KF3794.A35 800R76103 FIFTH DRAFT
1976a
INTERIM TREATMENT GUIDE
FOR THE CONTROL OF CHLOROFORM
AND OTHER TRIHALOMETHANES
by
James M. Symons
Major Contributors
J. Keith Carswell
Robert M. Clark
0. Thomas Love, Jr.
Richard J. Miltner
Alan A. Stevens
Water Supply Research Division
Municipal Environmental Research Laboratory
Office of Research and Development
Cincinnati, Ohio 45268
A ., in,* U.S. Environmental Protection Agency.
April iy/o .. ...
Region V, Library
230 South Dearborn Street,
Chicago, Illinois 60604
-------
i
TABLE OF CONTENTS
SUMMARY 1
INTRODUCTION 4
Background 4
Nomenclature 4
1975 Surveys 6
Treatment Research 8
Position of the Environmental Protection Agency 9
EXTENT OF PROBLEM 10
SUGGESTIONS FOR ALTERNATE TREATMENT 12
Change in Chlorination Practice 12
Changing the Point of Application of Chlorine 12
Use of Ozone or Chlorine Dioxide instead of Chlorine 13
Performance 13
Cost 16
Control'of Precursor Concentration 19
Granular Activated Carbon 19
Performance 19
Cost 21
Removal of Chloroform 29
SPECIFIC EXAMPLES OF ALTERNATE TREATMENT FOR DIFFERENT TYPES OF WATER
TREATMENT PLANTS 31
Chlorination Only 31
Aeration for Iron or Odor Removal 31
Coagulation, Settling, and Filtration for Turbidity Removal 31
Softening 32
Precipitation 32
Ion-Exchange 32
-------
ii
Treatment for Taste and Odor Control 33
Aeration 33
Powdered Activated Carbon 33
Chlorine Dioxide 34
Granular Activated Carbon 34
MONITORING 36
REFERENCES 39
ACKNOWLEDGMENTS 41
APPENDIX - CURRENT STATE OF KNOWLEDGE
-------
INTERIM TREATMENT GUIDE FOR THE CONTROL OF CHLOROFORM AND OTHER TRIHALOMETHANES
V
IN DRINKING WATER
April 1976
SUMMARY
On March 29, 1976 Administrator Train released a public statement that
said in part, "The recent test results and the fact that chloroform is prevalent
in the environment have convinced me that the prudent course of action at this
time is to minimize exposure to this chemical wherever it is feasible to do so."
The statement also said, "EPA will work with cities and States to evaluate certain
modifications to current treatment practices that can reduce the formation of
chloroform during the water treatment process, without lessening the effectiveness
of disinfection. EPA research has shown that changes in chlorination procedures
practiced by some water systems can result in reductions in the levels of
chloroform produced. EPA plans to share these initial findings on chloroform
reduction with the States and some cities encountering high chloroform levels, in
an effort to reduce human exposure as quickly as possible. This will also
allow EPA to gain added information to support the development of national
regulations to limit chloroform levels in water supplies."
The purpose of this Interim Guide is to provide utilities with the
information they will need to be able to assess their own particular circumstances
in conjunction with the technical assistance available from EPA.
-------
- 2 -
Data contained in this Interim Guide demonstrates that chloroform
concentrations can be lowered if chlorine is applied to the water with the
lowest possible organic content. Therefore, in locations where it is feasible,
a utility should consider moving the point of application of chlorine to the point
in the treatment process where the water should have the lowest organic content;
after filtration, or after coagulation and settling, if these unit processes
are employed. This should reduce the chloroform concentration of the finished
water. Utilities making such a change in disinfection practice should carefully
monitor the microbiological quality of their drinking water to make sure that
it has not deteriorated because of this change in practice.
Further reduction in chloroform concentration can be obtained if a
disinfectant such as ozone or chlorine dioxide is used instead of chlorine.
These two disinfectants do not produce chloroform, although they may produce
other organic by-products that have yet to be identified and evaluated for toxicit^
Furthermore, ozone does not produce a disinfectant residual, thus the addition
of chlorine may also be necessary. Under these circumstances, some chloroform
will be formed during passage through the distribution system.
Water containing very little organic matter can be produced when fresh
granular activated carbon is used as a medium for the adsorption of organic
compounds. This water can then be disinfected with either chlorine, ozone, or
chlorine dioxide and little chloroform or other organic by-products will be
produced because of the small quantity of organic matter available for reaction
with the disinectant. This treatment technique has the additional benefit of
removing many organic raw water contaminants, thereby providing consumers
with an additional margin of safety.
-------
- 3 -
The chief disadvantage of adsorption on granular activated carbon
as a treatment technique is that the adsorption'capacity of the material is
limited. For example, studies with Ohio River water have shown that granular
activated carbon is effective for removing chloroform precursor for about 1 month.
In other situations where the organic load is higher or lower than that in the
Ohio River this period of good performance would probably vary accordingly.
In general, however, the use of granular activated carbon for the control of
chloroform precursors means that the frequency of reactivation, will have to
be increased over that commonly used when taste and odor control is the
only objective.
A final problem with the proposal to change disinfectants or use granular
activated carbon is that of equipment and chemical availability. Unless these
changes are instituted in a phased program, temporary shortages may occur.
Discussions of various treatment techniques in detail including the cost of
treatment, specific suggestions for modification of several different
types of water treatment plants, and recommendations for monitoring are contained
in the Interim Guide that follows. The Appendix contains details of the
experimentation that lead to the treatment recommendations and a list of research
needed to provide additional information in this area.
-------
- 4 -
INTRODUCTION
Background
Reaction of chlorine with certain organic materials to produce chloroform
and related organic by-products has probably been occurring since chlorination
was first practiced as a disinfection procedure for drinking water. The presence
of these compounds in drinking water escaped detection until recently because,
having fairly low boiling points, they were lost during certain steps in the
procedures for performing typical organic analyses of water by gas chromatography.
1 2
Recently, however, both in this country and in The Netherlands, investigator
developed alternate organic analytic procedures that allowed' the measurement of
this type of organic compound. These investigators used the newly developed
analytic procedure to demonstrate that the concentration of chloroform and
related compounds were generally higher in finished water than in raw water,
1 3
indicating that they were being produced during the chlorination of water. '
Nomenclature
For those readers unfamiliar with organic nomenclature, the following
discussion defines some of the terms used later in the Guide. Although
methane gas is not invoved, the reaction of chlorine in water with certain organic
compounds (believed at this time to be primarily humic acids, part of the group
of organic materials associated with decaying vegetation) under certain conditions
produces a group of halogen-substituted single carbon compounds. These
compounds are named as derivatives of methane (CH,) and are listed below.
-------
- 5 -
1.
TABLE 1
FORMULAS AND NAMES OF THE TRIHALOMETHANES
H
1-1
Cl
1
- C - Cl
1
Cl
CHC13
2. Br
1
H - C - Cl
1
Cl
CHBrCl2
3. Br
I
H - C - Cl
i
Br
CHBr2Cl
Trichlorome thane
(Chloroform)
Bromodichloromethane
Dibromochloromethane
4. Br
I
H - C - Br
I
Br
CHBr3
Tribromomethane
(Bromoform)
5.
H - C - Cl
Cl
CHC12I
Dichloroiodomethane
H - C - Cl
I
Br
CHBrClI
Bromochloroiodomethane
7.
H -
Chlorodiiodome thane
8.
Dibromo iodomethane
9.
C - I
i
Cl
[cn2
H - C - Br
I
Br
CHBr I
H - C - Br
i
I
CHBrI2
Bromodiiodomethane
10.
H - C - I
Tr i iodome t hane
(lodoform )
-------
- 6 -
Under typical circumstances the trihalomethanes produced in drinking water
are dominated by compounds 1 and 2 above, but compounds 3 and 4 have been frequentl
found and 5 has been detected. The bromine- and iodine-containing trihalomethanes
have been shown to be formed by chlorine oxidizing any bromide and iodide in water
to bromine and iodine, these halogens then reacting with organic matter to form
the corresponding trihalomethanes. Fluorine-containing trihalomethanes are
neither formed during fluoridation nor can chlorine oxidize fluoride to
fluorine to produce them.
1975 Surveys
The increased interest in the organic content of drinking water generated
by studies of the New Orleans, Louisiana finished water, plus the information in
1 3
the literature cited ' prompted the Administrator of the U.S. Environmental
Protection Agency to announce a National Organic Reconnaissance Survey (NORS)
in November 1974. The purpose of this Survey was, in part, to determine
on a nationwide basis, the conditions under which trihalomethanes were formed
during water treatment. The NORS, in part, involved the sampling of raw and
finished water in 80 water utilities across the nation and determining the
concentration of compounds 1 through 4 in Table 1 on each sample.
This Survey, carried out during February to April, 1975, confirmed that all
of the chlorinated drinking waters investigated contained some chloroform,
ranging from less than 0.2 yg/£ (ppb) to 311 yg/£. One utility surveyed ozonated
as the only treatment and had <0.1 yg/£ chloroform in their drinking water.
A companion Survey (carried out during the same time period by EPA's
Region V Laboratory) of 83 utilities in the upper mid-West yielded very
a
similar data ranging from /chloroform concentration of <1 yg/£ to 366 vig/&.
Combining these two surveys, Figure 1, shows the median chloroform concentration
to be 20 vg/fc, with 10 percent of the drinking waters containing more than
-------
400
300
200
100
u
Z
O
u
ot
O
O
ex.
O
25
10
Til
COMBINED NORS AND REGION 2 SURVEY
152 UTILITIES
122 SURFACE, 30 GROUND
165 SAMPLES, FEB.-APRIL, 1975
5 10 20 40 60 80 90 95
PERCENT OF SAMPLES EQUAL TO OR LESS THAN GIVEN CONCENTRATION
FIGURE 1. FREQUENCY DISTRIBUTION OF CHLOROFORM
-------
- 8 -
105 ug/& of chloroform. Most of the finished waters also contained some of
the other three trihalomethanes measured. A similar survey will be made during
1976 in about 110 locations with samples collected in the Spring, Summer and Fall.
Prior to these Surveys some concern was expressed that other chlorinated
compounds detectable by this analytic technique were formed during chlorination.
This possibility was examined for three other compounds, 1,2-dichloroethane,
carbon tetrachloride and methylene chloride. Results showed that these compounds
were not formed during chlorination.
Treatment Research
In anticipation of the need for information on tiethods of controlling
trihalomethane concentrations, research on water treatment unit processes to
remove organic materials was intensified in September 1974. Over the 18 months,
up to March 1976 unit processes such as adsorption on granular activated carbon,
adsorption on powdered activated carbon, ozonation, the use of chlorine dioxide,
and aeration were studied to determine their effect on concentrations of
trihalomethane precursors and chloroform and other trihalomethanes.
compansion to this pilot plant research, controlled bench-scale studies were
carried out in an attempt to understand the influence of various parameters, such
as pH, temperature, chlorine residual, and the concentration and nature of
precursor on the reactions that produce chloroform and other trihalomethanes.
-------
- 9 -
Position of the Environmental Protection Agency
The release of the National Cancer Institute chloroform carcinogenicity
report , caused the Administrator of EPA to make a public statement on March 29, 1976
suggesting that water utilities take what steps they could to reduce the
chloroform concentration in their particular drinking water. To aid these
utilities in this effort, the Agency has prepared the Treatment Guide that follows
and wiE offer the utilities technical assistance. In addition, EPA will issue
an Advance Notice of Proposed Rulemaking to solicit public comment and information
regarding alternative regulatory strategys for organics in drinking water.
The purpose of the Guide is to provide utilities with the information they
wil need to be able to assess their own particular circumstances and
contains treatment suggestions for water utilities desiring to reduce the
concentration of chloroform in their drinking water, including data on cost and
effectiveness. The Appendix to this Guide summarizes the research data to
support the treatment suggestions and contains a list of research needs.
The Administrator's statement dealt with the problem of chloroform in
drinking water because at the present time chloroform is the only trihalomethane
that has been tested for carcinogenicity, and other physiological effects.
The other trihalomethanes measured in the two Surveys may, however, eventually
also be classed as health hazards. Therefore, the Guide, although emphasizing
chloroform, discusses treatment techniques for removing four of the trihalomethanes.
