You are on page 1of 11

Minerals Engineering 99 (2016) 818

Contents lists available at ScienceDirect

Minerals Engineering
journal homepage: www.elsevier.com/locate/mineng

Comprehensive examination of acid leaching behaviour of mineral


phases from red mud: Recovery of Fe, Al, Ti, and Si
Rachel A. Pepper, Sara J. Couperthwaite, Graeme J. Millar
Institute for Future Environments & School of Chemistry, Physics & Mechanical Engineering, Queensland University of Technology (QUT), GPO Box 2434, Brisbane,
Queensland 4001, Australia

a r t i c l e

i n f o

Article history:
Received 10 March 2016
Revised 6 September 2016
Accepted 23 September 2016

Keywords:
Red mud
Acid leaching
Metal recovery
Bauxite residue
Waste processing

a b s t r a c t
Red mud represents an environmental and economic liability for the alumina industry in the form of
wasted raw material. Although some leaching studies have been performed, there are deficiencies in
the current literature regarding the amounts of metals extracted relative to each other, and minimal
information regarding silicon contamination of extracts. There is also limited knowledge of extraction
efficiencies of different acids (particularly in the case of phosphoric acid) under the same experimental
conditions. This study focused on the leaching behaviour of the four most extractable elements present
within red mud (iron, titanium, aluminium and silicon). By varying the experimental conditions, acid
concentration, and type of acid, a comprehensive dataset of leaching trends was obtained. This allowed
for direct comparison of leaching efficiency for the four elements under the same conditions, which was
difficult previously due to the variation of experimental conditions and red mud composition between
studies. The patterns in recoveries were explained in terms of the reactivities of the mineral phases
within red mud and the interaction between the different acids and the reaction surfaces. Out of the four
acids studied (nitric, hydrochloric, sulfuric, and phosphoric) phosphoric and hydrochloric acids produced
some of the best recoveries for iron (7678%) and titanium (2324%), and phosphoric acid also produced
the highest recoveries for silicon (49%) and aluminium (50%). The differences observed between the acid
types and reaction conditions revealed potential for development of element selective extraction methods. Additionally, explaining leaching behaviour in terms of the mineral phases present allowed easier
prediction of expected leaching trends for these four elements, which made this study applicable to
red muds with a wide variety of compositions.
2016 Elsevier Ltd. All rights reserved.

1. Introduction
Metals present in industrial wastes represent significant losses
in raw material and are a significant environmental liability (Power
et al., 2011). Thus, development of practices to recover and reuse
metals from wastes or to treat industrial effluents is potentially
of great value. The alumina industry produces one of the largest
industrial waste streams in the form of red mud; for every tonne
of alumina extracted, between 1 and 1.5 tons of red mud is generated (Liu et al., 2009). Global red mud storage deposits are estimated at around 3 billion tons with an additional 120 million
tons produced per annum (Marcel Ct and Wilson, 2012). Due
to the highly caustic nature of the red mud slurry and the high proportion of heavy metals within it, disposal has been a major issue
(Power et al., 2011). Before 1970, disposal of red mud was achieved
Corresponding author.
E-mail address: graeme.millar@qut.edu.au (G.J. Millar).
http://dx.doi.org/10.1016/j.mineng.2016.09.012
0892-6875/ 2016 Elsevier Ltd. All rights reserved.

through marine discharge, where the slurry was pumped via a


pipeline into the deep ocean; or lagooning, where the slurry was
contained in inland ponds (Power et al., 2011). Recently, emphasis
has been placed on dry stacking disposal methods where the residue was thickened to a paste and pumped into pits (Power et al.,
2011). The disadvantages of dry stacking and lagooning methods
are that they use a lot of land area, which becomes unsuitable
for further use. Additionally, disposal of red mud by such methods
represents a loss of raw materials in the form of metals.
Red mud is predominantly comprised of oxides of aluminium,
iron, titanium, and silicon, with other mixed mineral phases and
trace heavy metals present (Evans et al., 2012); the proportion of
each species is highly dependent on the original source of bauxite
and the refinery conditions used (Evans et al., 2012). However,
typically high amounts of aluminium oxide (530 wt%), iron oxide
(560 wt%) and titanium oxide (0.315 wt%) make red mud a
potential candidate for metal recovery (Evans et al., 2012). In light
of this latter fact, considerable research has been undertaken to

R.A. Pepper et al. / Minerals Engineering 99 (2016) 818

develop methods of extracting these aforementioned metals (Liu


et al., 2009, 2012; Piga et al., 1995; Kitajima and Kasai, 1999;
Uzun and Gulfen, 2007; Kasliwal and Sai, 1999; Li et al., 2011;
Borra et al., 2015; Mishra et al., 2001; S
ayan and Bayramoglu,
2000, 2001; Papadopoulos et al., 2010). Tactics for recovering metals can be divided into two generic approaches, namely chemical
(leaching) Uzun and Gulfen, 2007; Borra et al., 2015; Sayan and
Bayramoglu, 2000, 2001; Papadopoulos et al., 2010 and physical
(magnetic separation and sintering) (Liu et al., 2009, 2012; Piga
et al., 1995; Kitajima and Kasai, 1999; Li et al., 2011; Mishra
et al., 2001).
Chemical leaching of metals from red mud is an attractive prospect due to its potential to allow selective recovery through
manipulation of experimental conditions. Nevertheless, the current literature is somewhat limited in this regard. Many previous
studies have focused on the extraction of a single element, with little regard to others, by varying parameters such as acid concentration, leaching time, leaching temperature, agitation rate and
amount of solid in regards to the amount of acid (Uzun and
Gulfen, 2007; Kasliwal and Sai, 1999; Borra et al., 2015; S
ayan
and Bayramoglu, 2000, 2001; Papadopoulos et al., 2010). Examples
of acids used in these leaching studies include sulfuric (S
ayan and
Bayramoglu, 2000, 2001), hydrochloric (Kasliwal and Sai, 1999)
and nitric (Borra et al., 2015), although phosphoric acid seems to
have been overlooked. There exists general agreement that higher
concentrations of acid result in greater levels of iron (Fe) and aluminium (Al) leaching Uzun and Gulfen, 2007; Borra et al., 2015;
Sayan and Bayramoglu, 2001, a trend which also extends to other
elements such as titanium (Ti) S
ayan and Bayramoglu, 2000,
2001 and rare earth elements (Borra et al., 2015). This latter trend
was also observed when elevated leaching temperatures were
employed (Uzun and Gulfen, 2007; Kasliwal and Sai, 1999; S
ayan
and Bayramoglu, 2000, 2001). Agitation of the sample during
digestion also seems to play a role in the amount of dissolution
of red mud. For example, Uzun and Gulfen (2007) found that
increasing the agitation rate was effective (up to 400 rpm) in accelerating the amount of leached metal, after which no benefit was
observed. Other studies (Sayan and Bayramoglu, 2000) have also
demonstrated an enhancement in leaching with higher agitation
rates, but the consensus seems to be that it is of lesser importance
than temperature or acid concentration. The final variable commonly tested by researchers is the solid to liquid ratio. A higher relative amount of acid compared to red mud has been correlated
with higher metal recovery (Borra et al., 2015). Iron extraction as
high as 97.46% using 6 M sulfuric acid at 105 C was reached by
Uzun and Gulfen (2007), while aluminium recoveries of 96.82%
were reported by S
ayan and Bayramoglu (2001) using 4 M sulfuric
acid at 90 C, with a 4 h reaction time and a solid:solution ratio of
0.04. The same study reported maximum titanium recoveries,
using the same conditions, of 96.33%.
The difficulty in evaluating results from these outlined studies
is that the reported recoveries are usually with regard to a specific
acid concentration and temperature, so it is difficult to compare
results. Additionally, the composition of red mud varies between
studies, with some experiments using Greek bauxite residue
(Borra et al., 2015) that had a higher proportion of iron oxides than
Turkish red mud (S ayan and Bayramoglu, 2000) (44.6 and 37.7%,
respectively). Such differences between the starting samples could
impact the final composition of the acidic extracts and adds further
complexity when comparing specific results from different studies.
Although the studies discussed above have reported good
recoveries of their metals of focus, there is little mention with
regards to the other elements present in the extract, which is vital
information if the extract is intended for further use. Multielement extraction from red mud has not been studied in detail,
particularly in the case of silicon. The hypothesis of this study is

