United States
Environmental Protection
Agency
Municipal Environmental Research
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-82-037 August 1982
Project  Summary
Alternative Water Disinfection
Schemes for  Reduced
Trihalomethane  Formation:
Volume  I.  Prototype Studies

Charles A. Sorber, Robert F. Williams, Barbara E. Moore, and Karl E. Longley
  The potential adverse health effects
of organohalides, particularly trihal-
omethanes, in drinking water are of
serious national concern. Trihalome-
thanes (THM) are formed primarily by
the reaction between organic precur-
sors  in the water and chlorine, the
halide most widely used for disinfec-
tion of potable water.
  The primary objective of this study
was to develop techniques to reduce
or eliminate THM in  finished water
without compromising the microbio-
logical quality of the  water. To this
end, a prototype rapid mixing system
was employed and a number of alter-
native  disinfection systems were
investigated. From these data, both a
general predictive model was devel-
oped and the role of algae and humic
acids as  THM  precursors was
evaluated.
  Results  of this study support the
contention that in disinfection prac-
tice a tubular plug flow reactor can be
used to advantage in reducing bacte-
rial populations in the water. An added
advantage is that use of this type reac-
tor will result in minimal THM forma-
tion  when chlorine is used as the
disinfectant.
  Experiments with standard plate
count (SPC) organisms and seeded
Escherichia co//showed that the disin-
fection systems of chlorine, chlorine
followed by ammonia, and  chlorine
dioxide were very effective when used
with hydraulic mixing. Chlorine diox-
ide was shown to be the best disinfec-
tant under the conditions of this study.
  Organisms used in this research var-
ied in their resistance to chlorination,
the least resistant being E. coli lys
147; next was £. coli C. SPC organ-
isms were more resistant than either
of the £. coli seed organisms, and bac-
teriophage f 2 was the most resistant of
all. SPC data were used to develop the
disinfection model presented in this
summary and the full report.
  An important finding was the sub-
stantial reduction in THM production
observed with the prototype system.
The lower chlorine doses required for
effective disinfection resulted in sig-
nificantly lower THM production. The
two best systems for minimizing THM'
production  were chlorine dioxide and
ammonia followed by chlorine, but
very substantial reductions of THM
concentration were achieved by chlo-
 rine followed by ammonia. For exam-
 ple, a 1.5 mg/L dose of chlorine alone
 produced 119 /ug/L of THM, whereas
 1.5 mg/L  of  chlorine followed by
ammonia produced about 7 //g/L of
THM. A dose of 1.5 mg/L of chlorine
 added after ammonia produced as low
 as 0.4/jg/L of THM, and 1.5 mg/L of
 chlorine dioxide produced 0.8 fjg/L of
THM.
   This Project Summary was devel-
 oped by EPA's Muncipal Environmen-
 tal Research Laboratory, Cincinnati.

-------
OH, to announce key findings of the
research project that  is fully docu-
mented in a separate report of the
same title (see Project Report ordering
information at back).

