-------
to groundskeeping or general labor might be utilized in the reactivation
activity. Such a joint use of labor might realize genuine savings, particu-
larly in a small plant. Table 20 displays the percentage of total costs
for plants with on-site reactivation which is made up of labor costs. It can
be seen that for small plants with on-site reactivation, labor costs account
for over 40 percent of the total cost.
SUMMARY AND CONCLUSIONS
It is obvious from the data presented in this report that chlorination
is the cheapest of all of the treatment technologies that might be used for
disinfection. Table 21 summarizes the values for a 1, 5, 10, 100, and 150
mgd plant for all of the treatment alternatives examined in this report.
As chlorination under certain conditions causes chloroform, a potential
carcinogen, in drinking water, planning and operating agencies must examine
alternatives to the chlorination process. These alternatives might take the
form of disinfection techniques other than chlorination, or of trihalomethane
removel techniques such as aeration, or of organic removal techniques such
as granular activated carbon. Hopefully, this report will assist in making
these evaluations.
92
-------
TABLE 20. LABOR COSTS FOR 1, 10, and 100 mgd GAG SYSTEMS REACTIVATING
ON-SITE (FILTER SHELL REPLACEMENT)
Plant
Capacity
5
10
100
Capacity
Factor
.7
.7
.7
Total Cost
$/yr
199,915.97
302,103.2
2,098,677.0
Labor Cost
$/yr
91,385
124,724
421,756
Percent
Labor Cost
46
41
20
93
-------
/- \
r-t
tO
00
o
o
o
rH
^v.
S^
CO
H
cn
en
o
2
S
<3
2
0
M
erf
a
0
m
rH
00
e
O
o
rH
'd
00
a
o
rH
1o
B
m
*O
00
a
rH
0)
4-1
CO
en
vo co oo oo in rH
o rH o o oo m
l-» CO 00 O\ O O
O i-H O O CT\ vD
CM r^ vo co oo r*»
• • • • • •
rH rH rH rH CM rH
rH rH
>*o co co in *«o in
• ••••• •
rH CM CM CM 0
•H tO
4J 4J
0 C
< 0
o
ll \^^
to
rH C
3 0
G jQ
tO P
P tO
o o
94
-------
REFERENCES
1. Breidenbach, Andrew W., "Regulations: Reactions and Resolutions,"
Journal of the American Water Works Association, Vol. 68, No. 2,
February 1976, pp. 77-82.
2. Bureau of Labor Statistics, "Chapter 11. Wholesale Prices," reprint
from the BLS Handbook of Methods (BLS Bulletin 1711), U. S. Department
of Labor, pp. 97-111.
3. Eilers, Richard G., and Smith, Robert, "Executive Digital Computer
Program for Preliminary Design of Wastewater Treatment Systems,"
November 1970, NTIS-PB222765 (report NTIS-PB222764 (card deck).
4. Fair, Gordon Maskew, and Geyer, John Charles, "Elements of Water Supply
and Waste Water Disposal," John Wiley & Sons, Inc., New York, pp. 480-481,
5. Federal Water Pollution Control Administration, "Sewer and Sewage
Treatment Plant Construction Cost Index," U. S. Department of the
Interior, Washington, D. C. 20242.
6. Finerty, Joseph M. (Editor), Employment and Earnings, April 1976,
U. S. Department of Labor, Bureau of Labor Statistics, Vol. 22, No. 10.
7. Love, 0. T., et al., "Treatment for the Prevention or Removal of
Chlorinated Organics in Drinking Water," submitted for publication to
the Journal of the American Water Works Association.
8. Miltner, R. J., "The Effect of Chlorine Dioxide on Trihalomethane in
Drinking Water," Master of Science Thesis, University of Cincinnati,
1976.
9. Patterson, W. L., and Banker, R. F., "Estimating Costs and Manpower
Requirements for Conventional Wastewater Treatment Facilities for the
Environmental Protection Agency," Black and Veatch, Consulting Engineers,
Kansas City, Missouri, 1971. .
10. Quarles, John R., Jr., "Impact of the Safe Drinking Water Act,"
Journal of the American Water Works Association, Vol. 68, No. 2,
February 1976, pp. 69-70.
11. Suindell-Dressler, "Process Design Manual for Carbon Adsorption,"
U. S. Environmental Protection Agency, Technology Transfer, October 1973.
95
-------
12. Symons, James M., "Interim Treatment Guide for the Control of Chloroform
and Other Trihalomethanes," June 1976, Water Supply Research Division,
Municipal Environmental Research Laboratory, Office of Research and
Development, Cincinnati, Ohio 45268, pp. 4-6.
13. Ibid, pp. 1-4.
14. Ibid, pp. 6-30.
96
-------
APPENDIX 2
Stevens, A.A., Slocum, C.J., Seeger, D.R. and Robeck, G.G., "Chlorination
of Organics in Drinking Water," Proceedings of Conference on the Environmental
Impact of Water Chlorination, Oak Ridge, Tennessee, October 22-24, 1975,
and submitted to the Journal of the American Water Works Association
for publication.
-------
CHLORINATION OF ORGANICS IN DRINKING WATER
Reprinted from: Journal American Water Works Association, 68:11, p. 615-620
(November 1976).
Water Supply Research Division
Municipal Environmental Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
Chlorination of
Organics in
Drinking Water
Alan A. Stevens,
Clois J. Slocum,
Dennis R. Seeger, and
Gordon G. Robeck
A paper- contributed to and selected by the
JOURNAL, authored by Alan A Stevens
(Active Member, AWWA), res chem , Clois J
Slocum, res, chem , Dennis R. Seeger, res
chem , and Gorden G Robeck (Honorary
Member, AWWA), dir., all of the Water Supply
Res Div , EPA, Cincinnati, Ohio
Bench- and pilot-scale investigations
revealed the influence of precursor
compound concentration, pH, type of
disinfectant, and temperature on triha-
lomethane formation. Implications of
the research for altering treatment
procedures to reduce trihalomethane
production are discussed.
Recently there has been great interest in
the stud) of organic compounds in drinking
water—interest that stems largely from the
results of a 1974 stud} of New Orleans
drinking water, and the publicity that
followed.1 About the same time, two
studiesj ; called attention to the presence in
finished drinking water of some tnhalo-
methanes (mostly chloroform) which were
not found in the respective raw waters at the
locations of stud) Both reports concluded
that the trihalomethanes were formed
during the chlonnation step of the water
treatment process.
The EPA undertook a survey of 80
selected cities to measure the concentrations
of six halogenated compounds in raw and
finished water Those six included four
trihalomethanes (chloroform, bromodichlo-
romethane. dibromochloromethane. bro-
moform) suspected of being formed during
chlonnation. plus carbon tetrachlonde and
1,2-dichloroethane, known contaminants at
New Orleans, but not necessarily formed on
chlonnation. During this National Organics
Reconnaissance Survey (NORS) a more
comprehensive organic analysis was also
performed in five of the 80 cities and has
just been completed in another five
The occurrence of trihalomethanes in
finished drinking water was demonstrated
to be widespread and a direct result of the
*Dehvered at th< \Tonf on the Environmental Impact of Water
Chlormation Oak Ridge Nat! Lab Oak Ridge Tenn (Oct 22-
24, 1975)
NOVEMBER 1976
chlormation practice. No hard evidence was
found in this regard with respect to 1.2-
dichloroethane or carbon tetrachlonde.
Based on the survey results, a theoretical
finished water with the median concentra-
tion of each compound would contain
about 21 jug/I of chloroform. 6 jtg/1 of
bromodichloromethane. 1.2 ;ug/l of dibro-
mochloromethane. and an amount less than
the detection limit for the method used1 of
bromoform (Fig 1) Although most of the
finished waters tested demonstrated this
decreasing order of concentration, this was
not always the case The finished water at
one location had a chloroform concentra-
tion of only 12 jug/1, but a bromoform
concentration of 92 jug/1 It was speculated
that this concentration reflected a relatively
high bromide concentration in the raw
water, with oxidation of bromide to hypo-
bromite by hypochlorite and subsequent
reaction of hvpobromite with precursor
compounds to form the bromine-substi-
tuted trihalomethanes
Recently workers at another EPA labora-
tory ' have adequately demonstrated this
effect by experimentally adding the hahdes
fluoride, bromide, and iodide in the form of
salts to Missouri River water and subse-
quently chlorinating that water The
detected reaction products included all ten
possible non-fluorine mixed and single
halogen-containing trihalomethanes. Final-
ly, the range of chloroform concentrations
was < 01-311 fig/1; bromodichlorometh-
ane. none found (NF)-116 jug/1; dibromo-
chloromethane, NF-100 jug/1, and bromo-
form, NF-92 jug/1
Although the health significance of triha-
lomethanes produced during chlonnation
of drinking water had not been completely
evaluated in 1975. understanding the
factors affecting the ultimate formation of
the trihalomethanes .was considered pru-
dent. The goal was then to develop general
conclusions applicable to rational modifica-
tion of water treatment processes if removal
of trihalomethanes was finally deemed
important for public health reasons. Basic
approaches to affect finished water triha-
lomethane concentrations considered for
study were reducing precursor com-
pound(s) concentration, changing disinfec-
tant (eg. to ozone, chlorine dioxide, etc.)
and removing trihalomethanes after forma-
tion. The last of these is being studied as an
alternative and has been discussed else-
where." Changing disinfectant without an
intense research input from studies of other
public health ramifications could be a
catastrophic step Therefore, because a
chlorine residual must be maintained with-
in the distribution system, removing precur-
sor compounds or controlling their reac-
tions with chlorine was considered the most
logical approach.
The foremost consideration in adjusting a
series of water-treatment processes to
remove an organic precursor is identifying
the compound(s) Bellar et al proposed
ethanol as the compound with oxidation by
hypochlorite to acetaldehyde. or acetalde-
hyde itself, followed by the classical halo-
form reaction as the mechanism of triha-
lomethane production.' Organic chemistry
texts typically cite acetone as the simplest
example of a methyl ketone that undergoes
the haloform reaction. Indeed, Fairless et al
have investigated the reactivity of simple
methyl ketones in water supplies and
consider them to play a major role in
trihalomethane production.7 Glaze and
Henderson have identified chlorinated ace-
tone derivatives that could be haloform
reaction intermediates in super-chlorinated
sewage effluents v These theories are attrac-
C
live because the precursor compounds
mentioned have been qualitatively iden-
tified during gas chromatographic-mass
spectrometnc (GCMS) analysis of Ohio
River water that contains the unknown
precursors that react to form trihalometh-
anes upon chlonnation.
In Dec. 1974 Rook proposed that natural
humic substances were responsible.-' Later
he discussed the probable role of the fulvic
acid fraction in trihalomethane production,
elaborating the thesis with examples of very
reactive w-dihydroxy aromatic compounds
suspected to be basic building blocks of the
humic (fulvic) acid structure."
A more recent article by Rook empha-
sizes after-coagulation treatment for remov-
al of trihalomethanes or their organic
precursors.1" However, a clarification of the
relative roles played by the two groups of
precursor compounds (humic materials vs.
acetyl derivatives of low molecular weight)
with inclusion of a consideration of the role
of pH will help to predict the success of
relatively simple water-treatment process
changes (such as optimizing existing coagu-
lation and sedimentation processes for
precursor removal or changing the ongoing
chlormation practice) designed to bring
about a reduction in the ultimate trihalo-
methane concentrations. The roles of other
treatment parameters such as NH, addition
with chlorine (free vs. combined chlorine)
and temperature should also be clarified.
Methods
Reagents. Chlorine was obtained in a
high purity grade.* Stock solutions were
prepared by passing the pure gas through
nitrogen-purged distilled water. Freshly
prepared stock solutions were standardized
by amperometric titration as described in
Standard Methods." Experimental mixtures
were prepared by appropriate volumetric
dilution of the stock solutions in the test
media.
Water for the various experiments was
obtained from the EPA's Municipal Envir.
Res. Lab. (MERL), Water Supply Res. Div
(WSRD) pilot plant facility at Cincinnati.
Ohio This plant has previously been
described in detail." Raw water was
obtained directly from the Ohio River
intake at the Cincinnati Water Treatment
*From Union Carbide, Ohio Valley Sales Cincinnati Ohio
A A STEVENS ET AL 615
-------
Plant. This water was used as an untreated
source water in all pilot-plant work. Settled
water was that obtained from the pilot plant
after alum coagulation and sedimentation
Dual-media filtered water was the settled
water after anthracite-sand filtration. Acti-
vated carbon filtered water was the same
settled water after passage through 1 5 m of
granular activated carbon (GAC).* Filtra-
tion rates through this plant were similar to
those found in a conventional water treat-
ment plant: 2-25 gpm/sq ft (5-6.25 m/hr)
Blank water for analytical purposes was
obtained by purging laboratory distilled
water exhaustively with helium gas.
The test precursor substances (humic
acid.f acetone,§ acetaldehyde ** and aceto-
phenoneft) were used as obtained from the
suppliers
Standard analytical solutions of chloro-
form, JJ bromodichloromethane,** dibro-
mochloromethane,§§ and bromoformtt
were prepared as described in the NORS
80-city report.'
Procedures. Analyses for the trihalo-
methanes were performed by a modifica-
tion of the volatile organic gas chromato-
graphic technique described by Bellar and
Lichtenberg1-' using specific halogen elec-
trolytic conductivity detection" as de-
scribed in the NORS 80-city report.'
Nonvolatile total organic carbon
(NVTOC) was measured using the method
and apparatus described in the NORS
80-city report ' Samples were acidified with
nitric acid and purged with carbon-free air
for about 10 min to remove carbon dioxide
before the actual analysis. Some volatile
organic materials were lost during this step.
NVTOC was defined as that organic carbon
remaining in the sample after this treat-
ment.
Briefly, the experimental procedure was
as follows: All reactions described were
carried out in the presence of phosphate
buffers. Reaction solutions were made up at
pH 7 and adjusted to the desired pH by the
addition of either hydrochloric acid or
sodium hydroxide. Reaction mixtures were
prepared with the appropriate source
water; buffer was added, and pH was
adjusted. The mixtures were then spiked
with the test compounds, and chlorine stock
solution was added. The reaction mixtures
were typically 1-2 1. Immediately after
mixing, zero time samples were taken by
pouring from the larger vessel into a 50-ml
serum vial containing an appropriate quan-
tity of 0.1 A' sodium thiosulfateft to halt the
reaction by removing chlorine. Samples for
storage (extended reaction time) were taken
in a similar manner without sodium thiosul-
fate. All vial's were sealed headspace-free
with TFE-faced septa immediately after
*Filtrasorb 200, from Calgon Corp Pittsburgh Pa
TFrom Pfdltz and Bauer, Flushing, N Y , or Aldntch Chemical
Co Milwaukee, Wise
^Nanograde, From Allmcrodt St Louis Mo
**Also from Aldntch Chemical Co
ttFrom Fisher Scientific, Pairlawn, N J
JiSpectroanalyzed by Fisher Scientific
§JtColumbia Chemical Co, Columbia, SC
filling as described in the NORS 80-city
reportA The sealed samples were stored at
the indicated temperature in either a water
bath or incubator controlled at ± 0.5C. At
the appropriate time, the vials were opened,
and aliquots were quickly transferred to a
30-ml vial containing sodium thiosulfate.
The smaller vial (headspace free) was then
sealed as described above. All preserved
samples were then stored under refrigera-
tion until analysis.
Results and Discussion
Precursor at pH 7. General. Trihalometh-
anes must result from a reaction or series of
reactions of chlorine with a precursor mate-
rial Simple methyl ketones react through
the classical haloform reaction mechanism
More complex substances, such as humic
materials, also react by this mechanism or
by some other mechanism that includes an
oxidative cleavage step. Because control of
trihalomethane production by precursor
removal or control of precursor reaction
rate was considered the best approach.
some knowledge of precursor identity was
required. Suggestions, as mentioned above,
as to identity of precursor varied from
complex humic materials to simple methyl
ketones or simple compounds with the
acetyl moiety.
This laboratory's earliest work with
precursor removal was simply an experi-
ment to determine whether GAC adsorp-
tion had any effect on precursor concentra-
tion In this work, samples of water taken
from the pilot plant were chlorinated at a
dose of 8 mg/1-that used at that time by the
Cincinnati Water Treatment Plant on the
same raw water to satisfy chlorine demand
and maintain a free residual in the distribu-
tion system In this experiment, not only
were settled and activated carbon filtered
water samples chlorinated to determine the
effect of the carbon, but dual-media filtered
and raw-water samples were also chlori-
nated at the same concentration for
comparison All four samples were buffered
at pH 7 The results in Fig. 2 show that
when the result of chlorination of fresh
GAC-filtered water was compared with the
result of chlorinating the settled water.
removal of precursor was indicated The
effectiveness of GAC filtration, however,
was shown later to be relatively short-
lived—a matter of only a few weeks under
conditions of pilot plant operation.1 The
other important aspect of this experiment
was the dramatic change in the rate of
chloroform formation when the results of
raw and settled water chlorination were
compared Conventional alum coagulation
and sedimentation caused the removal of
most of the precursor material from the raw-
water.
Paniculaies. The above experimental
results indicated that precursors are one or
more of the following some sort of particu-
late, a substance associated with particu-
lates. a substance reacting in association
with the particulates. or possibly a
substance that could be complexed with the
alum and precipitated with the floe. The
nature of the role of the particulates was
therefore further investigated. A simple
vacuum filtration of raw water through
Whatman No I filter paper was carried
out The filtrate, particulates trapped by the
filter (including filter paper) resuspended in
GAC filtered water, original raw water, and
GAC filter effluent with and without clean
filter paper were each chlorinated and
subsequently analyzed for trihalomethane
content after varying periods of storage.
Comparison of the reaction rate curves
for raw and filtered raw water shown in Fig.
3 illustrates a reduction of the rate of
trihalomethane production caused b} re-
moval of particulates. The rate curve for
GAC filter effluent with resuspended filter
paper and particulates from the raw water
indicates that essentially all of the differ-
ence between the raw and filtered raw water
rate curves can be accounted for by the
substances trapped on the resuspended
filter paper. The curves for GAC filter
effluent and GAC filter effluent plus filter
paper are simplv the appropriate controls
and are nearlv identical They indicate
essentially no reaction interference or
enhancement by the filter paper itself
According to these results, simple filtration
either removed some trihalomethane pre-
cursor from the raw water or the removal of
some of the paniculate matter reduced the
reaction rate of dissolved precursor The
paniculate matter, therefore, played some
direct role in trihalomethane production
when Ohio River water was chlorinated
To determine which of these mechanisms
was important, the effect of potentially
active surfaces was investigated by spiking
two sets of GAC-filtered water samples with
simple acetyl derivatives and then suspend-
ing Bentomte clay in one set and powdered
activated carbon in the other set Neither of
the two added particulates caused any
detectable increase in rate of tiihalometh-
ane formation. Therefore active surface
effects were not considered significant but
paniculate matter or substances strongly
sorbed on the paniculate matter were found
to be important precursors of trihalometh-
ane production at pH 7.
