United States
Environmental Protection
Agency
Water Engineering
Research Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/S2-87/100 Jan. 1988
&EPA Project Summary
The Fate of Chromium (III) in
Chlorinated Water
Dennis Clifford and Jimmy Man Chau
Although hexavalent chromium,
Cr(Vi), is not normally found in surface
water, the oxidation of trivalent chro-
mium, Cr(lll), to the more toxic Cr(VI)
in chlorinated water is thermodynam-
ically feasible and was the subject of
this study. It was found that free
available chlorine (FAC) readily con-
verts Cr(lll) to Cr(VI) at a rate that is
highly dependent upon pH, total
organic carbon (TOC), and chloride
concentrations while combined chlo-
rine (CAC) does not oxidize Cr(lll).
In deionized water and in a back-
ground of 0.010 M NaCI (355 mg/L
chloride), the highest oxidation rate
occurs in the 5.5-6.0 pH range. Results
with natural waters indicate that a
similarly fast oxidation rate can occur
in this same pH range. The aquatic
humus in natural water, however,
inhibits while chloride concentration
catalyzes the rate of Cr(VI) formation.
As expected, the initial Cr(lll) oxidation
rate' increases with increasing FAC
concentration and Cr(lll) level.
Monochloramine, a form of CAC, did
not oxidize Cr(lll) at any tested pH
between 6 and 8.5.
Finally, the results of this study
suggest that the oxidation of Cr(lll) to
Cr(VI) would rarely occur to a signif-
icant extent during drinking water
chlorination because of the presence of
naturally occurring organics (TOC)
present, the low concentrations of
Cr(lll) in natural waters, the probable
removal of insoluble Cr(OH)3 during
coagulation, and the increasing trend
to use combined chlorine for disinfec-
tion. However, in Cr(lll)-contaminated
waters that are relatively free of organic
contamination and have pH's in the 5-7
range, FAC readily converts Cr(lll) to
the more toxic hexavalent variety.
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
Chromium typically occurs in surface
waters in the trivalent (Cr(lll)) oxidation
state. The trivalent chromium cation is
always complexed with H2O, hydroxide
ion, or other common ligands including
chloride and fulvate anions. The indus-
trial sources of chromium are chrome
plating, leather tanning, and corrosion
inhibition. Additionally, many industries
use chromates as a cooling-water cor-
rosion inhibitor. Due to wastewater
discharges into receiving waters used for
drinking water supply, both Cr(VI) and
Cr(lll) can enter drinking water treatment
systems through raw water intakes.
However, due to the insolubility of Cr(lll)
and the reactivity of Cr(VI), serious
industrial contamination of drinking
water by chromium is quite rare.
Previous studies on the transformation
between these oxidation states have
been conducted. I n one such study it was
found the Cr(lll) was very slowly oxidized
to Cr(VI) by dissolved oxygen in Lake
Mendota (Wisconsin) water—2 //g/L
Cr(VI) was produced per week in filtered
lake water spiked with 125 fjg/l Cr(lll).
Oxidation by chlorination, the most
popular disinfection alternative, how-
ever, had been ignored as of the initiation
of this study (June 1983). The overall
-------
objective of the present study was to
determine to what extent Cr(lll) is
converted to Cr(VI) in chlorinated water
at the practical dosages of chlorine used
for disinfection, and in the presence of
aquatic humus, which is known to
complex Cr(lll) and react with chlorine.
The potential for Cr(lll) oxidation by
monochloramine was also studied
because of a trend to use chloramines
for drinking water disinfection to avoid
formation of trihalomethanes.
Initial experiments dealing with the
solubility of 200 iig/L total Cr(lll) in the
pH range of 5-10 indicated that, as
expected, the solubility was highly pH-
dependent. Furthermore, the longer the
Cr(lll)-spiked solutions aged, the lower
the Cr(lll) solubility. These observations,
added to the fact that water quality
criteria and effluent standards are
generally based on total (soluble and
particulate) chromium, led us to work
with total as opposed to soluble Cr(lll)
concentrations. In the usual test, the
background waters were spiked with 200
fjg/L Cr(lll), pH adjusted, and then
allowed to age for 1 or 4 days before
chlorine was added to oxidize Cr(lll) to
Cr(VI). Depending on pH and aging time,
approximately 5%-90% of the total Cr(lll)
present was insoluble as defined by
retention on a 0.45 Aim membrane filter.
