Characterization of Some Natural and Synthetic Materials with Silicate Structures

Home , Hydrophobic silica, Tetraethyl orthosilicate

ISSN: 1402-1757 ISBN 978-91-7439-XXX-X Se i listan och fyll i siffror där kryssen är

LICENTIATE T H E SIS

Department of Engineering Sciences and Mathematics Division of Materials Science Edwin Escalera Mejia Characterization of Some Natural and Synthetic Materials of Some Natural and Synthetic Materials Edwin Escalera Mejia Characterization With Silicate Structures

ISSN: 1402-1757 ISBN 978-91-7439-553-2 Characterization of Some Natural and

Luleå University of Technology 2013 Synthetic Materials With Silicate Structures

Edwin Escalera Mejia

Characterization of Some Natural and Synthetic Materials With Silicate Structures

Edwin Escalera Mejia

Luleå University of Technology Department of Engineering Sciences and Mathematics Division of Materials Science Printed by Universitetstryckeriet, Luleå 2013

ISSN: 1402-1757 ISBN 978-91-7439-553-2 Luleå 2013 www.ltu.se Abstract

The present thesis deals with characterization of silicate structures with a determined morphology and structure such as ordered mesoporous silica and layered silicates. Mesoporous silica groups are amorphous solids exhibiting highly ordered pore structures with narrow pore size distributions and large surface areas. Porous materials are used in various applications such as in adsorption, separation, catalysis, molds for templating, etc.

Another interesting group of layered materials are crystal silicates with minerals of natural origin. The silicates have a structure that consists of stacked layers in which planes of oxygen atoms coordinate to cations such as Si4+, Al3+, Mg2+, Fe3+ to form two dimensional sheets. The coordination of cations in adjacent sheets typically alternates between tetrahedral and octahedral. The properties and uses of the clays vary widely due to the differences in their structure and composition. Some important applications are paints, adsorption, intercalation, removal of pollutants from water and in ceramic industry.

The thesis consists of two parts. In the first study characterization of synthesized and functionalized ordered mesoporous silica were performed. Mesoporous silica with a large surface area on which organic functional groups are grafted was used to synthesize cobalt nanoparticles. Investigation by SEM and TEM showed hexagonal particles, with a pore size about 10 nm. The functionalization of the silica was studied by FTIR and TG/DTA techniques and the obtained nanoparticles were characterized by XRD, TEM and EDX analysis.

In the second study, an extended literature review on properties of clays is presented. Samples from three different clay deposits, Ivirgarzama (IC), Entre Rios (EC) and Uspha-Uspha (U) from Bolivia were characterized by different experimental techniques in order to assess their relevant features.

The chemical and mineralogical analysis showed that the clays consist mainly of kaolinite and illite along with quartz in different amounts. Also, certain amounts of feldspar, iron and magnesium are present in the clays and with predominance in the EC clay.

Thermal analysis (DSC/TG and dilatometer) and XRD were used to study the phase transformations and their microstructural evolution at sintering. The EC clay with a high alkali and iron content influenced both the onset of liquid formation and the onset of sintering. Mullite is a crystalline phase that strengthens the ceramics and it was formed in all the studied clays.

Based on these results, the EC and U clays provide required characteristics that enable them for use in the fabrication of products with red tonality, especially bricks, roofing tiles and rustic floor tiles. The IC clay with relatively low iron content and with relatively good refractoriness can be used for production of firebricks and also for partially replacing kaolin and silica in white firing ceramics. Thus, the clays from Ivirgarzama, Entre Rios and Uspha- Uspha are promising raw materials and they should be considered as valuable resources for the production of building ceramics.

iii

Preface

This licentiate thesis is a part of my PhD studies carried out at the division of Materials Science at Luleå University of Technology, Sweden.

The aim of the thesis was to characterize synthetic mesoporous silicates and some natural clays from Bolivia, which can be suitable as building materials.

The thesis is divided in two parts. The first study deals with the synthesis of mesoporous silica materials used as hard templates for synthesis of cobalt nanoparticles.

The second part is about characterisation of natural clay minerals from Bolivia and their thermal behaviour.

