CHARACTERISATION AND REUSE OF SOLID WASTES

By

Wentao Liang

Principal Supervisor: A/Prof. Cheng Yan

Associate Supervisor: Dr. Sara Couperthwaite

Submitted in fulfilment of the requirements for the degree of

[Master of Engineering (Research)]

Faculty of Science and Technology

Queensland University of Technology

2013

© 2013 Wentao Liang 1

Keywords

Acid stability, adsorption, bauxite residue, fluoride, red mud

© 2013 Wentao Liang 2

Abstract

The removal of fluoride using red mud has been improved by acidifying red mud

with hydrochloric, nitric and sulphuric acid, while affected by temperature. This

investigation shows that the removal of fluoride using red mud is significantly

improved if red mud is initially acidified, and thermally activated. The acidification

of red mud causes sodalite and cancrinite phases to dissociate, confirmed by the

release of sodium and aluminium into solution as well as the disappearance of

sodalite bands and peaks in infrared and X-ray diffraction data. The dissolution of

these mineral phases increases the amount of available iron and aluminium

oxide/hydroxide sites that are accessible for the adsorption of fluoride. The removal

of fluoride is dependent on the charge of iron and aluminium oxide/hydroxides (Fe2+

or Fe3+

/AlOH2+

or Al (OH) 2-

) on the surface of red mud. Acidifying red mud with

hydrochloric, nitric and sulphuric acid resulted in surface sites of the form ≡ SOH2+

and ≡ SOH. Optimum removal is obtained when the majority of surface sites are in

the ≡ SOH2+

as the substitution of a fluoride ion doesn’t cause a significant increase

in pH. Meanwhile, the minerals of red mud are changed through the thermally

treatment. This investigation shows the importance of having a low and consistent

pH, and the appropriate temperature for the removal of fluoride from aqueous

solutions using red mud.

© 2013 Wentao Liang 3

Table of Contents

Keywords ..................................................................................................................... 1

Abstract ........................................................................................................................ 2

Table of Contents ......................................................................................................... 3

List of Figure ................................................................................................................ 5

List of Tables................................................................................................................ 7

List of Publications ...................................................................................................... 8

Statement of Original Authorship ................................................................................ 9

The Proposed Supervisors and Their Credentials ...................................................... 10

Acknowledgments ...................................................................................................... 11

CHAPTER 1: INTRODUCTION .......................................................................... 12

Thesis outline .................................................................................................... 16 1.1

Significance ...................................................................................................... 18 1.2

CHAPTER 2: LITERATURE REVIEW .............................................................. 19

Bayer process review ........................................................................................ 19 2.1

2.1.1 Bauxite ore .............................................................................................. 19

2.1.2 Bayer process .......................................................................................... 22

2.1.3 Limitations of the Bayer process ............................................................. 25

Red Mud applications ....................................................................................... 26 2.2

2.2.1 Red mud issues ........................................................................................ 26

2.2.2 Red mud applications in materials .......................................................... 27

2.2.3 Red mud applications in water treatment ................................................ 29

Red mud treatment comparision ....................................................................... 33 2.3

2.3.1 Acid activated red mud ........................................................................... 33

2.3.2 Thermally activated red mud................................................................... 34

2.3.3 Combination treatment ............................................................................ 35

Fluoride applications ........................................................................................ 36 2.4

2.4.1 Fluoride issues ......................................................................................... 36

2.4.2 Red mud applications with fluoride ........................................................ 36

CHAPTER 3: EXPERIMENTAL AND CHARACTERISATION ..................... 39

Experimental techniques ................................................................................... 39 3.1

3.1.1 Fluoride adsorption ................................................................................. 39

3.1.2 Characterisation techniques..................................................................... 39

Characterisation techniques .............................................................................. 41 3.2

3.2.1 X-ray Diffraction (XRD) ......................................................................... 41

3.2.2 Infrared Spectroscopy (IR) ...................................................................... 43

3.2.3 Surface Area and Pore Size Analysis ...................................................... 44

CHAPTER 4: ACID ACTIVATED PROCESS ................................................... 51

© 2013 Wentao Liang 4

Experimental procedure .................................................................................... 51 4.1

4.1.1 Acid activated red mud ........................................................................... 51

Results and disscussion .................................................................................... 52 4.2

4.2.1 Elemental and mineralogical composition .............................................. 52

4.2.2 Leachants (ICP and reactions)................................................................. 55

4.2.3 Infrared spectroscopy .............................................................................. 58

4.2.4 Removal of fluoride using acid treated red mud ..................................... 64

CHAPTER 5: THERMALLY ACTIVATED RED MUD .................................... 68

Experimental procedures .................................................................................. 68 5.1

5.1.1 Thermally activated red mud................................................................... 68

Results and disscussion .................................................................................... 68 5.2

5.2.1 Red mud characterisation ........................................................................ 68

5.2.2 Thermally analysis .................................................................................. 73

5.2.3 Removal of fluoride using thermally activated red mud ......................... 77

CHAPTER 6: OVERALL CONCLUSIONS ........................................................ 83

CHAPTER 7 FUTURE WORK .............................................................................. 84

BIBLIOGRAPHY .................................................................................................... 85

© 2013 Wentao Liang 5

List of Figure

Figure 1: World bauxite reserves

Figure 2: Schematic representation of The Bayer process

Figure 3: Composition of red mud in different parts of the world

Figure 4: X-ray diffraction between planes of atoms

Figure 5: Classification of NDDT adsorption isotherm

Figure 6: Mineralogical composition of red mud washed with DI water

Figure 7: XRD patterns of washed and acid treated red muds (0.5 M)

Figure 8: SEM image of red mud reacted with concentrated HCl

Figure 9: ICP-MS of acid treated red mud filtrate

Figure 10: Infrared spectrum of washed red mud

Figure 11: Infrared spectrum of red mud treated with various concentrations of HCl

Figure 12: Infrared spectrum of red mud treated with various concentrations of

HNO3

Figure 13: Infrared spectrum of red mud treated with various concentrations of

H2SO4

Figure 14: Removal of fluoride (%) using different masses of acid treated red mud

and the associated changes in pH from the original 100ppm fluoride solution with pH

4.75

Figure 15: Removal of fluoride (%) using different masses of acid treated red mud

and the associated changes in pH from the original 100ppm fluoride solution with pH

7.99

Figure 16: X-ray diffraction pattern of red mud and corresponding reference patterns.

Figure 17: Infrared spectrum of red mud

Figure 18: DTG curve of red mud

© 2013 Wentao Liang 6

Figure 19: X-ray diffraction patterns of thermally activated red mud

Figure 20: Infrared spectra of thermally activated red mud

Figure 21: Fluoride adsorption (mmol/g) using red mud thermally activated to

1000°C for a 100mg/L fluoride solution with an initial pH of 4

Figure 22: Fluoride adsorption (mmol/g) using red mud thermally activated to

1000°C for a 100mg/L fluoride solution with an initial pH of 8

Figure 23: Fluoride adsorption (mmol/g) using red mud thermally activated to

1000°C for a 100mg/L fluoride solution with an initial pH of 4

Figure 24: Fluoride adsorption (mmol/g) using red mud thermally activated to

1000°C for a 100mg/L fluoride solution with an initial pH of 8

© 2013 Wentao Liang 7

List of Tables

Table 1: World bauxite mine production, reserves and reserve base

Table 2: Major mineralogical compositions of bauxites

Table 3: Major minerals composition of bauxite

Table 4: Different conditions for removing phosphorous by red mud

Table 5: d- values of iron oxides

Table 6: Types of isotherms, characteristics and examples

Table 7: Major elemental composition represented as (Al,Na,Si,Ti):Fe

Table 8: Oxide composition of red mud

Table 9: Physical properties of red mud, red mud heated to 500 and 1000°C

Table 10: ICP-OES of red mud after being immersed in water

© 2013 Wentao Liang 8

List of Publications

1. Wentao Liang, Sara J. Couperthwaite, Gurkiran Kaur, Cheng Yan, Dean W.

Johnstone and Graeme J. Millar, Effect of strong acids on red mud structural and

adsorption properties. Submitted to the Journal of Colloid and Interface Science

on the 13th September, 2013.

2. Wentao Liang, Sara J. Couperthwaite, Dean W. Johnstone, Cheng Yan, Gurkiran

Kaur and Graeme J. Millar, Adsorption properties of thermally activated red mud

for fluoride. Submitted to the Journal of Thermal Analysis and Calorimetry on the

11th October, 2013.

QUT Verified Signature

© 2013 Wentao Liang 10

The Proposed Supervisors and Their Credentials

Principle Supervisor: Associate Prof. Cheng Yan

Dr. Cheng Yan is an associate professor in the School of Chemistry, Physics

and Mechanical Engineering at QUT. He is an established expert in synthesis and

characterization of various materials. His current research interests include

nanomaterial for renewable energy, structural and functional Nano composites,

biomaterials and biomechanics, Nano mechanics and modelling at micro and Nano

scales. He has been awarded 2 competitive ARC fellowships, 10 ARC projects and

granted over $5 m research funds. He has generated more than 200 publications and

co-chaired several international conferences.

Associate Supervisor: Dr. Sara Couperthwaite

Dr. Sara Couperthwaite is a Vice-Chancellor Research Fellow in the School of

Chemistry, Physics and Mechanical Engineering at QUT. She established a

reputation in the recycling of industrial waste residues for water purification

techniques, with a focus on the alumina industry and coal seam gas water. Her

current research interests include the optimisation of a layered hydroxide compound

that is generated from alumina waste and seawater for the removal of heavy metal in

other mineral processing industries, and the removal of selected cations and anions in

coal seam gas water using commercially available ion exchange media. She has been

awarded an ARC DECRA grant and generated close to $0.5 million in commercial

research projects. She has generated more than 130 publications, presented at a

number of alumina refinery focused conferences and has numerous industrial

contacts in the alumina and coal seam gas industries.

© 2013 Wentao Liang 11

Acknowledgments

The financial and infra-structure support of the Science and Engineering Faculty and

the Institute of Future Environments, Queensland University of Technology is

gratefully acknowledged.

I would like to thank the following people and organisations without who this thesis

could not have been completed.

They include:

(i) My QUT supervisors: Associate Professor Cheng Yan and Dr Sara

Couperthwaite for providing a challenging research project and assistance

in the editing.

(ii) The entire Couperthwaite and Millar group, with a special thanks to

Mitch De Bruyn and Kenneth Nuttall for training and assistance in

numerous instrumental techniques.

(iii) The postgraduate community and staff of the Science and Engineering

Faculty who provided much needed support.

(iv) Mr. Anthony Raftery for advice and technical support with the XRD

instruments and preparation methods.

(v) Dr. Llew Rintoul for his assistance with the vibrational spectroscopy

instruments.

Finally, I would like to give a special thanks to my family and friends for their love,

support and encouragement, especially my parents.

© 2013 Wentao Liang 12

1CHAPTER 1: INTRODUCTION

At the end of 2010, around 3 billion tonnes of bauxite refinery residue (red mud) had

been produced globally using the Bayer process to convert aluminium oxides and

hydroxides in bauxite ore to alumina [1]. It is estimated that at least an additional

120 million tonnes is produced each year [1]. The magnitude of waste being

generated by this industry clearly demonstrates the need for future developments in

technologies that utilised this raw material. Due to the chemical complexity and

classification (hazardous material under the Basel Convention) [2] of bauxite residue,

numerous researchers are trying to utilise the waste residue [3-14].

Red mud (generally slurry) is comprised of iron oxides (20%-65%), titanium oxides

(2%-8%), silicon oxides (10%-15%) and undissolved alumina, along with a wide

range of other oxides depending on the country of origin [15, 16]. Trace levels of

metal oxides, such as arsenic, cadmium, chromium, copper, gallium, lead, mercury,

nickel and in some cases thorium and uranium, are of particular concern. Apart from

heavy metal contamination the alkalinity of red mud also possess restrictions on

viable applications due to the cost of neutralisation. Alkalinity in the residue exists in

both solid and solution as: 1) entrained liquor (sodium hydroxide, sodium aluminate

and sodium carbonate), 2) calcium compounds, such as hydrocalumite, tri-calcium

aluminate and lime, and 3) sodalite ((NaAlSiO4)6(Na2X), where X can be SO4

2-,

CO32-

, Al(OH)4- or Cl

-) [17].

The potential environmental implications of seepage, dam failures and flooding can

have a large negative impact on the surrounding water bodies, including groundwater,

lakes and rivers, when soluble caustic chemicals are released [18]. Mining industries

© 2013 Wentao Liang 13

take huge precautionary measures to minimise environmental risks (lining dams and

below capacity limits), however during natural disasters such as flooding there are no

measures that can be taken to prevent spillage. Apart from the potential risks of

tailing dams, there also exists the problem that large areas of land are being

transformed into landfill to contain the red mud residue [18]. The discovery of a

viable application (in-expensive and uses large quantities) for the reuse of red mud

will significantly minimise environmental impacts caused by tailings dams and the

associated costs of storage facilities (more than $80 million a year) [19].

In recent years, many researchers have focused on utilising red mud as an adsorbent

material, and have had success in the adsorption of heavy metals [20-22], arsenate

[23-28], phosphates [29] and to a lesser extent fluoride [30, 31]. Fluoride is naturally

found in groundwater due to the dissolution of fluoride bearing minerals over long

periods of time [32]. However, elevated levels of fluoride in groundwater can

generally be traced back to a number of industries, including but not limited to, glass

and ceramic production, electroplating, coal fired power stations, brick and iron

works, and aluminium smelters [33]. It is estimated that more than 200 million

people rely on contaminated drinking water containing more than 1.5mg/L of

fluoride (World Health Organisation safe level) [34]. Continual and excess exposure

to fluoride results in diseases such as osteoporosis, arthritis, brittle bones, and cancer,

infertility, brain damage, Alzheimer and thyroid disorders in humans [35].

Traditionally, contaminate fluoride drinking waters have been treated using lime,

which results in the precipitation of fluorite. However, other precipitating and

coagulating reagents include iron (III), alum, calcium and activated alumina [35].

© 2013 Wentao Liang 14

More involved processes, such as ion exchange, reverse osmosis and electro dialysis,

have also been explored, but have subsequent waste disposal issues and high

operating and maintenance costs [35, 36].

