Vous êtes sur la page 1sur 10

TECHNICAL REPORTS: HEAVY METALS IN THE ENVIRONMENT

Leaching Mechanisms of Cr(VI) from Chromite Ore Processing Residue


Mahmoud Wazne,* Santhi Chandra Jagupilla, Deok Hyun Moon, Christos Christodoulatos, and Agamemnon Koutsospyros
Stevens Institute of Technology
Batch leaching tests, qualitative and quantitative x-ray
powder diraction (XRPD) analyses, and geochemical
modeling were used to investigate the leaching mechanisms of
Cr(VI) from chromite ore processing residue (COPR) samples
obtained from an urban area in Hudson County, New Jersey.
The pH of the leaching solutions was adjusted to cover a wide
range between 1 and 12.5. The concentration levels for total
chromium (Cr) and Cr(VI) in the leaching solutions were
virtually identical for pH values >5. For pH values <5, the
concentration of total Cr exceeded that of Cr(VI) with the
dierence between the two attributed to Cr(III). Geochemical
modeling results indicated that the solubility of Cr(VI) is
controlled by Cr(VI)-hydrocalumite and Cr(VI)-ettringite at
pH >10.5 and by adsorption at pH <8. However, experimental
results suggested that Cr(VI) solubility is controlled partially
by Cr(VI)-hydrocalumite at pH >10.5 and by hydrotalcites at
pH >8 in addition to adsorption of anionic chromate species
onto inherently present metal oxides and hydroxides at pH <8.
As pH decreased to <10, most of the Cr(VI) bearing minerals
become unstable and their dissolution contributes to the
increase in Cr(VI) concentration in the leachate solution. At
low pH ( <1.5), Cr(III) solid phases and the oxides responsible
for Cr(VI) adsorption dissolve and release Cr(III) and Cr(VI)
into solution.

Copyright 2008 by the American Society of Agronomy, Crop Science


Society of America, and Soil Science Society of America. All rights
reserved. No part of this periodical may be reproduced or transmitted
in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system,
without permission in writing from the publisher.
Published in J. Environ. Qual. 37:21252134 (2008).
doi:10.2134/jeq2007.0443
Received 22 Aug. 2007.
*Corresponding author (mwazne@stevens.edu).
ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

illions of tonnes of COPR were deposited at numerous


urban areas in the United States, the United Kingdom
(UK), and elsewhere in the world during the rst half of the
last century (Burke et al., 1991; Weng et al., 1994; James,
1996; Farmer et al., 1999; Graham et al., 2006). The COPR
was benecially used as structural ll because of its favorable
structural quality as a granulated material. The deposited COPR
was produced by the high lime process where the chromite ore
was roasted at approximately 1200C to oxidize the chromium
in the ore from the trivalent to the hexavalent state. Hexavalent
chromium [Cr(VI)] was then chemically combined with the
sodium in the added soda ash to form Na2CrO4 (Allied Signal,
1982). Lime was added during the roasting process to act as a
mechanical separator allowing oxygen to react with the chromite
and sodium carbonate. Lime also served as a sequestering
agent combining with various ore impurities to form insoluble
compounds. The sodium chromate formed during the roasting
process was extracted with hot water as a weak yellow liquor
solution. The sodium chromate was then converted into sodium
dichromate by reaction with sulfuric acid. After draining, the
residue was discarded. The disposed COPR contains unreacted
chromite ore and unextracted chromate. Even though the high
lime process has ceased in the United States and the UK, it is still
being used in China, Russia, Kazakhstan, India, and Pakistan
(Darrie, 2001).
In the environment, chromium exists mainly in two oxidation states as hexavalent and trivalent chromium. The Cr(VI) is
highly mobile, severely toxic at moderate doses, and classied as
a respiratory carcinogen in humans. In contrast, Cr(III) is used
as a dietary element at low doses, and in most environmental
systems is immobile (Higgins et al., 1998). The COPR, which
contains both Cr(III) and Cr(VI), is not as benign as initially
thought; a yellow chromate solution is often observed to leach
from locations where COPR is deposited and elevated Cr(VI)
concentrations are measured in groundwater and water bodies
in the proximity of these sites (Farmer et al., 2002). In consequence, COPR has become a major contamination source of the
highly mobile and toxic Cr(VI) in many urban areas.
Leaching of Cr(VI) from soils and COPR has been studied fairly
extensively, however, COPR originating from various localities may
Center for Environmental Systems, Stevens Institute of Technology, Castle Point on
Hudson, Hoboken, NJ 07030.
Abbreviations: bgs, below ground surface; CAC, calcium aluminum oxide chromium
hydrates; COPR, chromite ore processing residue; DI, deionized; EDS, energy dispersive
x-ray spectroscopy; SEM, scanning electron microscopy; XRPD, x-ray powder diffraction

