United States
                    Environmental Protection
                    Agency
 Municipal Environmental
 Research Laboratory
 Cincinnati OH 45268
                    Research and Development
 EPA-600/S2-84-090 Sept. 1984
SERA         Project  Summary

                    Trihalomethane  Precursor
                    Removal  by  the  Magnesium
                    Carbonate  Process
                    J.S. Taylor, B.R. Snyder, B. Ciliax, C. Ferraro, A. Fisher, J. Herr, P. Muller,
                    and D. Thompson
                      A project was conducted to determine
                    and improve the ability to the magnesium
                    carbonate process to  remove trihalo-
                    methane (THM) precursors in treated
                    drinking water. The project was con-
                    ducted at a  drinking water treatment
                    plant in Melbourne, Florida, where the
                    process had  been installed in  the early
                    1970's (before THM regulation) to
                    reduce greatly the sludge produced
                    from water treatment.  The  process
                    involved recovering the magnesium
                    from the sludge by carbonation, reusing
                    the Mg(HCO3)2 liquor for coagulation,
                    recalcining the remaining CaCOa solids,
                    and reusing the recovered CaO for pH
                    control during coagulation. The project
                    consisted of seven phases,  including a
                    jar test, recycle recovery, oxidation,
                    THMFP model development,  distribu-
                    tion system, alternative disinfectant,
                    and granular activated carbon.
                      The magnesium carbonate process
                    showed no advantage over alum coagu-
                    lation with regard to the color remaining
                    in the water after treatment. With both
                    processes, the use of large doses ol
                    chlorine to  decolorize  the  finished
                    water and provide a disinfectant residual
                    in the distribution system resulted in the
                    formation of THM's in excess of the 0.1
                    mg/L regulation.
                     This Project Summary was developed
                    by EPA's  Municipal Environmental
                    Research Laboratory, Cincinnati,  OH,
                    to announce key findings of the research
                    project that is fully documented in a
                    separate report of the same title (see
                    Project Report ordering information at
                    back).
Introduction
  In  1980, the  U.S. Environmental
Protection Agency (EPA), the City of
Melbourne, Florida, and the University of
Central Florida entered into a cooperative
agreement to determine the capability of
the magnesium carbonate process to
remove trihalomethane (THM) precursors.
This process is unique to potable water
treatment in that the principal coagulant,
Mg(HCO3)2, was recovered and recycled.
Melbourne, Florida, was the only city that
had built a plant specifically for the
magnesium carbonate process.  No in-
formation had been gathered on the
ability of  this process to  remove THM
precursors, as THM control was not a
requirement or a concern at the time of
plant design. A schematic of the magne-
sium carbonate process at Melbourne
appears in Figure 1.
  Recovery of the Mg(HCO3)2 coagulant is
accomplished by carbonation  of the
Mg(OH)2 - CaCOs sludge taken from the
clarifiers. The form of magnesium is
changed to Mg(HCO3)2, which theoretically
allows complete magnesium recovery if
the total magnesium concentration in the
carbonation  recovery unit does not
exceed the solubility of MgCO3.H2O. The
carbonated slurry  is allowed to settle,
separating recovered Mg(HC03>2 liquor
from the remaining CaC03 solids. CaCO:,
is  not solubilized  because of the  low
solubility  product relative to  that of
Mg(HC03)2.3H2O,  1(T83 as opposed to
10~54. The CaC03 solids were eventually
calcined in  a lime kiln  and  reused,
theoretically producing no sludge for
disposal. The  principal reactions of the

-------
         Alum
         Carbon
    Lime
   Coag. Aid
                                                            CO2  (NaPOalt
  L. Wash/
   Pumps
                                                                                       CI2
           1
          1
 Flash
 Mixer
                 a
                 .a
                 •§•
                 o
a
                 o
                 I
     Floccu-
      lator
Clari-
 fier
                               Recycle Sludge
Ftecarb.
 Basin
Filters
                                             Recarb.
                                              Cells
                         Storage
I
                                             Thickener
                          Waste
                            Vacuum
                             Dryer
                          CaO
                         Storage
                             Furnace
         Backwash
          Storage
           Pond
figure 1.    Flow diagram {or Melbourne Water Treatment Plant.
Clear
Well
Storage
 DBT
magnesium carbonate process are shown
in Equations (1), (2), and (3). Equation (1)
represents  the  recycle of dissolved
Mg(HC03)2  and solid CaO which  is
solubilized and reacts to produce solid
Mg(OH)2 and CaCOs floe. Equations (2)
and (3) represent recovery of magnesium
and calcium by converting solid Mg(OH)2
to dissolved  Mg(HC03)2 and solid CaCOa
to solid CaO by recalcination. The solid
CaO is converted to dissolved Ca(OH)2 by
slaking (hydrolysis) before recycle.