that unless side by side extractions are carried out, the leaching
patterns for multiple elements relative to each other cannot be
determined. To confirm this hypothesis it was necessary to assess
the effects that variation of experimental parameters such as reaction time, temperature and solid:solution ratio had on the final
amount of each element leached from red mud, what effect acid
concentration and type had on the extraction of these four elements, as well as if any elements were favoured over others under
certain conditions.
This paper has addressed these questions through a series of
leaching experiments, and through analysis of these results some
leaching trends have been established for the four elements studied. The results represent valuable insight of leaching trends for
four of the most extractable elements found in red mud, particularly in the case of silicon, which has been previously overlooked.
The importance of knowing the composition of red mud extracts
should not be underestimated, particularly in the context of possible tailored extraction methods or re-use of the extracts for metal
scavenging or recovery.
2. Material and methods
2.1. Materials
Red mud used in this study was obtained from an Australian
alumina refinery in the form of a slurry. The slurry was subsequently filtered to isolate the solid red mud, washed with deionised water until the filtrate was a neutral pH, dried at 80 C and
finally crushed. The crushed red mud was then sieved to obtain
particles smaller than 212 lm. AR grade hydrochloric (32 wt/v%),
nitric (70 wt/v%), sulfuric (98 wt/v%) and phosphoric (80 wt/v%)
acids were used, and required concentrations for testing purposes
were made from these concentrated stock solutions along with
MilliQ ultrapure water through serial dilutions.
2.2. Experimental procedure
Acid leaching of red mud was performed using hydrochloric,
nitric, sulfuric and phosphoric acids. The general procedure for
these experiments involved: adding the desired amount of red
mud and acid to 50 mL falcon tubes; allowing the digestion to proceed under specific conditions as described below; and then collecting the solid and liquid phases upon completion of each test.
Separation of the solids from the liquid samples was achieved via
centrifugation (3000 rpm for 10 min, Eppendorf 5702 centrifuge).
All liquid samples were syringe filtered prior to storage and solid
samples were washed with MilliQ water, dried in an oven at
80 C and crushed before storage. ORP and pH measurements were
taken of the liquid extracts using a Cole-Parmer Ag/AgCl electrode
and a TPS pH probe. All testing series explored in this study were
conducted at four different acid concentrations (0.01, 0.1, 1.0 and
5.0 M) to compare the leaching and metal recovery profiles. Particulars for each experimental test are as follows:
Temperature leaching trials: A solid:solution ratio of 0.005 and a
leaching time of 24 h were used for all tests. The digestions were
performed in a pre-heated bead bath using pre-heated acids with
manual agitation. Temperatures tested were 25, 40, 55, 70 and
80 C.
Time leaching trials: Reactions in this series were performed for
1, 4, 24, 48 and 168 h with a solid:solution ratio of 0.005 at room
temperature (approximately 25 C). A rotary stirrer was used to
ensure constant, gentle agitation.
Solid:solution ratio leaching trials: Reactions in this series were
performed at room temperature on a rotary stirrer to ensure constant, gentle agitation for a period of 24 h. The solid:solution ratios
that were tested were 0.00125, 0.0025, 0.005, 0.0125 and 0.025.

10

R.A. Pepper et al. / Minerals Engineering 99 (2016) 818

2.3. Characterisation techniques


Solutions were analysed using a VISTA-MPX CCD simultaneous
Inductively Coupled Plasma Optical Emission Spectrometer (ICPOES) instrument using an integration time of 0.15 s with 3 replications, with the following wavelengths (nm): Al (308.215), Si
(212.412), Ti (334.94) and Fe (238.204). Instrument calibration
was performed using the method of external standardisation utilizing multi-element standards prepared in-house from Inductively Coupled Plasma Mass Spectrometry (ICP-MS) grade single
element stock solutions (High Purity Standards, Charleston, USA).
Method robustness, accuracy and precision was verified by continuing analysis of a number of Certified Reference Materials (CRMs)
covering a range of common matrices and analyte concentrations
(National Institute of Standards and Technology, Gaithersburg,
MD, USA, United States Geological Survey, Reston, VA, USA).
Powder X-ray diffraction (XRD) patterns of the red mud starting
product were collected using a Panalytical Xpert wide angle X-ray
diffractometer, operating in step scan mode, with Co Ka radiation
(1.7903 ). Corundum was used as an internal standard and was
incorporated into the red mud as 10% by weight prior to analysis.
Patterns were collected in the range 590 2h with a step size of
0.02 and a rate of 30 s per step. The XRD patterns were matched
with ICSD reference patterns using the software package HighScore
Plus. The profile fitting option of the software used a model that
employed twelve intrinsic parameters to describe the profile, the
instrumental aberration and wavelength dependent contributions
to the profile. Powder diffraction patterns were also analysed using
Total Pattern Analysis Solutions (Topas V5, Bruker) to obtain quantitative data. The Rietveld method was employed. A cobalt emission profile unique to the diffractometer was used to model the
incident X-ray beam. The background was modelled with a Chebyshev polynomial. A measured instrument function from NIST SRM
660a was calculated and used to model the profiles via refinement
of a Lorentzian crystallite size term for each phase. Corundum
(Al2O3) was set as the spiked phase (10 wt%); results were reported
as the wt% in the original sample and on an absolute scale. The
non-diffracting/unidentified component was the sum of all the
modelled phases subtracted from 100 wt%. This component was
the least accurately determined as its error was the sum of the
errors of the other phases. Other refined parameters included sample displacement, a scale factor for each phase and unit cell
parameters.
X-ray fluorescence (XRF) spectrometry was conducted on a
Spectro XEPOS 03 Energy Dispersive X-ray fluorescence (ED-XRF)
spectrometer. A micronized sample of red mud was placed in the
oven at 950 C for 2 h to remove moisture and carbonate species,
and to determine the loss on ignition (LOI). Fused glass discs were
prepared from the samples with a 50:50 flux ratio using the OX
Claisse Fluxer. 44-P, a certified reference material previously available from ORE (Ore Research and Exploration), was used as a check
standard. It was a certified reference material based mainly on a
composite of fully oxidised laterites and hematites from Australian
ore deposits, which had the advantage of having a similar matrix
and elemental composition to red mud.
3. Results and discussion
3.1. Characterisation of red mud
The XRD pattern for the red mud used in this study is shown in
Fig. 1. Primary phases identified in the red mud sample included
quartz (SiO2), anatase (TiO2), rutile (TiO2), hematite (Fe2O3),
calcite (CaCO3), boehmite (c-AlOOH), gibbsite (Al(OH)3), sodalite
(Na8(Al6Si6O24)Cl2) and kaolinite (Al2Si2O5(OH)4), all of which were
typical mineral components of red mud (Klauber et al., 2011). The