Humic acid Because humic substances
are more likely to be found in natural
waters as small particulates or sorbed on
clay particles" than are soluble simple
methyl ketones, a direct test of Rook's
hypothesis-1 was attempted using commer-
cially available humic acid, both suspended
at pH 7 and dissolved at higher pH. which
was later readjusted to pH 7. At concentra-
tions of humic acid representing an
NVTOC concentration similar to that
found for Ohio River water (approximately
3 mg/1 of NVTOC). the rate curve fo'r
formation of trihalomethanes was observed
to be very similar to that seen for chlorina-
tion of the natural water (Fig. 4). In addi-
tion, a filtration experiment (0.2 /xm pore
filter) similar to that earned out on the raw
616 WATER TECHNOLOGY/QUALITY
JOURNAL AWWA
-------
romodichloromethan
Per Cent Equal to of Less
Than Given Concentration
Fig. 1 Frequency Distribution of Trihalometh-
ane Data
50
I
25 -
Settled
Fresh GAC Filtered
30 60
Storage Time — hr
90
-J
120
Fig. 2 Effect of Treatments on Chloroform
Production—8 mg/l Chlorine Dose, 25C,
pH 7
Storage Time—hr
Fig. 3 Effect of Simple Filtration on Trihalo-
methane Production—Chlorine Dose 10 mg/l,
25C, pH 7
Fig. 7 Effect of pH on Chloroform Production,
Settled Water, 25C, 10 mg/l Chlorine Dose
Fig. 4 Comparison of Humic Acid, Raw Water
Reaction Rates at Similar NVTOC Concentra-
tions—10 mg/l Chlorine Dose
Dissolved at pH 11,
Readiusted, Filtered
(0 2 i»m), Chlorinated
n —
uspended, Filtered (0 2 Km). Chlorinated
10 20 30 40 50 60 70
Reaction Time—nr
Fig. 5 Filtered and Unfiltered 5 mg/l Humic
Acid Mixtures—pH 7, 10 mg/l Chlorine Dose
Fig. 8 Effect of Humic Acid Concentration on
Trihalomethane Production—pH 67, 25C, 10
mg/l Chlorine Dose
10 20 30 40 50 60 70
Reaction Time—hr
Fig. 6 Raw and Carbon Filtered Water Spiked
at 5 /xM With Low Molecular Weight Acetyl
Compounds—10 mg/l Chlorine Dose, pH 7
0 10 20 30 40 50 60 70
Fig. 9 Effect of pH on Trihalomethane
Production From 1 mg/l Humic Acid—25C,
Chlorine Dose 10 mg/l
NOVEMBER 1976
A A STEVENS ET AL 617
-------
Fig. 10 Effect of pH on Trihalomethane
Production From 1 mg/l Acetone—25C, Chlo-
rine Dose 10 mg/l
Fig. 11 Chloroform Production at Three
Temperatures of Raw Water—pH 7, Chlorine
Dose 10 mg/l
0 10 20 30 40 50 60 70 80
Reaction Time—hr
Fig. 12 Free vs Combined Chlorine and TTHM
Production With and Without NH, Addition-
pH 7
water described above was conducted on
suspensions and solutions of humic acid.
The results (Fig. 5) were similar to those
reported m Fig. 3. Thus, in terms of rate of
tnhalomethane formation on chlonnation,
the physical and chemical characteristics of
humic acid in suspension and solution at
these concentrations were found to be
similar at pH 7 to those of the unknown
precursor substances present in the Ohio
River.
Finally, attempts to react chlorine at pH 7
with simple acetyl compounds (acetone,
acetaldehyde, and acetophenone), when
these compounds were spiked at 5 /xmol/l
into raw- and GAC-filtered water, failed to
produce tnhalomethanes at rates signifi-
cantl) higher than those observed for the
blank samples (Fig. 6). Therefore, for chlo-
nnation of natural waters at pH values near
7, the humic acid precursor hypothesis of
Rook seemed the most valid.
Effect of pH on reaction rate and precursor
identity. General. Because the rate-deter-
mining step of the classical haloform reac-
tion is enohzation of a ketone. the rate of
tnhalomethane formation is pH dependent.
For example, the reaction of acetone with
hypochlorite to form chloroform proceeds
at a faster rate at pH 11.5 than at pH 6.5.
Experimentally, a sample of settled water
was buffered at pH 6.5 and another at pH
11.5; both were chlorinated at an initial
concentration of 10 mg/l The results (Fig.
7) show that the rate of formation of chlo-
roform increases with an increase in pH.
This could be explained simply by an
increase in the humic acid reaction rate, as
would be expected by the classical mecha-
nism Another possibility, however, is thato-
other compounds in the source water (set-
tled), such as acetone, that do not react
readily at pH 6.5. become significant con-
tributors to the overall reaction rate (chloro-
form formation) at pH 11 5. An indication
of the latter possibility was previously noted
in the work of Fairless et al—acetone was
shown to react at a significant rate at pH
9.5, but not at a pH near 7.7 Because
chlonnation is carried out at high pH in
some water supplies, especially where lime
softening or excess lime softening is prac-
ticed, further investigation of the effect of
pH was necessary.
Humic Acid Figure 8 illustrates the reac-
tion rate curves for formation of total triha-
lomethanes (TTHM) from three concentra-
tions of humic acid (0.1, 0.5, 1.0 mg/l)
spiked in GAC-filtered water in presence of
excess chlorine (10 mg/l with less than 10
per cent change during the course of the
experiment). An apparent first order rate
dependence on initial humic acid con-
centration is demonstrated; that is, at any
given time between any two curves, the
ratios of concentrations of TTHM produced
are equal to the respective ratios of initial
humic acid concentrations. The change in
rates with apparent exhaustion of reaction
sites can also be seen as nearly constant
TTHM concentrations are approached.
618 WATER TECHNOLOGY/QUALITY
In Fig. 9 the pH dependency of reaction
rate at one of these concentrations (1 mg/l)
is illustrated. The same curve characteristics
were observed at all pH values As noted
above, one can assume from the shape of
the curves that the reaction was nearly
complete at pH 6.7 or was proceeding very
slowly relative to the initial rate. Because
the reaction is essentially complete at pH
6.7 at the end of the experiment, the nearly
two-fold increase in final product concen-
tration at pH 9.2 can only be explained by
the presence of certain reactive sites on the
complex humic acid molecule that react at
insignificant rates at the lower pH. but are
reactive at higher pH. The concentration of
significant reactive sites in the reaction
mixture, when expressed as equivalents per
liter, is therefore at least twice as high at the
higher pH. Based on this analysis, and
considering humic acid to be 60 per cent
carbon, 0.7 per cent and 1.4 per cent of the
carbon present reacts ultimately to become
tnhalomethane at the low and high pH
values respectively.
Acetone. Reactions of acetone with chlo-
rine can be compared quantitatively with
those of humic acid in an evaluation of the
potential role of acetone as a precursor
because the similarity of the humic acid
reaction to that of the natural material in
the source water has already been demon-
strated (see Fig. 4). Figure 10 shows the pH
dependency of the rate of reaction of 1 mg/l
acetone. At pH 6.7 the TTHM concentra-
tion from acetone after 96 hr is about one
third of that observed from 1 mg/l humic
acid in the same 96-hr period (see middle
curve, Fig. 9). These numbers might seem
to indicate that acetone could be a signifi-
cant precursor at pH 6.7. Because the rate of
trihalomethane production from acetone
through the classical haloform reaction
mechanism is known to be proportional to
acetone concentration, however, 3 mg/l of
acetone would be required to give the same
TTHM concentration at 96 hr as would 1
mg/l of humic acid. Therefore, approxi-
mately 15 mg/l of acetone would be
required to give the concentration of chlo-
roform observed from the raw water (Fig.
4). Thus, if acetone were the important
precursor at pH 6.7, sufficient acetone
would be required in solution to account for
over 9 mg/l of NVTOC, which far exceeds
the 2 to 3 mg/l NVTOC usually found in
the source water (acetone is not easily lost in
the CO, stripping during NVTOC sample
preparation).
Furthermore, the reaction rate curve for
acetone at pH 6.7 is nearly a straight line
which indicates no change in rate during
the experiment. By again using the
assumption that acetone reacts by the clas-
sical haloform reaction mechanism and
from the final trihalomethane concentra-
tion observed, less than 1 per cent of the
acetone initially present was calculated to
have reacted. Because this change of
acetone concentration was insignificant, its
effect on reaction rate was not observed in
JOURNAL AWWA
-------
The occurrence of trihalomethanes in fin-
ished drinking water was demonstrated to be
widespread and a direct result of the chlonna-
tion practice Vintage installations, such as
the one shown, are still serving reliably
this experiment. An insignificant change
was expected, based on calculations using a
reported rate expression for acetone m the
haloform reaction.'* Therefore, if acetone
was the most important precursor and if its
concentration was high enough to account
for the observed rate of tnhalomethane
production from the source water, the char-
acteristic rate curve would be linear as
plotted. For these two reasons acetone is not
likely to be a significant precursor at pH
6.7.
At pH values much higher than 6.7,
however, the situation could be different
Figure 10 has been plotted on the same
numerical scale as Fig. 9, so that a direct
comparison of reaction rates between
acetone and humic acid at the various pH
values is possible A comparison of the
curves on these figures, representing the
trihalomethane formation rates at the
higher pH values, illustrates a much larger
increase in reaction rate of acetone with
changing pH than that observed with the
same concentration of humic acid. The 30-
fold observed increase (graphically measur-
ed) in acetone reaction rate was also
expected from calculations based on the
reported rate expression.1' A rate increase
of this magnitude could allow as httle as 500
,ug/l (15 mg/1/30) of acetone to account for
the trihalomethanes formed on raw water
chlonnation of pH 10.2. Therefore, low
molecular-weight compounds containing
the acetyl moiety that have haloform reac-
tion rates similar to that of acetone can
become significant contributors to total
trihalomethane production when chlorina-
tion is earned out at high pH Thus, both
possible explanations for the effect of pH
on reaction rate noted in the discussion of
Fig. 7 are valid.
The question of precursor identity is,
therefore, complicated because "precursor"
is actually a mixture of compounds with
differing reactivities at varying pH values,
solubilities, and other physical and chem-
ical characteristics. The relative contribu-
tions of the various constituents of a given
water depend somewhat on the treatment
practiced as well as on the source of the
water. The probable diverse nature of
precursor also may hamper efforts to find a
single general organic parameter for unit
process control that will predict effective
removal of precursor.
Temperature. The effect of temperature
on the rate of reaction of precursors present
in Ohio River water was investigated to
assess the potential effect of wide seasonal
temperature variations in raw and treated
waters. The wmter-to-summer water temp-
erature variation in Cincinnati. Ohio, raw
and finished water is approximately 26C
(from < 2C> 28C). The results presented
in Fig. 11 show that this temperature differ-
ential could easily account for most of the
winter-to-summer variation in chloroform
concentration (< 30 fig/1 to > 200 fig/i)
observed in Cincinnati tap water over the
past year when raw water chlonnation with
a three to four day chlorine contact time
was practiced. Some other factors, such as
seasonal variation in precursor concentra-
tion, certainly have some additional effect,
however
Disinfectant. Work is progressing with
measurement of the effects of the use of
oxjdants other than chlorine as disinfectants
(O,, C1OJ on trihalomethane production.
When completed, the results of these exper-
iments will be the subjects of future reports.
The work reported herein was confined to a
study on the effect of chlonnation practice,
given the presently recognized need for
maintenance of a chlorine residual in the
distribution system. Chlonnation in the
presence of added ammonia is practiced in
some locations in an attempt to maintain
residuals (as chloramme) for extended
periods of time. Figure 12 illustrates the
result of an attempt to form trihalometh-
anes with chlorine added in the presence of
added ammonia. Chlorine was added at 5.5
mg/1 (measured) to raw water and to raw
water spiked with 20 mg/1 NH_,C1 (ammon-
ia nitrogen, 5.2 mg/1). The results of the
measurements for trihalomethane produc-
tion and free- and combined- (mostly
NFLC1) chlorine residuals in Fig. 12 show
that when combined chlorination was prac-
ticed, trihalomethane production was min-
imized. Therefore, during chlonnation of
water where the ammonia breakpoint is not
achieved, trihalomethane production may
not be a problem At this time, however.
ammomation is not recommended as a
technique to avoid trihalomethane forma-
tion because of the relatively poor disin-
fecting power of chlorammes when com-
pared with that of free chlorine.
Full-scale plant operation. The reduction
of ultimate trihalomethane concentration in
finished drinking water is the primary goal
of on-going field research at a number of
water-treatment plants in the U S. Prelimi-
nary results of this work indicate that the
conclusions drawn above with regard to the
role of coagulation and settling in reducing
precursor concentration are valid. The
dramatic reduction of trihalomethane con-
centrations in finished water as a result of
simply changing the point of chlonnation
from before to after the first settling process
at a 160 mgd plant has recently been
reported."'
Summary and Conclusions
The precursor to trihalomethane produc-
tion during the chlonnation process in
drinking-water treatment is probably a
complex mixture of humic substances and
simple low-molecular-weight compounds
containing the acetyl moiety. The relative
importance and contribution to trihalo-
methane production of each of the specific
precursor compounds are pH dependent.
Where chlonnation following clarification
is earned out at pH values near 7, effective
coagulation and sedimentation may be
sufficient to reduce the precursor concen-
tration to levels where ultimate trihalo-
methane concentrations are below the yet
undefined adverse health effect levels.
Where chlorination is carried out at high
pH (as in a lime- or excess lime softening
plant), treatment for precursor removal is
more complicated. In these cases, removal
of relatively water-soluble low-molecular-
weight compounds (concentrations of
which would not be expected to be signifi-
cantly affected by coagulation and settling
processes) is also necessary before chlorina-
tion. Thus, the point of chlorination in the
treatment process, being a significant factor
in trihalomethane production, probably
represents the most important variable to
be considered for change in attempts to
reduce ultimate trihalomethane concentra-
tions in finished drinking water.
To date, GAC has been used with only
limited success to remove precursor com-
pounds. Because its effectiveness is limited
to only a few weeks after being placed in
filters, its use would require frequent activa-
tion or replacement cycles.
Work is continuing in an effort to deter-
mine ways to reduce the extent of tnhalo-
methane reaction through precursor remov-
al or control of reaction rates The final
evaluation of the success of this work must,
however, await more precise health effect
information regarding the significance of
the presence of trihalomethanes in drinking
water.
Acknowledgements
The authors acknowledge the assistance
of the Research Sanitary Engineers, O.T.
Love and J.K Carswell and accompanying
staff, who were responsible for pilot-plant
aspects of this work, B.L. Smith, Physical
Science Technician, for NVTOC analyses
and some chlorine residual measurements;
NOVEMBER 1976
A A STEVENS ET AL 619
-------
J.M. Symons and J.K. Carswell for review
of the manuscript, and Mrs. M. Lilly for its
preparation.
References
1 New Orleans Area Water Supply Study
(Draft Analytical Report) Lower Missis-
sippi River Facility. EPA, Slidell. La
(1974)
2 ROOK. J.J Formation of Haloforms During
Chlormation of Natural Waters Water
Treatment and Examination, 23' Part 2.234
(1974)
3 BELLAR, T A., LICHTENBERG. JJ . &
KRONER. R C. The Occurrence of Organo-
halide.s in Chlorinated Drinking Water.
Jour AWWA.66 II 703 (Dec. 1974)
4 SYMONS. J M.; BELLAR, T.A.. CARSWELL,
J K , DEMARCO, J . KROPP. K L . ROBECK.
GO.. SEEGER. D.R . SLOCUM, C J., SMITH.
B L . & STEVENS. A.A. National Organics
Reconnaissance Survey for Halogenated
Organics in Drinking Water Water Supply
Res Lab and Methods Development and
Quality Assurance Lab.. Natl. Envir Res.
Center, EPA, Cincinnati. Ohio Jour
AWWA. 67 11 634 (New 1975)
5 BUNN, WW.; HAAS, BB: DEANE, E.R.. &
KLEOPFER, R.D. Formation of Trihalometh-
anes by Chlormation of Surface Water.
Accepted for publication. Environmental
Letters, November Issue (1975)
6 LOVE, OT . JR , CARSWELL, J K , STEVENS.
A.A , & SYMONS, J M Treatment of
Drinking Water for Prevention and Remov-
al of Halogenated Organic Compounds (An
EPA Progress Report) Presented at the
AWWA 95th Annual Conf Minneapolis,
Minn (Jun 8-13, 1975)
7 FAIRLESS. B. Personal Communication EPA.
Region V. Central Regional Lab , Chicago.
Ill (1975)
8 GLAZE, W H Personal Communication
North Texas State Unvi , Denton. Tex
(1975).
9 ROOK, J.J Formation and Occurrence of
Chlorinated Organics in Drinking Water
Presented at the 95th Annual Conf
AWWA. Minneapolis, Minn (Jun 8.
1975)
10 ROOK. J.J. Haloforms in Drinking Water.
Jour AWWA, 68-3 168 (Mar 1976)
11 Standard Methods for the Examination of
Water and Waste Water APHA, New York
(13th ed 1971).
12 BELLAR. T.A & LICHTENBERG. JJ Deter-
mining Volatile Organics at the /eg/I Level
in Water by Gas Chromatography Jour
AWWA. 66.11 739. (Dec 1974)
13. STEVENS, A.A & SYMONS, J.M. Analytical
Considerations for Halogenated Organic
Removal Studies. Proc. AWWA Water
Quality Technology Conf, Dallas. Tex.
(Dec 2-3, 1975) p XXVI-1.
14 SCHNITZER, M & K.AHN, S V Hiimtc
Substances in the Environment Marcel
Dekker. Inc. New York (1972)
15 The Effect of Chlormation on Selected
Organic Chemicals. Manufacturing Chem-
ists Association, Final Report. Project 12020
EXG 03/72 EPA, Washington. D C
(1972).
16. KISPERT, EC Plant Modifications Minimize
Chloroform Formation at Cincinnati. Pre-
sented at the AWWA 96th Annual Conf
New Orleans, La. (June 23. 1976)
74054 4241.4310
620 WATER TECHNOLOGY/QUALITY JOURNAL AWWA
-------
APPENDIX 3
Love, O.T., Jr., Carswell, J.K., Stevens, A.A., Miltner, R.J. and Symons, J.M.,
"Treatment for the Prevention or Removal of Chlorination Organics in Drinking
Water," to be submitted to the Journal of the American Water Works Association.