Thus a utility desiring to remove bromine-containing compounds, will have the
benefit of the available research information on that topic. For example,
fourteen utilities in the NORS and Region V Survey had concentrations of
bromine-containing trihalomethanes in their finished water that exceeded the
chloroform concentration.
-------
EXTENT OF PROBLEM
Although 90 percent of the water utilities sampled in the two Surveys had
wintertime concentrations of chloroform in their drinking waters less than 105
a utility manager in a location not yet sampled'might want to know the concentratio
of chloroform and other trihalomethanes in that particular drinking water,
particularly in warmer weather. The analytic procedure has been published
Q
and can be performed onsite if qualified staff analysts and proper equipment are
available. Although an initial investment of from $7,060 to $8,000 is necessary,
purchase of such equipment will allow a water utility to monitor this important
parameter frequently, thereby providing their consumers with an additional assuranc
of safety. Many State, EPA Regional, or qualified private laboratories can perforn
this analysis. Using these laboratories does involve shipping of samples, and
some delay in receiving results, but is a satisfactory method of operation. Note:
Because the trihalomethane formation reaction will continue in the sample bottle if
chlorine is present, sodium thiosulfate should be added to dechlorinate the sample
upon collection if the concentration of trihalomethane at the time of collection
is desired. If the potential for additional trihalomethane formation, such as
might occur during distribution,is to be investigated, the sample should be stored
without the dechlorinating agent for a time and at a temperature similar to that
occurring in the distribution system and then dechlorinated.
Another method for estimating which utilities might produce water with high
chloroform concentrations would be a comparison with the nine utilities having the
highest chloroform concentrations in the National Organics Reconnaissance and Regl
Survey, approximately the upper 10 percent. These data, see Tables 2 and 3 indica
that high chloroform concentrations result when surface or shallow ground water wi
a high NPTOC concentration and a high chlorine demand is dosed with enough chlorin
to produce a high free chlorine residual, particularly if the water is somewhat ba
Water utilities with similar characteristics would be expected to have finished
water with relatively high chloroform concentrations. Controlled experiments have
confirmed that these factors tend to enhance the reaction of chlorine with
precursor to produce trihalomethanes.5,9
-------
11
TABLE 2
ANALYSIS OF NINE UTILITIES HAVING HIGH CONCENTRATIONS OF CHLOROFORM IN THE
NATIONAL ORGANICS RECONNAISSANCE SURVEY
Range Average
Chloroform Concentration 103 yg/£ - 311 yg/Jl 177 ug/Jl
Total Chlorine Dose 4.3 mg/Jl - 18 mg/Jl 9.0 mg/Jl
Combined Residual 0-1.7 mg/£ 0.5 mg/Jl
Free Residual 0-2.7 mg/Jl 1.2 mg/Jl
Raw Water NPTOC* 4.5 mg/Jl - 19.2 mg/Jl 8.4 mg/Jl
Finished Water NPTOC 2.3 mg/Jl =12.2 mg/1 4.7 mg/1
Chlorine Demand (Total Dose -
Total Residual) 2.8 mg/Jl - 15.7 mg/Jl 7.3 mg/Jl
Finished Water pH 7.3-9.5 (one unknown)
Number of Utilities with the following characteristics:
River Source - 5 Old Granular Activated Carbon - 2
Lake or Reservoir Source - 3 Raw Water Chlorination - 6
Shallow Ground Water Source - 1 Settled Water Chlorination - 3
Filtration - 9 Post-Chlorination - 6
Precipitative softening - 4
*Non-purgeable total organic carbon
TABLE 3
ANALYSIS OF NINE UTILITIES HAVING HIGH CONCENTRATIONS OF CHLOROFORM IN THE
REGION V SURVEY
Chloroform 127 yg/Jl - 366 ug/Jl 203 yg/Jl
Total Chlorine Dose 4.5 mg/Jl - 13 mg/Jl 7.4 mg/Jl
(2 unknown)
Number of Utilities with the following characteristics:
River Source - 8
Lake Source - 1
Raw Water Chlorination - 3
Post-Chlorination - 7 (2 unknown)
-------
- 12 -
SUGGESTIONS FOR ALTERNATE TREATMENT
In general terms, the reaction to produce chloroform is as follows:
Chlorine + Precursors > Chloroform + Other Trihalomethanes
This implies that three approaches to chloroform concentration control
are available:
1. Change disinfection practice.
2. Treat to reduce the precursor concentration.
3. Treat to reduce the chloroform concentration after formation.
In these studies changing the point of application of chlorine in a
treatment plant and the use of ozone and chlorine dioxide were evaluated
as techniques for changing chlorination practice. For control of precursor
concentration, adsorption on powdered- and granular activated carbon, ozonation,
and chlorine dioxide were investigated. Adsorption on powdered- and granular
activated carbon, ozonation and aeration were studied as methods for chloroform
removal.
Change in Chlorination Practice
Changing the Point of Application of Chlorine
Data developed during this study (Figures 5, 7*) show that less
chloroform is formed if chlorine is added to water with the lowest possible
organic content (highest quality). Therefore, the quickest and the least
expensive method of maintaining low chloroform concentrations in finished water,
would be for a utility to chlorinate the highest quality water possible.
*See the Appendix for Figures cited in this Section.
-------
- 13 -
If water is filtered, the highest quality water is filter effluent.
In many water treatment plants, however, chlorine cannot be added at this point
because insufficient contact time is present to permit adequate disinfection
before use, unless additional contact tanks were constructed. Therefore, chlorination
of settled water just prior to filtration may be the best alternative. This also
has the advantage of having a disinfectant pass through the filters, thereby
keeping them cleaner.
Chlorination of coagulated and settled water has disadvantages, however.
One is that the concentration of chloroform and other triha-lomethanes will
be reduced, but not eliminated in the finished water and will continue to be
organic
formed during distribution. Secondly, other/by-products produced during
cholorination may or may not be reduced in concentration, but will not be
eliminated. Finally, the absence of a disinfectant at the beginning of
treatment may cause problems because of the growth of algae, slimes and higher
forms in the early part of water treatment plant. Periodic shock chlorination
could possibly control these problems, but at those times some chloroform
formation would occur.
Use of Ozone or Chlorine Dioxide Instead of Chlorine
Performance
Neither ozone nor chlorine dioxide produced measurable quantities
of trihalomethane (Table IX, X)* when used as a disinfectant. Although this
appears favorable, uncertainties exist with both disinfectants. The use of
ozone does not produce a disinfectant residual to be carried throughout the
distribution system. Further, the health hazard, if any, of the by-products
of the reaction of ozone with organic matter occurring in water is not
known. The same situation in general, exists with chlorine. Except for
trihalomethanes, chloramines, and chlorophenols, little is known about
*See the Appendix for the Tables cited in this Section.
-------
- 14 - (Page number 15 not used).
the by-products formed during chlorination.
Finally, if chlorine is used to provide a disinfectant residual following
ozonation, trihalomethanes will then be formed and little improvement is
gained when compared to chlorination. Combined chlorine has been suggested
as a possibility (Figure 8), but its ability to control aftergrowth in the
distribution system is questionable.
Chlorine dioxide, on the other hand, does produce a residual, which is
an advantage. A disadvantage is the possible toxicity of the organic by-products
resulting from the reaction of chlorine dioxide with organic matter in water,
again, similar to possible undiscovered problems with chlorine. Furthermore,
a few citations in the literature have indicated concern over toxicity of
chlorite, a product of the reactions of chlorine dioxide when added to natural
10,11
water.
Another problem with chlorine dioxide is its generation. The reaction
of sodium chlorite (NaC109) and sulfuric or hydrochloric acid will produce
chlorine dioxide without chlorine being present, but this reaction is
inefficient. Therefore, chlorine dioxide is usually generated by reacting sodium
chlorite with chlorine. Because this reaction proceeds better at low pH,
excess chlorine is usually added to reduce the pH. This produces a chlorine
dioxide and chlorine mixture. The quantity of excess chlorine in chlorine
dioxide can be reduced by adjusting the pH with acid and carefully controlling
the ratio of chlorine to sodium chlorite. Although data indicate that the
resultant chloroform concentration will be lower when chlorine and chlorine
dioxide are used together when compared to the use of chlorine alone (Table X),
trihalomethanes will not be absent if excess chlorine contacts the water.
-------
- 16 - (Page number 15 not used).
Cost
In an effort to compare the cost of alternate disinfectants, calculations
were made assuming no disinfection facilities at a water treatment plant.
Therefore, costs of disinfection equipment, plus the cost of a disinfectant
12
contact chamber is included for chlorine, ozone and chlorine dioxide.
The summary of these data, Table 4, shows that the cost of all three disinfectants
are similar. Of course, in an existing plant, a change to ozone might mean
the abandoning of some of the existing chlorination facilities, so on that basis
the costs in Table 4 might not be too applicable. On the other hand, a water
treatment plant now using chlorine could add chlorine dioxide capability at a
small incremental cost.
In this cost analysis, a lower ozone and chlorine dioxide dose was
compared to the chlorination dose on the basis that both of these disinfectants
are more effective than chlorine and therefore less disinfectant would be
required. During the experiments described in the Appendix for example, 0.5 mg/Jl
of either ozone or chlorine dioxide was sufficient to adequately disinfect the
effluent from the dual-media filter in the pilot plant. A dose of 1.3 mg/£ was
required for chlorine to accomplish the same disinfection, although the
pilot plant was not performing too well at this time, so this might not be the
minimum chlorine dose. Finally, using a lower disinfectant dose should form
less non-trihalomethane organic by-products.
To determine the cost of disinfection to a typical household, a family of
four was assumed to use 200 gallons of water per day. If this rate of
consumption was steady throughout a calendar year, the annual usage would be
73,000 gallons. Therefore, multiplying any disinfection cost in cents per
1000 gallons times 73 would produce an estimate of the annual cost to a
typical household for a given treatment process. Using the costs in Table 4,
-------
- 17 -
TABLE 4
ESTIMATED COST OF DISINFECTION*
All Costs in Cents/1000 gallons
Design Capacity
Average Daily Flow
Chlorine 10^/lb
Chlorine 20«
-------
- 18 -
chlorination contributes from about $0.40 to $2 per year to the water bill of
an average household, depending on treatment plant size. On the same basis,
changing to ozone would make these figures about $0.50 to $4 per year, while the
use of chlorine dioxide would cost from about $0.70 to $2 per year. Note: Becausi
of the influence of local conditions these costs should be considered approximate^
In summary, changing disinfectants from chlorine to ozone or chlorine
dioxide will prevent chloroform formation during water treatment. As with
chlorine, the problem of yet undetected organic by-products being formed during
disinfection with these oxidants cannot be overlooked unless either of these
disinfectants is applied to low organic content water.
Combined chlorine is not as reactive as free chlorine for the formation
of chloroform, see Figure 8. Therefore, if a utility should add ammonia in
conjunction with chlorine addition, such that no free chlorine residual ever
existed, chloroform formation should be low. The mere presence, however, of a
combined chlorine residual in the finished water does not assure that free
chlorine was not present sometime earlier during the treatment of the water.
For example, many water utilities in the National Organics Reconnaissance Survey
that had a finished water with only combined residual did have substantial quantit
of chloroform in their water. In these cases, free chlorine residual must have
been reacting with chloroform precursor at some time during the treatment of the
water. In spite of this, combined chlorination as a primary disinfectant
disinfecting power to judge its
was not discussed in this sub-section because not enough is known about its /
13
value at this time. Combined chlorine may, however, have a potential as a
secondary disinfectant to provide a residual in the distribution system
following ozonation.
-------
- 19 -
Control of Precursor Concentration
Control of chloroform precursor concentrations was attempted by four
techniques — adsorption on powdered- and granular activated carbon, the use
of chlorine dioxide, and ozone. Unrealistically high doses of powdered activated
carbon only resulted in partial precursor removal (PagevA-28 ),the use of
chlorine dioxide was only moderately successful (Table X),and very high doses
of ozone were required to produce measurable results (Tables XI and XII).
Adsorption on granular activated carbon, on the other hand, was successful
(Figures 10 and 11).