The results of this study must be inter-
preted in light of these solubility
observations.
After acquiring data on the Cr(lll)
oxidation reaction at various pH levels
and in the presence of various concen-
trations of chloride, ammonia, and TOC,
the oxidation reaction was modeled. This
was done to facilitate the prediction of
the Cr(lll) oxidation rate in typical drinking
water sources contaminated with Cr(lll).
Analytical Methods
Chromium Analyses
Total Cr was analyzed using graphite
furnace atomic absorption (GFAA) spec-
troscopy as specified in U.S. EPA's
Methods for Chemical Analysis of Water
and Wastes (method 218.2). Prior to
analysis all samples were acidified with
4 mL HN03/L of sample (pH = 1.2). The
minimum detectable concentration was
approximately 1 .Qfjg/L and the precision
was ± 3% relative standard deviation
(rsd) at 100 Aig/L Cr. For purposes of this
study, soluble chromium is the Cr in the
filtrate of a previously-boiled and rinsed
0 45 A/m membrane filter.
Cr(VI) was determined by measuring
the absorbance of the red complex that
developed between diphenylcarbazide
and chromate (Cr(VI)) at pH 1.2. Absor-
bance was measured at 540 nm with a
Hitachi Perkin-Elmer spectrophotome-
ter* using a very long (10.0-cm) absorp-
tion cell. Using the method of standard
additions to avoid absorbance errors in
the presence of chlorine, the detection
limit was found to be 0.001 mg/L and
the precision was 3% rsd at 100 fjg/L
Cr(VI).
Chlorine Analyses
Free and combined chlorine were
analyzed using the DPD (N, N-diethyl-p-
phenylene-diamine) colorimetric method
and the 100-mL-sample-size DPD-FAS
(ferrous ammonium sulfate) titration
method, both of which are described in
Standard Methods^ 5th edition, sections
408E and 408D, respectively). The
methods were used concurrently to verify
one another.
TOC Analyses
Trace levels of TOC in the deionized
water produced by two Continental
Water Systems mixed-bed deionizers in
series were measured using a Dohrman
Model DC-54 Ultra-low Level TOC Ana-
lyzer utilizing persulfate-catalyzed, UV
oxidation. The high levels of TOC (> 1.0
mg/L) found in Houston tap and raw
waters were determined using a Beck-
man Model 915 B high-temperature
(950°C) TOC analyzer. The results indi-
cated 0.2 mg/L TOC in the deionized
water, 3.8 mg/L in the tap water, and
15.0 mg/L in the raw water before
treatment.
Batch Experiments for Cr(lll)
Oxidation
A 20-L, stirred batch reactor was
employed for the Cr(lll) solubility and
oxidation experiments. In a typical
experiment, an appropriate volume of
CrCb solution (200 mg/L Cr(lll)) was
pipetted into 20 L of continuously mixed
background water in a rectangular
polyethylene carboy (see Figure 1).
Within 1-2 min, the pH was adjusted to
the predetermined experimental value
using 1.0 N H2S04 or 1.0 N NaOH
solution. Control reactors with no Cr(lll)
or no chlorine were used as appropriate.
*Mention of trade names or commercial products
does not constitute endorsement or recommenda-
tion for use
Prior to the addition of chlorine, t
Cr(lll) added was allowed to hydrolyze
produce the species expected to
present in aged natural waters. To che
the extent of hydrolysis and Cr(OH)a
precipitation, solubility tests in deioniz
water and 0.01 M NaCI solution we
performed using the same 20-L reactc
by frequent analysis for both toi
(unfiltered) and soluble (filtered) Cr.