Fig. 1. X-ray diffraction pattern for red mud with corresponding reference patterns.

phase compositions and elemental compositions of the red mud


from the TOPAS refinement may be found in Table 1. The refinement revealed that the red mud samples were comprised of a large
proportion of amorphous material, which precluded precise phase
compositions to be determined. While the phase identification by
HighScore Plus suggested the presence of sodalite, the Rietveld
refinement showed a poor agreement between the fitted and
observed intensities. The reason for the latter intensity mismatch
was unclear, but may have been due to differences in reference
patterns between TOPAS and HighScore software, or from nonideal mineral compositions within red mud compared to naturally
occurring forms of these mineral phases (Santini, 2015). Problems
in identifying the mineral phases in red mud by XRD due to a high
amorphous content have been documented previously (Santini,
2015).
XRF data was also collected to obtain more information about
the mineral composition of red mud, which was deemed to be
comprised of Fe2O3 (29.82%), Al2O3 (21.69%), SiO2 (12.28%), ZrO2
(9.11%), TiO2 (6.88%), Na2O (5.19%), CaO (1.79%), MgO (1.483%),
SO3 (0.21%), K2O (0.18%), MnO (0035%) and P2O5 (<0.001%), with
a LOI of 11.06%. Comparison of the XRF information with the
XRD results revealed that the presence of some elements, such as
Al, Ca, Fe, O, Si and Zr in particular had been underestimated.
Therefore, it was inferred that these latter species were probably
present as amorphous material that was not detectable by XRD.
This structural analysis data agreed with typical elemental and
mineralogical results for red mud with the exception of Zr, which
has been observed before in red mud (Qu et al., 2013), although
not as prevalent as in the red mud used in this study. Red mud is
normally comprised of Fe2O3 (560%), Al2O3 (530%), TiO2 (0.3
15%), CaO (214%), SiO2 (350%) and Na2O (110%) Evans et al.,
2012. These wide variations in reported red mud compositions
reflect the large influence that the composition of the bauxite ore
and plant conditions has on the resultant red mud composition.

11

R.A. Pepper et al. / Minerals Engineering 99 (2016) 818


Table 1
Phase compositions and elemental compositions of red using TOPAS refinement.
Phase

Chemical formula

Phase composition (%)

Element

Elemental composition (%)

Quartz
Anatase
Rutile
Hematite
Calcite
Boehmite
Gibbsite
Sodalite
Kaolinite
Unidentified

SiO2
TiO2
TiO2
Fe2O3
CaCO3
c-AlOOH
Al(OH)3
Na8(Al6Si6O24)Cl2
Al2Si2O5(OH)4

1.95
3.16
1.96
26.85
0.69
8.00
2.51
4.63
2.02
48.23

0.370
0.067
0.022
0.456
0.038
0.133
0.065
0.218
0.161
0.634

Al
C
Ca
Cl
Fe
H
Na
O
Si
Ti
Unidentified

5.73
0.08
0.28
0.34
18.78
0.10
0.88
20.36
2.16
3.07
48.23

0.027
0.004
0.013
0.014
0.285
0.002
0.037
0.205
0.137
0.041
0.565

3.2. Elemental recoveries


Selected excerpts of the leaching study have been presented for
red muds four most extractable elements (Al, Si, Fe and Ti). The
selected data sets were chosen to allow easier comparison of leaching trends under different reaction conditions. Except where otherwise stated, the results presented in Figs. 25 used the following
conditions: 1 M acid, 25 C reaction temperature, 24 h reaction
time, and a red mud:acid ratio of 0.005.
3.2.1. Aluminium
The primary sources of aluminium (Al) in the red mud sample
were from boehmite, gibbsite, sodalite and kaolinite (Fig. 1), as
such the recovery of Al from red mud depended upon the dissolution of all the aforementioned phases which have varied reactivities. Fig. 2 shows the variation in aluminium recoveries under
different experimental conditions. The results showed that for
the most part, Al recovery did not exceed 45% which suggested
the dissolution of aluminium was incomplete. The bulk of the Al

that was dissolved in this study is proposed to be a result of the


dissolution of aluminosilicate phases since they are less stable relative to phases such as boehmite and gibbsite (Cama et al., 2002).
This phenomenon was observed in this study by investigation of
the red mud residue (see Supporting information). The major peaks
that were attributed to aluminosilicate materials disappeared after
treatment with the acids used in this study, while the peaks that
were associated with boehmite and gibbsite materials remained.
Similar results have also been noted in previous studies after washing red mud with mild acids (Liang et al., 2014). Albeit the percentage extractions for Al in this study were lower than those reported
previously (S
ayan and Bayramoglu, 2001), this study has made the
distinction of which mineral phases are likely to be responsible for
leaching of particular elements. This latter deduction allows a
degree of prediction regarding expected leaching behaviour based
upon the mineralogy of the red mud in question. Unfortunately,
due to the high amorphous content of the red mud used in this
study, an expected percentage recovery based upon dissolution
of all aluminosilicate materials is not feasible, but it is reasonable

Fig. 2. Variations in Al recovery under different experimental conditions; the acids used were hydrochloric, nitric, sulfuric, and phosphoric.

12

R.A. Pepper et al. / Minerals Engineering 99 (2016) 818

Fig. 3. Variations in Si recovery under different experimental conditions; the acids used were hydrochloric, nitric, sulfuric, and phosphoric.

Fig. 4. Variations in Fe recovery under different experimental conditions; the acids used were hydrochloric, nitric, sulfuric, and phosphoric.

R.A. Pepper et al. / Minerals Engineering 99 (2016) 818

13

Fig. 5. Variations in Ti recovery under different experimental conditions; the acids used were hydrochloric, nitric, sulfuric, and phosphoric.

to expect that bauxite residues with a higher proportion of aluminosilicate materials will leach more aluminium under conditions
similar to those used in this study than those with a higher proportion of boehmite or gibbsite. Incomplete characterisation of different red muds and varied experimental conditions that are used
across studies different studies lead to problems when attempting
to directly compare the recoveries of metals from different leaching studies. This highlights the value of a comprehensive study that
allows direct comparison between different elements for a larger
range of experimental variables and acid types.
The type of acid used for the extraction had some effect on the
recovery of Al. Phosphoric acid performed better than other acids
at higher acid concentrations. When the concentration of acid
was 0.1 M and below, recoveries for the four acids were relatively
similar, which may be due to the initial leaching of only the most
easily digestible Al phases at lower acid concentrations (Liang
et al., 2014). The highest recoveries were for phosphoric acid followed by sulfuric, then hydrochloric and nitric acids which exhibited similar recovery patterns. The similar recovery profiles for
nitric and hydrochloric acids across the four variables tested may
indicate that dissolution of Al in these systems is driven by the
interaction of the Al-bearing materials with the protons released
by these acids. This hypothesis is also supported by the greater
recoveries of Al that were obtained using higher concentrations
of these acids, and also those that were obtained through use of
sulfuric acid. Being a diprotic acid, sulfuric acid donated more protons to the solution at the same concentration relative to both
nitric and hydrochloric acids. Phosphoric acid however would not
be fully dissociated under the conditions used in this study, since
the pKa value for the first dissociation event of phosphoric acid
is 2.14 (Stone et al., 2016). Despite the incomplete dissociation of
the phosphoric acid, the pH measurements of the leaching experiments were still low enough for Al leaching to occur from aluminosilicate phases such as sodalite and kaolinite (Yang and Steefel,