Granular Activated Carbon
Performance
When fresh, granular activated carbon:
a) will adsorb trihalomethanes that have been formed by chlorination
practiced prior to granular activated carbon treatment (Figure 3);
b) will adsorb most trihalomethane precursors so that following granular
activated carbon treatment chlorination can be practiced without forming much
trihalomethane (Figure 10, 11);
c) will reduce the possibility of producing hitherto unknown organic by-
products during disinfection because little organic matter will be present
with which any disinfectant, chlorine, ozone, or chlorine dioxide can react, and,
d) will, beyond removing chloroform and trihalomethane precursors .produce
water with a low overall concentration of organic matter, thereby increasing the
likelihood of the removal of raw water organic contaminants that may be
of health concern now or in the future (Figure 11).
-------
- 20 -
In spite of the advantages listed above, this treatment technique is not
without its disadvantages. The performance noted above can be achieved when
granular activated carbon is fresh, but this effectiveness does not last for
a long time (Figure 11). For example, in a plant treating water similar in
character to the Ohio River (the source used in these experiments), removing
the filter media, replacing it with 2-3 feet of granular activated carbon, and
operating the filters at conventional approach velocities of 2-3 gpm/sq ft,
would provide the excellent performance described above from the granular activated
carbon for about one month. This time period will vary if water with an organic
content much higher or much lower than that in the settled water used in these
studies, about an NPTOC of 1 mg/£ in the winter, 1.1 rag/A in the spring, 1.4 mg/£
in the summer and 1.3 mg/£ in the fall, is applied to granular activated carbon.
Within these limits, however, with respect to performance, granular activated carbc
adsorption is the best technique for precursor removal yet investigated.
As noted above, although not directly related to the chloroform problem in
drinking water, granular activated carbon has the ability to adsorb many other orgc
Because adsorption is not complete, however, some uncertainty exists relating to t
exact organic content of the effluent from fresh granular activated carbon beds.
At this time, measurement of the total organic carbon content of fresh granular
activated carbon bed effluent is not possible. Although non-purgeable total organ
carbon concentrations in fresh granular activated carbon bed effluents are relativ
low, commonly less than the 0.1 mg/Jl detection limit of the analytic method,
the concentration of the "purgeable" total organic carbon fraction in fresh granul
activated carbon bed effluents is not known, although it is not expected to be hig
Nevertheless, granular activated carbon has the ability to adsorb many
specific organics of current concern, even when partially exhausted for the remove
of general organic compounds as measured by the NPTOC test. For example, taste ar
-------
- 21 -
odor causing compounds are removed for many months following the breakthrough
of NPTOC. Further, several years ago, partially exhausted granular activated carbon
was shown to remove dieldrin, lindane, 2,4,5-T, DDT and parathion. Finally,
granular activated carbon removed 30 pg/i. of naphthalene spiked into Cincinnati, Ohio
tap water for eight months even though other organics were penetrating the bed
after that time. This is not true for chloroform, however, which is much more
water soluble than naphthalene. Chloroform penetrated a granular activated carbon
bed treating Cincinnati, Ohio tap water about ten days to two weeks before NPTOC.
Therefore, although some uncertainties exist relating to the complete organic content
of the effluent from fresh granular activated carbon beds, past evidence indicates
that non-polar synthetic organics with low solubility in water are well adsorbed by
granular activated carbon. Also, the data presented in the Appendix of this Guide
do demonstrate that granular activated carbon can be very effective for removal of
precursors of chloroform and other trihalomethanes.
Cost
Attempts have been made to estimate the additional cost of water treatment using
granular activated carbon. The details of the assumptions used in the computer
program developed from the data in the "Technology Transfer Process Design Manual
for Granular Activated Carbon Systems " are presented in Table 5. Once calculated,
these costs were then analyzed to determine the influence of the following six
variables on the estimated cost: reactivation frequency, percent granular activated
carbon attrition loss per reactivation, granular activated carbon costs, fuel
cost, wage rate, and electric power cost. An analysis was also made of the cost of
off-site reactivation for small plants and the cost of using granular activated
carbon after filtration rather than as a combination filtration-adsorption medium.
Using the factors in Table 5, the costs for granular activated carbon
calculated to be as shown in Table 6.
-------
- 22 -
TABLE 5
FACTORS USED IN GRANULAR ACTIVATED CARBON COST CALCULATIONS
Activated Carbon cost $0.38/lb
Interest rate = 7%
Design life = 20 years
Construction Cost Index = 2.257 (Jan. 1975)
Wage Price Index = 1.675 (January 1975)
Hydraulic Loading Rate = 2 gal/min/sq ft
Contact time =4.5 min (Apparent)
Direct hourly wage rate = $4.730/hr (Jan. 1975)
Fuel cost - $1.26 per mil BTUs
Activated Carbon attrition loss per reactivation cycle = 10%
No loss in adsorptive capacity during reactivation
Volume of granular activated carbon per 1 mgd filter = 865 cu ft
Reactivation frequency = once per month at 100% of design capacity, once
per 1.4 months at 70% of design capacity (see text for discussion)
Activated carbon used as replacement for granular filter media
On-site reactivation furnace, multiple hearth
Plant production 70% of design capacity (A typical yearly average)
-------
- 23 -
TABLE 6
ESTIMATED COST OF GRANULAR ACTIVATED CARBON TREATMENT*
Design capacity, mgd 1 10 100 150
Average daily flow, mgd 0.7 7 70 105
Total cost, ^/lOOO gal. 40** 11 6 5
*These costs will vary at different locations (see Table 7) so should be
considered approximate.
Note: Of these costs, from 19 to 27% is capital cost and the remainder is
operation cost. Also, the Inventory of Public Water Supplies shows
that about 420 utilities using surface water as a source have a
design capacity of 10 mgd and greater. They serve about 77 million
customers.
**See page 25 for discussion of this cost.
These costs were calculated with the plant operating at 70 percent of
design capacity. If a plant were operating at a different percentage of design
capacity, unit costs might be expected to change proportionally to the flow
change. This does not happen, however, because as the hydraulic load on
granular activated carbonchanges,so does the organic load, if water quality
remains constant. Changing the organic load changes the time of organic breakthrough
thereby changing the time between reactivations. As shown in Table 7, this changes
the unit cost to somewhat modify any change caused by changing production.
For example, at a 100 mgd facility producing 100 mgd and reactivating
the granular activated carbon each month, the unit cost would not double when
the plant was operated at 50 percent of capacity, but would only increase 1.2
times because the granular activated carbon would only need to be reactivated
once every two months. Further, if water quality improved sufficiently during
this period of low demand to allow a three-month reactivation frequency, the unit
cost would not increase at all. In the calculations used for Table 6 the
-------
- 24 -
time between reactivation was assumed to increase from 1 month to 1.4 months
as the plant production decreased from 100 percent to 70 percent of design
capacity.
In most major metropolitan areas, the cost of drinking water, including
coagulation, flocculation, sedimentation, filtration and disinfection, is from
30 to 50 cents per 1000 gallons. Of this overall cost, only 5 to 8 cents per
1000 gallons is the cost of water treatment. EPA has nearly completed a study
that shows the average cost (not price) of drinking water in eleven major
utilities to be about 43 cents per 1000 gallons. Twelve percent of these costs
are for treatment, with the balance for acquisition of water, pumping, salaries
of employees, administration, amortization of distribution systems, and other
nontreatment costs. Using the previously discussed value of 73,000 gallons of
water usage per year for a household of four at 43 cents per 1000 gallons the
annual cost for water would be about $31, about $4 of which would be for
treatment.
The use of granular activated carbon adsorption as an additional treatment
process for controllingorganic contaminants would add from $4/year to $7/year
to the water bill for a household of four for treatment plants in the 10 to
150 mgd range, using the data in Table 6. Of course, in locations where the
organic content of the water is greatly different from that used in the studies
these
upon which / costs are based,the cost for treatment using granular activated
carbon will be different.
-------
- 25 -
The influence of six variables (reactivation frequency, percent granular
activated carbon attrition loss per reactivation, granular activated carbon
cost, fuel cost, wage rate and electric power cost) on cost can be classified
as minor, intermediate and major, see Table 7. Of these, variations in
fuel costs, and electric rates had a minor influence on the final cost of
treatment; the cost of the granular activated carbon, the percent granular activated
carbon attrition loss per reactivation cycle and wage rates (particularly
in smaller plants) had intermediate influence, whereas the most important
variable was reactivation frequency. Note that studies have not yet been
completed to determine the loss in adsorption capacity, if any, of granular
activated carbon for chloroform precursor removal during reactivation. If the
loss in capacity were high, this would increase the final cost.
The cost of on-site reactivation for a 1 mgd plant is very high, 40^/1000 gallons,
see Table 6. These costs could be lowered to more reasonable levels if the
exhausted granular activated carbon was transported to a central reactivation
facility. For example, the cost of granular activated carbon treatment could
be reduced to about 16^/1000 gallons by transporting the granular activated
carbon as far as 100 miles to a reactivation facility of the size equivalent
to that required at a 10 mgd water treatment plant. This makes the cost for a
1 mgd facility approach that of a 10 mgd plant. If granular activated carbon
treatment is desired for small water treatment plants, regional reactivation
facilities will be essential.
The final analysis was to determine the cost of constructing and operating
granular activated carbon contactors as a unit process following filtration.
The assumptions used in this calculation are given in Table 8.
-------
- 26 -
TABLE 7
INFLUENCE OF VARIATIONS IN FACTORS USED TO CALCULATE COSTS -
100 MGD PLANT
Relative Cost **
Variable
Fuel
Cost
($/BTU)
Electric
Cost
(j^/kw-hr)
Activated
Carbon
Cost
($/lb)
Activated
Carbon
Loss
(% per rea. cycle)
Direct
Hourly Wage
Rate ($/hr)
Reactivation
Frequency
Times per month
Level* High Intermediate
High 1.89 1.02
Int. 1.26 1.00
Low 0.63
High 1.5 1.02
Int. 1 1.00
Low 0.5
High 0.57 1.40
Int. 0.38 1-00
Low 0.19
High 15 1.40
Int. 10 1.00
Low 5
High 7.095 1.16
Int. 4.730 1.00
Low 2.365
High 2.0 2.20
Int. 1.0 1.00
Low 0.67
Low
0.98
0.98
0.60
0.60
0.84
0.60
*The intermediate level is that used in Tables 5 and 6. The high value was taken
as 150% and the low value as 50% of the intermediate level.
**The cost associated with the intermediate value is arbitrarily taken as 1.00
and the other costs calculated as a ratio based on the intermediate value cost.
-------
- 27 -
TABLE 8
ASSUMPTIONS FOR POST-ADSORPTION
Item
Number of contactors
2
Hydraulic loading (gal/min/ft )
Diameter contactors (ft)
Depth of contactors (ft)
3
Volume of granular activated carbon (ft )
AnDarent flnnt-art- t-ime fm-in . '\
10 mgd
3
10
15
20
10,000
8
100 mgd
15
10
20
20
94,200
8
Reactivation frequency (months) at 100% of
capacity 2 2
Reactivation frequency (months) at 70% of
capacity 2.9 2.9
Activated carbon attrition loss (% per
reactivation) 5 5
Average Daily Flow, mgd 7 70
Although more capital intensive, calculations indicate that using
granular activated carbon in a post-filtration mode is somewhat more economical
than replacing the granular media in the existing filter boxes with granular
activated carbon, see Table 9. The reason the overall cost is less is that
post-filtration contactors can be constructed of any size and shape and could
be made deeper. The longer contact time would permit longer periods between
reactivations. The resultant reduction in cost, plus the reduction in cost
that would accrue from a lower percentage of activated carbon loss per
reactivation cycle because the handling of the granular activated carbon from
these contactors would be facilitated, overcomes the increased cost of
construction new facilities.
-------
- 28 -
TABLE 9
COMPARISON OF CAPITAL REQUIREMENTS FOR EQUIVALENT ADSORPTION/FILTRATION AND
POST-FILTRATION ADSORPTION SYSTEMS FOR 10 AND 100 MGD AT 70% LOAD FACTOR
Capital Costs Operating and Total Cost
(«?/1000 gal.) Maintenance Cost (^/lOOO gal.)
10 100 10 100 10 100
mgd mgd mgd mgd mgd mgd
Adsorption/Filtration 31 85 11 6
Post-Filtration
Adsorption 53 52 10 5
*These costs will vary at different locations (see Table 7) so should be
considered approximate.
