United States
Environmental Protection
Agency
Water Engineering
Research Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/S2-88/007 Apr. 1988
Project Summary
Removal of Chromium from Ion
Exchange Regenerant Solution
Susan K. Siegel and Dennis A. Clifford
A pilot study in Scottsdale, AZ,
determined that chromate could be
efficiently removed from groundwater
using chloride-form anion exchange
resin. A typical run length exceeded 30
days after which the column could be
regenerated with 3-5 bed volumes of
1 N (6%) NaCI. The purpose of the
present study was to establish means
of removing the chromate (Cr(VI)) from
the spent regenerant solution, thus
rendering it a non-toxic brine. In this
bench-scale study, the Cr(VI) was
reduced to Cr(lll) and then precipitated
as Cr(OH)3(s). Sulfite, hydrazine, and
ferrous sulfate were tried as reduc-
tants. Sulfite and hydrazine operated
best at pH < 2 while ferrous sulfate
performed well in the neutral pH range.
Sludges from all of the reduction
processes settled well and settling was
improved with increasing NaCI concen-
tration. All three sludges passed the
extraction procedure (EP) toxicity test
when evaluated for chromium teacha-
bility. Ferrous sulfate reduction proved
to be the lowest cost treatment method
for the regenerant brine. For a 4 MGD
treatment system utilizing 70% bypass
flow and reducing the chromium from
the 0.050 mg/L to 0.035 mg/L in the
blended product water, the spent
regenerant brine treatment cost was
quite low— $1.50/million gal of pro-
duct water. For a 0.1 MGD treatment
system, the corresponding cost was
$4.60/million gal of product water.
These estimates include only the costs
for ferrous sulfate and Cr(OH)3 sludge
land-filling. Pilot-scale studies of the
entire chromate ion exchange removal
system including brine treatment and
possible reuse are recommended
before a full-scale system is designed.
This Project Summary was devel-
oped by EPA's Water Engineering
Research Laboratory. Cincinnati, OH.
to announce key findings of the
research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering
information at back).
Introduction
Hexavalent chromium of natural origin
is found in the ground water in several
Arizona locations. In one such location,
a 75-mi2 area in Paradise Valley encom-
passing Scottsdale, AZ, the highest
concentrations exceed 0.05 mg/L, the
maximum contaminant level (MCL) for
chromium in drinking water as specified
by the U.S. Environmental Protection
Agency (EPA). Three of Scottsdale's 10
city water supply wells were found to
exceed the MCL for chromium, and
several more were just below the limit.
Because these chromium-contaminated
wells constituted a major portion of the
Scottsdale water supply, it was consid-
ered important that their use be con-
tinued. In most cases blending with low
Cr(VI) wells was possible to meet the
MCL Looking to the future and the
possibility that the marginal wells could
eventually exceed the MCL for chro-
mium, treatment by ion exchange and
desalting processes were studied for
Cr(VI) reduction.
EPA-funded, pilot-scale chromate
removal studies performed by the Uni-
versity of Houston (UH) in the UH/EPA
Mobile Drinking Water Treatment Facility
showed that excellent Cr(VI) removal was
obtained by ion exchange using a mac-
roporous strong-base anion resin in the
chloride form. However, this treatment
process generated a waste brine contain-
ing 1.5%-12% NaCI and 100-400 mg/L
Cr(VI). In order to dispose of this poten-
tially toxic brine, it was necessary to
evaluate means of removing the Cr(VI)
-------
and rendering it non-toxic. The detoxified
brine could then be disposed of in an
evaporation pond or the local sanitary
sewer. It also would have a reuse
potential.
The bench-scale study of Cr(VI) re-
moval from brine described in this report
included three phases: (1) evaluation of
reducing agents and optimum pH con-
ditions for the reduction of Cr(VI) to Cr(lll),
(2) determination of optimum conditions
for precipitation of Cr(OH)a(s), and (3) an
economic evaluation of the process
including reduction and precipitation.