Introduction
  The use of chlorine as a disinfectant
in potable water has been based on its
long and successful history of rendering
water sufficiently free of pathogenic
organisms to minimize the probability of
waterborne disease among the  con-
suming public.  Because of  its  relative
ease of application and low cost, chlori-
nation has become synonymous for dis-
infection.  The  widespread   use  of
chlorine  has  been  one  of the  most
important public health control  mea-
sures of this century. Recently, potential
dverse health effects have been linked
to  organohalides,  particularly trihal-
omethanes (THM), in  drinking water.
These  organohalides are formed pri-
marily by the reaction between organic
precursors in the water and the chlorine
introduced into the water stream.
  The primary objective of this investi-
gation  was to  develop techniques to
reduce or  eliminate THM  and  other
organ chlorine compounds  in finished
water without compromising the micro-
biological  quality  of  the  water.  To
accomplish this goal, a  prototype  rapid
mixing system employing a plug flow
reactor was examined with several dis-
infectants. THM production and the dis-
infection efficiency resulting from the
use of alternative disinfectant schemes
were  compared with  those resulting
from the use chlorine.  The  alternative
disinfectant schemes  were  chlorine
dioxide, ammonia followed by the addi-
tion of chlorine, and chlorine  followed
by the addition of ammonia.
Procedure
Water Treatment Plant
  A new water treatment plant (WTP)
(at Boerne, Texas) located at a newly
constructed reservoir was the study
site.
  The treatment train of the 6056 mVd
(1.6 mgd) WTP consists of prechlorina-
tion, alum coagulation and sedimenta-
tion through  an   upflow  clarifier,
pressure sand filtration, and post chlori-
nation. Gas fed chlorinators are used for
chlorination, and facilities are available
for fluoridation. Sludge, filter backwash
water,  and  other plant  liquid  waste
streams  are piped to two evaporation
ponds for disposal.
  The  filtered, nondisinfected  water
was fed to a  1890-L (500-gal) storage
tank and pumped through the prototype
system at a f lowrate of 4.7 L/s (75 gal/
min) (Figure 1). Provisions were avail-
able for "seeding" the water stream
before  disinfection.  Seeding  require-
ments  for the prototype system  were
based on the levels of indigenousorgan-
isms found in the test water. Seeding
experiments were performed to evalu-
ate the prototype system since indigen-
ous organism levels were insufficient to
ensure  statistical  significance in the
evaluation of disinfection effectiveness.
  Aqueous chlorine, chlorine dioxide,
and  ammonia  were  peripherally  in-
jected into one of two plug-flow proto-
type mixers, which had internal throat
diameters of 20 mm (0.8 in.) and throat
lengths of 0.91 m (3 ft). Downstream
from each mixer, the water stream
passed through an energy recovery sec-
tion of 0.91 m (3 ft) into a 100-mm(4.0-
in.)  internal  diameter  pipe  (PVC,
schedule 80).  The total length of the
prototype system  was 37.36 m (122.6
ft). For a f lowrate of 4.7 L/s (75 gal/
min), the Reynolds numbers of the 20-
mm (0.8-in.) and 100-mm (4.0-in.) sec-
tions  are  370,000  and   74,200,
respectively. Studies involving adding

chlorine followed by ammonia or ammo-
nia followed by chlorine were done by
adding  the second chemical  into the
second mixer located 8.78 m (28.8 ft)
downstream from the first mixer. This
permitted  15-sec mean contact with
free chlorine  or ammonia before chlo-
ramine  formation  began. Total mean
contact time between the addition of the
disinfectant and the discharge of the
water  stream was approximately  60
seconds.
  Seed organisms (£. coli lys 147, £. coli
C, and phage fa) were added at the suc-
tion side of the system pump (organisms
were not introduced into the holding
tank).

Sampling
  Samples having  a  contact time of
approximately 1  sec were collected at
the end of the mixer throat, and samples
with a mean contact time of 60 sec were
collected at the discharge of the pipe
contactor. Samples with desired contact
times greater than 60 sec were col-
lected in glass or plastic containers, as
appropriate,   at  the discharge  point.
After the  predetermined contact time
had elapsed, a suitable reducing agent
was added to quench the disinfecting
agent. Additionally,  the sample col
lected at a sampling port immediate!
before the second mixer (the point o
ammonia or chlorine addition) allowei
the disinfectant residual and THM for
mation to  be  evaluated immediatel
before either chlorine or ammonia addi
tion for  the  chloramination  studies
Overall control samples were obtainei
from  a sample port immediately up
stream from the first mixer.
Bacteriological Analyses
  Bacteriological  samples  were  col
lected in sterile 500-ml polypropylene
bottles and placed in wet ice at 4°C unti
analysis. SPC organisms  were  mea
sured using the techniques in Standan
Methods.  Total  coliforms  (TC)  wen
determined by the membrane filtratior
technique (Standard Methods). The coli
form seed bacteria, E. coli lys 147, con
tained a phage that became active anc
lysed the bacteria after a 3 to 4 hour ex
posure at 42°C. These coliforms wer<
counted  by  a  technique . in  whict
plaques in an overlying £  coli laye
were enumerated  after 18  hours c
35°C incubation following the 42°C ex
posure.  Other plates without the £. co>
overlayer were heated to 42°C to lysi
the seed organism and then incubate*
for 48 hours at 35°C to determine SP(
less the £. coli lys 147. Bacteriophage f
was determined using soft agar overla'
plates with £ coli Hf r Hayes as the hos
organism.