In summary, although all desirable information is not available on
the performance of granular activated carbon, its capabilities are so favorable,
removing of trihalomethane precursor so that less than 5-10 yg/£ chloroform
is formed upon chlorination with few other by-products formed and producing
water with an NPTOC concentration of less than 0.1 to 0.2 mg/£, that its
benefits are worth the cost.
-------
- 29 -
Removal of Chloroform
Four techniques were studied to remove chloroform from water after it has
formed - the use of adsorption on powdered- and granular activated carbon,
aeration and ozonation. Unrealistically high doses of powdered activated carbon
(Table V) and ozone (Table VII) were required to effect substantial, although
not complete, removal of chloroform and other trihalomethanes. Preliminary
cost estimates indicate that both of these processes would be prohibitively
expensive for the removals obtained.
Both aeration (Table VI) and adsorption on granular activated carbon
(Figure 3) were effective for chloroform removed under certain circumstances.
Granular activated carbon treatment is, however, effective beyond just removing
chloroform as was discussed above. To determine the approximate cost of a
diffused-air aeration system for the removal of about 90 percent of the
chloroform formed during water treatment, an aeration basin constructed following
final disinfection with a 20-minute detention time and an air-to-water ratio
of 30 to 1 was assumed. Table 10 shows the cost of this form of treatment is
reasonable, from about $2/year to about $10/year additional cost for water for
a household of four assuming a household of four uses 73,000 gallons of water
annually.
TABLE 10
ESTIMATED COST OF AERATION*, 20-MINUTE DETENTION TIME, 30 TO 1 AIR-TO-WATER RATIO
Design Capacity, mgd 1 10 100
Average Daily Flow, mgd 0.7 7 70
Total cost, «5/1000 gal. 14 73
*These costs will vary at different locstions, so should be considered
approximate.
-------
- 30 -
which
The reaction of the chlorine residual/is not removed by aeration
(Table VI), with remaining precursor as the water passes through the distribution
system would probably raise the chloroform concentration before it reaches the
consumer, thereby somewhat negating this as an effective form of treatment.
Using Cincinnati, Ohio as an example of a situation where water containing both
a free chlorine residual and chloroform precursor leaves the treatment plant,
on May 9, 1975 water at the treatment plant contained about 80 yg/£ of chloroform.
After about three days passage through the distribution system the chloroform
concentration had increased to slightly over 120 yg/&. Further, aeration might
raise the dissolved oxygen level in the water, thereby aggrevating corrosion
problems in the distribution system. Therefore, a treatment technique for
the control of precursor concentrations or changing chlorination practice is
better than merely removing chloroform from finished water by aeration.
-------
- 31 ~
SPECIFIC EXAMPLES OF ALTERNATE TREATMENT FOR DIFFERENT TYPES OF WATER TREATMENT PLANTS
Chlorination Only
Many ground water supplies in the United States are not chlorinated and
many others draw water from a sufficiently high quality source that chlorination
is the only treatment. If chlorination is not practiced, no trihalomethanes
will be formed. If the water is of sufficiently high quality that it contains
little precursor, little trihalomethane will be formed upon chlorination.
If the choroform concentration is of concern to a utility practicing
disinfection only, and they desire to reduce the concentration in their
finished water, they must consider either using disinfectants such as ozone
or chlorine dioxide or <- building a treatment facility using
granular activated carbon to remove the chloroform precursors.
Aeration for Iron or Odor Removal
Although aeration is frequently practiced for the removal of such reduced
materials as ferrous iron, manganeous manganese, and hydrogen sulfide, this
aeration step is frequently before chlorination in the treatment process.
Therefore, although trihalomethanes can be removed from water during aeration,
this would not be the case in the situation described above. A utility wishing
to lower the chloroform content of their finished water in this situation
must consider using disinfectants such as ozone or chlorine dioxide or using
granular activated carbon for the removal of chloroform precursors.
Coagulation, Settling and Filtration for_Turbidity Removal
The utility treating water with the following unit processes: raw water
chlorination, rapid mix with coagulant addition, flocculation, sedimentation,
rapid granular filtration, and disinfection may have to consider several
alternatives to reduce the chloroform level in their finished water. For
example, changing the point of application of chlorine from the raw water
to a point just prior to the rapid granular filters should reduce the
-------
- 32 -
chloroform concentration in the finished water. The utility will have to
consider an alternate disinfectant such as ozone or chlorine dioxide to lower
the chloroform concentration further. Another alternative would be use of
granular activated carbon for the adsorption of organic matter, thereby
removing chloroform precursor.
Softening
Precipitation
A water utility practicing precipitative softening, whether lime-, excess
lime-, lime-soda ash-, or excess lime-soda ash softening, should consider
moving the point of application of chlorine to after recarbonation. This would
have the dual advantage of chlorinating a lower organic content water and
at a lower pH. Studies have shown that chlorination at higher
pH enhances the chloroform formation reaction. The utility will have to
consider using an alternate disinfectant such as ozone or chlorine dioxide
to further lower the chloroform concentration. Chlorination could continue
to be practiced if the utility used granular activated carbon as a treatment
technique for removing chloroform precursor by adsorption.
Ion-Exchange
In the National Organics Reconnaissance Survey only two utilities
employing ion-exchange for softening were included. Both had low concentrations
of chloroform in their finished water. In one case, this was not unexpected
because the raw water NPTOC concentration was low, but at the other utility
this was not the case, and the low chloroform concentration in the finished
water was somewhat surprising. Little information exists concerning the
behavior of chloroform in the presence of ion-exchange resins, but its
adsorptive properties would be expected to be low.
-------
33
Chloroform concentrations could be reduced, however, if chlorine is added
to the highest quality water in the treatment plant, commensurate with adequate
contact time for good disinfection. A disinfectant such as ozone or chlorine
dioxide should be considered if further reductions are desired. Chlorination
could be continued, however, if granular activated carbon were used as a
precursor removal unit process.
Treatment for Taste and Odor Control
Aeration
If treatment plants practicing aeration for taste and odor control have
the aeration process located following chlorination, some chloroform will be
lost to the atmosphere. Data have shown that higher than usual air to water
ratios will be necessary to achieve good removal of chloroform, therefore the
utility that has the capability of varying the intensity of aeration should
maximize it to help improve removal of chloroform.
Powdered Activated Carbon
Data show that powdered activated carbon is not particularly efficient
either for removing chloroform or chloroform precursors. Nevertheless, a water
treatment plant that has the capability of feeding the powdered activated carbon
should consider increasing the dose and possibly adding it at several points
through the treatment plant to increase its effectiveness. This may not be
the complete solution to a given problem, but if the dose can be raised to a
very high level,powdered activated carbon should help control chloroform
concentrations. This may, however, create a sludge disposal problem.
-------
- 34 -
Although both aeration and powdered activated carbon treatment will have
some effect on chloroform concentration in the finished water, powdered activated
carbon will have little effect on chloroform precursor concentration. The
effect of aeration on precursor concentration has not been studied, but it is
expected to be minimal. Therefore, the potential for the reaction of chlorine
and precursor would still exist following these unit processes and the
chloroform concentration would probably increase as the water passes through the
distribution system in the presence of free chlorine residual.
Chlorine Dioxide
Some utilites currently have equipment for generating chlorine
dioxide so that it can be used as an odor control process. In these circumstance!
chlorine dioxide reacts with phenol and prevents the formation of disagreeable
upon chlorination.
chlorophenolic tastes and odors/ Utilities with such equipment should consider
using chlorine dioxide both for odor control and for disinfection, if they desire
to reduce the chloroform concentration in their finished water. If this is
practiced care should be taken to generate the chlorine dioxide with
because
little or no chlorine in it, / this chlorine will react with chloroform
precursors and form some chloroform in the finished water.
Granular Activated Carbon
Some utilities currently have granular activated carbon adsorbers in their
treatment plant for the purpose of controlling the taste and odor causing
organics that previously had been creating consumer complaints. Analysis of
several of these water utilities in the National Organics Reconnaissance
Survey and Region V Survey indicated that, as currently operated, these
adsorbers are not controlling the chloroform concentrations in the finished
waters. To improve performance, these utilites should reactivate the granular
activated carbon more frequently to obtain its maximum benefit. Making this
-------
- 35 -
change in operation should allow these utilities to produce a water nearly
free of trihalomethane and low in general organic content.
-------
- 36 -
MONITORING
A utility attempting to control the concentration of chloroform should
consider a monitoring program along with any changes in treatment. This
monitoring program should be developed so that concentrations of chloroform and
other trihalomethanes can be determined, thereby evaluating the effectiveness
of the changes in treatment that were made to control chloroform.
The analytic technique for chloroform, involving purging the chloroform
from a sample with an inert gas prior to introduction into a gas chromatograph,
7 8
with a halogen specific detector has been described in the literature. '
Although this determination is not difficult, qualified technicians are
required to produce precise and accurate results. Any gas chromatographic procedur
requies some skill and knowledge to perform properly and this is no exception.
If such qualified technicians are not available on a utility's staff, samples
may have to be shipped to another laboratory where they can be analyzed. This
is less desirable than having the analytic capability in the utility, but is
an acceptable solution to the problem of monitoring in some circumstances.
If a utility is making these analyses itself, samples should be collected
at various stages of the water treatment process to determine where and what
quantity of chloroform is being formed. Further, because the concentration of
trihalomethnaes may change from the treatment plant to the point of use, samples
should be collected at various places in the distribution system to determine
the increase, if any, in chloroform concentration as the water moves from the
treatment plant to the consumer. Samples should be dechlorinated with sodium
thlosulfate to avoid changes prior to analysis.
-------
- 37 -
Studies have demonstrated that precursor concentrations change during
different seasons in some raw waters, thereby, changing the ultimate chloroform
concentration. Therefore, tap water samples should be taken frequently
enough to allow a utility to be aware of changes in raw water such that they
know the eventual chloroform concentration reaching the consumer at any time.
The initial cost of the equipment to make chloroform analyses is in the
range of $7,000 - $8,000 Once this initial purchase is made, the analysis
cost will mostly be the cost of the analyst's time. A typical analysis for
chloroform requires from 45 minutes to 1 hour with about 6 samples being able
to be processed in a typical working day, considering the time for standardization
of the gas chromatographic detector response.
A water utility practicing organic removal may also wish to monitor the
effectiveness of those unit processes. This is usually done by measuring
the total organic carbon concentration before and after such a unit process.
At the present time, equipment is available that will measure the portion
of TOG remaining in a sample that has been purged of carbon dioxide under
acidic conditions. The detection limit of this equipment is approximately
0.1 mg/£, with a precision of about - 0.1 mg/JL The cost of such equipment
is about $10,000 and also requires the same type of qualified technician
needed to make a chloroform analysis for reliable operation.
Another analytic determination currently under development would measure
total organic halogens (TOH). If the development of this test is successful,
it could be used advantageously to determine the concentration of all of the
halogen-containing organics (non-natural) in a single analysis. A variation
of this procedure that measures most of the total organic chlorine (TOC1)
is used in Europe to monitor the performance of granular activated carbon
beds. Samples are collected from within the bed at a point about six inches
-------
- 38 -
above the bottom. When the TOC1 concentration begins to rise in these
samples, breakthrough is near and the bed is reactivated. A similar approach
could be taken using NPTOC measurements.
-------
- 39 -
REFERENCES
1. Bellar, T.A., Lichtenberg, J.J. and Kroner, R.C., "The Occurrence of
Organohalides in Chlorinated Drinking Water," Jour. AWWA, 66:12:703
(December 1974).
2. Rook, J.J., "Production of Potable Water from a Highly Polluted River,"
Water Treatment and Examination, 21, Part 3, 259-274 (1972).
3. Rook, J.J., "Formation of Haloforms During Chlorination of Natural
Water," Water Treatment Exam., 23:2:234 (1974).
4. Love, O.T., Jr., Carswell, J.K., Symons, J.M., Stevens, A.A., Miltner, R.J.,
Kropp, K.L., Smith, B.L. and Keller, P.A., entitled "Treatment for the
Prevention or Removal of Chlorinated Organics in Drinking Water,"
to be submitted to the Journal of the American Water Works Association.