2125

dier in terms of leaching behavior and mineralogy due to variations in chromite ore composition, extraction process, and deposition practices. Some ores could be rich in silica, iron, aluminum,
or impurities which cause the resulting mineralogy to vary. For
example, the COPR at Glasgow has twice the amount of silicon as
the COPR at New Jersey. This may have very strong implication
on the mineralogy of COPR as it may favor the formation of cementitious silicious phases. Similarly, inecient extraction of sulfates during the sodium sulfate operation during processing or the
intrusion of sulfate due the oxidation of sulfur in certain kiln fuel
oil (Allied Signal) may change the resulting mineralogy of COPR.
The introduction of sulfates to COPR may cause the formation of
ettringite, a swell causing mineral in cement chemistry. Reports of
catastrophic heave and swell are abound at the Jersey site, however,
no heave was reported at Glasgow (Geelhoed et al., 2002; Dermatas et al., 2006a). Moreover, often times COPR is mixed with
indigenous soils at some deposition sites whereas at other sites it
is deposited as pure COPR. Even at pure COPR sites, the manner of deposition may have implications on COPR behavior. For
example, if COPR is deposited in homogenous layers, its behavior
may dier had it been deposited through conduits where the ner
material is expected to settle to the bottom of the heap and the
coarser material will collect on top similar to alluvial depositions at
deltas. The mineralogical composition of the ne and coarse materials is expected to be dierent (Dermatas et al., 2006b). Moreover,
due to its high alkalinity, COPR absorbs signicant amounts of
atmospheric CO2 where it combines with COPR constituents to
form various carbonate phases. The degree of atmospheric exposure
and manner of CO2 exposure may determine the amount of CO2
absorbed. For example, CO32 constitutes approximately 11.5% of
total mass of COPR at the Jersey site (total carbonate at Glasgow
is not reported). In addition, hydrotalcites were clearly identied
at the Jersey site but are not reported at Glasgow. This has implications on COPR mineralogy and leaching behavior, because hydrotalcites, which are magnesium aluminum carbonate hydrates, are
reported to sequester chromates through anionic exchange (Alvarez-Ayuso and Nugteren, 2005). Consequently, the leaching behavior and the resultant mineralogy will dier signicantly depending
on the parent ore, extraction process, and deposition practices.
Moreover, modeling techniques may dier in terms of scope
and methodology. For example, Yalin and nl (2006) reported
on a mechanistic model to simulate Cr(VI) leaching from COPR
obtained at an industrial plant in Turkey. The model incorporates
a sequence of batch (dissolution) and ushing (leaching) operations. It uses Runge Kutta method to solve the resultant coupled
dierential equations and determine the parameters of the leaching
reactor. The study does not address the geochemistry of COPR
leaching or mineralogy. Conversely, Geelhoed et al. (2002) reports
on the leaching of Cr(VI) in COPR originating from a Glasgow
site. The authors use leaching tests, XRPD, and equilibrium based
geochemical modeling to describe the leaching behavior of COPR.
The concentration of leached Ca, Al, Si, Mg, and Cr(VI) are used
to calibrate for the solubility constants of some solid phases that
were used in the model calculations. The mineral phases identied in the residues of the leaching tests and other phases that may
be present in COPR are used in their model. Even though the

2126

dissolution of COPR is kinetically controlled (dissolution time is


beyond experimental time frame) the authors used the ionic activity product instead of equilibrium solubility to account for the
discrepancy in the concentration of some elements between the
measured ones and the ones predicted by the model had they used
equilibrium constants as reported in the literature. For example,
they modied the equilibrium constant of Cr(VI)-ettringite to
make the model t their experimental data down to pH 8, even
though they did not identify Cr(VI)-ettringite in any sample
through XRPD or scanning electron microscopy (SEM)energy
dispersive x-ray spectroscopy (EDS). The discrepancy between the
measured and predicted concentrations could be due to the slow
dissolution rate of some COPR minerals such as brownmillerite
which may encapsulate Cr(VI) minerals as reported by Kremser
et al. (2005). Alternatively, Cr(VI) could be encapsulated in other
phases that were not identied in their study, such as hydrotalcites
which are thermodynamically stable down to pH 8.
In this study batch leaching tests complemented with
qualitative and quantitative XRPD analyses, SEM-EDS and
geochemical modeling are used to investigate the leaching
behavior of Cr(VI) in COPR originating from a site in Hudson County, New Jersey. The experimental and modeling
results are then explained in terms of potential mineral phases
that could be controlling Cr(VI) leaching and the kinetics of
dissolution of thermodynamically unstable minerals present
on site. The ndings of this study are relevant to assess the
potential of hexavalent chromium contamination emanating
from COPR sites in Hudson County, New Jersey. They are
also relevant for the development of remediation strategies
and chemical treatment of Cr(VI) in COPR.

Materials and Methods


Materials
The COPR samples, used in this investigation, were obtained from a study area in Jersey City, NJ. The samples were
collected from two stratigraphic layers (B1 and B2) during a
major bulk-sampling event from 14 borings across the site.
Layer B1 represents the upper most unsaturated zone and layer
B2 lies directly below layer B1. The water table level uctuates
within layer B2 throughout the entire site. A composite sample
B1B2 was prepared by mixing equal amounts of all samples
from layers B1 and B2 across the borings. The samples were
thoroughly homogenized before any COPR material was used.
All chemical reagents used were of ACS or higher-grade quality
and were obtained from Fisher Scientic (Suwanee, GA). All
stock solutions were prepared using deionized (DI) water.

Sample Characterization
The elemental composition of the COPR samples was determined by digesting the COPR material using USEPA Method
3051A (USEPA, 1996a) followed by USEPA Method 6010B
(USEPA, 1996b). Hexavalent chromium concentration was
obtained using USEPA Method 3060 (USEPA, 1996c) and
USEPA Method 7196A (USEPA, 1992). Total carbon was
measured using ASTM D5291 (ASTM, 2001) and it was as-

Journal of Environmental Quality Volume 37 NovemberDecember 2008

sumed to consist of inorganic carbonates. Silicon concentration


was determined by fusing the sample with Na2CO3 and precipitating Si as SiO2 which was then measured using the gravimetric method. All sulfur was assumed present as sulfate.

X-Ray Powder Diffraction Analyses


The XRPD was used for the qualitative characterization of the
particulate fraction of COPR. Samples were air dried for 24 h
and then were pulverized using a McCrone micromill for 5 min
with cyclohexane. The resulting slurry was air-dried and mixed
with 20% mass based corundum (Al2O3) as internal standard
and the resulting solids were subjected to XRPD analysis. Stepscanned x-ray diraction data were collected by the Rigaku DXR
3000 computer-automated diractometer using Bragg-Brentano
geometry. The diractometry was conducted at 40 kV and
40 mA using diracted beam graphite-monochromator with Cu
radiation. The data were collected in the range of 5 to 65 2
with a step size of 0.02 and a count time of 3 s per step. Qualitative analysis of XRPD patterns were conducted using the JADE
software version 7.1 and the International Centre for Diraction
database reference patterns (ICDD, 2002; ICSD, 2005; MDI,
2005). The JADE software was also used to quantify the identied crystalline mineral phases. JADE uses Whole Pattern Fitting which is based on the Rietveld method (Rietveld, 1969).

The pulverized COPR samples were added to a beaker with liquidto-solid ratio of 20. After a predetermined HCl acid dosage was
added to the samples, the beakers were capped with a plastic covers
and the contents were mixed with magnetic stirrers. Periodically
homogenous subsamples were withdrawn by a plastic syringe and
ltered using 0.45 m syringe lter. The ltrates were analyzed for
Cr(VI) and pH. Two sets of experiments were performed at acid
dosages of 5.5 and 9.2 [H+] eq/Kg COPR.