Experimental Details
Optimum pH. Reductant
Dosage, and Reaction Time
Jar Tests-
Acidic sulfite, ferrous iron, and acidic
hydrazine were tried as reductants. The
reduction reactions of Cr(VI) to Cr(lll)
using these reagents are shown below.
(1)
(2)
(3)
Table 1.
HCr04"
Cr3+
3Fe2*
3Fe3+
+ 7H+
4H20
+ + 3N2+16H20
In the three reactions above, pH adjust-
ment to 8.3 following the reduction
resulted in the formation of a CrfOHJafs)
precipitate, which was flocculated,
settled, and dewatered. Ferrous iron was
used in the neutral to alkaline pH range
and the precipitate formed during the
reduction step contained both Cr(OH)33(s).
All of the experiments in this study
were 750-mL jar tests conducted at room
temperature (22° to 25°C) with an initial
Cr(VI) concentration of either 100 mg/L
or 364 mg/L. Reagent grade chemicals
dried at 105°C were used for all the test
procedures. In the jar tests, solutions of
Cr(VI) in 2.0 M or 1 .0 M NaCI brines were
reduced to Cr(lll) following the addition
of sodium sulfite, ferrous sulfate, or
hydrazine. The composition of the arti-
ficial ion-exchange regenerant solution
is shown in Table 1. It is based on the
composition of an actual spent-
regenerant solution from the Scottsdale
pilot study. A magnetic stirrer was used
•to rapidly mix the samples at 150 rpm
for 5 min during addition of the reductant.
Then the pH of the Cr(lll) solution was
Makeup of the Artificial Ion Exchange (IX) Spent Regenerant Solutions for Us
in Cr(VI) Removal Experiments
364 mg/L Cr(VI); 2 M NaCI
100 mg/L Cr(VI); 1 M NaCI
mg/L
meq/L
mg/L
meq/L
Cations
Na*
/T
Total Cations
Anions
CrtO?
cr
HC03
Total Anions
TDS
47,725
273
47.998
757
71.000
4,575
76,332
124.329
2,075
7
2.082
7
2,000
75
2,082
24.725
75
24,800
208
35.500
4,575
40.283
65,083
1,075
1.9
1,076.9
1.9
1,000
75
1,076.9
Note: Potassium IK) was in the artificial spent regenerant solution because it was made u,
from KsCrzOj. Potassium would not be in a typical spent regenerant solution and it
presence here is considered unimportant.
increased to 8.3 by the addition of either
NaOH or Ca(OH)2 while the solution was
rapidly mixed on the magnetic stirrer for
approximately 5 additional min. The
Cr(lll) solutions were then flocculated by
mixing at 40 rpm for 20 min. Finally, the
samples were allowed to settle quies-
cently for at least 2 hr. The experimental
system is shown in Figure 1.
At the end of the settling period, a
supernatant sample was collected, fil-
tered through a 0.45 /urn membrane filter,
acidified with HN03 to a pH < 2.0, and
then stored in a plastic bottle at 4°C. The
concentration of total chromium in each
sample was analyzed with a Perkin-
Elmer Model 5000 graphite furnaci
atomic adsorption spectrophotomete
(GFAAS) with Zeeman backgroun
correction.*
Optimization Tests—
Jar tests using varying stoichiometri
amounts of sodium sulfite and ferrou
sulfate were conducted to determine th
optimum dosage. The dosages used wer
0.0, 0.75, 1.0, 1.25, 1.5, 1.75, and 2.I
times the theoretical stoichiometri
amount (SA). A higher level of hydrazin
Mention of trade names or commercial product
does not constitute endorsement or recommende
tion for use.
Floccculation 20 min at 40 rpm
CtD
Quiescent
Settling
6 Place Magnetic Stirrer
Figure 1. Jar test experimental apparatus schematic.
-------
was used because literature supplied by
the manufacturer, Olin Chemical Com-
pany, indicated that 3 times the SA of
hydrazine would be required to give
nearly complete chromium reduction
within 2 hr. Thus, hydrazine dosage was
not optimized.