The thesis is compiled of the following papers:

Synthesis of homogeneously dispersed cobalt nanoparticles in the pores of functionalized SBA-15 silica E. Escalera, M. A. Ballem, J. M. Córdoba, M-L. Antti and M. Odén Powder Technology 221 (2012) 359-364.

High temperature phase transformation in Bolivian kaolinitic-illitic clays E. Escalera, R. Tegman, M-L. Antti and M. Odén Submitted to Applied Clay Science (2013).

Acknowledgements

First of all, I would like to express my deep gratitude to my supervisor, Associate Professor Marta-Lena Antti, and my co-supervisor Professor Magnus Odén for their guidance and valuable ideas.

I wish to express my gratitude to Dr. Ragnar Tegman, for his knowledge, good suggestions, and also for all discussions concerning to the second paper.

I would like to thank Johnny Grahn for helping using the SEM instrument.

Likewise, I want to thank all my colleagues at the Division of Materials Science.

I want to thank Roberto Soto S., Coordinator of Project UMSS-ASDI-10 in Bolivia. And also I want to thank my colleagues at Chemistry-Department, FCyT-UMSS, for all good times shared.

I am deeply indebted to my parents for their unconditional love, encouragement and support.

I acknowledge the Swedish International Development Cooperation Agency, SIDA, for financial support for this project - Non Metallic Minerals as Resources for Development of Poor Bolivian Regions.

vii

PART ONE

Mesoporous materials SBA-15 and its application in the synthesis of cobalt nanoparticles

Contents 1. Introduction ...... 2 1.1 Mesoporous silica ...... 2 1.2 Uses and applications ...... 3 2. Theoretical Background ...... 4 2.1 Mesoporous silica SBA-15 ...... 4 2.1.1 Surfactants and silica precursors ...... 4 2.2 Synthesis of mesoporous silica ...... 5 2.2.1 Formation ...... 5 2.2.2 Hydrothermal treatment ...... 6 2.2.3 Removal of surfactants ...... 6 2.3 Functionalization of mesoporous silica ...... 7 2.4 Metal incorporation in functionalized mesoporous silica ...... 8 3. Materials and Methods ...... 9 3.1 Materials ...... 9 3.2 Experimental procedure ...... 9 3.3 Characterization methods ...... 10 3.3.1 Scanning Electron Microscopy (SEM) ...... 10 3.3.2 Fourier Transform Infrared Spectroscopy (FTIR) ...... 10 3.3.3 Termogravimetry and Differential Thermal Analysis (TG/DTA) ...... 10 3.3.4 Nitrogen adsorption/desorption isotherms...... 10 3.3.5 Transmission Electron Microscopy (TEM) ...... 10 3.3.6 X-ray diffraction (XRD) ...... 11 4. Summary of results ...... 12 5. Conclusions ...... 14 References ...... 15

viii

PART TWO

Characterization of Bolivian clay minerals Contents 1. Introduction ...... 20 1.1 Ceramics in Bolivia ...... 21 2. Theoretical background ...... 23 2.1 Clay minerals ...... 23 2.1.1 Main clay minerals groups ...... 23 2.1.2 Associated minerals of clays ...... 25 2.2 Decomposition and phase transformations ...... 27 2.2.1 Metakaolinite and spinel phases ...... 27 2.2.2 Mullite phase ...... 28 2.3 Sintering ...... 28 3. Materials and methods ...... 30 3.1 Raw materials ...... 30 3.2 Characterization methods ...... 31 3.2.1 Mineralogical and phase analysis by XRD ...... 31 3.2.2 Chemical analysis by ICP-AES ...... 31 3.2.3 Thermogravimetry and Differential Scanning Calorimetry analysis (TG/DSC) ...... 31 3.2.4 Dilatometry analysis (DIL) ...... 31 3.2.5 Bulk density and open porosity ...... 31 3.2.6 Scanning Electron Microscopy (SEM) ...... 32 4. Results and discussions ...... 33 4.1 Mineralogical analysis of raw clays (XRD) ...... 34 4.2 Chemical analysis of raw clays (ICP-AES) ...... 35 4.3 Morphology of clays (SEM) ...... 37 4.4 Thermal analysis ...... 38 4.4.1 Thermogravimetry - Differential thermal analysis (TG/DSC) ...... 38 4.4.2 Dilatometry analysis (DIL) ...... 40 4.4.3 Bulk density and open porosity ...... 40 4.5 Microstructural evolution analysis of fired samples (XRD) ...... 41 4.6 Microstructure of fired samples (SEM) ...... 43 5. Conclusions ...... 45 Future work ...... 46 References ...... 47