-------
TREATMENT FOR THE PREVENTION OR REMOVAL
OF TRIHALOMETHANES IN DRINKING WATER
by
0. Thomas Love, Jr., J. Keith Carswell, Richard J. Miltner,
and James M. Symons
with assistance and technical consultation from:
Paul A. Keller
Kenneth L. Kropp
Gordon G. Robeck
Dennis R. Seeger
Clois J. Slocum
Bradford L. Smith
Alan A. Stevens
Appendix 3
to
"Interim Treatment Guide for the Control of Chloroform
and Other Trihalomethanes"
-------
TABLE OF CONTENTS
Page
I. INTRODUCTION 1
II. REMOVING TRIHALOMETHANES AFTER FORMATION
A. Aeration 1
B. Adsorption
1. Powdered Activated Carbon (PAC) 5
2. Granular Activated Carbon (GAC) 5
C. Oxidation
1. Ozone (03) 13
2. Chlorine Dioxide (C102) 14
D. Summary of Studies for Reducing Trihalomethanes After
Formation 14
III. TRIHALOMETHANE PRECURSOR REMOVAL 16
A. Aeration 16
B. Adsorption
1. Powdered Activated Carbon 18
2. Granular Activated Carbon 18
C. Oxidation
1. Ozone 26
2. Chlorine Dioxide 29
D. Coagulation
1. Pilot Plant Studies 38
2. Field Studies 43
E. Summary of Trihalomethane Precursor Removal Studies 46
IV. ALTERNATIVES TO CHLORINATION 46
V. CONCLUSIONS 50
-------
I. INTRODUCTION
The issue of chlorinated organics formed in the treatment of drinking
water became a priority topic to the U.S. Environmental Protection Agency in
1975. The work of Rook in The Netherlands and Bellar, Lichtenberg, and
2
Kroner in the United States showed that chlorine used for disinfection reacted
with organic precursor(s) in the water and formed chloroform and other
halogenated organics. To assess the general situation across the United
States a National Organics Reconnaissance Survey was conducted (Symons, et al.,
3
1975). The predominant volatile chlorinated organics found in drinking water
were trihalomethanes - specifically, chloroform (CHCl^), bromodichloromethane
(CHBrCl ), dibromochloromethane (CHBr Cl), and bromoform (CHBr ). Of these
four trihalomethanes, chloroform appeared most frequently and in the highest
concentration. The basic equation — Chlorine + Precursov(s) -> Trihalomethanes
+ Other Chlorinated Organics — suggested the options for controlling the
concentration of these compounds were either to: 1) remove the trihalomethanes
after they were already formed, or 2) to prevent their formation by either
removing the precursor(s) before chlorination or by seeking an alternate
disinfectant. Both approaches have been studied in the Water Supply Research
456
Division laboratory ' ' and this paper summarizes the results to date.
II. REMOVING TRIHALOMETHANES AFTER FORMATION
A. Aeration
Chloroform is lost to the atmosphere when water is held in open vessels
o
or from a flowing stream accidentally contaminated by a chloroform spill.
In beakers standing open at room temperature, almost all of the chloroform was
lost from Cincinnati tap water on 3 days standing. At one European utility,
over 90 percent of the chloroform is lost during three weeks standing in a
deep holding reservoir just prior to water treatment. These data indicate that
-------
- 2 -
chloroform is "volatile" and will be lost from the water at any air-water
interface. Therefore, studying aeration as a unit process for removing
or reducing trihalomethanes after formation seemed logical.
A countercurrent-flow aerator was fabricated out of 3.7 cm (1.5 in.)
diameter glass tubing with a fritted glass diffusor. At an air to water
(volume to volume) ratio common to water treatment aerator design for
controlling taste and odor problems (1:1) the chloroform concentration in the
tap water was not significantly changed (see Table I). An increased air to
water ratio of 8:1 yielded a 58 percent chloroform reduction and a further
increase to 20:1 showed an 83 percent decrease. For perspective, the air
to water ratio used in the aeration basin of a conventional activated sludge
wastewater treatment plant is about 8:1 and the purging step in the volatile
9 10
organic analysis is approximately 44:1. Rook described the stripping
efficiency of a cascading tower aerator (for chloroform in water) that is
comparable to these diffused-air aeration results.
In situations where a free chlorine residual persists through a water
treatment plant, chloroform concentrations increase in spite of some loss to
the atmosphere because of a continuing rapid reaction of chlorine with
precursor.* For example, in one water utility the chloroform concentration
increased from 39 yg/Ji to 83 pg/£, and in another utility the increase was
* - The following are definitions of three terms used throughout this paper.
For details see Stevens and Symons. •'-
1. Instantaneous trihalomethane concentration — the concentration of
trihalomethanes in the water at the moment of sampling.
2. Terminal trihalomethane concentration — the concentrations of
trihalomethanes that occur when a sample of water is stored for a specified
time at a specified pH and temperature'.
3. Trihalomethane formation potential — the difference between the
terminal and instantaneous trihalomethane concentrations.
Mention of commercial products does not constitute endorsement by the U.S.
Environmental Protection Agency
-------
- 3 -
from 18 pg/£ to 63 pg/£ as the water flowed through the sedimentation basins.
Therefore, to fully evaluate aeration as a unit process, consideration of
trihalomethane formation subsequent to treatment must be included. To illustrate,
Figure 1 shows the effluent chloroform concentrations measured in one of the
aeration studies where duplicate samples were collected and one set held
for two days before analysis to simulate distribution storage. Additional
chlorine was added to insure a sufficient residual for the increased contact
time. A slight reduction in chloroform was. noted at the 1:1 air to water
ratio, however, the net chloroform concentrations after two days storage were
approximately the same for both the control sample and the 1:1 aerated sample.
At the 20:1 air to water ratio the chloroform reduction immediately after
aeration was about 85 percent, yet the net reduction based on a two-day
storage time was only 50 to 55 percent. Treatment for reducing the precursor
(trihalomethane formation potential) is further covered in Section III.
Table I
Reduction of Trihalomethane Concentrations in Drinking Water by Aeration
, Chlorine
Air: itfater Residual Trihaiomethanes,
Ratio
_
1:1
8:1
12:1
16:1
20:1
mg/£
1.3
1.2
1.2
1.2
1.2
1.1
CHC13
99
101
45
33
19
16
CHBrCl2
24
5
13
7
8
5
CHBr2Cl
5
5
3
<1
3
3
CHBr3
NFC
NF
NF
NF
NF
NF
a - Activated carbon filtered compressed air.
b - Cincinnati, Ohio tap ater, 10 min. contact time.
c - None found.
-------
- 4 -
•"71
cc
o
LL
O
cc
O
X
o
CHLOROFORM CONCENTRATION
AFTER AERATION
CHLOROFORM CONCENTRATION
AFTER AERATION, RECHLORINATION
AND TWO DAYS STORAGE @25°C
(UNREACTED CHLOROFORM
FORMATION POTENTIAL)
4:1 8:1 16:1
AIR TO WATER RATIOS
20:1
Fig. 1. Removal of Chloroform from Cincinnati, Ohio Tap Water by Aeration.
-------
- 5 -
B. Adsorption
1. Powdered Activated Carbon_(PAC)
Powdered activated carbon at a few milligrams per liter (mg/£) dosage
is often effective as a taste and odor control measure, but large dosages
are necessary to adsorb general organics, as measured by the Carbon
*12
Chloroform Extract (CCE-m) and non-purgeable total organic carbon (NPTOC).
latch studies using a jar test apparatus were conducted to determine the
PAC dosages required to reduce trihalomethane concentrations. The PAC was
added to the water samples, mixed at 100 rpm for two minutes, 50 rpm for five
minutes, settled for 30 minutes, centrifuged at 1500 rpm (480 gravities) for
20 minutes, then decanted and analyzed for trihalomethanes. Table II
shows the results of this procedure on Ohio River water that had been dosed
with alum and chlorine and stored in reservoirs for three days at the
Cincinnati Water Works. The highest PAC dosage (100 mg/£) reduced the
chloroform by 53 percent and the bromodichloromethane by 77 percent. The
initial dibromochloromethane concentration was only 2 yg/£. Bromoform was not
present. A PAC dosage of 100 mg/£ would be costly at a water treatment
plant and would generate a problem sludge.
2. Granular Activated Carbon (GAG)
Glass column 3.7 cm (1.5 in.) in diameter filled with different depths
and types of GAC (See Table III) were exposed to tap water at various
**
hydraulic loadings and contact times to determine the ability of GAC to
remove chloroform and the other three trihalomethanes. At a hydraulic loading
2
of 5 m /hr (2 gal/min/ft ) the trihalomethane reductions through 76 cm (30 in.)
* - The TOC concentration remaining after an acid purge which removes carbon
dioxide and possibly some organics.
** - Apparent contact time is the empty bed volume times the porosity of the
media divided by the flow rate.
-------
- 6 -
Table II
Reducing Trihalomethane Concentration in Prechlorinated Ohio River Water'
Using Powdered Activated Carbon
Powdered Act. Carbon
Trihalomethanes, yg/£
dosage, mg/£
0
1
2
4
8
16
32
64
100
CHC13
64
52
53
51
51
48
45
35
30
CHBrCl CHBr Cl
9 2
7 1
7 1
7 <1
8 1
8 <1
6 1
4 <1
2 <1
CHBr3
NFC
NF
NF
NF
NF
NF
NF
NF
NF
a - Alum and chlorine added and stored for 3 days off-stream in open
reservoirs.
b - Watercarb, Husky Industries, Dunnellon, Florida 32630
c - None found.
of a coal base and also a lignite base GAG are shown in Figures 2 and 3,
respectively. These columns were started at different times, but the
trihalomethane reduction patterns are similar. Chloroform was reduced 90
percent or more for about three weeks then the effluent chloroform concentration
steadily increased until it equalled the influent concentration at about the
ninth or tenth week. The trihalomethanes containing bromine were more
effectively reduced by the GAC. Positive reductions were observed for 26 to 30
weeks for bromodichloromethane, around 40 weeks for dibromochlormethanes and
greater than 40 weeks for bromoform. This may be because the brominated
compounds are present in lower concentrations than chloroform in Cincinnati, Ohio
-------
150-1
UJ
z
LU
O
I
E
TIME IN OPERATION, MONTHS
DIBROMOCHLOROMETHANE
1234
TIME IN OPERATION, MONTHS
BROMODICHLOROMETHANE
1234
TIME IN OPERATION, MONTHS
TEST PERIOD: FEB-MAY 1975
GAC TYPE: FILTRASORB 200
BED DEPTH: 76cm (30 INCH)
HYDRAULIC LOADING: 2gpm/ft2
APPARENT CONTACT TIME: 5 MIN.
BROMOFORM WAS NOT FOUND
Fig. 2. Removal of Trihalomethanes from Cincinnati, Ohio Tap Water by Coal-Base
Granular Activated Carbon.
-------
- 8 -
Z
cc
Z
UJ
o
Z
o
o
1501
100-
50-
UJ
Z
LU
5
o
_l
X
oc
CHLOROFORM
o-1
.'--EFFLUENT
1234
TIME IN OPERATION, MONTHS
DIBROMOCHLOROMETHANE
1234
TIME IN OPERATION, MONTHS
301
20-
10-
BROMODICHLOROMETHANE
1234
TIME IN OPERATION, MONTHS
TEST PERIOD: MARCH-JUNE, 1975
GAC TYPE: HD-10X30
BED DEPTH: 76cm (30 INCH)
HYDRAULIC LOADING: 2gpm/ft2
APPARENT CONTACT TIME: 5 MIN.
BROMOFORM WAS NOT FOUND
Fig. 3. Removal of Trihalomethanes from Cincinnati, Ohio Tap Water by Lignite Base
Granular Activated Carbon.
-------
_ 9 -
tap water and therefore, the lighter loading yields a longer life for the GAC,
or it might be that the brominated compounds are better adsorbed than the
chloroform. The latter is probably the most likely explanation as this has
also been suggested by Rook.
Although periods existed when the effluent trihalomethane concentrations
exceeded the influent (Note chloroform and bromodichlorome thane desorption in
Figure 4) , a materials balance after 40 weeks accounted for all but 6 percent
of the total trihalomethane loading in the coal base GAC system and 16 percent
in the lignite base GAC system. The total trihalomethane loading used in
the materials balance is a summation of the product of the averaged weekly
flows and the total trihalomethane concentrations. The total trihalomethane
concentration, expressed in terms of micromoles per liter, is the sum of
the individual trihalomethanes divided by their respective molecular weights.
For example, a water sample containing 50 yg/£ CHC1 , 26 yg/£ CHBrCl , 12 yg/£
SO 9 f\ 19 1
CHBr.Cl, and 1 yg/£ CHBr. would have ~0 + -ff- . + ±=-Q + ^ = 0.64 ymoles/liter
j j j_xy
total trihalomethanes.
Figure 5 summarizes a study where both the flow rate and GAC depth were
manipulated to give constant contact times. The rate at which the water was
applied to the GAC (i.e., contact time) had a direct effect on the life of the
bed. Doubling the contact time from 5 minutes to 10 minutes essentially doubled
the effectiveness of the GAC for chloroform removal. In other words, if the
flow rate is doubled and the depth remains the same, the life of the GAC bed
is reduced to one-half as shown in Figure 6.
TABLE III
Granular Activated Carbon Characteristics
Coal Base Coal Base Lignite Base
Filtrasorb 20Qa Filtrasorb 400a HD-10 x 30
Surface Area by^Nitrogen Gas
BET method, m /gm 850-900 1050-1200 600
Uniformity Coefficient 1.7 1.9 1.7
Effective Size, mm 0.55 - 0.65 0.55 - 0.65 0.9 - 0.9
Density, lbs/ft3 30 25 23.5
§- Calgon Corp., Pittsburgh, Pa.
~ ICI-US, Wilmington, Del.
-------
- 10
100-
50-
O
HI
>
O -200J
I***. 100-
\CHLOROFORM
LIGNITE GAC
S! ', r« ' COAL GAC 5°-
\ '";•' V"
\;<:
iii ° i:\l//^
a ' * /
• ..,! „ cn
•••**
\ * * * *
V.A
y
-/I
iVV. ,
f\i/ V»
V ''I
(
BROMODICHLORO-
METHANE
2 10 20 30 40 10
20
1-1
Vi 7
¥ i i
1 I
30 4
WEEKS IN SERVICE
WEEKS IN SERVICE
O
cr
1 oo-t*»- »v
50-
DIBROMOCHLOROMETHANE
° LIGNITE BASE GAC (HD-10x30)
. COAL BASE (FILTRASORB-200)
BED DEPTH: 76cm (30 INCH)
HYDRAULIC LOADING: 2gpm/ft2
APPARENT CONTACT TIME: 5 MIN.
BROMOFORM WAS 100% REMOVED
WHEN FOUND IN THE TAP WATER
10 20 30 40
WEEKS IN SERVICE
Fig. 4. Removal of Trihalomethanes from Cincinnati, Ohio Tap Water using Two
Types of Granular Activated Carbon.
-------
- 11 -
100-
Q
LU gO-
O
s
LU
EC
CINCINNATI, OHIO TAP WATER
AVERAGE APPLIED CHLOROFORM CONCENTRATION=24//g/L
TYPE: FILTRASORB 400
2 gpm/ft2
90cm (36 INCH)
4 gpm/ft2
180cm (72 INCH)
LU
O
OC
LU
60-
4o-
20-
1 gpm/ftz V*. \
90 cm (36 INCH) \ \
^~ 2 gpm/ft2
\ 180cm (72 INCH)
0. H
Q. ,
50% EFFECTIVE
= 12A(g/L)
10 15
TIME IN WEEKS
20
25
Fig. 5. Effect of Contact Time on Chloroform Adsorption on
Granular Activated Carbon.
-------
- 12 -
100
AVERAGE CHLOROFORM CONCENTRATION
IN APPLIED CINCINNATI, OHIO TAP WATER=
GAC DEPTH: 90 cm (36 INCH)
GAC TYPE: FILTRASORB 400
Q
LLI
HI
DC
UJ
O
DC
UJ
50% EFFECTIVE
°0 CHLOROFORM = 23 /ug/L
-20
234567
TIME IN WEEKS
Fig. 6. Effect of Hydraulic Loading on Chloroform Removal from
Tap Water.
-------
- 13 -
In field studies where the applied water and the effluent from GAG
beds were sampled, the findings regarding trihalomethane reductions were very
similar to the laboratory results. The actual operating GAG beds were
exhausted readily for chloroform yet the bromine containing trihalomethanes
were removed for longer periods as shown in Table IV.
TABLE IV
Trihalomethane Removal at Water Treatment Plants Using Granular Activated Carbon
Granular Activated Carbon Trihalomethanes, yg/£
Time in Service, months CHC13 CHBrCl CHBr Cl CHBr
b
Plant A - Settled Water
Filter No. 1 Effluent
Filter No. 2 Effluent
b
Plant B - Settled Water
Filter Effluent
Plant C - Settled Water
Filter Effluent
9
36 8
4 8
14
2-1/2 10
11
2 10
1.0
1.4
1.4
4.2
1.7
2.3
0.9
5.0
8.8
NF
1.4
<.l
0.8
0.3
i
NF
NF
NF
NF
NF
NF
NF
a - GAG used as both a filter and an adsorber that received chlorinated,
coagulated and settled water.
b - GAG filter influent.
c - None found.
C. Oxidation
1. Ozone (0,,)
_j
For these studies, a 3.7 cm (1.5 in) diameter glass counterflow ozone
contactor was fabricated. The contact time in the contactor could be varied
by adjusting the water depth. Ozone was generated by a Welsbach Model T-408
generator using "Aviator's Breathing" grade oxygen. The ozone-oxygen gas
mixture was dispersed through a fritted glass sparger in the bottom of the
contactor. Applied ozone concentrations were determined by an lodometric
Method. In an effort to maximize contact between the ozone-oxygen mixture
and the water, a small, high speed propeller mixer
-------
- 14 -
was positioned just below the water surface within the column. The propeller
caused almost complete dispersion of the rising bubble pattern, however, even
at an applied ozone dose of 25 mg/£ (4 to 5 minute contact time), attempts
to remove trihalomethanes from tap water were unsuccessful.
2. Chlorine Dioxide (CIO.,)
Chlorine dioxide is used either year-round or on a part-time basis in
approximately 100 water treatment plants world-wide for the control of tastes
and odors, iron and manganese removal, and, to a very limited degree,
*
disinfection. This phase of study examined C10~ prepared by reacting technical
grade (80% pure) sodium chlorite (NaClCL) with sulfuric acid, air-stripping
the chlorine dioxide from solution and trapping the gas in nitrogen purged
distilled water. Analyses for chlorine, chlorine dioxide and chlorite
(CIO ) were made using the DPD procedure. At dosages up to 10 mg/£ and
stored for 2 days , chlorine dioxide, like ozone, was ineffective in reducing
the trihalomethanes already present in drinking water.
Summary of Studies for Reducing Trihalomethanes After Formation
Studies conducted on removing trihalomethanes from drinking water
included the processes of diffused-air aeration, granular and powdered
activated carbon adsorption, ozonation, and treatment with chlorine dioxide.
Table V shows what conditions were found necessary to effect reductions
in chloroform concentrations. The bromine containing trihalomethanes were
more effectively reduced in concentration than was chloroform by granular
activated carbon however, the results of the other unit processes demonstrated
*EPA Grant No. R804385-01 "Status of Ozonation and Chlorine Dioxide Technologies
for Treatment of Municipal Water Supplies," Public Technology, Inc., Washington,
D.C., will attempt an accurate count of the water utilities using chlorine
dioxide.
-------
- 15 -
that the trihalomethanes are not easily removed from water. Therefore,
research was directed toward preventing their formation rather than removing
these materials after they are already formed.
TABLE V
Effectiveness of Various Unit Processes for Reducing Chloroform in Drinking Water
Initial Chloroform
Process Concentration, yg/£ 50% 25% 10%
Aeration
Air to Water Ratios
for diffused-air
aeration: 10 min.
contact time 100 6:1 15:1 25:1
Granular Activated Carbon
Expected life for
5 min. contact time 55 7 weeks 5 weeks 4 weeks
Powdered Activated Carbon
Dosage, mg/£ applied to:
a. Chlorinated Raw Water 64 95 mg/£ > 105 mg/£ > 105 mg/£
b. Chlorinated Coagulated
and Settled Water 44 27 mg/£ 90 mg/£ 105 mg/£
Ozonation
4 min. contact time Up to 25 mg/£ CL had no effect
on the chloroform concentration.
Chlorine Dioxide
Up to 48 hr. contact Up to 10 mg/£ C10? had no effect on the
time chloroform concentration.