Disinfectant Systems
   For all bacteriological and THM analy
ses, sodium thiosulfate was used t
quench the disinfectants at the desire
times. The quantity of reducing  ager
depended upon  the disinfectant dos
but was always in excess.
   Disinfectants were  introduced  int
the mixers through a vacuum ejectoi
and chlorine and-ammonia were fed a
gases. Chlorine dioxide was fed  as a
aqueous solution prepared by reactin
sodium chlorite with hydrochloric aci
in a contact  chamber filled with glas
beads (5-min detention time).
   Field  analyses  for residual chlorine
total chlorine, chloramines, and  chic
rine dioxide were performed by eithe
amperometric titration or colorimetri
determination  with  N,  N-diethyl
phenylene-diamine sulfate (DPD). Am
monia determinations in the field  wer
done colorimetrically  with   Nessler'
reagent.

-------
        Plug flow
   contacting sections
                                                         Test disinfectant
                                                    Pump
                                                                                                         Seed"
                                                                                                        storage
                                                                                                          tank
                                                                                           Filtered water storage
       Sample port (SP)

   Figure 1.  Schematic of prototype disinfection system.
 Chemical Analyses

  Temperature and pH were measured
 in the field at WTP sampling points. For
 prototype runs, these parameters were
 measured at the field location as well as
 after transport to the analytical labora-
 tory. All samples  for routine chemical
 analyses were collected in 1-L plastic
 bottles and stored  at 4°C for transport to
 the laboratory.
Trihalomethane Analyses

  Samples were collected in 40-ml
vials with (Instant THM) or without (Ter-
minal  THM)  sodium  thiosulfate and
sealed with Teflon®-faced* septums.
Instant THM samples were stored at 4°C
and Terminal THM samples were stored
at room temperature in the dark until
'Mention of trade names or commercial products
does not constitute endorsement or recommenda-
tion for use
analysis or until the designated times
for quenching by thiosulfate.
  THM samples were analyzed by gas
chromotography using the gas sparging
technique.  Samples   (5  ml)  were
sparged for 20 min at 25°C with a 30
ml/min flow of grade six helium carrier
gas and collected on a Tenax trap (30.5
cm  long; 0.318 cm OD; 0.293 cm ID).
Thermal desorption at 210°C with a 20
ml/min helium flow onto a 50°C analy-
tical Tenax column (60/80  mesh; 2.4
cm long; 0.31 cm OD; 0.216 cm ID) was
then  accomplished.  The  analytical
column was held  isothermally for  6
min  and temperature-programmed to
250°C at a rate of 10°C per min. Typical
retention times of 14.40, 16.38, 18.24,
and 20.02 min for chloroform, dichloro-
bromomethane, chlorodibromomethane,
and bromoform, respectively, were ex-
hibited for this analytical column. All
analyses were performed on a Varian
3700  gas  chromatograph  equipped
with a  flame ionization detector and  a
CDS 111C reporting integrator.

Results

Water  Treatment Plant

  Baseline  data collected during  the
normal operation of the WTP showed
little variability in chemical or bacterio-
logical parameters during the testing
period. The variations in operating con-
ditions were primarily because of prob-
lems associated with bringing the plant
on-line and developing experience to
provide  the most effective operation.
After this initial period of fluctuation
(approximately 6 months), the chemical
parameters showed almost no changes.
  Table 1 illustrates the ranges of bac-
teriological  and THM data observed at
the site since the sampling program
began in early 1979. All of the high coli-
form and SPC values were  observed
during the first 6 months of operation.
Since September 1979 there has been

-------
 Table 1.    Range of Bacteriological and Trihalomethane Data at Sampling
         Sites in Boerne WTP from February 1979 to August 1980

                            Total Coliform  Standard Plate Count   Total THM
Sampling Site
Reservoir (Site 1)
Pretreatment (Site 2)
Post-coagulation (Site 3)
Post-filtration (Site 4)
Post-chlorination (Site 5)
(cfu/1OOml)
<0.33 - 150
0.33 - 38
0.33 - 52
5-60
<0.33 - 4
(cfu/ml)
51 -9000
80 - 380
0.6 - 430
10- 1700

-------
             90
 c
 s
I

I
2
I
"5
B
             99
           99.9
          99.99
         99.999
        99.9999
                                       SPC
                         Dose (CLN) = 1.5 mg/L

                               pH = 7.6

                            Temp.  = 17.2°C
                                                      I
                                                               SPC
                                                          \
                                                                I
                                                                     I
                 0   10    20   30   40    50  60

               Contact time, seconds
                                                     70   0    20   40   60