PART ONE

MESOPOROUS MATERIALS SBA-15 AND ITS APPLICATION IN THE SYNTHESIS OF COBALT NANOPARTICLES

Part one of the thesis contains the synthesis, surface modification and characterization of mesoporous silica, SBA-15, with two-dimensional hexagonal arrangements. The obtained mesoporous material was then used as hard template for synthesizing cobalt nanoparticles.

Scope and objectives of part one In this study, the synthesis of the mesoporous silica SBA-15 and the subsequent surface modifications are presented. Both external and internal walls of the synthesized silica were modified through incorporation of organosilane groups, in order to enhance the synthesis of cobalt nanoparticles.

The objectives are to: - Synthesize mesoporous silica SBA-15. - Modify internal and external surfaces of mesoporous silica. - Synthesize cobalt nanoparticles.

The strategy used here is to synthesize nanoparticles with narrow size distribution using functionalized mesoporous silica as hard template.

1. Introduction

In the past few decades, the increasing routine use of advanced structural materials with defined and controlled porosity has led to deeper knowledge in the field of porous solids.

Nowadays, porous materials such as porous carbon, synthetic silicate zeolites, mesoporous silicates and ordered porous metal oxides are currently being studied for a large range of applications. Examples of applications are catalysts (Liu et al., 2005), new media for pollutant removal in air and water as well as in fuel production, and gas storage materials for energy technologies (Takagi et al., 2004) (Lee et al., 1999).

Porous materials are defined as solids containing pores i.e. voids, channels or interstices. Pore architectures such as size, shape, connectivity and the nature of the pore distribution, in combination with the chemical characteristics of the pore walls determine the properties and hence the possible applications for such materials.

The pores can be classified in closed and open pores, according to their accessibility to surroundings. Materials containing closed pores are mainly used for thermal and sonic insulation due to that they are completely isolated from their surroundings (Zdravkov et al., 2007). In contrast, the open pores have connectivity in between them which makes materials with open porosity suitable for adsorption, filters, catalysis, etc. Another classification of pores is based on the pore geometry. Pores can have different shapes such as spherical or cylindrical and they can be arranged in varying structures.

Materials with high open porosity normally have a large available surface area compared to materials with no or closed porosity. The porosity is the ratio of the pore volume to the total volume of the material.

The size of the pores in inorganic materials may range from the nano-scale to the macro-scale. According to the International Union of Pure and Applied Chemistry (IUPAC) porous materials can be classified into three classes based on their pore diameter (d), microporous d < 2 nm, mesoporous 2 ≤ d ≤ 50 nm and macroporous d > 50 nm (Sing et al., 1985) (Rouquerol et al., 1994).

1.1 Mesoporous silica Inorganic mesoporous materials such as mesoporous silica is one of the most investigated materials due to many applications in industrial fields such as filters and catalyst supports, as hard template for nanocasting of oxide nanoparticles (Yang and Zhao, 2005) and also in biochemical applications such as drug delivery system (Giri et al., 2007).

Ordered mesoporous silica may be readily synthesized under a wide range of pH from acidic to basic conditions, and also using cationic, anionic and neutral surfactants as well as a variety of commercially available copolymers (Muth et al., 2001).

An important type of ordered mesoporous materials (M41S) were discovered by scientists at Mobil Oil Corporation, who demonstrated remarkable features of this novel type of silica, and this opened up a new field of research (Beck et al., 1992). The M41S mesoporous family are often referred to as MCM materials. The most common MCM material is MCM-41which stands for Mobil Composition of Matter No. 41. MCM-41shows a highly ordered hexagonal array with a very narrow pore size distribution.

As shown in Figure 1.1, a variety of pore structures of this type of mesoporous silica can be synthesized, such as MCM-41with two-dimensional hexagonal alignment of mesoporous channels, MCM-48 with three-dimensional cubic order, and the layered material MCM-50 (Kresge et al., 1992).

a) Hexagonal b) Cubic c) Lamellar Figure 1.1 Structures of mesoporous silica materials.