Synthetic Hydrotalcite Chromate Exchange Capacity


An experiment was conducted to assess the capacity of synthetic hydrotalcite to exchange chromate from COPR leachates.
Synthetic hydrotalcite [Mg3Al(OH)8]2CO3nH2O was prepared
by co-precipitation as described by Alvarez-Ayuso and Nugteren
(2005). Briey, a solution containing 0.75 mol Mg(OH)26H2O
and 0.25 mol Al(NO3)39H2O in 250 mL of deionized water
was added to a vigorously stirred solution (500 mL) containing
1.7 mol NaOH and 0.5 mol Na2CO3. The resulting precipitate,
washed thoroughly with deionized water, was dried at 80C overnight. The prepared precipitate was characterized as hydrotalcite
by x-ray diraction. 0.1 g of the prepared synthetic hydrotalcite
was added to 20 mL of COPR leachate in 50 mL vial. After mixing for 1 wk the pH values of the mixtures were recorded and the
leachate was analyzed for Cr(VI).

Scanning Electron Microscopy Analyses

Geochemical Modeling

The samples were rst air-dried and attached to a substrate


using double-side carbon tape. Scanning electron microscopy
(SEM) analyses were conducted using a LEO-810 Zeiss microscope equipped with an energy dispersive x-ray spectroscopy (EDS), ISIS-LINK system.

The geochemical equilibrium computer program MINTEQ


(David and Allison, 1999; USEPA, 2006) was used to model
the concentration of Cr(VI) in the leaching tests and to determine the predominant solid phases in COPR. The solid phase
thermodynamic database of MINTEQ was augmented with
the thermodynamic data of the relevant solid phases cited from
the literature based on the XRPD results (Table 1). Pure hydrotalcite, 4MgOAl2O310H2O, was used in the model due to the
lack of thermodynamic data for other analogs. The 2-pk diuse
double layer surface complexation model was used to describe
the adsorption of Cr(VI). In the model simulation, iron and
aluminum were assumed to provide the adsorption sites. The
adsorbents were assumed to consist of active and inactive sites;
only the active sites were considered in the adsorption model
simulation. The concentration of the active sites was determined
from the model by minimizing the sum of the square of the
dierences between the experimental and the model predicted
values of the concentration of Cr(VI) (Jing et al., 2006). The
acid-base surface reactions and the relevant adsorption reactions
were adopted from the MINTEQ thermodynamic database. The
Davies equation (Snoeyink and Jenkins, 1980) was used to calculate the activity coecients in the model. The model calculated
ionic strengths were used for I < 0.5 M. For I > 0.5 M, the ionic
strength was xed at 0.5 M. The calculated ionic strength was
equal to or less than approximately 0.5 M down to pH 9, and it
was <0.85 M at pH values >5 (the lowest pH values in the model
simulations). The values of Cr(VI) concentrations predicted by
the model simulation did not change appreciably (<10%) when
the high ionic strengths were xed at 0.5 M.

Leaching Tests
The objective of these experiments was to study the quantities
of soluble chromium species ultimately released in COPR leachates
at various pH values for a mixing time of 1 wk. Representative
air-dried samples COPR were pulverized to ner than 150 m
(mesh-100). The pulverized COPR samples were mixed with DI
water using a liquid-to-solid ratio of 20. Incremental amounts of
concentrated HCl were added to cover a wide range of pH values.
The mixtures were left on an end-over-end mixer for 1 wk before
the pH values of the mixtures were recorded. The leachates were
analyzed for total Cr and Cr(VI). The residues were submitted for
XRPD analyses. Upon sample acidication, part of the carbonate
species degassed as CO2. The samples were allowed to degas before
they were capped. Duplicate samples were prepared with the same
acid dosage. The duplicate samples were air-dried and the total
carbonate contents were determined using ASTM D5291 (ASTM,
2001). The measured total carbonate concentrations at various pH
values were used in the model calculations.
Another set of leaching tests was conducted to study the release
and exchange of soluble chromium species in COPR leachate for
shorter time scale (minutes). The time scale of these experiments
varied in the range of 0 to 80 h. Representative COPR samples
were air-dried then pulverized to ner than 150 m (mesh-100).

Wazne et al.: Leaching Mechanisms of Cr(VI) from Chromite Ore Processing Residue

2127

Table 1. R eactions and parameters used in the model calculations


(phases added to Minteqa2).
Solid phase
Afwillite
Brownmillerite C4AF
C2AH8
CAH10
C4AH13
C4AH19
Gehlenite Hydrate C2ASH8
Hydrogarnet C3AH6
Hydrotalcite M4AH10
Monosulfate
Wairakite
Surface reactions:
= SOH + H+ = S-OH2+
4SOH = SO + H+
= SOH + CrO4 2 = SOHCrO42
= SOH + CrO4 2 + H+ = SCrO4 +H2O
= SOH + CO3 2 + H+ = SCO3 +H2O
= SOH + CO3 2 + 2H+ = SCO3H +H2O
= SOH + H4SiO4 = SOSiO2OH2 + 2H+ + H2O
= SOH + H4SiO4 = SOSi(OH)2 + H+ + H2O
= SOH + H4SiO4 = SOSi(OH)3 + H2O

Leaching Tests

Chemical formula
Log K
Ca3Si2O4(OH)6
64.92a (1)
Ca4Al2Fe2O10
139.13a (2)
Ca2Al2(OH)10.3H2O
58.19b (3)
CaAl2(OH)8.6H2O
37.99b (4)
Ca4Al2(OH)14.6H2O
107.25b (5)
Ca4Al2(OH)14.12H2O
103.68b (6)
Ca2Al2(OH)6SiO8H8.H2O 49.16c (7)
Ca3Al2(H4O4)3
78.27d (8)
4MgOAl2O310H2O
73.78c (9)
Ca4Al2(OH)12SO4.6H2O 71.97e (10)
CaAl2SiO10.(OH)4
18.07b (11)
7.29f (12)
8.93f (13)
3.9f (14)
10.85f (15)
12.78f (16)
20.37f (17)
11.69f (18)
3.22f (19)
4.28f (20)

Parameters used in the model: Best fit active adsorbent sites = 10.95 mmol/L.
gSurface site (SOH) density = 11 sites/nm2. gSpecific surface area = 600 m2/g.
Source(s): a: Common Thermodynamic Database Project (2004); b: EQ3/6
(Lawrence Livermore National Laboratory [1994]); c: Bennet et al. (1992); d:
Reardon (1992); e: Phreeqc Database, Parkhurst and Appelo (1999); f: David and
Allison (1999); g: Dzombak and Morel (1990).

