United States
Environmental Protection
Agency
Municipal Environmental
Research Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-84-093 Sept. 1984
Project Summary
Barium and Radium Removal from
Groundwater by Ion Exchange
Vernon L. Snoeyink, Julie L. Pfeffer, David W. Snyder, and Carl C. Chambers
A study was undertaken to develop
technology that can be used by small
water treatment plants to remove hard-
ness, barium, and radium"'. Special
emphasis was placed on finding an alter-
native to strong acid ion exchange
(which is used in the Na* form) because
this process adds large amounts to Na*
to the treated water. The primary objec-
tive of this study was to determine the
applicability of weak acid ion exchange
resin for removal of hardness, barium,
and radium from the types of ground-
water encountered in northern Illinois.
The capacity of the resin and the
regeneration requirements were to be
determined and compared with those of
strong acid resins for the same ap-
plication.
Additional tasks included (1) evalu-
ating the performance of the strong acid
resin now used at Crystal Lake, Illinois,
to remove hardness and barium, (2)
determining the barium and radium re-
moval efficiencies of strong acid ion ex-
change softeners used in homes, and (3)
modifying the surface of activated car-
bon to make it suitable for the selective
removal of barium.
Both strong and weak acid resin sys-
tems were very effective in removing
"*Ra and Ba". The weak acid system in
the H* form does not add Na* to the
water as does the strong acid system,
but the weak acid system will cost more
to use because of the need for acid-
resistant materials and the CO2
stripping.
This Project Summary was developed
by EPA's Municipal Environmental Re-
search 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
Naturally occurring barium in drinking
water exceeds the maximum contaminant
level (MCL) in some areas of northern Illinois
and northeastern Iowa. In these same areas
and in some parts of Florida, the concentra-
tions of radium228 exceed the MCL of 5 pico-
Curies (pCi)/L. Most of the contaminated
supplies are used by small communities,
many of which do not presently treat their
water to reduce the concentrations of these
substances. Both radium226 and barium are
alkaline earth metals, and both are found in
water as divalent cations. Their chemical
behavior is very similar, and it is much like
that of Ca2* and Mg2*, the principal com-
ponents of hardness in water. Thus, pro-
cesses used to soften water are very useful
for removing these contaminants from drink-
ing water.
A process particularly suited to small com-
munities where hardness, barium, and/or
radium are a problem is ion exchange. The
objective of this research was to investigate
the applicability of the ion exchange process
to this problem in northern Illinois. Strong
acid resins in the sodium form were to be
evaluated, but because their use results in
significant increases in the Ma* content of the
water, their performance was to be com-
pared with that of weak acid resins in the
hydrogen form. The latter can remove only
the equivalents of divalent cations equal to
the equivalents of alkalinity present, and ion
exchange must be followed by C02 stripping
and pH adjustment.
The resins were tested with an influent
water containing hardness of approximately
200 mg/L as CaC03, total alkalinity of 250
mg/L as CaC03,20 mg/L Ba2*, and 20 pCi/L
-------
226Ra. The tests used a regenerant dose
typical of softening operations.
In addition to investigating the applicability
of ion exchange to northern Illinois ground-
water, this project also (1) evaluated the per-
formance of the strong acid resin now used
at Crystal Lake, Illinois, to remove hardness
and barium, (2) determined the barium and
radium removal efficiencies of strong acid ion
exchange softeners used in homes, and (3)
attempted to modify the surface of activated
carbon to make it suitable for the selective
removal of barium.
Materials and Methods
The resins used in this study were Duolite
C-20 and Duolite C-433 manufactured by
Diamond Shamrock, Redwood City, CA.
C-20 is a strong acid resin with a polystyrene
matrix and sulfonate functional groups.
C-433 is a high-capacity weak acid resin with
a polyacrylic matrix and carboxyl functional
groups. Their capacities are 4.8 and 11.5
meq/g dry resin, respectively, and their prop-
erties are typical of resins made by a number
of manufacturers. The weak acid resin can
be used in solutions with pH > 5, and the
strong acid resin can be used for pH > 0.
The solutions used for most of the column
tests contained approximately 100 mg
Mg2*/ L as CaC03,100 mg Ca2*/ L as CaC03,
250 mg/L total alkalinity, 20 mg Ba2VL, and
20 pCi/L of 226Ra. Solutions for selectivity
determinations were prepared using reagent-
grade chemicals to the specifications re-
quired by the test.
Ion chromatography was used for quan-
titative analysis of cations. Hardness and
alkalinity were also determined in accordance
with Standard Methods for Analysis of
Water and Wastewater (15th edition, Amer.