Hexavalent chromium, chlorine, a
pH were frequently monitored during t
course of each chlorine oxidation bat
experiment. For all batch experimen
the aging time was defined as the tir
between Cr(lll) addition and chlori
addition to the background water. Af
aging, an oxidation batch test w
initiated by pipetting an appropric
volume of chlorine solution (3,000 m
L CI2) into both the control and the t<
batches. Then the pH of both batches w
quickly readjusted to the predetermin
initial pH by dropwise addition of 1.0
H2SO4 or 1.0 N NaOH solution.
Effect of pH on Cr(lll)
Solubility
The effect of pH on Cr(lll) solubility
deionized water and 0.01 M NaCI v\
investigated. Figure 2 shows that, at 1 £
200 fjg/L Cr(lll), the solubility of Cr|
in 0.01 M NaCI rapidly decreases w
time at pH's between 6.5 and 9.0. T
maximum precipitation rate occurred
the 7-8 pH range. The main Cr(lll) spec
in that pH range is believed to
insoluble Cr(OH3(s). When pH becorr
greater than 10, Cr(OH>3(s) redissolves
further complexation with OH" to yii
the anionic species Cr(OH)i. At aci<
pH's, the cation complexes Cr3+, Cr(OH
Cra(OH)4+, and Cr(OH)J are predomina
For all hydroxylated species, the octal
dral coordination sites not occupied
Propeller Mixer
Sampling Pipet
(25 or 50 mL)
Sampling ^
Spigot
20-L (5-ga
Polyethyle,
Carboy
Batch Reactor
Figure 1. Batch reactor for Crflll) chlorin.
experiments.
-------
200
760
1
120
SO -
40 .
Total Chromium
\ \ 180 min. j .
Soluble
Chromium
I (Dashed Lines)
10
11
Figure 2.
PH
Solubility of Cr(lll) in the control reactor containing 0.01 M NaCI as a function
of pH and time, no CrflVJ present
OH are presumed to be occupied by H20
or Cl~ depending on the Cl~ con-
centration.
A comparison of the solubility data in
deionized water (Figure 3) and 0.01 M
NaCI (Figure 2) suggests that the chloride
anion has no significant effect on the
Cr(OH)3 precipitation rate in the 7-10 pH
range. The presence of chloride does,
however, appear to reduce the Cr(OH)3(s)
precipitation rate in the 5-6 pH range
compared to deionized water. No appre-
ciable precipitation of Cr(lll) m 0.01 M
NaCI was observed between pH 5 and
6. This improved Cr(lll) solubility in the
presence of CI" at reduced (5-6) pH may
be explained by the fact that chloride ion
can substitute for the hydroxide ligand
to form a soluble complex chloro-
chromium(lll) molecule. This is also the
pH range in which the maximum oxida-
tion rate occurred using free chlorine.
Cr(lll) Oxidation in 0.01 M
NaCI and Deionized Water
It can be seen in the batch test data
for 0.01 M NaCI shown in Figure 4 that
50 /ug/L Cr(VI) was exceeded in approx-
imately 60 min contact time at pH 6. Also,
the percent Cr(lll) conversion for 1 day
was slightly increased in 0.01 M NaCI
(Figure 4) compared to deionized water
(not shown). Less oxidation of Cr(lll)
occurred in deionize;' water under
similar conditions. Compared to the
oxidation kinetics in deionized water, the
presence of chloride enhances the Cr(VI)
formation rate for the same 3.0 mg/L
chlorination starting with 180-200 jug/
L total Cr(lll).
The experimental results lead us to
suggest the following reactions to
explain the observed effects of pH and
the chloride ion on the reaction rate. In
acidic solution, e.g., pH < 5.0, the
presence of OH" can be ignored and the
oxidation of hydrated Cr(lll) by HOCI is
relatively fast; it can be written as
Reaction 1 .