Another common mesoporous silica (SBA) was discovered by Zhao et al., (1998). Mesoporous SBA which stands for Santa Barbara amorphous were first synthesized using triblock copolymers as surfactant. Since then, a variety of mesoporous SBA have been synthesized, such as SBA-15 which represents a two-dimensional hexagonal structure (Figure 1.1a) and SBA-16 with three-dimensional cubic structure (Figure 1.1b), etc.

Other families of mesoporous silica have also been reported in the literature, such as TDU-1 (Technische Universiteit Delft), first reported in 2001 by Maschmeyer et al., (2001), KIT, FDU and AMS (Fan et al., 2003), where the surfactants and synthesis conditions are variable.

1.2 Uses and applications Since the discovery of M41S and later SBA ordered mesoporous materials, there has been an increasing interest in the tailoring of this materials for many potential applications such as molecular sieves, drug delivery systems (Song et al., 2005), catalysis and for use as meso- reactors, adsorption and separation of biomolecules, host-guest chemistry, templates and as electrodes in solid-state ionic devices (Ishizaki et al., 1998). Morphology, pore size and surface area of the mesoporous silica can be tuned in many ways. This can be achieved by addition of salts (Qiao et al., 2006), co-surfactants (Li et al., 2006), oil (Lettow et al., 2000), solvents such as heptane as swelling agents and also by changing the reaction conditions (Johansson et al., 2010).

This wide range of applications are due to the unique structure of the materials which exhibits a regular array of uniform pore openings, uniform pores with narrow pore size distribution, high surface area and large pore volumes.

The application of mesoporous silica has been extended for use as hard templates for the synthesizing of metal nanoparticles, nanowires, nanorods, etc. with various applied approaches of synthesizing (Wu et al., 2006). By functionalization, the properties of mesoporous silica can be finely tuned by changing the organic groups on the surface making them suitable material for instance for metal adsorption (Aguado et al., 2009).

2. Theoretical Background

This chapter presents a literature review about mesoporous silica, functionalization of the external and internal walls and the incorporation of transition metal into the silica pores, which is presented in paper 1.

2.1 Mesoporous silica SBA-15 Since the discovery of family SBA mesoporous silica, there has been an increasing interest of these materials for use in many potential applications. This is due to their uniform pores, large surface area and high pore volume, which ensure their application in separation and adsorption processes, catalysis and for use as templates for synthesizing nanoparticles, nanowires, etc. There is a number of different types of mesoporous silica SBA reported, such as SBA-12 with a 3-dimensional hexagonal structure, SBA-11 with a cubic structure, and SBA-1, SBA-2, SBA-3, among others. The structure of mesoporous silica depends on the surfactant used (Kim and Ryoo, 1999).

One of the most studied mesoporous silica is SBA-15 (Santa Barbara Amorphous No. 15), with 2-dimentional hexagonal structure, space group (p6mm), which can be synthesized in large quantities from tetraethyl ortosilicate (TEOS) in the presence of tri-block copolymer and strong acidic media (Zhao et al., 1998). Synthesis of mesoporous silica is based on the well- known sol-gel process (Hench et al., 1990). Therefore surfactants, silica precursors, hydrothermal treatment, and removal of the surfactants are needed to form the final mesoporous material.

2.1.1 Surfactants and silica precursors Surfactants are known as structure directing agents. Surfactants are amphiphilic molecules that are composed by hydrophilic and hydrophobic parts. The hydrophobic part is often a hydrocarbon chain (Fröba et al., 2006).

The surfactants are classified by their head group. They can be anionic, cationic, non-ionic and amphoteric surfactants. Syntheses of M41S mesoporous material employ cationic alkylammonium surfactants and cethyltrimethyl ammonium bromide (CTA+Br-). It is an example of such a cationic surfactant commonly used to synthesize MCM-48 (Kresge et al., 1992).

The most common non-ionic surfactant used to synthesize several SBA mesoporous materials such as SBA-15, SBA-16 and SBA-12, is the triblock copolymer family which is commonly called Pluronics. There exist several different pluronics (EOxPOyEOx) with different molecular weights such as F108 (EO133PO50EO133), F127 (EO106PO70EO106), and P123 (EO20PO70EO20). The non-ionic surfactants consist of hydrophilic part of poly-ethylene oxide chains (EO) and hydrophobic part of poly-propylene oxide chains (PO). The initial letter P refers to paste and F refers to flakes.