Solution pH's during the reduction
reaction were next optimized by adjust-
ing the pH of the chromium solution to
1.2, 2, 3, 4, 5, or 6 prior to the addition
of the hydrazine or sodium sulfite.
Information obtained from the literature
had suggested that the reduction of Cr(VI)
by ferrous sulfate would be in the neutral
pH range. Therefore the pH's of the Cr(VI)
solutions prior to ferrous sulfate addition
were adjusted to 5, 6, 7, 8, 9, or 10.
The reaction times for the three
reductants were then optimized by
reducing the Cr(VI) for varying periods of
time: 0.25, 0.5, 1, 2, 3, 4, 5, and 6 hr.
Both optimum and non-optimum concen-
tration of hydrazine, sodium sulfite, and
ferrous sulfate were used. Finally, the
effect of the solution NaCI concentration
on the Cr(VI) removal was determined by
performing the tests with artificial spent
regenerants containing varying amounts
of NaCI. The brine concentrations used
in the tests were 0.0, 0.1, 0.25, 0.5, 1.0,
or 2.0 M NaCI. Optimum reductant
dosages and more than sufficient reac-
tion times were used.
Sludge Testing Procedures
The gravity-settled Cr(OH)3(s) or
Cr(OH)3(s) + Fe(OH)3(s) sludges were
dewatered by either centrifugation or
filtration. A weighed sample of the
centrifuged or filtered chromium sludge
was dried at 105°C for 48 hr and cooled
in a dessicator. The dried sludge was
then weighed and the percentage of
solids in the dewatered sludge was
determined by difference.
The various sludges were tested for
Cr(total) leachability using the extraction
procedure (EP) toxicity test as specified
by EPA. Briefly, the procedure comprises
the extraction of a 100-g sample of
centrifuged or filtered sludge at pH 5 in
the presence of an acetate buffer. The
extraction period is 24 hr, during which
the sludge buffer mixture is continually
stirred.
Results and Discussion
Optimization Test Results
Figure 2 shows that the optimum
dosage for both sodium sulfite and
ferrous sulfate was approximately 1.25
1000.0
100.0 -
A FeSO*2MNaCI
• /Va2S03/Vo/VaC/
• NazS032MNaCI
to
.c
o
o
10.0 '
0.10
Figure 2.
0.010
Stoichiometric Ratio of Reductant Added
The effect of Stoichiometric reductant concentration on residual Crftotal)
concentration during /VanSO3 and FeSOt reduction. All test solutions contained
364 mg/L Cr(VI). All supernatant samples were filtered through a 0.45 urn
membrane filter.
times the Stoichiometric amount (SA). As
mentioned previously, the minimum
hydrazine concentration was not opti-
mized but was estimated to be 3 times
the SA. The initial pH of the solutions
prior to sodium sulfite or hydrazine
addition was found to be an extremely
important factor in hexavalent chromium
reduction and removal. Figures 3 and 4
show that there was > 99.9% chromium
reduction and removal (following pH
adjustment and sedimentation plus
filtration of CrfOHMs)) if the initial pH
was < 2. With an initial pH in the range
of 2-4, there was no less than 80%
reduction and removal of Cr(VI) from the
test solutions.
Regarding the reaction time required
at pH < 2, hexavalent chromium was
completely reduced, as evidenced by an
overall 99.9% removal of total chromium,
in less than 15 min with sodium sulfite,
and in less than 1 hr with hydrazine. If
the solution pH was greater than 2 prior
to hydrazine or sodium sulfite addition,
increasing the reaction time did not
increase the amount of Cr(VI) reduced
and subsequently removed as Cr(lll).
Figure 5 shows that, for optimum Cr(VI)
reduction by ferrous sulfate, pH should
-------
7000
100
CO
.c
6
C
o
o
/.O
0.70
0.070
Na^SO3 (1.25 x Stoich. Amount)
5 min Reaction Time
Hydrazine (3.0 x Stoich. Amount)
1 hr Reaction Time
Hydrazine (3.0 x Stoich. Amount)
1.5 hr Reaction Time
1.0 2.0 3.0 4.0 5.0
Solution pH Before Reductant Addition
6.0
Figure 3. The effect of pH on the efficiency of Cr(VI) reduction by NasSOa or N2Ht. All
test solutions contained 364 mg/L Cr(VI) and 2 M NaCI. All supernatant samples
were filtered through a 0.45 t*m membrane filter.
be in the range of 5-8. As with optimum
dosages of sulfite and hydrazine, greater
than 99.9% of the Cr(VI) was removed
as Cr(OH)3(s) from the pH adjusted,
settled, and filtered solution.