2008) (see Supporting information). Therefore, although the pH


of the phosphoric acid series was sufficient for aluminosilicate dissolution, the higher recoveries obtained using phosphoric acid cannot be attributed solely to the amount of protons present in the
reaction. This suggested that other factors were attributing to the
higher observed recoveries with phosphoric acid, most likely interactions of the released. Al3+ ions with other species in solution.
The plot of Al extraction with time (Fig. 2D) shows the release of
Al from red mud to be a rapid process, with minimal increase in
additional Al extraction over the time period studied. This indicates the system was effectively at equilibrium with respect to Al
leaching for the data presented in Fig. 2. Therefore, interpretation
of this data in the context of reaction rate effects such as diffusion
of reaction products from the aluminosilicate surface (Hulbert and
Huff, 1970) is not appropriate. The stability of the aluminosilicate
dissolution products is a better explanation for the improved performance of some acids over others. Chloride, nitrate and sulfate
salts of aluminium are well within their solubility limits under
the conditions studied (Haynes, 2015), and have been shown to
form quite stable complexes within solution. In the case of the
reactions that utilised phosphoric acid, the extracted Al is proposed
to form a soluble complex in solution with dihydrogen phosphate
(Stone et al., 2016), rather than precipitating out of solution as
insoluble aluminium phosphate. This behaviour is a result of the
phosphoric acid being only partially dissociated under the conditions studied, since the pKa value for the full dissociation of phosphoric acid being 12.3 (Stone et al., 2016).
As stated above, the bulk of Al dissolution was achieved in a
rapid amount of time, which was a further reflection of the instability of aluminosilicate materials in acidic solutions. The most
effective acid for the dissolution of Al appears to be phosphoric
acid (Fig. 2C). This increased recovery may be due to the solution
chemistry and complex formation of Al in the phosphoric acid system as discussed above, or may be a reflection of the higher Al

14

R.A. Pepper et al. / Minerals Engineering 99 (2016) 818

recovery with phosphoric acid compared to the other acids used in


this study when the concentration of these acids was 1 M (Fig. 2A).
The higher recoveries obtained with phosphoric acid at this concentration were a result of gibbsite dissolution, which was not seen
for the other acids (see Supporting information). The change in Al
dissolution with temperature (Fig. 2C) showed a gradual rise in Al
dissolution as the temperature of the reaction increased. The
recoveries observed at 80 C were higher than those observed at
lower temperatures of the same concentration of acid (1 M), which
suggested that the extraction of Al from red mud is more favoured
as the temperature increased. This trend can be explained by
enhanced dissolution of aluminosilicate phases at higher temperatures; since dissolution of these materials has been shown to
increase with rising reaction temperatures (Carroll and Walther,
1990). The higher temperatures may have also had the effect of
allowing partial dissolution of some of the more stable aluminium
oxide phases that were present (gibbsite and boehmite), since
these phases are also dissolved more readily under higher reaction
temperatures (Castet et al., 1993).
The final variable that was tested in relation to red mud leaching was to examine the effect of solid:solution ratio. Al leaching
from using 1 M acid (Fig. 2B) showed that leaching was more efficient when lower amounts of red mud were used. The amount of Al
that was extracted declined as the mass of red mud in the reactions
increased, before reaching a fairly constant amount, except in the
case of the HNO3 extraction, where a lower recovery for the lowest
ratio was observed. The reason for this deviation is unknown, but
could potentially have been a result of solution interactions or
reaction product instabilities in these reactions.
The higher extractions at lower RM:acid ratios are proposed to
be due to an under-saturation of red mud in the lower ratio reactions. The term under-saturated refers to a system which has a
large amount of acid compared to the substance that is being leached (Cornell and Schwertmann, 2006) as is the case for the lower
ratio experiments. This use of lower amounts of red mud in a set
volume of acid effectively resulted in an increased amount of acid
that was available to react with Al-bearing phases in the red mud,
despite the reactions being done at the same concentration. The
higher recoveries observed for some of the lower ratio reactions
may also have been influenced by slightly better dispersal and
mixing at the lower ratios. A trend of decreasing recovery with
increasing red mud:acid did not continue across the whole data
set, which indicates that the aforementioned effects of undersaturation and mixing are not as suitable for controlling Al extraction from red mud as some of the other parameters that were
tested.
The results from the leaching experiments with regards to aluminium support the preferential leaching of aluminosilicate phases
over more stable Al-oxide phases such as boehmite or gibbsite.
This preferential dissolution of aluminosilicate phases could be
exploited as a means to selectively remove aluminium from red
mud or to enrich other elements whose phases were less reactive
under the same conditions. Due to the differences in composition
between different red muds, the percentage extractions reported
in this work may not extend to red muds with different origins.
However, the trends in extraction of the elements both individually and relative to each other provide valuable information in
regards to extracting metal values of interest or designing more
tailored leaching experiments.

The recoveries of silicon from red mud are proposed to be due


to the dissolution of sodalite and kaolinite since quartz does not
react with the acids used in this study (Knauss and Wolery,
1988). The dissolution of aluminosilicates to yield soluble silicon
is also supported by the high recoveries after short reaction times
(4145% after 1 h) and with low acid concentrations (4146% with
0.1 M acid); showing similar recovery patterns to those of Al.
A notable difference to Al however, was the sharp decline in Si
recovery as the concentration of acid exceeded 1 M. Since a similar
trend was not observed for Al, it was unlikely that this decline was
due to reduced aluminosilicate dissolution. An alternative explanation was that low recoveries of silicon could be explained by the
removal of extracted Si from the leachate solution. This instability
in leached Si may be a result of salting out effects (precipitation of
Si due to high ionic strength in the extract solution) or adsorption
onto the red mud residue (Milne et al., 2014). Adsorption of silica
onto the surface of red mud was not as likely due to unfavourable
interactions between the red mud surface and silica in solution.
The point of zero charge (PZC) for red mud has been quoted as
6.9 (Grfe et al., 2011) (although slight variations are expected
for different red mud compositions), and so the red mud surface
would have a net positive charge in the 5 M acid trials owing to
the protonation of surface hydroxyl groups. In acidic solutions, soluble silica is uncharged (Milne et al., 2014), and requires free OH
groups for adsorption to take place. Therefore, lower amounts of
silicon in solution at high acid concentrations may be due to a precipitation event that was influenced by the concentrations and
types of other ions in solution. Silica solubility has been shown
to rely heavily on the types and concentration of ions that are present in solution (Milne et al., 2014), with species such as Fe3+, Al3+,
and SO2
4 implicated in the formation of silica scale in industrial
systems (Amjad, 2013). Species such as these may have influenced
the stability of silicon for the 5 M tests since the recoveries of both
Al and Fe were enhanced under these conditions, especially when
hydrochloric, phosphoric and sulfuric acids were used.
Variations in recoveries for silicon with different acids did not
follow the same trends as for Al. For aluminium, recovery was
influenced to some degree by the amount of H+ available in the
reaction system, with di-protic acids performing better than
mono-protic acids under the same reaction conditions. For silicon,
this pattern did not hold although phosphoric acid out-performed
the other acids in the majority of tests that were performed. This
behaviour is proposed to be a result of the presence of phosphate
species which has a stabilising effect on soluble silicate (Milne
et al., 2014), that reduces the amount of silicon that would usually
precipitate. Sulfuric acid gave consistently lower recoveries due to
the higher amount of sulfate in these systems which enhanced the
precipitation of silicon. Thus, the differences between Si and Al
recoveries are proposed to be a result of different stabilities of Al
and Si in solution, and not from differences in leaching mechanisms. Since Si and Al were primarily leached from aluminosilicate
phases (Eqs. (1) and (2)), the leaching mechanism is the same for
both elements. The higher levels of Al that remain at higher acid
concentrations were a result of the Al3+ ion being stable in acidic
solutions.