Public Health Assoc., 1980). Some samples
were also analyzed by atomic absorption
spectroscopy. The 22eRa samples were ana-
lyzed at the Environmental Research Labo-
ratory, University of Illinois, using the radon
emanation method, and at the University
Hygienic Laboratory, University of Iowa, by
a technique involving coprecipitation of the
226Ra with barium sulfate and by alpha count-
ing with an internal proportional counter.
Experimental Results
Strong Acid Resins
Isotherms and column tests were used to
determine the capacity of the strong acid
resin with 4.8 meq/g dry resin (hydrogen
form). The selectivity sequence was Ba2* >
Ca2* > Mg2* > Na* > H*. Alkalinity had no
effect on capacity, but the capacity for
divalent cations decreased as the concentra-
tion of sodium increased.
Column tests were run using 2.5- x 62-cm
columns with the resin in both the Na* and
H* form. Application of the test solution at
rates of 2.5 to 5 bed volumes (BV)/hr to
virgin resin in the Na* form gave the
breakthrough curves shown in Figure 1. The
curves for the H* form of the resin were
similar except for the Na* concentration in
the effluent. In keeping with the selectivity
sequence, Mg2* is the first divalent cation to
appear in the effluent, followed by Ca2*, and
much later by Ba2*. Regeneration of the H*
form of the resin with 8 percent HCI at 1.6
BV/hr gave the results that appear in Figure
2; results for regeneration of the Na* form
- 120
Figure 1.
400
600 1200 1800
Volume of Influent (Bed Volumes)
Breakthrough curves for virgin strong acid resin in the sodium form.
500 7000 7500 2000
Cumulative Regenerant Applied (meq of HCI)
2500
Figure 2.
Mention of trade names or commercial products does
not constitute endorsement or recommendation for use.
Regeneration of the hydrogen form of the strong acid resin with HCI. AtA.B, C. andD.
the equivalents of HCI added per equivalent of divalent cation originally on the resin is
1. 2, 3, and 4, respectively.
-------
with NaCI gave very similar results. At a
regenerant dose equal to the equivalents of
divalent cation on the resin, 55 percent of
the ions were removed, including 80 percent
of the Mg2*, 67 percent of the Ca2*, and 15
percent of the Ba2*. A dose of three times
the equivalents of the divalent cations on the
resin was necessary to remove 50 percent
of the Ba2*.
Several exhaustion-regeneration cycles
were then carried out to determine the
behavior of the strong acid resin at steady-
state conditions. The influent was applied at
a rate of 17 BV/hr, and a typical NaCI
regenerant dose for softening (6.5 Ib/ft3, or
4.7 meq/g) was applied co-current at 1.7
BV/hr, followed by a slow rinse at 1.7 BV/hr
for 48 min and a fast rinse at 7 to 8 BV/hr
for 26 min. The Ba2* accumulated on the
resin over successive cycles and thus caused
the Ba2* to break through much earlier than
it did with virgin resin. The hardness and Ba2*
breakthrough curves that developed after
several cycles appear in Figure 3. The curves
show that approximately 225 BV of water
were processed before hardness and Ba2*
broke through at about the same time. The
regeneration efficiency (equivalents of di-
valent cations removed per equivalents of
NaCI applied) was 58 to 59 percent. Column
utilization (column capacity used/maximum
capacity) was about 58 percent; a large por-
tion of the remaining 42 percent was oc-
cupied by Ca2* and Ba2* that was not
removed by the regenerant dose.
The data in Figure 4 show the Ba2* con-
centration (when CaC03) effluent hardness
was approximately 40 mg/L as CaC03 as a
function of the number of cycles. The Ba2*
concentration leveled off at 1.7 mg/L for a
regenerant dose of 6.5 Ib NaCI/ft3 (4.7
meq/g). An HCI regenerant dose of 4.06 Ib
HCI/ft3 (4.7 meq/gj was tested on a second
column for a few runs, and similar results
were obtained. Increasing the regenerant
dose by 50 percent to 9.75 Ib NaCI/ft3 re-
duced the Ba2* to about 1.2 mg/L when the
hardness was 40 mg/L as CaC03. The re-
generation efficiency dropped to 46 percent;
however, the column utilization increased to
68 percent. For the regenerant dose of 6.5
Ib NaCI/ft3, 2.4 BV of brine with total dis-
solved solids (TDS) of 16,400 mg/L were
produced per 100 BV of product water.
The conclusion that Ba2* can be removed
to less than the MCL of 1 mg/L as long as
the strong acid resin is not saturated with
hardness was confirmed by observations at
the Crystal Lake, Illinois, municipal ion ex-
change system. The influent and effluent of
several home ion exchange softeners were
also sampled, and these too showed Ba2*
removal below the MCL.