-------
- 16 -
III. TRIHALOMETHANE PRECURSOR REMOVAL
Because no direct measurement exists for trihalomethane precursors,
the degree of precursor removal was judged by comparing trihalomethane
concentration upon chlorination of an untreated control (called the tri-
halomethane formation potential) to similar data collected on a treated
water after similar chlorination. For example, if the effluent from a sand
filter that was chlorinated and stored for two days yielded 50 yg/£ chloroform
and the same effluent passed through an adsorbent then chlorinated and stored
under similar conditions produced 25 yg/£ chloroform, the adsorbent was said
to be 50% effective in removing chloroform formation potential. This example
assumes that no instantaneous trihalomethanes are present in the filter
effluents.
A. Aeration
Aeration, already shown to be largely ineffective in reducing trihalomethane
concentrations, was evaluated for reducing trihalomethane formation potential
in Ohio River water. Using the diffused-air aerator described in Section II A,
river water was aerated at varying air-to-water ratios, then chlorinated
and stored at 25°C for two days in sealed vessels. The chlorine solution
used in all the precursors removal studies was made by bubbling gaseous chlorine
through nitrogen purged distilled water and the residuals were determined
amperometrically. The contact time of two days was selected as a matter
of convenience for maintaining free chlorine residuals and because experience
had shown that the rate of trihalomethane formation for this water was fastest
during the first few hours and then greatly reduced after 30 to 40 hours
contact time. A companion river water sample was chlorinated and stored without
being aerated. Table VI shows the influence of aeration on trihalomethane
formation potential to be insignificant even at an air to water ratio of 20
to 1, As shown earlier in Figure 1, the chloroform formation potential also
-------
- 17 -
TABLE VI
Effect of Aeration on Reducing Trihalomethane Formation Potential
Concentration, yg/£ after 2 day
Air :Water
Ratio
Ohio River Water -
Ohio River Water + 13 mg/£
Aerated* Ohio River Water 1:1
contact time
CHC10
NF
66
66
CHBrCl,,
NF
28.0
27.8
CHBr^Cl CHBr0
NF NF
8.0 <0.1
8.0 <0.1
+ 13 mg/£ Cl,
4:1
6:1
8:1
10:1
20:1
64
62
62
59
61
26.8
25.8
26.8
25.6
26.0
6.6
7.6
7.8
7.7
8.0
^Activated carbon filtered
compressed air, 10 min. contact time
NF = None Found
TABLE VII
Effect of Powdered Activated Carbon (PAC) on Trihalomethane Formation Potential
Concentration , yg/£ (2-day contact time)
PAC Dose
mg/£ CHC13 CHBrCl2
Settled
Settled
Settled
Water
Water + Cl
Water + PAC + Cl 2
4
8
20
50
100
NF
27
22
25
20
16
11
9
NF
15.2
15.1
16.4
15.8
16.9
13.0
9.5
CHBr Cl CHBr
NF NF
10.4 <.l
8.0 <.l
10.2 <.l
9.4 <.l
12.2 <.l
10.0 <.l
8.8 <.l
TTHM
NF
0.37
0.31
0.36
0.32
0.29
0.22
0.18
% Remov;
of TTHM
0
16
2
14
22
41
51
a - Alum coagulation
+ settling
b - Average of five experiments using three different commercial brards PAC.
c - Total Trihalomethane concentration.
-------
- 18 -
remained in an aerated tap water sample.
B. Adsorption
1. Powdered Activated Carbon
In the discussion on the removal of trihalomethanes, powdered activated
carbon was shown to reduce chloroform by 50 percent when dosed as high as
100 mg/£. To determine the effectiveness of PAC on removal of trihalomethane
formation potential, coagulated and settled water from the pilot plant
(described in Section IIIB-2) was dosed with varying quantities of PAC,
mixed at 100 rpm for 2 minutes, then centrifuged for 20 minutes at 1500 rpm
(480 gravities). The supernatant liquor was then decanted and chlorinated,
rapidly mixed for 2 minutes, then stored for two days at 25°C. As with the
trihalomethane removal studies, PAC had to be added in large doses before it
had much effect on reducing the trihalomethane formation potential. For
example, at the impractical PAC dose of 100 mg/£, 49 percent of the total
trihalomethane formation potential remained after treatment (see Table VII).
2. Granular Activated Carbon
A pilot water treatment plant was fabricated to provide a continuous
supply of treated, yet unchlorinated water for precursor removal studies.
To minimize contamination from structural materials the plant was built
almost entirely of stainless steel, Teflon and glass. Through the
assistance and cooperation of the Cincinnati Water Works, Ohio River water
was provided as a source of raw water. The pilot plant employed conventional
alum coagulation, flocculation, sedimentation, and originally had three
parallel filtration schemes — 76 cm (30 in.) dual media (anthracite and
sand); 76 cm (30 in.) of granular activated carbon (GAG) that acted as both
a filtering and adsorption media (which was termed filtration/adsorption); and
a dual media column followed by 76 cm (30 in) of GAC. The reason for this
latter arrangement (which was termed post-adsorption) was to see if the life
-------
- 19 -
of a GAG bed could be extended if prefiltered water was applied. The
2
nominal filter flow rates were 100 m£/min, which equaled a loading of 2 gpm/ft
(5 m/hr) with an apparent contact time within the GAG beds of slightly over
4 minutes. A simplified flow pattern is shown in Figure 7 and a detailed
schematic of the pilot treatment plant can be found in Reference 4. The
disinfection step (chlorine, chlorine dioxide, ozone or combinations of
these) followed filtration so that trihalomethane formation potential removal
could be monitored by measuring subsequent trihalomethane formation
after an appropriate contact time.
After 10 weeks of study the conclusion was reached that little was gained
in reducing the trihalomethane formation potential by prefiltering the
water before exposing it to GAG (see Figure 8). Although little difference
appears in performance between the two modes, there are other considerations
(such as GAG handling, flexibility in bed depths, etc.) that affect the
overall economics of the system that must be included in any comparative
evaluation.
Figure 9 shows the effect of 76 cm (30 in.) of coal base GAG on
trihalomethane formation and Figure 10 shows the effect on the same
parameter when a 152 cm (60 in.) lignite base GAG filtration/adsorption mode
replaced the post adsorption mode. Note that the apparent contact time for the
coal base GAG was five minutes and ten minutes for the lignite-base GAG.
As the GAG aged, at times the trihalomethane formation potential equalled
or exceeded the levels from the dual-media filter (see Figure 11). The
reason for this is not clear but it could have been caused by biological
activity within the GAG beds (because no disinfectant was applied before the
-------
- 20 -
COAGULATED & SETTLED WATER
COAL
SAND
f
D>
•^
i
CL2
CL02
te
}
1
COAL
GRANULAR
ACTIVATED
CARBON
i
t
1
v^i
*-
— ^
-»-
-^
,r
;.'.;.'.;;•
I
DUAL MEDIA
FILTRATION
FILTRATION/
ADSORPTION
POST
ADSORPTION
Fig. 7. Schematic ot Pilot Plant for Reducing the Trihalomethane Formation
Potential.
-------
- 21 -
100
POST ADSORPTION
FILTRATION/ADSORPTION
100
50-
123456789 10
TIME IN OPERATION, WEEKS
BROMODICHLOROMETHANE
100
DIBROMOCHLOROMETHANE
1234 56789 1O
TIME IN OPERATION, WEEKS
• 76 cm (30 INCH) FILTRASORB 200
^ DUAL MEDIA + FILTRASORB 200
HYDRAULIC LOADING: 2 gpm/ft2
APPARENT CONTACT TIME: 5 MIN.
TRIHALOMETHANE FORMATION POTENTIAL,ug/\_
AVERAGE RANGE
CHLOROFORM 14 8-23
BROMODICHLOROMETHANE 11 9-13
DIBROMOCHLOROMETHANE 4 0.2-8
BROMOFORM NOT FOUND
-50
8 9 10
Fig. 8. Comparison of Filtration/Adsorption vs. Post Adsorption for Removing
Trihaiomethane Formation Potential.
-------
- 22 -
GAC TYPE: FILTRASORB 200
HYDRAULIC LOADING: 2 gpm/ft2
APPARENT CONTACT TIME: 5 MIN.
CHLOROFORM FORMATION POTENTIAL enLORINATION CONDITIONS: 2-3 mg/L
75-i (CHLORINATED DUAL MEDIA EFFLUENT) CL2 FOR 4 DAYS
TEST PERIOD: MAR.-OCT., 1975
50-
25-
(x-
k\
/
J
TRIHALOMETHANE FORMATION POTENTIAL ,|Jg/L
_i.rvj -i W
O O Oi O
II 11
/~'~N-v- ' '•/ "\ /
IS\ .• 0.°, \, ~a »/ o-cf
or*--0'' * ^^ CHLORINATED GAC EFFLUE
^ ~.O«O*
5 10 15 20 25 30
BROMODICHLOROMETHANE FORMATION POTENTIAL
(CHLORINATED DUAL MEDIA EFFLUENT) •
^^/-•"\ o / o""\\
A /-^v0' ^
l\ 1 \ ,' *o
,/\ '°**-c7
•"*•*•*. //>*"'8>x*"*X««v^/
,°-°'°'0/ ^ ^" CHLORINATED GAC EFFLUENT
„* &
5 10 15 20 25 30
DIBROMOCHLOROMETHANE FORMATION POTENTIAL
(CHLORINATED DUAL MEDIA EFFLUENT)
^^/V \<' "V
/! \ A / »e\
y/ ° \
/ y^ytfi \
r$*<**'K CHLORINATED GAC EFFLUEN
5 10 15 20 25 30
TIME OF OPERATION, WEEKS
Fig. 9. Use of Adsorption on 76 cm (30 inch) Coal-Base Granular Activated Carbon
for Removing Trihalomethane Formation Potential.
-------
- 23 -
75-1
50-
25-
CHLOROFORM FORMATION POTENTIAL QAC TypE HD 1Qx30
(CHLORINATED DUAL MEDIA EFFLUENT) HYDRAULIC LOADING. 2
APPARENT CONTACT TIME: 10 MIN.
CHLORINATION CONDITIONS. 2-3 mg/L
FOR 4 DAYS
TEST PERIOD. MAY-DEC., 1975
a.q)-*.ra.ft^-' •Q.fl.--)-'' 'V vCHLORINATED GAC EFFLUENT
10
15
20
25
30
UJ
O
a
DC
O
LL
HI
z
LLJ
5
O
BROMODICHLOROMETHANE FORMATION POTENTIAL
(CHLORINATED DUAL MEDIA EFFLUENT)
30-i
15-
CHLORINATED GAC EFFLUENT
10
15
20
25
30
20-i
10-
DIBROMOCHLOROMETHANE
FORMATION POTENTIAL
(CHLORINATED DUAL MEDIA EFFLUENT)
\i CHLORINATED * GAC EFFLUENT
5 10 15 20 25
TIME IN OPERATION, WEEKS
30
Fig. 10. Use of Adsorption on 152 cm (60 inch) Lignite-Base
Granular Activated Carbon for Removing Trihalomethane
Formation Potential.
-------
- 24 -
-100
100
50-
CHLOROFORM
30
TIME IN OPERATION, WEEKS
-100
BROMODICHLOROMETHANE
10 20 30
TIME IN OPERATION, WEEKS
100
Q
LU
o
LU
DC
I-
lil
O
oc
UJ
Q.
50-
-200
DIBROMOCHLOROM ETHANE
TIME IN OPERATION, WEEKS
• 76 cm (30 INCH) FILTRASORB 200
TRIHALOMETHANE FORMATION POTENTIAL,
AVERAGE RANGE
CHLOROFORM 24 11-65
BROMODICHLOROMETHANE 17 8-32
DIBROMOCHLOROMETHANE 9 0.2-20
13152 cm (60 INCH) HD-10X30
TRIHALOMETHANE FORMATION POTENTIAL, yg/L
AVERAGE RANGE
CHLOROFORM 35 10-70
BROMODICHLOROMETHANE 23 8-34
DIBROMOCHLOROMETHANE 11 5-20
SEE TEXT FOR DISCUSSION OF BROMOFORM
APPARENT CONTACT TIME = 5 mm.
•• " •' = 10 min.
Fig. 11. Use of Granular Activated Carbon for Removing Trihalomethane
Formation Potential.
-------
- 25 -
filter) or may have resulted from clumps of organic materias working their
way down through the filters from the surface because no surface scrubbers
were in the filters to aid in backwashing.
The relative effectiveness of the GAG to prevent the formation of
trihalomethanes was highest for chloroform and lowest for bromoform.
Sufficient precursor was being passed through the GAG after one week to
produce a measureable amount of trihalomethane during a 4-day chlorine
contact time. A 4-day contact time was selected to approximate what would likely
be a maximum retention time in a distribution system. The 76cm deep GAG
column was exhausted (i.e., the trihalomethanes formed upon chlorination of the
GAG effluent equalled those levels formed in the chlorinated dual media
effluent) in about 13 weeks for chloroform, 8 weeks for bromodichloroniethane,
5 weeks for dibromochloromethane, and probably less than 2 weeks for bromoform.*
The reason for this might be that the GAG does not remove bromide effectively
so bromide in the effluent, plus the first traces of precursors will form
brominated trihalomethanes upon chlorination because the oxidation of
bromide to bromine by chlorine followed by bromination occurs faster than the
chlorination reaction. As the GAG ages and more and more precursors break
through (bromide still being present) chloroform is produced. This reasoning
is speculative, but work currently underway in the Water Supply Research
i
Division laboratory will provide a better understanding of the brominated
compounds and their behavior in a unit process.
*Bromoform was seldom detected in the pilot plant studies, however, when it was
formed in the dual media effluent it was detected in equal concentrations in
the GAG effluent.
-------
- 26 -
Field data to support these general findings on the removal of precursor
materials are lacking because the water utilities using GAG in the United
States chlorinate prior to the filters and thus the GAG is exposed to some
instantaneous trihalomethane levels. Granular activated carbon, as well
as other adsorbents will be further investigated in the future.
C. Oxidation
1. Ozone
Using the ozone contactor described earlier, a series of disinfection
studies were conducted to determine the minimum effective applied ozone dose
(mg 0_/liter HO) for disinfecting the pilot plant filtrates. The effluents
from the dual media filer and the 76 cm (30 in) deep GAG filter were ozonated
at various doses and then examined for bacteriological quality by both
coliform and standard plate count (SPG) methods. Figure 12 shows typical
SPG results obtained. The GAG effluent was more easily disinfected because of
the lower 0« demand. From these data, a minimum effective disinfection dose
of less than 1 mg 0,,/liter HO was selected as a starting point for precursor
removal studies.
Table VIII shows the trihalomethane results for an applied dzone dose
ranging from near the disinfection minimum to over 200 mg/£ on effluents
from the dual media and GAG filters. Although ozone alone would not form
trihalomethanes (under these conditions ozone did not oxidize chloride to
chlorine) ozonating for a few minutes contact time with small dosages followed
by chlorination produced as much (or more) chloroform as with chlorination alone.
This means that the trihalomethane formation potential was not reduced by
low level ozonation and subsequent chlorination to produce a disinfectant
residual in the distribution system would result in trihalomethane production.
The reason why low level ozonation plus chlorine produced more chloroform than
-------
- 27 -
1000-
o
X
00
100-
O
o
UJ
o
cc
<
Q
Z
(0
DUAL MEDIA EFFLUENT, pH=7.3
FILTRATION/ADSORPTION
GRANULAR ACTIVATED CARBON EFFLUENT
pH=7.9
GAC AGE: 8 WEEKS IN SERVICE
OZONE CONTACT TIME=6 MIN.
0.1
0.5
0.6
APPLIED OZONE, mg/L
Fig. 12. Post Disinfection with Ozone.
-------
- 28 -
TABLE VIII
Effect of Ozonation of Trihalomethane Formation Potential Removal
Continuous Flow Studies
0~ Contact Time = 5-6 Minutes
Sample
Dual Media
Effluent
If
M
II
II
Tl
IT
II
1!
Granular Activated
Carbon Effluent
tt
ii
n
..
"
n
Applied*
Ozone
Dose
mg/£
0
0.7
0
0.7
18.6
0
18.6
0
227
0
0.7
0
0.7
20.9
0
20.9
Chlorine
Dose
mg/£
0
0
8
8
0
8
8
8
8
0
0
8
8
0
8
8
Bromo-
dichloro-
Chloroform methane
Mg/£ yg/£
< 0.
< 0.
6
15
< 0.
12
14
91
62
None
None
2
3
None
4
5
2 None found
2 None found
14
8
2 None found
9
8
26
7
found None found
found None found
3
3
found None found
4
4
Dibromo
chloro-
methane
yg/£
None found
None found
4
3
None found
2
8
6
1
None found
None found
< 1
2
None found
2
2
Note: Bromoform was not found in any of these samples 'and all samples were
stored at 25°C for 6 days.
^Applied dose, continuous flow studies, mg/£ =
mg 03 standard liters of gas (0 + 0 )
standard liter of gas (0 + 0 ) minute
min.
liters, water
-------
- 29 -
chlorination alone is not known. It might be that ozone is altering some
material that would not normally participate in the haloform reaction as a
precursor, or possibly that because the ozone satisfies some of the oxidant
demand more chlorine is available for the haloform reaction. The applied
dose that was greater than 200 mg/£ may have completely oxidized some of the
trihalomethane precursors, thereby reducing the chloroform formation potential
from 91 to 62 yg/£. The reduction in bromine containing trihalomethane could
be the result of bromide losses because of ozonation.
To observe the effect of longer contact time s and generally higher ozone
doses, the contactor was used as a batch reactor rather than a continuous,
counter-current reactor as in previous runs. Figure 13 shows that the
trihalomethane formation potential can be reduced by ozone with contact
times that are probably unrealistic (1 to 2 hours or more) for water treatment.
The ozone application rate for this batch study was 43.5 mg 0,,/min
(applied to approximately 12.7 liters of river water). This application rate
is about 100 times greater than the minimum of 1 mg/£ applied ozone required
for effective disinfection in the pilot plant. In this batch test the calculated
gas to water ratio for the 6 hour contact time is approximately 14 to 1.
That indicates the effect is because of ozone and not merely gas stripping
as aeration alone at a 20 to 1 air to water ratio was ineffective for reducing
the trihalomethane formation potential (see Table VI and Figure 1).
2. Chlorine Dioxide
Chlorine dioxide is commonly generatedby mixing aqueous solutions of
sodium chlorite (NaClCL) and chlorine (which is purposely applied in excess to
insure complete consumption of the chlorite). Therefore, both C10~ and the
combined species of C10_ and Cl were examined for their effect on the
trihalomethane formation potential. Like ozone, a chlorine dioxide dose
-------
- 30 -
_i 0.6
<
I-
z
UJ
P S 0.5
I- o
si
cc
O z
u. o
-------
- 31 -
sufficient for disinfection was selected as the minimum dosage for these studies.
Figure 14 is a typical disinfection curve for CIO using effluent from the
dual media and granular activated carbon filters in the pilot plant. As
with the ozonation studies, the GAG effluent was more easily disinfected than
the dual media effluent.
