United States
       Environmental Protection  Office of Water          EPA 811-S-92-002
       Agency         (WH-550)            November 1992


wEPA  TECHNOLOGIES AND COSTS FOR

       CONTROL OF DISINFECTION

       BY-PRODUCTS



       EXECUTIVE SUMMARY

-------

-------
EXECUTIVE SUMMARY
TECHNOLOGIES AND COSTS FOR
CONTROL OF DISINFECTION BY-PRODUCTS
SCIENCE AND TECHNOLOGY BRANCH
CRITERIA AND STANDARDS DIVISION
OFFICE OF GROUND WATER AND DRINKING WATER
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C.
NOVEMBER 1992
                          MALCOLM PIRNIE, INC.

     One International Boulevard                2 Corporate Drive
     Mahwah, New Jersey 07495-0018             P.O. Box 751
                                        White Plains, New York 10602-0751

-------

-------
                                                 EXECUTIVE SUMMARY

BACKGROUND
      *

    •  The 1986 Amendments to the Safe Drinking Water Act (SDWA) require the United
States Environmental Protection Agency (EPA) to set maximum contaminant level goals
(MCLGs) for many contaminants found in drinking water.  These MCLGs must provide an
adequate margin of safety from contaminant concentrations that are known or anticipated
to induce adverse effects on human health.  For each contaminant, EPA must establish a
maximum contaminant level (MCL) that is as close to the MCLG as is feasible with the use
of best available technology (BAT).  Although the BAT identified for each contaminant
must be an economically feasible and proven technology under field conditions, systems are
not required to install BAT for purposes of meeting a corresponding MCL.
     .EPA is currently in the process of establishing MCLs and/or treatment techniques
for disinfectants and disinfection by-products (D/DBPs). The D/DBPs in current regulatory
focus include:

       •   Trihalomethanes  (THMs),  including  each  of the four individual species
           (chloroform, bromodichloromethane, dibromochloromethane, and bromoform);
       •   Haloacetic acids (HAAs), including dichloro-  and trichloroacetic acid;
       •   Chloral hydrate;
       •   Bromate;
       •   Chlorine;
       •   Chloramines; and
       •   Chlorine dioxide,  chlorite and chlorate.

      Total trihalomethanes (TTHMs) is the only contaminant currently regulated; the
MCL for TTHMs is 100 Mg/L. This MCL applies to the sum of the four chlorinated and/or
brominated  THMs  (chloroform,  bromodichloromethane,  dibromochloromethane,  and
bromoform).
                                     ES-1

-------
      The purpose of this document is to characterize the feasibility of treatment for DBF
control and to estimate costs for treatment alternatives that can then be used by utilities to
meet national regulations. Treatment criteria were developed through the use of a Water
Treatment Plant (WTP) simulation model for parameters critical to disinfection and DBF
control.
      The WTP simulation model was developed for EPA as a part of this  effort and
predicts DBP formation based on source:water quality and operational parameters for unit
treatment processes. The model predicts the removal of DBP precursors through treatment
processes, and DBP formation upon chlorination.  Processes for DBP precursor removal
that can be simulated include coagulation, GAC adsorption, and membranes; precipitative
softening is being developed at this time. The WTP simulation model can only predict the
formation of THMs and HAAs at  this time; predictions for bromate and bromoform
production following  ozonation are being  developed.  Limited verification  has been
performed on the WTP model predictions.
      The design criteria established through modelling were used to develop treatment
costs for DBP control These costs are used by EPA to determine national costs for various
potential DBP regulatory scenarios.

DBP PROPERTIES AND TREATMENT ALTERNATIVES

      This document includes a description of the chemical structures and physical/chemical
characteristics of those disinfectants and DBFs that are being considered by USEPA for
possible regulation.  DBP formation depends on many factors including the type(s) of
disinfectants, disinfectant dosagesr  and  water quality characteristics such as pH and
concentration of natural occurring organic material (NOM).
      Research has identified the following generic treatment alternatives for control of
D/DBPs:                                     .