Of particular interest, is the removal of fluoride using activated alumina and iron-

based materials, schwertmannite (Fe8O8(OH)6(SO4)•nH2O), granular ferric hydroxide

(Fe(OH)3), and goethite(α-FeOOH) [35, 36], as they are found in red mud and thus

advocates red mud as an adsorbent. Fluoride adsorption using iron-based adsorbents

are facilitated by exchange reactions involving F- and OH

- with FeOH surface groups.

A maximum Langmuir adsorption capacity of 7.0 mg/g of fluoride on granular ferric

hydroxide has been reported by Kumar et al. [37] in the pH range 6.0 to7.0.

Phosphate and sulphate (inner-sphere forming species) have been shown to have

negative effects on the loading capacity of fluoride, while outer-sphere forming

species (chloride and nitrate) improved fluoride removal slightly [35].

For activated alumina (amorphous gibbsite (Al(OH)3) or alumina (Al2O3)), Farrah et

al. [37] observed a maximum loading capacity of 9.72mg/g (Langmuir) in the pH

range 5.5 to 6.5. At lower pH, fluoride removal decreased due to the formation of

AlFx soluble species, while at higher pH, OH- displaced F

-. A study by Cengeloglu et

al.[30] investigated the adsorption capacity of untreated and hydrochloric treated red

mud. The maximum removal capacity was obtained using the acid treated red mud

and observed a Langmuir loading capacity of 3.574 mg/g at pH 5.5 after 2 hours.

Granular red mud prepared by Tor et al.,[37] obtained a Langmuir loading capacity

of 0.644 mg/g at pH 4.7 after 6 hours during a batch trial. However, this later study

showed an increased loading capacity for column studies (0.773-1.274 mg/g), which

© 2013 Wentao Liang 15

successfully regenerated using 0.2M NaOH. Previous studies have also shown that

the spent red mud satisfies the toxicity characteristic leaching procedure (TCLP)

used to classify inert wastes [37].

This investigation will assess the structural and resulting adsorption properties of

Australian red mud untreated and activated by acid and thermal methods. Particular

emphasis will be placed on determining the reactions involved during the activation

processes and how the remaining mineralogical composition correlates with the

fluoride loading capacities.

© 2013 Wentao Liang 16

THESIS OUTLINE 1.1

Red mud is a complex material that requires an understanding of different fields of

science, including physical chemistry, chemical engineering and material science, in

order to obtain an appreciation of the hurdles that need to be overcome for its reuse

as an adsorbent material. This thesis includes a chapter on relevant literature,

characterisation of red mud and subsequent changes in structural and adsorbent

properties for acid and thermally activated red mud.

I. Chapter 2: Literature Review

Chapter 2 presents a review of current literature on the development of red mud as an

adsorbent material. It will be divided into two main parts: 1) red mud and current

applications and 2) potential hazard of fluoride. A large portion of the review will

cover the most recent research on red mud as an adsorbent for heavy metals [20-22],

arsenate [23-28], phosphates [29] and fluoride [30, 31].

II. Chapter 3 : Experimental and Characterisation

Chapter 3 highlights the experimental procedures used throughout this investigation

and provides theoretical background on some of the physical characterisation

techniques, such as X-ray diffraction, infrared spectroscopy and surface analysis.

III. Chapter 4: Acid Activated Red Mud

Chapter 4 describes the process of acid activation. The main purpose of the acid

activation process is to remove species that do not participate in the adsorption

process as well as to reduce the influence of carbonate (CO32-

) [38]. All red muds

have been fully characterised and include untreated red mud and acid activated red

© 2013 Wentao Liang 17

mud using hydrochloric, nitric and sulphuric acid. Characterisation techniques

include infrared spectroscopy (IR), scanning electron microscopy (SEM) coupled

with energy dispersive X-ray spectroscopy (EDX) and X-ray diffraction (XRD). A

comparison of the structural changes and adsorption properties of the three acids

have been made.

IV. Chapter 5: Thermally Activated Red Mud

Chapter 5 presents the effects of temperature on the structural and adsorption

properties of red mud. Calcining red mud is used to reduce a number of the

mineralogical phases (aluminium oxy/hydroxides and sodalite) in order to increase

the surface area of hematite available for the adsorption of fluoride and therefore the

removal of sodalite and carbonate could be decomposed in a proper temperature

range. The composition of thermally treated red mud has been fully characterised,

and as a result structural and physical properties change. The impact on fluoride

removal varies significantly with the compositional changes that occur.

V. Chapter 6: Conclusions

This investigation has shown that the mineralogical phase involved in the adsorption

of fluoride is hematite. The activation of red mud through acid and thermal methods

has improved the adsorption capacity of red mud.

© 2013 Wentao Liang 18

SIGNIFICANCE 1.2

Due to the substantial amounts of red mud stored in tailing dams and their potential to

fail or release red mud waste during natural disasters, this investigation looks at the

potential of red mud as an adsorbent material for fluoride. The development of a suitable

application of red mud will not only minimise the potential environmental risks that

tailing dams pose but it will also reduce costs associated with the construction and

maintenance of these dams. In addition, the development of a material suitable for the

treatment of fluoride contaminated waters increase the significance of the work being

conducted. This thesis utilizes acid activated and thermally treated red mud to adsorb

fluoride, with an emphasis on understanding the molecular and structural changes that

occur during the treatment process.

© 2013 Wentao Liang 19

2CHAPTER 2: LITERATURE REVIEW

BAYER PROCESS REVIEW 2.1

2.1.1 Bauxite ore

Lateritic bauxites (silicate bauxites) are found mostly in the countries in the region of

the Earth surrounding the Equator (tropics). They are formed by lateralization

(tropical weathering) of various silicate rocks such as basalt, gneiss, granite, shale,

and syenite. It is a prolonged process of chemical weathering which produces a wide

range of ore compositions of varying thickness, grade, chemistry and mineralogy.

The formation of bauxites depends on intensive and long-lasting weathering

conditions of the underlying parent rock in a location with very good drainage.

Tropical weathering enables the dissolution of kaolinite (Al2Si2O5(OH)4) and the

precipitation of gibbsite (Al(OH)3). Zones with the highest aluminium content are

generally located in areas rich in iron and oxygen. Aluminium hydroxide in the

lateritic bauxite deposits is almost exclusively gibbsite in tropical regions, such as

Australia.

Bauxite contains aluminium in the form of three main minerals, 1) gibbsite

(aluminium trihydroxide), 2) boehmite (aluminium oxyhydroxides - γ-AlO(OH)),

and diaspore (α-AlO(OH)), with gibbsite and boehmite being the most dominate

aluminium compounds in all bauxite ores. Apart from aluminium, it also contains a

mixture of iron oxides (hematite - Fe2O3 and goethite - α-FeO(OH)), the clay mineral

kaolinite, and small amounts of anatase (TiO2). Currently, bauxite ores are the only

feasible resource for the production of alumina on a commercial scale; however

research into the extraction of alumina from clay is being undertaken.

© 2013 Wentao Liang 20

European and Chinese bauxite ores are composed almost entirely of the monohydrate

forms boehmite and diaspore, except for Greece which contains predominately

diaspore containing ores. Compared with Europe, the ores from the rest of the world

are composed primarily of gibbsite, with small amounts of boehmite and minor

amounts of diaspora [39].

In Australia, aluminium ores are predominantly composed of gibbsite, which are

consequently required lower digestion temperatures and caustic concentrations.

However, due to the poor quality of bauxites of Australia in terms of alumina content

and high organic content, the aluminium industries of Australia generally extract

aluminium using relatively mild conditions. Due to the fact that Australia has one of

the largest bauxite deposits in the world, Australian alumina industries are a

competitive supplier of alumina [40].

Australia is one of the top producers of bauxite, totalling almost one-third of the world's

production, followed by China, Brazil, India, and Guinea [41]. Although aluminium

demand is rapidly increasing, there are sufficient known reserves to meet the worldwide

demands for aluminium for many centuries. Resources of bauxites, the raw material

for aluminium, are not widespread throughout the world. There are only seven bauxite-

rich areas: Western and Central Africa (mostly, Guinea), South America (Brazil,

Venezuela, Suriname), the Caribbean (Jamaica), Oceania and Southern Asia (Australia,

India), China, the Mediterranean (Greece, Turkey) and the Urals (Russia) [41]. A table

outlining the current extractions of bauxite are summarised in Table 1, while Figure 1

summarised the bauxite world reserves.

© 2013 Wentao Liang 21

Table 1: World bauxite mine production, reserves and reserve base [41]

Country Mine production

2010 2011 (est.)

Australia 68,400 67,000

China 44,000 46,000

Brazil 28,100 31,000

India 18,000 20,000

Guinea 17,400 18,000

Jamaica 8,540 10,200

Greece 2,100 2,100

Other 22,880 27,080

Figure 1: World bauxite reserves[41]

© 2013 Wentao Liang 22

2.1.2 Bayer process

The one of methods of production of aluminium is the Bayer processes, which is

used for refining bauxite ore to smelting grade alumina (Al2O3). The process was

developed by Karl Josef in 1888, and has become the cornerstone of the aluminum

production industry. Depending on the quality of the ore, between 1.9 and 3.6 tonnes

of bauxite [39]. An investigation by Ostap [42], reported the major mineralogical

composition of bauxites over a large geographical area, including Africa, Australia,

India, Jamaica, Guyana and Brazil. This data is summarized in Tables 2 and 3.

Table 2: Major mineralogical compositions of bauxites[42]

Mineral Formula

Gibbsite Al2O3⋅3H2O

Boehmite Al2O3⋅H2O

Hematite Fe2O3

Aluminium goethite (Fe,Al)2O3⋅H2O

Anatase TiO2

Rutile Ti2O

Kaolinite, Halloysite Al2O3⋅2SiO2 ⋅2H2O

Quartz SiO2

Table 3: Major minerals composition of bauxite[42]

Constituent Percentage Composition

Caustic soluble Al2O3 30%-60%

Fe2O3 1%-20%Ti

TiO2 0.3%-10%

SiO2 (Clay) 0.1%-5%

SiO2 (Quartz) 0.1%-10%

© 2013 Wentao Liang 23

The Bayer process involves two distinct stages responsible for the several phase

transformations that enable alumina to be extracted from bauxite ore. The first

process requires increased temperatures that are dependent on the quality of the

bauxite ore (145°C for high-grade gibbsitic ores and 270°C for low-grade boehmitic

ores). Bauxite ore is crushed before the digestion process in concentrated sodium

hydroxide (NaOH). This part of the process transforms aluminium oxy/hydroxides

into a solution of sodium aluminate, while iron oxide compounds, quartz, sodium

aluminosilicates, calcium carbonate/aluminate and titanium dioxide remain insoluble.

After the dissolution of the aluminum oxy/hydroxides, the insoluble compounds are

removed by settling/filtration techniques. The clarification process is dependent on

the form of the aluminium present in bauxite ore. This waste residue is commonly

known as red mud in the alumina refining industry due to the distinct red colour

caused by the large iron concentration [43].

The next stage of the Bayer process involves the precipitation of gibbsite (Al(OH)3),

which is achieved by cooling the sodium aluminate solution under controlled

conditions. In order to achieve the desired gibbsite particle distribution, seeding

(addition of undersized gibbsite) is sometimes used to encourage the growth of

particles. The final stage of the process is to convert gibbsite to alumina by

calcination at temperatures of around 1000-1100°C in rotary or fluid flash calciners.

Chemically, the reactions of the Bayer process are depicted by the following

equations:

Extraction:

Gibbsite: Al(OH)3(s) + NaOH (aq) → Na+Al(OH)4

- (aq)

Boehmite: AlO(OH)(s)+ NaOH (aq)+ H2O Na+Al(OH)4

- (aq)

(1)

© 2013 Wentao Liang 24

Precipitation:

Na+Al(OH)4

- (aq) Al(OH)3(s)+ NaOH(aq) (2)

Calcination:

2Al(OH)3(s) Al2O3(s)+ 3H2O(g) (3)

A summary of all the processes involved in the Bayer process is illustrated in Fig.2

[39].

Fig.2 Schematic representation of The Bayer process [39]

© 2013 Wentao Liang 25

2.1.3 Limitations of the Bayer process

The two main chemical reactions that can limit the productivity and efficiency of

alumina refineries are; 1) the precipitation of gibbsite (Eq. 2) and 2) the

crystallization of sodium oxalate (Eq. 4). The co-precipitation of sodium oxalate with

gibbsite in the Bayer process results in the production of lower quality alumina due

to the incorporation of sodium as an impurity during calcination. Therefore,

numerous researches have been undertaken to minimize the amount of sodium

oxalate in the alumina stream through the use of materials such as lime or magnesia

that react with alumina to form layered cationic structures that capture and remove

oxalate [44].

Oxalate crystallization:

2Na+

(aq) +C2O42-

(aq) Na2C2O4(s) (4)

Precipitation of gibbsite from the pregnant liquor is a critical phase of the Bayer

process, which requires good crystal growths and purity. Precipitation is effected by

cooling the solution to 58°C, thus causing supersaturation and by introducing small

gibbsite particles as seeds. The crystalline gibbsite precipitate is separated into a fine

and a coarse fraction, whereby the fine fraction is recycled to the precipitators as

seed, while the coarse fraction represents the product to be calcined [40, 42].

Optimising the precipitation stage of the Bayer process is of great importance in

terms of productivity, however in terms of environmental issues and public relations

the optimisation of red mud chemistry and disposal is becoming a hot topic in the

scientific community.

© 2013 Wentao Liang 26

RED MUD APPLICATIONS 2.2

2.2.1 Red mud issues

The greatest downside of the alumina process is the large amount of waste that is

produced. For every ton of alumina produced, between 1-1.5 tonnes of red mud

needs to be treated and disposed [45]. The major chemical constituents of red mud

range from 20-65% Fe2O3, 10-27% Al2O3, 5-25% TiO2 and 2-8% Na2O, with

specific values depending on the source of the bauxite ore and digestion conditions

used [46].

According to Kumar et al. in 2006 [15], red mud had approached 90 million tonnes

annually, with the majority of waste residue comprising of high alkalinity with a pH

ranging from 10-13 [20]. The combination of the chemical complexity and alkalinity

of red mud poses difficulties in the safe disposal of the waste residue. There was a

major dam failure accident in Ajka, Hungary in October 2010 that caused significant

damage and loss of life because the red mud slurry was not treated before its disposal

[18]. Several environmental impacts were investigated after this disaster, with the

main focus on metal bioavailability and toxicity of contaminated soil and waterways.