Al2 Si2 O5 OH4 s 6H aq $ 2Al aq 2H4 SiO4 aq


3

2H2 O

Na8 Al6 Si6 O24 Cl2 s 24H aq


3.2.2. Silicon
Silicon containing mineral phases that were detected by XRD
included quartz, sodalite and kaolinite. Fig. 3 shows the leaching
behaviours of Si from red mud under the same conditions given
for Al in Fig. 2.

$ 6Al aq 6H4 SiO4 aq 8Na aq 2Cl aq


3

The effect of reaction temperature on Si extraction was not as


clear as for aluminium. The plotted results (Fig. 3C) showed a fluctuation in Si recovery, although the recoveries at 80 C were higher

R.A. Pepper et al. / Minerals Engineering 99 (2016) 818

for all acids tested than those at 40 C. Despite this, the large fluctuations in the recoveries observed at different temperatures
means that temperature may not be a reliable measure for controlling Si recovery from red mud. The absence of a clear trend in Si
recovery with increasing temperature as was seen for Al in Fig. 2C
suggested that the observed recoveries were due to the stabilities
of Si complexes in solution, rather than a fluctuation in aluminosilicate dissolution. As stated above, silica solubility is highly dependent on other species that are also present in solution, and so the
observed fluctuation may be a result of slightly unfavourable solution conditions at different temperatures. For example, the concentrations of extracted elements varied quite widely across the
temperature tests (see Supporting information), although specific
conclusions about the exact complexes that were forming in these
systems is difficult owing to their complexity. Study into the specific elemental speciation and complex formation in these systems is
certainly a worthwhile endeavour, but is deemed to be outside the
scope of this work, which was primarily to offer a direct comparison in leaching trends and to study the leaching behaviour of understudied elements such as silicon.
The trends for Si extraction with changes in time and with RM:
acid ratio (Fig. 3D and B, respectively) closely resemble those of
aluminium, which supports the notion that these elements are leached predominantly from aluminosilicate materials in the red
mud. The slight increase in Si recovery at the 24 h mark before a
lower plateau is attributed to the stability of the soluble Si species
in the reactions over time. The plateau in Si recovery likely indicates that equilibrium has been reached in the system, and it
appeared that this equilibrium state corresponded with a lower
recovery of Si. Therefore, if Si is an element of interest for extraction from red mud, shorter reaction times may be of benefit.
The leaching trends observed with respect to silicon show that
Si is leached exclusively from aluminosilicate phases, while Sicontaining minerals such as quartz remain in the red mud. The
instability of Si in more acidic extracts may provide a pathway
for either eliminating or recovering silicon from extracts of red
mud, and could be exploited to design targeted leaching or recovery methods.
3.2.3. Iron
The leaching behaviour of iron from red mud differed substantially from those of Al and Si. Rather than a sharp initial increase in
recovery with rising acid concentration or reaction time, Fe recovery was initially low and increased steadily with longer reaction
times, temperatures and acid concentrations (Fig. 4). These clear
differences to Si and Al behaviour were attributed in part to the relative reactivities of aluminosilicates and iron containing phases
within the red mud. Hematite accounted for most of the iron present in the red mud sample, which agreed with the observed
trends in Fe recovery. Hematite and most other iron oxides require
pH conditions of less than 1 for complete dissolution at temperatures as high as 70 C (Cornell and Schwertmann, 2006), so it
was expected to record lower levels of recovery for less concentrated acids and lower temperatures.
Increased acid concentration and time promoted the greatest
improvement in Fe recovery; however, the order of most efficient
acids differed between these two variables. For extractions using
5 M acids the order of the most effective acid was phosphoric,
hydrochloric, sulfuric and nitric, while for the 7 day extractions,
the order was sulfuric, phosphoric, hydrochloric and nitric. These
differences suggested that the dissolution of hematite from red
mud was influenced by the availability of hydrogen ions, the types
of counter-ions that were present in solution, and the stability of
the released iron complexes in solution. The dependence of iron
extraction from red on reaction with protons can be seen both in
the enhanced recoveries at higher acid concentrations but also in