Results and Discussion


Sample Characterization
The COPR material originating from layer B1 is black to
gray, ne to coarse sand sized with trace silt, silty sand, and sandy silt particles. The B1 layer was encountered from the ground
surface to approximately 2 to 3 m (79 ft) below ground surface (bgs). The measured water content ranged from approximately 4 to 23% with an average value of approximately 18%.
Conversely, the COPR material in B2 layer is gray, ne to
coarse sand sized with seams of silt and sandy silt particles trace
gravel size particles. The B2 layer was encountered beginning
from approximately 2 to 3 m (79 ft) bgs to approximately 4 to
5 m (1215 ft) bgs with thickness ranging from approximately
1 to 2 m (3.56.5 ft). Moisture contents ranged from approximately 20 to 38% with an average value of approximately 28%
The elemental composition of the COPR sample is shown in
Table 2. The relatively high calcium concentration at approximately 24% is due to the addition of lime during the extraction
process. The carbonate most likely was absorbed as CO2 by the
COPR material after the roasting process due to its alkaline
nature. The sulfate may have entered the system from the oxidation of sulfur in the kiln fuel oil (Bunker C fuel oil) during the
roasting process, or from the sodium sulfate operations at the site
(Allied Signal, 1982). Presence of sulfate is known to cause expansion in cement and cement-like material such as COPR.

The results of the short time scale leaching tests are shown in
Fig. 1. For both acid dosages, immediately after acidication the
pH dropped to approximately 3 for the sample with the high
acid dosage and 5 for the sample with the lower acid dosage.
However, the pH values started to increase as the acid is neutralized by the alkalinity of the COPR matrix. It appears that the
pH values stabilized at approximately 8 and 9 for the higher and
lower acid dosages respectively, after approximately 20 h.
The measured Cr(VI) concentrations in the leachates
15 min after the addition of acid were approximately 125 mg/L
and 200 mg/L for the high and low acid dosages, respectively.
Cr(VI) concentrations for both samples reached maximum
values after approximately 2 h. The Cr(VI) concentration of
the COPR sample with the smaller acid dosage (pH 9), peaked
at approximately 240 mg/L and then it decreased to approximately 200 mg/L after 76 h. Conversely, the Cr(VI) concentration in the COPR sample with the higher acid dosage (pH 8)
peaked at approximately 270 mg/L and then it decreased to
approximately 250 mg/L. The moderate decrease in Cr(VI)
concentration of approximately 17% for the lower acid dosage
(pH 9) could be attributed to incorporation of chromate in
stable COPR minerals at that pH. Adsorption is not expected
to play a signicant role at this pH range (Zachara et al., 1987).
In addition, had this decrease been due to adsorption, higher
removal rates would have been expected at the high rather than
the low acid dosage since the rate of adsorption of anions is
known to be greater at lower pH values. This information is
signicant for any remediation scheme incorporating leaching
techniques to solubilize Cr(VI) since no solid phase is reported
to control the solubility of Cr(VI) at pH <10.5.
The results of the long-term leaching tests (1 wk) are presented in Fig. 2 where the concentration of total and hexavalent
chromium (left ordinate) and amount of acid in [H+] eq/Kg
COPR (right ordinate) are plotted vs. pH. Concentration levels
for total Cr and Cr(VI) in the leachate are virtually identical at
pH greater than approximately 5. This indicates that all leached
total Cr in this pH region is in the hexavalent state. At pH values <5, Cr concentration is greater than Cr(VI) concentration.
The dierence between total Cr and Cr(VI) concentrations at
pH values less than approximately 5 is attributed to Cr(III).
Figure 2 indicates that almost all of the chromium was released
when the pH was below 1.5. Approximately 32 equivalents of
acid [H+] per 1 kg of dry COPR were needed to attain pH 1.5.
In addition, the majority of Cr(VI), quantied by alkaline digestion, was leached at pH 8. For example, Cr(VI) concentration
in the leachate was approximately 260 mg/L at pH of 8, with
a liquid-to-solid ratio of 20, meaning that 5200 mg/Kg Cr(VI)
was leached from the COPR sample; constituting approximately
93% of the total leachable Cr(VI). However, the entire Cr concentration was leached at pH < 1.5. The measured Cr concentration was 1100 mg/L which is equivalent to 22,000 mg/Kg.

Table 2. Elemental composition of composite chromite ore processing residue sample B1B2.
Element

Al

Ca

CO32

Cr(VI)

Cr

Fe

Mg

Mn

Na

Si

SO42

Percent mass basis

4.6

23.9

11.5

0.56

2.1

11.8

0.03

6.1

0.12

0.37

1.98

0.34

2128

Journal of Environmental Quality Volume 37 NovemberDecember 2008

Fig. 1. Plot of Cr(VI) concentration and pH as a function of time due to the additions of 5.5 and 9.2 eq/Kg [H+].

X-ray Powder Diffraction Characterization


The XRPD mineralogical analyses indicate that brownmillerite, brucite, calcite, quartz, hydrotalcite, and katoite are the
major mineral phases in the COPR samples (Fig. 3). The content
of brownmillerite ranged between approximately 38 and 20.5%
of the solid residue for the ANC tests, which indicated that it
is relatively stable down to pH 5.24. The amorphous content
as determined by the corundum internal standard was approximately 42% pH 12.03 and it increased to approximately 64.5%
at pH 5.24. The only known Cr(VI) bearing mineral identied
was Calcium Aluminum Oxide Chromium Hydrates (CAC)

also known as Cr(VI)-hydrocalumite, with molecular formula


(Ca4Al2(OH)12CrO4nH2O). The Cr(VI)-hydrocalumite is a
layered double hydroxide (LDH) mineral with chromate anions
held in the interlayers. The Rietveld quantication indicated
that CAC is present at 0.87% in the residue of the ANC test at
pH 12.03. A Cr(VI)-hydrocalumite content of 0.87% indicate
a Cr(VI) concentration of approximately 667 mg/Kg, which is
approximately 13% of total Cr(VI). This indicates that the majority of Cr(VI) is not encapsulated in the Cr(VI)-hydrocalumite
phase. The Cr(VI) could be present in other mineral phases such
as hydrogarnet (katoite) as reported by Hillier et al. (2007) or hy-

Fig. 2. Concentration of Cr and Cr(VI) in the leachates vs. pH after 1 wk of mixing with liquid-to-solid ratio of 20 for composite sample B1B2.

