United States
Environmental Protection
Agency
Municipal Environmental Research
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-83-042 Sept. 1983
&EPA Project Summary
Optimizing Coagulation-
Adsorption for Haloform and
Total Organic Carbon Reduction
M. J. Semmens, G. Hohenstein, A. Staples, G. Norgaard, K. Ayers, and M. P. Tyson
Coagulation and softening processes
were studied for their ability to remove
organics from Mississippi River water.
The goal was to use these processes to
reduce the organic contents of water
before expensive treatment with gran-
ular activated carbon (GAC) and there-
by extend the life of the carbon, improve
the product water quality, and reduce
the overall cost of water treatment
Miniature carbon columns were used
to evaluate the impact of various pre-
treatment strategies on the life and
performance of carbon for organics
reduction. Assessments were made of
(1) removal of specific trace organics
from the river water by coagulation, (2)
effectiveness of various coagulants,
coagulant aids, and lime softening in
reducing the organic contents of river
water, (3) the nature of the organics
removed by coagulation and softening,
(4) the effects of these processes on the
behavior of minature carbon columns,
and (5) characterization of the organics
removed by polarity and molecular
weight fractionation studies.
This Project Summary was developed
by EPA's Municipal Environmental Re-
search Laboratory, Cincinnati, OH, to
announce key findings of the research
that is fully documented in a separate
report of the same title (see Project
Report ordering information at back).
Removal of Specific Trace
Organics from Mississippi River
Water by Coagulation
Trace organic removal during coagulation
was evaluated using a conventional jartest
procedure. Four C14-labeled model com-
pounds were studied: phenol, benzoic
acid, salicylic acid, and octanoic acid.
These compounds were used to spike
river water samples so that removal studies
could determine how effective coagulation
was in reducing their concentration. Alum
and ferric sulfate were tested as primary
coagulants, and a study was conducted to
identify the value of a cationic polymer
(purifloc C31) when used as a coagulant
aid. All studies were conducted on the
same river water sample so that the relative
effectiveness of different operating condi-
tions could be tested.
Alum and ferric sulfate were equally
effective as coagulants on a weight basis
for the reduction of total organic carbon
(TOC) in river water, and the optimum pH
for these coagulants was approximately
5.0 and 4.0, respectively. The addition of
a polymer enhanced the effectiveness of
coagulation for TOC and turbidity removal;
it also eliminated the pH dependence of
turbidity removal so that good removals
were achieved over the entire range studied
(pH 4 to 9). The most benefit was gained
from a 1 -mg/L polymer dose. As polymer
doses were raised to 5 mg/L, the benefit
of the added polymer diminished. The
polymer improved the performance of
coagulation more when the primary coag-
ulant was ferric sulfate.
The removals of model organic com-
pounds (100 ijg/L initial concentration)
were relatively poor and in most cases did
not show a strong dependence on coagu-
lant dose. Removals of test compounds
were typically; 10%, except for salicylic
acid, which was removed slightly better
with higher alum dosages.
The influence of pH on the removal of
trace organics provided an interesting
result. Benzoic acid and octanoic acid
-------
showed either very slight or no pH depen-
dence, whereas phenol and salicylic acid
were removed best in the pH range of 5 to
7. Complex formation may best account
for the observed removal behavior.
A molecular weight fractionation study
indicated that coagulation removed the
higher-molecular-weight fractions of the
naturally occurring organic matter. Low-
molecular-weight (less than 100 nominal
molecular weight) organics were only re-
duced by 10% to 20%, even at high
coagulant dosages.
Organics Removal by Coagulant
Aids
A variety of jar test studies were con-
ducted on a single water sample to evaluate
the relative effectiveness of organics re-
moval during coagulation under various
operating conditions. The river water
sample tested was freeze concentrated
and reconstituted to ensure a constant
water quality during the study period.
Various coagulants and coagulant aids
were studied.
Polymers were studied in detail as pri-
mary coagulants and as coagulant aids.
Polymers were relatively ineffective coag-
ulants when used alone and were more
effective when used with alum or ferric
sulfate. Polymers were most beneficial
when used with low doses of coagulant,
and they improved the performance of
ferric sulfate more than that of alum.
Polymers with different charge densities
and chemical structures were evaluated.