                                                        Contact time, hours
Figure 3. Prototype disinfection experiment with chlorine followed by a gravimeter
          equivalent of ammonia after 15 seconds of contact time.
15-sec  flowtime  downstream of the
chlorine injection point (CLN). In both
experiments (CLN and NCL), the E coli
lys 147 population experienced a reduc-
tion  greater than  99.999% in 1 sec  or
less. The SPC organism population was
not  nearly as susceptible to  disinfec-
tion. A  dose of 1.5 mg/L of chlorine
followed by ammonia inactivated 99%
of the SPC organisms in 1 min, but 1.5
mg/L of chlorine  injected after ammo-
nia attained only 50% inactivation in 1
min. Both of these systems gave over
99.99% inactivation of  SPC organisms
in 24 hr. At a 5.0-mg/L chlorine dose,
however,  no  significant  differences
existed  between  the  SPC organism
                                        populations subjected  to  the  CLN  or
                                        NCL disinfection schemes.
                                          Disinfection results for  SPC  organ-
                                        isms  obtained in a single experiment
                                        using chlorine dioxide as the disinfect-
                                        ant at doses of 0.235, 0.390, and 1.49
                                        mg/L indicated that SPC organisms in-
                                        activation for  all doses was similar—
                                        90% inactivation was achieved in 1 min.
                                        After 24-hr  contact time,  however,
                                        99.3% inactivation was achieved with
                                        the 0.135 mg/L  CI02 dose, and greater
                                        than 99.8% inactivation was achieved
                                        with the 0.390 and 1.49 mg/L doses.
                                        Figure 4 depicts  a'single experiment in
                                        which 0.25 mg/L doses of chlorine di-
                                        oxide and chlorine were compared for
their efficacy as disinfectants. Chlorine
dioxide was the superior disinfectant for
SPC organisms achieving approximately
98.8% inactivation in 1 min  compared
with only 55% for chlorine. After 1-hr
contact time, the activation achieved by
chlorine dioxide  remained at 98% and
chlorine inactivation of the SPC organ-
isms  increased  to approximately the
same level.  Both disinfectants reduced
the seeded E. coli lys 147 populations by
greater than 99.999% in approximately
20 sec.
  The relationship between  disinfec-
tant  residual  and   SPC  organism
inactivation  was evaluated relative to
the residual's  interaction with contact
time.  Three  models were developed to
express the  residual: contact time rela-
tionship. When chlorine or chlorine fol-
lowed by ammonia were used in  the
disinfection scheme, the data were fit to
the models  separately using both free
and total residuals.
        Model A: y = b0Rblt       (1)
        Model B: y = b0(Rt)bl      (2)
        Model C: y = b0Rbnb2     (3)
where:  R = disinfectant residual, mg/L
        t =  contact time,  sec
        y = survival fraction of SPC
        organisms
        bo,  bi,  b2  =   regression
        coefficients
  The statistics  for  the  models  are
shown  in Table 2. Model A does not
describe the experimental data.  How-
ever,  Models B and C both provide signi-
ficant descriptions of the experimental
data as shown by associated correlation
coefficients  and Fisher (F) test values.
The correlation coefficient and Fisher
test value are both directly related to
error  in  model fit; however, both  are
shown  for  convenience.  For  those
experiments where chlorine or chlora-
mination was  used. Model B appears to
be superior to Model C since the expres-
sion,  (Rt)b , describes 40% to 63% of
"y";  whereas, the expression, Rbltb2,
associated with Model C describes only
16% to 40% of "y."
  A disinfection experiment with 4 to 6
x 106 pfu/ml of a virus seed (bacterio-
phage f2) was performed at two chlorine
doses. At the low chlorine dose,  1.5
mg/L, disinfection of the phage to  7
orders of magnitude took place in 1 min.
Disinfection of indigenous organisms
from  770 to 3.3 cfu/ml occurred in 15
sec.  At the higher chlorine dose, 5.0
mg/L, the phage was inactivated by  7
orders of magnitude within 30 sec and
the SPC was eliminated within 1  sec of
contact time.