Sodium chloride in the test solutions
clearly improved the formation and
settling properties of the Cr(OH)s(s) floe
formed using either sodium sulfite or
hydrazine as the reductant. In contrast,
very small Cr(OH)3(s) floe particles
formed in deionized water solutions, and
significant turbidity remained in the
supernatant solution after settling. The
same was true of the deionized water
system when ferrous sulfate was used
as the reductant, but the presence of a
large quantity of readily settleable ferric
hydroxide floe rendered the enhancing
effect of the NaCI less important.
Sludge Test Results
The Fe(OH)3(s) + Cr(OH)3(s) sludge
produced in the tests using ferrous iron
as the reductant settled much faster than
the Cr(OH)s(s) sludge produced following
reduction by sulfite or hydrazine. Add!
tion of 13-133 mg/L alum to the tes'
beakers increased the rate of settling o
the Cr(OH)3(s) sludge following sulfite
reduction, but it still settled significant^
slower than the FefOHMs) + Cr(OH)s(s
sludge.
Gravity settling for 4 days gave sludges
with less than 8% solids except for the
sludge containing CaCOsfs), which gave
12% solids consisting mostly (91%) o1
CaCOs(s). Centrifugation at 3,000 rpm foi
10 min gave better dewatering sludges
i.e., from 12% to 17% solids for the
Cr(OH>3(s) and Cr(OH)3
-------
1000.0
100.0
CO
.c
I
o
10.0
0.10
0.010
O NasSOa(1.25 x Stoich, Amount). pH 1.2
O Na2S03 (1.25 x Stoich. Amount). pH 2.0
A Na2SO3 (1.25 x Stoich. Amount). pH 4.0
• Hydrazine (3.0 x Stoich. Amount). pH 1.2
m Hydrazine (3.0 x Stoich. A mount), pH 6.0
pH1.2
pH1.2\
1.0 2,0 3.0 4.0
Reaction Time in Hours
5.0
6.0
Figure 4.
The effect of solution pH (before reductant addition) on the kinetics of Cr(VI)
reduction with SO3~ or N2Ht. All test solutions contained 364 mg/L Cr(VI) and
2 M NaCI. All supernatant samples were filtered through a 0.45 um membrane
filter.
3. The allowable Cr(VI) concentration
in the effluent blended water (raw
+ IX treated) is 70% of the MCL,
i.e., 0.035 mg/L.
4. The IX resin is regenerated after
treating 25,000 BV of raw water.
5. Regeneration of the resin bed is
accomplished using 5 BV of 1.0 N
NaCI and 5 BV of rinse water for
a total of 10 BV of spent regenerant
solution.
Based on the above assumptions, 30%
of the raw water must be treated by ion
exchange while the other 70% can be
bypassed and blended with the treated
water.
Two cases with differing flow rates
were analyzed. The first case is that of
a well with a capacity of 3.89 MGD (2,700
gal/min) like Scottsdale Well #32. The
second case is representative of a
smaller community, with a similarly
contaminated well having a flow rate of
0.1 MGD (69 gal/min). In both cases, an
empty bed contact time (EBCT) of 2-min
(3.74 gal/min ftj) was chosen for the
optimum design flow rate because it is
near the high end of the acceptable flow
rates (1 -5 gal/min ft3) recommended by
ion exchange resin manufacturers. Using
a 2-min EBCT for the 3.89 MGD well
yields approximately 61,000 L of spent
regenerant solution containing approx-
imately 100 mg/L Cr(VI), while the 0.1
MGD well yields approximately 1,580 L
of a similar spent-regenerant solution. In
both cases, the ion exchangers would
need to be regenerated only once in 35
days. With such a long run, prefiltration
or occasional backwashing would be
required to prevent IX column plugging.