15

the higher recoveries that were obtained when lower RM:acid


ratios were used. Reducing the amount of red mud in a reaction
effectively produced an under-saturated system with regards to
the red mud, which allowed more efficient extraction of iron to
take place. General reaction mechanisms for iron oxides such as
hematite suggest that dissolution occurs by a number of pathways,
one of which is the adsorption of protons to iron oxide surface sites
which results in promotion of Fe detachment from the bulk oxide
through weakening of the FeAO bonds (Cornell and Schwertmann,
2006). This particular mechanism is likely to have played a part in
the large increase in Fe recovery from red mud as the concentration of acid was increased (Fig. 4A), since the amount of protons
present in these reactions increased with rising acid concentration,
leading to more efficient adsorption of these protons onto iron
oxide phases and better dissolution.
The effect of acid counter-ions on the recovery of iron from red
mud was clearly seen from the comparison on the HNO3 and HCl
leaching efficiencies. If anions in solution did not have any effect,
recoveries using these two acids should have been similar since
the concentration of hydrogen ions would have been the same.
Instead, it can be seen in Fig. 4 that nitric acid consistently performed to a lower standard than HCl. This was proposed to be
due to interactions between the counter-ions of each acid and
the surface of hematite during dissolution. A previous study
(Majima et al., 1985) concerning the effects of different anions on
the dissolution rates of hematite has shown that ions such as chloride can enhance reaction rates over ions such as sulfate due to the
preferential adsorption of these ions at the reaction surface to
replace surface hydroxyl groups and assist with the release of Fe
from the bulk oxide (Cornell and Schwertmann, 2006). Despite
the results reported by Majima et al. (1985), sulfuric acid had better recoveries for all data series reported in Fig. 4 with the exception of the test that measured the effects of acid concentration. The
higher recoveries of iron when using sulfuric acid in the presented
data (Fig. 4BD) may be due in part to the higher recoveries
obtained with 1 M H2SO4 compared to those of HCl, HNO3, and
H3PO4, since the data presented in these figures used 1 M acids
to leach the red mud. Other considerations for the observed leaching behaviours are the nature and stability of the resulting soluble
species in the different acid solutions along with the formation of
stable complexes.
Speciation of the liberated Fe ions in the reaction solutions has
implications both for the stability of the reaction products and any
complex formation, as well as the mechanism by which Fe dissolution was taking place. The presence of Fe2+ ions during the dissolution of iron oxides also has the effect of enhancing dissolution
through acting as a catalyst for reductive dissolution (Cornell and
Schwertmann, 2006). The measured potentials of the extract solutions revealed that ferric ions were the predominant species in
solution for nitric and phosphoric acids, while the major soluble
iron species for hydrochloric and sulfuric acid extracts was found
to vary depending upon the concentration of acid that was used.
When the concentration of acid was less than 5 M, the potentials
of the H2SO4 extracts were below 0.771 V (refer to Supporting
information), which indicated that Fe2+ was likely to be the predominant species present (Takeno, 2005). These results may
explain the higher recoveries of Fe seen for sulfuric acid at concentrations below 5 M (Fig. 4A), since Fe2+ catalysed dissolution may
have been taking place in tandem with proton driven dissolution.
These results suggested that ferrous iron may be more stable in
the sulfuric acid digests than ferric iron, as seen from the relatively
lower recovery of iron in the 5 M H2SO4 digest compared to 5 M
HCl and 5 M H3PO4 (Fig. 4A).
The potentials of the HCl extracts at concentrations below 5 M
were found to be on either side of the critical value (0.771 V),
depending upon the conditions of the reaction; for example the

16

R.A. Pepper et al. / Minerals Engineering 99 (2016) 818

predominant species in the HCl extracts was generally Fe(II),


except when elevated reaction temperatures were used. This variation in the predominant Fe species for different acids may help to
explain some of the trends for Fe extracts that are seen in Fig. 4. In
the case of the effect of reaction temperature on the dissolution of
iron from red mud (Fig. 4C), extraction with all four acids was seen
to rise as the temperature of the reactions was increased. Sulfuric
acid was seen to give the highest recoveries, but recoveries
obtained using phosphoric acid became less than those obtained
using hydrochloric acid as the temperature increased. This promotion of iron recovery with hydrochloric acid at higher temperatures
corresponded with a shift in the major iron specie in the HCl
digests. At lower temperatures, iron was present in the ferrous
form, while the ferric form was favoured in the 70 and 80 C
digests (see Supporting information). The reason for this shift as
the temperature increased was potentially due to the parallel
release of other chemical species from red mud, such as cobalt,
which could act as oxidising agents for Fe in solution (Atkins and
de Paula, 2010). The higher recoveries obtained for the high temperature HCl digests when the Fe in solution was predominantly
in the ferric state highlights the complexity of these solutions
and the multiple variables that need to be considered when trying
to determine the processes that occurred in these solutions. The
higher recovery in these solutions is likely due to a number of variables; the first of these could be a reflection of the better performance of this acid under more extreme experimental conditions.
Hydrochloric acid is a popular reagent for the dissolution of iron
oxides, though the experimental conditions used to obtain complete or near-complete dissolution are often more extreme than
those employed in this study (Papadopoulos et al., 2010). Another
consideration is the formation of stable complexes in solution;
chloride has been shown to form stable complexes with ferric ions
in solution, which has the effect of ensuring the recovered iron
stays in solution, as well as potentially driving the reaction by
reducing the amount of free Fe ions in solution (Cornell and
Schwertmann, 2006). Clearly, the dissolution mechanisms of iron
from red mud and the resulting solution interactions of the liberated ions are quite complex; it appears that multiple processes
occur at once to result in the observed recoveries. In depth solution
studies to better understand these mechanisms and to determine
which complex species are forming may be worthwhile, but are
outside the scope of this work.
The dissolution of iron from red mud with respect to time
(Fig. 4D) showed an increase in iron recovery over the entire time
period studied, which suggested that the reaction was still continuing after 7 days. This behaviour highlights that the release of iron
from red mud by mineral acids is a much slower process than elements such as aluminium and silicon, which showed a rapid recovery profile due to their origin in less stable phases within the red
mud. Sulfuric acid gave the best recoveries over the time period
studied despite evidence of precipitation of an iron-sulfate phase
in the XRD pattern for the leached residues (see Supporting information). The phase was identified as ferricopiapite, which has a
chemical formula of Fe3+
4.67(SO4)6(OH)2(H2O)20 Majzlan et al., 2006.
The formation of this phase is associated with the generation of protons (Majzlan et al., 2006), although a large change in pH was not
observed over the reaction period. Despite this, pH measurements
for sulfuric acid were lower than those for the other acids that were
used in this study. This lower pH in the sulfuric acid reactions could
help explain the enhanced extraction of iron over the time period
studied compared to the other three acids, since the extra protons
present within the system should lead to enhanced dissolution via
the proton adsorption mechanism for the dissolution of iron oxides
(Cornell and Schwertmann, 2006) Other processes that could be contributing to the much higher Fe recoveries with sulfuric acid could
again be Fe2+ assisted dissolution as discussed above, although this

was not observed in the HCl digests, which were also in the ferrous
region. It is possible that the reaction conditions for hydrochloric
acid were too mild to efficiently dissolve the Fe-containing phases
(Cornell and Schwertmann, 2006), and that the complexation behaviour of chloride was not sufficient to overcome these thermodynamic barriers.
The conditions that favoured Fe dissolution in this study have
the disadvantage of also favouring the dissolution of other elements; which means that selective extraction of iron from red
mud may not be feasible without pre-treatment of the red mud
or treatment of the extract. This could involve enrichment of iron
containing phases in the red mud by a mild digestion, followed
by an extraction step to recover the iron from the red mud in a soluble and reactive from. The pre-treatment step also has the advantage of removing some species which may interfere with
downstream recovery of iron, such as aluminium or silicon
(Cornell et al., 1987). The leaching results also suggest that the
solution chemistry and stability of the leached Fe is a complex
and multi-faceted phenomenon, which would benefit from indepth study. Further study of the leaching behaviour may also help
to design better leaching methods that are more selective for iron.
3.2.4. Titanium
Leaching patterns of titanium from red mud may be found in
Fig. 5. Of the four acids studied, sulfuric acid was found to be the
most effective when leaching Ti, while phosphoric acid gave consistently lower recoveries with the exception of the 5 M digestion
(Fig. 5). These results were in contrast with those observed for the
other elements, for which phosphoric acid had consistently higher
metal recoveries.
Despite improved recoveries for Ti when sulfuric acid was used,
titanium recovery did not exceed 24% for any of the experiments.
XRD revealed the presence of anatase (3.16%) and rutile (1.96%),
both with the formula TiO2. The XRD analysis brought the total
TiO2 content to 5.12%, which differed from the value given by
XRF (6.88%) by 1.76%. This was a difference of 25.6%, which corresponded with the recoveries of Ti; suggesting that the recovery of
Ti was from the dissolution of amorphous Ti-containing phases.
XRD data collected after leaching showed that amorphous Ti was
leached preferentially over rutile and anatase, which were identified in the leached residues.
Both rutile and anatase tend to be more stable than other Fe and
Al containing minerals within red mud, with some studies resorting to the use of hydrofluoric acid (HF) to fully solubilise the anatase phase (Hanaor and Sorrell, 2011), although rutile remained
unreacted even after HF treatment. It has been previously reported
that amorphous metal oxides have better solubilities compared to
their crystalline counterparts (Schwertmann, 1991), so it is
expected that they would dissolve prior to any crystalline phases.
However, dissolution from predominantly amorphous phases presents a problem in terms of understanding the leaching mechanisms that are occurring, due to the nature of these amorphous
phases being largely unknown. It was likely that there are a number of amorphous phases that contributed to the overall amount of
leached Ti. These could include ilmenite or amorphous titania, both
of which have been observed in red muds before.
Unfortunately, due to the uncertainty regarding the nature of
amorphous phases that were responsible for the release of Ti for
the red mud sample, discussion of the leaching results in the context of reaction mechanisms is not feasible. Therefore, the general
trends that were observed in the data and the stability of the liberated Ti species will form the basis for the interpretation of the
leaching data.
Although the identities of the phases that contributed to Ti
release from red mud were unclear, it can be seen that the dissolution of these phases was promoted by increasing the concentration