The extent of environmental damage, the death of 10 people and countless more that

suffered caustic burns [47]. This was an unprecedented accident in the history of the

Bayer process. The accident proved that untreated red mud can pose a significant risk

to living organisms and the environment. Meanwhile, it is essential to note that the

cost for disposal can be up to 5% of alumina production costs [46].

The risks and costs associated with the large disposal areas for the storage of red mud

has received much attention, with many applications focusing on the utilizing of red

© 2013 Wentao Liang 27

mud as a secondary material in building materials and water purification. Pilurzu et

al. [48] successfully made bricks using red mud from the southwest of Sardinia,

however there are numerous innovations in the areas of adsorbents, ceramic glazes,

catalyst, pigments, and polymers [18]. However, due to the vast differences in the

composition of red mud, each application has to be designed with the original red

mud composition in mind. According to Wang et al. [49], the main composition of

red mud in Australia is crystalline hematite (Fe2O3), which makes up about 40% of

red mud, while in China, crystalline hematite (Fe2O3) is just less than 10%. The

Chinese red mud has a greater portion of quartz (SiO2), which makes up around 14%.

This vast chemical difference presents a huge challenge in the development of

applications.

2.2.2 Red mud applications in materials

I. Iron recovery

According to Liu et al. [20] , iron can be recovered from red mud by magnetic

separation, with the aim of using the recovered iron for building materials. They

analysed red mud with different chemical compositions, for the recovery of iron

based on the addition of additives followed by roasting at 1300°C for 110 minutes.

The study determined that the appropriate proportion of carbon powder to red mud to

be 18:100 and an additive ratio of 6:100. This method resulted in an iron recovery

rate of 81.40% depending on the initial iron content. The recovery of iron can be

explained by the following reactions:

Fe2O3 + 3C 3Fe + 3CO (5)

Fe3O4 + 4C 3Fe + 4CO (6)

FeO + C Fe + CO (7)

© 2013 Wentao Liang 28

II. Land fill and land reclamation

Red mud as a waste product should be disposed with as many precautions in place to

minimise the environmental impact. Although some alumina refineries implement a

neutralisation process before disposal, a lot of refineries still dispose without any

manipulation. Due to its very high alkalinity, it is essential that the residue is

neutralised prior to its use in alternative applications. It has been successfully done

by CVG-Bauxilum and Atomaer-KD Engineering Co., Inc. [50], which reduced the

pH of red mud from 12.20 to 7.60. The latest direction for the utilisation of red mud

is using it for filling mined or quarried areas, landfill cover [51], road-base and levee

material [48]. According to Brown and Kirkpatrice [51], red mud can be used as a

road-base and levee material, which is stacked and compacted by heavy equipment.

The red mud roads have a good condition even after a long period of time.

There is a prospective research field for agricultural land neutralization or

composting. Due to high alkalinity of red mud it can be mixed with manures to

neutralize the acidic agricultural land soil [48]. Meanwhile, according to Goldstein

and Reimers [52], red mud is a good disinfectant and stabilizing agent for sewage

treatment. It can be an organic composting material for agricultural uses after it has

been used to treat sewerage. It means it not only can reduce the water pollution, but

also can be a new organic composting material for sustainable agricultural

development.

III. Building materials and ceramic industry

Red mud has been used as bricks for building materials by Pilurzu et al. [53] in the

southwest of Sardinia. Smirnov et al. [54] treated red mud using sulphuric acid,

© 2013 Wentao Liang 29

which deactivates it along with the addition of Fe-Al coagulants before heat

treatment to produce red mud building materials meet specifications. The reuse of

red mud in the ceramic industry is also plausible. Balasubramanian et al. [55] mixed

red mud and fly ash to prepare glass-ceramic products that can be used to decorate

the exteriors in the building industry.

Red mud has been widely used in material industry and shows a lot of versatility.

However, the percentage of red mud used in metal recovery, building and ceramic

materials is limited, and in the scheme of the things makes no real impact on the

removal of red mud from tailing dams. For bulk use, landfill might be considered,

however extensive studies on the possible environmental implications long term need

to be further understood.

2.2.3 Red mud applications in water treatment

Red mud is not limited to landfill or building materials, but it can also be used to

produce valued materials for water treatment. In the past decade, there have been a

lot of applications of red mud in the water treatment industry, with some of the

studies showing promising results.

I. Coagulant

The composition of red mud from different parts of the world are summarised in

Figure 3 [48]. The typical composition of red mud has high contents of Fe and Al in,

which aligns well with coagulant applications that generally involve iron and

aluminium chemical compounds. According to Orešcanin et al. [56], red mud was

treated using sulphuric acid (30% wt), centrifuged, and then neutralised to pH with

waste caustic. This particular product could be used as a coagulant for heavy metal

© 2013 Wentao Liang 30

and turbidity removal in industrial wastewaters. The study revealed that 1 g of the

red mud coagulant could remove 99.3 mg of Cu2+

; 95.0mg of Zn2+

and 98.7mg of

Pb2+

from an initial solution containing 100 mg of each metal. The study proved that

it could be a plausible coagulant for this particular application.

Fig.3 Composition of red mud in different parts of the world [48]

I. Adsorbent

Beside fluoride adsorption, there are others studies on the removal of a number of

pollutants using red mud. Phosphate is one of the main focuses for its removal from

effluent [57-60] and 80-90% of phosphate could be adsorbed using activated red mud

in the pH range 3.2 to 5.5 [60]. In addition to the removal of phosphates, a lot of

studies have also been conducted on the removal of nitrates using raw red mud and

activated red mud. Acid activated red mud (10g with 200ml of 20% wt HCl) had an

adsorption capactiy of 123 and 380 mg/g for raw red mud and activated red mud,

respectively [61]. Studies focused on arsenic removal have been shown to have

adsorption efficiencies 35% higher for acid active red mud than raw red mud [62,

© 2013 Wentao Liang 31

63]. Heavy metals contamination occurs in many aqueous wastewaters, and red mud

has been widely used for surface precipitation and adsorption of heavy metals, such

as Pb2+

, Cd2+

and Zn2+

[64].

i. Phosphorus

Phosphorus is a very important index for water quality, if it is discharged in excess

to waterways it may cause eutrophication, which leads to algal and hydrophytes

blooms. Compared with activated carbon, which is a commonly used adsorbent that

is relatively cost-effective [65], however red mud has been shown to have a more

effective role in the adsorption process. Robert et al. [66] added red mud into a

commercial peat for removing phosphorus, and the result showed that phosphorus

removal was raised from 17% to 21%, with a 95% red mud mixture.

Sulphuric acid treated red mud has been investigated by Mohanty et al. [67], found

improvements in the adsorption capacity of red mud especially when the pH is

around 4.5. Investigations into the effects of heat treatment and the resultant

adsorption capacity of red mud has reported by Li et al.[68], which found red mud

had higher adsorption efficiencies when heated to 700oC for 2 hours. The adsorption

rate increased to around 99%.

Red mud as an adsorbent for removing phosphorous are depended its source and

treatment methods used. In order to obtain sufficient removals, past investigations

show that acidification and heat treatment are required. Table 4 shows the different

treatments used for removing phosphorous by red mud [49]

© 2013 Wentao Liang 32

Table 4: Different conditions for removing phosphorous by red mud [49]

Adsorbent Temperature (oC) P capacity (mg g

-1)

RM-HCl 30 23.2

RM - 40-45

Bauxite 25 0.67-0.82

Heated Bauxite 25 0.98-2.95

RM-H2SO4 40 7.4

RM 25 114

RM-HCl 25 162

Heated RM 25 345

Bauxsol 23 6.5-14.9

ii. Nitrate

Nitrate pollution has increased globally due to increased use of fertilizers and

livestock farming. It may cause high nitrate concentration in ground water, and high

nitrate-related problems in the source of drinking water, which could pose significant

risks to public health. Due to the increasingly nitrate pollution, Çengeloğlu et al. [61]

used red mud and HCl-activated red mud for the removal of nitrate. Compared with

the original adsorption capacity of red mud (1.86mmol nitrate g-1

), while an

adsorption capacity of 5.86 mmol nitrate g-1

was obtained for HCl-activated red mud.

The increased adsorption capacity is shown in the following reaction mechanisms:

RMOH + H+

↔ RMOH2+

(8)

RMOH2+

+ NO3- ↔ RMOH2-NO3 (9)

RMOH2+

+ NO3- ↔ RMNO3 + H2O (10)

(RM represents Red Mud- Fe, Al or Si of red mud surface)

© 2013 Wentao Liang 33

RED MUD TREATMENT COMPARISION 2.3

2.3.1 Acid activated red mud

Acid treatment has been tested on red mud to increase its adsorption capability [30,

58, 61, 62, 69]. For the purpose of removal arsenic, 0.25- 2.0M HCl activated 50g

dry red mud by stirring for 2 hours. The treated red mud was separated from the

solution by filtration, and then washed and dried at 105℃ for 4 hours. The adsorption

efficiencies of acid activated red mud for As (III) was 35% higher than raw red mud,

with a relationship between the concentration of HCl and adsorption capacity being

observed. The number of As (III) adsorption efficiency was 70, 88, 77 and 75% for

red mud activated by 0.75, 1.0, 1.5 and 2.0 HCl, respectively, with can be seen the

best concentration was 1.0M HCl. The reason of adsorption decrease may be due to

the adsorbent was blocked by sodalite leakage. As a result, it can be observed that

adsorption ability of red mud can be increased by acid treatment [62]. Another study

on acid activated red mud was reported by Çengeloğlu,et. al. [61], which involved

batch adsoprion techniques for the removal of nitrate using red mud. All the red mud

was from Kenya, with an average composition of 18.7% Al2O3, 39.7% Fe2O3 and

14.5% SiO2. It was found the adsorption capactiy of the activated red mud was

approximately 3 times better than untreated red mud.

A red mud study by Worsley Alumina, Australia, with a compositions of 60% Fe2O3

16% Na2O, 15% Al2O3, 5% TiO2 and 5% SiO2, also showed increased removal

capacities using acid activated red mud. This investigation used 2M HCl and 2M

HNO3 at liquid/solid ratios of 20ml/g. The study also included the characterisation of

the acid activated red muds (RM-HCl and RM-HNO3), using surface analysis

techniques (BET) and X-ray diffraction (XRD). The BET surface area showed raw

© 2013 Wentao Liang 34

red mud had a surface area of 22.7m2/g, but acid active red mud had a larger surface

area; 28.48m2/g for RM-HCl and 38.15m

2/g for RM-HNO3. The total pore volume

was found to be 0.0566, 0.0779 and 0.0658cm3/g for raw red mud, RM-HCl and RM-

HNO3, respectively. All the results showed that acid activated red mud by HCl and

HNO3 increased the adsorption capacity of red mud, with HCl treated red mud

performing better than HNO3 [59].

Most applications focused on acid treatment by hydrochloric [61, 62], phosphoric

and nitric acid [59], with few reports on the use of sulphuric acid as an acid activator.

Heavy metal removal using red mud generally follows four steps: 1) surface

precipitation, 2) flocculation in the form of hydrolytic products, 3) chemical

adsorption and 4) ion exchage. Based on surface-complex formation, the metal ions

are removed by –OH group on suface or as synthetic cation exchanger [69].

2.3.2 Thermally activated red mud

Heat treatment is another widely used method to increase the adsorption capacity of

red mud. Previous investigations focused on phosphate removal using thermally

activated red mud have shown increases in adsorption capacities [57-59]. The study

by Li et al. [57], looked at the adsorption capacities of seven thermal activation

temperatures (ambient, 200, 500, 600, 700 800, 900 and 1000°C) for phosphate

removal and reported capacities of 9.19, 12.00, 14.27, 16.50, 19.05, 14.38, 14.13,

13.75 mg/g, respectively. Increases in adsorption capacities was reported as an

increase in surface area of red mud started towards 500°C, due to the expulsion of

water leading to the formation of a more porous structure [70]. The maximum

adsorption capacity of red mud was achieved at 700°C (19.05mg/g), which was more

© 2013 Wentao Liang 35

than double untreated red mud with an adsorption capacity of 9.19mg/g. Above

700°C, phosphate removal became less efficient, which has been explained as the

decomposition of some oxy/hydroxyl minerals and to a certain extent the

decomposition of calcite.

Another application has shown the thermal behavior of red mud and the chemical

components changes between 120 and 1400°C [71]. Various complementary

techniques, including X-ray diffraction analyses and thermal analyzer (gas-

chromatographic/mass spectrometer) were used in this investigation, which found

that red mud continued to undergo chemical changes up until 900°C. At this

temperature, H2O vapour and CO2 are lost due to the decomposition of aluminum

hydroxides and silico-alumino-carbonates, respectively. There were more changes

between 900 and 1100°C, assigned to Ca3Al2O6 and NaAlSiO4 due to the alkaline

oxides such as CaO and Na2O. Ferric iron (Fe3+

) was reduced to Fe2+

, and combined

with –TiO2 phase at these elevated temperatures.

2.3.3 Combination treatment

Literature has shown that the combination of acid and heat treatment influence the

adsorption capacity of red mud [58, 59]. However, a comprehensive review of the

effect of different acid and heat treatments on the removal of fluoride does not exist.

An investigation that activated red mud by treating it at 700°C for 2 hours washed

with 0.25M HCl (RM-HCl-700°C). A comparison of thermally activated red mud at

700°C (RM-700°C) with acid and thermal activation (RM-HCl-700°C) showed that

the combination of activation techniques had a higher loading capacity (26.4mg/g

phosphate compared to 19.1mg/g) [58]. While an investigation by Liu et, al. [58],

© 2013 Wentao Liang 36

reported the absorption capacity for phosphate as: RM-HCl > RM-HNO3-700 °C ≈

RM-HNO3 > RM-HCl-700°C > RM. Therefore, it is apparent that the concentration

of acid used in the acid activation requires more focus, as a number of studies report

the optimum calcination temperature to be at around 700°C [20, 57, 64, 72].

FLUORIDE APPLICATIONS 2.4

2.4.1 Fluoride issues

Fluoride is found naturally as sellaite (MgF2), fluorspar (CaF2), cyolite (Na3AlF6) and

fluorapatite (3Ca3(PO4)2Ca(F,Cl2)) [73]. Over time the weathering of these minerals

results in fluoride existing naturally in groundwater [74]. However, elevated levels of

fluoride in groundwater can generally be traced back to a number of industries, including

but not limited to, glass and ceramic production, electroplating, coal fired power stations,

brick and iron works, and aluminium smelters [75, 76]. It is estimated that more than 200

million people rely on contaminated drinking water containing more than 1.5mg/L of

fluoride (World Health Organisation safe level) [77]. Continual and excess exposure to

fluoride results in diseases such as osteoporosis, arthritis, brittle bones, cancer, infertility,

brain damage, Alzheimer and thyroid disorders in humans [76].