The silica precursor is another essential component in the synthesis of mesoporous silica. Several types of silica precursors can be used for synthesis of mesoporous silica. The most common used are alkoxides, which are hydrophobic molecules such as tetramethyl orthosilicate (TMOS) or tetraethyl orthosilicate (TEOS). Another alternative silica precursor could be sodium silicate which is cheaper and it is often used in combination with other alkoxides (Matos et al., 2002).

2.2 Synthesis of mesoporous silica The synthesis of mesoporous silica involves three steps. The first step is the formation of the mesoporous structure using surfactants and silica precursors. The second step is the hydrothermal treatment at moderate temperatures. The final step is removal of surfactants from the mesoporous silica in which various techniques can be applied.

2.2.1 Formation The micelle formation begins when the surfactant is dispersed in an aqueous solution due to the interactions between themselves. The micelles consist of a hydrophobic (PO) core surrounded by hydrophilic (EO) chains which form a corona around the core (see Figure 2.1b). a) b) Pluronic: P123 EO PO EO

Silicon source: TEOS

Figure 2.1 Chemical structures of (a) surfactant P123 and the silica precursor TEOS. (b) Micelle formation showing hydrophobic core (PO) and hydrophilic corona (EO).

The formation of micelles is determined by the nature of the surfactant and conditions in the solution such as concentration of surfactants and temperature of the solution. When the silica precursor is added to the solution containing micelles, it hydrolyses and the silica network is formed. The transition from micelles to gel is gradually making the micelles become elongated, and this is known as polymerization of the silica (Fröba et al., 2006).

Two possible mechanisms have been proposed for the formation of these materials: a) True liquid-crystal templating (TLCT) and b) Cooperative self-assembly of the P123 and TEOS that together can develop a liquid-crystal templating phase with hexagonal arrangement, (see Figure 2.2) (Kresge et al., 1992).

Figure 2.2 Mechanism of the formation of mesoporous silica by surfactans. a) True liquid crystal templating, b) cooperative self-assembly.

In the TLCT mechanism it is observed that the concentration of the P123 is relatively high and that under the right conditions of temperature and pH a lyotropic liquid-crystalline phase is formed without requiring the presence of TEOS (Attard et al., 1997). In this case the polymerization process begins in the core/corona region interface. In addition the polymerization is simultaneous to the elongation of the micelles.

2.2.2 Hydrothermal treatment The hydrothermal treatment is a good way of tuning the properties in terms of pore size, micropore volume and surface area of the synthesized mesoporous silica. These properties are dependent of time and temperature of the hydrothermal treatment (Liu et al., 2008).

The hydrothermal treatment begins when the formation step of the mesoporous silica is finished. By increasing the hydrothermal treatment temperature for instance, there is an increased pore size, reduced microporosity and as consequence also reduced surface area. Similar effect, but not as pronounced, is obtained by increasing the hydrothermal treatment time. It was also pointed out that the hydrothermal treatment decreases the shrinkage of the silica walls upon calcination. This is an advantage in order to get large pores for determined applications such as functionalization and metal incorporation into the formed channels.

2.2.3 Removal of surfactants The final step of the synthesis is the removal of surfactants. Surfactants are often removed by calcination, but there are alternatives such as chemical removal or decomposition by microwave activation and digestion using acids (Gallis et al., 2001).

Normally the calcination is carried out in oxidizing conditions by increasing the temperature from room temperature to 500 °C, and holding time approximately 6 hours to decompose the surfactant P123 completely (Yamada et al., 2002). The removal of surfactants by calcination produces mesoporous silica with narrow pore size distributions and highly ordered mesostructures.

In addition, the surfactants can also be removed by wet chemical oxidation using hydrogen peroxide (Yang et al., 2005), sulphuric acid, hydrochloric acid or perchlorates under acidic conditions (Cai and Zhao, 2009). Also, mixtures can be used such as hydrochloric acid and ethanol to remove surfactants. Thereby only the surfactants are removed and the pore size distribution of the mesoporous silica structure remains intact.