5. Stevens, A.A., Slocum, C.J., Seeger, D.R. and Robeck, G.G., "Chlorination
of Organics in Drinking Water," submitted to the Journal of the American
Water Works Association for publication and Proceedings of Conference on
the Environmental Impact of Water Chlorination, Oak Ridge, Tennessee,
October 22-24, 1975.
6. Report on the Carcinogenesis Bioassay of Chloroform, Carcinogen Bioassay
and Program Resources Branch, Carcinogenesis Program, Division of Cancer
Cause and Prevention, National Cancer Institute, Bethesda, Maryland.
7. Bellar, T.A. and Lichtenberg, J.J., "The Determination of Volatile Organic
Compounds at the yg/£ Level in Water by Gas Chromatography," Journal of
the American Water Works Association, 66, 739 (Dec. 1974).
-------
- 40 -
8. Stevens, A.A. and Symons, J.M., "Analytical Considerations for
Halogenated Organic Removal Studies," Proceedings American Water
Association Water Quality Technology Conference, pp. XXVI-1 - XXVI-6
(December 1974), American Water Works Association, Denver, Colorado (1975).
9. Rook, J.J., "Haloforms in Drinking Water, Journal of the American
Water Works Association. 68, 168-172 (March 1976).
10. Samdal, J.E., "Water Treatment and Examination in Norway, Water
Treatment and Examination. 21, 309-314 (1972).
11. "Clinical Toxicology of Commercial Products." Gleason, Gosselin, Hodge
and Smith, 3rd Edition (1969).
12. Clark, R.M. and Guttman, D., "Cost Calculations of Water Treatment Unit
Processes," Water Supply Research Division, Municipal Environmental
Research Laboratory, U.S. Environmental Protection Agency, Cincinnati,
Ohio (March 1976), Mimeo.
13. Kruse, C.W., Oliveri, V.P. and Kawata, K., "The Enhancement of Viral
Inactivation by Halogens," Water and Sewage Works, 118, 187-193 (June 1971)
14. Process Design Manual for Granular Activated Carbon Adsorption,
Technology Transfer, U.S. Environmental Protection Agency, October 1971.
15. Symons, J.M., Bellar, T.A., Carswell, J.K., DeMarco, J., Kropp, K.L.,
Robeck, G.G., Seeger, D.R., Slocum, C.J., Smith, B.L. and Stevens, A.A.,
National Organics Reconnaissance Survey for Halogenated Organics,
American Water Works Association, 67, 634-647 (Nov. 1975).
16. Miltner, R.J., "The Effect of Chlorine Dioxide on Trihalomethane in
Drinking Water," Master of Science Thesis, University of Cincinnati, 1976.
-------
- 41 -
ACKNOWLEDGMENTS
The author wishes to thank those who reviewed this Guide for all of
the helpful suggestions. These reviewers were: J.K. Carswell, R.M. Clark,
J. DeMarco, O.T. Love, Jr., A. A. Stevens, Gordon G. Robeck, J. Cotruvo,
G. Coad and J. Hoffbuhr. The author also expresses his appreciation to
Ms. Maura M. Lilly who typed the manuscript so promptly.
-------
APPENDIX
CURRENT STATE OF KNOWLEDGE
-------
- A-i-
TABLE OF CONTENTS
SURROGATE MEASUREMENTS A-l
TREATMENT RESEARCH A-5
Introduction A-5
Removal of Chloroform from Water A-6
Granular Activated Carbon A-6
Powdered Activated Carbon A-10
Aeration A-ll
Ozone A-12
Alternate Disinfection Procedures A-14
Point of Application of Chlorina A-14
Ozone A-22
Chlorine dioxide A-24
Chlorination plus Ammoniation A-26
Prevention of Chloroform Formation by Removal of Precursors A-28
Powdered Activated Carbon A-28
Ozone A-28
Chlorine Dioxide A-33
Granular Activated Carbon A-33
Chloroform Precursor Removal A-33
Removal of Precursors for Other Trihalomethanes A-34
FUTURE RESEARCH A-38
-------
- A-l -
SURROGATE MEASUREMENTS
Because the analysis for chloroform is a gas chromatographic procedure
requiring skilled operators and about one hour' to complete, a simple rapid
surrogate measurement that would predict chloroform concentrations seemed
desirable. Chlorine reacts with some organic precursors to form chloroform
and related products, therefore a test that would measure the precursor
concentration in the raw water would be useful for anticipating finished
water chloroform concentrations. No tast for trihalomethane precursors
exists ,so a test for general organic content was considered as an alternative.
The difficulty with using a general organics test as a measure of chloroform
precursors is that the precursor concentrations are not a constant percentage
of the general organic content.
Nevertheless, in the report of the National Organics Reconnaissance
Survey three general organic tests were proposed as surrogate parameters,
non-purgeable total organic carbon (NPTOC), ultra-violet absorbance, and
emission fluorescence scan. Note: Because turbidity interfered with the two
opitcal measurements, ultra-violet absorbance and fluorescence, they were
performed on the finished, rather than the raw water. The report suggested
that raw water NPTOC concentrations were related to total trihalomethane
concentrations in the finished water.
To review this suggestion, these data were statistically analyzed by
three different methods; Spearman Rank Correlation; Linear Regression Analysis,
and Log-Log Regression Analysis. The high level of confidence of the Spearman
Rank Correlation Coefficients in Table I shows that the level of the surrogate
measurements does rise and fall as the chloroform concentration rises and falls.
Table II, however, shows that although the linear regression correlation coefficient
for NPTOC and ultra-violet absorbance with chloroform concentration is fairly
high, the percent of the chloroform concentration variation explained by the
-------
- A-2 -
TABLE I
SPEARMAN RANK CORRELATION OF SURROGATE MEASUREMENTS
Chloroform
Surrogate
Raw Water
NPTOC
Number
of Observations
82
Correlation
Coefficient
0.57
Level of
Confidence
>99%
Finished Water
Ultraviolet
Absorbance
Finished Water
Fluorescence
81
82
0.48
0.42
>99%
>99%
TABLE II
LINEAR REGRESSION OF SURROGATE MEASUREMENTS
Finished water
Ultraviolet
Absorbance
Finished Water
Fluorescence
% of CHC13 95% Confider
Variation Explained Limits Aroui
Chloroform
Surrogate
Raw Water
NPTOC
Number Correlation by Surrogate the Arithmet
of Observations Coefficient Variation Mean (&T.7
82 0.74 54.5 - 79.8 yg/£
81
82
0.54
0.13
29.2
1.8
- 99.9
- 117.4
-------
- A-3 -
variation in NPTOC concentration or ultra-violet absorbance is fairly low,
indicating that other factors are important in determining the chloroform
concentration in a given water. Finally, the scatter of the data is shown by
the magnitude of the 95 percent confidence limits around the mean, again indicating
that these surrogate measurements are poor predictors of chloroform concentrations.
In an effort to improve the usefulness of these surrogate measurements
as chloroform concentration predictors, they were reanalyzed after making a log
transform. The data in Table III shows that making a log transform does, in
general, improve the usefulness of these surrogate measurements, but not sufficiently
to be considered satisfactory. Therefore, the previous suggestion of the
relation of raw water NPTOC concentrations and finished water total trihalomethane
concentrations was overstated.
One other possibility would be to make a multi-variant analysis to take
into account some of the other factors thought to influence chloroform production
such as free or combined chlorine residual, raw water chlorination and so
forth. Even if this would be successful, the resultant equation might be
too complicated to be useful. Therefore, although further research on the three
surrogate analytic procedures might improve them, at this time, the best method
of determining the chloroform concentration is to obtain the necessary equipment
and technical staff to perform the analysis directly.
This is particularly true because special analytic equipment and skilled
operators are required to make the surrogate measurements. These analyses do
have a place in water treatment, however. If an organic removal unit process
is being used by a utility these general organic content measurements are
good process control determinations.
-------
- A-4 -
TABLE III
LOG-LOG REGRESSION OF SURROGATE MEASUREMENTS
Percent of CHCli
95% Confidence
Variation Explained Limits Around
Chloroform
Surrogate
Raw Water
NPTOC
Finished Water
Ultraviolet
Absorbance
Finished Water
Fluorescence
Number of
Observations
82
81
82
Correlation
Coefficient
0.67
0.51
0.41
by Surrogate
Variation
45.5
26.0
16.8
the Geometric
Mean (16.2 ug/£
±7.7
±9.6
± 10.6
-------
- A-5 -
TREATMENT RESEARCH
Introduction
Intensified research on the subject of organic removal has been underway
by the Water Supply Research Division for the past 18 months. Although this
research is not completed, much progress has been made in understanding the
influence of various water treatment unit processes on the organic content of
water. This Section of the Guide will review these findings in an effort to
disseminate what is known about the subject as promptly as possible. As phases
of the work are completed they will be published in detail in the open
4
literature.
Within the context that some of these data are preliminary, the following
three subsections will summarize the current understanding of treatment for the
control of chloroform and related trihalomethanes. These studies have been
carried out in three general areas:
1) Techniques to remove trihalomethanes once they have been formed.
The techniques studied have been adsorption on granular activated carbon and
powdered activated carbon, aeration, and ozonation.
2) Changes in disinfection practice to reduce the amount of trihalomethanes
formed. The changes studied include changing the point of application of
chlorine, the use of alternate disinfectants, either ozone or chlorine dioxide,
and the practice of chlorination plus ammoniation.
3) Methods of treating water so that chlorination will not form
trihalomethanes. The techniques studied were adsorption on powdered activated
carbon, ozonation, use of chlorine dioxide and adsorption on granular activated
carbon.
-------
- A-6 -
Although much of this work covers trihalomethane by-products of
disinfectants (chlorine, ozone and chlorine dioxide), another related problem
with any disinfectant is the possibility of forming by-products other than
trihalomethanes. The use of new analytic procedures revealed hitherto
unknown by-products from chlorination, the trihalomethanes. Other analytic
techniques such as high pressure liquid chromatography and gas chromatography-
mass spectrometry will probably reveal other by-products of chlorination and
undetermined by-products of any other disinfectant that might be considered for
use. Further, the health significance of these organic compounds will probably
be unknown. Therefore, in the discussion below, the problem of unknown
by-product formation, possibly hazardous, must be kept in mind when considering
the behavior of any disinfectant.
Removal of Chloroform from Water
Granular Activated Carbon
In an attempt to determine whether granular activated carbon beds would
be effective for the removal of chloroform and other trihalomethanes, Cincinnati,
Ohio tap water containing various concentrations of four trihalomethanes was
passed through two small glass granular activated carbon columns. The
granular activated carbon was about 30 inches deep, and the water was
2
passed through the beds at a rate of approximately 100 ml/min (2 gpm/ft ). This
results in an apparent (media in place) contact time of 4 to 5 minutes, (8-10
minutes empty bed contact time). One of the columns was filled with a
coal-based granular activated carbon and the other was filled with a
lignite-based granular activated carbon with the characteristics in Table IV.
-------
- A-7 -
TABLE IV
GRANULAR ACTIVATED CARBON CHARACTERISTICS
*•
Coal-base Lignite-base
2 2
Surface Area by Nitrogen Gas 850-900 m /gin 600 m /gm
BET Method
Uniformity Coefficient 1.7 1.7
Effective Size 0.6 mm 0.8 - 0.9 mm
Density 30 lbs/ft3 23.5 lbs/ft3
Figure 3 shows that some chloroform began to appear in the coal-base
column effluent after 3-5 weeks of operation and that the granular activated
carbon was exhausted for chloroform removal after about 8 weeks of operation.
First breakthrough of chloroform occurred after 2 weeks for the lignite-
base granular activated carbon with exhaustion after 8 weeks. In both cases,
the bromine-containing trihalomethanes, occurring at lower concentrations than
chloroform in Cincinnati, Ohio tap water were removed for a few months.
Three operating water treatment plants using granular activated carbon
as adsorption/filtration unit processes were tested fon chloroform removal.
At one plant four weeks after installing fresh granular activated carbon about
88 percent of the chloroform was being removed. Three weeks later, however,
the removal had declined to 28 percent. At the second plant, after ten weeks
of operation the granular activated carbon was only removing nine percent of
the chloroform. The third plant was sampled when the granular activated carbon
was 16 weeks old and no chloroform was being removed. At both the second and
third utility when the granular activated carbon was 7 and 10 weeks old,
respectively, more than 60 percent of the bromine-containing trihalomethanes
-------
- A-8 -
were being removed. In general, these data confirm the data cited above.