The trihalomethanes formed after dosing untreated and treated pilot plant
water with chlorine-free chlorine dioxide (generated with sulfuric acid as
discussed earlier in this paper) are shown in Table IX. Of the four trihalomethanes,
only chloroform was detected in the CIO. treated samples. Because the differences
between instantaneous and terminal chloroform concentrations were always
less than 0.2 Mg/£ and the precision of the volatile organic analysis at
these low concentrations is approximately - 0.2 ug/&, chlorine-free chlorine
dioxide was concluded not to form trihalomethanes, thus, acting as an
oxidizing agent rather than a chlorinating agent. These results were
encouraging, but because chlorine is usually present with chlorine dioxide in
practice, another study was "undertaken to examine the effect of the combined
species, or what is sometimes referred to as chlorine dioxide with excess
chlorine.
Figure 15 is a schematic of the CIO generating and feeding scheme.
As before, technical grade sodium chlorite salt was dissolved in nitrogen
purged distilled water to form the aqueous NaCIO,., solution. The concentration
was checked occasionally by the DPD method. The chlorine solution was
prepared by passing high purity grade chlorine gas through nitrogen purged
distilled water as described earlier. Sufficient chlorine was introduced,
so that the effluent from the generator contained chlorine dioxide and chlorine
-------
- 32 -
10000
DUAL MEDIA EFFLUENT
CIO2 CONTACT TIME = 30 min.
pH = 7.0 - 8.1
TEMPERATURE = 22 - 26°C
FILTRATION/ADSORPTION
GRANULAR ACTIVATED
CARBON EFFLUENT
GAC AGE: 24 WEEKS IN
SERVICE
NONE DETECTED AT
0.2 mg/l CI02 DOSE
0.1 0.2 0.3 0.4 0.5
APPLIED CHLORINE DIOXIDE, mg/l
Fig. 14. Post Disinfection with Chlorine Dioxide.
-------
- 33 -
TABLE IX
Trihalomethane Formation Using Chlorine Dioxide
cio2
Dose
Sample mg/£
Untreated Ohio
River Water3 1
2
2
II II o
n n 9
Dual Media Filter
Effluent3 2
" " 2
" 2
" "b 0
5
" 5
Granular Activated
Carbon Filter
Effluentb 0
" " 0
1
6
6
.4
.7
.7
.7
.7
.0
.0
.0
.5
.5
.5
.4
.9
.0
.6
.6
Residual
mg/£
0.
1.
1.
0.
0.
1.
-
1.
0.
4.
2.
0.
0.
0.
4.
3.
7
5
3
8
3
3
1
1
0
3
2
2
4
0
6
Contact Time
Hours
0.
0.
6.
18
42
0.
6.
19
0.
0.
5
5
0
5
5
5
5
114
0.
0.
0.
0.
5
5
5
5
114
Chloroform , yg/£
Instantaneous
< 0.
< 0.
< 0.
< 0.
< 0.
< 0.
< 0.
< 0.
0.1
0.1
0.1
0.2
0.4
0.2
0.4
0.4
1
2
2
2
2
2
2
2
Terminal
0
0
0
0
<
<
<
<
0
0
0
0
0
0
0
0
.2
.2
.1
.1
0.
0.
0.
0.
.1
.1
.1
.3
.3
.3
.3
.3
2
2
2
2
a - Batch dosing in headspacefree bottles.
b - Continuous dosing in pilot plant.
c - Brotnodichlormethane, dibromochloromethane and bromoform were not found.
Water pH range: 7.2 - 8.1 Temperature Range: 23-26°C
-------
34 -
WATER FROM PILOT PLANT
SAMPLING |
TAP-
SAMPLE PORT TO
CHECK FLOW, CIO 2
STRENGTH, AND
MEASURE EXCESS
CHLORINE
1=1XXXXXX
'X
EFFLUENT
TO CONTACT
CHAMBER
STATIC
MIXER
PERISTALTIC PUMP
SODIUM CHLORITE
SOLUTION*
CHLORINE SOLUTIONS
ADJUSTED TO pH 2 - 3
WITH H2S04
'CONCENTRATIONS BASED ON GRANSTROM AND LEE,
JAWWA, 50: 1453, 1958.
Fig. 15. Schematic of Chlorine Dioxide Generator used in Pilot Plant Studies.
-------
- 35 -
with no measurable chlorite. The production of chlorine dioxide in this
manner is based on the work done by Grantstrom and Lee (1958).
The trihalomethanes that were formed after adding chlorine alone or
chlorine dioxide with chlorine to the effluent from the dual media filter
in the pilot plant can be compared in Table X. Although trihalomethanes
were formed in the presence of chlorine dioxide and chlorine, the levels
were less than with chlorine alone. For example, referring to Table X,
3 yg/£ chloroform were formed after 22 hours contact with 1.3 mg/£
chlorine dioxide and 1.5 mg/£ chlorine. On the other hand, 17 yg/Ji
chloroform were formed under the same conditions when 1.5 mg/£ chlorine
was added with no chlorine dioxide. Therefore, chlorine dioxide, although
it did not form trihalomethanes, affected the haloform reaction. After
numerous experiments of this type, as the chlorine dioxide to chlorine ratio
was increased, the formation of trihalomethanes was found to decrease. In
Figure 16, the chloroform formation is plotted against the C1CL to Cl~ ratio.
For these experiments the formation of chloroform was concluded to be
minimized if the chlorine dioxide to chlorine ratio is kept above 2. Similar
plots were obtained for bromodichloromethane and dibromochloromethane formation.
Ohio River water that had been treated with 2 mg/£ chlorine dioxide
such that all of the chlorine dioxide had been consumed in 48 hours was
subsequently chlorinated at 8 mg/£ as a follow-up experiment to determine
if chlorine dioxide was reducing the trihalomethane formation potential as
the data in Figure 16 would indicate. Only 50 to 70 percent of the trihalomethanes
were formed in 1 to 2 days when chlorine dioxide treatment preceeded
chlorination as compared to the trihalomethane formation when 8 mg/£ chlorine
was used alone. Similar results were obtained when a study was conducted
with humic acid solution. Therefore, chlorine dioxide was concluded to be
reducing the precursor concentration.
-------
- 36 -
X
w
t-3
pQ
<3]
H
C
CU
3
rH
MH
m
W
j_j
(U
4J
, |
•H
Pn
CO
•H
rrj
(1)
1
rH
CO
a
o
o>
s3
•H
O
rH
U
cn
cn
0)
a
X
4-1
3
0
42
4J
•H
13
(3
cfl
42
4J
•H
**^s
0)
13
•H
O
•H
P
CD
a
•H
o
HH
O
a
o
•H
4J
•H
13
13
?
•^^,
60
B
'"O 'O T3 *^3 *T3 T3
(3 (3 (3 C 13 G
3 3 3333
O O O O O O
<4H <4H MH 14H <4H <4H
QJ 1) 0) QJ 01 CU
C C CN (3 C (3 S3
O rH O O O O O
JscNi— icsi-3-in!3r-ir-H25!2;S3>2
rrj
c
3
0
a) m
C m • CN m
O ^O rH CO O
!H;vOrH-i
in romro CNCNCNCN
• ••• ••••
ml IrHrHrHl 1 l
-------
- 37 -
25
DC
O
I
00
DC
LLJ
Q
HI
5
DC
£
DC
O
LL
O
DC
O
I
O
20
15
10
0.5 _ mg/l CIO2
2.0 ~ mg/l Cl 2
!.i?
I
1 7
2.0
/v20
045 l| \\
,135 M / \
\ 4.2
\ 22!
13 \»
1 5 «v
1 25 •• 1.3
165
175 4,7
1 9
\ 0.9
345
0.8
05
Fig.
0 123456
RATIO OF CHLORINE DIOXIDE (mg/L) TO CHLORINE (mg/L)
16. Formation of Chloroform in Dual Media Effluent by Chlorine Dioxide with
Excess Chlorine.
-------
- 38 -
D. Coagulation
1. Pilot Plant Studies
Early in these studies, samples were collected before and after the various
unit processes within the pilot plant and analyzed for NPTOC (non-purgeable
total organic carbon). The relative NPTOC results shown in Figure 17 are
typical and generally as expected for similar results have been demonstrated
18
in a full-scale water treatment plant. Coagulation, flocculation and
sedimentation and filtration have a marked effect on the NPTOC concentration
(approximately 60% reduction). To determine whether or not trihalomethane
precursor was also removed during conventional treatment, raw water, coagulated
and settled water, and dual-media filtered water from the pilot plant were
chloinrated in closed containers to determine the production pattern of
19
trihalomethanes. These experiments revealed that the pattern for reducing
the chloroform formation potential was similar to that of NPTOC reduction as
shown by the typical trihalomethane formation curves for the various qualities
of water in Figure 18. Conventional treatment, however, had much less effect
on preventing the formation of bromine containing trihalomethanes. Bromide
may not be significantly affected by coagulation and remains available for
oxidation to bromine and then the haloform reaction. A detailed explanation
of these studies and the results of follow-up investigations into the
19
particulate nature of the trihalomethane precursor is given by Stevens, et al.
Additionally, Rook has described studies on the character of the trihalomethane
precursor and treatment for trihalomethane reduction.
-------
- 39 -
1.0
0.75
LU
o
lit
> 0.5
K
UJ
CC
0.25
0
-
L-
•
R
N
C
.
RAW WATER
NPTOC RANGE = 2.2 - 3.9 mg/l
COAGULATION AND
SEDIMENTATION BASIN EFFLUENT
DUAL MEDIA
FILTER EFFLUENT
Fig. 17. Relative IMon-Purgeable Total Organic Carbon Removal during
Water Treatment in the Pilot Plant.
-------
- 40 -
150
O)
2 100
UJ
O 50
z
O
O
UJ
CHLOROFORM
40 80 120
TIME, HOURS
DIBROMOCHLOROMETHANE
160
60
40
BROMODICHLOROM ETHANE
40 80 120
TIME, HOURS
160
A UNTREATED OHIO RIVER WATER
• COAGULATED AND SETTLED WATER
• DUAL MEDIA FILTERED WATER
CHLORINE DOSE: 5 mg/L
SEE TEXT FOR DISCUSSION
OF BROMOFORM
0 40 80 120 160
TIME, HOURS
Fig. 18. Trihalomethane Formation Potential in Various Qualities of Water.
-------
- 41 -
EPA pilot plant studies where chlorine was applied continuously at
various points within the treatment train further demonstrate the importance
of the role of coagulation and the point of chlorination in effecting reduced
trihalomethane concentrations in treated water. In one series of experiments,
river water was chlorinated (see Figure 19, point 1) then held for two days
to simulate off-stream storage (or a long raw water transmission line), then
it received either aluminum or ferric sulfate coagulation, flocculation,
sedimentation, and filtration through dual media. A finished water sample was
collected and stored two days at 25°C before analysis for trihalomethanes.
The raw water chlorine dose (10 mg/£) was sufficient to maintain a free
chlorine residual in the finished water sample for the two-day contact time.
After three days of operation in this mode of treatment, the point of
chlorination was moved to just prior to the coagulation/flocculation basin,
(Figure 19, point 2). As a control, a raw water sample was collected daily,
chlorinated in a closed container and likewise stored at 25°C for two days.
This provided a trihalomethane formation potential measurement. After each
change in point of chlorine application, two to three days were allowed for
the plant to stabilize.
When the raw water was chlorinated and stored before further treatment,
the terminal chloroform formation was maximized. However, when the point
of chlorination was moved to the coagulation/flocculation basin two events
began simultaneously — trihalomethane formation caused by the chlorine and
precursor reduction by the coagulant. The net effect is shown in Figure 19,
point 2. Points 3 and 4 show some additional improvement in reducing
terminal chloroform formation when the clarification step is allowed to proceed
further prior to chlorination. Further experiments will be conducted to
determine if ferric sulfate is consistantly more effective than alum for
trihalomethane precursor removal as these studies indicate.
-------
- 42 -
RELATIVE CHLOROFORM
FORMATION IN FINISHED
WATER AFTER 2 DAYS AT
GIVEN POINT OF CHLORI-
NATION COMPARED TO
CHLOROFORM FORMATION
POTENTIAL IN RAW WATER
1.0
.75
.5
.25
FERRIC
SULFATE
COAGULANT
ALUMINUM
SULFATE
COAGULANT
RIVER
DISTRIBUTION
*~ SYSTEM
2 DAYS RAW
WATER STORAGE
COAGULATION,
FLOCCULATION,
SETTLING
FILTRATION
Fig. 19. Chloroform in Finished Water Relative to Point of Chlorination.
(Pilot Plant Studies)
-------
- 43 -
2. Field Studies
The EPA Laboratory and pilot plant experimental results encouraged the
water utility of Cincinnati, Ohio to attempt to reduce the chloroform content
in their finished water by moving the point of chlorination so that clarified
water was chlorinated rather than raw water. Figure 20 is a schematic diagram
of the Cincinnati Waterworks. The City of Cincinnati pumps its water from
the Ohio River into two large uncovered reservoirs with approximately three
days retention time. For the past several years the practice has been to
add alum to the water going to these reservoirs and sufficient chlorine to
carry a free residual through the reservoirs and the treatment plant to the
extremities of the distribution system. A consultant was hired to conduct
the study and much effort went into collecting additional samples to insure
the bacteriological integrity of the system during the experimentation.
In mid-July 1975 the point of chlorination was moved from point A to the
headworks of the treatment plant, point B. 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 (T.U.) to approximately 2 T.U. as the water entered the
treatment plant.
The sharp decline in chloroform concentrations in the tap water following
moving of the point of chlorine application from Point A to Point B in
mid-July is attributed to the change in chlorination practice (Figure 20).
The change had little effect on the bromine containing trihalomethanes. During
the summer and early fall of 1975, other changes such as the addition of
small quantities of powdered activated carbon, variations in coagulants, and
moving the point of chlorination to follow filtration, were made, but none
of them had a significant observable effect on the chloroform concentration.
Some uncertainty exists regarding the total influence of all the treatment
-------
- 44 -
APPROX. 3 DAYS
OFF-STREAM STORAGE
POINT B
WATER
TREATMENT
PLANT
MIX
FLOC, SETT.
FILTERS
CHLORINE, ALUM
POINT A
itn
DISTRIBUTION SYSTEM
Fig. 20. Schematic of Cincinnati Waterworks.
-------
- 45 -
cjuu
280
260(
_, 240
O)
^•220
z"
fl 200
cc
£ 180
ui
Z 160
O
O
uj 140
Z
I 120
UJ
I 100
O
^ 80
E
"~ 60
40
20
KIP
IMr
RAW WATER CHLORINATED
AND STORED:
7 MOVE CHLORINATION FROM ,-«.„... -,.,«..• r^-n-n-ri A •
POINT A TO POINT B FORMATION POTENTIAL
- FIGURE 19, JULY 14, 1975 Q 4 DAY CHLOROFORM
/>^ FORMATION POTENTIAL
7 ^ A 6 DAY CHLOROFORM
FORMATION POTENTIAL
•* UHLUKOrOKM
A
O
©
-
-
_
BROMODICHLOROMETHANE
Ll X
', X DIBROMOCHLOROMETHANE
\AX^ A
^CA\ -/V xAr\ A- A"'
/\ X ^•a ~"lt f ^ ^L X" S'**^. ^'^ Jf'<**.
\~. *. - - _ • * *• *^ — — — — - -^ ^*N ' -"s.^^^ *^ ^ •+**•
\'~— -/ if'*- _^—^"\x>
JUL AUG SEP | OCT NOV DEC JAN | FEB MAR APR M
1975 1976
-
-
E
-
a _
_
-
r
i _
' X A^
^/"C^"' /_
AY JUN JUL
Fig. 21. Trihalomethanes in Cincinnati, Ohio Tap Water.
-------
- 46 -
changes because often only one or two samples were collected under each
*? n 9i
condition. Details on this experimentation have been presented. '
Other water utilities have also made in-plant modifications in treatment
9 ?
and monitored the subsequent effects. One water utility using lime softening
23
showed about a 75 percent reduction aid anotfier utility reported a substantial
reduction in the terminal chloroform concentration when chlorine was added
after recarbonation instead of at the headworks prior to the high pH of lime
softening. The importance of the pH control because of its effect on
11 19
trihalomethane formation has been reported. '
E. Summary of Trihalomethane Precursor Removal Studies
Experiments examining the prevention of trihalomethane formation were
conducted with a pilot water treatment plant having conventional coagulation and
sedimentation. The settled water was then divided between a dual media and
two different depth granular activated carbon (GAG) filters. Following
filtration, the effluent was disinfected with either chlorine, ozone, chlorine
dioxide, or combinations of these oxidants. Bench-scale studies with powdered
activated carbon (PAG) and aeration were also conducted. Table XI lists
the effectiveness of these various unit processes for reducing the chloroform
formation potential.
IV. Alternatives to Chlorination
Because the trihalomethanes are formed during the chlorination step, one
possibility would be to suggest an alternative disinfectant, such as
chlorine dioxide or ozone. Although chlorine dioxide prevents or minimizes
trihalomethane formation, the by-products of these reactions are not now (1976)
fully known. Mass spectrometric determination of these compounds is beginning
and hopefully health effects studies of the identified species will follow.
With precaution, therefore, in certain situations, chlorine dioxide may be
-------
- 47 -
TABLE XI
Effectiveness of Various Unit Processes for Reducing Chloroform Formation Potential
Chloroform Forma-
tion Potential
Chloroform
Formed
Process
Remarks
Aeration followed by
chlorination
66'
Coagulation, Sedimentation 48
and Dual-Media Filtration
followed by Chlorination
Coagulation, Sedimentation 48
Filtration/Adsorption by
Granular Activated Carbon
(5 min. contact time)
followed by chlorination
Powdered Activated Carbon 27C
added after Coagulation
and Settling followed
by Chlorination
Ozone only 48
Ozone followed by chlorination 48
66C
13
< 1
< 10
20a
9
None found
48
Diffused-air aeration with
air to water ratios up to
20:1 did not reduce
chloroform formation
potential (10 min. Contact
Time)
GAG would be effective for
3 weeks
GAG would be effective for
8 weeks
at PAC dosage = 8 mg/£
at PAC dosage = 100 mg/£
PAC contact time = 2-20 min.
0 neither forms
trihalomethanes, nor removes
precursors at disinfection
dises
Disinfection doses
(^ 1
Chlorine Dioxide only
74
Coagulation, Sedimentation and
Dual Media Filtration
followed by:
1. Chlorination -
2. Chlorine dioxide with
chlorine
< 1 C10_ does not form
trinalomethanes
1.3
and 1.5
3 Trihalomethane formation
decreases as the ratio of
C109 to Cl? increases
All tests performed on Ohio River water. Chloroform Formation Potential
is the amount of chloroform formed when raw water is chlorinated past break-point
and stored at 25°C for a specified contact time.
a - chlorine contact time = 48 hours
b - chlorine contact time = 96 hours
c - chlorine contact time = 22 hours
d - contact time for combination of chlorine dioxide with chlorine - 22 hours
-------
- 48 -
a promising alternative to chlorine because a disinfectant residual can be
maintained in the distribution system. This residual is reportedly comparable
24
to chlorine as a biocide, and if the concentration of excess chlorine used
to generate the chlorine dioxide can be kept low, the trihalomethane concentrations
in the finished water will also be low.
Some question exists concerning the health effects of ingesting chlorite.
25
Chlorite converts hemoglobin to methemoglobin, and the Norwegian Health
fy tr
Authority has recommended the absence of chlorite in drinking water.
Health effects studies on the toxicity of chlorite are underway within EPA.
At present (1976) no restrictions are imposed on the use of chlorine dioxide
for treating drinking water in the United States.
