Ozone combined with ultraviolet radiation
upon the amount of ozone used and the contact time. does oxidize chloroform to produce chloride ion, but
Organic sulfides will oxidize to sulfones, sulfoxides no identified organic oxidation product. Unsaturated
and su Ifonic acids at slower rates tha n the su Ifide ion aliphatic or alicyclic compounds react with ozone.
itself. The sulfide ion will oxidize to sulfur, to sulfite
and to sulfate. Pesticides. Ozonation of parathion and malathion
produces paraoxon and malaoxon, respectively, as
Nitrogen Compounds. Organic nitriles, nitroso com- intermediates, which are more toxic than are the
pounds, and hydroxylamines will be oxidized de- starting materials. Continued ozonation degrades the
pending on the amount of ozone used and the contact oxons, but requires more ozone than the initial reac-
conditions. The oxidation reaction of ammonia is first- tion. Ozonation of heptachlorprQduces a stable
order with respect to the concentration of ammonia product not yet identified. Aldrin and 2,4,5,-T are
and is catalyzed by OW over the pH range 7-9 (13). At readily oxidized by ozone, but dieldrin, chlordane,
an initial ammonium concentration of28 mg/l as N, a lindane, DDT, and endosulfan are only slightly
pH of 7.0, and a contact time of 30 minutes, an 8 affected by ozone.
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 Humic Acids. Humic materials are resistant to
percent reduction was observed. Narkis reports that ozonation, requiring lengthy times of ozonation to
total oxidation of organic nitrogen and ammonia was produce small amounts of acetic, oxalic, formic and
never achieved, even at a pH of 12, and at a pH of 6 terephthalic acids, carbon dioxide, and phenolic
nitrates were not produced (14). compounds. Ozonation of humic materials followed
by immediate'chlorination (within eight minutes) has
Nitrite ion is oxidized very rapidly to nitrate ion. This been shown to reduce trihalomethane formation in
reaction can have a significant effect on ozone some cases. Ozonized organic materials generally are
disinfection capability when incomplete nitrification more biodegradable than the starting, unoxidized
occurs. Venosa reported that as much as 2 mg/I of compounds. .
ozone was required to oxidize 1 mg/I of nitrite-
nitrogen (15).
6.2.3 Ozone Disinfection Reactions
Iron and Manganese. The reaction with ozone and the Transfer of ozone into the wastewater is the first step.
ferrous and manganous ions will form an insoluble in meeting the disinfection objective, since ozone
precipitate. The ferrous ion will be oxidized to ferric, must be transferred al1d residual oxidants produced
which will react with OW to form an insoluble before effective disinfection will occur (16). Once
precipitate. Similarly, manganeous ions will form transferred, the residual oxidants, such as ozone,
manganic ions which will react with OH- to form a hydroxide. or peroxide, must make contact with the
precipitate. organisms in order for the disinfection action to
proceed. Therefore, similar requirements and kinetic (16,18,19). Additional work has shown that it was
relationships used for (:hlorine disinfectants can also possible to empirically relate effluent fecal coliform
be used for ozone disinfection. concentration to the product of contactor off-gas
concentration and liquid contact time (20). The data
Contact time has been the subject of much clntro- shown in Figures 6-3 and 6-4 indicate that a better
versyfor ozone disinfec:tion. Studies have shown that correlation occurred for the product of off-gas ozone
effective disinfection c:an occur at contact times as concentration times time than occurred for off-gas
short as one minute (17), but most existing ozone ozone concentration alone. Additional information
disinfection systems helve contact times that are 10 to regarding the use of off-gas ozone concentration as a
16 minutes. Effective ozone disinfection is due to the process control parameter is presented in Section
combined results of high transfer efficiency, good 6,.3.3.
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 analysis has not been practical.
As a result, kinetic relationships in terms of trans- Figure 6-3. Effluent focal coliform concentration versus
off-gas ozone concentration (20).
ferred ozone or residual concentrations have been
::r 10 6 .-----,....,--r---.-.--r-r-rr-r---r--r--r-r"T"1:-T1..,
The transferred ozone, dosage versus disinfection :2 Dt:: 5.25 min
performance kinetic relationship is further discussed 8 10 5 At:: 9.75 min
in Section The relationship between dis-
infection performance and residual oxidants is shown c
in Figure 6-2 (17). Thesl9 types of relationships can be
used graphically or Etmpirical 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 thl:lt ozone disinfection can also
be related to contactor off-gas ozone concentration

Flguro 6-2. Effluent total coliform concentration versus

total resldua:1 oxidants and residUal ozone (17).

103 ~--"----':--'_""';r----""---Ti----'"

Figure 64. Effluent fecal coliform concentration versus

'\ product of off-gas ozone concentration times
time (20).

\ Dt = 5.25 min
\ Log Y =- 0.387X + 3.14
At:: 9.75 min

Residual Ozone\

2.2 ---- - --~\..._--
100~_~ ~. !:-1_ _~I_ _~I::--_,--,
o 4 6 8 10 12
Residuals, mg/I

6.2.4 Ozone Terminology co'nstituents in the wastewater or decomposes to
Throughout this manual the most common units of hydroxyl radicals (9). Some of the transferred ozone
expression are used. The standard temperature and may not react or decompose and will exist in the
pressure are 20C (68F) and 101 kPa (1 atmosphere), contactor as residual ozone (C 2 ). Ozone not trans-
respectively. All gas volume values are corrected to ferred exits the contact basin in the off-gas flow (G2)'
standard temperature and pressure, unless otherwise
The off-gas flow rate (G 2 ) can be quite different from Measured Ozone Parameters the feed-gas flow rate (G 1 ). When oxygen is used as
A simplified line diagram of a wastewater ozone the feed-gas, from 5 to 10 percent of the oxygen may
disinfection system is shown in Figure 6-5. The dissolve in the wastewater and cause the off-gas flow
system consists of the four components of feed-gas rate to be lower than the feed-gas flow rate. If an
preparation, ozone generation, ozone contacting and exhaust blower is used to pull the off-gas out of the
ozone destruction. The feed-gas source may be either contact basin, then the off-gas flow rate could be
air or high purity oxygen. In unique applications higher than the feed-gas flow rate due to air getting
recycled oxygen has been utilized. The feed-gas flow into the tanks through cracks, etc. The off-gas flow
rate (G 1 ) is usually measured before the ozone rate may be measured before or after the ozone
. generator so more precise instrumentation may be destruct unit since the two flows are identical. the
used without the added expense of providing ozone ozone concentration in the off-gas (Y 2 ) is measured
resistant materials. prior to the ozone destruct unit and the ozone
concentration in the exhaust-gas (Ys ) is measured
The concentration of ozone in the feed-gas (Y1) is after the ozone destruct unit. The wastewater flow
measured before the ozone contact basin. In the into (L 1 ) and out of (L 2 ) the contact basin is identical,
contact basin most of the ozone is transferred to the and may be measured at either point. The influent
wastewater and reacts with the ozone demanding coliform concentration (No) is sampled prior to the

Figure 6-5. Simplified ozone process schematic diagram.

Ga Ya Ozone
+--- Destruction

Feed Gas Preparation
Oxygen Production
G1 Ozone 'G 1 Y1
- N
Oxygen Storage Generation
Contact l2 C2
AirlOxygen Treatment
r" Basin


G1 = Feed-Gas Flow Rate (ma/min)

Y1 = Feed-Gas Ozone Concentration (g/m 3 )
G2 = Ga = Off-Gas and Exhaust-Gas Flow Rate (m 3 /min)
Y2 = Off-Gas Ozone Concentration (g/m 3 )
Ya = Exhaust-Gas Ozone Concentration (g/m 3 )
l1 = l2 = Wastewater Flow Rate
C1 = Residual Ozone Concentration in Wastewater Influent (mg/l)
C2 = Residual Ozone Concentration in Wastewater Effluent (mg/l)
No = Influent Coliform Concentration (#/100 mIl
N = Effluent Coliform Concentration (#/100 ml)

ozone contact basin and the effluent coliform con- applied ozone dosage (D), ozone transfer efficiency
centration (N) is sampled after the contact basin. (TE), and transferred ozone dosage (T). The termi-
nology, symbols, units of expression and formulas for
The metric units of expression and English unit the calculated ozone parameters are shown in Table
conversion factors for typical ozone related measure- 6-6.
ments are presented in Table 6-4. The English unit
conversions for ozone concentration in the gas are The formulas shown assume that metric units of
dependent upon the standard temperature and expression are used for the individual parameters.
pressure conditions selected. The conversion factors The most common English units of expression are
shown are for a standard temperature of 20C (68F) percentwtforY1 andY 2 , scfmfor G 1and G2 , mgdfor l1
and a standard pressure of 101 kPa (1 atmosphere). If and L2 , and Ibid for P. To calculate P, the gas flow rate
other standards are used, then the conversion factors is typically converted to Ibid and is multiplied by the
may be changed using the formulas shown in Table ozone concentration in percent wt.
6-4. The English unit equivalents for typical ozone Pin Ib ozonelday = G1 in Ibid * Y1 in % wt/100
concentrations encountered in ozone disinfection
facilities are presentc3d in Table 6-5. The applied ozone dosage, in English units, is equal to
the ozone production in Ibid divided by liquid flow in Calculated Ozone Parameters mgd divided by a conversion factor of 8.34 Ib/ga!.
Ozone measurements described in the previous
section can be used to calculate ozone production (P), D in mgll = P in Ib ozone/d/L 1 in mgd/8.34

Table 6-4. Terminology for Measured Ozone Pal'aml3ters

Parameter Symbol Metric Units English Conversions
Flow Measurements
Feed-gas m 3/min cfm x 0.02832
Off-gas m 3/min cfm x 0.02832
Exhaust-gas m 3/min cfm x 0.02832
Wastewater m 3/min mgd x 2.629
gpm x 0.003785

Ozone Concentration
Air Feed-gas
Feed-gas g/m 3 O/OVol x 19.96
Off-gas g/m 3 %Wt x 12.10
Exhaust-gas g/m 3 ppm Vol x 0.001996
ppm Wt. x 0.001210

Oxygen Feed-gas
. %Vol x 19.96
O/OWtx 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 20C 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/m 3 = (O/OVol)(Wj(10)
,,;, (M)(W)(1 000)/M-48) + (48001% Wt))
.= (ppm Vol)(W)(0.001)
= (ppm Wtj(Wj(M/48j(0.001)
%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 glmole
Air = 29 glmole .
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 glmole
Molar volumEI = 0.08205 UmolerK
Molar volume at temperature of 20 0 e =(273.15 + 20 0 K)* (0.08205 L/moleOK)
= 24.503 Llmole
W = (48 g/mc.le)/(24.053 Llmole) = 1.996 giL

Table 6-6. English Unit Equivalents for Ozone Concentration
Standard Pressure ...................... 14.696 psi (101 kPa)
Standard Temperature ........................... 68F (20C)
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 UmolerK

Metric English
Wt Ozone Percent Weight PPM Weight PPM
to Vol Percent
Gas Air Oxygen Volume Air Oxygen Volume
(g/m 3 ) (%) (%) (%) (ppm) (ppm) (ppm)
2 0.17 0.15 0.10 1,658 1,503 1,002
4 0.33 0.30 0.20 3,313 3,004 2,004
6 0.50 0.45 0.30 4,967 4,503 3,007
8 0.66 0.60 0.40 6,618 6,001 4,009
10 0.83 0.75 0.50 8,267 7,498 5,011

12 0.99 0.90 0.60 9,914 8,993 6,013

14 1.16 1.05 0.70 11,559 10,486 7,015
16 1.32 1.20 0.80 13,201 11,978 8,018
18 1.48 1.35 0.90 14,842 13,469 9,020
20 1.65 1.50 1.00 16,480 14,958 10,022

22 1.81 1.64 1.10 18,116 16,446 11,024

24 1.98 1.79 1.20 19,750 17,932 12,026
26 2.14 1.94 1.30 21,382 19,417 13,029
28 2.30 2.09 1.40 23,012 20,900 14,031
30 2.46 2.24 1.50 24,640 22,381 15,033

32 2.63 2.39 1.60 26,265 23,862 16,035

34 2.79 2.53 1.70 27,889 25,340 17,038
36 2.95 2.68 1.80 29,510 26,818 18,040
38 3.11 2.83 1.90 31,129 28,294 19,042
40 3.27 2.98 2.00 32,746 29,768 20,044

42 3.44' 3.12 2.10 34,362 31,241 21,046

44 3.60 3.27 2.20 35,974 32,712 22,049
46 3.76 3.42 2.31 37,585 34,182 23,051
48 3.92 3.57 2.41 39,194 35,651 24,053
50 4.08 3.71 2.51 40,801 37,118 25,055

Table 6-6. Terminology for Calculated Ozone Parameters

Parameter Symbol Units Calculation Formula
Ozone Production P g/min G1 x V1
Applied Ozone Dose D mg/l G1 X V1/L1
Transfer Efficiency
Precise TE % 100 * [(G 1 * V1) - (G 2 * V 2 )]/(G 1 * V1)
Approximate TE % 100 * (V1 - V2 )1Y 1
Transferred Ozone Dose T mgtl V1 x TE * G1tL,