Later in the pilot plant test, which lasted a total of 10 months, the
concentration of trihalomethanes in the effluent occasionally exceeded that
in the column influent, somewhat evident in Figure 3. This means that either
previously adsorbed materials were being desorbed or that the granular
activated carbon catalyzed the reaction between chlorine and the precursors
and was actually producing trihalomethanes. A mass balance calculated for the
entire experiment, however, shows that trihalomethanes were not produced in
the granular activated carbon columns, but that desorption was periodically
sufficient to cause the effluent concentration to exceed the influent
concentration.
-------
- A-9 -
0>
•fc
Z
O
<
U
z
O
-------
- A-10 -
Powdered Activated Carbon
In an effort to determine whether or not powdered activated carbon
could be used to remove chloroform and the other trihalomethanes, several jar
tests were performed using Ohio River water. Chlorinated water from the
effluent of the Cincinnati Waterworks pre-plant settling basin containing a
free chlorine residual of 2.2 mg/£ and a chloroform concentration of 64 yg/Jl
was studied in a jar test with 2 minutes rapid mix, 5 minutes slow mix
and 30 minutes settling. Eight doses of powdered activated carbon ranging
from 1 mg/& up to 100 mg/£ were added to the jars in parallel, allowed to
contact the water without coagulant, and the powdered activated carbon
removed by centrifugation. Analysis of the data, Table V, shows that 30 yg/£
of chloroform remains even after the highest dose of 100 mg/£ of powdered
activated carbon was used. Further, if these dosages were used a troublesome
sludge may be created.in full scale practice.
TABLE V
REMOVAL OF TRIHALOMETHANES FROM CHLORINATED SETTLED WATER USING POWDERED
ACTIVATED CARBON (PAC)
PAC
Dose,
mg/£
0
1
2
4
8
16
32
64
100
Cl
Residual
2.2
2.0
1.9
1.9
1.7
0.9
0.5
0.1
0.1
Chloroform
64
52
53
51
51
48
45
35
30
Bromodi-
c hi or ome t hane
yg/£
9
7
7
7
8
8
6
4
2
Dibromo-
chloromi
yg/£
2
1
1
0.9
1
0.8
1
0.7
0.6
Note: Bromoform was not found in any of these samples.
-------
- A-ll -
Aeration
A diffused-air aeration device was built to evaluate the effectiveness of
this technique for removing trihalomethanes from water. Because aeration
or purging is used in the analytic procedure to remove trihalomethanes from
water, this treatment technique was anticipated to be successful. The aerator
used was a continuous flow device, a glass column with a 10-minute contact
time. Table VI shows that a typical air to water ratio used in water treatment
for the removal of taste and odor causing compounds (1:1) did not remove any
measurable amount of chloroform. High air to water ratios were required to remove
84 percent of the chloroform. Because the analytic procedure using a very high
gas to water ratio of approximately 44 to 1 is successful, trihalomethanes can
be purged from water, but not effectively at lower gas to water ratios.
TABLE VI
REMOVAL OF TRIHALOMETHANES FROM TAP WATER BY AERATION
Air:H20
Ratio
0
1:1
8:1
12:1
16:1
20:1
C12
Residual
1.31
1.22
1.21
1.20
1.22
1.09
Chloroform
ygM
99
101
45
33
19
16
Bromo-
dichloro-
me thane
24
5
13
7
8
5
Dibromo-
chloro-
me thane
5
5
3
<1
3
3
Note: Bromoform was not found in any of these samples.
-------
- A-12 -
Chloroform is lost to the atmosphere when water is held in open vessels.
In beakers, standing open at room temperature, almost all of the chloroform
was lost from Cincinnati, Ohio tap water on 3 days standing. In Rotterdam,
The Netherlands, over 90 percent of the chloroform is lost during three
weeks standing in a 20-foot deep holding reservoir just prior to the water
treatment plant. Further, at Whiting, Indiana, the chloroform concentration
declined from 4 pg/£ to 1 yg/£ through the settling tank. In this situation,
ammonia is added at the same point as the chlorine, just prior to sedimentation,
so no free chlorine residual exists through the sedimentation basin. These data
indicate that chloroform is "volatile" and will be lost from the water at any
air-water interface.
In situationswhere a free chlorine residual does persist through the
sedimentation basin, however, chloroform levels increase, in spite of some loss
to the atmosphere, because of a faster reaction of chlorine with precursor.
For example, in Huntington, West Virginia, the chloroform level increased
sedimentation
from 39 yg/£ to 83 yg/£ through the/tanks.
Ozone
A series of tests were made to determine whether or not ozonation was
an effective method of removing chloroform from water. This test was conducted
in a small glass column through which Cincinnati, Ohio tap water passed at
a rate of approximately 80 ml/min, creating a contact time of 5-6 minutes with
the ozone. In an effort to improve the contact between the ozone-oxygen
mixture and the test substrate, a high-speed propeller mixer was placed in the
counter-flow column. The rotation of the propellor increased the downward
flow of water against the rising ozone-oxygen bubbles, causing an almost
complete dispersion of the rising bubble pattern. Table VII shows
that none of the 8 test conditions had noticeable influence on the
-------
- A-13 -
concentrations.
chloroform and the other trihalomethane/ The lack of removal using gas
alone, at a gas-to-water ratio of 0.5 to 1 is in agreement with the aeration
sub
data presented in the previous Section.
TABLE VII
OZONATION OF CINCINNATI OHIO TAP WATER FOR CHLOROFORM REMOVAL
GAS TO WATER RATIO 0.5 to 1
5-6 MINUTE CONTACT TIME
Sample
Tap Water
Mixer Only
Oxygen Only
Air Only
Ozone Only
Mixer + Q
Mixer + Air
Mixer + 0^
*Applied Dose,
mgO_
Applied 0 *
Dose, mg/I
0
0
0
0
25
0
0
25
continuous flow
Chloroform
ygM
10
11
12
12
11
11
12
11
studies, mg/£ =
standard liter
Bromo-
dichloro-
methane
ygM
9
10
10
8
10
9
8
9
gas (0 +
Dibromo-
chloro-
me thane
ygM
6
7
7
4
7
6
6
6
Bromoform
ygM
1
1
l
0.8
0.9
0.5
1
0.9
00) minute
Standard liter of gas (0 +
minute
liters, water
This parameter may not be directly related to the actual oxidation of organic
compounds because of unaccounted for variations in mass transfer or chemical
reaction or both.
-------
- A-14 -
Alternate Disinfection Procedures
This /section will describe research results on the effect of several
alternate disinfection procedures on chloroform formation. When considering
alternate disinfection techniques, caution must be exercised such that no
deterioration in microbiological quality occurs because of the use of any
alternate procedure. Thus, any utility attempting to reduce chloroform
concentrations by altering its disinfection should carefully monitor its
finished and distributed water for microbiological populations to assure itself
and any Regulatory Agency that no decline in quality has occurred.
Point of Application of Chlorine
Research on this topic was conducted in a 150-gallon/day pilot plant
in which Ohio River water was treated with alum coagulant, flocculated, settled,
and filtered. Filters of both dual-media,(coal-sand^ and granular activated
carbon were used. The pilot plant tanks were constructed of stainless steel,
and the filter columns of glass to avoid introduction of extraneous organic
coiqnounds into the treated water through use of plastics. Samples could be
collected after each unit process and water of various qualitites could be
disinfected to determine the influence of a particular treatment process
on the production of chloroform.
Figure 4 shows the relative concentrations of NPTOC after coagulation and
sedimentation and after dual-i-media filtration. To determine whether or not
chloroform precursor was also removed during coagulation, sedimentation and
filtration, raw water, coagulated and settled water, and dual-media filtered
water from the pilot plant were chlorinated in closed containers to determine
the production pattern of chloroform.
-------
- A-15 -
0>
E
< 3
Q i
LU LL.
S "
O
^
O
n
LU
o
U
O
O u-
O
il
=: z
D ui
O %
o2
(J CO
I
m
O
o
S z
00 ^
< D
m
cs
4 O o.
O ^ H.
z S o
O ui
oa I
a: U —
LU
oi
31VDS 3AllV13il
O
-------
- A-16 -
Figure 5 shows that although the initial rate of chloroform production
was similar for all three qualitites of water,'the final concentration of chlorofo
was lower in coagulated and settled water and even lower in filtered water after
7 days of storage. Therefore, chloroform precursor was removed during
coagulation, sedimentation and filtration. For reasons not completely understooc
at this time the rate of formation and ultimate concentration
of dibromochloromethane was almost identical with all three qualities
of water.
In another effort to determine the influence of chlorination practice
on trihalomethane formation, the data in the National Organics Reconnaissance
Survey was sorted with respect to chlorination treatment. During the survey
each water utility was asked whether they pre-chlorinated their water.
Unfortunately, the definition of pre-chlorination varies, at some utilities
this is chlorination of raw water, at other utilities it is chlorination
prior to the filters, chlorination of settled water. The data above
indicate this difference is important. Nonetheless, within the uncertainty
noted, Table VIII shows that water utilities practicing "pre-chlorination"
in general had higher relative trihalomethane concentrations than those
plantspracticing post-chlorination (just before entering the distribution system)
only.
-------
- A-17 -
o
es
O
o
o
00
O
•O
o
es
oo
3
O
u
<
H-
K O
K u
LU
z
u
00 2
ot
o
u.
O
O
es
|/BT/'NoliVcJlN3DNOD WelOdOHOIHD
-------
- A-18 -
TABLE VIII
CHLORINATION PRACTICE INFLUENCE
National Organics Reconnaissance Survey Data
Raw NPTOC Range, mg/Jl
0-1 >l-2 >2-3 >3-4 >4-5 >5
Rel. Rel. Rel. Rel. Rel. Rel.
TTHM*** TTHM TTHM TTHM TTHM TTHM
Category n* Cone. n Cone. n Cone. n Cone. n Cone. n Cone.
All Locations 14 1.00 16 1.00 13 1.00 15 1.00 9 1.00 13 1.00
Pre-Chlorination** 4 2.00 14 0.96 9 1.06 13 1.06 7 1.09 11 1.03
Post-Chlorination 10 0.67 2 1.18 4 0.94 4 0.65 2 0.70 2 0.87
*Number of locations
**See text for explanation.
***TTHM is total trihalomethane, the sum of all of the trihalomethane occurring
in a given water. The values given are all relative to the "All Locations"
value ,arbitrarily taken as 1.00.
These findings encouraged the water utility of Cincinnati, Ohio to attempt
to reduce the chloroform content in their water by moving the point of
chlorination so that clarified water was chlorinated rather than raw water.
Figure 6 is a schematic diagram of the Cincinnati Waterworks. Because of the
addition of a coagulant prior to entering the off-stream storage reservoirs
(Point A) at the time of the study the raw water turbidity was reduced from
approximately 11 Turbidity Units to approximately 2 T.U. entering the
treatment plant.
Although the total trihalomethane production potential of the raw water
was not being determined at this time as a control, the sharp decline in
concentration following moving of the point of application of chlorine from
Point A to Point B in mid-July was attributed to the change in treatment
practice(Figure 7).
-------
to
O
Z
Z
u
Z
u
LL.
O
X
-------
- A-20 -
Over the next few months other changes were made in water treatment
practice, as indicated on Figure 7, but none of. them had as significant effect on
the chloroform concentration as did the first change. The rise in chloroform
concentration after the discontinuance of ferric sulfate and the reduction in
lime dose in the rapid mix (Point 1 - Figure 7), may have been caused by a
decrease in precursor removal during coagulation and settling. Only 1 datum
point was collected under these conditions, however, so firm conclusions about
the influence of adding coagulant on precursor removal are difficult to make.
Similar uncertainty exists related to the decline in chloroform concentration wher
chlorination was moved :T?O after the filters (Points 3 and 4 on Figure 7).
Details on this experimentation will be presented at the 96th Annual AWWA
Conference in New Orleans, Louisiana in June 1976.
At this writing other water utilities are considering making similar
changes in disinfection practice in an attempt to determine the effect of
such changes under their particular circumstances. For example, one utility
changed the point of application from after primary coagulation and settling
ash
to after excess lime-soda/softening (just before filtration) and reduced
the chloroform concentration in the finished water from a little over 200 yg/A
to a little over 100 yg/£. When additional data are available, a better
judgment as to the general applicability of this technique for trihalomethane
concentration reduction will be available.