Wazne et al.: Leaching Mechanisms of Cr(VI) from Chromite Ore Processing Residue

2129

Fig. 3. The x-ray diffraction patterns of the residues of the leaching tests at various pH values.

drotalcites through anionic substitution. Even though Cr(VI)ettringite is reported to control the solubility of chromate at
pH >10.4 (David and Allison, 1999; Jing et al., 2006; USEPA,
2006) it was not identied in any of the COPR samples. Conversely, hydrotalcite phases such as quintinite and sjogrenite were
identied in many COPR samples as evidenced by XRPD in Fig.
3 and Table 3, and also by SEM as shown in Fig. 4.
The SEM images indicated the presence of hydrotalcite
phases with extensive chloride substitution in the interlayers
probably due to the use of HCl in the ANC tests. However, the
chromium peak at approximately 5.4 keV could be due to the
substitution of CrO42, Cr3+ or both in the crystal structure.
Hydrotalicte has platy like hexagonal structure as shown clearly
in Fig. 4. The crystals were clearly identied at pH 10.9, but as pH
decreased it seems that the crystal size decreased and there appears
to be some disorder in addition to some coating. In general, hydrotalcite crystals were more dicult to identify at lower pH values.

Modeling
The diuse double layer surface complexation model (Dzombak and Morel, 1990; USEPA, 2006) is used to describe Cr(VI)
adsorption data. Carbonate adsorption is incorporated into the
model since carbonate is reported to adsorb onto iron hydroxides (Zachara et al., 1987; Van Geen et al., 1994; Wijnja and
Schulthess, 2001). Silicate has also been reported to adsorb onto
iron hydroxides (Meng et al., 2000) and therefore it is included
in the model. The total concentration of iron and aluminum in
the leaching solution are 98.21 and 87.03 mM, respectively.
The calculated best t for the active adsorbent concentration is 10.95 mM, which indicates that the molar ratio of the
active sites to the total iron and aluminum concentration is
<6% It is worth noting that the adsorption sites are not assumed to associate with any specic solid phase.
Brownmillerite was the major mineral phase observed in the
residues of the leaching tests. However, brownmillerite is not

2130

predicted by the model as it is not thermodynamically stable in


aqueous solution. Tamas and Vertes (1972) reported complete
hydration of synthetic brownmillerite in less than 2 wk. It appears that brownmillerite hydration is kinetically inhibited in
the COPR material and that may explain its persistence some
80 yr after its deposition. The inclusion of magnesium in the
crystal structure of brownmillerite was reported to slow its hydration as compared to pure brownmillerite (Jupe et al., 2001);
this may explain the persistence of brownmillerite since magnesium constitutes approximately 6% of the COPR material.
Moreover, the rate of brownmillerite hydration cannot be determined since the original composition of the COPR minerals after the quenching process is not known and there is no reliable
estimate of the initial brownmillerite content and its hydration
byproducts. Also, it is not clear whether brownmillerite underwent further hydration after deposition. The hydration byproducts of brownmillerite are katoite (a hydrogarnet), portlandite,
and hematite (Tamas and Vertes, 1972). Katoite was predicted
in the residue of the COPR leaching tests at pH > 11.5. The
model also predicted the sequestration of magnesium in hydrotalcite; however, the experimental data showed that magnesium
was sequestered in brucite and periclase in addition to hydrotalcites. Brucite and periclase could be meta-stable states that will
transform into hydrotalcite if and when sucient aluminum is
released from brownmillerite hydration. Hematite and diaspore
predicted by the model are the crystalline phases of the amorphous iron and aluminum hydroxides, respectively.
The model predicted the concentration of Cr(VI) to be controlled by adsorption at pH < 8 and by precipitation at pH >
10.5 (Fig. 5). It predicted Cr(VI) sequestration in Cr(VI)hydrocalumite and Cr(VI)-ettringite mineral phases at pH > 11
and 10.5 < pH < 11.5, respectively (Fig. 6). In the pH region
9 < pH < 10.5, chromate was predicted almost to be 100% in
the dissolved phase (Fig. 5). However, the experimental results
indicated that the model overpredicted Cr(VI) concentration.

Journal of Environmental Quality Volume 37 NovemberDecember 2008

Table 3. Rietveld quantification of minerals and phases in the residues of the leaching tests.
pH
Mineral phases
Calcium aluminum oxide chromium hydrate
Brownmillerite
Brucite
Calcite
Hydroandradite
Katoite
Periclase
Quartz
Quinitinite-2H
Sjoegrenite
Albite
Gibbsite
Percent crystalline phases
Noncrystalline phase
Percent total

12.03
11.04
10.58
9.35
8.30
7.76
6.42
5.24
%
0.87
26.82
2.84
8.53
3.19
2.61
2.21
3.19
1.22
1.45
5.17

0.69
33.84
3.64
6.04
3.98
3.77
2.40
3.64
1.65
1.85
7.00

28.47
2.14
6.38
2.31
2.09
2.09
2.73
1.66
1.23
4.50

29.66
1.87
6.84
1.62
0.69

37.16
1.41
7.18
1.59
0.47

37.94
0.74
6.84

33.99
0.81
5.01

20.57

3.15
0.64
4.72
4.72

3.95

4.27

4.42

4.41

1.47
5.59

1.14
6.04

0.86
8.78

58.11
41.94
100.06

68.50
31.37
99.86

53.62
46.38
100.00

49.18
50.82
100.00

58.83
41.11
99.94

56.97
43.03
100.00

53.86
46.14
100.00

0.60
7.11
1.67
35.53
64.47
100.00

The only stable mineral in that pH region that may contain


chromate through anionic exchange is hydrotalcite (Fig. 5 and
6). Hydrotalcites are reported in the literature to be capable of
removing chromate from solution through anionic exchange
(Goswamee et al., 1998; Lazaridis and Asouhidou, 2003; Terry,

1.17

2004; Alvarez-Ayuso and Nugteren, 2005). Specically, the


data of Terry (2004) indicates Cr(VI) removal rates of approximately 96 and 95% from solutions with initial concentrations
of 5 and 20 mg/L, respectively using hydrotalcite concentrations of 5 g/L at pH values of 2.0 and 2.1. Lower Cr(VI) re-

Fig. 4. Scanning electron microscopy (SEM) images of the leaching residue at pH 10.9 and pH 9.4.