To accurately calculate ozone TE several parameters the approximate calculation is adequate and elimi-
must be measured (See Table 6-6). However, a nates the need for installing a separate flow meter for
simplified and acceptable estimation can be obtained off-gas flow measurement. However, when precise
ifthe off-gas flow rate is assumed to approximate the data are required, for example for verification of
feed-gas flow rate, as shown in Table 6-6. -' design specifications, the precise method of calcu-
lating TE should be used. .
The error that occurs using the approximate TE
calculation is less than two percent for most design' Auxiliary Measurements and Calculations
situations (Le., actual TE is usually greater than 85 Other importa;lt parameters .in ozone terminology are
percent and off-gas flow rate.ls usually within 10 temperature, pressure, energy; power, moisture
percent of the feed-gas flo~ rate). In most applications content, and dew point. Moisture content and dew
point of the feed-gas are most important when air or design. The moisture content increases dramatically
recycled oxygen is used because the feed-gas mois- as the dew point temperature increases. For example,
ture content must bl3 reduced to very low levels in the initial moisture content of an air feed-gas with a
order to prevent damage to the internal components standard dew point of 20C (68F) is 17.7 g/m 3 (1.11
ofthe ozone generator. Moisture removal is typically Ib/1 ,000 ft\ whereas the desired moisture content
not required when hil~h purity oxygen is the feed-gas, ofthe treated feed-gas is only 0.008 g/m 3 (0.000507
because its moisture content is usually very low. Ib/1 ,000 ft 3 ) (dew point temperature of -76F (-60C)).
The dew point temperatures shown are for a standard
The relative moisture content ofthe feed-gas is called pressure of 101 kPa (14.7 psi), and must be adjusted
dew point temperatul!e, or "dew point." Dew point is for operating pressure conditions. The precise calcu-
the temperature at which a gas (at a specific pressure lation for dew point and moisture content is by means
condition) is saturated with water. If the temperature of psychrometric formulas or by use of psychrometric
of the gas decreases e,r pressure increases from those charts (21). However, the calculation may be approx-
conditions, the water will condense. Calculating dew imated by an inverse proportional relationship of
point is an important 131ement in ozone system design absolute pressures (gauge plus atmospheric pres-
because the size of some of the air treatment sure). A summary of atmospheric pressures at
equipment is a function of the feed-gas dew point different elevations is shown in Table 6-8 to assist in
temperature. the determination of absolute pressure.
The relationship between moisture content in air and
In ozone process design the moisture loading to the
dew pointtemperature at standard pressure is shown
desiccant dryer should be calculated. The desiccant
in Table 6-7 (21). The data is for the range of dew point
dryer is used to reach the required dryness of the
temperatures typically encountered in ozone process
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
Tablo 67. Moisture Content of Air for Air Temperature Example 6-1.
from -80 to 40C (21)
Moisture b Example 6-1. Calculation of Moisture Loading to
Moisture Weight a Content
Air Temperature Desiccant Dryer
Weight Volume Ib H2 0 per
re) (OF) Ib H2 0/lb air Ib/ft3 1,000 fe
Assume that a plant is at an elevation of 915 m (3,500
-80 -112 0.0000003168 0.07526 0.00002384 ft) above sea level (atmospheric pressure is 89.1 kPa
-75 -103 0.0000007713 0.07526 0.00005805
-70 -94 0.000001640 0.07526 0.0001234 (12.93 psi)); the feed-gas flow rate is 600 scfm (17
-65 -85 0.000003342 0.07526 0.0002515 m 3 /min); the gauge pressure is 93 kPa (13.5 psig); the
-60 -76 0.000006743 0.07526 0.0005074 inlet air temperature is 30C (86F); the relative
humidity is 60 percent; and the compressed air
-55 -67 0.00001311 0.07526 0.0009866
-50 -58 0.00002464 0.07526 0.001854 temperature is cooled with an after-cooler to 20C
-45 -49 0.000::14455 0.07526 0.003352 (68F). Determine the ambient air moisture content;
-40 -40 0.000'07925 0.07526 0.005964
-35 -31 0.0001381 0.07526 0.01039'3

-30 -22 0.000:2344 0.07526 0.017640

-25 -13 0.000:3903 0.07526 0.029373 Table 6-8. Atmospheric Pressure at Different Altitudes.
-20 -4 0.00015731 0.07526 0.050665
-15 5 0.0011l20 0.07526 0.07676 Altitude Atmospheric Pressure
-10 14 0.001'506 0.07526 0.1209
Feet Meters psi kPa Atm
-5 23 0.002485 0.07526 0.1870 0 0 14.70 101.325 1.000
0 32 0.003:788 0.07526 0.2851 500 152 14.43 99.49 0.982
5 41 0.005421 0.07526 0.4080 1,000 305 14.16 97.63 0.964
10 50 0.0071158 0.07526 0.5763 1,500 457 13.91 95.91 0.947
15 59 0.010/,9 0.07526 0.8045 2,000 610 13.66 94.18 0.930
2,500 762 13.41 92.46 0.912
20 68 0.014j'5 0.07526 1.1100 3,000 914 13.17 90.80 0.896
25 77 0.020~6 0.07526 1.5171 3,500 1,067 12.93 89.15 0.880
30 86 0.027~11 0.07526 2.0552 4,000 1,219 12.69 87.49 0.864
35 95 0.036i'3 0.07526 2.7642 4,500 1,372 12.46 85.91 0.848
40 104 0.04911 0.07526 3.6959 5,000 1,524 12.23 84.32 0.832
6,000 1,829 11.78 81.22 0.802
Air welghtlvolume is correc:ted to 68F and 1 atmosphere pres-
7,000 2,134 11.34 78.19 0.772
sure 8,000 2,438 10.91 75.22 0.742
9,000 2,743 10.50 72.39 0.714
-Air weightlvolume '" 1,205 g/m 3
bg/ m 3 ';"lb/1,000 ft3 * 16.01:2 10,000 3,048 10.10 69.64 0.687

the compressed, cooled feed-gas moisture content; The gas and liquid flows separate after the contact
and the moisture loading of the feed-gas to the basin. The off-gas is treated to remove the un-reacted
desiccant dryer. ozone and is reused, recycled, or discharged to the
atmosphere. The disinfected wastewater is dis-
Ambient air moisture content. From Table 6-7 the charged to the receiving water. Any residual ozone in
moisture content ofthe air at standard pressure and a the wastewater is reacte,d or is reverted back to
temperature of 30C (86F) is 32.1 g/m 3 (2.055 oxygen.
Ib/1 ,000 ft 3). This value must be adjusted for ambient
pressure and relative humidity conditions, and may 6.3.1 Feed-Gas Flow Schematic
be approximated as follows:
Ozone disinfection processes are typically distin-
2.055 Ib/1,000 fe * 14.70 psia/12.93 psia * guished by the type of feed-gas used. The most
0.60 = 1.401 Ib/1 ,000 ft3 common types are air-fed and oxygen-fed. Oxygen-
recycle systems have been used on some occasions.
or A diagram showing the components of these proc-
24.29 g/m * 1 atm/.88 atm * .60 =16.56 g/m 3 esses is presented in Figure 6-7. The air-fed system is
most common in plants where oxygen is not available.
Compressed, cooledfeed-gas moisture content. From Oxygen-fed systems are typically used in conj unction
Table 6-7 the moisture content of the feed-gas at a with an oxygen activated sludge treatment system,
temperature of 20C (68F) is 17.8 g/m 3 (1.110 where the unused oxygen from the ozone disinfection
Ib/1 ,000 fe). The absolute pressure is 182 kPa (26.43 system is lIsed in the biological treatment process.
psia) (i.e., 13.5 psi + 12.93 psi). The approximate
moisture content is: The quality of the feed-gas is critically important for
the electric discharge ozone generators. The feed-gas
1.110 Ib/1 ,000 ft3 * 14.70 psia/26.43 psia = must be oil-free, particle-free, and dry. To achieve
0.617 (b/1 ,000 ft 3 these characteristics the air-fed (and oxygen-recycle)
systems must pre-treat the gas to remove moisture
The moisture content of 9.88 g/m 3 (0.617 Ib/1 ,000 and particulates, and oil, if present. The once-through,
fe) is the maximum amount of water the feed-gas can oxygen-fect system typically pretreats for particulates
"hold" at 20C (68F) and 182 kPa (26.43 psia). This only, because theail and moisture content ofthe high
moisture content is 41 percent of the moisture purity oxygen from.the oxygen production facilities is
content ofthe ambient air. The remaining 59 percent negligible.
of the moisture will condense out of the air stream
and will be removed through a "moisture trap." The ozone generation equipment and power require-
Although not part of the question, the standard dew ments are about 50 percent lower for an oxygen-fed
point of the compressed air can be appf"Oximated system than for an air-fed system because approx-
using Table 6-7 as about 1 DoC (50F). imately twice as much ozone is produced with high
purity oxygen at a given input of power to the
Moisture loading to the desiccant dryer is approx- generator. However, this lower cost of producing
imated as follows: ozone is often offset by a higher cost of obtaining high
purity oxygen to feed the generator. On the other
0.617 Ib/1 ,000 ft 3 * 600 scfm * 60 min/hr/1 ,000 = hand, when the oxygen can also be used in the
22.21 Iblhr biological system the economics are more favorable
or using high purity oxygen.
3 3
9.88 g/m * 17 m /min= 167.9 g/min The ozone disinfection contacting units are similar for
all feed-gas systems shown in Figure 6-7. In the
6.3 Process Flow Schematics' contact basin some disinfection action occurs through
The ozone disinfection process consists of a liquid direct contact of the microorganism with the ozone in
and a gas flow scheme, as shown in Figure 6-6. The the gaseous phase; however, "effective" disinfection
feed-gas is typically air or high purity oxygen. A small appears to exist only when residual oxidants or an
amount (from 1 percent wt in air to 4 percent wt in ozone residual is present (16,17). In order to obtain
oxygen) of the feed-gas is converted to ozone within residual oxidants or residual ozone, a sufficient
the ozone generator. The ozone c'ontaining feed-gas amount of ozone must be transferred to the waste-
is directed to a contact basin where it combines with water.
the wastewater. Typically, a high percentage of the
ozone, but only a small percentage of the total feed~ Several investigators have documented that an
gas flow, is transferred to the liquid in the contact excellent correlation exists between the log-trans-
basin. The disinfection action proceeds in the contact ferred ozone dosage (log(T)) and log-coliform survival
basin. ratio (Iog(N/N o )) (17,18,22-25). Stover (17) reported
Figure 66. Ozone disinfection process gills and liquid flow diagram.

Gas Flow Stream Liquid Flow Stream

FeedGas Wastewater
Ozono Physical,
Generation and/or


Ozone To Off-Gas
Disinfection Treatment
Contact and
Basin Discharge,
or Re-Use
L ..... Wastewater Effluent
To Discharge

Figure 6-7. Diagrams showing feed-gas flow of typical ozone disinfection processes.

Air Fad

K . _
Destruction Vent

Oxygen Fed

Ozone Oxygen
High Purity Ozone Ozone Vent
Generator Activated
Oxygen Contacting Destruction Sludge

Oxygen Rec/cle
Dew Pont Ozone Ozone Ozone
Treatment ~ Generator f-+ Contacting f-+ Destruction ~ Vent Activated f-l. , Vent

.L- Make-up Oxygen

Recycle Oxygen

that this relationship exists irrespective of the type of oxygen requirement of 1.5 kg 02/kg BOD. The
feed-gas used; thus, the decision to use air or oxygen example also shows that the activated slUdge gas to
as the feed-gas is typically based on economic liquid ratio balances with the ozone process G1/L1
considerations of providing the oil-free, particle-free ratio at a feed-gas ozone concentration of 40 g/m 3 (3
dry gas to the generator and not on the disinfection percent wt) and an applied ozone dosage of 6 mg/l,
performance capability associated with the feed-gas. assuming 10 percent loss in the ozone contactor.
A common application for oxygen-fed ozone disinfec-
The G 1/L 1 ratio required for the ozone system is
tion is in conjunction with an oxygen activated sludge
shown as 10 percent greater than required for the
process where the oxygen is reused. Only about 10
activated sludge process, because about 10 percent
percent of the oxygen in the feed-gas is "used-up" in
of the oxygen is used up in the ozone process through
the ozone process; thus, 90 percent or more is
transfer of ozone and oxygen to the wastewater. For
available for the biological treatment process. The
the example in Figure 6-8, the adjus~d G 1 /L 1 ratio for
major design consideration is to balance the oxygen
the ozone system isO.15. Knowing the G 1 /L 1 ratio for
requirements of both processes. This can be ad-
ozone, either the applied ozone dosage can be
dressed by balancing the oxygen gas (G 1 ) to waste-
determined for a given feed-gas ozone concentration,
water liquid (L 1 ) flow rates to both processes.
or the required feed-gas ozone concentration can be
An example approach to balancing the gas to liquid estimated for a given applied ozone dosage. If the
ratios to both processes is shown in Figure 6-8. The oxygen balancing analysis indicated that significantly
first step is to estimate the applied BOD concentration more oxygen was required for ozone disinfection
to the oxygen activated sludge process. The activated because the ozone concentration was too high, then
sludge oxygen gas to wastewater liquid ratio can then an oxygen recycle process could be provided. If the
be approximated by estimating the applied oxygen analysis indicated that more oxygen was required for
requirement. The example in Figure 6-8 shows that the activated sludge process, then an oxygen bypass
the activated sludge G 1 /L 1 ratio is 0.135 for an pipe around the ozone disinfection system could be
applied BOD concentration of 120 mg/l and an provided.