Finally, the cold weather data, November - December 1975, collected while
chlorine was being added at Point B, Figure 6 is compared in Figure 7 to similar
data collected in January - February 1975 before the change and shows somewhat
lower chloroform concentrations in finished water after the treatment change
even in the winter.
-------
- A- ''I -
I/BT/ 'NOUVaiN3DNOD
H31VM Q3HSINH
-------
- A-22 -
Ozone
In conducting the experimentation on the use of ozone as a possible
alternate disinfectant to chlorine, the assumption was made that post-chlorinatic
would be required in order to maintain a disinfectant residual throughout
the distribution system. Therefore, in the studies reported below, ozonation
was followed by a sufficient dose of chlorine to maintain a chlorine residual
throughout the time of storage of the samples, simulating passage of the
water through a distribution system.
The ozonation equipment used in these experiments was a small glass
counter-flow column plumbed so that effluent from the pilot plant described
above could be passed through it, contacting the applied ozone for about
5-6 minutes, the minimum reported in the literature that was adequate for
disinfection. The first studies that were made were to determine the "applied
ozone dose" required to achieve good disinfection. These studies indicated
that an "applied ozone dose1'of 0.5 mg/£ for 5-6 minutes reduced the standard
plate count in the effluent from the dual-media filter from about 550/ml
to I/ml. No total coliforms could be found. The ozonation studies
therefore were made with an "applied ozone dose" of about 0.7 mg/1 to determine
if a minimum dose of ozone at a short contact time would have an effect on
chloroform concentrations.
Table IX shows that ozonation of the dual-media filter effluent does not
produce chloroform, but ozonation followed by chlorination to produce a
chloroform, actually
residual throughout the distribution system formed/more chloroform than did
disinfection with chlorine alone. The reason more bromodichloromethane was
formed in the chlorinated, but not ozonated sample is not known. The chlorine
dose was chosen as 8 mg/1 to insure that sufficient chlorine was present to
exceed any chlorine demand during the storage period in all samples.
-------
- A-23 -
Unless some other means of providing the residual disinfectant throughout the
distribution system can be found, ozonation for disinfection followed by chlorination
eliminate
to provide a chlorine residual will not . the chloroform concentrations
reaching the consumer. One alternate possibility would be add chlorine and ammonia
to provide a combined chlorine residual. An example of this is Whiting, Indiana.
Data collected on May 29, 1975 showed that the raw water contained 6 yg/£
of chloroform because of chlorination practiced by a nearby industry. After
ozonation, this concentration was essentially the same and at this point in the
treatment chlorine plus ammonia/aSded. After passing through the remainder of
the treatment plant, sedimentation and filtration, the chloroform concentration
was 0.8yg/JL A sample taken at a dead-end at one extremity of the distribution
only contained 2 yg/Jl of chloroform, a very small increase. Although this appears
favorable, the ability of a combined chlorine residual to control microbiological
populations in the distribution system is not well defined.
TABLE IX
OZONATION OF DUAL-MEDIA FILTER EFFLUENT
NPTOC - 1.0 mg/fc
Contact Time: 5-6 minutes
Bromo-
Applied* Applied Storage dichloro-
Ozone Dose Chlorine Dose Time Chloroform methane
Sample mg/l, mg/£, Days yg/Jl
Effluent only
Effluent + 0-
Effluent + Cl.
Effluent 4- 0_ -
3
*Applied Dose,
0
0.7
0
1- Cl 0.7
continuous flow
mg 03
0
0
8
8
studies, mg/£
standard
6
6
6
6
=
liters of
<0.2
<0.2
6
15
gas (0 + 00)
None Found
None Found
14
8
min.
V
standard liter of gas (0 + 0 ) minute liters,water
*This parameter may not be directly related to the actual oxidation of organic
compounds because of unaccounted for variations in mass transfer and/or
chemical reaction rates.
-------
- A-24 -
Chlorine dioxide16
Another disinfectant considered as an alternate was chlorine dioxide.
A chlorine dioxide generator was assembled so that both "chlorine-free"
chlorine dioxide and chlorine dioxide containing excess chlorine used in the
generation of the chlorine dioxide could be added to the effluent from the
dual-media filter in the pilot-plant described above. Microbiological
determinations were made to demonstrate that maintaining a chlorine dioxide
residual of 0.5 mg/1 for 30 minutes was adequate to disinfect the dual-media
filter effluent as indicated by the absence of total coliform and low standard
plate count.
Table X shows that if chlorine dioxide is added to dual-media filter
effluent without any chlorine being present, no chloroform is formed, Test 1.
Test 2, comparing chlorination alone/with use of chlorine dioxide in which
(2b)
chlorine dioxide was added in the presence of excess chlorine^ shows that
chloroform was produced, but less was produced than when chlorination only
was practiced. This may imply that the chlorine dioxide has some effect on
trihalomethane precursor such that it is less available for reaction with
3a and 3b
chlorine. Test /, confirms this phenomenon and demonstrates that the more
chlorine dioxide that is present for an equal amount of chlorine, the lower
the eventual concentration of chloroform. This would imply that if chlorine
dioxide was generated such that only small amounts of chlorine were present,
the amount of chloroform formed might be quite low. Note in these
experiments rather high doses of chlorine and chlorine dioxide were used
in an attempt to demonstrate whether or not certain reactions would proceed
under extreme conditions.
-------
X
w
h4
CQ
^j
H
H
pj
^
ij
fa
W
B3
w
H
rJ
n
2
@
1
rJ
}"*>
Q
O
H
W
P
H
X
0
Q
M
Pi
3
PC
CO
P*H
O
z
o
rH
H
M
Q
1
O
e
o
o
•H
O
Q
O
V4
PQ
6
O
M-l
O
0
03
1 0)
0 fi
O 4=
rH 4J
43 Q)
0 6
1
O
t-l (U
0 C
rH CO
0 W
•r1 CD
T3 6
e
O
14-1
O
^4
o
rH
CJ
(U
60
n)
M (U
o a
J-» T-l
W EH
"«
-a
•H
CMW
rH CO O
O rH rH rH
co r^ ^o
O rH rH
f^ CO rH
O rH CM
CM
«
O CT\
.^
CM -1-1 1 1 O 1
O cO
rH
U
se
a
CO
>>l
CO
Q
rH
*x^
60
6
I— {
**«^.
60
e
rH
^^
60
e
H
^N^,
60
U
0)
H
K
m -"i* sj~ o
• • • •
r^ r^» r^^ r^
"^
in
f^
«
»a- rH CM CM
in
O O CM CO
a>
0 0 fH
m in m
o
rH rH CM
in
CO CO
• •
CM O O O
in co
• ,
m o rH o
cO J3 co
rH CM CM CO
•a
c
3
O
fa
0)
c
o
m
•
*^*
^.
CM
1^
rH
0
•
fNfc
CM
m
•
H
m
•
CM
O
•
ro
c\l
•
-------
- A-26 -
Chlorination Plus Ammoniation
The practice of combining ammonia with chlorine to produce chloramines
for disinfection is poorly understood in water utility practice. Many
investigators think that chloramines are a poor biocide and therefore do not
13
provide adequate protection to consumers from microbiological hazards
Many water utilites that had practiced chloramine disinfection have changed to
free chlorine disinfection in an attempt to provide a better quality water to
their consumers. Nevertheless, many utilities continue to practice chloramine
disinefection and still maintain a microbiological quality of water at the
consumer's tap that meets the requirements of the appropriate Regulatory
Agencies.
To investigate the reactivity of chloramine with respect to chloroform
formation, raw Ohio River water was dosed both with free and combined chlorine.
Figure 8 shows that less than 7 percent of the total chloroform formed after
70 hours when free chlorine was added to the river water was produced when
chloramines were used. Although this information appears favorable, caution must
be exercised in its application to avoid any sacrifice of microbiological
quality.
-------
- A-27 -
120
110
100
_ 90
o>
4.
Z
o
O
u.
o
X
u
80
i- 70
£*
K-
Z 60
UJ
U
Z
O 50
40
30
20
10
2.8 mg/l
FREE CI2
O^ RESIDUAL
OHIO RIVER WATER
5.5 mg/l Cl 2 AND 5.2 mg/l NH3-N ADDED
COMBINED CHLORINE RESIDUAL
9.A mg/l COMBINED CU RESIDUAL
10 20 30 40 SO 60
REACTION TIME, HOURS
70
FIGURE 8. CHLOROFORM PRODUCTION
WITH FREE AND COMBINED
CHLORINE RESIDUAL
-------
- A-28 -
Prevention of Chloroform Formation by Removal of Precursors
Powdered Activated Carbon
To determine if powdered activated carbon was effective for chloroform
precursor removal, coagulated and settled Ohio River water from the pilot
plant was dosed with varying quantities of powdered activated carbon, mixed for
two minutes at 100 rpm and then centrifuged for 20 minutes at 1500 rpm.
The supernatant liquor from the various samples was decanted and chlorinated with
4.5 mg/& of chlorine. The chlorine was mixed rapidly for 2 minutes, then the
samples were stored at 25°C for two days. After this reaction time the samples
were analyzed for trihalomethanes. Under these conditions the chloroform
formation was slightly over 30 yg/£ in the sample which had no powdered
activated carbon added. In the sample with the highest dosage of powdered
activated carbon, 100 mg/Jl, the chloroform production was 11 yg/&. This
extremely high dose of powdered activated carbon only resulted in a 63 percent
reduction in precursor concentration as meausred by the chloroform produced
upon chlorination. Note, because no test exists for chloroform precursor, the
degree of precursor removal is judged by comparing the chloroform concentration
upon chlorination of an untreated control to similar data collected on a
treated water after similar chlorination.
Ozone
The first studies with ozone were to determine its effectiveness for
reducing the concentration of general organic parameters, as chloroform
precursors might be removed under the same conditions. These studies were
conducted in a glass counterflow reactor, approximately 13 feet high. Ozone
was introduced at the bottom of this column through a sintered stainless
steel sparger. The test substrate was Cincinnati, Ohio tap water which
flowed down through the column at various flow rates. At a 1 gpm flow
rate, the contact time in the column was approximately 19 minutes.
-------
- A-29 -
Table XI presents the results from the first series of tests showing
that the removal of carbon chloroform extract -(CCE-m) and NPTOC was related,
in general, to both the applied ozone dose and the amount of ozone utilized
while passing through the column. Organic compounds measured in the
fluorescence test used, however, were not removed in any definitive pattern.
These data show that very high ozone doses are required to effect substantial
removal of organics as measured by various general organic parameters and
indicated the chloroform precursor removal by ozonation might be difficult.
TABLE XI
EFFECT OF OZONATION ON GENERAL ORGANIC PARAMETERS
Applied
Ozone
Dose*
rag/ A
Air Only
Oxygen
Only
8.9
19.2
26.8
38.2
46.7
56.5
71.0
140.0
*Applied
m?
Ozone Contact
Utilized** Time
mg/A Minutes
8.8
18.0
17.7
19.9
39.7
26.5
29.0
62.5
Dose, continuous
OT
19
38
95
19
38
95
19
38
flow
CCE-m
Reduction
%
0
25
33
67
73
60
90
75
83
80
studies, mg/£
standard
Fluorescence
Reduction
%
3
3
11
30
11
25
8
19
29
29
liter of gas (0_
NPTOC
Reduction
%
0
0
18
24
43
-
50
50
-
75
+ 0 „) min.
standard liter of gas (0_ + 02)
minute
liters,water
This parameter may not be directly related to the actual oxidation of organic
compounds because of unaccounted for variations in mass transfer and/or chemical
reaction rates.
**Applied ozone dose minus ozone escaping the top of the contactor.
-------
- A-30 -
Chloroform precursor removal studies were carried out on both dual-media
filter effluents from the pilot plant described above, and on raw Ohio River
water. Table XII presents the results of several different continuous flow-
and batch- studies on chloroform precursor removal showing that whatever
organic compounds are reacting with chlorine to form chloroform can be altered
by ozone so that chloroform production is reduced upon chlorination. In all
cases, however, very high doses of ozone were required to accomplish this, and
the organic by-products created during this massive ozonation have not been
identified at this time.