Other studies have begun to determine the inorganic end-products when
chloride dioxide is used in water treatment. For example, when 1.3 to 4.0 mg/£
chlorine dioxide was applied to dual media filtered pilot plant effluent
and stored for 3 days at 25°C, on a mg/£ as Cl basis, 50 percent of the
chlorine dioxide remained, 25 percent had become chlorite (CIO,,) , 9 percent
had become chlorate (CIO ) and 14 percent was reduced to chloride. The
unaccounted for 2 percent was assumed to be analytical error. These preliminary
\
results indicate that chlorite would be ingested by the consumer at concentration
depending on the original chlorine dioxide dosage. Further studies on chlorine
27
dioxide end-products are underway and will be reported on in detail.
Ozone is another effective disinfectant that does not produce trihalomethanes,
-------
but it fails to provide a residual disinfectant in the distribution system,
and, like the other alternatives to chlorine little is known about its
organic by-products. Unless some other means of providing the residual
disinfectant throughout the distribution system can be found, ozonation for
disinfection followed by chlorination to provide a residual will not eliminate
the trihalomethane concentrations reaching the consumer. Chloramines do not
19
form trihalomethanes to any great extent and one alternate possibility would
28
be to add chlorine and ammonia to provide a combined chlorine residual.
The microbiological quality of the finished product is the foremost consideration
in treating drinking water. Any disinfectant, therefore, must be an effective
barrier against problem organisms. Further, it should be economical to use,
but not at the expense of creating toxic or potentially toxic end-products or
by-products. Future studies evaluating disinfectants will have to include all
of these aspects.
-------
- 50 -
Conclusions
1. If trihalomethanes are already present in drinking water (i.e., there
exists an instantaneous trihalomethane concentration), diffused-air aeration,
powdered activated carbon, ozonation, or addition of chlorine dioxide are in
general, ineffective for reducing these concentrations. Granular activated
carbon will reduce the chloroform concentration for a few weeks, but will
adsorb bromine containing trihalomethanes for several months.
2. If trihalomethanes are not present in the water, but would be formed
upon chlorination (i.e., there exists a trihalomethane formation potential):
a. diffused-air aeration, powdered activated carbon, or ozonation are
in general, ineffective processes for reducing that formation potential.
b. chlorine dioxide, coagulation/sedimentation/filtration, or granular
activated carbon adsorption reduce the potential for forming chloroform
but are generally ineffective for reducing the potential for forming
bromine containing trihalomethanes.
3. Neither chlorine dioxide (without excess chlorine) nor ozone form
trihalomethanes, however, cost, residual disinfectant, end-products or organic
by-products, are considerations which must be included if these oxidants
are to be evaluated as an alternative to chlorine for insuring the bacteriological
safety of drinking water.
4. Water superintendents should evaluate their existing operations to
ascertain what improvements can be made in coagulation and settling to effect
better particulate removal in their plant. This might include considering
coagulant and flocculant aids, more frequent jar testing and thus varying
coagulant dosage, or even changing coagulants during certain seasons to
compensate for changes in raw water quality. These tests followed by a
determination of trihalomethane formation potential could guide future action.
-------
- 51 -
5. Water purveyors and design engineers should critically review their
chlorination practices on a case by case basis and determine if the point of
chlorine application can be moved further into the treatment train without
compromising the bacteriological safety of the drinking water. This is
particularly applicable at those utilities which are prechlorinating or
plan to prechlorinate off-stream stored raw water or long raw water transmission
lines. It must be recognized, however, that treatment modifications, other than
simply changing the point of chlorine application may be necessary to insure
a continual supply of safe and palatable water.
ACKNOWLEDGMENTS
Appreciation is expressedto the following individuals who provided valuable
assistance during these experiments: Mr. Carl Negli and his staff at the
Cincinnati Water Works pumping station for supplying the source water for
the pilot plant activities; Mr. Charles Bolton, Superintendent, Cincinnati
Water Works and his staff - Mr. Edward Kispert, Mr. George Hicks, and Mr. James
Ohleur; Dr. Riley Kinman and Mrs. Janet Richabaugh, University of Cincinnati
for allowing us to share the partial results of their study on chlorination
modifications at the Cincinnati Water Works: Mr. Raymond Taylor, Dr. Martin
Allen and their staff at EPA furnished microbiological support; Mr. Paul A.
Keller, Mr. Kenneth L. Kropp, Mr. Dennis R. Seeger, Ms. Clois J. Slocum and
Mr. Bradford L. Smith of the EPA Water Supply Research Division assisted in the
experiments and provided technical consultation; Mr. Gordon G. Robeck and
Mr. Alan A. Stevens furnished technical direction and manuscript review.
Special thanks is given to Ms. Maura M. Lilly and Ms. Virginia D. Maphet
for typing this manuscript and to Mr. Jesse M. Cohen who provided valuable
editorial and technical comments.
-------
- 52 -
1. Rook, J.J., "Formation of Haloforms During Chlorination of Natural
Waters." Water Treatment and Examination, 2^ (Part 2); 234 (1974).
2. Bellar, T.A., Lichtenberg, J.J. and Kroner, R.C., "The Occurrence
of Organohalides in Chlorinated Drinking Water . " Journal American
Water Works Association, ^:703 (1974).
3. 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 in
Drinking Water." Water Supply Research Laboratory and Methods
Development and Quality Assurance Laboratory, National Environmental
Research Center, USEPA, Cincinnati, Ohio, JAWWA, 67_ (11): 634-647 (1975).
4. Love, O.T., Jr., Carswell, J.K., Stevens, A. A. , Sorg, T. J. , Logsdon, G.S.,
and Symons, J.M., "Preliminary Assessment of Suspected Carcinogens in
Drinking Water - Interim Report to Congress. Appendix VI." USEPA
Report, Washington, D. C. , (June 1975).
5. Love, O.T., Jr., Carswell, J.K., Stevens, A. A. and Symons, J.M.,
"Treatment of Drinking Water for Prevention and Removal of Halogenated
Organic Compounds (An EPA Progress Report) . Presented at the 95th
Annual Conference of the American Water Works Association, June 8-12, 1975,
Minneapolis, Minnesota.
6. Love, O.T., Jr., Carswell, J.K. , Sterns, A. A., and Symons, J.M. ,
"Pilot Plant Studies and Measurement of Organics," Presented at 1975
Water Quality Technology Conference, American Water Works Association,
Atlanta, Georgia, December 8-10.
7. Dilling, W.L., Tefertilier, N.B. and Kallos, G.J., "Evaporation Rates
and Reactivities of Methylene Chloride, Chloroform, 1,1,1-Trichloroethane,
Trichloroethylene , Tetrachloroethylene and Other Chlorinated Compounds
in Dilute Aqueous Solutions," Environmental Science and Technology, 9^, (9):
833 (September 1975).
8. Neely, W.B. , Blau, G.E. and Alfrey, T. , Jr., "Mathematical Models Predict
Concentration-Time Profiles Resulting from Chemical Spill in a River,"
Environmental Science and Technology, 10, (1): 72 (January 1976).
9. Bellar, T.A. and Lichtenberg, J.J., "Determining Volatile Organics at
the ug/£ Level in Water by Gas Chromatography . " JAWWA, 66:739-744
(December 1974) .
10. Rook, J.J., "Haloforms in Drinking Water," JAWWA, 68^, (3): 168 (1976).
11. Stevens, A. A. and Symons, J.M., "Protocol for Measuring Concentrations
of Trihalomethanes and Their Precursors at Water Treatment Plants,"
Appendix IV of "Interim Treatment Guide for the Control of Chloroform
and Other Trihalomethanes," by Symons, J.M. , USEPA, Water Supply
Research Division, Cincinnati, Ohio (1976).
-------
- 53 -
12. Love, O.T., Jr., Carswell, J.K., Stevens, A.A. and Symons, J.M.,
''Evaluation of Activated Carbon as a Drinking Water Treatment Unit
Process." USEPA, Cincinnati, Ohio, 17 pp. mimeo, (March 3, 1975).
13. Basic Manual of Application and Laboratory Ozonation Techni^ues^,
p. 21, The Welsbach Corporation, 3340 Stokley Street, Philadelphia, Pa.
14. Palin, A.T., "Methods for the Determination in Water of Free and Combined
Available Chlorine, Chlorine Dioxide and Chlorite, Bromine, Iodine and
Ozone, Using Diethyl-o-phenylene diamine (DPD), J. Inst. Water Engr., .21,
537 (August 1967).
15. Clark, Robert M., Guttman, D., Machisko, J. and Crawford, J., "Cost
Calculations of Water Treatment Unit Processes," Water Supply Research
Division, Municipal Environmental Research Laboratory, U.S. Environmental
Protection Agency, Cincinnati, Ohio (March 1976).
16. Granstrom, M.L. and Lee, G.F., "Generation and Use of Chlorine Dioxide
in Water Treatment," JAWWA, 50: 1453 (Nov. 1958).
17. Miltner, R., "The Effect of Chlorine Dioxide on Trihalomethanes in Drinking
Water," M.S. Thesis, University of Cincinnati, August 1976.
18. Inhoffer, Wendell, R., "Use of Granular Activated Carbon at Passaic Valley
Water Commission," Proceedings, Third Annual AWWA Water Quality Technology
Conference, Atlanta, Ga., (1975).
19. Stevens, A.A., Slocum, C.J., Seeger, D.R. and Robeck, G.G., "Chlorination
of Organics in Drinking Water," Conference on the Environmental Impact
of Water Chlorination, October 22-24, Oak Ridge, Tennessee (1975).
20. Kispert, Edward, "Getting the Most out of Your Treatment Plant," presented
at the 96th Annual Conference of the American Water Works Association,
June 20-25, 1976, New Orleans, Louisiana.
21. Kinman, Riley, N. and Rickabaugh, Janet, "Study of In-Plant Modifications
for Removal of Trace Organics from Cincinnati Drinking Water," A report
prepared for the City of Cincinnati (1976).
22. Harms, Leland L. and Looyenga, Robert W., "Formation and Removal of
Halogenated Hydrocarbons in Drinking Water - A Case Study at Huron,
South Dakota," EPA Grant No. R008128010, Final Report (October, 1976),
U.S. E.P.A.
23, Singley, J.E., Beaudet, B.A., Brodeur, T.P., Thurrott, J.C. and Fisher, M.E.,
"Minimizing Trihalomethane Formation in a Softening Plant," Final Report,
EPA Contract No. CA6992948-A, MERL, Cincinnati, Ohio 45268.
-------
- 54 -
24. Benarde, Melvin, A., Israel, Bernard M., Oliver!, Vincent P. and
Granstrom, Marvin L., "Efficiency of Chlorine Dioxide as a Bactericide,"
Applied Microbiology, Vol. 13, No. 5, 776-780 (1965).
25. Musil, J., Knotek, Z. , Chalupa, J. and Schtnit, P., "lexicological Aspects
of Chlorine Dioxide Application for the Treatment of Water Containing
Phenols," Scientific Papers from Institute of Chemical Technology,
Prague, 8_, 327 (1964).
26. Myhrsted, J.A. and Samdal, J.E., "Behavior and Determination of
Chlorine Dioxide," JAWWA, 61, 205 (1969).
27. Miltner, R.J., "Measurement of Chlorine Dioxide and Related Products,"
In preparation to be presented at the IV Annual AWWA Water Quality
Technology Conference, San Diego, California (Dec. 1976).
28. Symons, J.M, "Interim Treatment Guide for the Control of Chloroform
and Other Trihalomethanes," U.S. Environmental Protection Agency,
Municipal Environmental Research Laboratory, Water Supply Research
Division, Cincinnati, Ohio (1976).
-------
APPENDIX 4
Stevens, A.A., "Determination of Chloroform Formation Potential in Water,"
To be submitted to the Journal of the American Water Works Association.
-------
MEASUREMENT OF TRIHALOMETHANE AND PRECURSOR
CONCENTRATION CHANGES OCCURRING DURING WATER TREATMENT
AND DISTRIBUTION
A.A. Stevens
and
J.M. Symons
Appendix 4
to
Interim Treatment Guide for the Control of Chloroform and other Trihalomethanes
-------
TABLE OF CONTENTS
INTRODUCTION 1
BACKGROUND 2
DEFINITIONS 3
METHODS 6
Measurement of Instantaneous Trihalomethane Concentrations. ... 6
Procedure for Instantaneous Trihalomethane Determinations ... 6
Measurement of Terminal Trihalomethane Concentrations and
Trihalomethane Formation Potential. . . 8
Effect of Time 9
Maintenance of Chlorine Residual 12
Effect of Temperature 14
Effect of pH 14
Loss of Volatile Species 15
Effect of Bromide or Iodide Contamination 16
Effect of Precursor Contamination 16
Procedure for Terminal Trihalomethan? and Trihalomethane
Formation Potential Determinations 17
EXAMPLES OF THE USE OF METHODS - INTERPRETATION OF RESULTS 19
Simple Chlorination 19
Conventional Treatment 21
Lime Softening. , 21
Granular Activated Carbon Filtration/Adsorption 26
Summary of Examples 30
SUMMARY 30
ACKNOWLEDGMENTS 32
-------
INTRODUCTION
Because of recent findings concerning the carcinogenicity of chloroform
and the confirmation of the ubiquity of chloroform in chlorinated drinking
2
water, many purveyors of potable water are interested in sampling their
product to determine the extent of their individual chloroform problems and
resolve them when possible. Additionally, as a direct result of the announced
U.S. Environmental Protection Agency policy regarding the initiation of a
3
"Voluntary Nationwide Chloroform Reduction Program," other water utilities are
anticipated to attempt to reduce concentrations of chloroform reaching the
consumer through modification of the treatment process.
In these treatment modification and surveillance programs, difficulties
often arise concerning what considerations should be made when selecting
sampling and analysis techniques to best evaluate the extent of the problem
or the success or failure of efforts to reduce that problem. This paper
discusses the necessary considerations for developing a method for evaluation
of treatment. The method is based on the physical and chemical factors
controlling production of chloroform and reviews the influence of these factors
on the concentrations of chloroform and other trihalomethane (TKM) species that
are observed in a sample at the time of analysis. These factors must not be
overlooked during planning of chloroform reduction projects. The need during
such studies for uncontaminated glassware, for head space free samples, and
for an adequate analytical technique for the THM measurements will be reviewed
245
only briefly because they have been discussed in detail elsewhere. ' "
-------
- 2 -
In addition to physical and chemical considerations, adequate bacterio-
logical monitoring of finished waters during chloroform reduction programs
must also be included. Any research program in which disinfection practice is
a variable requires careful attention on the part of the utility operators
to ensure that water of adequate microbiological quality reaches the consumer.
BACKGROUND
Chloroform results from the generalized reaction:
(Chlorine + "Precursor" -> Chloroform).
This occurs to some extent in any water treatment plant where chlorination for
disinfection is practiced. The reaction is not instantaneous and occurs over
a period of a few days until either chlorine or precursor is exhausted. '
In the presence of natural bromide, the reaction products include some mixed
halogen trihalomethane {THM) species (bromodichloromethane, dibromochloromethane)
r o
and bromoform. This occurs in most chlorinated drinking water, even where
bromide concentrations in the source water are small. Iodines-containing species
have also been observed, presumably because of the presence of natural iodide.
Because the chemical reactions for formation of these bromine- (and iodine-)
containing trihalomethanes are probably mechanistically similar to that for
formation of chloroform, the trihalomethanes, including chloroform, can be
discussed as a group* for treatment evaluations.
*The discussion of THM Species as a group should not be confused, however, with
the term "total trihalomethane'' (TTHM) that was used in the National Organics
Reconnaissance Survey (NORS) report. For certain purposes in the report TTHM
was calculated by converting the weight concentration for each of the THM species
(yg/£) to micromoles per liter by dividing each species' weight concentration
by the appropriate molecular weight and adding them together to obtain an index
of total trihalomethane. Whether or not the three bromine-containing trihalo-
methanes have the same, less, or more toxicological significance as chloroform
is not known. In addition, the factors that control the relative quantities
of these species obtained during treatment are not well known. Thus, much
information can be lost by reducing data to a single index. Nevertheless, the
decision of how to treat the data depends upon the treatment research project
goals and therefore may vary. Because the considerations for sampling and data
analyses for all of the trihalomethane species are the same, for simplicity,
trihalomethanes will be discussed in the group sense throughout this paper.
-------
_ 3 -
Because their formation is not instantaneous, THM concentrations increase
in the water as it flows through a water treatment plant (unless removed during
treatment) to reach some value higher than that which would be observed if
an analysis for THM species was performed immediately after sampling at the
first point of chlorination. Further, the consumer is likely to receive water
with THM concentrations higher than those leaving the plant because the
reaction proceeds in the distribution system. Additionally, not only are the
concentrations of THM time dependent, but the rate of the reaction is dependent
on pH, precursor concentration, nature of precursor(s), temperature, and to some
/o 10 "17^
degree free chlorine concentration early in the chlorination process. ' '
Finally, the ratio of chloroform to other trihalomethanes is highly dependent
f -to I (\\
on the bromide content of the source water. ' These factors are discussed
in detail below.
DEFINITIONS
Basic to this discussion are these three important definitions:
1. Instantaneous THM (InstTHM) Concentration: The concentration of
THM in the water at the moment of sampling. This may be expressed in terms
of the individual species or their sum as TTHM (see p. 2 footnote).
2. Terminal THM (TermTHM) Concentration: The concentration of THM
that occurs at the termination of the measurement of this parameter. To
measure TermTHM concentration, chlorine-precursor reaction conditions are
selected according to the treatment practiced at the particular plant being
evaluated. In general, a sample of water is chlorinated under these plant
conditions and chloroform and other THM species are measured after a specified
time period. This concept will be discussed in detail in the next two sections
of the paper.
-------
- 4 -
TermTHM concentration is equally important as a parameter for evaluating
consumer risk as is the InstTHM concentration, but because this parameter is
a measure of the sum of the amounts of THM species already present (instantan-
eous) and those formed during the reaction time, a third parameter must be
defined that is useful for evaluating unit process performance for removal
of unreacted precursor.
3. THM Formation Potential (THMFP); THMFP is measured as the increase
in THM concentration that occurs during the storage period in the determiantion
of the TermTHM concentration. The THMFP is obtained by subtraction of the
InstTHM concentration from the TermTHM concentration either when TTHM or when
the individual species data are used. THMFP is a measure of the port-ion of
the total precursor material of most concern to the consumer remain-ing in the
water at a given point in the treatment train. This parameter, when measured
on unit process influent and effluent samples, can be used to determine the
efficiency of that process for removal of that pertinent fraction of precursor
material.
The distinction between THMFP and a Total Precursor1 parameter is important.
Total precursor concentration i* the concentration of all organic THM precursor
materials present in the water that oould react with halogen species under
conditions that maximize the yield of trihalomethanes.
Because the identities of these organic compounds are not precisely known
at this time (1976), Total THM Precursor concentration could also be expressed
as concentrations of THM or concentration of TTHM obtained from that reaction.
However, no standardized procedure for measuring this parameter exists, and
considerable research will be required to establish the optimum conditions to
assure the complete reaction of all precursor(s) to yield maximum trihalo-
methane concentrations.