Figure 6-8. Oxygen requirement for ozone disinfection compared to oxygen requirement for activated sludge.

0.6 r-----r--r-----,:----r---,---r-......- ......-....,.-,....--r---r---,....-r-----,:----r-~-__rr_ ........,..._r_-.,

10% Adjustment
0.5 for 02 Loss


a: 0.4



o 40 80 120 160 200 0 2 4 6 8 10

Applied BODs Concentration, mg/L Applied Ozone Dosage. mg/L

When oxygen is not used in the biological portion of water and a properly designed disinfection contact
the wastewater treatment process, then the cost of basin is provided (17). However, there appears to be
an air-fed ozone process may be compared with the economic considerations for the ozone disinfection
cost of a total oxygen-recycle system to determine the process associated with the type of wastewater
most economical alternative. The main disadvantages treatment processes selected prior to the ozone
of the oxygen-recycle system were the cost of disinfection system. The main effect of the waste-
obtaining oxygen-ri(:h feed-gas and the additional water treatment processes used is the impact on the
attention required tl::l handling an oxygen-enriched amount of ozone required to obtain the desirEld
gas. The main advantages were the reduced cost of disinfection level.
the ozone generaticm equipment and the reduced
energy consumption for ozone generation. Most Wastewater treatment schemes prior to ozone dis-
oxygen activated sludge plants in the United States infection that have been evaluated by various re-
that use ozone use the once-through, oxygen-fed searchers are shown in Figure 6-9 (17,24,25). These
system. As indicated in Table 6-1, all wastewater schemes range from simple "fine screening" to
treatment plants that use ozone disinfection, but do biological systems followed by chemical treatment.
not use oxygen activated sludge, use the air-fed Given and Smith et al. (24) evaluated ozone disinfec-
system. tion using the effluent from four types of wastewater
treatment process schemes, including fine screening
The ozone destruct units shown in Figure 6-7 are effluent, rotating biological contactor effluent, an-
fundamentally the sa:me for all feed-gas systems. The aerobic lagoon effluent, and a "strong" waste flow
size of the ozone destruct unit is based on the volume stream. The coliform survival results obtained over a
of gas directed to thEl unit. range of transferred ozone dosages is shown by the
dose/response curves in Figure 6-10.
6.3.2 Wastewater Flow Schematic
Ozone will disinfect to a very high degree if a The major factor influencing the amount of ozone
sufficient amount of ozone is applied to the waste- required to achieve a desired reduction in fecal

Figure 6-9. Exampisl treatment schemes using ozone disinfection (17,24,26).

Preliminary Treatment

SC:::~~ing J----------------------------I..-_D_is_?n_Zfo_e~_~_io_n ~ __

Secondary Treatment

Biological JI--------------'--__ _ o_z_o_n_e

Disinfection L.

Tertiary Treatment-Filtrlltion

- - - - - - - - - - - - - - I . . -__
Bi_O_IO_'.". J~--- L.._ _ F_i1_tra_t_io_n_ _ I-----:.IL..__D_i_S?_:f_~n_C~_io_n __~

Advanced Treatment

B'o'o,'' ' JI----I Chemical

Disinfection 10-

Figure 6-10. Fecal coliform survival for rotating biological Scheme C1 and Scheme C2. Scheme C1 had a nitrite-
contactor effluent, screened effluent, anaero- nitrogen concentration greater than 1.0 mg/I and
bic lagoon effluent, and strong wastewater
Scheme C2 a nitrite-nitrogen concentration less than
1.0 mg/1. The major factors'influencing the amount of
ozone required for disinfection were the influent
l!- =(~)-2'9 dissolved chemical oxygen demand (DCaD), the
No 1.8 influent nitrite-nitrogen concentration, andthetarget
n = 331 or desired effluent coliform concentration. The effect
r= -0.89 of the target effluent coliform concentration was
0.7 1 1.8 3.4 12.5 evaluated by comparing the ozone dosage required to
l~---,,"-T-'I,,_,.....\;:1-r--,...,...,..,.,...~I- ....---r--,-, 0 meet the former EPA standard of 200 fecal coliforms
" .... &
'" " oJ'&. <. per 100 ml and the more stringent California standard
',e~ ~
/ .~~ 90.i
= (CU/12.5)-2 99

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'

u 10-3
l!- = (CU
No 0.7
r 9

"" /
.... , n=16
,,~ = -0.93
'99.9 u.
California standard 100 percent of the time at an
ozone dosage of 6 mg/1. The other pre-treatment
effluents did not meet this high coliform standard
evenat ozone dosages of 10 mg/1. The major effect of
n = 13 NINo =(Cu/3;4r 4 .6
10-4 r = -0.82
carbon adsorption pre-treatment was to reduce the
n = 30 DCaD concentration of the wastewater and thus
r = -0.92 reduce its ozone demand.
10-5 ~----'~--'---'--::-.1-J....L..l-:-'=-
_ _..J...,..---l'--.L.....I 99.999
1 5 10 50 The less stringent disinfection standard of 200 fecal
Ozone Utilized, mg/L coliform/100 mL was achieved 100 percent of the
time for all wastewater treatment schemes evaluated,
Initial Conditions
except when the nitrite-nitrogen concentration was
greater than 1.0 mg/I (Scheme C1 ). The high nitrite-
Type of Waste nitrogen concentration of the wastewater influent to
Parameter RBC Screened Lagoon Strong the contact basin was caused by al1 upset in the
FC 6.5 x 104 1.3 x 106 3.8 x 105 2.8 X 106 activated sludge plant. The ozone reacted with the
TEMP 8.9 8.6 6.9 5.4
TURB nitrite causing an increase in ozone demand and a
5 42 67 151
pH 7.9 7.8 7.6 significant reduction in disinfection performance (25).
BOD 13 95 92
S5 13102 121 1010 The specific relationships between disinfection per-
V55 8 57 47 '620 formance and the DCaD 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/I, the best disinfection performance was achieved
with an influent nitrite-nitrogen concentration less
than 0.15 mg/I and a Dca 0 concentration less than
coliform concentration was the "initial ozone de-
12 mg/1. Based on these results, to keep the ozone
mand" of the wastewater, which was 0.7, 1.8 , 3.4 ,
dosage less than 10 mg/I, tertiary treatment of the
and 12.5 mg/I for the RBC, screened, anaerobic
wastewater may not be necessary to meet the former
lagoon and "strong" effluents, respectively (24). The
EPA standard (25). However, tertiary treatment to
slopes ofthe dose/response curves varied from -2.9: 1
reduce the ozone demand due to DCaD and nitrite-
to -4.6:1. Soll'fe difference in the slope may be due to
nitrogen appears necessary to meet the stringent
the type of wastewater treated. but the data set for all
California standard of 2.2 total coliforms per 100 ml.
wastes except the screened waste was too Ii mited (1 3
to 16 data points) to confirm a cause for the variation
in slope: Stover et al. (17) also evaluated the ozone dosage
requirements to meet two different disinfection
Gan et al. (25) evaluated ozone disinfection per- standards of 2.2 and 70 total coliforms per 100 ml.
formance of an activated sludge effluent and three The evaluation was completed for three different
tertiary treatment schemes following activated sludge wastewater treatment schemes consisting of a
treatment including: coagulation, sedimentation, and filtered activat,ed' sludge secondary effluent, a nitri-
filtration (Scheme A); coagulation and filtration fied effluent, and a filtered nitrified effluent. The most
(Scheme B); and carbon adsorption (Scheme C). The significant factors affecting the dosage requirement
carbon adsorption data set was further divided into were the organic quality of the wastewater as

Figure 6-11. Effect .)f water quality and performance criteria on ozone dosage requirement (25).

rn Ozone Residual =2 mg/L
., .....c Secondary Eff. Feed
N0 2 -N:50.15 mg/L
~., 80 Z 10-1
.~ "-

::IE z


1: 20
10 6 10 6
California Former EPA
Standard Standard
mg/L mg/L

E:ffects of Pretreatment Residual Dissolved COD, mg/L

em Meeting Coliform
Elacteria Sta ndards Effect Residual DCOD After Ozonation
on Total Coliform Survival Ratio

...: .....c
.= Tot. Ozone
Z Dosage = 6 mg/L z
a.. 10-1 a. 1
::!E DCOD ~2 mg/L ::!E 10-
Z 10- Z 2
a. a. 10-
::!E ::!E
0 .g-
0.;:; 10-3 <ll
a: Secondary
iii iii
> > Eff. Feed
:::l 10-4 Suspended Solids
rn rn
E E 023 mg/L
~ ~ 10-5 .11 mg/L
:= 10-5 (5
u u Ozone Dosage
C5 10-8 ~
I- 10- 6
0 0.4 0.8 0 2 3 4 5 6
Nitrite Nitrogen, mg/L Contact Column Number

E:ffect of Nitrite-Nitrogen on Comparison of Disinfection

Disinfection with Ozone Performance Between the High
and Low Suspended Solids

measured by COD concentration, the target effluent factors is illustrated in Figure 6-12, where the
coliform concentration and the influent coliform effluent total coliform concentration is shown to
concentration. The magnitude of th~ effect of these decrea~e as the transferred ozone dosage increases.

Stover et al. (17) evaluated the effect of the organic first two stage's "indicates that removal of suspended
concentration of the wastewater influent to the solids prior to ozonation may not be essential to
contact basin by comparing the transferred ozone achieve disinfection of wastewater."
dosage requirement for secondary and nitrified
effluents to obtain similar effluent total coliform
concentrations. For both disinfection standards eval- Figure 6-12. Total coliform concentration versus transfer-
uated, significantly more transferred ozone dosage red ozone dosage for various effluents (17).
was required for the filtered secondary effluent (COD
of 40 mg/I) than was required for the nitrified 105 r----,.---....---.----y---,------,
effluents (COD of 20 mg/I). Filtered Nitrified Nitrified
Log\( = 0.121X+2.179 LogY = 0.146X+3.12
Stover also showed that a more stringent effluent r = 0.63 r = 0.78
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 Secondary
filtered nitrified wastewater treatment scheme five Log Y = 0.088X + 3.47
times more transferred dosage was required to meet r = 0.77
the 2.2 per 100 ml standard (15 mg/I) than was
required to meet the 70 per 100 ml standard (3 mg/I).

Stover analyzed the effect ofthe influerittotaI coliform

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 effl uent 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
influent total coliform concentration was incorporated 2.2 ---------
in the analysis in Figure 6-13, where the log reduction 101..----..L._ _..L._ _....L_ _.......I l.,-_---I
of total coliform (Log No/N) is plotted against the log o 7 14 21 28 35 42
transferred ozone dosage. The results indicate no
difference in performance. Stover's conclusion was Transferred Ozone Dose, mg/L
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 Figure 6-13. Total coliform reduction versus log transferred
ozone dosage for nitrified effluents (17).
suspended solids tested (average 2.5 mg/l for the
filtered nitrified and '5.7 mg/l for the nitrified ef-
fluents). 6 r--.----,----.----,.--,--,-----,----,

Gan et.al. (25) also addressed the subject of filtration

for suspended solids removal prior to an ozone 5
disinfection process. The results ofGan's comparison .~
between a high (23 mg/I) and a low (11 mg/I) ~
"0 4
suspended solids influent to the ozone disinfection Cl)'
a: 0
0 0
contact basin are shown in Figure 6-11, where total E do 0

coliform survival is shown as a function of contact l3 3


column or stage in the disinfection contact basin. As "0 0

u 0 0 0

shown, the total coliform survival ratio was lower m 2

'0 8 e
(Le., better kill of coliforms) when the suspended 0 r9
t- o
solids concentration was lower. However, the effect C>
0 8
oo !!J o Nitrified
o 0 o Filtered Nitrified
of suspended solids diminished as. disinfection pro-
gressed through the six-stage contact basin. The
Log Y = 2.51 Log X + 0.76
r = 0.76
effect of suspended solids was greatest in the first OL--_--'-_----l._ _..I..--_--L.._--ll.-_...L-_--L_----I
two stages, and minimal effect was evidenced by the o 0.50 1.00 1.50 2.00
sixth stage. Gan concluded that the removal of (3.2) (10.0) (31.6) (100.0)
suspended solids (via flotation) that occurred in the Log Transferred Ozone Dosage (mg/I)