Wazne et al.: Leaching Mechanisms of Cr(VI) from Chromite Ore Processing Residue

2131

Fig. 5. Model prediction of Cr(VI) concentration as function of pH for the leaching tests.

moval rates were obtained at higher pH values. Approximately


30% Cr(VI) removal rates were obtained at pH value of approximately 9 for solutions with initial Cr(VI) concentration of
20 mg/L. It is worth noting that hydrotalcite is not expected to
be stable at pH 2, and the high Cr(VI) removal rates that Terry
attributed to hydrotalcite may in fact be due to the adsorption
of chromate anion onto aluminum hydroxide generated by the
dissolution of hydrotalcite. Terry did not investigate the stability of hydrotalcite at low pH values or present any data to suggest that it is stable. On the other hand the removal rates that
were achieved by Terry at higher pH values are similar to the
rates reported in this study; that is, 27% removal rate at pH of
approximately 9.5.
On the other hand, Alvarez-Ayuso and Nugteren (2005,
Wat. Res., 39, 25352542) reported that Cr(VI) adsorption at
high pH values decreased when much longer times than those
required to reach equilibrium were used. Alvarez-Ayuso and
Nugteren deduced that chromium release occurred under such
conditions. Accordingly, these phenomena can be attributed
to the slow dissolution of calcinated hydrotalcite and/or to
the replacement of previously sorbed chromium by carbonate.
However, Alvarez-Ayuso and Nugteren did not report similar
desorption behavior for the noncalcinated hydrotalcite, which
was used in this study.
Therefore, to our knowledge chromate exchange from
COPR leachate by hydrotalcites or chromate substitution in
hydrotalcites in COPR has not been reported in the literature. Hydrotalcite phases such as sjoegrenite, Mg6Fe2(CO3)
(OH)145H2O, and quintinite-2H, MgAl2(OH)12(CO3)4H2O,
were identied in the COPR samples down to pH of 8 (Fig. 3
and 4, Table 2). These belong to a class of layered compounds
derived from the structure of mineral brucite, Mg(OH)2.
Brucite is comprised of layered sheets of neutral octahedrons
of magnesium hydroxide. When a fraction, x, of Mg2+ ions
are isomorphously substituted by Al3+ ions, the sheets acquire
the composition [Mg(1x)Alx(OH)2]x+ (0.2 < x < 0.33) and
become positively charged (Radha et al., 2005). Anions, (Cl,

2132

NO3, CO3 2 and others), are incorporated in the interlayer


region along with water molecules to balance the charge decit
and restore stability. Many divalent ions such as Ni, Co, Cu,
and Zn form LDHs with trivalent ions such as Al, Cr, Fe, and
Ga produce a very large class of isostructural compounds (Trave
et al., 2002; Bontchev et al., 2003).
The decrease in Cr(VI) concentration at pH > 9 (Fig. 1) as
the leaching experiment progressed may provide an indication
of chromate substitution in hydrotalcites. To further assess the
capacity of synthetic hydrotalcite to remove chromate from
COPR leachate an experiment was conducted using synthetic
hydrotalcite. Approximately 27% of chromate was removed
from the COPR leachate when approximately 386 mg/L
Cr(VI) COPR leachate was mixed with 0.1 g/L synthetic hydrotalcite at pH 9.5 for 1 wk. The calculated amount of Cr(VI)
exchanged is 20.8 mg/g which is smaller than the value reported by Lazaridis and Asouhidou (2003) of 33 mg/g at pH 10
from 10 mg/L synthetic chromate solution. This indicates that
chromate in the COPR leachates can substitute for OH or
CO32 in the interlayers of hydrotalcite. This may indicate that
for treatment schemes pH has to be decreased to <8 to destabilize hydrotalcite phases.
The underestimation of Cr(VI) concentration at pH 8
could be due to the competitive adsorption of silicates (stronger than that predicted by the model) which strongly adsorbs
in that pH region. This may have consumed most of the available adsorption sites and caused higher Cr(VI) leaching rates.
The COPR mineralogy as it exits the roasting process is
comprised of a mixture of brownmillerite (Ca4Fe2Al2O10), periclase (MgO) and quicklime (CaO), which are considered to be
the parent COPR minerals, in addition to minor impurities
(Allied Signal, 1982). Based on experimental observations,
COPR consists mainly of two groups of minerals that behave
dierently: fast reactants, mainly hydrates and carbonates,
and slow reactants, such as brownmillerite and periclase. Fast
reactants imply that the reactions will occur within minutes
and slow implies months or years to complete. Cr(VI) is

Journal of Environmental Quality Volume 37 NovemberDecember 2008

Fig. 6. Model prediction of mineral phases, present in the residues of the leaching tests for composite sample B1B2, at various equilibrium pHs.

found mainly in calcium aluminium oxide chromium hydrates


(CACs), (monochromate or Cr(VI)-hydrocalumite) and is
partially present in hydrotalcites (this study) and hydrogarnets
(Hillier et al., 2007) through anionic substitution. The experimental results indicated that Cr(VI) bearing minerals react
within minutes, whereas brownmillerite hydration may take
years to complete. However, the two groups of minerals are
sometimes partially intertwined. Modeling of the leaching data
should allow for the incorporation of chromates into various
mineral phases through anionic exchange with phases such as
hydrogarnets, hydrotalcites, and other phases (once identied).
Moreover, modeling should also account for the slow dissolution of brownmillerite and hence the availability of Cr(VI)
when chromate phases are encapsulated into the brownmillerite
nodules. A major nding of this study is the strong evidence
about the presence of chromates in hydrotalcites. However,
more research is needed to determine the percentages of readily
available and nonreadily available (encapsulated in brownmillerite nodules) chromates present in CAC(s), Cr(VI)-ettringite,
hydrogarnts and hydrotalcites.
Finally, the complete hydration of brownmillerite and the
subsequent dissolution of its byproducts are expected to release additional alkalinity according to the following equation:
Ca4Fe2Al2O10 +7H2OCa(OH)2 +Ca3Al2(OH)12 +Fe2O3 [1]
This may cause pH to stabilize at values higher than those
measured in Fig. 2. For example, for a brownmillerite content
of 25% of the COPR matrix, the complete dissolution of
brownmillerite requires additional 4[H+] eq/Kg. The resulting
mineralogy at equilibrium at various pH values is shown in Fig.
5. However, the complete dissolution of brownmillerite is not
expected to signicantly aect the model prediction of aqueous