2.4.2 Red mud applications with fluoride

There are many different technologies that have been used to remove fluoride from

aqueous systems on a laboratory scale. Granular ferric hydroxide (GFH) has been

shown to be effective, with 95% fluoride removal being reported and a maximum

adsorption capacity of 7.0mg/g [78]. Factors that cause a reduced adsorption capacity

for fluoride included ionic strength, pH value, surface loading and a number of co-

existing anions. Ionic strength had a limited impact on fluoride adsorption in

© 2013 Wentao Liang 37

comparison to the pH, which shows optimum fluoride removal between 3- 6.5. Based

on batch methods, major anions reduced fluoride adsorption in the following order:

Cl < SO42-

< HCO3- < H2SO4 [79]. Metal oxides have also been shown to have good

adsorption capacity for fluoride. An Al-Ce oxide hybrid has been reported to have an

adsorption capacity of 91.4mg/g at 25 ℃ and pH 6.0. Aluminium titanate (AT) and

bismuth aluminate (BA), have been reported to have a fluoride loading capacity of

0.85mg/g and 1.55mg/g, respectively at 30 °C [80, 81]. More recently, a potassium

manganese oxide (KMnO4), iron oxide-hydroxide nanoparticles and activated

alumina have been investigated for their effectiveness in the adsorption fluoride from

aqueous media [82-84].

Due to the vast quantity of red mud available numerous studies have been conducted

on the potential application of red mud for a range of wastewater treatment options.

Commonly, red mud has been activated prior to use through either acid and/or

thermal treatment methods. Acid activated red mud significantly enhances fluoride

removal from aqueous solutions, demonstrated in the work by Çengeloğlu et.al., [61]

who used 20 wt% hydrochloric acid (HCl) as the acid activator. The results clearly

showed the activated form had a higher loading capacity than untreated red mud,

with maximum removal of fluoride being obtained at pH 5.5 [30].

In the case of phosphate adsorption, HCl activated red mud has also been report to

increase the amount of phosphate that can be removed compared to untreated red

mud with the highest adorsption capacity reported as 26.4mg/g (HCl contration at

0.25mol/L) [58]. Except for acid activated, temperature also influences the activation

of red mud. The maximum adsorption capacity of phosphate by calcined red mud can

© 2013 Wentao Liang 38

achieve 19.1mg/g, compared with other samples under different temperatures [58].

Another study relates thermal red mud, which heated red mud at 700°C for 2 hours,

and 99% phosphate removed [57]. Most reports showed activated red mud is better

than untreated red mud for the remove of number of anions such as fluoride [30],

nitrate [61], and phosphate [57, 58].

© 2013 Wentao Liang 39

3CHAPTER 3: EXPERIMENTAL AND CHARACTERISATION

EXPERIMENTAL TECHNIQUES 3.1

3.1.1 Fluoride adsorption

Red mud from an Australian alumina refinery was dried over a period of 2 days at

105°C before being crushed to a fine powder using a Fritsch agate ball mill. Using a

Retsch sieving stack consisting of 10 sieves, ranging from 4mm to 64µm, red mud

was processed to give a size fraction < 250µm. A known amount of red mud was

then placed in a SEM furnace that had been previously heated to the desired

temperature. The sample of red mud was placed ion the furnace and heated for 2

hours. After heating, it was allowed to cool to room temperature before being re-

weighted so that the mass loss could be calculated. The red mud was then placed in a

desiccator until it was used for adsorption studies.

3.1.2 Characterisation techniques

Fluoride analysis was obtained using a TPS uniPROBE fluoride (F-) ion selective

electrode (ISE). A fluoride ISE buffer was prepared using 1 M of sodium chloride

(NaCl) and 1 M sodium citrate dehydrate (Na3C6H5O7·2H2O) dissolved in

approximately 1.5 L deionised water. Sodium hydroxide (NaOH) was use to adjust

the solution pH to 5.5 before making the solution up to 2 L. Calibration standards of

1, 10, and 100 mg / L fluoride stock solutions were prepared from AR grade sodium

fluoride (NaF). Standards and samples were analysed by combining 10 mL of the

calibration standard solution (or sample) and 10 mL of the buffer solution whilst

being stirred. The concentration of fluoride was then measured using the fluoride

ISE.

© 2013 Wentao Liang 40

The oxide composition of red mud (RM) and red mud heated to 500ºC (RM500) was

determined using XRF. Particle size analysis of the components of RM and heat

treated RM was obtained by the granulometric method. It was obtained by sieving

methods and by sedimentation analysis. The specific surface areas of all samples

were determinate by BET/N2 adsorption methods. The pH of the red mud samples

was determined by mixing dry red mud (2g) in water (10mL) with a TPS WP40 pH

meter calibrated with pH 7.00 and 10.01 buffers.

X-Ray diffraction patterns were collected using a Philips X'pert wide angle X-Ray

diffract meter, operating in step scan mode, with Co K radiation (1.7903 Å).

Patterns were collected in the range 5 to 90° 2 with a step size of 0.02° and a rate of

30s per step. Samples were prepared as Vaseline thin films on silica wafers, which

were then placed onto aluminium sample holders. The XRD patterns were matched

with ICSD reference patterns using the software package HighScore Plus. The

profile fitting option of the software uses a model that employs twelve intrinsic

parameters to describe the profile, the instrumental aberration and wavelength

dependent contributions to the profile.

Samples of the residual acid solutions, after centrifugation, were analysed using an

Agilent ICP-MS 7500CE instrument. The samples were diluted by a factor of 20

using a Hamilton dilutor with 10 and 1mL syringes. A certified standard from

Australian Chemical Reagents (ARC) containing 1000ppm of aluminium,

magnesium, calcium, iron and sodium were diluted to form a multi-level calibration

curve and an external reference that was used to monitor instrument drift and

accuracy of the results obtained. Results were obtained using an integration time of

© 2013 Wentao Liang 41

0.15 seconds with 10 replications. Calibration curves had an r2 value of 0.998 or

higher.

Infrared spectra were obtained using a Nicolet Nexus 870 Fourier Transform infrared

spectrometer (FTIR) with a smart endurance single bounce diamond ATR

(attenuated total reflectance) cell. Spectra over the 4000-525 cm-1

range were

obtained by the co-addition of 64 scans with a resolution of 4 cm-1

and a mirror

velocity of 0.6329 m/s.

CHARACTERISATION TECHNIQUES 3.2

3.2.1 X-ray Diffraction (XRD)

The crystal structure of each material is unique and depends on how its atoms are

ordered in planes and axes. The intercept of these planes with the symmetry axis

determines their position within the crystal lattice. The atoms interact with the

electromagnetic properties of the X-rays causing them to scatter. The majority of the

waves are made extinct through destructive interference: however, some are

enhanced by constructive interference. The direction of these latter waves is related

to the distance between the atomic planes, known as d-values, and the Bragg angle

(Figure 4). is the angle at which the waves hit the crystal producing the maximum

interference. The Bragg equation links this angle of scattering to the d-values as

follows [85]:

© 2013 Wentao Liang 42

Fig.4: X-ray diffraction between planes of atoms

= (11)

Where λ= wavelength used (nm)

= order of reflection

The output is the observed intensity plotted against Bragg angle, from which the d-

values are calculated. The type of radiation usually used for iron samples is either

CoK or CuK . If CuK is used, the observed intensity is much lower, as iron

absorbs much of the radiation.

I. D-values of common iron oxides

Each crystalline product has characteristic d-values that can be found on reference

cards in the Joint Committee on Powder Diffraction Standards (JCPDS) mineral

powder diffraction file. A compound can be identified by comparing the calculated

d-values against the JCPDS diffraction file of iron oxides summarised in Table 5.

The values are in order of intensity; the greatest is listed first

© 2013 Wentao Liang 43

Table 5: d- values of iron oxides[85]

Mineral Most intense d-values (Å)

Hematite 2.70, 3.68, 2.52

Magnetite 2.53, 1.49, 2.97

Goethite 4.18, 2.45, 2.69

Akaganeite 3.33, 2.55, 7.47

Lepidocrocite 6.26, 3.29. 2.47, 1.94

Ferrihydrite 2.54, 2.24, 1.97, 1.73, 1.47

Feroxyhyte 2.54, 2.22, 1.69, 1.47

3.2.2 Infrared Spectroscopy (IR)

Infrared spectroscopy (IR spectroscopy) is the spectroscopy that deals with the

infrared region of the electromagnetic spectrum, that is light with a longer

wavelength and lower frequency than visible light. It covers a range of techniques,

mostly based on absorption spectroscopy. As with all spectroscopic techniques, it can

be used to identify and study chemicals. A common laboratory instrument that uses

this technique is a Fourier transform infrared (FTIR) spectrometer.

The infrared portion of the electromagnetic spectrum is usually divided into three

regions; the near-, mid- and far- infrared, named for their relation to the visible

spectrum. The higher-energy near-IR, approximately 14000–4000 cm−1 (0.8–2.5 μm

wavelength) can excite overtone or harmonic vibrations. The mid-infrared,

approximately 4000–400 cm−1 (2.5–25 μm) may be used to study the fundamental

vibrations and associated rotational-vibrational structure. The far-infrared,

approximately 400–10 cm−1 (25–1000 μm), lying adjacent to the microwave region,

has low energy and may be used for rotational spectroscopy. The names and

© 2013 Wentao Liang 44

classifications of these subregions are conventions, and are only loosely based on the

relative molecular or electromagnetic properties.

Infrared spectroscopy exploits the fact that molecules absorb specific frequencies

that are characteristic of their structure. These absorptions are resonant frequencies,

i.e. the frequency of the absorbed radiation matches the transition energy of the bond

or group that vibrates. The energies are determined by the shape of the molecular

potential energy surfaces, the masses of the atoms, and the associated vibronic

coupling.

In particular, in the Born–Oppenheimer and harmonic approximations, i.e. when the

molecular Hamiltonian corresponding to the electronic ground state can be

approximated by a harmonic oscillator in the neighbourhood of the equilibrium

molecular geometry, the resonant frequencies are associated with the normal modes

corresponding to the molecular electronic ground state potential energy surface. The

resonant frequencies are also related to the strength of the bond and the mass of the

atoms at either end of it. Thus, the frequency of the vibrations is associated with a

particular normal mode of motion and a particular bond type.

3.2.3 Surface Area and Pore Size Analysis

The specific surface area is a key functional parameter for an adsorbent. Summaries

of adsorption theory, Langmuir and Brunauer–Emmett–Teller (BET) equations are

provided in subsequent sections. The resulting isotherms from nitrogen adsorption

can be classified and the pore size distribution extracted from the data. The types of

isotherms and hysteresis loops are outlined, in addition to the various methods used

for pore size distribution determination.

© 2013 Wentao Liang 45

I. Adsorption Theory

When a disperse solid is exposed to a gas in an enclosed space at a known pressure,

the solid will absorb the gas. After a period of time, the pressure will become

constant. The extent, of gas adsorption on a porous solid depends on the temperature,

pressure and the molecular size of the gas and available surface area. The amount of

gas adsorbed can be calculated from the change in pressure, if the volume of the

vessel and solid are known. Hence for a known gas held at a specified temperature,

the amount of gas adsorbed is solely dependent on the pressure and therefore an

adsorption isotherm can be constructed (volume adsorbed vs. relative pressure,

(pressure/ saturated vapour pressure, / )). The surface free energy is reduced by

adsorption, hence entropy is reduced. It is an exothermic process, with the heat of

physical adsorption being similar in magnitude to the heat of condensation [86].

a) Langmuir equation

Langmuir [87] was the first to present a kinetic approach to adsorption was based on

a flat surface with zero accumulation on the surface. i. e. the rate of adsorption is

equal to the rate of desorption. It is assumed that the energy of adsorption is constant;

implying that the surface is uniform and monolaver coverage is reached at saturation.

b) BET equation

The BET equation (12) is based on kinetics and is a modified version of the

Langmuir isotherm with multilayer, rather than monolayer, adsorption [88]. The

previously adsorbed layer serves as adsorption sites for the next layer and so on.

Generalising the Langmuir equation maintains the disadvantage of the assumption

that the surface is energetically uniform, i.e. all adsorption sites are equivalent and

© 2013 Wentao Liang 46

that there is no interaction between adsorbed molecules. It is also assumed that at

saturated pressure the number of layers is infinite.

The BET equation is restricted in its validity to a relative pressure ( / ) in the range

0.05-0.35 [89]. The lower limit is set so as not to include the heat of adsorption for

the first layer of molecules, which is much greater than subsequent layers. The BET

equation breaks down at relative pressures > 0.35 and over-predicts adsorption by

computing a high entropy value [90].

=

+

∙

(12)

Where = adsorption capacity (mmol/g)

= monolayer adsorption capacity (mmol/g)

= pressure (Pa)

= saturation pressure (Pa)

= BET contant

The specific surface can be calculated if the ‘knee’ (point B in Figure 5) on the

isotherm is well defined. This would be characterised by a high value of the BET

constant , represents the monolayer adsorption capacity and is calculated from

the BET equation (12), assuming one molecule of nitrogen occupies 16.2 Å2.

II. Isotherm classification

The isotherm produced from nitrogen adsorption can usually be identified as one of

five classes proposed by Brunauer, Deming, Deming and Teller (BDDT) [91], in

© 2013 Wentao Liang 47

addition to a stepwise isotherm (see Figure 5). Table 6 provides a summary of the

characteristics of each isotherm, degree of porosity and an example system.

The Type IV isotherm which describes the adsorption for ferric hydroxide can be

used to calculate the specific surface of the material and the pore size

distribution[91]. Type IV is characteristic of a material possessing a pore size

distribution in the range of up to hundreds of angstroms. The definition of pore size

as recommended by IUPAC is an arbitrary classification, developed using nitrogen

as an adsorbent at its normal boiling point [92]. Pores widths than 20 Å are given the

term micropores, between 20 and 500 Å are mesopores and greater than 500 Å are

macropores. Each classification has a particular significance in terms of the

adsorption isotherm (see Table 6).