3HOCI + 2[Cr(H20)6]3+ - 3CI~
(D
It is evident from the proposed reaction
that CI" and H* ions should inhibit the
rate since they are products. But the
observed effects of both these ions in the
5-8 pH range is just the opposite, i.e.,
increasing concentrations of Cf and H*
speed up the rate. Presumably, both
these ions hinder the formation of
Cr(OH)3(s) and Cr(lll)-hydroxide com-
plexes, which are quite stable with
respect to chlorine oxidation. The exper-
imental results indicate that Reaction 2,
occurring in the 8-10 pH range, is very
slow.
30CI" + 2Cr(OH)3 + 40H~
3CT +
(2)
If Reactions 1 and 2 are considered alone,
chloride ion should theoretically inhibit
the Cr(VI) formation rate, and ye. it does
not do so even at very high (0.6 M)
chloride concentrations. At low pH, the
expected pH effect on Reaction 1 did
occur in all the tests, i.e., lowering the
pH to below 5.0 in deionized water and
in 0.01 M NaCI did eliminate or substan-
tially reduce the Cr(lll) oxidation rate.
Cr(lll) Oxidation in Houston
Tap Water (3.8 mg/L TOC)
The lack of Cr(lll) oxidation in pre-
viously filtered Houston tap water aged
1 day after Cr(lll) spiking is represented
in Figure 5 as a function of pH. Compared
to the rates in deionized water and 0.01
-------
200
160
\
s
5
a
o
V)
'20
40
Total Chromium
Soluble
c/,™/n/um
(Dashed Lines!
figure 3.
5 6 7 8 9 10 11
Solubility of Cr(lll] in the deionized water control reactor as a function of time
and pH, no CrfVII present.
M NaCI, also shown in Figure 5, the Cr(VI)
formation rate in tap water at 3.0 mg/
L chlorination was insignificant through-
out the 5-8 pH range. Undoubtedly, the
elimination of the oxidation of Cr(lll) by
free chlorine resulted, in part, from the
lowering of the FAC concentration by
rapid substitution and addition reactions
between chlorine and humic material
(3.8 mg/L TOC). These reactions
between chlorine and humic material
occurred because an analysis of aged
Houston tap water showed insignificant
concentrations of other chlorine-reactive
species such as ammonia.
An additional factor, Cr(lll) complexa-
tion, is responsible for the much slower
oxidation of Cr(lll) in TOC-contaminated
tap water. For example, in the pH 6
experiments with 12 vs. 3 mg/L chlori-
nation, the average FAC concentration
in the tap water during the first 2 hr
following 12 mg/L chlorination was
roughly 6 mg/L, i.e., double the 3 mg/
L FAC in the deionized water tests. Yet
the Cr(lll) oxidation rates were essentially
identical. Here the decrease in thf
expected oxidation rate of Cr(lll) exposet
to 6 mg/L FAC (avg) is attributed t(
formation of stable Cr(lll) complexes wit!
the humic material in the tap water. Th(
existence of these chromium-organii
complexes in fresh and salt water is wel
documented and is the subject of con
tinuing investigations by othe
researchers.
When examining the results in Figun
5, keep in mind the fact that in deionize(
water and in 0.01 M NaCI there was <
variable amount of soluble Cr(lll) presen
due to formation of Cr(OH)3(s) as ;
function of pH and aging time (set
Figures 2 and 3). With Houston tap wate
an additional factor, Cr(lll) complexatioi
by organic matter is responsible for <
further decrease in the amount of Cr(lll
available for oxidation by chlorine.