40
o 30
o
20
i
10
Total Hardness - D
Barium - •
MCL
-J_
2.0
1.5
1.0
0.5
50 700 750 200
Volume of Influent (Bed Volumes)
250
Figure 3. Breakthrough curves for hardness and barium on a strong acid resin in the sodium
form after several exhaustion-regeneration cycles.
2.0
1.5
5,0
I
I
| 0.5
Regenerant Dose:
^ 6.5Ib NaCI/'ft* i 9.75 Ib NaCI/ft*
^4.61 or 4.06 Ib HCI/ft*' "~
MCL
I I I
I
o — Na Form Column
A — H Form Column
(HCI Regenerant)
I I I I I i I
8 10 12
Number of Cycles
14
16
18
Figure 4.
Barium effluent concentrations for the strong acid resin exhaustion-regeneration
cyclic runs. The barium values shown are those when the effluent hardness
concentration is 40 mg/L.
Weak Acid Resins
The maximum capacity of the weak acid
resin was determined to be 11.5 meq/g dry
resin, but the extent of swelling of the resin
determined how much of the capacity could
be used. For example, exposure of the resin
in the H* form to a solution containing only
BaC03 salt resulted in a capacity of only 7.5
meq/g, whereas carbonate salts of Ca2* and
Mg2*, which cause the resin to swell more
-------
because of their larger hydrated radii, used
the full 11.5 meq/g. Addition of salts to the
Ba2+ solution increased the swelling and
allowed Ba2* to use the full capacity.
Application of test solution to 2.5- x 62-cm
columns of the virgin resin in the H+ form
gave the breakthrough curves that appear in
Figure 5. The equivalents of alkalinity per liter
in the product water exceed the equivalents
of divalent cations per liter, so the initial
removal of Ca2+, Mg2+, and Ba2+ was com-
plete. Their order of appearance in the ef-
fluent was in reverse order of the selectivity
series, Ba2+ ^ Ca2* > Mg2t, and the
breakthrough curves were less steep than
those for the strong acid resin, a result
caused by a slower rate of exchange by weak
acid resins.
A major advantage of using a weak acid
resin is the ease with which it can be re-
generated by strong acid. The data in Figure
6 show that application of one equivalent
HCI per equivalent of divalent ion on the resin
resulted in removal of 90 percent of the ions,
and 1.5 equivalents HCI per equivalent of
divalent cation gave essentially complete
removal.
Several exhaustion-regeneration runs
were then made to determine resin behavior
under continuous operation. Influent was ap-
plied at about 17 BV/hr; three cycles were
completed, each with a different regenerant
dose. The breakthrough curve in Figure 7
resulted after steady-state behavior was ob-
tained for the cycle using 8.5 meq HCI/g.
More than 650 BV of product water were ob-
tained before effluent hardness reached a
level of 40 mg/L as CaC03. The Ba2+ con-
centration leveled off at 0.2 mg/L, as shown
in Figure 8. For this regenerant dose, column
utilization was about 70 percent, and regen-
eration efficiency was about 95 percent. The
30 percent of column capacity that was not
used was attributable to slow exchange
kinetics rather than to buildup of Ca2* ions,
as it was for the strong acid resin. Approxi-
mately 1.2 BV of spent brine was produced
with a TDS of 19,900 mg/L per 100 BV of
product water.
Application of less regenerant than the
equivalents of divalent cations removed led
to a higher Ba2+ concentration at hardness
breakthrough and to a high leakage of Ba2*
and hardness at the start of the subsequent
run.
Radium Removal
Radium226 removal efficiency for both the
strong and weak acid resins was excellent.
The concentration was reduced from the 20
pCi/L in the influent to much less than the
MCL of 5 pCi/L, even after the resins were
saturated with hardness and Ba2+. Analysis
of the 226Ra in the spent regenerant showed
900 7800
Volume of Influent (Bed Volumes)
2700
Figure 5. Breakthrough curves for weak acid resin in the hydrogen form.
800
500 1000 A 1500 B 2OOO
Cumulative Regenerant Applied (meq of HCL)
2500
Figure 6.
Regeneration of weak acid resin with HCI. A and B represent 1.0 and 1.5 equivalents
of HCI applied per equivalent divalent cation on the resin.
that all226 Ra removed during the exhaus-
tion cycle was extracted from the resin dur-
ing regeneration. Some 226Ra accumulated
on the strong acid resin during the first few
exhaustion-regeneration cycles, but after
several cycles, no additional accumulation
was apparent, and the resin still performed
satisfactorily.