Fig.5: Classification of NDDT adsorption isotherm [91]

© 2013 Wentao Liang 48

Table 6: Types of isotherms, characteristics and examples

Type Characteristics Example System References

I. Microporous with small internal

surface

Oxygen on activated carbon [91]

II. Reversible non-porous or

macroporous

Nitrogen on iron catalyst [89]

III. Adsorben-adsorbate interaction

is weak

Water vapour on non-

porous carbon

[92]

IV. Hysteresis loop with a plateau,

mesporous

Benzene on iron oxide gel [89]

V. Asorben-adsorbate interaction is

weak porous

Water on charcoal [91]

VI. Stepwise multi-layer, non-

porous

Argon on graphitised

carbon black

[92]

During the adsorption process, the adsorbate diffuses through the pores of the solid.

Initially this is through the smallest, or micropores. These sites have the most energy

and have greatest polarity. As the pressure increases, the adsorbate molecules diffuse

to larger and therefore less energetic pores. In monolayer coverage, all the molecules

are in contact with the surface. For multilayer adsorption, only the lowest layer is in

contact with the surface [92].

III. Pore size distribution

There are several methods for determining the pore size distribution (PSD) from

isotherm data. Some have a mechanistic basis (e. g. Barrett, Joyner and Halenda

(BJH) and Dubinin-Stoeckli (D-S)), Horvarth-Kawazoe (11-K) uses a quasi-

thermodynamic approach and the density functional theory (DFT) is generated by

statistical thermodynamics. The advantages and disadvantages of these different

methods are outlined below.

© 2013 Wentao Liang 49

a) Mesopore determination

PSD based on the Kelvin equation, such as the BJH method [93] assume there are no

micropores and that the material exhibits a regular pore structure [92]. The BJH

method is only suitable for mesoporous size analysis and is incompatible with the H-

K, D-s and t-plot methods [94].

b) Micropore determination

Much of the work in this area of determining micropore volume was completed by

Dubinin and his co-workers [95, 96]. Other methods include the t-plot modification

and the H-K method [97]. However there has been some doubt as to their ability to

describe micropore adsorption realistically [98].

More recently interest has developed in a model using a density functional theory

(DFT) to calculate model isotherms. From this data a pore size distribution can be

generated using a deconvolution technique. A slit-like or cylindrical shape pore

geometry is assumed [99]. This has enabled a method to describe the entire pore size

range, rather than that limited by the Kelvin equation (13). For more details

regarding the DFT approach see [100].

+

=

)

(13)

Where = saturation vapour pressure (Pa)

= critical pressure (Pa)

= surface tension of liquid condensate (N m

-1)

= molar volume of the liquid condensate (cm3 mol

-1)

© 2013 Wentao Liang 50

= principle pore radii of curvature in the liquid meniscus (Å)

R = universal gas constant (J, mol-1

K -1

)

T= TEMPERATURE (K)

There is a characteristic drop in the PSD using DFT model between 8-10 Å, due to

the packing effect of adsorbate molecules in the pores. Initially a monolayer is

formed in the narrowest pores, a process known as volume filling. After a critical

pore value, a second layer is required to fill the pore. This leads to an. abrupt increase

in adsorbate uptake, indicating the onset of multilayer formation [99]

This model does not take into account the heterogeneity of the pore walls, namely

geometric and energetic differences [101]. In addition, electrostatic interactions are

not considered in the DFT model, which assumes spherical adsorbate molecules and

interaction solely by van der Waals forces [98]. This leads to an underestimation of

the PSD.

© 2013 Wentao Liang 51

4CHAPTER 4: ACID ACTIVATED PROCESS

The majority of researches investigating acid treated red mud with a focus on the use

of hydrochloric acid for the activation process. However, this could reduce the iron

content available for subsequent adsorption applications. Consequently, this chapter

of the thesis outline the structural changes and resulting adsorption capacities of

Australian red mud untreated and treated with hydrochloric, nitric and sulphuric acid.

Particular emphasis will be placed upon determining the reactions involved during

acid activation and how the remaining mineralogical composition correlates with

fluoride removal efficiencies.

EXPERIMENTAL PROCEDURE 4.1

4.1.1 Acid activated red mud

Red mud from an Australian alumina refinery was dried over a period of 2 days at

105°C before being crushed to a fine powder using an agate ball mill (Fig. 7 and fig.

8). Using a Retsch AS200 sieving stack (Fig.9) consisting of 10 sieves, ranging from

4mm to 64µm, red mud was processed to give a size fraction < 250µm. Known

concentrations of hydrochloric (HCl), nitric (HNO3) and sulphuric (H2SO4) acid were

prepared from concentrated AR reagents (Rowe Scientific). Red mud (12.5g) was

measured into 250mL Nalgene bottles and then reacted with 200mL of DI water and

each acid. Each bottle was placed on a Ratek rotary stirrer for 1 hour at 200rpm.

These samples were subsequently centrifuged at 400rpm for 10 min using a C2041

Centurion centrifuge. The red mud was placed in the oven to dry (90°C), while the

solution was stored for analysis.

© 2013 Wentao Liang 52

RESULTS AND DISSCUSSION 4.2

4.2.1 Elemental and mineralogical composition

The elemental composition of this Australian red mud comprised of predominately

iron (Fe2O3), titanium (TiO2) and aluminium (Al2O3 and AlO(OH)) mineralogical

compounds. Although the elemental analysis of red mud has been reported in

numerous papers, the composition of each red mud sample differs, due to the original

composition of the bauxite ore and the operating conditions used to extract alumina.

The elemental abundance in bauxite residues generally follow Fe > Si ~ Ti > Al > Ca

> Na [102]. The red mud used in this work has a particularly high content of

aluminium, suggesting that the operating conditions were not optimised. The phase

composition of untreated red mud (Figure 6) comprised of majorly hematite (Fe2O3),

gibbsite (Al(OH)3), boehmite (γ-AlO(OH)), sodalite (Na8(Al6Si6O24)Cl2 or

Na8(Al6Si6O24)CO3), TiO2 (anatase and rutile), quartz (SiO2) and possibly cancrinite

(Na6Ca2Al6Si6O24(CO3)2). These mineralogical phases agree with the elemental

analysis results (Table 6). The broadness of the peaks in the XRD pattern align with

the following remarks from Grafe et al., [102] that approximately 70% (by weight)

of bauxite phases are crystalline, while the remaining 30% are amorphous materials.

© 2013 Wentao Liang 53

Table 7: Major elemental composition represented as (Al,Na,Si,Ti):Fe

Red mud treatment Al:Fe Na:Fe Si:Fe Ti:Fe

DI water 1.27 1.04 0.73 0.22

Conc. HCl 0.98 1.04 0.71 0.21

0.5M HCl 0.58 0.27 0.22 0.22

Conc. HNO3 0.76 0.27 0.41 0.19

0.5M HNO3 0.96 0.57 0.40 0.22

Conc. H2SO4 1.07 0.97 0.73 0.00

0.5M H2SO4 0.82 0.37 0.32 0.18

A comparison of the washed and acid treated red mud XRD patterns (Figure 7)

shows some phase intensity changes, and suggests that a portion of the mineralogical

phases are unstable in acidic media. In particular, the sodalite peaks at around 16.2,

28.1 and 40.2 °2θ significantly decrease in intensity indicating the dissolution of this

phase. Multiple decreases in intensities are observed between 30 and 35 °2θ, and are

associated with quartz (HNO3 and H2SO4 predominately) and cancrinite. It appears

that 0.5M acid solutions have a limited effect on the overall mineralogical structure

of iron and titanium oxide phases. This is ideal, as a high iron content in red mud has

been found to be beneficial in the removal of fluoride [103].

© 2013 Wentao Liang 54

Figure 6: Mineralogical composition of red mud washed with DI water

Figure 7: XRD patterns of washed and acid treated red muds (0.5 M)

© 2013 Wentao Liang 55

4.2.2 Leachants (ICP and reactions)

I. SEM-EDX

A comparison of the amount of Al, Na, Si and Ti to the amount of Fe in the washed

and acid treated red muds is provided in Table 1. Interestingly, a higher

concentration of acid doesn’t necessarily mean that a greater amount of a particular

element will be removed from the solid phase. In the case of HCl and H2SO4, the

0.5M acids had a greater effect on the dissolution of compounds containing

aluminium, sodium and/or silica than their respective concentrated acid counterparts.

This can be explained by the formation of additional phases between the dissolution

species and excess Cl- ions when using concentrated HCl, for example that forms

NaCl. This is confirmed by XRD (Figure 7), which showed halite peaks (~ 36.9 °2θ).

The SEM image of red mud reacted with concentrated HCl (Figure 8) also clearly

shows the formation of an addition mineralogical phase, and based on the XRD

pattern it can be assumed that halite (NaCl) has coated the exterior of the red mud

particles. The formation of this phase is not expected to hinder the removal of

fluoride as it is highly soluble, however due to its solubility it will increase the

salinity of any treated solutions.

II. ICP-MS

After the treatment of red mud with each of the 1M acids, the filtrate was analysed

using ICP-MS to determine the concentration of major ions being released into

solution (Figure 9). It was found that sodium containing compounds are the most

susceptible to acidic solutions. An increase in Na+ is also observed for the washed

red mud sample indicating that about a third of sodium released into solution is due

to the dissolution of residual NaOH or NaOH trapped in sodalite aggregates. The

© 2013 Wentao Liang 56

general formula for sodalite is Na8 (Al6Si6O24) CO3, which has a Na:Al mole ratio of

1.3. By subtracting the amounts of Na and Al ions released by the dissolution of

NaOH (based on red mud washed values) from the concentrations of ions in the

leachant solution of HCl, H2SO4 and HNO3 gives the following Na:Al ratios: 1.52,

1.26 and 1.27, respectively. This indicates that the majority of sodium and

aluminium ions released in solution are due to the dissolution of sodalite. A lower

amount of sodium is released into solution for HCl treated red mud due to the

formation of NaCl. The release of calcium is proposed to be due to Ca substituted

sodalite and/or cancrinite (Na6Ca2Al6Si6O24 (CO3)2). The dissolution of sodalite is

predicted to form the following products [104]:

Na8 (Al6Si6O24) CO3 + 13H2SO4 →

4Na2SO4 + 3Al2 (SO4)3 + 6Si (OH) 4 + H2O + CO2↑ (14)

Na8 (Al6Si6O24) CO3 + 24HCl →

8NaCl + 6AlCl3 + 6Si (OH) 4 + H2O + CO2↑ (15)

Na8(Al6Si6O24)CO3 + 24HNO3 →

8Na(NO3) + 6Al(NO3)3 + 6Si(OH)4 + H2O + CO2↑ (16)

© 2013 Wentao Liang 57

Figure 8: SEM image of red mud reacted with concentrated HC

Figure 9: ICP-MS of acid treated red mud filtrate

© 2013 Wentao Liang 58

4.2.3 Infrared spectroscopy

I. Red mud

Infrared spectroscopy has been used to monitor changes in bonding environments of

the numerous components of red mud with the addition of different acids. These

changes include both shifts in band position (strength of bonds) and also

decreases/increases in intensity due to the dissolution/formation of phases. Washed

red mud (Figure 10) will be used as a baseline to any changes that have occurred

with the addition of different strengths of acid. A broad band between 3650 and 3000

cm-1

is associated with multitude overlapping hydroxyl-stretching bands, in

particular metal-OH groups and water. Based on the XRD pattern of red mud it is

believed that the multiple of bands between 3650 and 3300 cm-1

are due to the ν

(OH) stretching modes of gibbsite [105-107]. Boehmite peaks are observed at

around 3235 and 3124 cm-1

[105, 106]. Bands associated with surface hydroxyl

groups of hematite appear in the range 3700, 3635, 3490, 3435 and 3380cm-1

.[108]

In many cases these bands are not observed because the surface hydroxyl groups are

removed during the drying process. The overall broadness of the band is due to

multiple water stretching modes. Corresponding water bending modes are observed

as a low intensity broad peak centred at 1655 cm-1

.

In the lower wave number region, several bands are observed between 1550 and

1350 cm-1

, predominately due to carbonate ions in different bonding environments.

The only carbonate containing minerals identified in the XRD patterns are sodalite

and cancrinite, however some form of carbonate mineral may also be present in the

amorphous content of red mud. The most intense peak in the infrared spectrum (987

cm-1

) is believed to be due to stretching vibrations of Si(Al)-O in sodalite and

© 2013 Wentao Liang 59

cancrinite [109]. The small band at 696 cm-1

is also thought to be associated with the

Si-O-Al framework of sodalite [110]. It is also possible that nitrate is incorporated in

these cage structures due to the presence of bands at around 1430 cm-1

[111].

II. Acid treated red mud

The dissolution of sodalite (disappearance of the intense band at 987 cm-1

and the

bands between 1400 and 1500 cm-1

) is clearly observed in Figure 10 with the

addition of hydrochloric acid. These observations coincide with the interpretation of

the elemental analysis and the proposed dissolution reactions. In the absence of

sodalite, bands associated with Si-O vibrations (possibly quartz) are observed at

1036 cm-1

, which gradually decreases and is confirmed by XRD patterns. The band

profile also indicates the formation of SiO2·xH2O (formed from dissolved silica –

Si(OH)4), which has characteristic bands at 800 (w), 948 (w), 1090 (vs) 1190 (s,sh)

and 3330 cm-1

(m) [112] . This dissolution product appears to be relatively stable.

Minimal changes in the higher wavenumber region indicates that the major iron and

aluminium oxide/hydroxide components of red mud remain relatively unscathed until

the concentration of acid reached 1M (significant decrease in intensity suggesting the

initial stages of dissolution).

The infrared spectra of red mud treated with nitric acid (Figure 11) and sulphuric

acid (Figure 9) show many similar bands as those described for hydrochloric acid

treated red mud. Nitric acid treated red mud showed additional bands at 1406 and

1352cm-1

due to Al hydroxylated nitrate and ν3 NO3- co-adsorbed with H2O on the

red mud particles, respectively [111].Sulphuric acid treated red mud also showed an

additional band (broad shoulder on the 1074 cm-1

) ascribed to sulphate vibration

© 2013 Wentao Liang 60

modes found in the dissolution product thenardite (Na2SO4). The broadness of the

higher wave number region is proposed to be water adsorbed to thenardite.

Figure 10: Infrared spectrum of washed red mud

© 2013 Wentao Liang 61

.