The cationic polymers tested were superior
to anionic and nonionic polymers, which
caused little or no improvement in organics
removal during coagulation. Several
cationic polymers were selected for their
differences in chemical structure and
molecular weight and were tested. Though
the improvements in organics removals
caused by the use of these polymers were
very similar, a slight improvement was
noted with increasing molecular weight
and increased charge density.
The order of polymer and coagulant
addition did not influence organics removal
signficantly. The slow addition of polymer
resulted in slightly better performance
than a single, one-shot addition of polymer
of equal dose.
Activated silica was tested as a coagulant
aid m conjunction with 20 mg/L alum and
provided no benefit at pH 5.0.
Multistage coagulation studies were
conducted at pH 5, 7, and 9. A marginal
benefit was observed for multistage coag-
ulation at the higher pH values, but the
extra capital costs required to implement
the process would outweigh any benefit
gained in improved organics removal.
Organics removal by alum was a largely
reversible process. Coagulant added at pH
9 may remove more organics if the pH is
later adjusted toward 5. Conversely, a
coagulant used at a low pH may release
organics into solution again if the pH is
raised.
Softening Studies
Batch softening studies were conducted
to identify the impacts of operating condi-
tions on the effectiveness of organics
removal from Mississippi River water. In
addition, hardness removals were docu-
mented, and the relationships between
hardness removal objectives and organics
removal were identified.
Organics and hardness removals were
not influenced by the way in which the
lime was added. Lime added as a dry
powder gave the same performance as a
lime slurry addition.
The rate of organics removal was quite
rapid, and more than 95% of the observed
removal was complete within 30 minutes.
The Ca 00% precipitated during recarbona-
tion did not contribute to organics removal.
Alum addition with the lime improved
the removals of both hardness and organics,
but if too much alum was added, the
resulting drop in pH caused a deterioration
in performance. The order of alum and
lime addition was unimportant.
A cationic polymer, ARCO 6320P, was
tested with lime plus alum. The polymer
caused a slight deterioration in removal
performance for both hardness and organics.
Organics removal during softening was
most strongly influenced by the operating
pH. Softening at a high pH gave better
organics removal.
The primary mechanism of organics
removal appeared to be precipitation of
the organics as calcium and magnesium
complexes at high pH. Evidence suggests
that adsorption of organics to a Mg (OH)2
precipitate is not a major mechanism for
organics removal.
Nature of Organics Removed by
Chemical Pretreatment
A single large water sample was collected
from the Mississippi River and subjected
to various treatments before application to
carbon columns. The pretreatments in-
cluded alum coagulation, iron and polymer
coagulation, and lime softening. Water
samples were taken at three stages: (1)
the raw river water, (2) following pre-
treatment, and (3) following carbon ad-
sorption. The samples were analyzed to
determine the character of organics re-
moved by each of the processes.
The organic content of each sample was
fractionated according to molecular weight
and polarity; 40% of the organic content
of the river water was hydrophobic and
60% was hydrophilic. In both fractions,
the organics were predominately acidic
and neutral in character. The molecular
;weight distribution of the river water for •
ithefractions<1 K,* 1 to 10K, 10to 100K, •
iand >100K was 23%, 50%, 10%, and :
1 7%, respectively.
Coagulation either with alum or with iron
and polymer gave similar performances.
Hydrophilic acids were completely removed,
and hydrophobic acids were reduced by
approximately 60%. The neutral fractions
were also reduced. Lime, by comparison,
was not as effective in removing the acidic
fractions.
Following GAC treatment where the pH
was adjusted to favor adsorption, the
organic content of the water was essentially
independent of the pretreatment provided.
Pretreatment reduced the higher-mole-
cular-weight fractions most effectively
and did not reduce the <1 Kfraction. Indeed,
lime addition increased the concentration
of organics that were less than 1 K. GAC
completely removed the remaining organics
in the 1 to 1 OK fraction and about 60% of
the organics in the <1 K material.
Trihalomethane(THM) precursors were
associated mainly with the acidic and
neutral fractions with a molecular weight
<10K. The removal of THM precursors
bore a strong correlation with the removal
of TOC from the different fractions.
A fingerprint of low-molecular-weight,
volatile organic compounds was made for
each water sample tested following a
technique developed by Suffete?al. [Environ.