Influent and effluent samples taken from
home ion exchange softeners using NaCI
regeneration also showed that they removed
226Ra to below the MCL.
Cost
The cost of using strong acid ion ex-
change softening with NaCI as the regen-
erant and the type of water used in the
laboratory study was estimated to be
$1.36/1000 gal and $0.38/1000 gal for a 0.1
and 1 MGD facility, respectively, for the type
-------
I
40
Barium
Hardness
MCL /^
200 400 600
Volume of Influent (Bed Volumes)
800
Figure 7.
Breakthrough curves for barium and hardness on a weak acid resin in the hydrogen
form after several exhaustion-regeneration cycles.
2.0
1.5
i
I
1.0
0.5
Regenerant Dose
• 9.0 meq/g
D 7.6 meq/g
8.5 meq/g
0246
Number of Cycles
Figure 8. Barium effluent concentration
for the weak acid resin exhaus-
tion-regeneration cyclic runs.
The barium values shown are
those that occur when the efflu-
ent hardness concentration is
40 mg/L as CaCO2-
of water used in the laboratory study. The
additional cost for using a weak acid system
with HC\ regeneration will be $0.15 and
$0.08/1000 gal, respectively.
The system designs were based on the
performance observed when 4.7 meq
NaCI/g and 8.5 meq HCI/g were used to
regenerate the strong and weak acid resins,
respectively. The flow scheme for the weak
acid system included the resin columns, a
CO2 stripping column, pH adjustment, and
chlorine addition. The strong acid resin
system was similar except that the C02 strip-
ping column was excluded. The costs were
based on cost data supplied by the Illinois
Water Treatment Company in Rockford, Il-
linois. The strong acid ion exchange units in-
cluded the columns, the resin, regenerant
day tanks, brine tanks, and interface piping.
The weak acid unit included all of these plus
a carbon dioxide stripping tower. A com-
parison of the two types of exchange sys-
tems was used to evaluate the additional
costs required for weak acid exchange
systems at these different treatment ca-
pacities. Complete treatment costs for strong
acid ion exchange softening were deter-
mined from previously published data and
appropriate price indices.
The costs do not include the costs of
pumping raw, in-plant, or finished water.
Brine disposal was also not included, since
costs are very site specific. They do include
the cost of adjusting product water pH,
however, and this may not be necessary in
all cases.
Conclusions
After several exhaustion-regeneration
cycles at a regenerant dose typical of soften-
ing operations, Ba2* and 228Ra were effec-
tively reduced well below their MCL's. For
both types of resin, significant concentra-
tions of Ba2* appeared in the effluent at
about the same time as did hardness, but
226Ra continued to be removed even after
saturation of the resins with hardness. These
conclusions were consistent with the perfor-
mance of an operating municipal ion ex-
change plant in northern Illinois and with the
results obtained using home ion exchange
softeners.
Though both strong and weak acid resin
systems can effectively remove hardness,
226Ra, and Ba2*, the weak acid system in the
H* form does not add Na* to the water as
does the strong acid. The weak acid system
will cost more to use, however, because of
the need for acid-resistant materials and the
C02 stripping. But this cost is offset as the
plant size increases. The additional cost for
the weak acid system declines from about
$0.15/1000 gal for a 0.1-MGD plant to
$0.08/1000 gal for a 1-MGD facility. In ad-
dition, the weak acid resin will remove only
the divalent cations that are balanced by an
equivalent amount of alkalinity. If a water
contains divalent ions in excess of the al-
kalinity, the hardness left in the water may
be desirable because it should make the
water less corrosive during distribution.
Whether the weak acid resin will remove Ba2*
from such a water must still be shown,
however.
Conversion of an existing system that uses
strong acid resin in the sodium form to one
that uses weak acid resin in the hydrogen
form must be done with care. In most cases,
acid-resistant materials will have to be in-
stalled, and provision must be made to neu-
tralize the spent acid before its discharge.
Additional design considerations include that
the capacities of the two resins are different
and that the weak acid resin is subject to
more swelling. Use of a weak acid resin will
also require the installation of a C02-stripping
process.
*USGPO: 1984-759-102-10663
-------
Vernon L Snoeyink, Julie L Pfeffer, David W. Snyder, and Carl C. Chambers are
with the University of Illinois, Urbana, IL61801.
Richard Lauch is the EPA Project Officer (see below).
The complete report, entitled "Barium and Radium Removal from Groundwater by
Ion Exchange." (Order No. PB 84 -189 810; Cost: $ 14.50, 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
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES P/
EPA
PERMIT No G-35
Official Business
Penalty for Private Use $300
U'U H
ST
60604
------- |