Venosa et al. (26) also studied the effect on disinfec- 3. Wastewater Quality Characteristics
tion of filtered and non-filtered secondary effluent.
The data indicated that improved disinfection ef- The ozone demand of the wastewater signi,f-
ficiency occurred as a result of filtration, but was icantly increases the ozone dosage requir~
more related to the reduction in tota I chern ica I oxygen ments. A plant with a large industrial
demand (TCOD) than to removal of TSS. The con- contribution may have a large ozon~ dosage
clusion was that filtration to remove TSS may not be requirement. Pilot testing to establish ozone
necessary when the TCOD concentration is low. dosage requirements in these plants is highly
The EPA Water Engineering Research Laboratory has Incomplete nitrification and a high concen-
also evaluated several secondary treatment plant tration of nitrite-nitrogen will significantly
effluents to detormine the relationship between increase the ozone demand and thus the
ozone dosage and total coliform reduction (27). The ozone dosage requirement. Thenitrite-nitro-
most significant f,actor influencing the ozone dosage gen concentration preferably should be less
requirement to achieve a desired effluent total than 0.15 mg/I to optimize disinfection
coliform concentration was the TCOD conc'entration performance.
of the effluent. At five plants where the TCOD of the
secondary effluent was less than 40 mg/I, a total 6.4 Ozone Equipment Design
coliform concentration of 1,000 per 100 ml could be Considerations
achieved with ozclne dosages between 4 and 7 mg/I.
Ozone generation is an established process; but its
However, when Meckes et al. (27) evaluated a plant
use in wastewater disinfection is relatively new. ,In
which treated a significant amount of industrial
this section the current state-of-the-art equipment
waste (TCOD =74 mg/I), a dosage greater than 12
design considerations to develop an ozone disinfe,c-
mg/I was projected in order for the process to meet
tion system that consistently achieves desired levels
the 1,000 total coliforms per 100 mllimit.
of performance is discussed.
Based on the data presented above, it appears that
there is no technical basis for excluding the use of 6.4.1 Ozone Generation Equipment
ozone following any treatment scheme. However, The basic components that comprise an ozone
depending upon the type of wastewater treated generator are depicted in Figure 6-14. The com-
and/or the effluent disinfection requirement, the ponents include an electrical source that suppli\3s
wastewater treatment scheme may be an important high voltage, alternating current across a discharge
economical consideration. A summary of the issues gap where the oxygen containing feed-gas passes; a
to consider when selecting the liquid flow schematic dielectric material that prevents electrical sholt-
prior to ozone disinfection is presented below: circuiting; and a heat removal mechanism that
prevents rapid decay ohhe ozone molecule back to an
1. Required Emuent Target oxygen molecule. Heat removal is required because
85 to 95 percent of the electrical energy supplied to
To meet the former EPA standard of 200 fecal the ozone generator produces heat (1 ,7).
coliforms per 100 ml, tertiary treatment may
not be necessary. Ozone Generation Theory
To meet more stringent standards, such as 14 Electrical power used to generate ozone is received
fecal coliforms per 100 ml, tertiary treatm~nt from a voltage regulator, and in some generators' a
should be considered. frequency regulator is also used. The altered current
To meet a standard of 2.2 total coli:forms per
100 ml, cldvanced treatment unit processes , .
prior to the ozone disinfection process may be Figure 6-14. Cross-section view of principal elements of a
required. Corona discharge ozone generator.

2. Influent Coliform Concentration Heat Removal

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,
Treatment processes that reduce the influent
coliform c:oncentration (such as filtration) will
decrease the ozone dosage required to
acheive a specific effluent standard. Heat Removal

is supplied to a number of ozone generation cells Figure 6-16a. Schematic diagram of a typical power supply
connected in parallel, as shown in Figure 6-15. Each to an ozone generator (1 I.
cell acts as a capacitor, as illustrated in Figure 6-15b.
The capacitance ofthe cell is a function ofthe width of
the gas space and the electrical conductivity of the Ozone
dielectric material. Because the dielectr.ic material /---+--'iI' ..r Generation
(glass or ceramic) is the major componentof the cell, Cells
the ozone generating cell is often ca lied a "dielectric." Electrical Regulation of Connected
Power Voltage/Frequency I in Parallel
Supply L __ I~_'-_...,
Ozone is produced when the ionization potential of
the dielectric is reached. The system's electrical Ground
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 Figure 6-15b. Schematic diagram of an ozone producing
the ionization potehtial is reached a flow of electrons cell, a "Dielectric" (1 I.
(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).
High . , Low
When the voltage is greater than the ionization Tension Electrode liT ,",100 "",,,ad.
potential of the dielectric the electrons travel from
one electrode to the other electrode within the Dielectric -----' L Discharge Gap
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 Figure 6-16. A free flow of electrons in the discharge gap
causes various reactions with the' oxygen
formed will vary from none at all to a maximum oftwo molecule (1).
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.
e_ Results in (a)


An alternating electrical current must be used in

ozone generation; thus, the voltage will cycle above
e..- (b)
and below the ionization potential of the cell (1).
Ozone production occurs when the voltage is greater
than the ionization potential. Ozone production is or
terminated in the portion of the cycle when the
voltage is below the ionization potential, as shown in
Figure 6-17. The amount oftime ozone is formed is
dependent upon the frequency of the power supply.
0--0 = Oxygen Molecule
The number of free electrons discharged is a function
~ = Free Electron
of the applied, peak voltage and its electrical fre-
quency.lfthe peak voltage is significantly higher than
the ionization potential of the dielectric, a propor-
A = Ozone Molecule
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 Design Considerations for Ozone
supply increases the amount of time that the ioniza- Generation Equipment
tion voltage will be reached, which also increases the The amoun,t of ozone produced by an ozone generator
ozone production rate if the temperature of the is affected by the physical characteristics of the
ozonized gas is acceptable. equipment, the power supply to the generator; the
Figure 6-17. Ozone formation occurs when the v4)ltage the square of the voltage. Frequency or voltage
level is sufficient to create a free flow of changes are used to adjust the ozone production rate
electronls within the discharge gap (1). for a given ozone generator. The production rate also
increases with a greater number of dielectrics in
service. Because of these relationships, almost any
Period desired production rate can be supplied by the ozone
I Period of equipment manufacturers.
~---......-.....'- ['eionization
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
ofthe discharge gap may vary slightly from generator
Correspondingly. in the Discharge Gap: to generator, even with generators of the same
0-0 '" Oxygen manufacturer. A very slight change in either param-
..... '" Free Electron & Direction of Travel eter has a large influence on the ozone production
V '" Ozone rate. As such" the ozone production rate may vary
g.r--=-II-----;rl--.--t-7--;::---;::-t---+::--;r---;:;--;:-r- slightly for similar sized ozone g~nerators, and from
<; ~b~ ~ ~ ~ ~~ ~ t t i: ~g~ ~ ~~i generator to generator.
~ \D..oJ' i '\; 6 ~ cx:8 8 0.0 o-egp ~g 0.0
.fi 0-0 ~: i ~ ~~ o-D t fJb ~~ ~6" ~ g: Ozone Generator Power Supply. Three primary ways
i5 Cycle in which the power is supplied to the ozone generator
r3 Repeats are shown in Figure 6-18 (1). Method A is a low
E, frequency (typically 60 hertz), variable voltage system.
(J) This system is most common because high voltage
Q) N Ql III transformer technology preceded the technology in
0) 0 -Ql
'i..E! 0.0 high frequency transformers. Method B is a medium
E, "'" EO.
C) u o E frequency (up to 600 hertz), variable voltage power
<..) m u 8 supply system. This process has been used to
c: III c: Ql
C) III 00 increase the production rate of installed ozonte
'';:; ~ '';:::: M
~J .~ 130 equipment(1). Method C is a variable high frequency,
'~n! '2
.2 E
constant voltage unit. Method Cis used by various
'(j).~ Ql 0
CIO ocn ozone generator manufacturers in lieu of the variable
voltage process. No one method has a clear advantage
over the other.
moisture content and dust content of the feed-gas; The most common ways in which voltage and
the temperature of thel ozonized gas, and the feed-gas frequency to the ozone generator are controlled are
oxygen content. Each factor is further discussed. shown in Figure 6-19 (1). .The voltage controlled
system uses single phase power, which will cause an
Physical Characteristics of Ozone Generator. The imbalance in the amperage of a three-phase poWer
relationship of the factors affecting the ozone produc- supply system unless three or multiples of three
tion rate of the generator dielectric is shown in ozone generators are operating simultaneously.
Equation 6-2 (7). However, this capability is seldom available; thus, the
P- 4k f Va [Cd(Vo-Va) - VaCdj (6-2) impact of an unbalanced electrical load on the overaU
plant three-phase power supply should be evaluated.
where: To balance the electrical load a Scott transformer is
P = Ozone production typically used (29).
Va = Voltage applied to the discharge gap
Vo = Extreme or peak value of the applied voltage The power supply to the ozone system is typically at a
Cd =
Capacitance of dielectric frequency of 60 cycles and potential of 480 volts. For
Ca =
Capacitance of the discharge gap a variable voltage system the voltage may be in- .
f = Frequency of current supply creased by applying the line current to one of a series
k = Constant of tappings on the primary side of the potential
transformer, thereby changing the transformer ratio
The most important point in the equation is that the and the resultant secondary voltage (1). Another
ozone production ratel increases with frequency and method is to use a variable autotransformer to feed

Figure 6-18. Schematic diagram of three power supply Figure 6-19. Typical ways for varying voltage and frequency
systems typically used for ozone generation to an ozone generator (1 ).
(1 ).

Method A
Fixed Line Low Frequency (60 Hz), Variable Voltage ( V'd'~.
~ '" Ozo.ne

Ozone '---v----'
\ = ~.J,~=
Generator Line Voltage Auto Inductance High Tension
Power Powerstat High Tension at 60 Hz Transformer to Correct Transformer
Source or Variable Step-Up (Sometimes Power Factor
Transformer Transformer Connected to a

[v1ethod B Variable Voltage

Fixed Medium Frequency (600 Hz), Variable Voltage Four Thyristers

Adjust Rectified
Power and Change
to Desired Frequency
,Ozone (This is Where Control
HIT ;( is Exercised)
Frequency High Tension
Converter Step-Up '/ Parallel
Method C
Fixed Voltage Variable Frequency

Ozone '---v----'
Rectifier Bridge \ Ozone
HITI----' Generator High Tension
(Convert AC to DC) Generator
Power High Tension Transformer
Source Step-Up
Transformer Variable Frequency

