P- t 7/6 ".
CHEMICAL, MICROBIOLOGICAL, AND MUTAGENIC EFFECTS OF USING
 ALTERNATIVE DISINFECTANTS  FOR  DRINKING  WATER  TREATMENT

                            by

                 Benjamin W. Lykins, Jr.
                   Kathleen M. Schenck
            Drinking  Water  Research Division
          Risk  Reduction  Engineering Laboratory
           U.S.  Environmental  Protection  Agency
                 Cincinnati, Ohio    45268
                    Wayne E. Koffskey
                Department  of Public  Works
                     Jefferson Parish
               Jefferson, Louisiana   70121

                           and

                      Mark  H. Griese
                Water Quality and Research
             Evansville Water 4 Sewer Utility
                Evansville, Indiana   47740
                 The University of Kansas
        42nd Environmental Engineering Conference
                     Lawrence,  Kansas
                     February 5, 1992
           RISK  REDUCTION  ENGINEERING  LABORATORY
            OFFICE OF RESEARCH AND DEVELOPMENT
           U.S.  ENVIRONMENTAL PROTECTION AGENCY
                  CINCINNATI,  OHIO  45268

-------
          CHEMICAL,  MICROBIOLOGICAL,  AND MUTAGENIC EFFECTS OF USING
            ALTERNATIVE DISINFECTANTS FOR DRINKING WATER TREATMENT
                            Benjamin  W.  Lykins,  Jr.
                                     Chief
                      Systems  and  Field  Evaluation Branch
                              Kathleen  M.  Schenck
                                   Biologist
                      Systems  and  Field Evaluation  Branch
                       Drinking Water Research Division
                     Risk Reduction Engineering Laboratory
                     U.S. Environmental Protection Agency
                            Cincinnati,  Ohio   45268
                               Wayne  E.  Koffskey
                                 Chief Chemist
                Jefferson  Parish  Department  of  Public  Utilities
                      Jefferson Parish,  Louisiana  70121
                                      and
                                Mark H. Griese
                                    Manager
                          Water Quality and Research
                      Evansville Water And Sewer Utility
                         • Evansville, Indiana  47740
INTRODUCTION
     Future Federal regulations for disinfection/disinfection by-products will
potentially effect most water treatment plants in the United States.  In antici-
pation of proposing regulations in  June 1993, the U.S. Environmental Protection
Agency is in the process of assessing the  performance and potential health risks
of several drinking water  disinfectants and their by-products.  In order to pro-
vide field data for this assessment and to more thoroughly understand the inter-
actions of disinfection and disinfection by-product control  regarding the advan-
tages and disadvantages of a  selected direction,  studies were done at two drink-
ing water utilities. This paper will discuss some of the results from these stu-
dies which consisted of a  pilot plant evaluation at Jefferson Parish, Louisiana
and bench-scale evaluations at Evansville, Indiana.

REGULATIONS

     Existing disinfection/disinfection by-product regulations apply only to tri-
halomethanes with  a maximum  contaminant  level (MCL)  of 0.10 mg/L.   Also, this
regulation only applies to systems serving more than 10,000 people.  Possible

-------
future disinfection/disinfection by-product requirements will apply to all public
water systems including non-community systems.  Compounds  likely to be regulated
with MCLs when these regulations become effective are:

     •    Trihalomethanes    (chloroform,    bromodichloromethane,
          chlorodibromomethane, bromoform)
     •    Haloacetic acids (trichloroacetic acid, dichloroacetic acid)
     •    Chloral hydrate
     •    Bromate
     •    Chlorine, chloramines, chlorine dioxide, chlorate, chlorite

     Over twenty compounds have been considered for regulation and those excluded
at this time were done so for basically two reasons:   (1) insignificant health
risk at levels that occur in drinking water such as haloacetonitriles and chloro-
picrin, and  (2) health risks not adequately characterized in time for the reg-
ulation such as  the  aldehydes.  As far as  the  tri hal omethanes  are  concerned,
there are several options being considered.  One option is to develop an MCL for
each of the  four  trihalomethanes.   Other  options include:   (1)  an MCL for the
total trihalomethanes or, (2) an MCL for each of  the trihalomethanes and an MCL
for the total trihalomethanes.

     The  reasons  that these  trihalomethane  options  are  being  considered are
because the  health risks for the individual  trihalomethanes are significantly
different, technical feasibility for limiting their formation can vary for each
compound, and hopefully,  the  trihalomethane regulation can act as a surrogate to
limit other halogenated by-products. Traditionally, drinking water  standards for
contaminants are set at the lowest possible number which is technically and eco-
nomically feasible.   However,  one  has  to  ensure  that  drinking  water  is micro-
biologically safe which  may  mean that  a greater  risk  will have to be accepted
from disinfectants and disinfection by-products than for other contaminants.  An
example of this is shown  in Figure  1 where risk trade-offs have to be considered
between microbial control and  production of disinfection by-products.