-------
Because the chlorination conditions for TermTHM concentration measure-
ment are somewhat less than optimum for THM formation, the TermTHM concentra-
tion in that determination obtained will be somewhat less than the theoretical
maximum THM concentration. Thus, the value obtained for TermTHM under these
conditions will yield by subtraction of the InstTHM concentration, a THMFP
smaller than the theoretical "total precursor" parameter. Although the value
obtained (THMFP) is not the total precursor concentration, it is an index of
the concentration of materials of most concern relative to THM formation at a
particular water treatment plant. Also, because controlling parameters
(under treatment plant conditions) are easily measured at the operating plant,
TermTHM concentration (and therefore THMFP) is a practical measurement.
Figure 1 presents the four parameters discussed above graphically.
InstTHM CONCENTRATION
THM FORMATION POTENTIAL
| | + y///\ TermTHM CONCENTRATION
| |
TOTAL PRECURSOR
FIG. 1. GRAPHICAL REPRESENTATION OF FOUR
TRIHALOMETHANE PARAMETERS
-------
METHODS
Measurement of Instantaneous THM Concentrations
For this measurement the reaction of chlorine with precursor materials
must be halted at the time of sampling with the goal being to measure only
trihalomethanes present at the time of sampling. In the WSRD laboratory and
in others, a small amount of reducing agent is added to the sample to react
with the chlorine and thus render it unavailable for oxidation or substitution
reactions. A small increase in trihalomethane concentrations upon storage
after addition of reducing agent is usually observed, and this is probably
caused by a slow hydrolysis of certain trihalo-irttermediates. The hydrolysis
step, of course, does not require the presence of chlorine. The distinction
should be made between this minor effect on the InstTHM concentration and
the continued THM formation reaction when no reducing agent is added. The
increase in THM concentration during storage after the addition of a reducing
agent has, in our experience, amounted to only a few percent of the total
value. The effect should be most noticeable at neutral pH because the hydrolysis
step would be accelerated at high pH and be near completion soon after the
intermediates are formed. Therefore, intermediates would be present in low
concentration when the reducing agent is added (little change would be observed
after dechlorination).
Procedure for InstTHM Determinations: Normally, the sample is taken,
2
head space free in muffled vials exactly as described in the NORS report
except that sufficient sodium thiosulfate is added to the vial prior to
filling to completely reduce any chlorine present in the sample. In the
WSRD laboratory 1 ml of 0.1N sodium thiosulfate for each 100 ml of sample has
been used successfully. This is 0.5 ml or about 10 drops for a 50 ml serum
•
vial. The amount used is not critical because on a stoichiometric basis this
is an excess of reducing agent, and the volume used is insignificant compared
-------
to the size of the sample. The filled bottle is then sealed with a teflon faced
septum held in place by a crimped top or screw cap and placed under refrigeration
to retard microbiological activity while the samples await analysis. This
procedure has been used routinely for about two years in this laboratory ' '
without any apparent problems. Potassium ferrocyanide, sodium sulfite, and
ascorbic acid, have also been used successfully as reducing agents for this
.. _. 11,12,13
application.
An alternate suggestion to obtain the InstTHM concentrations is to analyze
the sample immediately after sampling, providing a chlorine residual will not
interfere with the analysis, but this is often inconvenient or impossible.
14
Another suggested option was recently described , in which the THM species
are separated from the water sample upon sampling by sorption on a macroreticular
resin column. Whether or not the hydrolysis of trichloro-intermediates will
affect this result is not known, however.
The actual method of determination of the THM concentrations is not
critical, and acceptable procedures vary widely. At the WSRD laboratory the
2
method described in the NORS report has been used continuously in research
since that survey because of its convenience and reliability. For this
analysis, the sealed sample prepared as above is brought to 25°C prior to
opening in order to obtain reproducible purging efficiencies. A 5 ml aliquot
is then removed and transferred to a glass purging apparatus wherein the
trihalomethanes are stripped from the aqueous phase by passage of a flow of
helium upward through the sample. The trihalomethanes stripped in this manner
are collected on a sorbant porous polymer material contained in a stainless
-------
steel trap that is placed in series with the purging device. The tri-
halomethanes are then thermally desorbed from the trapping material onto
a gas chromatographic column. Finally, temperature programmed gas
chromatography is carried out, and the concentrations of trihalomethanes
are measured by use of a halogen^specific detector,
Measurement of Terminal THM Concentration and THM Formation Potential:
These two parameters are discussed together because, as mentioned above,
the measurement for TermTHM concentration together with the InstTHM
concentration yields the THMFP by subtraction. The TermTHM concentration
is measured by reacting chlorine with THM precursors in a given sample
under certain controlled conditions that affect yield and rate of formation
of the trihalomethanes and subsequently measuring the concentrations
THM species produced. Because for the reasons described earlier, this
is not a Total Precursor concentration raasurement, the selected
conditions for this measurement must be che same as those experienced
at the water treatment plant under study and be reproducible from sample
to sample. Critical conditions to consider are: (1) time of reaction
(time elapsed before halting the halogenation reaction with a reducing
agent), (2) maintenance of a free chlorine residual, (3) temperature,
(4) pH, (5) prevention of loss of the volatile products during the time
of reaction, and (6) avoidance of contamination of reagents. These
will be discussed below.
-------
- 9 -
Effect of Time: Although a single measurement of THM concentrations
after a storage period of several days in a bottle under appropriate conditions
can give a useful determination of TermTHM concentration for that specified
time, much more information can be gained from the reaction rate curve(s)
obtained by plotting THM concentrations vs time. ' ' The rate curve(s)
obtained by periodic measurement of THM concentrations of properly stored
finished water can be used to estimate the future THM concentrations at any given
time after water leaves the treatment plant. This is particularly important
when the goal of the analyst is to estimate ultimate consumer exposure to
THM at different points along the distribution system. The THM concentration
vs. time curve is especially useful where the utility has a large variation
in the time water is in various parts of the distribution system. The rate curve
can also be used to estimate THM concentrations at any given time after
water is taken from a sampling point within the plant when thepurpose is to
use the concentration obtained to calculate the THMFP at that point in
treatment for evaluation of unit process effectiveness.
In any system, it is preferable to generate a rate curve, at least
initially, so that the nature of the reaction that occurs at that location can
be determined. For example, Figure 2 shows two hypothetical curves describing
the rate of chloroform formation that might be expected for finished waters
of distinctly different quality after leaving typical water treatment plants.
Curves A and B in Figure 2 represent two extreme situations that might
occur. Although at time T, the chloroform concentrations are the same for
both waters, the short term chloroform concentration is greater in Plant A
and long term chloroform concentration is greater in Plant B. A Plant A
curve would be expected where chloroform formation potential is relatively
low but the precursor present is of the type that reacts quickly under the
given conditions (see ref. 8); that is, the final concentration of chloroform
-------
- 10 -
UJ
UJ
cc
o
LLI
O
z
o
o
CC
O
u_
o
cc
o
X
o
Inst.
CHCI3
CONC.
PLANT B
4-
PLANT A
1 1
0 T
T
1
T
1
2T
TIME (DAYS)
FIG. 2. FORMATION OF CHLOROFORM UNDER WIDELY DIFFERENT
TREATMENT PLANT CONDITIONS
-------
-Il-
ls reached early. A Plant B curve would be expected where chloroform formation
potential is high, but the reaction with chlorine is slow, because of nature of
precursor or reaction conditions. Thus, these curves are more informative
than a single chloroform determination performed at time T and the single
measurement at each plant could easily be misinterpreted to mean that the
plant situations were the same.
Good approximations of both curves are obtained by the selection of three
or four points beyond time 0 (instantaneous value) such as 1/2T, T, 2T as shown,
where 2T is equal to or slightly longer than the maximum distribution system
residence time. These added analyses could cause the generation of rate
curves to be time-consuming, especially if conditions are such that reactions
are slow and the distribution system residence time is long. If the development
of the rate curve is beyond the capability of a utility, the time for the
determination of TermTHM concentration should be the longest residence time in
the distribution system as this represents the most stringent condition for that
utility.
19 8,19
Preliminary results from recent work indicate that the kinetics
of the THM formation reaction may be able to be sufficiently well described
under a controlled set of conditions that a limiting value for TermTHM
concentration may be predictable by measurement of only a few values for THM
concentration early in the sample holding period. This approach to limiting
value calculations is still under investigation.
-------
- 12 -
Maintenance of Chlorine Residual: In conventional U.S. water tratment
practice, maintenance of a free chlorine residual throughout the distribution
system is often recommended or required. The continued reaction of precursor
with chlorine to yield trihalomethanes depends on the maintenance of a free
o
chlorine residual . Thus one of the prime conditions necessary for THM
formation is widely maintained. Again, with chloroform as an example, the "raw"
water curve presented in Figure 3 shows the abrupt cessation of chloroform
production as the chlorine became depleted. The 24-hour-and all later samples
gave the same chloroform concentrations, and chlorine residual determinations
confirmed the lack of chlorine. Thus the 24-hour-and later chloroform
concentrations could be misleading, assuming one of the conditions in the water
utility under investigation was maintenance of a chlorine residual throughout
the distribution system. For example, a single 48-hour chloroform determination
without an accompanying chlorine residual measurement would give a misleadingly
low Terminal chloroform concentration and chloroform formation potential.
Thus, for evaluation of systems where free chlorination is practiced, to assure
that these misleading results are not obtained, a chlorine residual measurement
must always be performed at the time of THM analysis to assure that a free
residual is present.
Work at the WSRD laboratory seems to indicate that TermTHM concentrations
are not affected significantly by the amount of free chlorine present. This
may be only because the concentrations usually are limited by the amounts of
precursors present. Some work has indicated that TTHM formation rates may
be dependent on free chlorine concentration where the reaction is not
12
precursor limited. In addition, Rook has published data showing an increase
in chloroform formation rate with an increase in chlorine dose. Since Rook
did not show TTHM concentrations, whether this was caused by an increase in
overall trihalomethane formation rate or simply a change in chlorine to
-------
— 13 —
o
III
J
E
CO
Q
• UJ
(0
3
<
X
X
UJ
' CM
O
^
•^
\
—
^>.
v -^
•s. t
' 1 1 1 i i ~ ~~ r — - — -L
5 x> o M A ' ' f
3"SSSi20l«
Mi-,2;-!Oh"inc\
^ i ^™ i^ ^_ —•• »^
1 /fill 'kl/*\l 1 WLJ 1 kl— U^M^ki-k IAI1 irf-k IJ^m *^k^. .^.
O
UJ
oc
UJ
J
u.
1 ^
<
1 C5
FRESH
|
1
1
1
\
\
\
V
c
o
-rCM ^
^ O)
E
00
z
O*~**
_ CO
s^
< 0)
5 >-
-§ o«'
1 1 ^
"- o>
15
9E z
.• OUJ
1
60
RAGE TIME, HRS
1 ON CHLOROF
C, pH-7 (STEV
0 Wo
h- H™ Lrt
w zS
UJ ,
5m
UJ
1- (0
0 mS
"CO "J Q
cc
1- UJ
°i
52
UjO
U.
UJ
O
5 CO
d
-------
- 14 -
bromide ratio cannot be determined (this effect of bromide concentration is
discussed below). Because of the uncertainty of the effect of chlorine con-
centration on reaction rate, the dose used in the TermTHM determination should
be nearly the same as that used at the treatment plant, and because that dose
is adequate for maintenance of a distribution system residual, it should be
adequate to supply the required residual for the duration of the test.
Effect of Temperature: Upon chlorination of a natural water approximately
twice as much chloroform can be formed in a given period of time at 25°C as is
Q TO
formed at 3°C, ' This range of temperature is not uncommon, summer to winter,
in U.S. surface waters. A need for close temperature control during the deter-
mination of TermTHM concentration is, therefore, indicated. Temperature is
largely seasonally controlled, and, for a given system, an estimated average
temperature of the distribution system is the logical choice for the controlled
reaction temperature.
Effect of pH; The trihalomethane formation rate has been shown to
o -I 9
increase with an increase in pH. ' This increase is expected because the
haloform reaction is base catalyzed. The selection of the pH for the controlled
reaction during the TermTHM concentration determination is not as straight-
forward as that for reaction time and temperature discussed above, however.
The variation of pH through an operating water treatment plant can be quite wide
and the variation is operationally controlled.
If the determination of the TermTHM concentration and the THMFP for the
finished water only is desired, pH selection is not a probelm. The samples
should be stored at the finished water pH, If, however, a comparison of the
THMFP of the finished water with that of the source (raw) water, or with
water at any stage of treatment to evaluate success of a unit process in
-------
- 15 -
reducing THMFP is desired, the selection of pH is more difficult.
The analyst must be sure that the same portion of the Total Precursor
Q
concentration (pH dependent) is reacting at each point, and that the reaction
Q
rate of chlorine with that material (also pH dependent) is the same at each
point. Thus, all of the samples from each of the various sampling points
must be chlorinated and stored buffered at a single selected pH value.
Therefore, because the THMFP test is designed to measure the portion of the
Total Precursor that is significant in a given water as it leaves the treat-
ment plant, the logical selection of the single reaction pH value is that of
the finished water entering the distribution system, as with the choice of
temperature.
This choice can lead to seemingly anomolus results, however. For example,
where pH is high through a unit process the THMFP at that pH might be higher
than THMFP for the same water measured at distribution system pH, Because
the reactions in the distribution system occur at a lower pH, reactions which
occur at the high pH at higher rates and involving a different portion of
the Total Precursor are not important to the consumer. Therefore, that
part of the total precursor should not be measured as THMFP and will not be
if the samples are buffered at the distribution system pH.
Loss of Volatile Species; To prevent misleading losses of trihalo-
methanes produced during the reaction period, the reactions must be carried out
in sealed, head-space free, containers. Container materials should be all glass
or glass with teflon lined caps. Standard glass stoppered reagent bottles
filled to overflowing so as to wet the stopper surface or the teflon septum-
sealed serum vials used for sampling for InstTHM determinations (see above)
have been found suitable.
-------
- 16 -
Effect of Bromide or Iodide Contamination; As mentioned earlier,
bromide or iodide present in the water can, as a result of first reacting
with chlorine, cause formation of THM species other than chloroform. In the
19
case of bromide, the velat-ive amounts of THM species formed has been shown
to be highly dependent on the bromide content of the water and the chlorine
dose, presumably because these determine the ratio of bromine to chlorine
available for competing reactions. Although the product ratios change, the
effect of bromide may be small when TTHM is calculated, however.
Preliminary work indicates that equal amounts of bromine and chlorine
substitution as trihalomethanes would be expected if the original bromide
19
concentration is as little as 2% of the chlorine dose. Clearly, any
bromide (or iodide) contamination of reagents used will cause a different
ratio of THM species to be formed than would normally occur on chlorination
of that water under plant conditions. Where individual THM species data
are used, this effect could be the cause of misinterpretations of data. For
example, if only chloroform concentrations are reported and reagents are
contaminated with bromide the TermTHM and THMFP tests would give lower
chloroform concentrations than those observed in the plant or in distribution.
However, measurement of all THM species would reveal that the bromine-containing
species were present in higher concentrations and the TTHM concentrations
approximately the same. As mentioned earlier, a change of chlorine dose
where bromide content of the water is constant could cause the same effect.
Effect of Precursor Contamination: In the WSRD laboratory distilled-
deionized- carbon filtered water has been used for "blank" water for reagent
preparation. At pH 7 the contribution of precursor in reagents has been
small- At higher pH, however, blank values tend to be higher. Care should
be taken to minimize volumes of reagents used in TermTHM measurements in
order to avoid this contribution to the THM concentrations obtained.
-------
- 17 -
Procedure for Terminal THM, and THM Formation Potential Determinations.
A test for TermTHM concentrations and THMFP can be standardized
in approach, but the conditions for sample treatment and storage, and therefore
the portion of the Total Precursor concentration measured, will vary
widely from system to system depending upon:
1. distribution system residence time
2. total chlorine demand of the sample
3. ambient temperature of the system
4. pH of the finished water in the particular system under investigation
as these variables must be chosen to match those in the system.
In work at the WSRD laboratory, a large (1-3 liters) sample of water
is collected and the pH is adjusted to that selected with an appropriate
inorganic (e.g., phosphate or borate) buffer. The final buffer strength is
about . 01M. The sample is then chlorinated, if needed, by the addition of
a previously standardized chlorine or hypochlorite solution. (Sufficient
chlorine is added at this time to maintain a free residual for the duration
of test period. Water leaving the treatment plant should already contain
sufficient free chlorine and the chlorination step is not needed). Several
sample bottles are filled and capped head space free - two bottles for each
point to be determined on the rate curve. For example, four bottles are
needed if only two points on the curve are to be determined, that is,
initial (zero time) and final values. Two more bottles are required to
determine each intermediate point. One of the "zero" time sample bottles
containr sodium thiosulfate to immediately reduce the chlorine so only the
InstTHM is measured. The other "zero" time sample has no reducing agent
and is used for measurement of the chlorine residual. This entire sequence
from sample collection to the capping of the bottles should be done as
quickly as possible to avoid loss of InstTHM during the manipulations.
-------
- 18 -
The samples, except "zero" times are then stored at the selected
temperature. After the preselected time, or times, if the reaction rate curve
is to be determined, one sample bottle is opened and an aliquot is transferred
by pouring into a smaller bottle containing sodium thiosulfate to prevent
further reaction of precursor with chlorine. This smaller bottle is then
quickly sealed head space free to await THM analysis. This measurement
determines a THM concentration for the respective time on the rate curve, and
the analytical procedure is exactly as described above for InstTHM determination.
A second bottle is opened at the same preselected time(s) and the chlorine
residual is measured. More details of this procedure are given in References
2, 4 and 8.
After the above determinations are completed the THM concentration
measured from the "zero" time sample is the InstTHM concentration; the THM
concentration from the bottle stored for the longest time is TermTHM
concentration. The difference between the InstTHM concentration and the
TermTHM concentration yields the THMFP.
A much simpler approach to the determination of "Free"(Inst.) and
20
"Potential" (Term) haloforms has recently been suggested by Nicholson.
Nicholson uses gas chromatograhy with direct aqueous injection of both
purged and unpurged samples. The introduction to the hot injection port
o fthe aqueous sample results in hydrolysis of halogenated intermediates
to yield trihalomethanes. Thus the direct aqueous injection method is said
to measure "Total Potential Haloforms," and subtraction of the pre-purged
sample value gives the "Free haloform" (Inst.) concentration. The method
does not, however, taken into account the effects of a free chlorine residual
over a period of days, nor has the relationship of the high temperature
hydrolysis to treatment plant conditions been established.
-------
- 19 -
EXAMPLES OF THE USE OF METHODS - INTERPRETATION OF RESULTS
Some hypothetical examples will help to demonstrate the use of the two
experimental determinations and the calculated THMFP to estimate both consumer
exposure to trihalomethanes resulting from the chlorination process and the
efficiencies of the various unit processes within the plant for removing
precursor compounds during treatment. The efficiency of unit processes for
removing chloroform or other trihalomethanes can also be estimated.
Simple Chlorination; The first example (Figure 4) represents the simplest
case — a water treatment plant with chlorination only. Figure 4 depicts
the relative values for the parameters that might be obtained if analyses
were conducted for the InstTHM concentration and TermTHM concentration for
source water (A), plant clearwell (B), and a theoretical point at the maximum
residence time in the distribution system (C). Note: Recall that for
simplification the trihalomethanes are being discussed here as a group.