Cr(lll) Oxidation in Filtered
Raw Water (10 mg/L TOC)
The results of Cr(lll) oxidation ir
Houston raw water prior to treatment ty
coagulation, flocculation, and filtration a
the Houston Water Treatment Plant are
shown in Figure 6, which includes the
results of three batch tests at pH C
following 3, 6, or 12 mg/L chlorination
The original filtered raw water sample.1
were diluted with deionized water t(
prepare the 10 mg/L TOC raw water. I
is clear from Figure 6 that the presence
of a high concentration of humic materia
influenced the Cr(lll) oxidation by chlo
rine in the raw water containing 10 mg/
L TOC. No Cr(VI) was found at the en<
of 1 day following initial dosages of 3.(
or 6.0 mg/L chlorine. Substantial Cr(VI
production did, however, occur with 1 '<
mg/L chlorination. Unlike tap wate
containing only 3.8 mg/L TOC, the FAC
concentration resulting from 12 mg/l
chlorination decayed to an undetectable
concentration after 1 day. The concen
tration of Cr(VI) in this high TOC, raw
water was substantially lower ever
though there were nearly equal residua
free chlorine concentrations (5 mg/L
maintained during the first 4 hr o
contact. This again suggests that Cr(lll
is strongly complexed by the organii
matter in the raw water This is not t<
suggest that Cr(VI) production is always
inhibited at 3-6 mg/L chlorination it
natural waters containing TOC. Cr(VI
production is possible at lower TO(
levels, e.g., 1 -2 mg/L, and in fresh Cr(lll
solutions in which the Cr(lll) has not ha<
time to hydrolyze and form Cr(OH)3(s]
4
-------
200
160
I
6 720
2
5
80
40
( ) Soluble Chromium (III)
120
1440
Time, min
Figure 4. Cr(lll) oxidation in 0.01 M NaCI at pH 6, 4 day Cr(llll aging before 3 mg/L
chlorination.
Oxidation of Cr(\\\) under such conditions
was observed in a separate EPA study
by Ulmer.
Attempted Monochloramine
Oxidation of Cr(lll) in 0.01 M
NaCI
Ammonia-N (NH3-N) was added, using
NH4CI, to give twice the NHa-N concen-
tration needed to produce monochlora-
mme. The experimental data showed that
monochloramine, a combined form of
chlorine which is produced almost
instantaneously from the reaction of
ammonia with chlorine, cannot oxidize
Cr(lll) under the experimental conditions.
No Cr(VI) was found even after 24 hr of
oxidation at pH levels 6, 7, and 8.5 in
0.01 M NaCI spiked with 200 fjg/L total
Cr(lll).
The Effect of Chloride Ion on
the Cr(lll) Oxidation Rate
It is evident from Figure 7 that the
previously mentioned enhancement of
the Cr(lll) oxidation rate due to chloride
continues far beyond 0.01 M (3550 mg/
L chloride). It is still evident at a chloride
concentration of 0.6 M (21,200 mg/L
chloride), typical of sea water. Further-
more, the enhancement is seen in very
dilute chloride solutions, e.g., 0.001 M
(35.5 mg/L CI"). Ligands other than
chloride, e.g., sulfate, may produce
similar rate-enhancing results.
Reaction Rate Modeling
As expected from the simplest and
most complicated models of Cr(lll) chlo-
rination, the rate increases with increas-
ing Cr(lll) and free chlorine concentra-
tions. Using a simple rate model with no
back reaction, the Cr(lll) oxidation rate
was third order in Cr(lll) concentration
at pH 6.0 in 0.01 M NaCI, but only 0.8
order in Cr(lll) at pH 4.5. Using the same
model, the Cr(lll) oxidation rate was found
to be 0.56 order in HOCI concentration
at pH 5.5 in 0.01 M NaCI. Such simple
models quantify the relative sensitivity
of the rate to Cr(lll) and HOCI concen-
tration but should not be taken as truly
representative of the true chlorination
reaction involving various complexes of
Cr(lll) with CI", OH", and H20.
Conclusions
The fundamentals of the Cr(lll)-
chlorine oxidation reaction in deionized
water, dilute NaCI, and natural water
have been elucidated by this experimen-
tal study utilizing aged Cr(lll) solution
containing soluble, insoluble and com-
plexed Cr(lll). Cr(lll) can be oxidized to
Cr(VI) by free chlorine at a rate that is
highly dependent on pH, chlorine dosage,
chloride concentration, and TOC concen-
tration. Combined chlorine in the form
of monochloramine cannot oxidize Cr(lll)
to Cr(VI), even under the most favorable
conditions of pH and chloride
concentration.