2.3 Functionalization of mesoporous silica Functionalization is the addition of functional groups onto the surface of a material by chemical methods. Organic functionalization of mesoporous silica permits precise control over the surface properties and pore size for specific applications, such as chromatography, catalysis and adsorption.

Functionalization is employed for surface modification of mesoporous silica in order to achieve desired surface properties such as water repellent coatings, (hydrophobicity) (Rao et al., 2007). On the other hand, the surface modification can be used to generate a monolayer of charged groups into the pore surface and it can facilitate a uniform distribution of ion- exchanged metal precursors into the channels of mesoporous silica SBA-15 (Yang et al., 2003).

The physical properties of functionalized silica can vary within a wide range depending on the nature of the silylating agent used. There are a large number of functional groups that can be used for different applications in various materials. The most common groups of silanes to functionalize mesoporous silica are alkoxysilanes and alkylsilanes families. The incorporation of the functional groups into the mesoporous silica can be obtained either during the synthesis (co-condensation) or after the synthesis (grafting) (Zhu et al., 2002).

The functionalization process by grafting is a method based on slow hydrolysis of organic functional groups and condensation of the compounds with free OH-binding sites at the silica surface, thus forming new covalent -Si-O-Si- bonds (Rao et al., 2007).

Equation (1) shows the reaction between silanol group and organosilane agent.

4-n(≡Si–OH) + (R’O)4-nSiRn → (≡Si–O)4-n-Si-Rn + (4-n)R’OH (1) silica surface organosilane agent surface modified silica alcohol

The reaction shows the substitution of the hydrogen in OH groups by replacement of the organic functional group of type (R’O)4-nSiRn where (R’O) is a hydrolyzable group, such as methoxy, ethoxy or acetoxy, and R is an organic functional group such as alkyl, amino, etc.

The co-condensation method is an alternative that directly incorporates functional groups into the mesoporous silica simultaneously with the synthesis of mesoporous silica. It is carried out by the co-condensation of tetraalkoxysilanes such as TEOS and TMOS with trialkoxyorganosilanes of the type (R’O)3SiR in the presence of surfactants, leading to mesoporous silica with organic anchored covalently to the pore walls. Typically, by grafting method the hydrophilic silica surface can be modified to hydrophobic surface. This process is carried out by reaction of chlorosilanes ClSiR3 with the free silanol groups on the surface of the silica (Hitzky and de Juan, 2000). However, various silylating agents can be used such as mono-, di- alkyl and tri alkyl. The most conveniently used are those with tri alkyl silylating agents, due to high surface modification. Although, a more hydrophobic silica surface has been achieved using tri alkyl agents, compared with those achieved after use of mono and di alkyl agents (Rao et al., 2007).

Both grafting and co-condensation methods can be used to functionalize the internal pores of mesoporous silica. However, the co-condensation method has a number of disadvantages. In general, the degree of mesoscopic order of the mesostructure decreases with increasing concentration of organosilane groups in the reaction mixture, which leads to totally disordered mesostructure. However, after applying the grafting method the surface silica remains intact when organosilane groups are grafted to the silica walls. Another disadvantage of the co- condensation method is that care must be taken not to destroy the organic functionality during removal of the surfactant, thus only extractive methods should be used instead such as calcination at elevated temperatures (Fröba et al., 2006).

2.4 Metal incorporation in functionalized mesoporous silica Nanostructured materials represent a transition between the individual molecules and bulk solids. The synthesis of metals and metal oxide nanoparticles has received increased attention due to their unique physiochemical properties which make them desirable in many technological applications such as optics, magnets, electronics, catalysis, sensors (Belkacem et al., 2008), and hydrogenation processes (Takagi et al., 2004).

Within the past years, several techniques have been developed for synthesizing of nanoparticles. One of the most common routes used is based on introducing a suitable metal or metal oxide precursor into a selected template. Different kinds of mesoporous silica can be used as templates to synthesized nanoparticles. The most frequently used mesoporous silica is SBA-15 due to pore size larger than other mesoporous materials such as MCM-41and FDU-1. For instance, metal nanoparticles of Ag, Pt, Au, Pd, nanowires of Cu, Ni (Lin et al., 2008), and also alloys of Pt-Au and Au-Ag have been synthesized in the channels of the mesoporous silica (Liu et al., 2008). Also by using SBA-15 silica as template different oxides such as Fe2O3, Co3O4 and In2O3 have been synthesized (Fuertes, 2004).