Figure 11: Infrared spectrum of red mud treated with various concentrations of

HCl

© 2013 Wentao Liang 62

Figure 12: Infrared spectrum of red mud treated with various concentrations of

HNO3

© 2013 Wentao Liang 63

Figure 13: Infrared spectrum of red mud treated with various concentrations of

H2SO4

© 2013 Wentao Liang 64

4.2.4 Removal of fluoride using acid treated red mud

The effect of acid activated red mud and pH on the removal of fluoride from aqueous

solutions was investigated and found that red mud treated with sulphuric acid has the

greatest efficiency in fluoride removal (approximately 70% removed from an initial

concentration of 100ppm), independent of the initial solution pH (Figures 14 and 15).

It is clear from the results that the best removal percentages are achieved at low pH <

4.5. The study by Cengeloglu et al.,[103] reported similar removal percentages using

red mud, however maximum adsorption was reported to occur at a pH around 5.5.

Deviations in results between the studies are believed to be due to the different

activation processes used.

The removal mechanism of fluoride using red mud primarily involves neutral

(≡SOH) and protonated (≡SOH2+) sites on the oxide/hydroxide components (such as

hematite and gibbsite) when the pH is less than 7[113]. The increased removal

efficiencies of acid treated red mud compared to wash red mud are consistent with

the protonation of the surface hydroxyl groups (17). Increased removal efficiencies

for sulphuric acid treated red mud are due to 2 protons being available to protonate

the surface hydroxyl groups.

≡SOH + H+ → ≡SOH2

+ (17)

The benefit of the protonated sites in acid treated red mud is that the replacement of a

proton (H+) with fluoride (F

-) releases water (18), while the substitution of a proton

with a neutral site (19) releases a hydroxyl unit and in turn causes the pH to rise. This

is observed for the DI washed red mud sample (neutral pH and thus has ≡SOH sites)

© 2013 Wentao Liang 65

that showed promising removal efficiencies at low mass to volume ratios, however

with the continued release of OH- units, as F

- ions are adsorbed, the pH became

alkaline and resulted in the deprotonation of the surface sites (≡SO-), 20. This can be

further confirmed by comparing the removal efficiencies of washed red mud;

whereby the fluoride solution with an initial pH of 4.75 showed reasonable fluoride

adsorption (approximately 35%) followed by a sharp decrease as the pH rose above 6

(Figure 13), while a maximum of 10% fluoride removal is observed for the fluoride

solution with an initial pH of 8 (Figure 14). Fluoride adsorption is hindered when the

pH is greater than 6 because of the increasing repulsive forces between the

negatively charged surface (≡SO-) and fluoride ions.

≡ SOH2+ + F

- → ≡SF + H2O (18)

≡SOH + F- → ≡SF + OH

- (19)

≡SOH + OH- → ≡SO

- + H2O (20)

The efficiency of fluoride adsorption is highly dependent on the pH and any

fluctuations in pH. This is clearly observed in Figure 14 for the adsorption of fluoride

using sulphuric acidified red mud. This relationship shows that for consistent

fluoride adsorption, a constant pH needs to be maintained to avoid any sudden

shocks to the surface adsorption sites. It is also highly possible that the formation of

AlFx complexes, in particular aluminium trifluoride (AlF3), occurs when the pH of

solution is less than 4. The formation of AlF3 would account for some of fluctuations

in fluoride removal percentages. The study by Cengeloglu et al.[103] reported a

decline in fluoride removal at pH below 4 and did not report the formation of any of

these AlFx phases. This could be accounted for the difference in the quantity of

© 2013 Wentao Liang 66

gibbsite of the different red mud sources used, whereby this study had a greater

amount of Al2O3 (24.0%) than Cengeloglu (18.7%) [103]. The formation of this

phase is a result of gibbsite in red mud reacting with HF that forms under these

highly acidic conditions (21).

Al2O3 + 6HF → 2AlF3 + 3H2O (21)

Figure 14: Removal of fluoride (%) using different masses of acid treated red mud

and the associated changes in pH from the original 100ppm fluoride solution with pH

4.75

© 2013 Wentao Liang 67

Figure 15: Removal of fluoride (%) using different masses of acid treated red mud

and the associated changes in pH from the original 100ppm fluoride solution with pH

7.99

© 2013 Wentao Liang 68

5CHAPTER 5: THERMALLY ACTIVATED RED MUD

EXPERIMENTAL PROCEDURES 5.1

5.1.1 Thermally activated red mud

Red mud from an Australian alumina refinery was dried over a period of 2 days at

105°C before being crushed to a fine powder using an agate ball mill. Using a Retsch

sieving stack consisting of 10 sieves, ranging from 4mm to 64µm, red mud was

processed to give a size fraction < 250µm. A known amount of red mud was then

placed in a SEM furnace that had been previously heated to the desired temperature

(100, 200, 300, 400, 500, 600, 700, 800, 900, 1000°C). The sample of red mud was

heated for 2 hours. After heating, it was allowed to cool to room temperature before

being re-weighted so that the mass loss could be calculated. The red mud was then

placed in a desiccator until it was used for adsorption studies.

RESULTS AND DISSCUSSION 5.2

5.2.1 Red mud characterisation

The mineralogical composition of red mud and the corresponding reference patterns

are provided in Figure 16. The identified mineral phases in the sample are hematite

(Fe2O3), gibbsite (γ-Al(OH)3), boehmite (γ-AlO(OH)), sodalite (Na8(Al6Si6O24)Cl2),

anatase (TiO2), rutile (TiO2) and quartz (SiO4). The chemical composition of red

mud is presented in Table 1, with the main component of red mud being iron oxide

(27.77%). The combined oxide content of Fe2O3, Al2O3 and SiO2 in the red mud

sample makes up about 70%. Although the elemental analysis of red mud has been

reported in numerous papers, the composition of each red mud sample differs, due to

the original composition of the bauxite ore and the operating conditions used to

extract alumina. The elemental abundance in bauxite residues generally follow Fe >

© 2013 Wentao Liang 69

Si ~ Ti > Al > Ca > Na[114]. This particular red mud had a higher than usual

abundance of aluminium, suggesting that operating conditions were not optimal at

the time.

Figure 16: X-ray diffraction pattern of red mud and corresponding reference patterns.

Infrared spectroscopy has been used to monitor changes in bonding environments of

red mud after thermal activation. The infrared spectrum of red mud is provided in

Figure 17. The untreated red mud presents a broad band between 3650 and 3000 cm-

1, a number of overlapping medium strength bands between 1650 and 1300 cm

-1, a

strong relatively broad band centred at 968 cm-1

, and several weak bands ranging

from 800 to 400 cm-1

. The broad band centred at around 3300 cm-1

is assigned to a

multitude of overlapping hydroxyl-stretching bands, in particular metal-OH groups

and water. Based on the XRD pattern of red mud, bands associated with aluminium

species gibbsite, boehmite and hematite are expected. The multiple bands observed

© 2013 Wentao Liang 70

between 3650 and 3300 cm-1

are due to the ν(OH) stretching modes of gibbsite [105-

107] while the band at 3277 cm-1

is assigned to ν(OH) stretching modes of boehmite

[105, 106]. Bands associated with surface hydroxyl groups of hematite have been

reported in literature at around 3700, 3635, 3490, 3435 and 3380cm-1

[108]. In many

cases these bands are not observed because the surface hydroxyl groups are removed

during the drying process, and this appears to be the case in this investigation. The

broadness of the band indicates that the samples contain water, which is further

confirmed by water bending modes being observed as a low intensity band at around

1630 cm-1

.

Table 8: Oxide composition of red mud

Oxide RM (%)

Fe2O3 27.77

Al2O3 23.44

SiO2 19.00

Na2O 5.71

CaO 3.72

TiO2 3.05

MgO 0.24

MnO 0.09

Amorphous / other 17.01

The main band observed in the infrared spectrum at 968 cm-1

is assigned to Si(Al)-O

stretching vibrations in sodalite [109]. Weak bands at 709, 670 and 646 cm-1

are

assigned to vibrations associated with the Si-O-Al framework of sodalite [110]. A

number of bands between 1500 and 1300 cm-1

(1458, 1436, 1402 and 1318 cm-1

) are

associated with C-O, C-N and C-C vibration modes. Based on XRD, the only

carbonate or nitrate containing mineral identified is sodalite, however some form of

© 2013 Wentao Liang 71

carbonate mineral may also be present, such as calcium carbonate. The band at 1436

cm-1

indicates that nitrate is incorporated in sodalite [111].

Figure 17: Infrared spectrum of red mud

The differential thermalgravimetric curve for red mud shows a number of

decomposition steps (Figure 3).The initial mass losses (56 and 82°C) are due to

physio-sorbed water and account for a total mass loss of 3.30% of a total mass loss of

11.65%. Up to 600˚C, the observed weight loss is proposed to be related to the

dehydroxylation of AlO(OH) phases. The largest mass loss of 4.70% is due to the

dehydroxylation of gibbsite (22) and boehmite (23) [115, 116]. At higher

temperatures, up to 700˚C, the observed weight loss could be related to the

decomposition of sodalite group phases attributed to the evolution of CO2 and

formation of mixed oxide phases [117]. It has been reported the mass loss in this

temperature range is due to the decomposition of calcium carbonate [116, 118],

© 2013 Wentao Liang 72

however this mineralogical phase has not been identified in the XRD patterns (Figure

19) but is possible as XRF found 3.72% CaO exists in the samples (Table 8). The

XRD pattern also identified the formation of nepheline (NaAlSiO4) in red mud

activated above 800°C, which corresponds to a mass loss of 0.93% observed in the

DTG curve (Figure 18). Nepheline forms as a result of the decomposition of the

sodalite cage structure (24) [119].

Dehydroxylation of gibbsite

Al(OH)3(s) → AlO(OH)(s) + H2O(g) (22)

Dehydroxylation of boehmite

AlO(OH)(s) → Al2O3(s) + H2O(g) (23)

Formation of magnetite

Na8Al6Si6O24Cl2(s) → 6NaAlSiO4(s) + 2NaCl(g) (24)

Figure 18: DTG curve of red mud

© 2013 Wentao Liang 73

5.2.2 Thermally analysis

The mineralogical composition of red mud thermally activated up to 1000°C is

provided in Figure 19. The patterns are dominated by hematite (Fe2O3) peaks for all

temperatures investigated, with only a very weak magnetite (Fe3O4) peak observed at

1000°C. The major changes in mineralogy with increased temperatures is the

disappearance of sodalite and boehmite peaks and the appearance of nepheline,

potassium aluminium oxide and minor amounts of other mixed metal oxides.

Boehmite peaks are no longer visible at temperatures greater than 500°C, while

sodalite peaks are no longer observed at 1000°C. Upon the disappearance of sodalite

peaks the identification of nepheline is observed, indicating that sodalite has

transformed to nepheline as described in Equation 24 for the DTG curves.

© 2013 Wentao Liang 74

Figure 19: X-ray diffraction patterns of thermally activated red mud

The physical properties of red mud and red mud thermally activated at 500 and

1000°C are provided in Table 9. The mean particle size of the red mud particles

increase as the temperature increases from 26.8µm for washed red mud up to 34.2µm

© 2013 Wentao Liang 75

for thermally activated red mud at 1000°C. Apart from the disappearance and

appearance of mineralogical phases in the XRD pattern, the sharpness of the most

intense bands (predominately hematite) increases as the temperature increases. Thus,

thermal activation improves the crystallinity of these phases. A similar observation

has been reported by Wu and Liu.[116] The surface area of red mud increases to

around 500°C (increase of around 30m2/g), however reduces to almost the same

surface as originally observed for the washed red mud sample. The increase in

surface area correlates with the decomposition (Figure 18) of the aluminium

hydroxide phases gibbsite and boehmite. It is proposed that the increase in surface

area is due to an increase in small pores forming as free water is removed during the

decomposition of gibbsite and boehmite. The reduction in surface area for red mud

activated at 1000°C is believed to be related to an increase in the pore diameter[118]

and/or the change in morphology, whereby an increase in the average size of

particles is observed due to agglomeration formation[116]. The report by Wu and

Liu,[116] presents a graph that shows the increase in average particle size up to

1350°C. This data shows a gradual increase in particle size up to 600°C, with

minimal size changes observed between 600 and 750°C, and a more rapid increase in

particle size up to 1200°C (38-39 µm).

© 2013 Wentao Liang 76

Table 9: Physical properties of red mud, red mud heated to 500 and 1000°C

pH BET (m2/g) Mean diameter of particles (µm)

Red mud 10.19 14.51 25.8

RM 500°C 10.21 44.01 31.6

RM 1000°C 9.30 17.36 34.2

Infrared spectra of red mud thermally activated at 500 and 1000°C clearly show the

decomposition of aluminium hydroxide species such as gibbsite and boehmite

(Figure 20). The multiple bands observed between 3650 and 3000cm-1

and at around

1600cm-1

decrease significantly for RM 500°C and are no longer visible at 1000°C.

The decomposition of these phases has been confirmed by XRD (Figure 19). The

intense Si(Al)-O stretching vibrational band centred at 969cm-1

is present in all

spectra, however the nitrate bands (1400-1475cm-1

) decrease in intensity for RM

500°C and are absent for RM 1000°C. These observations provide additional

evidence of the transformation of sodalite (Na8Al6Si6O24Cl2) to nepheline (NaAlSiO)

because both compounds will show intense Si(Al)-O stretching vibrations. The

disappearance of nitrate species indicates that the sodalite structure has nitrate in its

cages. After red mud had been heated to 1000°C more detailed information at the

lower wavenumber region could be deduced. These bands are believed to be

associated with the Fe-O and Fe-O-Fe bending modes of hematite. Bands at around

400 and 650cm-1

are characteristic of the A2u out-of-plane modes, while the in-plane

modes of hematite are located at around 525 and 400cm-1

.[120] The 414cm-1

band is

observed for all red mud samples, but at increased temperatures the other hematite

bands become visible; 678, 620 and 539cm-1

.

© 2013 Wentao Liang 77

Figure 20: Infrared spectra of thermally activated red mud

5.2.3 Removal of fluoride using thermally activated red mud

The effect of thermal activation is shown in Figures 21 and 22, whereby washed red

mud and red mud activated to 1000°C performed better than red mud activated at

© 2013 Wentao Liang 78

500°C. Figure 21 also shows the different pH regions and relative fluoride adsorption

potentials of the different red muds. At higher solution pH, red mud activated to

1000°C had the greatest fluoride adsorption but decreases as the pH increases. This is

also observed for the other red mud samples. RM1000°C shows a greater resistance

to pH increases, which has also been shown in Table 9 to only have a residual pH of

9.30 compared to 10.20. It is also proposed that RM1000°C performed better than

RM500°C due to an increase in the ratio of hematite to other mineralogical phases

(Figure 19 shows hematite as the primary phase), which increased the probability of

fluoride adsorbing to its surface.