        •   Removal of NOM prior to disinfection;
        •   Use of alternative oxidants or disinfectants that do not create DBFs at levels
            considered adverse to human health; and
        •   Removal of DBFs after they are formed.
                                       ES-2

-------
Because  of the uncertainty relative  to the occurrence and  health risk  for DBFs from
alternative oridants and disinfectants, the first alternative listed above is considered the most
desirable for reducing the D/DBPs being considered for regulation.
      Further, several studies using a variety  of waters throughout  the country have
demonstrated that although the use of alternative disinfectants may decrease some DBFs,
other DBFs can be increased through the reaction of the alternative disinfectant with
precursors present in the raw  water.  Therefore, the impacts  of  treatment  process
modifications and water quality characteristics on DBFs must be viewed in light of potential
risk trade-offs. The risk trade-off becomes especially difficult since little is known about the
health effects of many D/DBPs.
      A major concern in the water  industry is how to provide adequate disinfection to
inactivate microorganisms while, at the same time, minimize DBF formation. In addition,
it is  important to control known DBFs without increasing risks from, as yet,  undefined
DBFs. This document discusses the following technologies available for D/DBP control:

       •   Coagulation/filtration;
       •   Precipitative  softening;
       •   Adsorption processes;
       •   Oxidation processes;
       •   Air stripping;
       •   Membrane processes;
       •   Reduction processes; and
       •   Biological processes.

REMOVAL OF DBF PRECURSORS

      NOM is  a generic term  for naturally occurring organic material that contains
precursors which react with disinfectants to form DBFs.  NOM consists of humic substances,
amino acids, sugars, aliphatic acids, aromatic acids  and a large number of other organic
molecules. NOM has been shown to bind with metals and  synthetic  organic chemicals
                                       ES-3

-------
(SOCs), thereby allowing these contaminants to proceed through treatment processes not
designed for NOM removal.  These conditions can contribute to the following:

      - •   Increased disinfectant demand, requiring higher disinfectant dosages.
       •   Increased substrate for microorganism growth in distribution systems.
       •   Competition with SOCs for activated carbon adsorption sites; and
       •   Higher coagulant dosages.

      Because the characteristics of NOM are widely varied on a chemical and physical
basis, surrogate parameters must be used .to measure NOM levels.  Commonly  used
surrogates to measure NOM or DBF precursor concentrations  include:

       •   THM formation potential (THMFP);
       •   Total and dissolved organic carbon (TOC and DOC); and
       «  Ultraviolet absorbance at a wavelength of 254 nm  (UV-254).

      These surrogates may be used to screen raw water sources for DBF precursor content
and to determine the performance of unit processes  for the removal of DBF precursors.
The following processes were evaluated as technologies for NOM (DBF precursor) removal:

        •   Coagulation/filtration and precipitative softening;
        • '  Adsorption processes such  as  granular  activated carbon  (GAG), powdered
            activated carbon (PAC) and resin adsorbents;
        •   Oxidation processes such as ozone and chlorine dioxide;
        •   Membrane processes; and
        •   Biological degradation.
                                       . ES-4

-------
      Based on the evaluations in this document, the following processes are considered
most effective for NOM removal:

       '•    Coagulation/filtration, particularly at low pH and high coagulant dosages;
        •    GAC adsorption; and
        •    Membrane processes.

ALTERNATIVE DISINFECTANTS

      In addition to treatment technologies to remove NOM, alternative disinfectants were
evaluated for D/DBP control. Any disinfection alternative implemented at a treatment
plant should:

        •    Provide adequate disinfection to control pathogens at the treatment plant and
            in the distribution system;
        •    Limit the formation of regulated DBFs to concentrations lower than the MCL;
        «    Limit the formation of unregulated DBFs to concentrations lower than those
            of potential concern; and
        •    Achieve adequate color removal, iron oxidation and taste and odor control.