Recycle: Mg(HC03)2 + CaO(s) + H20—
     Mg(OH)2
-------
 increased the THMFP, TOC, and  color
 following coagulation, settling, and filtra-
 tion (CSF). However, increasing doses of
 doses of conventionally prepared MgfHCOgfe
 recycled coagulant increased removal of
 THMFP, TOC, and color following CSF. Of
 the three magnesium sources tested,
 unused MgS04.7H20 was the  best,
 removing 5%  to 10% more color,  TOC,
 and THMFP than either the saline or con-
 ventionally prepared Mg(HC03)2 recycled
 coagulant. Thus recycling the Mg(HCOa)2
 coagulant decreased the efficiency of
 CSF. Typically, magnesium coagulation
 at pH 11.3 to  12.0 accompanied 80% to
 98% color removal,  20% to 40% TOC
 removal, and 40% to 65%THMFP removal.
 Optimum THMFP reduction was always
 accompanied by optimum TOC and color
 reduction. Approximately a 2%  increase
 in color, TOC, and THMFP removal was
 obtained for every 0.1 pH increase from
 pH 11 to 12. Not all of this increased re-
 moval was attributable to increased mag-
 nesium precipitation, as that was essen-
 tially complete at pH 11.5. No significant
 differences occurred in THMFP, TOC, or
 color removal  among the three different
 lime sources  (commercial, recalcined,
 and reagent) that were used for pH con-
 trol in magnesium coagulayion.
  The reaction of color, TOC, or THMFP
 with floe that  was removed by filtration
 through a 0.45-/um filter was independent
 of  variations in  slow and rapid mixing
 times. But the removal of formed turbidity
 during coagulation was maximized at less
 than 0.75 min  of rapid mixing and 19 min
 of slow mixing. These results indicated an
 instantaneous reaction of color, TOC, or
 THMFP with floe and (as expected) a mix-
 ing  dependent relationship  for turbidity
 removal. Alum used as a coagulant aid at
 pH  11.5 increased THMFP, TOC,  and
 color removal by approximately 10% and
 enhanced floe settling. An  increase in
 dissolved aluminum in the settled water
 was not found until more than 2 mg/L
 Al+3 was  dosed  in  the  alum form.
 Statistical correlation of color, TOC, and
 THM indicated that not all TOC produced
 THM's or color.


 Recycle Recovery
  The color,  TOC, and THMFP of the
 recovered  MgJHCOafe liquor typically
 would exceed 3,000 chloroplatinate units
 (CPU), 500 mg/L, and 50,000 fjg/L,
 respectively. The purpose of the recycle
 recovery subtask was to determine
 whether recovered Mg(HC03)2 could be
 improved as  a coagulant for removing
i THMFP, TOC,  or color by ozone or CIO2
 oxidation. These studies indicated that
neither  Oa nor ClOa oxidation  of the
recovered Mg(HCO3>2 liquor would improve
the color, TOC, or THMFP removal when
the Mg(HCOs)2 was recycled.


Oxidation
  The purpose of the oxidation phase was
to determine whether oxidation with Os
or CIO3 could improve removal of color,
TOC,  and THMFP in the  magnesium
carbonate process before chlorination.
In this process, Oa or CIO2 oxidation could
be executed  either in  the  raw water
before coagulation, after coagulation  at
pH 11.5, or after recarbonation at pH 8.3
and was investigated on a bench scale
under these conditions.
  Ozone applied before coagulation
reduced color but did not reduce TOC  or
THMFP  of the raw water. When the
ozonated raw water was coagulated, the
final color was slightly lower, but the TOC
and THMFP were unchanged when com-
pared with the unozonated water. Ozone
applied to the recarbonated water after
coagulation also reduced the color but did
not change the TOC or THMFP of the
unozonated treated water. The color  in
either case could be reduced to less than
the 15 CPU standard,  but  Oa doses  of
more than 10 mg/L were required. Ozone
applied  after  coagulation  at pH 11.5
reduced color, TOC, and THMFP more
effectively than  at  other  pH's  and
demonstrated that ozonation is used
most effectively at  pH 11.5 after coagula-
tion.
  ClOz applied before coagulation did
reduce the TOC and THMFP but did not
reduce the color of the coagulated water.
The same effect was found when CI02
was applied after coagulation at pH 11.5.
In each case, a CIOz dose of 10 mg/L
reduced the TOC 7%and the THMFP 35%.
CIOz applied after recarbonation did not
significantly affect TOC or color, but it
reduced THMFP 23% at the same CIO2
dose.