ALTERNATIVE  DISINFECTANT STUDIES

Jefferson Parish, Louisiana

     For this study,  raw  river water from the Mississippi River was  pumped to the
full-scale  plant where it was  clarified with diallyldimethyl ammonium chloride
or dimethyl amine-type cationic polymers.  After fluoridation, but before disin-
fection a portion of the  clarified water was filtered through pressure sand fil-.
ters before going to a 10  gallon-per-minute pilot plant.  This pilot plant was
operated  in a conventional mode with disinfection followed  by sand filtration.
Post  treatment  consisted of granular  activated  carbon after  sand filtration.
Figure 2  shows  the flow  diagram for this mode of operation.

      Each disinfectant contact chamber was 12.75 inches  (320 mm) outside diameter
and  constructed of stainless  steel.   The chlorine,  chloramine,  and chlorine
dioxide contact chambers were 10 feet (3m) in height, producing approximately 30
minutes  disinfectant  contact time  with a  flow of 2 gpm (0.13 L/s).   The ozone
contact  chamber was  11  feet (3.3m). in height,  with  countercurrent  operation
consisting  of water entering at  the top of the contact chamber and ozone gas

-------
CD
D)
                  .2 .2
               T  o "cS
                  CD C
                  **•" T"
                  = O
                  w IE
                  00
                             CO
                            •*-*
                             o
                             o
                            ol
                             i
                            CO
                             c
                            ,g
                            *•+-«
                             o
   CO
O
0)
                            CO  CO
                            bcr
                             CD

                             g>
                            LL

-------








V
1
1
I
1


c
— Js
5 .52
ZO


t


CM
I



f
1
CM





t


o



t
I
CO
O


t
_J

^tf ^-
«-g
§1
VM/^ ^^J



^^j k*
s|
O/T
O




^vd? ^™
«•§
§|
Of*





s|
§5
v^ ^j









































73 »-
03 ~
COU.





^3 ^m
!•• 0^
COU.






73 ^
f. fl)
"3 -t—
CO r=
COU.





C 0)
<^Q!




eft
doil




































O
<.
•
O





o
o






o
o





o
o




o
0


J2
Q_
CL

<
 c3
CL

 O
 CO
 CD
 O
 OJ

 CD

 O
CO


 o
[E

cvi
u_

-------
entering at the bottom.   The water and ozone gas influent lines were oriented so
that  the  influent water  would be  in  contact with  the ozone gas  stream for
approximately 30 minutes.

     Ozone was generated from compressed dry air using  an electrically powered
ozone  generator with  a maximum  output capacity  of  0.25  Ib/d  (0.11  kg/d).
Chloramine was produced  by adding  hypochlorous acid  to the  water stream a few
seconds prior to the addition  of ammonium hydroxide.  Chlorine dioxide was gen-
erated using two solutions,  one containing  sodium chlorite and sodium hypochlo-
rite and the other containing sulfuric acid.  The final pH of the solution was
4. Chlorine was added in  the form of hypochlorous acid.   Residual concentrations
after disinfection addition were maintained through the contact chambers to be
comparable to those found in many drinking water utilities at the  time the study
was initiated.  Average residuals were 1.0 mg/L chlorine, 2.1 mg/L monochlora-
mine, 0.5 imj/L chlorine dioxide, and 0.5 mg/L ozone.   The nondisinfected process
stream was identical to the disinfected process streams except for elimination
of the contact chamber.  The major objectives of this study were to evaluate: (1)
the microbiological effectiveness  of the disinfectants, (2) the control of halo-
genated by-products, and (3) potential health effects associated with the use of
these disinfectants.

Microbiological Effectiveness

     All of  the disinfectants at  the  dosages used,  reduced the col i forms to
acceptable levels.   However,  with heterotrophic bacteria,  all  disinfectants
except chloramines reduced the levels to below 100.  Other organisms may not have
been as effectively controlled by all disinfectants.   For instance,  during one
study  when  MS2 Coliphage was  spiked into  the pilot  plant, chloramines  were
ineffective  for this  virus  indicator as shown in Table  1.   Future  studies at
Jefferson  Parish  will   apply  the  concentration  x   time  (Ot)  concept  for
determining disinfectant efficiency.  An example of these C»t values that have
been developed for  inactivation of various microorganisms by the major disin-
fectants is shown in Table 2.