Cr(VI) concentration at various pH values because the dissolution


and precipitation of Cr(VI) minerals are relatively fast (minutes);
similarly, the adsorption and desorption of chromate are also
relatively fast. A model simulation with 25% additional adsorbent
sites shifted the adsorption edge <1 pH unit to the right at pH
values <9 (Fig. 6); however, the additional adsorbent had no eect
on the model simulation at pH > 9 where anion exchange and
precipitation are expected to dominate Cr(VI) solubility.

Conclusions
The leaching mechanism of Cr(VI) from COPR was investigated using the XRPD analyses and geochemical modeling.
The model was able to simulate the adsorption edge at approximately pH 5 and the precipitation edge at approximately
pH 11. The experimental and modeling results suggested the
presence of a mineral phase controlling the solubility of Cr(VI)
in the pH region 8 < pH < 11. Experimental results indicated
that hydrotalcite may be controlling the solubility of chromate
though anionic exchange in that pH region. The COPR, at the
deposition sites, is expected to continuously leach Cr(VI) at a
concentration equivalent to the solubility of Cr(VI) with respect Cr(VI)-hydrocalumite. If pH were to drop below pH 11,
signicant increase in Cr(VI) leaching will ensue; however, this
seems unlikely due to the high buering capacity of COPR.
Finally, all chromium was released when 32[H+] eq/Kg of acid
was added to the COPR matrix. This information is very useful
for stabilization or recovery of chromium in COPR matrices.

Acknowledgments
The authors wish to thank Honeywell International Inc.
for the nancial support of this study.

Wazne et al.: Leaching Mechanisms of Cr(VI) from Chromite Ore Processing Residue

2133

References
Allied Signal. 1982. Process description-Baltimore works. Honeywell,
Morristown, NJ.
Alvarez-Ayuso, E., and H.W. Nugteren. 2005. Purication of chromium(VI)
nishing wastewaters using calcinated and uncalcinated Mg-Al-CO3hydrotalcite. Water Res. 39:25352542.
American Society for Testing Material. 2001. Test methods for instrumental
determination of carbon, hydrogen, and nitrogen in petroleum
products and lubricants. ASTM D 5291-96. Annual Book ASTM
Stand., 236240. ASTM, West Conshohocken, PA.
Bennet, D.G., D. Read, M. Atkins, and F.P. Glasser. 1992. A thermodynamic
model for blended cements: II. Cement hydrate phases; thermodynamic
values, and modeling studies. J. Nucl. Mater. 190:315325.
Bontchev, R.P., S. Liu, J.L. Krumhansl, J. Voigt, and T.M. Neno. 2003.
Synthesis, characterization, and ion exchange properties of hydrotalcite
Mg6Al2(OH)16(A)x(A)2-x 4H2O (A, A = Cl-, Br-, I-, and NO3-, 2 x 0)
derivatives. Chem. Mater. 15:36693675.
Burke, T., J. Fagliano, M. Goldoft, R.E. Hazen, R. Iglewicz, and T. McKee.
1991. Chromite ore processing residue in Hudson County, New Jersey.
Environ. Health Perspect. 92:131137.
Common Thermodynamic Database Project. 2004. CTDP homepage.
Available at http://www.cig.ensmp.fr/~vanderlee/ctdp/index.html
(veried 6 Aug. 2008).
Darrie, R.G. 2001. Commercial extraction technology and process waste
disposal in the manufacture of chromium chemicals from ore. Environ.
Geochem. Health 23:187193.
David, S.B., and J.D. Allison. 1999. Minteqa2, an equilibrium metal speciation
model: Users manual 4.01. Environmental Res. Lab., USEPA, Athens, GA.
Dermatas, D., R. Bonaparte, M. Chrysochoou, and D.H. Moon. 2006b.
Chromite ore processing residue (COPR): Contaminated soil or
hazardous waste? J. ASTM Int. 3(7): doi: 10.1520/JAI13313.
Dermatas, D., M. Chrysochoou, D.H. Moon, D.G. Grubb, M. Wazne, and
C. Christodoulatos. 2006a. Ettringite-induced heave in chromite ore
processing residue (COPR) upon ferrous sulfate treatment. Environ.
Sci. Technol. 40(18):57865792.
Dzombak, D.A., and F.M.M. Morel. 1990. Surface complexation modeling:
Hydrous ferric oxide. John Wiley & Sons, New York.
Farmer, J.G., M.C. Graham, R.P. Thomas, C. Licona-Manzur, E. Paterson, and
C.D. Campbell. 1999. Assessment and modeling of the environmental
chemistry and potential for remediative treatment of chromiumcontaminated land. Environ. Geochem. Health 21:331337.
Farmer, J.G., R.P. Thomas, M.C. Graham, J.S. Geelhoed, D.G. Lumsdon,
and E. Paterson. 2002. Chromium speciation and fractionation in
ground and surface waters in the vicinity of chromite ore processing
residue disposal sites. J. Environ. Monit. 4:235243.
Geelhoed, J.S., J.C.L. Meeussen, S. Hillier, D.G. Lumsdon, and R.P. Thomas.
2002. Identication and geochemical modeling of processes controlling
leaching of Cr(VI) and other major elements from chromite ore
processing residue. Geochim. Cosmochim. Acta 66(22):39273942.
Goswamee, R.L., P. Sengupta, K.G. Bhattacharyya, and D.K. Dutta. 1998.
Adsorption of Cr(VI) in layered double hydroxoides. Appl. Clay Sci.
13:2134.
Graham, M.C., J.G. Farmer, P. Anderson, E. Paterson, S. Hillier, D.G.
Lumsdon, and R. Bewley. 2006. Calcium polysulde remediation of
hexavalent chromium contamination from chromite ore processing
residue. Sci. Total Environ. 364:3244.
Higgins, T.E., A.R. Halloran, M.E. Dobbins, and A.J. Pittignano. 1998.
In-situ reduction of hexavalent chromium in alkaline soils enriched
with chromite ore processing residue. Air Waste Manage. Assoc.
48(11):11001106.
Hillier, S., D.G. Lumsdon, R. Brydson, and E. Paterson. 2007. Hydrogarnet:
A host phase for Cr(VI) in chromite ore processing residue (COPR)
and other high pH wastes. Environ. Sci. Technol. 41(6):19211927.
Inorganic Crystal Structure Database. 2005. Fachinformationszentrum,
Karlsruhe, Germany.
International Centre for Diraction Data. 2002. Powder diraction le,
PDF-2 database release. International Centre for Diraction Data,
Newtown Square, PA.