3. Materials and Methods 3.1 Materials Pluronic P123 (EO20PO70EO20, Aldrich), tetraethyl orthosilicate (TEOS) (98%, Aldrich) and hydrochloric acid (37%, p.a., Fluka) were used for the synthesis of mesoporous silica SBA- 15. Trimethyl-chlorosilane (TMCS ≥ 99%, Aldrich), 3-aminopropyl-trimethoxysilane (APTMS 97%, Aldrich) and toluene (anhydrous 99.9%) were used for external and internal functionalization of the silica walls. Cobalt (II) sulfate heptahydrate (CoSO4·7H2O, 99%, Aldrich) was used as source of cobalt and sodium borohydride (NaBH4, 99%, Aldrich) for the chemical reduction process. Sodium hydroxide (purity ≥ 97%, p.a., Fluka), was used to dissolve mesoporous silica.

3.2 Experimental procedure For the synthesis of mesoporous silica SBA-15 using surfactant P123 as a structure directing agent and TEOS as a silica source a detailed procedure was used developed by Sayari et al., (2004).

The schematic illustration of the synthesis of mesoporous silica SBA-15 and the sequence step followed in this work to obtain cobalt nanoparticles is shown in Figure 3.1 (I) and (II), respectively.

(I) (II)

Pluronic (P123) (a) Mesoporous silica template (as-SBA-15)

∞ HCl sol. (b) External surface functionalization (TMCS) TEOS

∞ P123 + (c) Removal template HCl (P123)

Aging 100˚C (d) Internal surface functionalization 24h (APTMS)

Filtration and drying (e) Metal incorporation and chemical deposition Mesoporous silica (as-SBA-15)

Figure 3.1 Schematic illustrations: (I) Synthesis of mesoporous silica SBA-15 and (II), gradual synthesis procedure of the cobalt nanoparticles.

The synthesized mesoporous silica (as-SBA-15) (a), has been used to follow the sequence of surface functionalization of the silica walls as illustrated in Figure 3.1(II). These sequence steps include the grafting of alkyl hydrophobic (-Si(CH3)3) groups on the external silica surface (b), extraction of the surfactant P123 by calcination at 300 °C for 5 hours (c), and grafting of amine (-Si(CH2)3-NH2) groups inside the pores (d).

The obtained functionalized mesoporous silica was impregnated with cobalt sulphate aqueous solution and subsequent chemical reduction by sodium borohydride aqueous solution (e). Detailed information about the functionalization and metal incorporation procedures is addressed in paper 1.

3.3 Characterization methods Several methods were used to characterize the structure and textural properties for the samples at different synthesis steps using SEM, FTIR, TG/DTA, N2-physisorption, TEM and XRD.

3.3.1 Scanning Electron Microscopy (SEM) Scanning electron microscopy was used to examine the morphology and estimate the size of the synthesized mesoporous silica. The sample was placed on a carbon tape and coated with a thin layer of gold before being inserted into the microscope. The SEM images were recorded with FEI Magellan 400 field emission XHR-SEM.

3.3.2 Fourier Transform Infrared Spectroscopy (FTIR) FTIR spectroscopy is a technique that provides information about the chemical bonds between molecules such as -OH, Si-OH, Si-O-Si, C-Si and C-H. The surface chemical modification of the mesoporous silica was studied by Fourier transform infrared spectroscopy (FTIR) with a Shimadzu FTIR-8400S spectrometer, using KBr pellets. To prepare the pellets, 0.8 mg of sample and 120 mg of KBr powder were ground and mixed to remove scattering effects, and finally the powder mixture was pressed to a pellet size, suitable for the instrument.

3.3.3 Termogravimetry and Differential Thermal Analysis (TG/DTA) TG/DTA was used to study the decomposition of polymers of mesoporous silica before and after functionalization. The analyses were performed using a Netzsch STA 449C Jupiter instrument. The measurements were carried out in air. Approximately 20 mg of material was placed in a sintered alumina crucible and the temperature was increased from room temperature to 700 ºC at a heating rate of 10 ºC min-1.