      The most prevalent disinfectants for primary disinfection (pathogen inactivation) in
the United States include chlorine (C12), chlorine dioxide (C1O2) and ozone (O3). Secondary
disinfection  is provided by maintaining a disinfectant residual throughout the distribution
system; candidate secondary disinfectants in the United States include  CL, chloramines
(NHjCl), and G1O,.
      To achieve both primary and secondary disinfection, utilities may use a combination
of disinfectants. Primary and secondary disinfectant combinations that are capable of
meeting the Surface Water Treatment Rule (SWTR) are summarized in Table ES-1.
                                       ES-5

-------
                                .   TABLE ES-1
          PRIMARY AND SECONDARY DISINFECTION ALTERNATIVES
Primary Disinfectant
Chlorine
Chlorine
Ozone
.Ozone
Chlorine Dioxide
Chlorine Dioxide
Chlorine Dioxide
Secondary Disinfectant
Chlorine
Chloramines
Chlorine
Chloramines
Chlorine
Chloramines
Chlorine Dioxide
      Although  the  SWTR does  not  specify primary  disinfection  credit  for the
ozone/hydrogen peroxide process, the process can be used as a primary disinfectant if the
utility can demonstrate that adequate levels of primary disinfection are maintained.  It also
should be noted that primary disinfection credit can be achieved with chloramines.  Few
utilities, however, are expected to continue to use chloramines in this capacity because of
the relatively poor disinfecting capacity and large CT values (the product of disinfectant
concentration in mg/L and disinfection contact time in minutes) required by the SWTR.
Because ozone does not maintain a residual in the distribution system over time, another
disinfectant, must  be applied to achieve secondary disinfection.  The presence of a
disinfectant residual continues the formation of DBFs in the distribution system, the extent
of which depends on the type of disinfectant and treated water characteristics.
      In the United  States, the combination of disinfectants most commonly  used are
chlorine/chlorine and chlorine/chloramine for primary/secondary disinfection. In waters
with high THMFP, some utilities have used chloramines throughout the treatment plant.
With the promulgation of the SWTR, however, these utilities are evaluating whether this
practice can be continued, given the large CT values required for disinfection credit using
chloramines as a primary disinfectant.
                                       .ES-6

-------
      Other utilities, particularly in the southeast, midwest and Texas, have pursued the use
of chlorine dioxide as a primary disinfectant.  Utilities typically have  used chlorine or
chloramines as a  secondary disinfectant when chlorine dioxide is used as  a  primary
      »•
disinfectant.
      Finally,  ozone is increasing in popularity as a primary disinfectant in the United
States. Although chlorine dioxide is used as a residual disinfectant when ozone is applied
in parts of Germany and Switzerland, free chlorine and more commonly chloramines are
typically used as secondary disinfectants in the United States.
      Each  of these  disinfectant combinations produce DBFs.  Therefore, the use of
alternative disinfectants requires considerable care.  A modified disinfection scheme may
decrease the formation  of  some DBFs  while  increasing  the presence of others.  As
previously indicated, the rate and extent of DBF formation is strongly related to the type,
concentrations  and characteristics of the NOM  present,  the type of disinfectant,  the
locations of disinfectant application, residence time in the system and other water quality
characteristics, such as pH, temperature, and bromide concentration.

REMOVAL OF DBFs AFTER FORMATION

      Removing DBFs  before  the finished water enters the distribution system is the
remaining DBF control strategy discussed in  this document.  The strategy for removing
DBFs after their formation is limited by the following factors:

       •   The amount of  DBFs formed in  the treatment plant relative to the amount
           formed in the distribution system; and
       •   Costs for the required treatment.

      The following technologies may be applicable for removing various DBFs:

       •   GAC adsorption;
       •   PAC adsorption;                         '
       •   Air stripping;
       •   Conventional treatment;
                                       ES-7

-------
       •   Oxidation;
       •   Membranes;
      * •.   Reducing agents; and
       •   Biological treatment.

      For this approach to be feasible from a process standpoint, a significant proportion
of the DBFs must be formed before the water leaves the treatment plant.  In addition, the
application of a given technology  may be specific to only a small portion of the DBFs
formed, and therefore, is relatively costly compared to removing precursors for a wide range
of DBFs.