THMFP Model
  A log-variant model for THMFP forma-
tion of CSF waters was developed from
more than 600 measured THMFP's from
three  different sources chlorinated with
5, 10, 20, and 40 mg/L Cl at pH 6.0,7.5,
and 9.0, and at temperatures of 8°, 16°,
and 32°C for times of 0.005,1,6,24,48,
and 144 hr. The r2 and a values for the
model were 0.98 and 10~4. The model is
given  in Equation (4) and shows that the
effect of TOC > time > Cl >Temperature
> OH on  THM formation. The average
molecular weight of the THM's formed in
this study was 127 - approximately 85%
HCCk 14%  HCCIzBr, and 1% HCBr2CI.
This model, shown in Equation (4), could
be used to estimated THMFP during plant
design or operation:
    THMFP = 2.70 (TOC)124 (CI)'
                            L0.24
         (OH)010 (T)017 (t)026
                       (4)
where:
THMFP=/umol/LTHM as THM
   TOC=mmol/L TOC as C
     CI=mmol/LCl2asCI
    OH=mmol/LOH"asOH"
     T=°C
      t=hr.
  This model can provide useful informa-
tion for THM control in similar waters.
Assuming a water temperature of 20°C, a
TOC of 10 mg/L, a pH of 9, and a chlorine
dose of 20 mg/L,  THM's would be 547
/ug/L  after 96 hr in the distribution
system. The TOC precursors would have
to be  reduced to 2.6 mg/L to have pro-
duced no more than 0.10 mg/LTHM'sfor
the same conditions. Such  a  level is
impractical for most highly colored
surface supplies. If disinfection is changed
to chloramination with a chlorine dose of
10 mg/L before ammonia feed under the
same  conditions,  free chlorine contact
time would have to be less than 9 min to
avoid  THM formation greater than 0.10
mg/L.
Distribution System
  The water quality in the Melbourne
distribution system was  monitored to
determine the effects of two  process
changes which  were instigated at  the
water treatment plant to control THM's.
These were: (1) changing from Mg(HCOa)2
coagulation to alum coagulation; and (2)
changing from chlorine disinfection to
chloramination.  Using alum resulted in
the removal of approximately 10% more
color, 13% more TOC,  and 10% more
THM  precursors in the coagulation
process.
  The change from chlorine to chlora-
mines in  June  1983 resulted in: (1)
THM's at the extremities of the distribution
system of less than  50jug/L, (2) disinfec-
tant residuals in water with  summer
temperatures exceeding 85°F at distant
points where they  had  seldom, if ever,
been  documented, (3) lower standard
plate counts (SPC) by orders of magnitude,
and (4) a significant reduction in consumer
complaints. As a result of these changes,
Melbourne is now experiencing  the
highest residuals,  lowest SPC's, and
lowest THM's that have been recorded in
the distribution system.

-------
Alternative Disinfectants
  The purpose of the alternative disinfec-
tant phase was to determine the capability
of CI02, NH2CI,O3, and CI02, singularly or
in pairs, to remove trace color, reduce
THMFP, and provide residuals asfunctions
of time and temperature. NH2CI was not
effective for color reduction, but CIO2, O3,
and C\z could effectively reduce 1 to 2
CPU from 1 L of CSF water for every mg/L
dosed. Temperatures did not significantly
affect  color  removal.  Color reduction
became increasingly difficult as the final
color decreased.
  The rate of disinfectant dissipation was
03  >  CI2 > CIO2 > NH2CI.  This rate
generally increased with  temperature
and exhibited exponential decay. When
dosed at 3 to 5 mg/L, NHaCI could maintain
the required residual of 0.6 mg/L for 7
days  at 32°C. CIO2  and  O3 could not
maintain a residual at doses of 10 mg/L
and 20 mg/L, respectively, under these
conditions.
  Neither CI02, O3, nor NH2CI produced
THM's. But if chlorine were used for trace
color  removal, the  dose  and contact
periods would have to be closely controlled
-typically less than 10 mg/L and 9 minto
avoid violation of the  0.1  mg/L  THM
standard.  The combination of O3 with
NH2CI and CI2 with NH2CI were effective
for: (1) reducing Term THM,  SPC, and
trace color, and (2) producing a residual in
the distribution system.  O3 with CIO2 and
CIO2 with NH2CI would also be effective if
CI02 doses higher than 2 mg/L could be
used, but such doses are not permitted in
Florida.