Ha'logenated By- Product Control

     During the one-year operation of the pilot plant,  two surrogate parameters
that  give  an indication  of  organic concentrations  including halogenated by-
products were evaluated:  total organic  carbon (TOC) and total  organic halide
(TOX).   The  average  TOC concentrations in  the disinfectant contact  chamber
effluents were 3.1,  2.9,  3.2,  3.2  and  3.2  mg/L for the nondisinfected,  ozone,
chlorine dioxide,  chloramine, and chlorine process streams, respectively.  There
was  some  indication that ozone was reducing the TOC  concentration  while the
concentration of TOC was fairly constant for the other disinfectants.  Further
in the treatment system after the ozone  sand column, the TOC was reduced an addi-
tional 0.8 mg/L when compared to the nondisinfected system.  Based on the levels
of heterotrophic bacteria in the effluents of the ozone contact chamber and ozone
sand column,  the  reduction of  TOC across the ozone contact  chamber  appears to
have resulted primarily from oxidation while that across the ozone sand column
can be attributed to biodegradation.
     With an average nondisinfected influent concentration  of  25 (tg/L, TOX con-
centration increased significantly after 30 minutes of disinfectant contact time

-------
                TABLE  1.  JEFFERSON  PARISH MICROBIOLOGICAL DATA

                                   MS2
                    RESIDUAL   COLIPHAGE     COL IFORM       HPC
STEAM                (mq/U       (ml)        HOP ml)       (ml)
NONDISINFECTANT      '  -        400,000         640      43,000
OZONE                 0.3          <1             0          14
CHLORINE DIOXIDE      0.5   .        1             0          57
FREE CHLORINE         1.0          <1             0          44
CHLORAMINE            1.9        46,000           0         270

-------
           TABLE 2.  SUMMARY OF C«t VALUE RANGES FOR INACTIVATION OF
                VARIOUS MICROORGANISMS BY DISINFECTANTS (2'3'4)
Disinfectant
Micro-
oraanism
L. Coli
Polio 1
Rotavirus
Phage f2
G. lamblia
cysts
G... muris
cysts
Free Preformed
Chlorine Chloramine
oH 6 to 7 DH 8 to 9
0.034-0.05
1.1-2.5
0.01-0.05
0.08-0.18
47->150
30-630
Crvptosporidium 7200
-------
for chlorine dioxide, chloramine, and chlorine to 86,  99,  and  246  fjg/l, respec-
tively.  The same trend seen for TOC was also observed for TOX where  an average
reduction of 33 percent occurred  in  the ozone contact chamber producing an aver-
age effluent concentration of 16  pg/L with a further reduction  to 11 fjg/l across
the ozone sand column.  Treatment of the sand filtered effluents  with  free chlo-
rine followed by 5 day  storage to simulate the distribution system  significantly
increased TOX concentrations for  all process streams (557,  540, 339, and 379 vg/L
for nondisinfected, chlorine, ozone, and chlorine  dioxide, respectively).

     The trihalomethanes (THMs)  reacted as expected with no significant concen-
trations (1 fjg/L average)  observed in the disinfectant contact chamber and sand
column  effluents  of the  nondisinfected,  ozone, and  chlorine dioxide  process
streams.  An average THM concentration of 3 fjg/L occurred  in the  chloramine dis-
infectant contact chamber and sand  column effluents while  that  in the chlorine
contact chamber effluent  averaged  39 /*j/L and increased  to 49 /*j/L across the
chlorine sand  column.   By  maintaining  a chloramine  residual  for 5  days,  the
terminal THM concentrations increased slightly  to  an  average  of 8.5, 3.2, 4.2,
and 9.4 //g/L  for the nondisinfected, ozone,  chlorine dioxide,  and chloramine
process streams, respectively.  Similar treatment and storage with  free chlorine
produced relatively high terminal THM concentrations for the nondisinfected and
chlorine process streams with average concentrations of 236 and  225 fjg/l.  When
compared to the nondisinfected sand  column effluent,  terminal  THMs were 35 and
41 percent  less when ozone  and  chlorine dioxide were used during pretreatment
resulting  in  average concentrations  of 154  and  138 tig/I, respectively.   As
expected, granular activated carbon  (GAC) reduced  all concentrations until the
columns became saturated in 60 to 80 days.

     The haloacetic acids followed the same trend as seen  with THMs except at a
lower concentration. The highest concentrations were formed using  free chlorine
which mainly produced dichloroacetic acid (DCAA),  trichloroacetic acid (TCAA),
and bromochloroacetic  acid  (BCAA).   Chloroacetic acid (CAA),  bromoacetic acid
(BAA), and dibromoacetic acid (DBAA) were also  formed to some extent. Average
instantaneous concentrations for dichloroactic acid were 0.9,  1.7,  3.7, 1.1, and
13 fjg/l for the nondisinfected, chlorine dioxide, chloramine, ozone, and chlorine
process streams after sand  filtration, respectively.  Five-day terminal  values
for the same process streams with free chlorine were 60, 44, 38,  and 60 fjg/l for
the nondisinfected,  chlorine dioxide, ozone, and chlorine streams, respectively.