In anticipation of studies on the use of ozone to treat raw water to
enhance coagulation and thereby ^enhance chloroform precursor removal in the
pilot plant, a larger glass ozone contactor was fabricated to produce a
theoretical detention time of 30 minutes at the pilot plant influent flow rate
of approximately 400 ml/min. To begin to collect some data on the influence
of raw water ozonation on coagulation and sedimentation, and on chloroform
precursor removal, Ohio River water was treated in a jar test with 3 different
doses of alum coagulant. The chloroform formation upon chlorination of these
effluents was compared to a similar test in which the Ohio River water had
been ozonated in the device described above prior to coagulation and sedimentatio
Figure 9 shows that at the lower coagulant' doses, a reduction in eventual
chloroform formation did occur either because coagulation and sedimentation was
enhanced by ozonation or because the ozone affected the organic precursors
in some manner, or both. Because settled water turbidities were not determined
the effectiveness of coagulation was difficult to judge. The slight lowering
of pH might also have had an effect. More studies of a similar nature will be
conducted in an attempt to further understand the influence of raw water
ozonation on the clarification,
-------
- A-31 -
TABLE XII
EFFECT OF OZONATION OF PRECURSOR REMOVAL
CONTINUOUS FLOW-STUDIES
CONTACT TIME -5-6 Minutes
Sample
Dual-Media
Effluent
M ii
ii ii
ii ii
« it
Applied
Ozone
Dose**
mg/1
18.6
0
18.6
0
227*
Chlorine Storage
Dose, Time,
mg/1 Days
0
8
8
8
8
6
6
6
6
6
Bromo- Dibromo-
dichloro- chloro-
Chloroform, methane methane
Ug/& yg/£ Mg/&
<0.1
12
14
91
62
None
9
8
26
7
found None found
2
8.2
6
1
BATCH STUDIES
Applied
Ozone Chlorine Contact Storage
Dose** Dose Time Time
Bromo Dibromo
dichloro- chloro-
Chloroform methane methane
Sample
Dual-Media
Effluent
ii ii
it it
ii ii
Ohio River
Water
ii it
ii ii
mg/1
0
53
106
212
0
56
112
mg/1
8
8
8
8
8
8
8
Hrs.
-
0.5
1.0
2.0
-
0.5
1.0
Days
6
6
6
6
6
6
6
yg/£
42
41
32
19
163
66
40
ml*<
18
4
4
5
18
7
8
ug/£
4
2
None
None
2
< 1
< 0.
found
found
6
Note: Bromoform was not found in any of these samples.
*Mixer to increase ozone - water contact operating.
**Applied Dose, continuous flow studies, mg/A =
mg 03
standard liters of gas (00 + 0~)
standard liter of gas (0 + Q-)
Applied Dose, batch studies, mg/5, =
mg 0-3
minute
min.
litersfwater
standard liter of gas (0» + 0?)
standard liters of gas (0 +_09) contact time,min.
X • J fc X
minute
liquid volume in
reactor, lit^r.s
These two applied doses are not, in general, comparable because the latter
reflects contact time while the former does not. In addition, neither parameter
may be directly related to the actual oxidation of organic compounds because of
unaccounted for variations in mass transfer and/or chemical reaction rates.
-------
- A-32 -
180
OHIO RIVER WATER
2 MIN. RAPID MIX
20 MIN. FLOCCULATION
1 HR. SEDIMENTATION
8 mg/l CHLORINE
6 DAY SAMPLE STORAGE
pH 6.5 - 7.1
20 30 40
ALUM DOSE, mg/l
FIGURE 9. INFLUENCE OF OZONATION ON
CLARIFICATION AND PRECURSOR
REMOVAL
-------
- A-33 -
Finally, an attempt was made to completely eliminate chloroform precursor
by ozonating 14 liters of water for 6 hours at the rate of 43.5 mg of ozone
added per minute. This calculates to an "applied ozone dose" over the 6-hour
period of over 1,100 yg/&, which is admittedly impractical. After this intense
ozonation, raw Ohio River water only produced 5 yg/£ of chloroform when
chlorinated with a chlorine dose of 8 mg/£ and the sample stored at 25°C
for 6 days. Therefore, chloroform precursor can be virtually eliminated from
raw water by ozonation, but not under conditions that would make this method
practical for water utilities.
Chlorine Dioxide
As indicated previously in this Guide, chlorine dioxide appears to reduce
chloroform precursor. The reader is referred to Table X for the data supporting
this conclusion. Further support of this suggestion was obtained when Ohio
River water that had been treated with chlorine dioxide such that all of the
chlorine dioxide had been consumed was subsequently chlorinated. Depending
on dose and contact time only 20 to 50 percent of the chloroform was produced
when chlorine dioxide treatment preceeded chlorination as compared to
chlorination only.
Granular Activated Carbon
Chloroform Precursor Removal
To determine the effectiveness of adsorption on granular activated carbon
for the removal of chloroform precursor, two granular activated carbon columns
were arranged to operate in parallel with the dual-media granular filter
described previously as part of the pilot plant. The characteristics of the
granular activated carbon used are given in Table IV. The coal-based granular
activated carbon was 30-inches deep with 4-5 minutes apparent contact time,
and the lignite-based granular activated carbon was 60-inches deep, with 8-10
minutes apparent contact time. Both granular activated carbon columns received
settled water containing some carryover floe. Previous experimentation had
indicated that filtering settled water through a dual-media filter prior to
applying it to the granular activated carbon columns did not improve the
performance of the granular activated carbon.
After fresh granular activated carbon was placed in both columns, a
sample was collected each week from the effluent of each of the three
columns. The dual-media filter effluent was the control for the study. These
samples were tested for NPTOC concentration, and were chlorinated with 2-3 mg/&
of chlorine, then stored for 4 days at 25°C to simulate passage through
-------
- A-34 -
a distribution system.
The effluent data from both granular activated carbon columns, Figure 10,
shows that the breakthrough of chloroform precursors was sooner in the effluent
of the shorter coal-base column, but the shorter coal-base granular activated
carbon was somewhat more effective for chloroform precursor removal than the deep
lignite-base material early in the test (chloroform concentration of <1 to 2 yg/
vs. 2-3 yg/£).
The passage of chloroform precursor was studied at two water treatment plant
using granular activated carbon as a combination adsorption/filtration media.
At one utility, four weeks after the first granular activated carbon had been
installed, chlorination of the filter effluent increased the chloroform concentra
1.6 times. Three weeks later chlorination of the filter effluent increased the
chloroform concentration 2.5 times. At another utility, when the granular activa
carbon was 10 weeks old, chlorination of the filter effluent increased the
chloroform concentration 3.4 times. These data show the breakthrough of
chloroform precursor and confirm the pilot plant data reported above.
Removal of Precursors for Other Trihalomethanes
To determine the effectiveness of adsorption on granular activated carbon fo
the prevention of bromine-containing trihalomethanes and to determine if NPTOC
measurements would be an adequate surrogate for precursor breakthrough under thes
circumstances, the total trihalomethane data collected from the granular activate
carbon was compared to NPTOC concentrations each week. To calculate the total
trihalomethane concentration in each sample the individual weight concentrations
were converted to micromole concentrations by dividing by the appropriate molecul
weight and arithmetically summing the values. Figure 11 shows that the breakthro
of total trihalomethane precursor and NPTOC concentration is roughly parallel,
occurring after about 4 weeks for the 30-inch coal-base granular activated carbon
after about 5-6 weeks for the 60-inch lignite-base granular activated carbon.
-------
- A-35 -
\
z
o
LU
a.
z
o
00
eg
o
LLJ
-
ID *
Z eg
< O
eg co
0 eg
- 2 O ^
_
8°
CO ft
2 °
< _l
u_ X
O <-»
LLI Cg
CO
eg
^
O
|/6jy'NOIlViliN3DNOD
-------
-- A-36 -
I/Bui 'DOldN
t 3
> <
* at
o O
O
< O)
m E
CO
6
cs
ci
O
I
_
Z
O
6
Z
o
CO
*
<
u
|/|oui/r'3NVH13WO1VHiai IViOi
u
CO
Z
o
•o
o.
o
a: .
O •
o «±-
-o
o
co
cs
o
•—. O CO CM "~. O
0 6 d °
|/|oiurf'gNVHi3WO1VHiai 1V1O1
-------
- A-37 -
The difference in chloroform precursor breakthrough (Figure 10) and total
trihalomethane precursor breakthrough (Figure 11) is caused by the bromine-
containing trihalomethanes. For reasons not understood at this time,the
concentration of the bromine-containing trihalomethanes was reduced less than
chloroform when comparing the chlorinated granular activated carbon effluents
to the chlorinated effluent from the dual-media filter.
Figure 11 also shows that the deeper lignite-base granular activated
carbon column produced water with an NPTOC concentration below the detection
level of analytic method (0.1 mg/£)for 6 weeks before starting to rise.
Although this column was twice as deep as the coal-base column, this activated
carbon does have a smaller surface area as measured by the nitrogen gas BET
method (Table IV) and this may be the reason it did not produce very low NPTOC
concentration water for twice as long as the coal-base granular activated carbon.
In summary, in spite of some uncertainties, Figures 10 and 11 show that
during the first month of operation both granular activated carbon columns
produced an effluent with very low concentrations of NPTOC, chloroform, and
total trihalomethane.
-------
- A-38 -
FUTURE RESEARCH
The following list of topics are areas in which further research is
needed to respond to some of the unanswered questions raised earlier in this
Guide.
1. Field trials of the alternate treatment processes discussed in this
Guide, including an investigation of the microbiological quality of the
distributed water.
2. Studies on higher surface area granular activated carbon and other
adsorbents.
3. Investigation of moving, deep-bed upflow granular activated carbon
adsorbers as a method for maintaining good performance for a longer period
of time, thereby reducing operating cost.
A. Determine the loss in capacity of granular activated carbon for
organic removal (particularly chloroform precursor) following reactivation.
These studies will also include an evaluation of other types of reactivation
beyond the multiple-hearth furnace.
5. Investigation of techniques for removing or preventing the formation
of bromine-containing trihalomethanes.
6. Evaluation of the use of potassium permanganate and other oxidants
for chloroform precursor removal.
7. Improving coagulation and settling to increase trihalomethane
precursor removal.
8. Determination of non-trihalomethane by-products of disinfection
with chlorine, ozone, chlorine dioxide, ultra-violet radiation, and combined
chlorine.
9. Detailed organic analyses of the effluent from fresh granular
activated carbon beds.
-------
- A-39 -
10. Development of a low-concentration total organic carbon analyzer.
11. Determination of the nature of chloroform precursor so that it
could be measured directly and problems thereby anticipated.
12. Development of a total organic halogen analyzer.
13. Further investigation of treatment unit process costs.
14. Toxicological evaluation of fresh granular activated carbon
adsorber effluent to establish the relationship between low NPTOC concentrations
and the absence of adverse health effects.
15. Toxicological evaluation of organic non-trihalomethane by-products
formed during chlorination, ozonation and the use of chlorine dioxide in
water. Toxicity of chlorite must also be investigated.
16. Evaluation of ozone, chlorine dioxide, ultra-violet radiation,
and combined chlorine as a viricide.
U.S. Environmental Protection Agency
Region V, Library
230 South Dearborn Street
Chicago, Illinois 606041 ^
-------
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
SUBJECT: Future Plans for Control of Chloroform
FROM:
TO:
DATE May 11, 1976
Gordon G. Robeck, Coordinator, EPA Drinking Water
Research Activities
Recipients of Interim Treatment Guide
Attached is a copy of the Fifth Draft of the Interim Guide
as it was given to the Public Advisory Council on April 22, 1976.
This draft is currently under revision to reorganize it somewhat,
update the information slightly, and to provide a detailed Appendix
on cost and a more complete Appendix on Current Knowledge. This
revision should be available in July and will be a more polished
reference document as opposed to the working document now available.
Although we are collecting more information all the time, I see
no reason not to start applying what we know already.
We would appreciate it if you would keep us informed of any
data gathered so they can be used along with the field data we will
be collecting this summer to help establish a final treatment
recommendation if the Agency decides to regulate chloroform levels.
P
C i V '£
2r' 1976
r-.
fOJRMWW ~
mum,
------- |