Each bar could represent the single group index (TTEM), any one of the
individual species, or be subdivided horizontally into four bars of different
heights to represent all four commonly found trihalomethanes. According to
the bar graph, trihalomethanes were absent in the untreated source water
(InstTHM was not found on analysis of source water), but the full THMFP was
present and equal to the TermTHM concentration obtained experimentally. At
the clearwell, some of the precursor measured as THMFP had reacted to form
trihalomethanes (measured as InstTHM in finished water) leaving a smaller
remaining THMFP. The remaining THMFP, plus InstTHM concentration equals
the TermTHM concentration determined originally on the source water.
-------
20 -
o
UJ
O
z
o
o
InstTHM CONCENTRATION
THM FORMATION POTENTIAL
TermTHM CONCENTRATION
SOURCE
CHLORINE
CLEAR
WELL
k
END OF DISTRIBUTION
SYSTEM
FIG. 4. TRIHALOMETHANES FORMED DURING WATER
TREATMENT BY CHLORINATION ONLY
-------
- 21 -
At point C the entire original THMFP had reacted to give an InstTHM concen-
tration identical to the TermTHM concentration. No unit process exists at
this plant that was effective for reduction of either TermTHM or InstTHM
concentrations. The practice of chlorination itself converted THMFP to
InstTHM, thereby causing a reduction in the THMFP, In assessing the THMFP
removal by any unit process, care must be taken to separate removal of THMFP
by conversion to InstTHM by chlorination and removal of THMFP by the unit
process itself. Only at a point closer to the treatment plant than the
maximum length in the distribution system is consumer exposure to THM lower
than the TermTHM concentration shown in Figure 4.
Conventional Treatment: Shown in Figure 5, during conventional treatment with
raw water chlorination, some THM is formed during rapid mixing and throughout
the following treatment stages in the presence of chlorine. Thus, the
InstTHM concentration increases as the water passes through rapid mixing,
settling, and filtration — points B, C, and D. Coagulation and settling do
reduce THMFP (i.e,, precursor removal) so that parameter as well as TermTHM
concentration declines from point B to C, Filtration removes a little more
precursor material that is associated with the carryover floe so the THMFP
declines slightly again from point C to D. The remaining THMFP is converted
by chlorination to THM from point D to E and therefore the InstTHM concentration
determined for a sample taken at that point in distribution equals the TermTHM
concentration of the sample from point D,
Lime Softening; A case more complex than conventional treatment is illustrated
in Figure 6, This example treatment plant employs lime softening, recarbonation
after settling, and rapid sand filtration for treatment of the same water in
the example shown in Figure 5, Two alternate points of chlorination are
shown in Figure 6, The probable effect of raw water chlorination only is
-------
- 22
DC
I-
Z
LU
o
z
o
o
InstTHM CONCENTRATION
THM FORMATION POTENTIAL
TermTHM CONCENTRATION
IB
ISOURCEJK
CHLORINE J
COAGULANT-1
>
RAPID
MIX
-*|SETTLING|
m
m
END OF DISTRIBUTION
SYSTEM
FIG. 5. TRIHALOMETHANES FORMED DURING CONVENTIONAL
TREATMENT WITH RAW WATER CHLORINATION
-------
- 23 -
Z
O
UJ
O
z
o
o
5
ISOURCELp.
RAPID
MIX
SETTLING
CI2 LIME
f
C02
END OF DISTRIBUTION
SYSTEM
FIG. 6a. TRIHALOMETHANES FORMED DURING TREATMENT BY LIME
SOFTENING WITH RAW WATER CHLORINATION
InstTHM CONCENTRATION
THM CONCENTRATION
IA'
|SOURCEki»
^^^•H
c
c
U THM FORMATION POTENTIAL
U + ^TermTHM CONCENTRATION
^
//%'A yfy
IB' |c' D' IE'
RAPID
MIX
JFII TFRiU— •> FMn nr
SYSTEM
DISTRIBUTION
LIME
CO
FIG. 6b. TRIHALOMETHANES FORMED DURING TREATMENT BY LIME
SOFTENING WITH SETTLED WATER CHLORINATION
-------
- 24 -
shown in the upper bar graph (Figure 6a) 5 and chlorination after recarbonation
only is shown in the lower bar graph (Figure 6b)» The general explanation
for the bar graphs of Figure 6 is the same as that for Figure 5, except
that some additional interesting effects of treatments are demonstrated.
Point C (after settling, before recarbonation) shows the effect of
high pH on the reaction rate. In Figure 6a a large proportion of the original
THMFP has been converted to THM as compared to Figure 5 where the pH in
the settling basin was lower, Because the trihalomethanes produced through
the lime softening process are carried through the treatment process, the
TermTHM concentration measured after storage of samples D and E will include
whatever InstTHM concentrations were formed at the accelerated rate during
treatment. This point will be elaborated on immediately below and when
Figure 7 is discussed. The sum of THMFP and InstTHM concentrations (the TermTHM
concentration) has declined between B and C because the settling during lime
softening removes some precursor. From C to D, after recarbonation and sand
filtration the rate of increase in InstTHM concentration is slowed as the pH is
lowered so only a slight increase in InstTHM concentration is shown. Further,
the filters remove a small additional amount of precursor(s) that is associated
with floe particles, resulting in a slight decrease in the THMFP. The InstTHM
concentration increases with time at the expense of THMFP between D and E.
In the lower bar graphs in Figure 6b, because of the location of the point
of chlorination, no InstTHM concentration is shown until point D after
recarbonation and chlorination. The InstTHM concentration increases during
flow through the distribution system to Point E . Note that the values for
TermTHM concentration in the Figure 6b are lower at point BT through E' than
those observed for their respective counterparts (B through E) in Figure 6a.
-------
- 25 -
The first reason, as noted before, is that precursors that are other^
wise removed during softening and settling are converted rapidly to trihalo<-
methanes (shown as InstTHM) during the treatment process depicted in Figure 6a.
The trihalomethanes are then carried through the treatment process.
The second reason is more directly related to the high pH during treatment
and explains why the TermTHM concentration is higher at point B than at B'
and is also higher at B than that measured at point A. Precursor materials
that react at insignificant rates at the lower pH of the distribution system
and during the TermTHM concentration test conducted at points A, A and B
are converted rapidly to trihalomethanes during the treatment depicted by
Figure 6a. This contribution of InstTHM concentration to TermTHM concentration
at B does not appear in the TermTHM concentration measurement on water samples
from A, A' or B', These trihalomethanes are carried on to the subsequent
sample points.
On the other hand, the THM Formation Potentials at points C* and D
are slightly higher than those at the corresponding points in Figure 6a.
This is also caused by the rapid reaction of precursors at high pH. In this
case, precursor materials represented by the THMFP in the lower scheme (test
conducted at distribution system pH) reacted at a rapid rate during softening
to form InstTHM during the upper treatment process. That part of the THMFP
material was therefore no longer available to appear as part of the calculated
THMFP value on the upper bars C and D, This reduction of THMFP might be
expected to cause a reduction in TermTHM concentration through the upper
treatment train, but the lessening of C and D THMFP is exactly compensated
for by the corresponding increase in InstTHM concentration and therefore
does not reduce the TermTHM concentration in that treatment mode. Of
course, as noted above, settling and filtration does reduce the TermTHM
concentration.
-------
- 26 -
The above effects of the reaction of chlorine with precursor otherwise
reduced in concentration during the softening process and the high reaction
rate of chlorine at high pH with precursor unreactive at distribution system
pH combine to cause the TermTHM concentrations to have higher values when
source water chlorination is practiced, A more dramatic demonstration of
these effects is shown in Figure 7. Because the TermTHM concentration
determination is carried out at the distribution system pH, the InstTHM
concentration after softening could theoretically exceed the TermTHM
concentration established in the source water. This would occur if the
detention time at high pH through softening and settling was long enough.
Therefore, with sufficient contact time, the InstTHM concentration at the
process effluent could exceed the TermTHM concentration measured under distri-
bution system conditions at the influent to that process. This type of result
would not be inconsistent with the rationale for selecting testing conditions
described in this paper, however. Reduction of THMFP through the unit process
can still be directly calculated and that is the goal of the determination.
The increase of InstTHM through the process is also easily calculated from the
influent and effluent InstTHM concentrations,
Granular Activated Carbon Filtration/Adsorption; Figure 8 illustrates effects
that might be observed at points through a water treatment plant that employs
alum or iron coagulation, settling, and granular activated carbon (GAG)
filtration/adsorption. Two alternate points of chlorination are shown. Again,
as in the lime-softening case, the TermTHM concentration at C' in Figure 8b is
lower than the same parameter at point C in Figure 8a. Although the THMFP is
similar at point C' to that at point C, the TermTHM value at point C is higher
than at point C' because of the InstTHM concentrations resulting from the
chlorination reaction through treatment (A-C). Here, however, the
-------
27 -
llnstTHM CONCENTRATION
CUTHM FORMATION POTENTIAL
TermTHM CONCENTRATION
[SOURCE
r
-HIND OF DISTRIBUTION
SYSTEM
CHLORINE LIME
CARBON DIOXIDE
FIG. 7. TRIHALOMETHANES FORMED DURING TREATMENT BY LIME
SOFTENING WITH LONG DETENTION TIME AT HIGH pH
-------
- 28 -
THM CONCENTRATIOr
A
SOURCE^
D A DID
MIX
??
Yf
B
/S.
^
C
f* Af*
FILTER
^
^J
D
t
4
|E
> CKin r»c niCTDioi ITIOM
' CINU VJr Ulo 1 rilDU 1 lUPJ
SYSTEM
CHLORINE COAGULANT
CHLORINE
FIG. 8a. TRIHALOMETHANES FORMED DURING TREATMENT BY GRANULAR
ACTIVATED CARBON ADSORPTION WITH RAW WATER CHLORINATION
JCENTRATIOf
^,
O
O
2
X
1
|A'
SOURCEhHRAPID
' TlMix
••«
^ InstTHM CONCENTRATION
CUTHM FORMATION POTENTIAL
[ZD+E32 TermTHM CONCENTRATION
B'
•jJ*5FTTI IMCl
C
f
FILTER
D'
tlsv
I
I
E;
D OF DISTRIBUTION
STEM
COAGULANT
CHLORINE
FIG. 8b. TRIHALOMETHANES FORMED DURING TREATMENT BY GRANULAR
ACTIVATED CARBON ADSORPTION WITH POST-CHLORINATION ONLY
-------
- 29 -
instantaneous value at C was caused only by the chlorination in the treatment plant
of precursor that was present at points A, B, A', B' but was no longer present
when the TermTHM test was carried out on the chlorinated settled water at
C1 and was not influenced by a pH effect as noted previously. Because these
precursors were removed during settling they were not available to contribute
to the TermTHM concentration when sample C' was chlorinated.
The removal efficiencies of trihalomethanes and precursors vary widely
9 10
with time in service of GAG filters ' and the efficiencies illustrated for
this treatment step in Figure 8 are completely arbitrary. Note that although
a 50% removal of precursor (as THMFP) and of THM (shown as InstTHM concentration)
in the upper graph (Figure 8a) can seeminly be estimated (C minus D concentrations
divided by C concentration x 100)_, care must be taken in making such interpretations
if this was a real observation. THM is being produced during the time in a
unit process at a rate dependent upon the physical and chemical factors mentioned
previously and is removed at a competing rate according to the effectiveness
of the process in use. The net rate is what is actually being measured through
process. For example, from Figure 8a, between points C and D, whether some amount
of precursor THMFP between 0 and 50% reacted to produce THM that was efficiently
adsorbed or a true 50% removal of precursor (THMFP) occurred cannot be
determined. In other words, in the case of GAG adsorption systems determing
the separate adsorption efficiencies of THM and precursor in the mixed dynamic
system is difficult. At best, the effects can be described as 50% reductions
as opposed to removal in InstTHM, TermTHM concentrations and THMFP with no
connotation as to mechanism for this occurrence. This is not so complicated
in the example shown in Figure 8b where no InstTHM is formed until after the
filter adsorbs.
-------
- 30 -
Summary of Examples: These four hypothetical examples are not designed as
predictions of the success or failure of certain unit processes in a treatment
train or to be indications of the relative effectiveness of those processes.
The examples do serve to indicate the kinds of results that might be obtained
when a plant is sampled for measurement of InstTHM- and TermTHM concentrations
and THMFP and when the results of these measurements are compiled for unit
processes or whole plant evaluations. They also serve to illustrate the
complexity of the design of in-plant studies of treatment processes for
reduction of these parameters and some of the more important considerations
pertinent to analysis of data obtained as a result of those studies.
SUMMARY
Instantaneous Trihalomethane concentrations in chlorinated drinking water
may be measured in samples where chlorination reactions were stopped by addition
of a suitable reducing agent at the time of sampling. The trihalomethanes
are then separated from the aqueous phase and subjected to an acceptable form of
measurement.
Terminal Trihalomethane concentration is a measure of trihalomethanes
formed as a result of sample storage under conditions closely approximating
those of the distribution system corresponding to the plant under study.
The parameter can be used to estimate consumer exposure to trihalomethanes
as well as provide a route to the calculation of Trihalomethane Formation
Potential remaining at any stage of treatment.
Trihalomethane Formation Potential is a useful measure of pertinent
unreacted precursor material. The value is obtained by subtraction of
the Instantaneous Trihalomethane concentration from the Terminal Trihalomethane
concentration in a given sample.
-------
Total Precursor concentration measured as maximum trihalomethane
produced on chlorination is not a viable parameter because establishing
completeness of the reaction is rather difficult, and the measurement would
invariably give trihalomethane concentrations higher than those actually
reaching the consumer.
Generation of the trihalomethane formation rate curve, although not always
necessary, provides useful background information for plant and unit process
evaluations. The curve, when generated for finished water samples, provides,
a useful estimate of trihalomethane concentrations for any given Lime after
the water leaves the treatment plant.
The proper measurements of Instantaneous Trihalomethane and Terminal
Trihalomethane concentrations and calculation of Trihalomethane Formation
Potential in conjunction with a carefully planned sampling program can be
used to determine in-plant sources of trihalomethanes as well as to evaluate
whole plant and unit processes efficiencies in removal of precursors and removal
of trihalomethanes formed. The measurement can be used to determine success
or failure of efforts designed to reduce trihalomethane concentrations reaching
the consumer through modification of water treatment practice.
-------
- 32 -
ACKNOWLEDGMENTS
The authors wish to thank those who reviewed this manuscript for all
of the helpful suggestions. These reviewers were: J.K. Carswell, O.T. Love.,
J. DeMarco, G. G. Robeck, H.J. Brass and F. C. Kopfler. The authors also
express their appreciation to Ms. M. Lilly and Ms. P. Pierson who typed the
drafts and this version of the manuscript. The work of the entire organics
removal staff which made this writing possible is also acknowledged.
-------
- 33 -
LITERATURE CITED
1. Report on the Carcinogenesis Bioassay of Chloroform, Carcinogen Bioassay
and Program Resources Branch, Carcinogenesis Program, Division of Cancer
Cause and Prevention, National Cancer Institute.
2. 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.,
1975 National Organics Reconnaissance Survey for Halogenated Organics in
Drinking Water, Water Supply Research Laboratory and Methods Development
and Quality Assurance Laboratory, National Environmental Research Center,
USEPA, Cincinnati, Ohio; Jour. AWWA, 67^634.
3. Train, R.E., USEPA News Release, March 29, 1976.
4. Bellar, T.A. and Lichtenberg, J.J., 1974, The Determination of Volatile
Organic Compounds at the yg/£ Level in Water by Gas Chromatography, USEPA,
National Environmental Research Center, Cincinnati, Ohio, EPA-670/4-74-009.
See also:
Bellar, T.A. and Lichtenberg, J.J., 1974, Determining Volatile Organics
at the yg/£ Level in Water by Gas Chromatography. Jour. AWWA, 66:739.
5. Stevens, A.A. and Symons, J.M., 1975, Analytical Considerations for
Halogenated Organic Removal Studies. In: Proc. AWWA Water Quality
Technology Conference, December 2-3, Dallas, Texas, pp. XXVI-1,
6. Rook J.J., 1974, Formation of Haloforms During Chlorination of Natural
Waters. Water Treatment and Examination, 23: Part 2, 234.
7. Bellar, T.A., Lichtenberg, J.J., and Kroner, R.C., 1974, The Occurrence
of Organohalides in Chlorinated Drinking Water, Jour. AWWA, 66:703.
8. Stevens, A.A., Slocum, C.J., Seeger, D.R., Robeck, G.G., (1975),
Chlorination of Organics in Drinking Water. Presented at the Conference
on the Environmental Impact of Water Chlorination, Oak Ridge National
Laboratory, Oak Ridge, Tennessee, October 22-24.
9, Love, O.T., Jr., Carswell, J.K., Stevens, A.A., Symons, J.M., 1975,
Treatment of Drinking Water for Prevention and Removal of Chlorinated
Organic Compounds, An EPA Progress Report, Presented at 95th Annual
Conference AWWA, Minneapolis, Minnesota, June 8-13.
10. Love, O.T,, Jr., Carswell, J.K., Stevens, A.A. and Symons, J.M., 1975,
Pilot Plant Studies and Measurement of Organics. Presented at 1975 Water
Quality Technology Conference, American Water Works Association, Atlanta,
Georgia, Dec. 8-10.
-------
- 34
11. Kopfler, F.C., Melton, R.G., Lingg, R.D. and Coleman, W.E., GC/MS
Determination of Volatiles for the National Organics Reconnaissance
Survey (NORS) of Drinking Water, in ''Identification and Analysis of
Organic Pollutants in Water," 1st ed., Keith, L.H., Ed., Ann Arbor
Science Publishers, Inc., Ann Arbor, Michigan, 1975, Chapter 6.
12. Rook, J.J., (1976) Haloforms in Drinking Water, Jour. AWWA, 68, 168.
13. Kissinger, L.D., Fritz, J.S,, 1976, Analytical Notes - Analysis of
Drinking Water for Haloforms, Jour. AWWA, 68, 435.
14. Fritz, J.S., as reported in C&EN, April 12, 1976, p. 35. See also:
Stevens, A,A. and Kopfler, F.C., "Analyzing Drinking Water," C&EN,
June 21, 1976, p. 5.
15. Coleman, W.E., Lingg, R.D., Melton, R.G., Kopfler, F.C., The Occurrence
of Volatile Organics in Five Drinking Water Supplies Using Gas
Chromatography/Mass Spectrometer (GC/MS), In "Identification and
Analysis of Organic Pollutants in Water," 1st ed., Keith, L.H,, Ed.,
Ann Arbor Science Publishers, Inc., Ann Arbor, Michigan 1975, Chapter
21.
16. Gould, E.S., "Mechanism and Structure in Organic Chemistry," Holt,
Reinhart & Winston, New York, 1964.
17. Seeger, D.R., 1976, USEPA, Cincinnati, Personal Communication.
18. Bunn, W.W., Haas, B.B., Deane, E.R. and Kleopfler, R.D., 1975,
Formation of Trihalomethanes by Chlorination of Surface Water,
Environmental Letters, 1(3(3), 205-314 (1975).
19. Moore, L., 1976, USEPA, Cincinnati, Personal Communication.
20. Nicholson, A.A, and Meresz, 0., "The Occurrence and Determination of
Free and Total Potential Haloforms in Drinking Water," Presented at
the 27th Pittsburgh Conference on Analytical Chemistry and Applied
Spectroscopy, Cleveland, Ohio, March 1976.
iiUSGPO: 1977 — 757-056/5473 Region 5-11
-------
-------
-------