     Similar treatment of the sand filter effluents with chloramine and a 5 day
storage period  resulted in only slightly elevated DCAA  concentrations.   The
slight concentration Increase is similar to that seen for THMs during similar
treatment  with  chloramine  suggesting that  both were formed  by free chlorine
during  in-s'itu formation of chloramine.   GAC adsorption  produced  continued
removals of 80% or greater  of DCAA  after steady-state which was about 150 days
of operation.  GAC  steady-state  was  reached  in  about 250  days  for  DCAA after
storage for five days  with  a free  chlorine residual.   After GAC steady-state,
average removals were  48,  73,  53,  46, and  51  percent for the nondisinfected,
ozone, chlorine dioxide, chloramine,  and chlorine process streams, respectively.
Similar observations for all unit processes were also seen for TCAA.

     Relatively low concentrations of haloacetonitriles were formed across each
process stream with  chlorine producing the highest levels which  averaged 3.1
total haloacetonitriles. Less than  1 pg/L average  of  haloacetonitriles was
                                                                             8

-------
observed across  the  other process  streams.   The predominant haloacetonitrile
(HAN) was dichloroacetonitrile (DCAN) followed by bromochloroacetonitrile (SCAN),
dibromoacetonitrile (OBAN), and trichloroacetonitrile (TCAN).  Treatment of the
chlorine sand column  effluent with additional free chlorine and subsequent 5-day
storage produced an average concentration of 1.9 pg/L of DCAN which was  the same
as that  of the  sand column  effluent suggesting that  all  DCAN  precursors had
reacted.   No  consistent  breakthrough of the HANs was observed through  any GAC
column except that of the chlorine process stream, which was still  removing over
95% of the influent HANs  at the end  of  the  one-year operational period.

     Only  two haloketones, 1,  1, 1-trichloropropane  (TCP) and  1,  1-dichloro-
propane  (DCP) were detected with the highest concentration (1 to  2  /*3/L) being
observed in the  chlorine  process stream.  Post chlorination of the  sand column
effluents  followed by 5-day storage produced similar TCP concentrations in each
process stream with average concentrations of 2.1, 2.5, 4.2, and 2.5 fjg/l for the
nondisinfected,  ozone, chlorine dioxide,  and chlorine,  respectively.   Although
consistent breakthroughs  of these   haloketones  were observed across  the GAC
columns, removals remained above 85 percent throughout the one-year operational
period.

     Chloral hydrate (CH) was formed predominantly in the chlorine process stream
with  an  average contact  chamber effluent concentration of 2.9  fjg/l which in-
creased  to 4.5 j/g/L  across the sand filter  because  of an  additional 30 minutes
chlorine contact time.  The CH concentrations in the effluent  of  the chloramine
contact  chamber  and  sand  column were identical averaging  0.25 //g/L. CH was de-
tected  intermittently  in the contact chamber and sand column effluents of the
chlorine dioxide,  ozone,  and nondisinfected process  streams at average concen-
trations  ranging from 0.01 to  0.07 /*j/L.  Post-treatment of the  sand column
effluents  with free chlorine and storage for 5 days produced relatively similar
CH concentrations averaging 79,  55, 45, and 75  /*}/L for the nondisinfected,
ozone,  chlorine dioxide,  and chlorine process streams, respectively.   Similar
treatment  with  chloramine produced  average  concentrations of  0.03 to 0.3 fjg/l.
GAC  adsorption  removed  all  of the  CH  throughout the  project  in all  process
streams.

      The concentrations of chloropicrin  (CP) in the effluents of the disinfection
contact  chambers and the sand columns  were the same, averaging  0.004, 0.004,
0.015,  0.038, and 0.43 ijg/l for  the nondisinfected, ozone,  chlorine  dioxide,
chloramine,  and  chlorine  process streams, respectively.  Chlorination  and  5-day
storage  of the sand column effluents produced concentrations averaging 1.3, 7.7,
1.4,  and 1.3  ^g/L  for the nondisinfected, ozone, chlorine dioxide, and  chlorine
process  streams, respectively.   Pre-ozonation appears to  have produced an in-
crease  in CP precursors.  Only slight  increases in CP  concentrations  were ob-
served  after similar chloramine treatment and storage of the sand column efflu-
ents with  average  concentrations of 0.03, 0.04,  0.11,  and  0.09 /xi/L for the non-
disinfected,  ozone,  chlorine dioxide, and chloramine, respectively.  No consis-
tent breakthrough of CP above  0.003 //g/L was  observed across  the GAC  column  of
any  process  throughout the operational  period.