the primary sfde of the main transformer (1 ). When a ever, the specifications should require that the
variable frequency system is used both the frequency frequency and voltage transformers be high quality
and voltage must be increased. The voltage must be units designed for ozone service. The ozone generator
increased to a level above the ionization potential of supplier should be requested to provide a record of
the dielectric (about 10,000 volts). The frequency is successful electrical equipment performance.
increased up to a maximum of 2,000 hertz (1 ).
Another item to consider in the design of the ozone
Because the ozone generator uses high voltage and in generation system is power factor. An operating
some cases also high frequency electrical current, ozone generator can decrease the power factor to 0.3
special electrical design considerations must be to 0.5, depending on the generator setting (29).
implemented. For example, special insulation must Corrections will normally be cost effective, since
be provided for the electrical wire; a cool environment utilities that supply electrical power typically impose
forthe high voltage transformers should be provided; penalties for a low power factor.
and the electrical transformers should be protected
from ozone contamination due to minute ozone leaks Power factor is the ratio between the apparentpower
t/:lat could occur on a periodic basis. (kW) measured by a watt-hour meter and actual
power (kVA) measured in terms of voltage and
The electrical considerations for an ozone system amperage. This relationship is shown in Equation
should receive special attention. For example, a 6-3.
number of problems have been reported with dry-type Power Factor = Apparent powerI Actual power
potential transformers (1). Oil-cooled transformers pf = kW/kVA . (6-3)
apparently have performed more reliably. In view of
the dependence of ozone generation on high fre- The power factor is UNITY when the voltage and
quency or high voltage electrical energy, the ozone current of an alternating current power supply are
generator supplier should be responsible for design- "in-phase" with each other, for example in a purely
ing and supplying the electrical subsystems. How- resistive circuit like a heating element. In a purely
capacitive circuit, like an ozone generator that has not electrics must be cleaned before the generator is
reached the ionization potential of the dielectric, the placed back into service.
voltage and current are 90 degrees out of phase. In
this case the powel: factor is ZERO. For a generator A high moisture content of the feed-gas not only
producing ozone the voltage and amperage will be causes damag~ to ozone generator dielectrics and
somewhere betweein 0 and 90 degrees out of phase; increased generator maintenance, but also will
thus, the power factor will be less than 1.0. The actual decrease the ozone production rate at a given power
power factor will val'y depending on the power supply setting of the generator. Ozone production begins to,
to the ozone genemtor and the amount of electrical decrea.se when the dew point temperature exceeds
resistance developed within the electrical circuit. -51C (-60 0 F)(31 ,32). To prevent the negative effects
of moisture content on ozone production and gen-
The power factor may be corrected by installing erator maintenance, several authors suggest a min-
inductors in the electrical circuit or by using the imum dew point temperature of the feed-gas.
inductance created by the operation of motors in
other areas ofthe treatment plant. However, caution Manley and Niegowski (7) suggest that the feed-gas
must be exercised when using other plant equipment moisture content should be less than 0.02 to 0.03
for power factor correction, because of the variable grams of water per cubic meter of air (i.e., dew point
operating conditions of the equipment from hour to temperature less than -53C (-63F. Varas et al. (33)
hour and from day to day. The inductance of other recommend that the dew point temperature be less
equipment should only be used to control the low than -60C (-76F). Robson (34) suggests that less
power factor of the ()zone generator when consistent maintenance and prolonged dielectric life occurs
equipment operation can be assured. when the dew pdint temperature of the feed-gas is
-60C (-76F) or lower. Gerval (35) states that the
Moisture and Dust Content of the Feed-Gas. The "highest value to be guaranteed for dew point should
moisture content of the feed-gas to the ozone be -50C (-58F) ." Chapsal (36) recommends that the
generator has a twel-fold influence on ozone produc- dew point temperature range between -50 and -80C
tion. A high moisture level not only decreases the (-58 and -112F). Damaz (37) recommends a dew
ozone production rate, it also increases the rate of point of -60C (-76F) to protect the generator equip-
contamination of the ozone generator dieh3ctrics. ment.
Contamination occurs whether the feed-gas is high
purity oxygen or air. If oxygen is the feed-gas, .Based on these sources it appears that the dew point
hydrogen peroxide isformed in the presence of water temperature should not be higher than -50C, in
vapor and forms dl3posits on the dielectrics; these order to achieve maximum output from the ozone
deposits can be removed by scrubbing with soapy generator. However, in order to reduce maintenance'
requirements and prolong dielectric life, a dew point
water (7).
temperature equal to or less than -60C appears
If air is the feed-gas, about one mole of nitrogen' warranted. To achieve either dew point condition,
pentoxide (N 2 0 s) will develop for everyone hundred extensive feed-gas treatment is requiredfor air-fed
moles of ozone formed (30). Nitrogen pentoxide can and oxygen recycle ozone generation systems.
decompose to nitrogen dioxide (N0 2 ), which interferes
directly with ozone output (30). When nitrogen Damaz (38) presents other design considerations to
pentoxide is in the presence of water vapor, nitric acid prevent moisture contamination of the ozone gener-
will form (7). The nitric acid will be deposited on solid ator. Moist air back-flow from the contact basin
surfaces inside the ozone generator and piping, and should be avoided by installing a leak-proof check
will enhance cormsion of metal surfaces (7). In valve in the piping from the ozone generator to the
addition, the nitric acid can create a heatsink on the ozone contact basin. Also, the ozone generator must
glass or ceramic dielectrics and will increase the be purged with dry feed-gas prior to start-up of th13
potential for dielectric breakage. ozone generator. Most operators purge with dry feed-
gas for a minimum cif eight hours prior to start-up of
In normal practice (i.e., air dew point less than -50C the ozone generator. It should be noted that this time
(-58F)), about 3g to 5g of nitric acid are formed for period does not include the time required to obtain the
every 1,000 g of ozone formed (30). Nitric acid can be dry feed-gas from the dessicant dryers.
easily removed by scrubbing with soapy water. This
task should be considered as routine maintenance on Dust and organics in the feed-gas can also create
a 12-01onth interval. Significantly greater quantities operating problems with the ozone generator (35).
of nitric acid will fOl"m when the dew point temper- The dust can collect on the dielectrics, decrease
ature of the air is higher than -50C (-581=). If a generator efficiency, increase dielectric stress and
malfunction occurs and excessive moisture is dis- cause unnecessary dielectric breakage. A filter
charged to an operating ozone generator, the di- should be installed prior to the ozone generator to

capture the dust and organics. According to Gerval ozone generator cooling water is often treated to
(35), the filter(s) should be able to remove 99 percent obtain "boiler" quality water, in order to prevent
ofthe particles greater than 1 micron in diameter and scaling or corrosion in the cooling loop. For example,
98 percent of the particles greater than 0.4 micron in water with a high chloride concentration h~s been
diameter. Rakness et al. (18) also recommend two reported to "attack" stainless steel. A cost-effective
filters in series, the first a 1 micron filter and the analysis can be completed to determine if it is
second a 0.3 micron filter. It is emphasized that filters economically feasible to recycle the cooling water, or
should be installed for oxygen as well as air feed-gas to simply purchase, use and discharge the ozone
systems. generator cooling water. In instances where the
source of cooling water is too hot, a refrigeration
Ozone Generator Cooling. Ozone production increas- cooling system has been used (1 ).
es when the temperature of the ozonized gas is
minimized. In addition, reduced heat build-up will Air cooled ozone generators require that the gener-
increase life expectancy of the dielectrics. The feed- ator room ventilation system be an integral part ofthe
gas flow rate and the heat removal capability of the ozone system design. Experience from operating
ozone generator are factors influencing the temper- plants indicates that the cooling airmust be dust-free
ature of the ozonized gas. The feed-gas flow rate is and oil-free in order to prevent electrical short-
typically established by the design engineer through circuiting on the high voltage electrical supply
the selection of the maximum ozone concentration. system. The heated air may be used for plant heating,
The heat removal capability of the ozone generator is but care must be taken to avoid ozone contamination
governed by the ozone generator equipment manu- of the plant buildings in case of ozone leakage.
facturer and by the temperature of the cooling
medium. Feed-Gas Oxygen Content and Flow Rate. The feed~
gas flow rate and oxygen content can be used to
Because cooling is a major aspect of the energy change the ozone concentration from the ozone
efficiency of an ozone generator, the cooling methods generator, and thus change the rate of ozone
utilized by the ozone generator manufacturers are production. When the oxygen concentration of the
highly competitive. The discussion of generator feed-gas is greater, a higher concentration of ozone
cooling presented in this manual addresses general can be obtained for a given generator power setting.
design considerations. The various ozone generator The ozone concentration increases because the
manufacturers must be contacted to determine the potential increases for collisions between the elec-
specific cooling requirements for their individual trons released across the discharge gap and the
ozone generators. oxygen molecules within the gap. In some cases
oxygen enriched air has been used to upgrade the
Depending on the type of ozone generator, cooling is capacity of an existing ozone generator (39). However,
accomplished with either water, oil or freon plus the most common use of an oxygen-fed ozone
water, or air. In order to optimize electrical energy disinfection system is direct use of high purity oxygen
efficiently, the cooling water temperature should be and re-use of the oxygen in the actiyated sludge
20C (68F) or less. Approximately 3 to 4 liters (0.75 process.
to 1.0 gal) of 20C (68F) cooling water is required for
each gram of ozone produced. Different ozone A lower feed-gas flow rate will increase the ozone
generators will vary on the amount of cooling water concentration. However, the specific energy con-
required. A common design parameter is that the sumption will be greater because of less cooling
temperature rise of the cooling water should not capability by the air flow. This relationship is shown in
exceed 5C (41F). It is noted that some ozone Figure 6-20 for a typical air-fed ozone generator. As
generators have operated successfully with cooling shown, the' specific energy consumption gradually
water temperatures as high as 40C (104F). How- increases from an average 12.1Whl g (5.5 kWh/lb) at
ever, the design engineer must be aware that ozone an ozone concentration of 4.8 g/m 3 (0.4 percent wt)
production capacity is lowered, electrical energy to an average 23.2 Whig (10.5 kWh/lb) at an ozone
efficiency is reduced, and dielectric glass is stressed concentration of 30.2 g/m 3 (2.5 percent wt).lt should
during operation with warm water temperatures. be noted that operation at a lower ozone concentra-
Equipment capacity must be de-rated for operation tion tends to reduce ozone generat.ion specific energy,
with warm temperatures. . but will require a higher feed-gas flow rate and higher
air treatment cost. A balance between these two
The cooling water source can be potable water or opposing factors may be evaluated. The maximum
good quality non-potable water. The cooling system ozone concentration is typically 1.5 percent wt, where
may be closed-loop or once through and discharge. the average specific energy is 17.6Wh/g (8 kWh/lb).
Typically, potable water in a closed-loop cooling
system is used with the potable water re-cooled with In Figure 6-20 a range of specific energy consump-
water from the plant non-potable water supply. The tions is shown to account for different cooling
capabilities of various; ozone generators and different Figure 620. Specific energy consumption versus ozoll1e
feed-gas temperaturos that may be encountered. It concentration for an air-fed ozone generator.
should be noted tha1t the data shown is for a high Ozone Concentration (% by Weight)
quality feed-gas thmugh clean ozone generators;
thus, Figure 6-20 can be used to estimate energy 30 0;:-_---=0:.;.5=--_ _=_.:1.;:.0_ _--=1:.;.:.5=--_--=2:;..0=--_--=2:.;.5
consumption for typical ozone generation systems. 13
The manufacturer should be contacted for more 12 :Q
precise information on performance characteristics ~
25 ::::::
11 .r.: .
of individual ozone gEmerators. ~ 10 ~
'g, 20 9 .'J
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 15 7 ~
may be used to evalUl3te specific energy requirement u 6 u
relative to ozone production and assist in the selection Ei 5 ~
of feed-gas flow rate. An example g'enerator mappi ng ~ 10
w 4 Ln
curve based on actual test results is shown in Figure ~ <>
6-21 (17). The approach to developing the generator '(j 3 ~
~ 5 OJ
mapping curve is as ~ollows: (1) Select a desired gas (IJ 2 (~
flow rate, (2) Calculate the ozone production at a 1
selected ozone concEmtration, (3) Plot the data point o 01::-------!-5--1,-l0,----.J1L5 -----:2:l::0-----,2~5,------:3,.,J0 0
(i.e., calculated ozone production at specific energy
requirement for corrElsponding ozone concentration), Ozone Concentration (g/m 3 )
(4) Repeat steps 1-3 for various ozone concentrations,
and (5) Select another gas flow rate and repeat steps
1-4. Figure 6-21. Example ozone generator mapping curve
using air feed-gas (17).
The design ozone concentration for an air-fed ozone 600 r--.--,---.--,-----,--,-----,--,-----,--,-----,----.
generator typically rSlnges from 12 to 24 gl m (1 to 2
percent wt). A concentration of 18 g/m (1.5 porcent
wt) is most common. While individual ozone gener- .c
ators may achieve concentrations greater than de- '-
sign, it is recommended that the discharge ozone ;400
concentration from all ozone generators be main- .~ - - - - - 2 . 8 lis
tained at or below design conditions in order to -6o 300 2,1 lis
maximize generator cooling effectiveness. Als,o, it is ll:
recommended that individual ozone generators not ~ 200
be operated at power settings greater than 75 percent 2
of maximum, unless necessary. For example, rather 100
than operate one ozone generator at 100 percent of
capacity, 2 ozone generators should be operated at 50 o L-~_L.____L._-----JL-....l-__.l_...L-____L_-!-_L_.L----l

percent capacity. These operating prac~iceswillcause 12 14 16 18 20 22 24 26 28 30 32 34 36

less stress on gene'rator dielectrics, will decrease Specific Energy (Whig)
generator maintenance problems, and will usually
minimize electrical consumption. .
consumption of the ozone disinfection system. More
The typical range of ozone concentration and specific information on energy consumption on auxiliary
energy consumption for an oxygen-fed ozone gener-
3 equipment is discussed in Section 6.5.2.
ator is shown in Figure 6-22. The typical (40 g/m )
design ozone concentration for an oxygen-fed ozone Ozone Generator Monitoring andAlarms. The follow-
generator at 3 percent wt' is twice the concentration' ing are required to adequately monitor ozone genel's-
for an air-feed systom. Also, the specific energy is tion equipment:
about half of an air-feed system.
a. Inlet feed-gas flow rate-Monitor system load-
It should be noted that the energy consumption ing.
information shown in Figure 6-20 and Figure 6-22 is
for the ozone genel'ator only. These data must be b. Inlet feed-gas temperature-Monitor system
coupled with the enEtrgy consumption ofthe auxiliary operation. Also provide alarm and shut-down
equipment in order to determine the overall energy device.
Figure 6-22. Specific energy consumption versus ozone formance. This meter and recorder may be
concentration for an oxygen fed ozone gen- considered optional equipment on very small
erator. systems; but the operators should at least be
provided with the equipment and procedures for
Ozone Concentration (% by Weight) determining ozone concentration using wet-
0 1.0 2.0 3.0 4.0 chemistry methods.

6 :0- k. The cooling water system should be designed to

Ci :::::: achieve adequate cooling for the worst case
:2 12.5 .r:

~ 5
g conditions.
c: c:
0 0
aE 10

Co I. Cooling water system monitoring equipment

4 E
O~'l~e{\ . ::J should provide detailed information regarding
c: c: system operation:
0 7.5 0
u 3 u>-
Water temperature-Monitor system opera-
c: 5 c: tion. Also provide alarm and shut-down.
0 2 wu Inlet water flow rate-Monitor system opera-
;;::: ;;:::
'0 '0 tion. Also provide alarm and shut-down.
Ql Ql
2.5 Co
(J) Outlet water temperature-Monitor system
Heat exchanger cooling water inlet and outlet
00 0 temperature and inlet water flow rate-
10 20 30 40 50 60
Ozone Concentration (g/m 3 ) . Monitor system operation.