Figure 21: Dependence of fluoride adsorption (mmol/g) with pH using a 100mg/L

fluoride solution with an initial pH of 4.

© 2013 Wentao Liang 79

Figure 22: Dependence of fluoride adsorption (mmol/g) with pH using a 100mg/L

fluoride solution with an initial pH of 8.

The main removal mechanism for fluoride is through its adsorption onto the surface

of iron oxide/hydroxides, and as such it is pH dependent. The best adsorption occurs

in the lower pH region. This pH dependency is observed in Figures 23 and 24. For a

fluoride solution of 100mg/L with an initial solution pH of 4, a maximum fluoride

adsorption of 3.73 mmol/g for RM1000°C is obtained, however this rapidly declines

as the pH increases as the solution becomes more alkaline. Above pH 7, the amount

of fluoride removed remains between 0.07 and 0.05 mmol/g. The fluoride solution

with an initial pH of 8 shows minimal fluoride removal, however this is not

unexpected due to the strong competition that hydroxide ions have for the iron

surface. A maximum adsorption of 0.07 mmol/g is observed (Figure 24). The study

by Çengeloğlu et al, [30] on the removal of fluoride using red mud showed a

© 2013 Wentao Liang 80

maximum removal at a pH of around 5.5. Deviations in results between the studies

are believed to be due to the different activation processes used.

Figure 23: Fluoride adsorption (mmol/g) using red mud thermally activated to

1000°C for a 100mg/L fluoride solution with an initial pH of 4.

© 2013 Wentao Liang 81

Figure 24: Fluoride adsorption (mmol/g) using red mud thermally activated to

1000°C for a 100mg/L fluoride solution with an initial pH of 8.

The removal mechanism of fluoride using red mud primarily involves neutral

(≡FeOH) and protonated (≡FeOH2+) sites on the oxide/hydroxide components (such

as hematite and gibbsite) when the pH is less than 7 [113]. The increases in

adsorption at lower pH are due to the protonation of the iron surface sites The

increases in adsorption at lower pH are due to the protonation of the iron surface sites

(25).

≡FeOH + H+ → ≡FeOH2

+ (25)

Hematite neutral sites only show promising removal efficiencies at low mass to

volume ratios. However, at higher mass to volume ratios the continual release of OH-

units as F- ions are adsorbed causes the pH to become more and more alkaline. This

© 2013 Wentao Liang 82

results in the fluoride adsorption process to be hindered due to increasing repulsive

forces between the negatively charged surface (≡FeO-) and fluoride ions (26).

≡FeOH + OH- → ≡FeO

- + H2O (26)

The relationship between pH and fluoride adsorption shows that for consistent

fluoride adsorption, a low, constant pH needs to be maintained to avoid any sudden

shocks to the surface adsorption sites. At pH levels less than 3, it is also highly

possible that AlFx complexes form, in particular aluminium trifluoride (AlF3). The

formation of AlF3 would account for some of the fluctuations observed in the

fluoride loading capacities. The formation of this phase is a result of gibbsite in red

mud reacting with HF that forms under these highly acidic conditions conditions

(27).

Al2O3 + 6HF → 2AlF3 + 3H2O (27)

© 2013 Wentao Liang 83

6CHAPTER 6: OVERALL CONCLUSIONS

Red mud is comprised of a number of mineralogical phases; however the most

important phases in the removal process are hematite and gibbsite. The treatment of

red mud with acid improved the adsorption properties of red mud in 2 ways: 1)

transformed ≡SOH / ≡SO- sites to ≡SOH2

+ and 2) increased the availability of metal

oxide/hydroxide sites through the removal of sodalite and cancrinite phases. In order

to achieve reasonable removal efficiencies for fluoride a pH < 4.5 needs to be

maintained, with sulphuric acid producing the best removal efficiencies. Red mud

treated with sulphuric acid gave the best removal efficiencies for fluoride due to 2

protons being available to protonate the surface hydroxyl groups. Sudden changes in

pH have shown to have negative effects on the removal efficiencies and thus need to

be controlled.

In the case of thermally activated red mud, the effect thermal activation has shown

washed red mud and red mud activated at 1000°C performed better than red mud

activated at 500°C. Meanwhile, at higher solution pH, red mud activated to 1000°C

had the greatest fluoride adsorption but decreases as the pH increases as observed for

the other red mud samples. RM1000°C shows a greater resistance to pH increases

once introduced to solution, but only have a residual pH of 9.30 compared to 10.20.

Hence, according the data analyse, keep in 1000°C and pH value around 4 can

achieve the greatest fluoride adsorption.

© 2013 Wentao Liang 84

7CHAPTER 7 FUTURE WORK

This study has shown that red mud as an adsorption material can be used to

successfully remove fluoride. Due to red mud is a by-product of Bayer Process,

brings high risks for environment, the research could be continually focused on

balancing thermally activated and acid treatment, trying to arrange a best point for

fluoride adsorption.

In this case, thermally activated should be treated as a single factor to test the

minerals changes of red mud, 100 °C to 1000°C temperatures are reasonable.

Meanwhile, the orders of treatment might have different results. Hence, acid

treatment is following thermally treated is another factor should be considered.

All the research should be expected that the red mud resources differences, as it

mentioned above, which will bring the different research story.

© 2013 Wentao Liang Page 85

Bibliography

1. Hind, A.R., S.K. Bhargava, and S.C. Grocott, The surface chemistry of Bayer process

solids: a review. Colloids and Surfaces a-Physicochemical and Engineering Aspects,

1999. 146(1-3): p. 359-374.

2. Fergusson, D.L., The Conversion and Sustainable Use of Alumina Refinery Residues:

Global Solution Examples, in Essential Readings in Light Metals. 2013, John Wiley &

Sons, Inc. p. 965-973.

3. Burton, E.A. and L.M. Walter, Relative precipitation rates of aragonite and Mg calcite

from seawater: Temperature or carbonate ion control? Geology, 1987. 15(2): p. 111-

114.

4. Calbeck, J.H. and H.R. Harner, Particle Size and Distribution by Sedimentation

Method1. Industrial & Engineering Chemistry, 1927. 19(1): p. 58-61.

5. Chvedov, D., S. Ostap, and T. Le, Surface properties of red mud particles from

potentiometric titration. Colloids and Surfaces A: Physicochemical and Engineering

Aspects, 2001. 182(1–3): p. 131-141.

6. Fitch, B., WHY PARTICLES SEPARATE IN SEDIMENTATION PROCESSES.

Industrial & Engineering Chemistry, 1962. 54(10): p. 44-51.

7. Palmer, S.J., R.L. Frost, and T. Nguyen, Hydrotalcites and their role in coordination of

anions in Bayer liquors: Anion binding in layered double hydroxides. Coordination

Chemistry Reviews, 2009. 253(1–2): p. 250-267.

8. Palmer, S.J., T. Nguyen, and R.L. Frost, Synthesis and Raman spectroscopic

characterisation of hydrotalcite with CO32− and VO3− anions in the interlayer. Journal

of Raman Spectroscopy, 2007. 38(12): p. 1602-1608.

9. Plummer, L.N. and E. Busenberg, The solubilities of calcite, aragonite and vaterite in

CO2-H2O solutions between 0 and 90°C, and an evaluation of the aqueous model for

the system CaCO3-CO2-H2O. Geochimica et Cosmochimica Acta, 1982. 46(6): p.

1011-1040.

10. Saylor, C.H., Calcite and Aragonite. The Journal of Physical Chemistry, 1927. 32(10):

p. 1441-1460.

11. Shirure, V.S., B.P. Nikhade, and V.G. Pangarkar, Intensification of Precipitation Using

Narrow Channel Reactors:  Case Study of Hydrotalcite Precipitation. Industrial &

Engineering Chemistry Research, 2006. 46(10): p. 3086-3094.

12. Shirure, V.S., A.S. Pore, and V.G. Pangarkar, Intensification of Precipitation Using

Narrow Channel Reactors:  Magnesium Hydroxide Precipitation. Industrial &

Engineering Chemistry Research, 2005. 44(15): p. 5500-5507.

13. Steinour, H.H., Rate of Sedimentation. Industrial & Engineering Chemistry, 1944. 36(9):

p. 840-847.

14. Ward, H.T. and K. Kammermeyer, SEDIMENTATION IN THE LABORATORY Design

Data from Laboratory Experimentation. Industrial & Engineering Chemistry, 1940.

32(5): p. 622-626.

15. Kumar, S., R. Kumar, and A. Bandopadhyay, Innovative methodologies for the

utilisation of wastes from metallurgical and allied industries. Resources, Conservation

and Recycling, 2006. 48(4): p. 301-314.

16. Wang, S., H.M. Ang, and M.O. Tade, Novel applications of red mud as coagulant,

adsorbent and catalyst for environmentally benign processes. Chemosphere, 2008.

72(11): p. 1621-35.

17. Walter, L.M. and E.A. Burton, The effect of orthophosphate on carbonate mineral

dissolution rates in seawater. Chemical Geology, 1986. 56(3–4): p. 313-323.

18. Szépvölgyi, J., A chemical engineer's view of the red mud disaster. Nachrichten aus der

Chemie, 2011. 59(5): p. V-VII.

19. Brecht, P., Mud, glorious mud. Nurs Stand, 1997. 11(45): p. 20-1.

© 2013 Wentao Liang Page 86

20. Liu, Y., C. Lin, and Y. Wu, Characterization of red mud derived from a combined Bayer

Process and bauxite calcination method. Journal of Hazardous materials, 2007. 146(1):

p. 255-261.

21. Lombi, E., et al., In situ fixation of metals in soils using bauxite residue: chemical

assessment. Environmental Pollution, 2002. 118(3): p. 435-443.

22. Thiyagarajan, C., et al., Plant-available manganese in bauxite residue sand amended

with compost and residue mud. Soil Research, 2012. 50(5): p. 416-423.

23. Genc, H., et al., Adsorption of arsenate from water using neutralized red mud. J Colloid

Interface Sci, 2003. 264(2): p. 327-34.

24. Genç-Fuhrman, H., H. Bregnhøj, and D. McConchie, Arsenate removal from water

using sand–red mud columns. Water Research, 2005. 39(13): p. 2944-2954.

25. Genc-Fuhrman, H., et al., Comparing the arsenic sorption capacity of Bauxsol-tm and

its derivatives with other sorbents, in Natural arsenic in groundwater. Occurrence,

remediation and management. 2005, Balkema Publishers, A.A. / Taylor & Francis The

Netherlands: Leiden, NL. p. 223-235.

26. Genç-Fuhrman, H., J.C. Tjell, and D. McConchie, Adsorption of arsenic from water

using activated neutralized red mud. Environmental science & technology, 2004. 38(8):

p. 2428-2434.

27. Genc-Fuhrman, H., J.C. Tjell, and D. McConchie, Increasing the arsenate adsorption

capacity of neutralized red mud (Bauxsol). J Colloid Interface Sci, 2004. 271(2): p. 313-

20.

28. Kandori, K., et al., Effects of silicate and phosphate ions on the formation of ferric oxide

hydroxide particles. Journal of Materials Science, 1992. 27(3): p. 719-728.

29. Huang, W., et al., Phosphate removal from wastewater using red mud. J Hazard Mater,

2008. 158(1): p. 35-42.

30. Çengeloğlu, Y., E. Kır, and M. Ersöz, Removal of fluoride from aqueous solution by

using red mud. Separation and Purification Technology, 2002. 28(1): p. 81-86.

31. Mohan, D., et al., Fluoride Removal from Water using Bio-Char, a Green Waste, Low-

Cost Adsorbent: Equilibrium Uptake and Sorption Dynamics Modeling. Industrial &

Engineering Chemistry Research, 2011. 51(2): p. 900-914.

32. Banks, D., et al., Natural concentrations of major and trace elements in some

Norwegian bedrock groundwaters. Applied Geochemistry, 1995. 10(1): p. 1-16.

33. Shen, F., et al., Electrochemical removal of fluoride ions from industrial wastewater.

Chemical Engineering Science, 2003. 58(3): p. 987-993.

34. Water, S. and T. Health, Guidelines for Drinking-water Quality, Volume 1 :

Recommendations (3rd Edition). 2003, Albany, NY, USA: World Health Organization

(WHO).

35. Bhatnagar, A., E. Kumar, and M. Sillanpää, Fluoride removal from water by

adsorption—A review. Chemical Engineering Journal, 2011. 171(3): p. 811-840.

36. Ayoob, S., A.K. Gupta, and V.T. Bhat, A Conceptual Overview on Sustainable

Technologies for the Defluoridation of Drinking Water. Critical Reviews in

Environmental Science and Technology, 2008. 38(6): p. 401.

37. Kumar, E., et al., Defluoridation from aqueous solutions by granular ferric hydroxide

(GFH). Water Res, 2009. 43(2): p. 490-8.

38. Tor, A. and Y. Cengeloglu, Removal of congo red from aqueous solution by adsorption

onto acid activated red mud. Journal of Hazardous Materials, 2006. 138(2): p. 409-415.

39. Kirke, E., Recent Trends in Bayer Precipitation Practice. Chem. Eng. Aust., ChE 7,

1982(1): p. 35-39.

40. Lee, M., et al. Characterization of aluminum trihydroxide crystals precipitated from

caustic solutions. in ACS symposium series. 1997. ACS Publications.

41. Survey, G., Mineral Commodity Summaries: 2012. 2012: Government Printing Office.

42. Addai-Mensah, J., Surface and structural characteristics of gibbsite precepitated from

pure, synthetic Bayer liquor. Minerals engineering, 1997. 10(1): p. 81-96.

43. Liu, Y., R. Naidu, and H. Ming, Red mud as an amendment for pollutants in solid and

liquid phases. Geoderma, 2011. 163(1): p. 1-12.

© 2013 Wentao Liang Page 87

44. Perrotta, A., Nanosized corundum synthesis. Material Research Innovations, 1998. 2(1):

p. 33-38.

45. Kumar, R., et al., Mechanical activation of bauxite—potential and prospects in the

Bayer process. Light Metals, 2005: p. 77-79.