-------
MUTAGENICITY STUDIES  (Jefferson  Parish.  Louisiana)

Satnole Disinfection

     For the health-related studies at Jefferson Parish, a portion of the clar-
ified water  after fluoridation was filtered through pressure sand filters and
split into four  process streams as shown in Figure 3.  Three of  these  streams^
were disinfected with either  chlorine,  chloramine,  or ozone while  the fourth
stream was not disinfected and served as a control.   Chlorine dioxide  was not
evaluated at this time because it appeared that chlorine dioxide would not be an
acceptable disinfectant when the disinfection/disinfection by-product  regulation
was promulgated.  The system used for  the  three disinfected streams consisted of
a contact chamber followed in  series by a  sand column and a 55-gallon  stainless-
steel drum fitted with a spiral stainless-steel baffle.  The modified drum served
as an additional  contact chamber.  The system for the nondisinfected  stream was
similar except that  the initial  contact chamber  was  omitted.s The contact time
in  the  contact  chamber  was approximately 30 min  for  each of the disinfected
streams and  the  empty-bed contact time for the sand column was approximately 20
min.  The ozone stream was split after the sand column and post-disinfected with
either chlorine  or chloramine.  As a  result of  stream  splitting,  the flow rate
to the two post-disinfected drums was  decreased.  Therefore, the contact time for
post-disinfected drums was  approximately  150-180 min while the contact time for
the non-ozonated chlorine and chloramine streams and the  nondisinfected stream
was approximately 85-100 min.   The  amount  of chlorine and chloramine added to the
two ozonated streams was adjusted so that the residual disinfectant concentration
leaving  the  drums matched  the corresponding final residuals observed  for the
chlorinated  and chloraminated (non-ozonated) streams  (0.5-1.0 ppm C12 and 0.8-1.5
ppm NH2C1).

SAMPLE COLLECTION

     Macroreticular resins (Amber!ite XAD-8 and XAD-2) were used to collect and
concentrate  the  organic compounds from  the  process streams for  chemical and
toxicological testing.  The concentration procedure described by Miller et al.
was  used  with  the following modifications.   Ethyl  acetate was utilized instead
of acetone to elute the organic material from the XAD resins and the ratio of
sample volume to resin volume was  decreased from 760:1 to 150:1.  The change in
ratio was made to minimize the potential breakthrough of MX, 3-Chloro-4 (dichlor-
omethyl)-5-hydroxy-2(5H)-furanone, and compounds with similar  solubilities in
water.  The  XAD  resins were cleaned by Sohxlet extraction with ethyl acetate and
methanol  for 24  hr  and stored in methanol.  Prior  to sample collection, each
column  pair containing  5 L of XAD-8  and  XAD-2 resin was  flushed  with 200 L of
water to remove  the  residual  methanol.   Fifteen  hundred  liters of each of the
process  streams  were passed through the column pairs serially at a flow rate of
60 L/min.  The pH of the column influent was maintained at pH 2 by the in-line
addition of 9N hydrochloric acid.

      Following sample collection, each column was filled with sufficient ethyl
 acetate to provide a standing head. The columns were then  agitated to completely
wet the resin and allowed to  equilibrate for 15 minutes.  Each column pair was
 then eluted  serially with 15 L of ethyl  acetate.  Residual  water was removed from
 the ethyl  acetate extract by allowing the water to separate and then draining off
 the water in a separatory funnel.  Each extract was concentrated to <1.5 L under
                                                                             10

-------





o
'CT
___
o
X
0













j
/
/
/















1
1
I

C'
«

!
CO!
CO '


,
1
I
1
1
«*-M
1
\
ml
"il
I
I
I
1
















(}
(I



t









,<5
y***
CO
O






















(0

3
*5











































•



T3
C
CO
CO w
QJ dl
D£
en11
o
ol


L
















^
,








•




i
'
i
•



t








t
<



































CD
0
O
1 f
z •--
D.12
2Q







A
1
O"~

CM




t
I
CM










t
1
J?
O


t
_J






^ *•"
°5
-^^ Vmrn

O O





^^ t^
•§1
OiR
o









•*** m
(O-g
s|
^"^ ^J






















































CO ^***
<*A ii
wj UL.






^ i-
«3 •*-*
cou.











SO)
€
COLL




Col
i4^
COLL




...

















C
1
I



c










'




	 1





o







o
o

•
>
N
1
.... J


5-> -•



0
rff
d




0
o




l/ ^ cf
^
1
•mmJ
-
^ {/)

03
o
LL.
C
r) 1 ^ CD
j r^ j75
CD

H—
^ °
rrt
vvJ
rt\
QJ
o \+ i
0
CD
Wy
cn
'1 ^ w .C
rrt
^u
*/%
CO
^ -
.0
O"™~

O
X
o

CO
— ^ CD
rr>
\Jt
LL
\ \ ^
U J ^
11

-------
vacuum at 40°C by  rotary evaporation.  The samples were each adjusted to exactly
1.5 L with ethyl  acetate before  storing  at  4°C.   Immediately  prior  to chemical
or toxicity testing, aliquots of the 1000X concentrates were further concentrated
by rotary concentration at 40°C.