The laboratory test results for reduc-
tion of 100 mg/L Cr(VI) under optimum
conditons were extrapolated to estimate
the amounts of chemicals necessary to
treat the 1,580 L(417 gal) and the 61,000
L (16,200 gal) volumes of regenerant
solution.
The costs of reducing the Cr(VI) to Cr(lll)
with subsequent removal from the large
and small volume regenerant solutions
are shown in Tables 2 and 3. The tables
list the costs for the acidic sulfite and
ferrous iron methods of Cr(VI) reduction.
The relative costs of removing the Cr(VI)
from the two different size systems are
shown graphically in Figure 6. Cr(VI)
reduction using FeSO* is more econom-
ical than that using Na2S03, principally
because the sulfite reduction method
requires larger quantities of acids and
bases (relatively expensive) for pH
adjustmentthan the ferrous iron method.
In fact, the major expense for sulfite
treatment is for pH adjustment, while the
major cost for Fe(ll) treatment is due to
the amout of ferrous iron required to
reduce the Cr(VI) to Cr(lll). FeS04 is more
cost effective in dollars spent per kg Cr(VI)
removed than NaaSOa for both the 0.1
MGD and the 3.89 MGD well systems.
Another consideration in addition to
the cost of Cr(VI) reduction is the ease
of handling the chromium sludge pro-
duced. Using FeSO4 as the reductant also
has the added advantage of producing
a heavier, quicker settling Cr(OH)3
-------
<0
<5
to
.c
a
26.0
24.0
22.0
20.0
75.0
76.0
74.0
72.0
70.0
8.0
6.0
4.0
2.0
18.6)
Added FeSO<-7H2O and
Collected Supernatant
Added FeS04-7H2O, Increased
pH to 8.3 at 5 min with
INNaOH, and Collected
Supernatant
Figure 5.
Table 2.
Item
Solution pH Before Adding FeSOt-7H2O
Effect of initial pH (pH 5. 6. 7. 8. 9. 10) on reduction of CrfVI) FeSOt. 1.25X
stoichiometric amount FeSOt. All solutions contained an initial 100 mg/L Cr(VI)
concentration. 1 M NaCI. 75 meq/L HCOi- All supernatant samples were filtered
through a 0.45 fjm membrane filter.
Chemical and Disposal Costs for Treating 61.000 L Spent Regenerant Solution*
(Large Community Case)
Reduction by
Na2SO3
Reduction by
FeSOt
Wt. Filtered Sludge
20% Solid, kg
Chemical Cost. $
Cost to Dispose of
Sludge
Chemical + Disposal Costs
Total. $
1.000 gal product WzO
81
121.91
0.51
363
56.39
2.30
122.42
0.003
58.69
0.0015
^Assumptions include: 3.89 MOD product water; 70% bypass flow; 0.05 mg/L CrfVI) in feed
and 0.035 mg/L CrfVI) in blended water; 2 min EBCT; 25.000 BV run length (34.7 days);
10 BV of spent regenerant solution containing 100 mg/L Cr(VI). Capital and labor costs are
not included and sludge disposal costs do not include transportation.
hexavalent chromium (Cr(VI)) to trivalent
chromium (Cr(lll)) providing the dosage,
pH, and reaction time are properly
controlled. The optimum dosage of both
sodium sulfite and ferrous sulfate was
1.25 times the theoretical stoichiometric
amount. Hydrazine dosage was not
optimized. Rather, it was used at a 3-
times-stoichiometric concentration 1
reduce the time required for Cr(V
reduction. For hydrazine and sodiui
sulfite, the optimum solution pH for th
chromium reduction reaction was les
than 2.0. Ferrous sulfate gave goc
reduction of CrfVI) to Cr(lll) if the solutio
pH was between 5 and 8. The CrfVI) wa
reduced to Cr(lll) in less than 15 mi
when using sodium sulfite or ferrou
sulfate, and in less than 1 hr usin
hydrazine, when all reduction reaction
occurred at optimum conditions.