R.A. Pepper et al. / Minerals Engineering 99 (2016) 818

of acid used (Fig. 5A). Increased extraction efficiency of Ti with


more concentrated acids can be attributed to the higher concentration of protons in the reaction which facilitates the dissolution of
Ti-bearing materials (Agatzini-Leonardou et al., 2008). Pourbaix
diagrams for titanium suggest that leached Ti is most likely present
in the form of TiO2+ (Chen et al., 2005) which could form complexes in solution with anions contributed to solution from the
acids used, such as with sulfate or chloride (OSullivan and Tyree,
2007). Complexation of the liberated TiO2+ could potentially
enhance the dissolution of Ti-containing phases from red mud in
order to maintain equilibrium as free TiO2+ is removed from the
system (OSullivan and Tyree, 2007).
The effects of RM:acid ratio on the extraction of Ti (Fig. 5B)
showed an initial decrease in Ti recovery as the amount of red
mud was increased. This was proposed to be due to an undersaturation of the red mud compared to the amount of acid, in a
similar manner to the other elements addressed above and the
trends observed when the concentration of acid was increased.
The higher efficiency of sulfuric acid over others may be a reflection of the better performance of sulfuric acid at the chosen concentration (1 M), or may indicate that the released Ti is more
stable in the sulfuric acid solution, potentially as a result of the
aforementioned complex formation. The specific speciation of Ti
in red mud leaching solutions has not been discussed in detail in
the current literature. As mentioned above for iron, the leaching
solutions are likely to be complex which makes determining specific solution chemistry somewhat difficult without in depth further
study which is not within the scope of this study.
The variation in Ti recovery with temperature presented in
Fig. 5C showed a gradual increase over the tested range. The difference in the amount of titanium leached over the temperature
range was not as pronounced as for other elements examined in
this study, which suggested that other factors such as acid concentration or reaction time may have had more of an impact on Ti
leaching from amorphous phases in red mud. Trends in Ti leaching
with regards to reaction time (Fig. 5D) show a slow increase in the
amount of extracted Ti as reaction time progresses. Once again, use
of sulfuric acid gave higher recoveries than those observed for
other acids due to the better performance of this acid for the concentration at which the data is presented. The continued extraction
of Ti may also have been stimulated by the lower pH and corresponding higher concentration of protons in the sulfuric acid
reactions.
The recovery patterns of titanium were similar to those of iron,
and so the conditions that were shown to favour Ti recovery also
tended to favour the extraction of the other elements studied.
The fact that the leached Ti in this study came predominantly from
amorphous phases highlights the importance of red mud characterisation, including assessment of amorphous material. This is
not common practice in the literature, but provides additional
information that can be used to predict expected leaching behaviour of red mud in acid solutions.
4. Conclusions
Recovery of four major elements from red mud by means of acid
digestion was found to be significantly dependent on the experimental conditions used. Nitric acid was discovered to be a poor
choice for the recovery of elements present mainly as metal oxides
(Fe, Ti), but was acceptable for elements such as Al or Si, which
were present as a more soluble aluminosilicate material. An additional advantage of using HNO3 to extract Al and Si is that the transition metals (Fe and Ti) would remain within red mud, allowing
for later recovery with further digestions using a more effective
acid such as 5 M HCl, H2SO4, or H3PO4 at elevated temperatures.
Metal recovery was facilitated by higher acid concentrations, ele-

17

vated reaction temperatures, longer reaction times, and an excess


of acid relative to red mud. Conversely, Si recovery was low for
high concentrations of acid, due to the instability of soluble silicate
under such conditions.
The range of tests that have been performed provide a useful
guide as to the behaviours of various mineral phases within red
mud under different conditions; allowing abundance of elements
within red mud extracts relative to each other to be evaluated.
Additionally, since this work related the observed leaching behaviour of each element to their respective mineral phases, prediction of leaching behaviour and expected recovery based upon the
mineralogical composition of red mud was possible. This latter
insight was advantageous when dealing with red muds of vastly
different compositions, since expected extract compositions could
be estimated based upon the proportion and identity of mineral
phases present within the red mud sample. The experiments also
showed that phosphoric acid, which has been previously overlooked, was a viable candidate for red mud leaching. This study
provided a platform for designing more targeted extractions of
Fe, Al, Si and Ti.
Future experimentation should focus on refining experimental
parameters to enhance selectivity within extracts, as well as determining speciation of the extracted elements. Avenues to explore
could include a combination of different extraction methods in a
sequential manner or treatment of the extract by pH adjustment
to recover species as solids. It would also be beneficial to determine the extent to which the simultaneous dissolution of multiple
phases impacts the recovery of metals from red mud, which would
allow easier comparison between studies that may use red mud
from different sources.
Acknowledgements
The financial and infra-structure support of the Energy and Process Engineering Discipline of the Science and Engineering Faculty,
Queensland University of Technology, and the Central Analytical
Research Facility (CARF) is gratefully acknowledged. Queensland
Government is thanked for supporting acquisition of some of the
instrumentation used and funding Dr. Couperthwaite through an
Accelerate Fellowship grant (20142017). The technicians and laboratory staff of CARF have offered invaluable help during the
experiments.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.mineng.2016.09.
012.
References
Agatzini-Leonardou, S., Oustadakis, P., Tsakiridis, P.E., Markopoulos, C., 2008.
Titanium leaching from red mud by diluted sulfuric acid at atmospheric
pressure. J. Hazard. Mater. 157 (23), 579586.
Amjad, Z., 2013. Mineral Scales in Biological and Industrial Systems. CRC Press.
Atkins, P., de Paula, J., 2010. Atkins Physical Chemistry. OUP Oxford.
Borra, C.R., Pontikes, Y., Binnemans, K., Van Gerven, T., 2015. Leaching of rare earths
from bauxite residue (red mud). Miner. Eng. 76, 2027.
Cama, J., Metz, V., Ganor, J., 2002. The effect of pH and temperature on kaolinite
dissolution rate under acidic conditions. Geochim. Cosmochim. Acta 66 (22),
39133926.
Carroll, S.A., Walther, J.V., 1990. Kaolinite dissolution at 25, 60 and 80 C. Am. J. Sci
290 (7), 797810.
Castet, S., Dandurand, J.-L., Schott, J., Gout, R., 1993. Boehmite solubility and
aqueous aluminum speciation in hydrothermal solutions (90350 C):
experimental study and modeling. Geochim. Cosmochim. Acta 57 (20), 4869
4884.
Chen, C.-C., Chen, J.-H., Chao, C.-G., Say, W.C., 2005. Electrochemical characteristics
of surface of titanium formed by electrolytic polishing and anodizing. J. Mater.
Sci. 40 (15), 40534059.