Sci. Technol. /2(12), 1 31 5 (1 978) ], but
no distinctive fingerprints were obtained,
probably because the extracts were not
methylated. The results show little dif-
ference in character for waters receiving
different pretreatments, and the analytical
methods used were not capable of deter-
mining the actual character of the various
samples.
Activated Carbon Column
Performance Studies
Through the use of miniature carbon
columns containing 3 to 7 mL of 50 x 80
mesh GAC (WestvacoWV-G), breakthrough
curves were obtained for TOC, ultraviolet
absorption (UVA), and THM precursors in
1 0 to 30 days. By operating four columns
in parallel and comparing the column per-
formance for different pretreatment condi-
tions on the same water sample, the
*K= thousand
-------
influence of pretreatment on carbon ad-
sorption was identified. In some studies,
the removal behavior of synthetic resins
was compared to that of activated carbon
for pretreated river water.
The removals of UVA, TOC, and THM
precursors by activated carbon all followed
the same trends. The relationship between
these parameters varied somewhat with
different levels of treatment, but each
parameter showed the same trends in all
the studies. Thus a pretreatment that was
best for UVA removal was also the best for
TOC and THM precursor removal. This
result suggests that simple analysis of
UVA removal might be used as an indicator
of the relative removal of TOC and THM
precursors once a detailed correlation is
established.
The removal performance of activated
carbon was strongly influenced by both
solution pH and coagulant dose. Effective
carbon adsorption capacity improved as
the pH was reduced. At pH 5, the capacity
was approximately three to four times
greater than that at pH 8 if column opera-
tion was terminated at 1 mg/L effluent
TOC. If no coagulation was provided, the
carbon was unable to reduce the TOC of
the raw river water to an acceptable level,
and high TOC concentrations broke through
immediately. When coagulation was em-
ployed as a pretreatment, however, the
organic content of the river water was
significantly reduced and the carbon
columns were able to produce an effluent
containing very low concentrations of TOC.
In addition, carbon service life increased
with increasing coagulant doses, because
higher doses caused a greater reduction in
the influent organic concentration. A simple
cost analysis was conducted and used to
illustrate the concept of an optimum pH
value and coagulant dose for pretreating
Mississippi River water before GAC treat-
ment.
Lime softening improved the carbon
adsorption capacity for organics removal
from Mississippi River water. As with the
coagulation studies, increased levels of
organic removals in pretreatment lead to
longer carbon service life. Carbon adsorp-
tion capacities increased with increasing
lime dose during pretreatment, provided
the pH was lowered for adsorption.
For Mississippi River water, lime soften-
ing was a much less effective pretreatment
process for organics removal than coagula-
tion. The data obtained indicate that lime
softening with 1 30 mg/L of high calcium
quicklime and 7 mg/L of alum (as practiced
at the Minneapolis Water Treatment Plant)
produced organics removals roughly com-
parable with those yielded by 20 mg/L of
alum at pH 5. A coagulant dose of 80
mg/L of alum gave far superior removal
performance and extended the useful life
of the GAC columns.
Two weakly basic resins in the free base
form were inferior to activated carbon for
adsorption of TOC and THM precursors
from coagulated river water. These studies
were conducted at pH 7, however, and the
resins' performance might be improved by
operating at different pH values.
A column of strongly basic anion ex-
change resin (IRA 904 in the chloride
form) was as effective as activated carbon
for both TOC and THM precursor removal
from coagulated river water samples. The
resin was also regenerated effectively with
sodium chloride, but further studies are
needed identify the best regeneration
conditions.
The full report was submitted in fulfill-
ment of Cooperative Agreement No. CR-
806377 by the University of Minnesota
under the sponsorship of the U.S. Environ-
mental Protection Agency.
M. J. Semmens, G. Hohenstein, A. Staples, G. Norgaard, K.Ayers, andM. P. Tyson
are with the University of Minnesota, Minneapolis, MN 55455.
Leown Moore is the EPA Project Officer (see below).
The complete report, entitled "Optimizing Coagulation-Adsorption for Haloform
and TOC Reduction," (Order No. PB 83-210 757; Cost: $20.50, 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
irUS GOVERNMENT PRINTING OFFICE 1983-659-017/7175
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
RETURN POSTAGE GUARANTEED
PS 0000329
U S ENVIR PROTECTION AGENCY
REGION 5 LIbRAKY
230 S DEARBORN STREET
CHICAGO IL 6060a
------- |