-------
            90
 §
1
I
I
I
."*
i
I
          99.9
         99.99
        99.999
       99.9999
                                             o-
1 \ Symbol Organism     pH

H    O  SPC          7.5

 \\   •  SPC          7.6

 \\   A f. CG////S 147 7.5

  \\  A f. Colilys147 7.6
Temp. °C

  21.0

  23.5

  21.0

  23.5

Disinfectant

    C12

   CI02

    Clt

   C102
                                                         ft
               0    /O   20   30   40   50  60

               Contact time, seconds
                                                   70  0   2O    40   60

                                                       Contact time, minutes
Figure 4.   Prototype disinfection experiments with a 0.25 mg/L dose of chlorine
           dioxide compared with a 0.25 mg/L dose of chlorine.
Chemical Parameters—
  The same chemical parameters that
were determined for the normal opera-
tion of the WTP were determined for the
prototype operation.  Regardless  of
disinfectant system or dose, there were
no substantial differences between the
values obtained for the prototype system
and those obtained for the WTP. The
only chemical parameter that was sig-
nificantly affected by operation of the
prototype system, compared with that of
the WTP, wasTHM production. Compar-
ison of the THM  values with those
observed for the normal plant operation
show lower overall values for the proto-
                                        type system. Similar  results were ob-
                                        served  for  the chlorine  followed  by
                                        ammonia, ammonia followed by chlo-
                                        rine,  chlorine  dioxide,  and chlorine
                                        (lowest effective disinfectant dose) dis-
                                        infectant systems. In all cases, effective
                                        disinfection   occurred  and  lower
                                        concentrations of Terminal THM were
                                        observed.
                                          The THM production was also deter-
                                        mined for those prototype experiments
                                        in which seed organisms were present.
                                        No differences were observed in the
                                        presence or absence of microorganism
                                        seed.  Table 3 shows the Total THM
                                        values obtained in several disinfectant
systems. The duplicated data for chlo-
rine and chlorine followed by ammonia
at doses of 0.5 and 1.5 mg/L illustrates
the THM production observed without
added seed organisms.
  As with disinfection,  several models
were examined for THM production. The
model best expressing THM production
relative  to time and chlorine residual is
given by equation 4.

        Z = bo (I R,t) b1           (4)
              i = 1
where:  Z = THM concentration, //g/L
        R, = free chlorine  residual at
        end of contact  interval, mg/L
        t, = contact interval, sec
        M = number of contact
        intervals
        Bo' bi = regression coefficients
Application of test system data to  the
above model yielded equation 5.

        Z = 0.281  (I R,t,} 0466      (5)
              i = 1

The correlation coefficient for 25 data
points was 0.815, significant at greater
than the 99% level, and the regression
coefficient,   bi,  was  significant  at
greater  than the 99% level. Obviously,
use of the model is specific for the test
system from which it was derived.

Discussion
  Analysis of the chemical data for the
Boerne  Reservoir  indicates that  the
reservoir resembles a young, oligotro-
phic lake with  no significant  waste
inputs that currently affect water qual-
ity. Examination of the present and pro-
jected population and land use patterns
indicates that no significant industrial
or agricultural growth is expected in the
water shed and that future development
would   be primarily  residential  and
recreational.  Bacteriological data col-
lected   also  indicate  no  significant
human  pollution at this  time. Conse-
quently,  the THM  produced  upon
clorination  arose  only from  natural
organic  material that was delivered to
the reservoir by runoff.
  The overallTOC levels observed in the
reservoir  were  low,  thus relative
similarity  between  systems indicates
that the  rapid mixing techniques and the
addition of a biological seed does  not
appreciably  alter the chemical param-
eters characteristic of the water.
  The effect of the prototype rapid mix-
ing system upon the added bacteriologic
seeds (E. coliC,E. coli\ys 147, and bac-
teriophage f2> was pronounced. Opera-
tion  of  the prototype  without