RESULTS AND DISCUSSION

     Three rounds of water sample  concentrates were collected during September
1989, March 1990,  and July 1990.  An examination  of the seasonal effects  (early
fall-89, early spring-90, and mid summer-90)  of finished water quality and  surro-
gate parameters for the control stream (non-disinfected + sand)  showed that the
early spring  sample had the lowest  values for: pH  (7.3),  dissolved  solids (184
mg/L), TOC (2.8 mg/L),  and TOX  (15.7 ug/L).   The early spring sample also had the
lowest concentrations for eight volatile  organics examined.  This latter finding
may simply be a consequence of  the lower TOC  levels  in the spring sample that may
result from  a flushing of   naturally occurring .organic matter from the  source
water over the course of the winter.  However,  it  should be noted that TOX  values
observed  for  the  (C12  +  sand)  and (NH2C1 +  sand) samples were highest for the
spring sample.  This suggested  that the nature  of  the precursor material was dif-
ferent and that more  semivolatile  and  nonvolatile halogenated compounds were
formed.   This idea is  supported by  the  observed increase in the formation of
haloacetic acids.

     Significantly lower values were obtained for the  spring sample  with  regard
to brominated compounds (bromodichloromethane, dibromochloromethane, and  bromo-
form).  The most likely causes  being either a decrease in the bromide content in
the source water  or an increase in the ammonia level  of the  source water that
would result  in the formation of combined bromine rather than  free bromine.  As
for the effect of seasonal  changes on the formation of  halogenated by-products,
the differences appeared to be less well  defined.  However,  the over-all  forma-
tion of brominated DBFs was spring  < summer <  fall.

     Many of the halogenated by-products  of  interest were analyzed  for each dis-
infectant stream. <6'r'8> Of special  interest was  the potential concentration of
these by-products when water is  delivered to  the customer.  This  was evaluated
by storing samples for  5 days with a disinfectant residual to simulate concentra-
tions in the distribution system.  An example of  the effects of using different
disinfection  alternatives for  the  September 1989 sampling  is  shown  in Table 3.
These data show a definite trend.   For example,  with dichloroacetic acid, when
chlorine  is added prior to and after sand filtration about 45  ^g/L,  on average,
was detected.  If  ozone is added prior to  sand  filtration and chlorine after sand
filtration, this average concentration was reduced to 32 #3/L.  If however, mono-
chloramine is used prior to and after sand filtration,  the average  concentration
is reduced to about 8  fjg/L.  Further reductions  were seen when ozone was added
before sand filtration and monochleramine after sand filtration  (4.6 ^g/L aver-
age).  This  trend was  seen for other prevalent halogenated  by-products such as
trichloroacetic acid,  bromochloroacetic  acid, chloral hydrate, and trichloro-
methane.

     A similar trend  is  seen for mutagenicity studies on samples collected at
Jefferson Parish as shown in Figure 4. From these data, one might  conclude that
ozonation prior to sand filtration and chloramination after sand filtration will
                                                                            12

-------
TABLE 3.  JEFFERSON PARISH DISINFECTION BY-PRODUCT FORMATION
      POTENTIAL FOR SELECTED STREAMS (SEPTEMBER 1989)
                    Concentration in
BY-PRODUCT
Chloroacetic
Acid
Dichloracetic
Acid
Trichloro-
acetic Acid
Bromoacetic
Acid
Dibromoacetic
Acid
Bromochloro-
acetic Acid
Dichloro-
acetonitrile
Trichloro-
acetonitrile
Dibromochloro-
acetonitrile
Bromochloro-
acetonitrile
1,1-Dichloro-
propane
1,1,1-Trichloro-
propane
Chloropicrin
Chloral Hydrate
Trichloromethane
Bromodichloro-
methane
Dibromochloro-
methane
Tribromo-
methane
TOX
NONDIS.
0.
0.5
0.2
0.
0.
0.2
0.1
0.
0.
0.
0.1
0.
0.
0.
1.0
0.4
0.1
0.1
27.
SAND-
C12
16.0
44.9
39.8
1.2
0.8
28.7
1.6
0.1
0.3
0.7
0.1
1.8
0.9
37.5
156.
35.0
8.
0.
505.
OZONE-
SAND -
_CJ2
17.0
32.3
19.0
1.0
1.2
20.2
1.3
0.
0.4
0.7
0.1
1.7
4.2
32.6
121.
29.0
10.0
1.0
328.
NH2C1 -
SAND-
_NH2CJ_ .
0.4
8.2
1.7
0.1
0.1
3.8
0.
0.
0.
0.
0.
0.
0.
0.1
5.3
1.2
0.1
0.
55.
OZONE -
SAND-
NH2C1
0.6
4.6
0.7
0.1
0.1
1.4
0.
0.
0.
o.
0.
0.
0.
0.
1.8
0.5
0.1
0.1
29.
                                                                    13