DEVELOPMENT OF DESIGN CRITERIA AND UPGRADE COSTS

      The overall approach for the development of design criteria and upgrade costs for
selected DBF control alternatives assumes that there are five basic types of treatment
practiced in the United States at the present time:

       •   Surface Waters
                 Coagulation/filtration systems
                 Precipitative softening systems
                 Unfiltered systems (including those that will be required to filter under
                 the SWTR)
       •   Ground Waters
                 Unfiltered systems
                 Precipitative softening systems

      For  the  analysis  presented  in  this  document,  only  the  surface  water,
coagulation/filtration category was evaluated. This category was analyzed first because: 1)
surface water systems are generally more sensitive to DBF formation than ground waters,
and 2) this category represents the largest population served.
                                       ES-8

-------
      As stated previously in this section, the three basic alternatives for control of D/DBPs
are:

      - •    Removal of NOM prior to disinfection;
       •    Use of alternative disinfectants; and                                 ,
       •    Removal of DBFs after formation.

      Design criteria and upgrade costs were developed for specific treatment schemes
employing the first two control alternatives.  Costs for removal of DBFs after formation
were not developed because it is almost always more cost efficient to remove the precursors
before they are formed.
      Based on overall effectiveness, expected economic feasibility and practical full-scale
experience, the most promising and effective processes for the removal of NOM are:

       •    Increasing the coagulant dosage (only for coagulation/filtration systems);
       •    Installing GAC adsorption; and
       •    Installing membrane filtration.

      For the use of alternate disinfectants, only two primary disinfectants are considered;
C12 and O3. The most applicable secondary disinfection alternatives for the control of DBFs
are C12 and NH2C1. Therefore, the disinfection alternatives in Table ES-2 were considered
in this evaluation.

                                   TABLE ES-2
                  DISINFECTION ALTERNATIVES EVALUATED
Primary Disinfectant
Chlorine
Chlorine
Ozone
Secondary Disinfectant
Chlorine
Chloramine
Chloramine
                                       ES-9

-------
      Chlorine dioxide is not considered as a primary disinfectant in this analysis because
equations were not available for predicting chlorine dioxide decay and formation of the
inorganic by-products chlorite and chlorate.  Although chlorine dioxide is not evaluated in
this do'cument, upgrade costs for  chlorine dioxide disinfection are provided in another
document; Technologies and Costs for Ground Water Disinfection.
      Treatment plants may require one, or a combination of, the control alternatives listed
above in order to  meet  future D/DBP standards.   A  summary of D/DBP control
alternatives  and combination of alternatives for which design criteria and upgrade costs
were generated is provided in Table ES-3.

                                   TABLE ES-3
           SUMMARY OF DBF CONTROL ALTERNATIVES EVALUATED
NOM Removal Process
•* ' ji-
Original Treatment Process Train
Increase Coagulant Dosage
Install GAC Adsorption
Install Membranes
Disinfection Strategy
0,/Cl,
x«
X
X
X
0,/NHiCl
X
X
X
X
0,/NHjCI
X
X
X
NE<2)
Notes:
(1)Basc plant process with
     -  Not evaluated.
                            disinfection.
      Upgrade costs for each D/DBP control alternative are designed to represent the costs
for an existing plant to improve treatment to meet potential D/DBP standards.  Upgrade
costs were generated by calculating the difference in total cost between a completely new
treatment plant without D/DBP control and a completely new treatment plant with D/DBP
control  The treatment plant without D/DBP control, also referred to as the "base plant",
is assumed to be a facility which utilizes a Clj/Cl, disinfection strategy and which currently
meets the requirements of the SWTR.
                                       ES-10

-------
      To estimate upgrade costs for any D/DBP control alternative, design criteria must
first be developed.  For this analysis, design criteria were developed  using the  WTP
simulation  model and accepted  engineering practices.  The WTP model  simulates NOM
removal, DBF formation, and  disinfectant levels  in  a  water treatment  plant and its
distribution system.  Using this model, water quality-related design parameters such as
chemical dosages, contact basin  size and sludge production were predicted.  Other design
criteria common to all alternatives, such as clarifier overflow rates and filtration loading
rates, were developed based on accepted engineering practice.
      Raw water quality plays a significant role in determining design criteria for a given
treatment plant.  As a result, design criteria for treatment parameters such as chemical
dosages, contact basin size and solids production were developed for a wide range of water
qualities which were assumed  to be representative of  the  treatment  category  under
consideration (surface water treatment plants using coagulation and filtration). The median
values generated in this analysis were used as design criteria for each control alternative.
      Through the development of design criteria and  upgrade costs  for  the control
alternatives listed in Table ES-3,  it  was found that the  upgrade costs for different
combinations of alternatives was nearly equal to the sum of the individual  costs for each
upgrade. For example, the cost for installing GAC adsorption and switching to chloramines
was nearly  equal to the  sum of individual upgrade cost for adding GAC and the upgrade
cost for switching to chloramines. As a result, this document provides costs for the following
NOM removal and alternate disinfection processes:

        •    Using  monochloramine  (as  opposed  to free  chlorine)  as a  secondary
            disinfectant;
        •    Increasing coagulant dosage to improve NOM removal;
        •    Using ozone as a primary disinfectant  and monochloramine as a secondary
            disinfectant;
        •    Installing post-filter GAC adsorption; and
        •    Installing membrane filtration.
                                       ES-11

-------
      A summary of the key design criteria and assumptions used to develop upgrade costs

are presented in Table ES-4.

                                   TABLE ES-4

                   DESIGN CRITERIA AND KEY ASSUMPTIONS
    Control Alternative
        Design Criteria and Key Assumptions
 Using Chloramines
• 4:1 chlorine residual to ammonia ratio.
• 0.8 mg/L ammonia dose.
 Increase Coagulant Dose
• Alum as coagulant.
• Increase dosage to 50 mg/L from 10 mg/L.
• Lagoons used for dewatering.
• Land for additional lagoons available on site.
 Using •
 Ozone/Chloramines
  5 mg/L ozone dose.
  Ozone generation system and contact chamber sized
  for design flow.
  4:1 chlorine residual to ammonia ratio.
  0.8 mg/L ammonia dose.
 Install GAC Adsorption
  EBCTs of 15 and 30 minutes.
  180-day regeneration frequency.
  Replacement of GAC for Flow Categories 1 to 6(1).
  On-site GAC regeneration for Flow Categories 7 to
 Install Membranes
  Nanofiltration assumed.
  Sized for design flow.
  Molecular weight cutoff = 200.
  Recovery rate = 85 percent.
  Operating pressure = 80 psL
Note:

WSee Table ES-5 for Flow Category description.



      As shown in Table ES-4, design criteria and upgrade costs for membrane filtration

were based on nanoffltration.  Nanofiltration is capable of achieving significant removals of

NOM as well as providing excellent disinfection. Ultrafiltration is also a membrane process

capable of providing DBF and disinfection control Although Ultrafiltration is very effective

in removing pathogens and costs less than nanofiltration, it typically does not remove as

much NOM as nanofiltration.  Systems may be able to use a combination of Ultrafiltration

for pathogen removal and chlpramines  for  reduced DBF formation when used as a

secondary disinfectant. A potential problem with this approach, however, includes failure
                                       ES-12

-------
of the membrane, allowing pathogens to enter the distribution system without necessary
removal  or inactivation  (chloramines  do  not provide an  adequate backup primary
disinfection capability in the event of catastrophic membrane failure).
      »•
      Upgrade cost estimates were prepared for each control alternative for water supply
systems of several sizes, based on USEPA's 12 flow categories. These categories were
divided  into two groups; small systems  having design flow of less than 1 mgd and large
systems having design flow greater than 1 mgd. The median population served, average flow
and design capacity for each flow category is presented in Table ES-5.

                                   TABLE ES-5
                            USEPA FLOW CATEGORIES
: USEPA Flow
'•«.''" Categories
Median Population Served
Average Flow
(mgd)
Design Capacity
(mgd)
Small Systems - Design Flow < 1 mgd
1
2
3
4

5
6
7
8
9
10
11
12
57
225
750
1,910
0.0056
0.024
0.086
0.23
0.024
0.087
0.27
0.65
Large Systems - Design Flow > 1 mgd
5,500
15,000
35,000
60,000
88,000
175,000
730,000
1,550,000
0.70
2.1
5.0
8.8
13
27
120
270
1.8
4.8
11
18
26
51
210
430
      For these systems, the cost presented  in this document apply separately to each
 treatment facility within a given water system.  For example, some large systems have
                                       ES-13