2134

James, B.R. 1996. The challenge of remediating chromium-contaminated


soil. Environ. Sci. Technol. 30:248A251A.
Jing, C., S. Liu, G.P. Koratis, and X. Meng. 2006. Leaching behavior of
Cr(III) in stabilized/solidied soil. Chemosphere 64:379385.
Jupe, A.C., J.K. Cockcroft, P. Barnes, S.L. Colston, G. Sanker, and C. Hall.
2001. The site occupancy of Mg in the brownmillerite structure and
its eect on hydration properties: An X-ray/neutron diraction and
EXAFS study. J. Appl. Crystallogr. 34:5561.
Kremser, D.D., B.M. Sass, G.I. Clark, M. Bhargava, and C. French. 2005.
Environmental forensics investigation of buried chromite ore processing
residue. Paper E-12. In B.C. Alleman and M.E. Kelley (conf. chairs) In
Situ and On-Site Bioremediation2005. Proc. Of the 8th Int. In Situ
and On-Site Bioremediation Symp., Baltimore, MD. 69 June 2005.
Battelle Press, Columbus, OH.
Lawrence Livermore National Laboratory. 1994. EQ3/6 Database. Version 8,
Release 6. Lawrence Livermore Natl. Lab., Livermore, CA.
Lazaridis, N.K., and D.D. Asouhidou. 2003. Kinetics of sorptive removal
of chromium(VI) from aqueous solutions by calcinated Mg-Al-CO3
hydrotalcite. Water Res. 37:28752882.
Materials Data Inc. 2005. Jade, Version 7.1. Materials Data Inc., Livermore, CA.
Meng, X., S. Bang, and G.P. Koratis. 2000. Eects of silicate, sulfate,
and carbonate on arsenic removal by ferric chloride. Water Res.
34(4):12551261.
Parkhurst, D.L., and C.A.J. Appelo. 1999. Users guide to Phreeqc (Version 2)-A
computer program speciation, batch-reaction, One-dimensional transport
and inverse geochemical calculations. U.S. Geol. Surv., Denver, CO.
Radha, A.V., P.V. Kamath, and C. Shivakumara. 2005. Mechanisms of the
anion exchange of the layered double hydroxides (LDHs) of Ca and Mg
with Al. Soil State Sci. 7(10):11801187.
Reardon, E.J. 1992. Problems and approaches to the prediction of chemical
composition in cement/water systems. Waste Manage. 12:221239.
Rietveld, H.M. 1969. A prole renement method for nuclear and magnetic
structures. J. Appl. Crystallogr. 2:6571.
Snoeyink, V.L., and D. Jenkins. 1980. Water chemistry. John Wiley & Sons,
New York.
Tamas, D.F., and A. Vertes. 1972. A mossbauer study of the hydration of
brownmillerite. Cement Concrete Res. 3:575581.
Terry, P. 2004. Characterization of Cr ion exchange with hydrotalcite.
Chemosphere 57:541546.
Trave, A., A. Selloni, A. Goursot, D. Tichit, and J. Weber. 2002. First
principles study of the structure and chemistry of Mg-based
hydrotalcite-like anionic clays. J. Phys. Chem. B 106:1229112296.
U.S. Environmental Protection Agency. 1992. Chromium, hexavalent
(colorimetric), Method 7196A. USEPA, Washington, DC.
U.S. Environmental Protection Agency. 1996a. Microwave assisted acid
digestion of sediments, sludges, soils, and oils. Method 3051A. USEPA,
Washington, DC.
U.S. Environmental Protection Agency. 1996b. Inductive coupled plasmaatomic emission spectrometry. Method 6010B. USEPA, Washington, DC.
U.S. Environmental Protection Agency. 1996c. Alkaline digestion for
hexavalent chromium. Method 3060A. USEPA, Washington, DC.
U.S. Environmental Protection Agency. 2006. Visual MINTEQ Version 2.5.
USEPA, Washington, DC. Available at http://www.lwr.kth.se/English/
OurSoftware/vminteq/ (veried 7 Aug. 2008).
Van Geen, A., A.P. Robertson, and J.O. Leckie. 1994. Complexation of carbonate
species at the goethite surface: Implications for adsorption of metal ions in
natural waters. Geochim. Cosmochim. Acta 58(9):20732086.
Weng, C.H., C.P. Huang, H.E. Allen, A.H.D. Cheng, and P.F. Sanders.
1994. Chromium leaching behavior in soil derived from chromite ore
processing waste. Sci. Total Environ. 154:7186.
Wijnja, H., and C.P. Schulthess. 2001. Carbonate adsorption mechanism on
goethite studied with ATR-FTIR, DRIFT, and proton coadsorption
measurements. Soil Sci. Soc. Am. J. 65:324330.
Yalin, S., and K. nl. 2006. Modelling chromium dissolution and leaching
from chromite ore processing residue. Environ. Eng. Sci. 23(1):187201.
Zachara, J.M., D.C. Girvin, R.L. Schmidt, and C.T. Resch. 1987. Chromate
adsorption on amorphous iron oxyhydroxide in the presence of major
groundwater ions. Environ. Sci. Technol. 21:589594.

Journal of Environmental Quality Volume 37 NovemberDecember 2008