The cooling water system is vital to the

operation of the water-cooled ozone genera-
c. Inlet pressure-Monitor system operation. tors. The operators should be provided with
the instrumentation to be able to make control
d. Discharge ozonized gas temperature-Monitor adjustments, as necessary, to maximize
system operation. Also provide alarm and generator cooling. The alarm and shut-down
shut':down. on the water flow rate and temperature are
provided to protect the generator dielectrics
e. Discharge ozone concentration with in-line from breakage due to heat stress.
ozone meter-Monitor system performance.
Also provide a recorder. Monitor may be con-
sidered optional for smaller installations. m. Cooling air system monitoring equipment should
provide detailed information regarding system
f. Discharge ozone concentration using wet-chem- operation.
istry procedures (40)-Monitor system concen-
tration. Ambient air temperature-Monitor system
g. Inlet voltage-Monitor system loading.
The air cooling system is primarily dependent
h. Inlet amperage-Monitor system loading. upon the room temperature. The "cooling"
fans typically operate at a constant speed.
i. Inlet frequency-Monitor system loading. The cleanliness ofthe cooling air is important
to proper cooling, although no direct measur-
j. System watt-hour meter-Monitor system ing is typically done to monitor this parameter.
energy. May be considered optional for smaller The air should be as cool as possible to
installations. maximize generator cooling, and as clean as
possible to prevent dust accumulation on the
The operator shou Id know the various aspects of cooling surface. Also, no oil should be used,
system loading, performance and operation to because an oil film will develop on the cooling
be able to determine production levels and surface.
operating characteristics. The alarm and shut-
down on the inlet and discharge gas temperature
is used to protect the generator from damage. n. Inlet feed-gas temperature with alarm and shut-
The recorder on the discharge ozone concentra- down-Monitor system operation and protect
tion meter is used to monitor generator per- dielectrics.

o. Inlet feed-gas dew point with alarm and shut- Figure 6-23. Details of a horizontal tube, voltage controlled,
down-Monitor system operation and protect water cooled ozone generator (1).

p. Safety inter-lock to prevent power energizing B E

during generator cleaning or repair. 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. Thle cooling medium may be water,
water plus oil or freon, and air. The physical ar-
rangement ohhe dielectrics is typically either vertical C
or horizontal.

Horizontal Tube, Voltage Controlled, Water Cooled.

The horizontal tube, voltage-controlled, water-cooled
ozone generator iB 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 siteel jacket acts as the low tension
electrode, and contains multiple, cylindrical tubes
where glass dielectrics are inserted. The internal side
ohhe glass dielectric is coated with a metallic coating H
which acts as the high tension electrode. The feed-
gas passes betweEIO 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 G J F E K
60 hertz, although operation at frequencies from 600
to 800 hertz has been practiced (1). The horizontal Legend
tube, voltage-controlled, water-cooled ozone gener-
A-Air Inlet G-Tube Support
ators are installed at a number of wastewater B-Ozonized Air Outlet H-H.V. Terminal
treatment plants int~luding: Frankfort, Kentucky; Vail, C-Coolant Inlet I-Port
Colorado; Pensacola, Florida; Murfreesboro, Tennes- D-Coolant Outlet J-Metallic Coating
see; Brookings, South Dakota; and Olympia, Wash- E-Dielectric Tube K-Contact
ington. F-Discharge Zone

Vertical Tube, Voltage Controlled, Water-Cooled. The

vertical tube, voltage controlled, water-cooled ozone
generator utilizes the cooling water both as the which is created between the tubular stainless steel,
grounding electrode and the coolant. Details of the high tension electrode and the glass dielectric, low
generation system eire shown in Figure 6-24 (1 ).Input tension electrode. The ozonized gas is discharged into
gas is "pulled" through the dielectrics by means of a the middle compartment, from where it is drawn to
vacuum created in 'the ozone contacting system. the ozone contactor.

The ozone generator consists ohhree compartments. The vertical tube, voltage-controlled, water-cooled
The feed-gas is drawn into the upper compartment ozone generator is coupled with a proprietary,
where it enters hollow metal tubes which are the high aspirating turbine mixer contacting system. The
tension electrodes. The gas is drawn downward process was installed at the first wastewater treat-
within the tubular metal electrode to emerge into the ment plant to use ozone disinfection in the United
closed end of a glass dielectric tube. The feed-gas States (Indiantown, Florida), andwas evaluated in an
then passes upward through the corona discharge, EPA research study (17).

Figure 6-24. Details of vertical tube, voltage controlled, Figure 6-25. Details of a vertical-tube, frequency control-
water cooled ozone generator (1 ). led, double-cooled, ozone generator (1).

r - - - - - ,...... Cooling Water

~( Dry Air Inlet Discharge

... E
o. m
, 0"f Aidnlet I ~ r Ozone-Rich
Gas Outlet
Glass Dielectric ' \

~E High Tension
o High Voltage
u Fuse

Ozone-Rich Gas
IIII / Cooling
~~"li;~""~,---.jlllllllf"""_"",_~,,,,,_-~ Water Level
E Glass Tube
Dielectric Gas Flow in Ozone Generator Tube
E c:
8... I _
Dry Air.l,
Cooling IFjF==jI=lf=lI==IJ:=li=t1 u.
.~ . Water --:':Ozone-Rich
. Spacer Pm Inlet -.~ .


Ozone Generator Tube Typical Ozone Generator

Vertical Tube, Frequency-Controlled, Double-Cooled.

The vertical tube, frequency-controlled, double-
Stainl... Sta.,
.Ground Electrode
cooled ozone generator was invented by Bollyky (U.S. Cooling Water Ozone Out
Patents 3,766,051; 3,833,492). The generator utilizes Inlet
both water and a /Jon-conducting liquid (oil or freon)
for cooling. The unit incorporates a complex dielectric
design to achieve additional cooling, as illustrated in Lowther Plate, Frequency-Controlled, Air-Cooled.
Figure 6-25 (1). The high tension electrode is cooled The basic elements of the Lowther plate, frequency-
by the non-conducting liquid that is contained in a controlled, air-cooled ozone generator are shown in
closed-loop cooling system. The low tension elec- Figure 6-26 (1). The dielectric consists of an alum-
trode is cooled by water. inum heat dissipator, a steel electrode coated with a
ceramic material, a silicone rubber spacer to establish
The dual-cooled generator is typically operated at the discharge gap, a second ceramic coated steel
10,000 volts, a frequency of 2,000 hertz, and a electrode with gas inlet and ozonized gas outlet, and a
pressure up to 137 kPa (20 psig). A disadvantage of second aluminum heat dissipator. Thirty to forty
this unit is the higher capital cost associated with dielectrics are manifolded to form a module. Atypical
more involved cooling equipment. However, the Lowther plate module is also shown in Figure 6-26.
generator energy efficiency is reported to be better The unit uses ambient air for cooling and operates at
due to the additional cooling provided and is reported an upper frequency range of 2,000 hertz, at a
to require fewer electrodes. potential of 9,000 volts. The maximum operating
pressure is 103 kPa (15 psig).
Wastewater treatment plant installations for the
vertical tube, frequency-controlled, double-cooled A Lowther plate ozone generator was installed at
ozone generator include Indianapolis, Indiana; Springfield, Missouri; Holland, Michigan; Moorhead,
Granby, Colorado; Alburndale, Florida; West Knox- Minnesota; and Concord, North Carolina. Another
ville, Florida; Summers, New York; and Hunter, New type of air-cooled generator was installed at the
York. Mission Viejo plant near Denver, Colorado (41).

Agul'8 6-26. Details of an air-cooled, Lowther plate type a critical aspect in ozone system design. The major
ozone lIenerator (1 I. equipment design considerations are the quality of
the treated air and the reliability of the treatment
equipment. 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 ((b/1,000 fe), but the design spec-
ification for moisture content is typically expressed as
8'7" dew point temperature. The relationship between
, moisture content and dew point temperature is
Control discussed in Section
Panel Solid State Excessive moisture content in the feed-gas to the
~ ~
Behind Louvered
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 maintle-
Air Intake
:l1;'o6" ~"
nance problems, it is recommended that the dew
point temperature of the feed-gas to all corona
Silicone discharge ozone generators not exceed-60oe (-76F),

Ground unless absolutely n'ecessary. Operation at a dew
High Voltage _ A rSteel Separator point temperature greater than -50 oe (-58F) should
Steel ElectrOde) Electrode be avoided altogether.
Aluminum ->- 03 -0(
The primary function of the air treatment system is to
Heat Disslpator It[)iSCha~~e Gap remove moisture from the air. Moisture can be
Ceramic ~ ~~~~~ removed by increasing pressure, by cooling, or by
adsorption techniques (21). Equipment used to rt3-
Dielectric L.. A Silicone Section A-A move moisture is further described. The procedure for
Separator calculating moisture content at various temperature
and pressure conditions is disQussed in Section Specific Manufacturers' Ozone Generators
Ozone generators manufactured by different com- The particu late content of the feed-gas should aIso be
panies have unique characteristics, but also have minimized in order to optimize generator production
some common requirements. For example, all ozone efficiency and reduce maintenance. Prior to the ozone
generators should have a similar quality feed-gas, generators, filters which remove particles greater
and most ozone genElrators can operate independent than 1 micron in diameter followed by filters which
of the type of contae:t basin used. A listing of ozone remove particles greater than 0.3 to 0.4 micron in
generator manufacturers can be obtained from the diameter are recommended (18,35).
International Ozone Association (lOA). Air Feed-Gas Treatment Systems
6.4.2 Feed-Gas Supply and Treatment Equipment Air feed-gas systems are typically classified by their
Feed-gas to an ozone generator may be air, once- operating pressure. The most common is a low
through oxygen, or rl3cycled oxygen, as discussed in pressure system, which operates at a pressure
Section 6.3.1. Typically, either air or once-thl'ough ranging from 69 to 103 kPa (10 to 15 psig); although
oxygen is used. Once-through oxygen systems typ- pressures up to 275 kPa (40 psig) have been reported
ically do not require further moisture removal. when the pressure is reduced prior to the ozone
However, filters to I'emove particulates should be generator. High pressure systems operate at a
provided. Air and oxygen recycle feed-gas ozone pressure ranging from 480 to 690 kPa (70 to 100
generation systems must provide treatment to re- psig), reduce the pressure prior to the ozone gen-
move both excess moisture and particulates. erator, and are typically used in "small to medium"
sized applications. Either system may be used in
Inadequate air or oxygen recycle feed-gas treatment, conjunction with most of the ozone generation
especially inadequatl3 moisture removal, has caused equipment discussed in Section, and all ofthe
several ozone system problems and failures of ozone ozone contacting systems discussed in Section 6.4.3.
treatment processes (4,42). As such, air treatment is It should be noted that the decision to use a high or

low pressure system is often based on a qualitative Figure 6-27. Example low pressure air feed-gas treatment
evaluation of potential maintenance requirements, in schematic.
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
r--- --...,
l __ ..:-Optional r--~ ---.,
for the desiccant dryers. .

c. The high pressure air pre-treatment system

: After-Cooler:
1-_ - -C_-_-_-J__............-+-

-_~r- __
After-Cooler :
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 Heat-Reactivated
at a negative or in some cases a slightly positive Desiccant Dryers
pressure. The nominal pressure process is a propri-
etary process used in conjunction with the Kerag ~ ~
ozone generator and aspirating turbine mixer ozone
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 r
To Ozone Generator To Ozone Generator
flexibility design considerations is discussed in
The pre-compressor filters are provided to protect the
air compressors from damage due to large particles. The compressed, cooled aIr IS directed to a pre-
The air compressors are typically positive displace- desiccant filter, which is used to remove dust and dirt
ment, oil-less units. Positive displacement compres- particles greater than 3 to 5 micron in diameter.
sors are used in order to obtain constant air flow at Particulate removal prior to the desiccant dryers
variable operating pressures. Variable pressures are reduces plugging in the desiccant medium.
often encountered due to variable pressure losses in
downstream equipment and processes such as filters Probably the most important component of the air
and ozone contact basins. Oil-less compressors are treatment process is the desiccant dryer. The desic-
used to eliminate oil contamination of the down- cant dryer consists oftwo towers containing moisture
stream desiccant dryer medium and ozone generator adsorbing media. One tower operates while the other
dielectrics. Liquid-seal and rotary lobe compressors tower is regenerating. The low pressure system
have been used most frequently. desiccant dryer uses heat for reactivation of the
The compressors may be followed by an after-cooler
or a refrigerant dryer. These components are depicted The post-desiccant filters are installed to remove
by dotted lines in Figure 6-27, which indicate that particulates less than 0.3 to 0.4 micron in diameter.
they are optional. Typically, either orie or the other Two-stage filtration is preferred. The first stage filter
option is provided. These cooling mechanisms are removes particulates greater than 1 micron and the
used to remove moisture in the air at minimal second stage removes particulates less than 0.3 to
operating expense. 0.4 micron in diameter (18,33,35).