Granular Activated Carbon
  Four three-column GAC pilot plants were
continuously operated at the Melbourne
Water Treatment Plant. Each column was
glass, end  plates and  valves were
stainless steel, and all  interconnections
were made with Teflon* tubing. Each 5-ft-
high, 4-in.-diameter column had an empty
bed contact time (EBCT) of 10 min. Ozone
and CI02 contactors, made from Plexiglas,
were baffled into six compartments and
had a total 30-min detention time.  Flow
was constant at 190 gallons per day and
bed depth in all columns was 2 ft.
  Four different influents were compared
for  color, TOC, and THMFP removal by
GAC adsorption. The influents were raw
water, CSF water, CSF and ozonated water,
and CSF and CIO2 treated water. These
systems  are denoted as RW/GAC,
CSF/GAC, CSF/03/GAC, and CSF/CICV
GAC. The  CSF/03/GAC  system was
more effective than  the  other three
systems (see Table 1),  removing 11%
more THMFP, 10% more TOC, and 16%
more color during the 473 days of
operation than the next most effective
system (CSF/GAC). The additional re-
movals occurred entirely in the O3 con-
tactor and were due solely  to ozonation.
The effluent SPC's and dissolved oxygen
(DO) decreases from the columns indi-
cated a  higher biological activity in the
CSF/Os/GAC  columns compared with
that in the CSF/GAC columns. Thus
biological activity in the GAC columns did
not appear to be a significant means of
color, TOC, or THMFP removal.  03 and
CI02 pretreatment alone  did increase
removals of THMFP, TOC, and color, but
they also reduced the absorbability of the
organics  and therefore  the overall
efficiency  of the  GAC  systems.  Ozone
was 2 to 3 times as effective as CIO2 for
reducing THMFP, TOC, and  color. Chang-
ing to alum coagulation removed more
THMFP such that the GAC efficiency was
increased fivefold. This and the relative-
ly poor  performance of the  RW/GAC
system compared with the CSF/GAC
system indicate that optimum treatment
by GAC could be achieved by positioning
GAC columns following coagulation and
before any oxidation processes, and by
maximizing coagulation  for color, TOC,
and THM precursor removal. The percen-
tage of color removal was generally twice
the TOC or THMFP removal in the GAC
columns.
  DO and  SPC  measurements taken
during the  GAC pilot plant experiments
indicate that significant biological activity
existed in  all the GAC columns but that
the  removal of color,  TOC, or THM
precursors generally decreased with
time. Complete exhaustion of the RW/GAC
columns for TOC removal occurred even
though significant biological activity was
always present in the  columns. These
observations suggest that  a  removal
mechanism other than biological activity.
Table 1.    GAC Capacities to Meet Color Standard and THM Regulation Under Melbourne Test
          Conditions
      principally physical adsorption, is respon-
      sible for color, TOC, and THM precursor
      removal, and that eventual exhaustion of
      these columns would be  reached at
      Melbourne.

      Conclusions
        Because chlorination must be used to
      remove color remaining after coagulation,
      the magnesium carbonate process  will
      produce excessive THM's. Thus  it will be
      difficult to use this system as a potable
      water process and still meet the THM
      regulation.  Oxidation of the recycled
      Mg(HCO3)2,  the  raw water, or  the CSF
      water could not reduce the InstTHM to
      the 0.10  mg/L  regulation. The THMFP
      model did accurately predict THM forma-
      tion  and showed  that, as  a  primary
      disinfectant, chlorine could only be used
      for a short  contact  period to form  less
      than 0.1 mg/L THM's. Chlorine  followed
      by NH2CI  produced a  water with  the
      lowest SPC, THM's belowO.10 mg/L, and
      the highest residuals ever recorded at the
      extremities of the Melbourne distribution
      system. Biological activity  on GAC was
      not indicated to be a major mechanism of
      color, TOC,  or THMFP removal.
        The  full  report was submitted in
      fulfillment of Cooperative Agreement No.
      807704 by the Unrversity  of  Central
      Florida under the sponsorship of the U.S.
      Environmental Protection Agency.
                                        Capacities to Meet 15 CPU
                                Color Standard and 0.10 mg/L THM Regulation
System
Color (CPU/gm GAC)
THMFP (mg/gm GAC)
'Mention of trade names or commercial products
does not constitute endorsement or recommendation
for use.
CSF/GAC
CSF/Os/GAC
CSF/CIOZ/GAC
RW/GAC
CSF/GAC*
       240
       265
       264
       140
       252
       0.56
       1 04
       0.54
       0.48
       2.49
                                        ^Received only alum coagulated influent.
*USGPO:  1984-759-102-10666

-------
      J. S. Taylor, B. R. Snyder, B. Ciliax, C. Ferraro. A. Fisher, J. Hen, P. Muller, and D.
        Thompson are with the University of Central Florida. Orlando, FL 32816.
      J. Keith Carswell is the EPA Project Officer (see below).
      The complete report, entitled "Trihalomethane  Precursor Removal by the
        Magnesium  Carbonate Process," (Order No.  PB 84-191 147; Cost 25.00,
        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
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300

-------