The optimum pH range for Cr(lll)
oxidation by free chlorine is 5.5-6.0. It
is theorized that above this pH range, the
increasing presence of OH" ligands
causes the formation of insoluble
Cr(OH)3
-------
o
4
I
60
50
40
30
20
10
0 \-
Deionized Water
O.OIMNaCL
Tap Water
pH
Figure 5. CrfVI) production during 1 hr following 3 mg/L chlorination as a function of
pH in deioniied water, 0 01 M NaCI and Houston tap water containing
approximately 200 ttg/L total Crflll). All Crflll)-spiked solutions were aged 1-3
days prior to chlorination. No measurable oxidation occurred in tap water
containing 2.8 mg/L TOO.
L chloride) under optimum conditions.
The rate-enhancing influence of the
chloride ion on the chlorine oxidation of
Cr(lll) continued throughout the 0.001-
0.6 M NaCI range. At 3 mg/L chlorine
and a pH nearthe optimum, the oxidation
rate was found to be 0.35 order in
chloride concentration. Apparently the
chloro-Cr(lll) complexes that form in the
presence of excess chloride ions are
more susceptible to oxidation by free
chlorine than are the aquo- and hydroxyl-
complexes.
The rapid Cr(lll) oxidation rate observed
in deionized water was not reproduced
in raw or treated natural waters contain-
ing TOC due to the presence of aquatic
humus. The fulvate and humate anions
appear to complex the Cr(lll) strongly and
protect it from chlorine oxidation Addi-
tionally, the TOC is oxidized by free
chlorine, thereby reducing the free
available chlorine concentration. No
observable CrfVI) was produced from 200
j/g/L total Cr(lll) at pH 7 5 and a 3 mg/
L chlorine dose. If enough chlorine is
added, however, Cr(lll) can be oxidized
to Cr(VI) in spite of the TOC present.
Furthermore, if the Cr(lll) solutions are
chlorinated immediately after Cr(lll)
addition, higher levels of CrfVI) are
expected due to the higher concentra-
tions of non-hydrolyzed Crflll).
Starting with a relatively high (200 /ug/
L) concentration of total Crflll), in a
solution aged for 1 or more days, it is
unlikely that a significant concentration,
e.g., 50 /jg/L, of CrfVI) would be produced
as a result of typical (e.g., 1 -3 mg/L) free
chlorine dosages to natural surface
waters containing 2-10 mg/L TOC.
Experimentally, no Cr(VI) was produced
even after 24 hr following application of
3 mg/L free chlorine to Houston tap
water aged 1 day following Crflll) spiking.
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.
-------
40
30 -
I
20-
10 .
0 4-
• 3 rng/L Dose
j^| 6 mg/L Dose
r~\ 12 mg/L Dose
24
Time, hr
Figure 6.
Cr(lll) oxidation in raw water containing 10 mg/L TOO as a function of time following
3.6 and 12 mg/L chlorination. pH = 6 and Cr(lll) initial = 200 /ug/L. No measurable
Crflll) oxidation occurred at 3 or 6 mg/L chlorination. All Cr(lll)-spiked solutions
were aged 1 day prior to chlorination.
400
300 -
-J
I
IT
•o
200 -
700
70,000 75,000 20,000 25,000
Chloride Cone , mg/L
Figure 7. Initial Crfl/l) oxidation rate as a function of chloride concentration at pH 6.0 and
3 0 mg/L chlorination.
-------
Dennis Clifford and Jimmy Man Chau are with the University of Houston,
Houston, TX 77004.
Thomas Sorg is the EPA Project Officer (see below).
The complete report, entitled "The Fate of Chromium(lll) in Chlorinated Water,"
(Order No. PB 88-130 992/AS; Cost: $14.95) 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:
Water Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
\
/J3tS30C| ; ft « 'I
« * i „ u . j i.
Official Business
Penalty for Private Use $300
EPA/600/S2-87/100
0000329 PS
U S eUVIJ? PROTECTION ASfNCY
REGION 5 LIBRART
230 S DEARBORN STREET
CHICAGO IL 60604
------- |