High Pressure Systl'Jm Flow Schematic. A schematic remove particulate matter less than 0.3 to 0.4 micron
of a high pressure air treatment system is shown in in diameter. The high pressure system also has a
Figure 6-28. The pre-compressor filters are IJsed to , pressure reducing valve to regulate operating pms-
remove larger particulates and protect the air com- sures in the ozone generator.
pressors. The compressors are typically oil-less units;
however, oil-seal cClmpressors followed by extensive Nominal Pressure System Flow Schematic. A sche-
oil removal equipmlmt have been utilized. matic diagram of the nominal pressure air treatmElnt
system is shown in Figure 6-29. The nominal
Figure 6-28. Example high pressure air feed-galS traatment pressure system typically operates at a pressure
IchomEltic. slightly below ambient pressure conditions. In cases
where a chiller is used to cool the air, a blower may be
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.

r - - - - -----....,
r,==--.=Fan--=-=-~' :
II .....
L l- Optional
__- '
I r-- - -
I I Fan- - - , II
- -r - __ oJ I I '- - - :L..:- - _oJ I
I .-- - - - .L. - -, I
II I' - - _ _
_ _ I II

...J \ [=--Chil~~=J \
L --:J
Cleslccant Dryers
~/ ~ Tower B

Heat Reactivated

~sicant Dryers

To Ozone Generator To Ozone Generator

The after-coolers following the high pressure com- Generator
pressors are essential, as they are used to remove the
heat of compression. The filter(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). To Ozone Destruct To Ozone Destruct

The high pressure system desiccant dryer consists of

two towers with moisture adsorbing media. One Theprimary difference between the nominal pres-
tower operates whilEl the other tower is regenerating. sure system and the other air treatment process
Regeneration is accomplished without additional schematics is the method of moving the air. With the
heat; thus, the high pressure desiccant dryers are nominal pressure system the air is drawn through the
called heat-less units. The post-desiccant filters air treatment process and through the ozone gener-
ators with the vacuum created by the operation of the Discharge pressure-Monitor system opera-
turbine mixer. The pre-desiccant filter is used to tion.
remove particulates greater than 5 microns in Discharge temperature-Monitor system per-
diameter, and the post-filter removes particles greater formance. Also include system alarm and
than 1 micron in diameter(33). The nominal pressure shut-down;
system uses a heat-reactivated desiccant dryer. Discharge flow-Monitor system perform-
The energy input andair supply to the turbine mixing Seal water temperature, when liquid-seal
contacting unit can be controlled by a variable speed compressors are used-Monitor system oper-
drive on the mixerto adjust its pumping action, and by ation.
adjusting the orifice size of the water inlet to the Seal water flow-Monitor water availability.
mixer. The specific details of achieving this flexibility Also include system alarm and shut-down.
can be arranged with the manufacturer.
f. Instrumentation that should be provided to
monitor and control after-cooler operation is: Air Compressor Design Considerations
Air compressors are used for the low and high Feed-gas temperature-Monitor system load-
pressure air treatment processes. Typically, oil-less ing.
compressors are used. However, for high pressure Dischargetemperature-Monitor system per-
applications oil-seal compressors have been used formance.
when they are followed with extensive oil removal Cooling water inlet and outlet temperature
equipment. -Monitor system operation.
Cooling water flow rate-Monitor system
Low Pressure Air Compressors. Low pressure air
compressors operate at a pressure ranging from 69 to
103 kPa (1 0 to 15 psig), although operation up to 276 High Pressure Compressors. The following design
kPa (40 psig) has been reported when the pressure to considerations are applicable for high pressure air
the ozone generator is reduced. The following design compressors:
considerations are applicable for low pressure air
compressors: a. Non oil-lubricated, high pressure compressors
(piston type compressors that use a Teflon seal)
a. Positive displacement compressors have been should be designed with a duty cycle less than
used most frequently because good control over 60 percent and should be coupled with pre-
air flow rate is achieved at variable operating planned methods of replacing the Teflon seal on
pressures encountered. a periodic interval. A replacement such as once
per year may occur.
,b. The air compressors should be located in a
remote area to avoid noise disruption, or should b. High pressure compressors may be oil-lubri-
be located in a sound-proofed housing. cated if followed by a sophisticated oil removal
process. Successful performance has occurred
c. Liquid-seal compressors often have been used with a cyclonic moisture/oil separation device,
for ozone, despite a higher capital cost and in followed by an impingement filter, followed by a
some cases a slightly higher energy require- coalescing filter. The final filter should remove
ment. These disadvantages were acceptable oil particles greater than 0.03 micron in diam-
considering an overall design objective.of reli- eter. At the same time, it should be anticipated
able, continuous performance. It should be that the medium in the desiccant dryers may
noted that when liquid-seal compressors are need to be replaced quite frequently (e.g., every
used the size of the downstream moisture 12 months).
removal equipment does not change because
that equipment is designed to treat saturated air c. Instrumentation should be provided for process
no matter what type of compressor is used. monitoring and control, as discussed previously
in the Low Pressure Air Compressor section.
d. Air compressors should be followed by an after-
cooler or by refrigerant dryers.Water-to-air lRefrigerant Dryer Design Considerations
after-coolers are generally used, rather than air- The value of refrigerant dryers has been questioned
to-air after-coolers. at some operating installations. Some problems that
have occurred with their use include sensitivity to
e. Instrumentation that should be provided to icing because operating temperatures arequite close
monitor and control air compressor equipment to the freezing point of water and non-availability of
operation is: maintenance expertise required to repair the refrig-

eration equipment (42). If these problems can be gel is used. After several hours of operation the
overcome, refrigerant dryers can be very effective desiccant becomes saturated with moisture and is
because they can remove excess moisture in the air unable to ma inta in the desi red dew poi nt. At that ti me
for a small amount of energy consumption. The the desiccant must be regenerated. The cycle valve(s)
decision to use a refrigerant dryer is typically based switch the operating tower so the "used" medium
on cost savings that can result in reducing the size of can be regenerated. It should be noted that the switch
the desiccant dryer bocause ofthe moisture removed valve is an important component in the dessicant
by the refrigerant dryer (21). dryer. Stainless steel valves have been reported to
work best. Also, routine maintenance is necessary to
The following design considerations are applicable make sure the valves switch every time.
for refrigerant dryers::
Figure 6-30. Diagram of a heat-reactivated desiccant dryor
a. If the desiccant dryers are reduced in size such with internal heating coils.
that operation of the refrigerant dryer is essen-
tial to the overo311 treatment process, bal::k-up
equipment and spare parts must be provided to Pressure
"Wet-Gas" Pressure
insure continuEld operation of this important, Relief Valve Relief Valve
sensitive-to-operate piece of equipment.

b. Plant maintenal1ce personnel must be trained

on refrigeration equipment, or refrigeration
equipment service and repair must be available Purge
on a responsive contract basis. Air
c. Instrumentation and control equipment that
must be provided to prevent overload of the

downstream desiccant dryers are: Tower
Inlet temperature-Monitor unit loadin!~.
Outlet temperature-Monitor unit perform-
ance. Also provide an alarm and shut-down
uuU Tower
Shroud Valve
for high temperature. e~
Feed-gas flow rate-Monitor system loading. purge"" E co
Air 8> Desiccant Dryer Design Consideratio,ns
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, P.e.s.ic-
cant dryers used felr ozone systems are special
application units, since "off-the-shelf" dryers typical-
Iy are capable of achil:lving dew points in the range of
To Ozone Generator
only -40C (-40F) or higher. Dew points of -60C
(-76F) and lower are required for ozone generation. It
should be noted that dew point temperature is 81 Regeneration involves heating the desiccant to a
function of pressure, and recorded dew point temper- temperature between 90 and 260 0 C (200 and 500F)
atures always should be referenced to pressure. In and purging with dry air (21). The typical design is for
this manual all dew point temperatures are based on heating to a temperature between 120 and 170C
standard pressure ,conditions, unless otherwise (250 and 350F) for 1 to 2 hours, and cooling for a
noted. Refer to Sectil:m for the procedure to minimum of 6 hours. Heating enhances moisture
adjust dew point temperature to standard conditions. evaporation. The evaporated moisture is removed by
purge air. After heating is completed the purge airwill
Heat-ReactivatedDesiccant Dryers. Heat-reactivated continue to flow through the tower and cool the
desiccant dryers are used in nominal pressure and desiccant before the regenerated tower is placed back
low pressure air treatment systems. A schematic of a into service. It is extremely important that the
heat-reactivated desiccant dryer is shown in Figure desiccant be cooled before being placed back into
6-30. Wet-air is directed to the operating tower service so that a "temperature spike" does not occur.
where moisture in the air is adsorbed onto the An aftercooler may be considered after the dessicant
desiccant. The desiccetnt is typically activated alumina dryer and before the dew point analyzer, with alarm,
and molecular sieves (21). In some instances silica in order to insure that a "temperature spike" does not

occur. The dew point analyzer is used to prevent to end of operating mode). However, the design cycle
possible moisture contamination due to a leak in the time should be longer (e.g., preferred 16-hr and
aftercooler or excessive moisture from the dryer. minimum 12-hr) because the desiccant loses its
moisture adsorbing capability in a few months of
From 5 to 20 percent of the dry air is recycled (called operation. At initial start-up the cycle time should be
purge air) to the regenerating tower. If the purge air longer in order to give the operator flexibi Iity to reduce
flow is inadequate, hot air and excess moisture will the cycle time as the desiccant loses its moisture
remain in the regenerated tower. This moisture laden adsorbing capacity. When the minimum, 8-hr cycle
air will be discharged to the ozone generator when time is reached, the operator must replace the
the regenerated tower begins its operating cyCle, arid desiccant and can then re-set the cycle time to the
will cause damage to generator components. To original design settings.
prevent unexpected or unnoticed loss of purge air
flow, it is essential that the purge air flow be Figure 6-31. Schematic of a heat-reactivated desiccant
monitored and controlled. It is importantto note that dryer with external heating equipment.
this is typically an option that must be specified in
Relief Valve "Wet-Gas" Relief Valve
The purge air is filtered to protect the equipment, Inlet
particularly the air supply control or to weir switching
valves. The schematic diagram in Figure 6-30 shows
that the source of the purge airjs 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
Desiccant Desiccant
importance of constant availability of purge air flow. Media Media
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 To Drain
capacity(Le., pounds of desiccant per Ibid of moisture
loading), and method of heating. Control Control
Valve Valve
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-
To O~one Generator,
ferred because better distribution of heat throughout
the desiccant is provided (37), but this system is more
complicated because of the additional equipment The regeneration cycle may be controlled on a timed
requirements. or on a demand basis. The timed basis uses a pre-
establish,ed time to initiate the regeneration cycle,
The amount of desiccant used and the regeneration whether or not the operating (drying) tower has
cyCle time are the most important design considera- exhausted its capability to remove moisture and
tions. The cycle time must be long enough to obtain achieve the desired dew point. The timed cycle
sufficient cooling of the regenerating tower, and control is adjusted by changing the timer settings.
short enough to maintain the desired dew point by the
operating tower. The recommended minimum cycle A regeneration cycle based on demand is more
time, because of cooling considerations, is eight expensive to install because of the controls involved.
hours (i.e., time from beginning of regeneration mode However, the system has potential O&M cost savings.

Using the demand regeneration system the cycle is final design criteria because the type and quality of
initiated when a pre-set maximum value for the dew desiccant are also very important.
point of the dischargl~ air from the dryer is reached.
Dew point temperatUl:e is monitored on a continuous A recommended design approach is to specify that a
basis. Energy savings; occur because the cycle time, minimum dew point of -60C (-76F) (corrected to
and corresponding hl~ating time would be activated standard pressure conditions) will be achieved during
only when the tower has exhausted its drying a minimum cycle time of 16 hI' while being operated
capability. Further, the life expectancy of the desic- at the maximum expected moisture loading condi-
cant is prolonged beccluse it would not be subjected to tions (Le., maximum temperature, maximum flow and
the stress of reactivation (Le., heating and cooling) as minimum pressure). In addition, a minimum amount
often. 'of desiccant for each operating tower should be
Demand regeneration control has some definite A summary of the design considerations for heat-
advantages, but has important operational considera- reactivated desiccant dryers is listed below:
tions that must be addressed, namely the reliability of
the dew point monitor and sensitivity of the control a. The tower switching mechanisms should be
logic. Ozone equipmc~nt suppliers, desiccant dryer designed for long life and must be maintained
manufacturers, and E~xisting ozone plant operating per the manufacturer's recommendations to
personnel may be contacted to determine which insure continued, reliable performance.
meters have been successfully used and how sensi- b. The desiccant dryers should be designed with
tive the control logic must be. Successful perform- sufficient room around the units to allow for
ance has been obtained from various systems, but inspection or repair of the heating elements, for
problems have been encountered with others. All in- maintenance and repair of the tower switching
line dew point monitors should be routinely checked mechanism, and for ease of replacement ofthe
by performing a "dew point cup" test (32,42). desiccant.