18

R.A. Pepper et al. / Minerals Engineering 99 (2016) 818

Cornell, R.M., Schwertmann, U., 2006. The Iron Oxides: Structure, Properties,
Reactions, Occurrences and Uses. John Wiley & Sons.
Cornell, R., Giovanoli, R., Schindler, P., 1987. Effect of silicate species on the
transformation of ferrihydrite into goethite and hematite in alkaline media.
Clays Clay Miner. 35 (1), 2128.
Evans, K., Nordheim, E., Tsesmelis, K., 2012. Bauxite residue management. Light
Met., 6366 (Warrendale, PA: TMS)
Grfe, M., Power, G., Klauber, C., 2011. Bauxite residue issues: III. Alkalinity and
associated chemistry. Hydrometallurgy 108 (12), 6079.
Hanaor, D.A., Sorrell, C.C., 2011. Review of the anatase to rutile phase
transformation. J. Mater. Sci. 46 (4), 855874.
Haynes, W.M., 2015. CRC Handbook of Chemistry and Physics. CRC Press.
Hulbert, S., Huff, D., 1970. Kinetics of alumina removal from a calcined kaolin with
nitric, sulphuric and hydrochloric acids. Clay Miner 8 (3), 337.
Kasliwal, P., Sai, P.S.T., 1999. Enrichment of titanium dioxide in red mud: a kinetic
study. Hydrometallurgy 53 (1), 7387.
Kitajima, T., Kasai, E., 1999. Carbothermic reduction of bauxite residue. Shigen Sozai
115 (8), 611617.
Klauber, C., Grfe, M., Power, G., 2011. Bauxite residue issues: II. Options for residue
utilization. Hydrometallurgy 108 (12), 1132.
Knauss, K.G., Wolery, T.J., 1988. The dissolution kinetics of quartz as a function of
pH and time at 70 C. Geochim. Cosmochim. Acta 52 (1), 4353.
Li, Y., Wang, J., Wang, X., Wang, B., Luan, Z., 2011. Feasibility study of iron mineral
separation from red mud by high gradient superconducting magnetic
separation. Phys. C (Amsterdam, Neth.) 471 (34), 9196.
Liang, W., Couperthwaite, S.J., Kaur, G., Yan, C., Johnstone, D.W., Millar, G.J., 2014.
Effect of strong acids on red mud structural and fluoride adsorption properties.
J. Colloid Interface Sci. 423, 158165.
Liu, W., Yang, J., Xiao, B., 2009. Application of Bayer red mud for iron recovery and
building material production from alumosilicate residues. J. Hazard. Mater. 161
(1), 474478.
Liu, W., Sun, S., Zhang, L., Jahanshahi, S., Yang, J., 2012. Experimental and simulative
study on phase transformation in Bayer red mud soda-lime roasting system and
recovery of Al, Na and Fe. Miner. Eng. 39, 213218.
Majima, H., Awakura, Y., Mishima, T., 1985. The leaching of hematite in acid
solutions. Metall. Trans. B 16 (1), 2330.
Majzlan, J., Navrotsky, A., McCleskey, R.B., Alpers, C.N., 2006. Thermodynamic
properties and crystal structure refinement of ferricopiapite, coquimbite,
rhomboclase, and Fe2 (SO4)3 (H2O)5. Eur. J. Mineral. 18 (2), 175186.

Marcel Ct, G.C., Kristy, Wilson, 2012. Orbites Red Mud Remediation and Mineral
Recovery Process. Orbite, Editor.
Milne, N.A., OReilly, T., Sanciolo, P., Ostarcevic, E., Beighton, M., Taylor, K., Mullett,
M., Tarquin, A.J., Gray, S.R., 2014. Chemistry of silica scale mitigation for RO
desalination with particular reference to remote operations. Water Res. 65,
107133.
Mishra, B., Staley, A., Kirkpatrick, D., 2001. Recovery and utilization of iron from redmud. Light Met. (Warrendale, PA, US), 149156.
OSullivan, D.W., Tyree, M., 2007. The kinetics of complex formation between Ti (IV)
and hydrogen peroxide. Int. J. Chem. Kinet. 39 (8), 457461.
Papadopoulos, N., Karayianni, H.-S., Hristoforou, E., Metaxa, E., 2010. Valuable
Products Obtained from Red Mud, Greece, pp. 21.
Piga, L., Pochetti, F., Stoppa, L., 1995. Application of thermal analysis techniques to a
sample of Red Mud a by-product of the Bayer process for magnetic
separation. Thermochim. Acta 254, 337345.
Power, G., Grfe, M., Klauber, C., 2011. Bauxite residue issues: I. Current
management, disposal and storage practices. Hydrometallurgy 108 (12),
3345.
Qu, Y., Lian, B., Mo, B., Liu, C., 2013. Bioleaching of heavy metals from red mud using
Aspergillus niger. Hydrometallurgy 136, 7177.
Santini, T.C., 2015. Application of the Rietveld refinement method for quantification
of mineral concentrations in bauxite residues (alumina refining tailings). Int. J.
Miner. Process. 139, 110.
S
ayan, E., Bayramoglu, M., 2000. Statistical modeling of sulfuric acid leaching of
TiO2 from red mud. Hydrometallurgy 57 (2), 181186.
S
ayan, E., Bayramoglu, M., 2001. Statistical modelling of sulphuric acid leaching of
TiO2, Fe2O3 and A12O3 from red mud. Process Saf. Environ. Prot. 79 (5), 291296.
Schwertmann, U., 1991. Solubility and Dissolution of Iron Oxides, in Iron Nutrition
and Interactions in Plants. Springer, pp. 327.
Stone, K., Bandara, A.M.T.S., Senanayake, G., Jayasekera, S., 2016. Processing of rare
earth phosphate concentrates: a comparative study of pre-leaching with
perchloric, hydrochloric, nitric and phosphoric acids and deportment of
minor/major elements. Hydrometallurgy 163, 137147.
Takeno, N., 2005. Atlas of Eh-pH diagrams. Geol. Surv. Jpn. Open File Rep. 419, 102.
Uzun, D., Gulfen, M., 2007. Dissolution kinetics of iron and aluminum from red mud
in sulfuric acid solution. Indian J. Chem. Technol. 14 (3), 263268.
Yang, L., Steefel, C.I., 2008. Kaolinite dissolution and precipitation kinetics at 22 C
and pH 4. Geochim. Cosmochim. Acta 72 (1), 99116.