-------
Table 2. Statistics Summary for Disinfection Models
Correlation Fisher
Data Model Regression Coefficients Coefficient Test
Source" Number b0 bi b2 r(R) F
CI2.-FAC A 3.66xW'3 6.223 — 0.163 0.527
CI2:TAC A 2.90x1Q-3 0.477 — 0.114 0.250
CLN.FAC A 9.10x10~A -1.114 — -0.430 2.043
CLN:TAC A 5.26x10'3 -2.681 — -0.649" 6.544*
CI02 A 7.50x1Q-3 1.659 — 0.279 1.353
CI2:FAC B 0.629 -0.301 — -0.587° 9.443C
CI2:TAC B 0.590 -0.275 — -0.558" 8.117*
CLN.-TAC B 0.447 -0.588 — -0.744" 11.179C
CLN.-TAC B 0.432 -0.481 — -0.611* 5.356*
CI02 B 0.214 -0.280 — -0.623C 10.1 39C
CI2:FAC C 0.398 -0.849 -0.277 0.731" 9.749"
CIZ:TAC C 0.376 -0.870 -0.240 0.693" 8.056"
CLN.-FAC C 0.163 -1.323 -0.357 0.943" 31.957"
CLN.-TAC C 0.354 -2.219 -0.165 0.981" 102.606"
CIOz C 0.222 -0.247 -0.278 0.673" 4.757"
a Disinfectant Systems: Clz - chlorine only; CLN - chlorine followed by ammonia;
CIOz - chlorine dioxide.
Residual Chlorine: FAC - free available chlorine; TAG - total available chlorine.
"95% level.
"99% level.
Table 3. THM Production from Several Disinfectant Systems
THM Production fag/L)
Disinfectant Disinfectant Total
Dosefmg/L) System CHCh CHCkBr CHCIBrz CHBr3 THM
0.00 None <0.1 <0.1 <0.1 <0.1 <0.1
0.25 Chlorine 1.06 1.80 1.43 <0.1 4.29
Chlorine Dioxide* 0.24 0.11 <0.1 <0.1 0.35
0.50 Chlorine 1.87 2.17 1.18 1.07 6.29
4.17 6.46* 8.22" 2.42" 21.6*
1.20 0.22 0.50 1.70 3.99
Chlorine + Ammonia* 0.99 0.87 0.61 <0.1 2.47
0.40 0.53 
-------
other bacterial species present in most
raw or potable waters. In this  study,
both E. coli C and F. coli lys 147 were
rapidly  inactivated within  15 sec after
exposure to all test disinfectants. Similar
inactivation kinetics  have  been
observed by others. The SPC bacteria,
on  the other hand, comprise a diverse
group of organisms that was found to be
considerably more resistant to disinfec-
tion during this study. Thus,  the use of
SPC bacteria appears to be more desir-
able in defining the efficacy of potable
water disinfectants.
  Of  importance  was  the substantial
reduction in THM production observed
with the prototype system. As shown in
Table 3, lower disinfectant doses pro-
duced lower Total THM levels for chlo-
rine. There was,  however,  less of  a
difference between Total THM formed
in the chlorine followed by ammonia
system (0.93 /jg/L to 10.24 fjg/L) than
was observed with the chlorine system
(6.29 fjg/L to 179.3 /jg/L) for the same
doses. The 15 sec of contact with free
chlorine before ammonia addition pro-
duced higher levels of Total THM when
compared with the addition of ammonia
followed by chlorine at the same dose
but the  15 sec  of contact  with free
chlorine  also showed better   disin-
fection.  Thus, the ammonia addition
was able to prevent additional THM
from being produced.
  The two best systems for minimizing
THM production were chlorine dioxide
and ammonia followed by chlorine. The
low level THM observed  for chlorine
dioxide may have  been due to low level
chlorine contamination in the system
produced  because of  the  method  of
generation  of  the chlorine dioxide.
Significant reduction of THM were ob-
served with  chloramine  and chlorine
dioxide  as  the disinfectants. A 17-fold
reduction in THM production was ob-
served at a 1.5 mg/L dose of chlorine
followed by ammonia when compared
with the same dose of chlorine alone. A
275-fold reduction in THM production
was observed for ammonia followed by
chlorine under the same conditions.
  The prototype  system was operated
with two chemical seeds to determine if
the rapid mixing system affects the THM
production in a  predictable manner.
Three doses of humic acid and one dose
of  chlorophyll-a  were  added to the
prototype  system during  the  study
period.   Both  seed  systems  show
increased THM levels over the back-
ground  indicating that these materials
do  contribute to  the THM production.
The smaller increases of THM observed
for the chlorophyll-a  experiment  are
consistent with the  low level of seed
used. Experiments with humic acid did
not exactly parallel  either the normal
plant operation nor the static laboratory
experiments with regard to time course
of THM production. The prototype sys-
tem did appear to affect the THM pro-
duction.1  This observation  may   be
explained by the kinetics of the reaction
of the disinfectant with the THM pre-
cursors. If diffusion of the disinfectant
to  the  organic precursor  is  rate-
controlling, increasing the mixing rate
will increase the reaction rate up to a
plateau value. If the reaction is not con-
trolled by diffusion,  however, the reac-
tion rates will be independent of mixing.
For the present system, the THM pro-
duction is slower than that observed in
other systems. This can be interpreted
as a build-upof higher concentrations of
intermediate chlorinated compounds by
achieving the diffusion plateau. These
intermediates then  break down  more
slowly to produce THM because of self-
interaction or then  follow other path-
ways to materials other than THM. At
long disinfectant residence  times,  the
levels of THM are comparable between
the  laboratory and  prototype experi-
ments.
Conclusions
 1. Chlorine dioxide was a better disin-
   fectant  than chlorine for  standard
   plate count  (SPC) organisms in the
   prototype rapid mixing system.
 2. For  contact times greater than 1
   hour, the chlorine followed by am-
   monia disinfection system proved to
   be a better disinfection system than
   the  chlorine dioxide  and chlorine
   disinfection  systems on a gravimet-
   ric basis.
 3. Bacteriophage f2 was  effectively in-
   activated with  chlorine under the
   rapid mixing conditions used.
 4. Highly turbulent mixing of chlorine
   followed  by  ammonia  addition
   resulted in reduced THM levels (10-
   20//g/L with THM formation poten-
   tial of approximately 130//g/L) with
   effective SPC organism reduction.
 5. The prototype mixing  system effec-
   tively  reduced  THM   initially and
   maximized  disinfection. Prolonged
   contact  chlorine  produced  addi-
   tional THM  if a free  residual per-
   sisted.
 6. Soluble  humic  materials may  be
   responsible for a significant portion
   of THM formation potential in the
   Boerne reservoir water.
7. Chlorophyll-a was responsible for a
   small but real portion of THM forma-
   tion  potential   in  the  Boerne
   reservoir water.