A word of caution is noted in operation of all dew point c. The heating coils used in internal heating, heat-
probes. When the measuring probe gets wet due to a reactivated desiccant dryers should not be
moisture spike, it takes several hours of operation contacted directly with the desiccant. A protec-
under dry conditions before the probe will accurately tive shroud should be provided.
reflect the correct dew point temperature reading.
d. The decision to use a timed or demand reactiva-
tion cycle control system should be based upon
The amount of desicccmt and the operating pressure economic considerations. Demand reactivation
(i.e., absolute pressure) controls the moisture removal cycle systems are generally desireable, espe-
capability of the prCIcess. Once the type of air cially for larger systems. For either control
treatment process is: selected (Le., low or high system the operators must routinely check and
pressure), operating pressure is not considered calibrate the in-line dew point monitor to assure
further in design. This leaves amount of desiccant as reliable performance from this critically impor-
a most important factor. Inadequate desiccant can tant instrument.
cause operating proble\ms in a short period oftime. At
one plant with insufficient desiccant, dew point break- e. The minimum pperating cycle time of a heat-
through occurred within 5.5 hours of an 8-hr drying reactivated desiccant dryer should be 8 hI', in
time and the desiccant was only 3 months old. The order for the regenerated bed to cool down
performance objective was a dew point temperature before being placed back into operation. The
of -50C (-58F). At this plant the desiccant moisture design cycle time should be 16 hI' (minimum 12
loading was 13 Ib desic:cant/lb of moisture loading in hI') to allow for decreasing this time as the
an 8-hr period. desiccant gradually deteriorates.
Apparently, 13 Ib desic:cant/lb of moisture loading is f. The process air compressors must be sized to
insufficient. Varas suggests that at least 18 Ib also handle the desiccant dryer purge air flow
desiccant be provided per Ib of moisture received rate, which may be asmuch as 20 percent of the
during one cycle of operation (33). Damez sugg,ests ozone generation process air flow rate. Reliable
that at least5.5 Ib of desiccant be provided per Ibid of process monitoring and control instrumentation
moisture received, which corresponds to 16.5 Ib that should be provided are:
desiccant/lb of moisture received during an B-hr
regenerating cycle (37). These suggested desiccant Feed-gas flow rate-Monitor system loading.
amounts may be used as guidelines for the design of Inlet temperature-Monitor system loading.
desiccant dryers, but should not be considered the Outlet temperature-Monitor system per-
formance. Also include alarm and system
shut-down capability.
Inlet and outlet pressure-Monitor system Figure 6~32. Pressure swing (heat-less) high pressura dasi-
operation. ccant dryer in purging mode.
Purge air flow rate-Monitor system opera-
Discharge feed-gas dew point, in-line meter-
Monitor system performance. Also include Moist Air
recorder and system alarm and shut-down. Intake
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 Operating Regenerating
a heat-less desiccant dryer is a pressure swing
desiccant dryer because the unit operates with
t Tower

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 Dry Air
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 CD Wet Air Inlet Adsorption Desiccant
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
CD Regenerated Air
Outlet Valve
Orifice Plate

pressure conditions and purged with "dry" air from

the operating tower. The moisture that has adsorbed Silencer Check Valve

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 ). heat-reactivated desiccant dryers discussed in the
previous section.

6.4.3 Ozone Contacting

The purge air flow rate for a heat-less desiccant dryer The capability of the ozone contacting unit is critical to
is normally 15 to 25 percent of the process air flow the successful performance of the ozone disinfection
rate. It is important to note that the compressors must system. It is also important to the economical opera-
be sized to accommodate this additional flow require- tion of the process. Two important ozone contact
ment. The amount of moisture that is removed and basin design considerations are the capability to
the corresponding dew poiht is lowered when the achieve good disinfection and capability to achieve
amount of desiccant is increased and/or the cycle good ozone transfer. In a well designed ozone contact
time is decreased. basin, good ozone transfer will exist if good disinfec-
tion occurs. However, the opposite may also occur if
The heat-less desiccant dryers have a history of the contactor is poorly designed; good ozone transfer
reliable performance, if properly sized and main- can exist when disinfection is less than satisfactory.
tained. The tower switching mechanism must be
maintained per the manufacturer's instructions, and At some of the plants visited during development of
the amount of desiccant must be adequate. Over a this design manual; poor disinfection was noted
period of time (1 to 5 years) the desiccant will lose its qespite good ozone transfer. Excessive short-circuit-
moisture adsorbing capability and will have to be ing was considered the primary cause of the poor
replaced. This process will occur gradually and must disinfection performance. Design considerations for
be monitored closely. The design considerations for both ozone transfer and ozone disinfection are
heat-less desiccant dryers are similar to those for discussed in this section of the manual.

Figure, 633. Ozone transfer efficiency decreases as applied Ozone Tranisfer Design Consideratiolns ozone dosage increases and as ozone demand
The bubble diffuser contactor is the most common of the wastewater decreases (17).
type of contacting ~iystem used. Other OZOnl:l con-
tactors, such as tho aspirating turbine mixer con-
tactor, are usually coupled with a proprietary ozone 100 .--.~~-r------,r---..,-----r---....,
generator (Kerag), alnd ozone transfer efficienl~y (TE)
Information can be clbtained from the manufae}turer.
The characteristics of the bubble diffuser ozone
contactor are presented herein. 96

The transfer of ozo,ne into wastewater has been

evaluated by several researchers (43-45). All con- ?fi
clude that ozone transfer into wastewater can be f 90
described by the tWI:>-film theory. In this theory the Cl)
mass transfer of ozone per unit time is a function of ffi.
the two-film exchan!~e area, the exchange potential, Cl)

and a transfer coeffident. The exchange area for the ~ 86

bubble diffuser contactor is the surface area of the E!
bubbles, which is discussed later. The exchange Cl)
potential is called th(:! "driving force" and is depend- 2 Example Relationship
ent upon the differElnCe between saturation ozone 080 Do Not Use as Design Values
concentration minus residual ozone concentration.

Ozone transfer can be described by the two-film

theory, but in practicc3 contactor basins have not been 75 ~--~---+--~~;----7-----:-!
designed utilizing this theory. The theoretical basis
o 10
has been avoided because the design coefficients Applied Ozone Dose, mg/I
have not been well documented, not because the
theory is unsound. When the design coefficients are Wastewater quality will affect ozone TE, and is
well documented, contactor design indeed may be important to keep in mind when evaluating the
established using thEI two-film transfer model. In this performance of existing contactors, pilot-scale con-
section of the manual the current state-of-the-art of tactors, and newly installed contactors. Wastewater
ozone contactor design is used. The most important quality is typically not used as a basis to modify the
factors affecting desi!;}n are discussed, and the design physical characteristics of the contactor. A summary
rationale is presented. of the important water quality considerations on
ozone TE design is listed below:
Ozone TE is primarily influenced by the physical
characteristics of the contactor and the quality of the a. Ozone TE will decrease as applied ozone dosage
wastewater. At a given applied ozone dosage, a increases. A specified minimum design TE
wastewater of poor quality will have a high ozone should be coupled with a specified applied
demand and the con1actor will exhibit a high TE:. The ozone dosage.
high TE is due to the disappearance of ozone in
oxidation reactions (i.e., ozone demand reactions). An b. Ozone TE will increase as wastewater quality
example of the effee}t of water quality on transfer deteriorates (I.e., ozone demand increases). A
efficiency is iIIustratEld in Figure 6-33 (17). The TE of specified minimum design TE should be coupled
the same contactor was higher when treating sec- with a specified description of the wastewater
ondary quality wastelwater than when treating ter- quality.
tiary quality effl uent. The differences in TE were more
pronounced as applie!d ozone dosage increased. 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-
The chemical quality of the wastewater also affects ison of TE of existing full-scale and pilot-scale
ozone TE, especially pH and alkalinity. A high pH results should consider differences in waste-
andlor a low alkalinity will cause a lower ozone water chemical quality.
residual (i.e., other factors being constant) because
the hydroxyl radicals will be maximized (refer to The physical characteristics of the ozone contactor
Section The lower residual will increase the are the most important considerations for the design
exchange potential, 01' driving force, and will increase engineer because the engineer controls this element
TE. of the process. The most important physical char-

acteristics for optimizing ozoneTE are depth of the tion. Deep diffusers will increase TE because the
contactor and type and location of diffusers. The saturation ozone concentration and associated driv-
contactor physical characteristics for optimum dis- ing force will increase. However, a deep contactor will
infection performance are discussed in Section also incmase capital cost. A shallow contactor will reduce capital cost, but will decrease TE. A high
elevation of the contactor will also decrease TE. Hegg
Ozone transfer theory states that the mass transfer of et al. (42) reported that a 3.6-m (12-ft) deep contactor
ozone is dependent upon the exchange area of the at an elevation of 2,290 m (7,500 ft) above sea level
gas/liquid film. The exchange area for the bubble achieved a TE of 50 percent while disinfecting a
diffuser contactor is primarily a function of the tertiary effluent.
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 Most of the bubble diffuser ozone contactors at plants
is between 2 and 5 mm in diameter. and preferrably visited during development of this manual had a
between 2 and 3 mm in diameter (30). The primary diffuser depth between 4.9 and 6.1 m (16 and 20 ft)
factors affecting bubble size include diffuser perme- deep. These contactors achieved a TE between 80
ability and air flow rate. A lower permeability and and 95 percent when treating a high quality sec-
lower air flow rate will decrease the size of the ondary effluent at an applied ozone dosage equal to or
bubble. less than 6 mg/1.

Another consideration to optimize ozone TE is to,

optimize the ozone gas to wastewater liquid ratio A summary of important design considerations for
(G 1 /L 1 ) and ozone gas to contactor volume ratio the physical characteristics of the ozone contactor are
(G/V). Stover et al. (17) observed a decrease in TE as listed below:
the G 1/L 1 ratio increased, as depicted in Figure 6-34.
a. The diffusers in the bubble diffuser ozone
A minor decrease in TE occurred as the G 1 /L 1 ratio contactor should beat least4.9 m(16 ft) deep for
increased from 0.2 to 0.5. The decrease in TE was
a plant located near sea level treating a high
more pronounced as the G 1 /L 1 ratio approached 1.0. quality secondary effluent at an applied ozone
Grasso (46) and Given and Smith (24) reported a
dosage less than 6 mg/1. The contactor should
decrease in TE as the G/V ratio increased from 0.005
be deeper if the wastewater is of higher quality,
to 0.05. They reported that for best results the G/V if the applied ozone dosage is higher, or if the
ratio should not exc;:eed 0.03.
plant is located at a higher elevation.
Figure 6-34. An increase in the gas to liquid ratio causes a
decrease in ozone transfer efficiency. b. Both rod- and disc-shaped porous stone dif-
fusers have been usedin bubble diffuser ozone
c. The air flow rate to the diffuser must be within
the range recommended by the manufacturer.
90 Detrimental Effect 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 I/s (4
... 2
For the approximately 260-cm 2 (40-in ) disc
~ diffusers the air flow rate per diffuser should
c: 70
not exceed 0.6 I/s (1.25 cfm).
2 2
For the approximately 390-cm (60 in ) disc
Applied Dosage = 10 mg/L diffusers the air flow rate per diffuser should
not exceed 0.8 lis (1.8 cfm).

d. Each porous stone diffuser should be secured

50.!.--=':::---:='-=-_ _--:-~,._-----------J
o 0.2 0.5 . 1.0 2.0 with stainless steel holders and sealed with
Gas/Liquid Ratio ozone-resistant gaskets such as Hypalon, Viton,
Teflon or Silicon.
The depth ofthe diffusers in the contactor is probably
the most important design issue because it repre- e. The diffusers should be able to achieve a
sents a cost consideration as well as aTE considera- estimated bubble size of 2 to 3 mm in diameter.

f. Porous stone diffusers that have the following Mixing is necessary to bring the residual oxidants into
criteria have typically been used in bubble contact with the microorganisms, but back-mixing
diffuser ozone contactors. can increase the potential for short-circuiting. For
example, in a 'bubble diffuser ozone contact basin
Permeability of the porous media typically back-mixing in each stage of the contactor is suffi-
has ranged from 12 to 20 cfm/ft 2 /in at 2 cient to change the liquid flow characteristics from a
inches water clDlumn. Note: Permeability is plugflowto a near complete mix pattern in that stage.
defined by a test of a porous stone plate that is Therefore, to minimize the effect of short-circuiting,
12 in by 12 in by 1 in thick, tested under 2 in multiple stages that are positively isolated from each
water column pressure. other should be provided. The optimum number of
Porosity ofthe porous stone diffuser has been stages has not been documented.
between 35 and 45 percent.
Wastewater detention time in the contact basin is
g. Diffusers have typi:cally been located from 15 to
also a factor affecting disinfection performance with
30 cm (6 to 12 in) from the bottom of the
a wide range of contact times reported to achieve
acceptable disinfection. Gan reported a 6-log reduc-
h. All diffusers should be installed at the same tion in coliform organisms for a high quality activated
elevation in order to evenly distribute the ~Ias carbon effluent with 2 minutes liquid contact time
flow to each diffuser. and an applied ozone dosage of 12 mg/1. However,
only a 1-log reduction was achieved with the same
i. Diffusers have typically been installed in the conditions when secondary effluent was treated. Gan
liquid's downflow stages of the contactor