-------
    0)
        > to
                                         ±= ffi
                                            £
                                         O) 
-------
solve the halogenated by-product problem.  This is true but one also has to eval-
uate the total  impacts of disinfection.  For instance, although ozone shows pro-
mise and may be  a  viable disinfection alternative, assimilable organic carbon
(AOC) increase in the distribution system has to be controlled.  This can be done
by biostabilization  of the water during treatment  before distribution.  An ex-
ample of this is shown in Table 4.   In this example, after ozonation,  162 (jg/L
of AOC is present.   Sand filtration  provides  some  biostabilization  by  reducing
the AOC to 38 fjg/l while the GAC effluent  is 4 ^g/L;  comparable  to  the  AOC con-
centration (3.7 /jg/L)  in the chlorine contact  chamber effluent.  Ozonation by-
products such  as  aldehydes,  ketones, and acids  are  also  a concern.   Although
chloramines may be  the disinfectant of choice to provide a disinfection residual,
it may not provide the desired disinfection effectiveness and there have been
some health concerns  with ingesting  chloramines.

EVANSVILLE. INDIANA  STUDIES

     Another potential disinfectant for controlling halogenated  by-products,  is
chlorine dioxide which was further evaluated  beyond  the scope of the Jefferson
Parish project.  At a drinking water utility on the Ohio River, a pilot plant was
used to compare chlorine  dioxide disinfection with chlorine disinfection'  .  The
addition of chlorine dioxide to the raw water with delayed chlorination permitted
coagulation/sett!ing/filtration and  oxidation  to  remove trihalomethane precur-
sors, thereby reducing the amount of trihalomethanes formed during post-treatment
chlorination.  A comparison  of  the average trihalomethane concentrations for the
two disinfectant modes showed a reduction of approximately 60 percent when chlo-
rine dioxide was used.   Although chlorine dioxide  disinfection  can  reduce tri-
halomethane concentrations, control  of the  metabolites  (chlorite and chlorate)
is essential  before chlorine dioxide can  be  considered a viable  disinfection
alternative.   Equipment is now available  to  produce chlorine dioxide that  is
virtually free of chlorine and  chlorate.  Bench-scale studies have recently been
completed which  evaluated  the  reduction of chlorite and chlorine dioxide by  a
reducing agent.c>

     All of the samples used for these studies were collected from the effluents
of pilot plant secondary  settling basins after approximately 200 minutes of chlo-
rine dioxide contact time.( }  These grab  samples were analyzed  immediately for
residual chlorine  dioxide (C102), chlorite  (C102), and chlorate (C103), treated
with the respective reducing agent for a variety of contact times, and reanalyzed
for the  same chlorine dioxide  species.

     The major reducing  agents considered for these studies consisted of sulfur
dioxide, sodium thiosulfate, sodium metabisulfite,  and ferrous iron.  For sulfur
dioxide,  complete  reduction of residual  chlorine dioxide and chlorite ion was
achieved  but  a marked increase in chlorate ion concentration was  consistently
observed as shown in Figure 5.  This same phenomenon occurred with sodium metabi-
sulfite.  The degree of C102 reduction with sodium thiosul fate depended on pH and
contact  time.  Removal of chlorite ion was demonstrated  with little or no chlo-
rate formation.  However, the dosages required for complete reduction appear  to
make this  reducing agent impractical.   Ferrous iron  appears  to  be  an effective
methodology for the removal  of  chlorine dioxide and chlorite ion residuals.  Both
of these  undesirable finished  water  residuals were eliminated in minutes at  pH
6.0 to 7.0 and preliminary results indicate that excess  reductant may be easily
controlled by  pre-filter chlorination.
                                                                             15

-------
       TABLE 4. AVERAGE AOC CONCENTRATION
                AT JEFFERSON PARISH,  LA
     TREATMENT                     AOC.uq/L
NON DISINFECT                         13.5
03CONTACT CHAMBER                    162.
Oj-SAND                               38.2
03-SAND-GAC                           4.0
C12CONTACT CHAMBER                    3.7
C12-SAND-GAC                          2.6
                                                              16

-------
                               Q-



o S
Jt± CO
^- *— c
o .0
.c .c
0 O
$ <{> <


<



<
4



> c

[


> I



J


	 1 	 1 	 1 	 1 	 1 	 1 	 1 	 ^ 	 1 	 1 	 1 I | 1 1 1 f"""^
^r
_ 50
in
" w
I °
: co

u tO
^ CNJ
•

L o
: CM
.

•
i- !^
1 o
- 10
> o



^)
CO
D)
CO
O
Q
0
T5
'x
O

b
u.
D
«*—
3


CO        CM         i

  (I/BUJ) eO!0 - 3OIO
                                           0
                                           12
                                           'x
                                           g

                                           b
                                           CO
                                            g

                                            O'
                                            CD   ->
                                           tr .2
II
O LJJ


in

 2
 ^
 O)
LL
                                               17

-------
     Bacterial mutagenicity was evaluated during a dual  pilot plant study using
ferrous iron  in one  train  with chlorine  dioxide disinfection and ozone in the
other train (Figure 6).  Figure 7 shows the  results of these tests using the TA
100  strain.   Regardless  of whether  chlorine  dioxide  or  ozone  is used  as  a
predisinfectant, when GAC  is incorporated into  the  treatment scheme, the muta-
genicity is low when compared to the schemes incorporating dual media filtration.