46. Kumar, R., S. Srikanth, and P. Ramachamdrarao, Phase stability in the Fe2Or-Al2Or-

Na2CO3-C system and its implications in the processing of red mud. Metallurgy of

Nonferrous and Rare Metals, Edited by LI Leontiev, AI Kholkin and VV Belova,

Russian Academy of Sciences, Moscow, 2002: p. 137-153.

47. Gräfe, M. and C. Klauber, Bauxite residue issues: IV. Old obstacles and new pathways

for< i> in situ</i> residue bioremediation. Hydrometallurgy, 2011. 108(1): p. 46-59.

48. Paramguru, R., P. Rath, and V. Misra, Trends in red mud utilization–A review. Mineral

Processing & Extractive Metall. Rev., 2004. 26(1): p. 1-29.

49. Wang, S., H.M. Ang, and M.O. Tadé, Novel applications of red mud as coagulant,

adsorbent and catalyst for environmentally benign processes. Chemosphere, 2008.

72(11): p. 1621-1635.

50. Galarraga, R.A., et al. CVG-bauxilum red mud neutralization. in LIGHT METALS-

WARRENDALE-PROCEEDINGS-. 2002. TMS.

51. Kirkpatrick, D., Red mud product development. LIGHT METALS-WARRENDALE-,

1996: p. 75-80.

52. Goldstein, G. and R. Reimers, Trace element partitioning and bioavailability in red mud

synthetic freshwater sediment. LIGHT METALS-WARRENDALE-, 1999: p. 19-24.

53. Pilurzu, S., et al. New research proposals for utilization and disposal of bauxitic red mud

from Bayer process. in Minerals, Metals and Materials Society/AIME, REWAS'99:

Global Symposium on Recycling, Waste Treatment and Clean Technology. 1999.

54. Smirnov, D., T. Molchanova, and L. Vodolazov, Associated components of bauxites and

ways for their utilization. Tsvetnaya Metallurgia, 1999. 4: p. 31-34.

55. Balasubramanian, G., M. Nimje, and V. Kutumbarao, Conversion of aluminium industry

wastes into glass-ceramic products. V. Recycling of Metals and Engineered Materials,

Eds. D. Steward, R. Stephensen, JC Daley, TMS, 2000: p. 1223-1228.

56. Orešcanin, V., et al., Red mud and waste base: raw materials for coagulant production.

Journal of trace and microprobe techniques, 2001. 19(3): p. 419-428.

57. Li, Y., et al., Phosphate removal from aqueous solutions using raw and activated red

mud and fly ash. Journal of Hazardous Materials, 2006. 137(1): p. 374-383.

58. Liu, C.-j., et al., Adsorption removal of phosphate from aqueous solution by active red

mud. Journal of Environmental Sciences, 2007. 19(10): p. 1166-1170.

59. Huang, W., et al., Phosphate removal from wastewater using red mud. Journal of

Hazardous Materials, 2008. 158(1): p. 35-42.

60. Altundoğan, H.S. and F. Tümen, Removal of phosphates from aqueous solutions by

using bauxite. I: Effect of pH on the adsorption of various phosphates. Journal of

Chemical Technology and Biotechnology, 2002. 77(1): p. 77-85.

61. Cengeloglu, Y., et al., Removal of nitrate from aqueous solution by using red mud.

Separation and Purification Technology, 2006. 51(3): p. 374-378.

62. Altundoğan, H.S., et al., Arsenic adsorption from aqueous solutions by activated red

mud. Waste Management, 2002. 22(3): p. 357-363.

63. Soner Altundoğan, H., et al., Arsenic removal from aqueous solutions by adsorption on

red mud. Waste Management, 2000. 20(8): p. 761-767.

64. Castaldi, P., et al., XRD, FTIR, and thermal analysis of bauxite ore-processing waste

(red mud) exchanged with heavy metals. Clays and Clay Minerals, 2008. 56(4): p. 461-

469.

65. Babel, S. and T.A. Kurniawan, Low-cost adsorbents for heavy metals uptake from

contaminated water: a review. Journal of hazardous materials, 2003. 97(1): p. 219-243.

66. Salters, V.J. and J. Longhi, Trace element partitioning during the initial stages of

melting beneath mid-ocean ridges. Earth and Planetary Science Letters, 1999. 166(1): p.

15-30.

© 2013 Wentao Liang Page 88

67. Mohanty, S., et al., Removal of phosphorus from aqueous solution using alumized red

mud. International journal of environmental studies, 2004. 61(6): p. 687-697.

68. Yanjun, L., Zhang Zhaoyuan~ 2 Han Yuexin~ 1 (1. Northeast University; 2. Anshan

Mining Co. of Anshan Steel Group); Prospect and Application State of High Grade Iron

Concentrate [J]. Metal Mine, 2006. 11.

69. Apak, R., et al., Heavy metal cation retention by unconventional sorbents (red muds and

fly ashes). Water Research, 1998. 32(2): p. 430-440.

70. Cornell, R.M. and U. Schwertmann, The iron oxides: structure, properties, reactions,

occurrences and uses. 2003: Wiley. com.

71. Sglavo, V.M., et al., Bauxite ‘red mud’ in the ceramic industry. Part 1: thermal

behaviour. Journal of the European Ceramic Society, 2000. 20(3): p. 235-244.

72. Smiljanić, S., et al., Rinsed and thermally treated red mud sorbents for aqueous Ni<

sup> 2+</sup> ions. Chemical Engineering Journal, 2010. 162(1): p. 75-83.

73. Mohapatra, M., et al., Review of fluoride removal from drinking water. Journal of

Environmental Management, 2009. 91(1): p. 67-77.

74. Banks, D., et al., Natural concentrations of major and trace elements in some

Norwegian bedrock groundwaters. Appl. Geochem., 1995. 10(1): p. 1-16.

75. Shen, F., et al., Electrochemical removal of fluoride ions from industrial wastewater.

Chem. Eng. Sci., 2003. 58(3-6): p. 987-993.

76. Bhatnagar, A., E. Kumar, and M. Sillanpaeae, Fluoride removal from water by

adsorption. Chem. Eng. J. (Amsterdam, Neth.), 2011. 171(3): p. 811-840.

77. W.H.O., Guidelines for drinking water quality. 2004, Geneva.

78. Kumar, E., et al., Defluoridation from aqueous solutions by granular ferric hydroxide

(GFH). Water Research, 2009. 43(2): p. 490-498.

79. Tang, Y., et al., Fluoride adsorption onto granular ferric hydroxide: Effects of ionic

strength, pH, surface loading, and major co-existing anions. Journal of Hazardous

Materials, 2009. 171(1–3): p. 774-779.

80. Wajima, T., et al., Adsorption behavior of fluoride ions using a titanium hydroxide-

derived adsorbent. Desalination, 2009. 249(1): p. 323-330.

81. Tomar, V. and D. Kumar, A critical study on efficiency of different materials for fluoride

removal from aqueous media. Chemistry Central journal, 2013. 7(1): p. 51.

82. Ku, Y. and H.-M. Chiou, The Adsorption of Fluoride Ion from Aqueous Solution by

Activated Alumina. Water, Air, and Soil Pollution, 2002. 133(1-4): p. 349-361.

83. Daifullah, A.A.M., S.M. Yakout, and S.A. Elreefy, Adsorption of fluoride in aqueous

solutions using KMnO4-modified activated carbon derived from steam pyrolysis of rice

straw. Journal of Hazardous Materials, 2007. 147(1–2): p. 633-643.

84. Raul, P.K., et al., Removal of Fluoride from Water Using Iron Oxide-Hydroxide

Nanoparticles. Journal of Nanoscience and Nanotechnology, 2012. 12(5): p. 3922-3930.

85. Schwertmann, U. and R.M. Cornell, Iron oxides in the laboratory. 2008: John Wiley &

Sons.

86. Shaw, D.J. and B. Costello, Introduction to colloid and surface chemistry: Butterworth-

Heinemann, Oxford, 1991, ISBN 0 7506 1182 0, 306 pp,£ 14.95. 1993, Elsevier.

87. Langmuir, I., The adsorption of gases on plane surfaces of glass, mica and platinum.

Journal of the American Chemical society, 1918. 40(9): p. 1361-1403.

88. Brunauer, S., P.H. Emmett, and E. Teller, Adsorption of gases in multimolecular layers.

Journal of the American Chemical Society, 1938. 60(2): p. 309-319.

89. Do Duong, D., Absorption Analysis: Equilibria and Kinetics. Vol. 2. 1998: Imperial

College Pr.

90. Allen, T., Particle Size Measurement: Volume 2: Surface Area and Pore Size

Determination. Vol. 2. 1997: Springer.

91. Gregg, S. and K.S. Sing, Adsorption, Surface Area, and Porosity. 1983.

92. Pierotti, R. and J. Rouquerol, Reporting physisorption data for gas/solid systems with

special reference to the determination of surface area and porosity. Pure Appl Chem,

1985. 57(4): p. 603-619.

© 2013 Wentao Liang Page 89

93. Barrett, E.P., L.G. Joyner, and P.P. Halenda, The determination of pore volume and area

distributions in porous substances. I. Computations from nitrogen isotherms. Journal of

the American Chemical society, 1951. 73(1): p. 373-380.

94. Ravikovitch, P.I., et al., Unified approach to pore size characterization of microporous

carbonaceous materials from N2, Ar, and CO2 adsorption isotherms. Langmuir, 2000.

16(5): p. 2311-2320.

95. Dubinin, M., The potential theory of adsorption of gases and vapors for adsorbents with

energetically nonuniform surfaces. Chemical Reviews, 1960. 60(2): p. 235-241.

96. Dubinin, M. and H. Stoeckli, Homogeneous and heterogeneous micropore structures in

carbonaceous adsorbents. Journal of Colloid and Interface Science, 1980. 75(1): p. 34-

42.

97. Kawazoe, H., Method for the calculation of effective pore size distribution in molecular

sieve carbon. Journal of Chemical Engineering of Japan, 1983. 16(6): p. 470-475.

98. Dombrowski, R.J., D.R. Hyduke, and C.M. Lastoskie, Pore size analysis of activated

carbons from argon and nitrogen porosimetry using density functional theory.

Langmuir, 2000. 16(11): p. 5041-5050.

99. Olivier, J.P., Modeling physical adsorption on porous and nonporous solids using

density functional theory. Journal of porous materials, 1995. 2(1): p. 9-17.

100. Feeley, J., et al., Determination of surface properties and flow characteristics of

salbutamol sulphate, before and after micronisation. International journal of

pharmaceutics, 1998. 172(1): p. 89-96.

101. Olivier, J.P., The determination of surface energetic heterogeneity using model

isotherms calculated by density functional theory, in Fundamentals of Adsorption. 1996,

Springer. p. 699-707.

102. Grafe, M., G. Power, and C. Klauber, Review of bauxite residue alkalinity and

associated chemistry, C. Minerals, Editor. 2009: Karawara, WA, Australia.

103. Cengeloglu, Y., E. Kir, and M. Ersoz, Removal of fluoride from aqueous solution by

using red mud. Sep. Purif. Technol., 2002. 28(1): p. 81-86.

104. Fortin, S. and R. Breault. Bayer process heat exchangers cleaning efficiency: Optimizing

the acid dissolution of sodalite scale. in Engineering Conferences International. 2003.

Sante Fe, New Mexico.

105. Farmer, V.C. and Editor, Mineralogical Society Monograph 4: The Infrared Spectra of

Minerals. 1974: Mineral. Soc. 539 pp.

106. Palmer, S.J., B.J. Reddy, and R.L. Frost, Characterisation of red mud by UV-vis-NIR

spectroscopy. Spectrochim. Acta, Part A, 2009. 71A: p. 1814-1818.

107. Balan, E., et al., Surface modes in the infrared spectrum of hydrous minerals: the OH

stretching modes of bayerite. Phys. Chem. Miner., 2008. 35(5): p. 279-285.

108. Rochester, C.H. and S.A. Topham, Infrared study of surface hydroxyl groups on

hematite. J. Chem. Soc., Faraday Trans. 1, 1979. 75(5): p. 1073-88.

109. Castaldi, P., et al., Study of sorption processes and FT-IR analysis of arsenate sorbed

onto red muds (a bauxite ore processing waste). Journal of Hazardous Materials, 2010.

175(1–3): p. 172-178.

110. Castaldi, P., et al., XRD, FTIR, and thermal analysis of bauxite ore-processing waste

(red mud) exchanged with heavy metals. Clays Clay Miner., 2008. 56(4): p. 461-469.

111. Ross, M.W. and T.C. DeVore. Desorption of nitric acid from boehmite and gibbsite.

2008. American Chemical Society.

112. Miller, F.A. and C.H. Wilkins, Infrared spectra and characteristic frequencies of

inorganic ions. Anal. Chem., 1952. 24: p. 1253-94.

113. Raul, P.K., et al., Removal of fluoride from water using iron oxide-hydroxide

nanoparticles. J. Nanosci. Nanotechnol., 2012. 12(5): p. 3922-3930.

114. Gräfe, M., G. Power, and C. Klauber, Review of bauxite residue alkalinity and

associated chemistry. 2009, CSIRO Document DMR-3610.

115. Chen, Q. and W. Zeng, Calorimetric determination of the standard enthalpies of

formation of gibbsite, Al(OH)3(cr), and boehmite, AlOOH(cr). Geochim. Cosmochim.

Acta, 1996. 60(1): p. 1-5.

© 2013 Wentao Liang Page 90

116. Wu, C.-S. and D.-Y. Liu, Mineral phase and physical properties of red mud calcined at

different temperatures. J. Nanomater., 2012: p. 628592, 6 pp.

117. Nickels, D.L., M.E. Sawayda, and T.E. Weyand, Production of high purity magnetite by

reduction of hematite. 2007, Pittsburgh Iron Oxides, LLC, USA . p. 23 pp.

118. Tong, W., et al., Effects of surface properties of red mud on interactions with

Escherichia coli. J. Mater. Res., 2013. 28(17): p. 2332-2338.

119. Stormer, J.C., Jr. and I.S.E. Carmichael, Free energy of sodalite and the behavior of

chloride, fluoride, and sulfate in silicate magmas. Amer. Mineral., 1971. 56(1-2): p. 294-

308.

120. Rendon, J.L. and C.J. Serna, IR spectra of powder hematite: effects of particle size and

shape. Clay Miner., 1981. 16(4): p. 375-84.