The presence of NaCI in the regenerar
solution gave better Cr(OH)3(s) flo
formation and decreased floe settlin
times following pH adjustment to 8.3 fc
all three reductants. Even though th
NaCI concentration varied among th
regenerant solutions, there was alway
good floe formation when the solutio
contained from 0.25 to 2.0 M NaCI (1.5%
12% NaCI). The voluminous amount c
Fe(OH)s(s) floe present when ferrou
sulfate was used as the reductant als
increased the Cr(OH)a(s) floe size ani
decreased the settling time.
Using NaOH or Ca(OH)2 for adjustmen
of the solution pH to 8.3 gave equall
good Cr(OH)3
-------
Table 3.
Chemical and Disposal Costs for Treating 1,500 L Spent ftegenerant Solution*
(Small Community Case)
Item
Reduction by
Na2S03
Reduction by
FeSO*
Wt. Filtered Sludge
20% Solid, kg
Chemical Cost, $
Cost to Dispose of
Sludge
Chemical + Disposal Costs
Total, $
2.10
5.28
0.01
9.40
4.80
0.06
5.29
4.86
1,000 gal produc
0.0050
0.0046
"Assumptions include: 0.1 MOD product water; 70% bypass flow; 0.05 mg/L CrfVI) in feed
and 0.035 mg/L CrfVI) in blended water; 2 min EBCT; 25,000 BV run length (34.7 days);
10 BV of spent regenerant solution containing 100 mg/L CrfVI). Capital and labor costs are
not included and sludge disposal costs do not include transportation.
days to chromate breakthrough. The cost
advantage for ferrous iron reduction is
mainly due to lower chemical costs.
Sodium sulfite is a less expensive
reductant than ferrous sulfate but the
total chemical costs for sulfite reduction
include large amounts of acid and base
for pH control. In both cases, the
estimated cost of the non-toxic sludge
disposal was less than 4% of the total
treatment and disposal costs. Thus, the
four times greater amount of sludge
produced using FeS04 does not
significantly impact the cost of treatment
plus disposal.
Recommendations
For a large Cr(VI)-contaminated well
like Scottsdale Well #32, the following
simple cost-effective procedure for
treatment and disposal is recommended
for the spent ion-exchange regenerant
using a batch reactor-settler.
1. Add approximately 1.3 times the
SA of FeSO« and adjust pH to 5-
7 using acid or base if necessary.
2. Mix rapidly for 30-60 sec and then
mix slowly (flocculate) for 20-30
min.
3. Settle for 2-24 hr.
4. Filter the supernatant.
5. Filter or centrifuge the sludge.
6. Dispose of the filtrate into the
sanitary sewer or an evaporation
pond.
7. Dispose of the non-toxic filter cake
(Cr(OH)33(s)) in a landfill.
Before implementation on a municipal
scale, the entire chromate removal
process, including chloride-form anion
exchange followed by spent regenerant
treatment using ferrous sulfate, should
be demonstrated over a period of at least
1 yr on a pilot scale. During this time,
tests should be performed on reuse of
the spent regenerant following Cr(VI)
reduction and removal.
The full report was submitted in
fulfillment of Cooperative Agreement No.
807939 by The University of Houston
under the sponsorship of the U.S.
Environmental Protection Agency.
Dennis Clifford and Susan Siegel are with the University of Houston. Houston,
TX 77004.
Thomas Sorg is the EPA Project Officer (see below).
The complete report, entitled "Removal of Chromium from Ion Exchange
Regenerant Solution," (Order No. PB 88-158 084/AS; Cost: $14.95, 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 Officer can be contacted at:
Water Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
•fcU.S.Government Printing Office: 1988 — 548-158/67095
7
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PAIC
EPA
PERMIT No G-35
Official Business
Penalty for Private Use $300
EPA/600/S2-88/007
OOQC329 PS
JWiSSTifKHP10" ""*
230 S OfABBflRK STffffT
CWICAfiO It 60404
------- |