SUMMARY

     Regulations to control contaminants  in  drinking water  in the United States
are expected to  become more stringent and more prevalent. Forthcoming disinfec-
tion and disinfection by-product regulations will effect  virtually every commun-!
ity water system.  There are various ways to control disinfection by-products,
one of which is  to use an alternative to  chlorine.   However, when this is done,
one should consider various aspects of each disinfectant.  There appears to be
no single disinfectant that is  applicable for all  situations.

ACKNOWLEDGEMENTS

     The authors acknowledge the assistance of Robert Miller,  Paul Ringhand, and
Ron Dressman for their help in  sample collection, analyses,  and data interpreta-
tion.  The authors also thank Sandra Dryer, Pat Underwood, and Steve Wai trip for
helping prepare this  paper.

     This  paper  has  been  reviewed  in accordance  with  the U.S.  Environmental,
Protection Agency's peer  and administrative review policies and approved for pre-
sentation and publication.  Mention of trade names  or commercial products does
not constitute endorsement or  recommendation for use by the USEPA.
                                                                            18

-------
                         Secondary
                         Settling
Raw Water
    '  /*
    -4$*
lum


 cil tt|  I ct| I  t±3
     cioa
   V
        2 Stage     Flocculators
        Rapid Mix                  Tube
                         Secondary
                         Settling
                     W I  W
        2 Stage     Flocculators     _ .  .
        Rapid Mix                  Tube
                                                          GAG
                                              Cla
                                              Contact
                                              Chambar
                                                          GAC
     Figure 6.   Evansville,  Indiana Pilot Plant
                                                           19

-------
cvj   co
 I
o
CO
 CM


Jein
                                       (0
                                       Q
                                              0)  ,-
                                              I
                                              1.1
                                                 S
                                               o o
                                               c^
                                               CD C
                                                 .
                                               O) Q.

                                               '"S CC
                                               £13
                                                2

                                               I
                                                20

-------
REFERENCES

 1.  Lykins, Jr. B.W., Koffskey,  W.E.,  and  Miller,  R.G.,  "Chemical Products and
     Toxicologic Effects of Disinfection",  Journal AWWA, 78:11, 66-75, 1986.

 2.  Hoff, J.C., "Inactivation of Microbial Agents by Chemical Disinfectants",
     U.S. EPA, EPA/600/2-86/067, 1986.

 3.  National Primary Drinking Water Regulations,  Final Rule, Federal Register,
     40 CFR Parts 141 and 142, Vol. 54, No. 124, Thursday, June 29, 1989.

 4.  Korich, D.6., Mead, J.R., Madore, M.S., Sinclair, N.A., and Miller, R.G.,
     et al,  "Results  of lexicological  Testing  of Jefferson Parish Pilot Plant
     Samples", Environmental Health Perspectives, Vol. 69, 129-139, 1986.

 5.  Miller, R.G., et al, "Results  of Toxicological Testing of Jefferson Parish
     Pilot Plant Samples",  Environmental  Health Perspectives, Vol. 69, 129-139,
     1986,

 6.  Lykins, Jr.,  B.W., et al "Ozone  for  Trace Organic  Contaminant Removal",
     Proceedings of 1990 International Ozone Association Conference, Shreveport,
     Louisiana, March 27-29, 1990.

 7.  Lykins, Jr. B.W., Moser,  R., and DeMarco,  J.,  "Treatment Technology in the
     United States: Disinfection and Control of Disinfection By-Products", the
     2nd  Japan  -  U.S. Governmental  Conference  on Drinking  Water  Quality
     Management, Tokyo, Japan, July 24-26,  1990.

 8.  Lykins, Jr.  B.W.,  "Disinfection of Drinking  Water  and  Control  of Disin-
     fection By-Products in the United States",  U.S./U.S.S.R. Bilateral Project,
     Cincinnati, Ohio,  October 15,  1990.

 9.  Lykins, Jr.  B.W. and  Griese,  M.H., "Using Chlorine Dioxide  for Trihalo-
     methane Control",  Journal AWWA, 78:6,  88-93,  1986.

10.  "Disinfection  Byproduct  Control   Using   Chlorine   Dioxide",   U.S.' EPA
     Cooperative Agreement No. CR-816837,  1990.

11.  Griese, M.H., et al.,  "The Use  of Reducing Agents For the Total Elimination
     of  Chlorine  Dioxide and  Chlorite Residuals  in  Drinking  Water",  Journal
     AWWA, 83:5,  56-61,  1991.
                                                                             21

-------