-------
treatment facilities at multiple locations. The total costs for such a system can be obtained
by adding together the costs for each individual treatment facility.
      Estimated costs were developed using the WATER model for small systems and the
      IF
WATERCOST model for large systems.  Where necessary, these computer models were
supplemented with costs from GAC cost models and vendor costs. Estimated total upgrade
costs consist of operation and maintenance (O&M) and annual debt service on the capital
cost (i.e., 10 percent interest, 20-year design life). The cost basis is June 1991. Some cost
indices have increased insignificantly between June 1991 and September 1992; others have
decreased.
      Upgrade costs are shown graphically in Figures ES-1 and ES-2.  Figure ES-1 shows
upgrade costs for NOM removal strategies (i.e., increasing coagulant dosage, installing GAC
and installing membrane filtration).  Although ultrafiitration was not  evaluated in the
document,  upgrade costs  (capital  and  O&M) may  be 10  to  20  percent less  than
nanofiltration.  Figure ES-2 shows upgrade costs for alternate disinfection strategies (i.e.,
CyNH,Cl and O3/NH2C1).
      Each  upgrade cost represents  the cost to install  and operate the given  control
alternative; no consideration is given in this document to the costs  for retrofitting the
existing plant.  Retrofit  costs can increase the upgrade cost from  10 to greater than 100
percent depending on site-specific factors such as unknown interferences, extra piping, site
constraints, hydraulics and the requirement to maintain the existing plant in operation. The
type of control technology also affects retrofit costs. For example, retrofit costs would most
likely be greater for GAC alternatives compared to the addition of an ammonia feed system.
Therefore, it is recognized that costs for individual utilities may vary depending upon unique
site constraints and design criteria.
      Ifc must also be recognized that upgrade costs for the control alternatives presented
in this document are developed based on specific design criteria and assumptions. As a
result of some of these assumptions, upgrade costs were found to be independent of selected
DBF limits.  For example, the cost for a system installing GAC adsorption with a 15 minute
empty-bed contact time and a regeneration frequency of 180 days will be the same whether
a particular system is attempting to meet a TTHM goal of 100 pig/L or 25 ng/L, because
the criteria are specified.  The ability of such a  treatment plant to meet selected D/DBP
goals using GAC adsorption depends, to a large degree, upon raw water quality. Although
a wide variety of raw water quality was considered during the development of design criteria
                                       ES-14

-------
for plants in this treatment category (surface water systems using coagulation and filtration),
it is not the purpose of this document to evaluate the ability of a particular treatment plant
to meet a DBF limit using a given treatment technology.  Rather, this document presents
the estimated costs for each control alternative based on the specified design criteria.
      A more detailed description of the selection of design criteria and costs can be found
in the document entitled "Technologies and  Costs for  the Control of Disinfection By-
Products."  The upgrade costs presented in this document are used as the basis to generate
national costs for compliance with different disinfection/disinfection by-product  goals.
                                        ES-15

-------

-------
                                                       FIGURE ES-1
UPGRADE COSTS FOR IMPROVED NOM REMOVAL
                         SMALL SYSTEMS
   2.00CL


   1,500
O_
T—
^ 1,000


2  500
                                              Treatment Upgrade
                                           Increased Coagulant Dosage
                                          Addition of GACjEBCT= 15 min)
                                          Addition of GACJEBCT = 30 min)
                                             Addition of Nanofiltration
         A..
                            '•A	
      0.02    0.03    0.05        0.1         0.2    0.3
                            Design Flow (mgd)

                         LARGE SYSTEMS
                                                      0.5
  250
sr-200
(0
03
   150
tn

-------
                                                FIGURE ES-2
UPGRADE COSTS FOR ALTERNATE DISINFECTION
  600
^500
O)

  400
4- 300
CO

  200
5 100
  40
8
ga
8
st (0/1 0
S 8
  15

  10
                      SMALL SYSTEMS
                                            Treatment Upgrade
                                           CB/NH2CI
                                           O3/NH2CI
                                                  -A
0.02   0.03    0.05        0.1        0.2    0.3
                    Design Flow (mgd)

                 LARGE SYSTEMS
                                               0.5
                                            Treatment Upgrade
                                              O2/NH2CI
                                              O3/NH2CI

                                                  A
                5    10    20     50   100   200
                       Design Flow (mgd)
                                               500  1.00C

-------