Recommendations
 1. The chlorine, chloramination, anc
   chlorine  dioxide  disinfection  sys-
   terns  evaluated  for  bacteria
   disinfection efficiency in this stud\
   should be further evaluated for virus
   disinfection efficiency.
 2. The effects of mixing on mactivatior
   of virus should be evaluated for dif
   ferent disinfectant systems.
 3. Specific  mixing parameters shoulc
   be  evaluated  to  determine the
   appropriate geometry and  energy
   input parameters applicable to en
   gineering design of the mixing sys
   tern; this would  permit design of  £
   cost effective mixing system.
 4. Isolation of  endogenous  organic
   materials in raw reservoir water anc
   the examination of their THM forma
   tion potentials is needed.
 5. The effect of higher added doses oj
   chlorophyll-a on  the production  o'
   THM should be evaluated.
  The full report was submitted in fulfill
ment of Grant No. R-806046-01 -0 ty
the Center for Applied Research anc
Technology,  the University of Texas a
San  Antonio, under sponsorship of the
U.S.  Environmental Protection Agency
                                   8

-------
Charles A. Sorber, Robert F. Williams, Barbara E. Moore, and Karl E. Longley
  were with the Center for Applied Research and Technology. The University of
  Texas, San Antonio, TX 78285. (Sorber and Longley are presently with the
  College of Engineering. The University of Texas, Austin, TX 78712 and Strauss
  and Roberts are with Consulting Civil Engineers, Inc., Porterville, CA 93257.)
Gary S. Logsdon is the EPA Project Officer (see below).
The complete report,  entitled "Alternative Water  Disinfection Schemes  for
  Reduced Trihalomethane Formation: Volume I. Prototype Studies," (Order No.
  PB 82-227 471; Cost: $ 12.00, subject to change) will be available only from:
       National Technical Information Service
       5285 Port Royal Road
       Springfield,  VA 22161
        Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
       Municipal Environmental Research Laboratory
       U. S. Environmental Protection Agency
       Cincinnati, OH 45268
                                                                                                 *USGPO:1M2-559-092-460

-------
o  »-•
T>  Z20
CO  <:o
        ru
  Oi-TJ-O
oxxo
9-:
o
  m
       m
       Tt
       O
                         CD
                         T



                         1
                         O
                         *H

                         O
                       9.5
                       il
                       0) 2
                       2. 0
                       0 =
                       VI
                       N>
                       a>
                       oo
                m > TJ m
                CO O O  O -
                W< S. 3 c

-------