3.3.4 Nitrogen adsorption/desorption isotherms The specific surface area, pore size and pore volume of the samples at different synthesis steps were measured using physisorption with N2 gas. Nitrogen adsorption-desorption measurements were performed at 77 K using a Micromeritics ASAP 2020 surface area and porosity analyzer. The samples were degassed at 373 K for 9 hours before the measurement. The specific surface area was determined by BET model (Brunauer et at., 1938), at a relative pressure (P/Po) range of 0.08-0.2; the pore size distribution was derived from the adsorption isotherm branch using KJS method (Kruk et al., 1997). Finally the total pore volume was calculated from the amount of adsorbed N2 at P/Po = 0.975. 3.3.5 Transmission Electron Microscopy (TEM) TEM micrographs were used to investigate the mesoporous silica with cobalt nanoparticles inside the pores and the particle size distribution of the nanoparticles. TEM was performed with an FEI Tecnai G2 microscope operated at 200 kV, and the chemical composition

determined by energy dispersive X-ray spectroscopy (EDX) in the TEM. For preparing TEM samples, the material of interest was dispersed in acetone and then deposited onto carbon copper grids and allowed to dry before analysis.

3.3.6 X-ray diffraction (XRD) Powder X-ray diffractometer (Siemens D 5000) was used for determination of the phase of crystalline cobalt structure, using Cu Kα radiation over 30° ≤ 2θ ≤ 70°.

4. Summary of results

In this study, a procedure via wet chemical process was successfully carried out to synthesize cobalt nanoparticles at room temperature by reducing cobalt sulphate heptahydrate with sodium borohydride in aqueous solution using functionalized SBA-15 mesoporous silica as a hard template. Silica template dissolution by NaOH aqueous solution resulted in well dispersed Co nanoparticles ranging in size from 2 to 4 nm.

It was shown that the synthesized mesoporous silica SBA-15 has hexagonally ordered mesoporous channels running along the length of the particles, with a pore size of about 10 nm, see Figure 4.1a. The SEM micrograph in Figure 4.1b shows particle as rod-like hexagonal shaped agglomerate with a diameter of 0.4-0.5 µm and length 1-1.5 µm.

(a) (b)

Figure 4.1 Micrographs of mesoporous SBA-15 silica: (a) TEM micrograph showing the channel structures and pore arranged in a hexagonal order (inset). (b) SEM micrograph showing the particle morphology size by agglomeration of various porous spheres.

Likewise, it was found that both external and internal functionalization of silica walls play a crucial role on the infiltration and reaction of the reagents in the silica framework. This process was performed step by step, as described in the experimental part of this thesis.

The external silanol groups (-OH) of the silica was then first modified with TMCS (Cl- Si(CH3)3) groups, as can be seen in Equation (2).

≡Si–OH + Cl–Si(CH3)3 → ≡Si–O–Si(CH3)3 + HCl (2) external surface hydrophobic agent silica external functionalization

Hence, a highly hydrophobic surface was achieved which proved to be sufficient to avoid formation of large cobalt particles on the outside of the silica particles. The incorporation of these groups on the SBA-15 surface has been qualitatively confirmed by FTIR analysis.

The results have shown that the absorption intensity of the Si-OH groups decreases due to replacement of H from silanols (-OH) with Si to form a new O-Si bond. On the other hand, the absorption intensity of the C-H and Si-C groups was increased due to presence of -Si- (CH3)3 groups attached on the silica surface.

The surfactant P123 was efficiently removed from the pores by calcination at 300 °C without affecting the alkyl groups recently anchored to the external silica surface. It was also demonstrated by FTIR and TG/DTA results.

The internal functionalization of the mesoporous silica was successfully achieved with APTMS molecules, as can be seen in Equation (3).

3(≡Si–OH) + (CH3O)3–Si(CH2)3–NH2 → (≡Si–O)3–Si(CH2)3–NH2 + 3CH3OH (3) internal surface silica hydrophilic agent internal silica functionalized

By anchoring amino (-Si-(CH2)3-NH2) groups inside the pore silica a stronger negative charge than silanol groups was achieved thereby enhancing the attraction of cobalt ions into the silica pores. By using aqueous solution of NaBH4 as a strong reducing agent (Özkar et al., 2005) the cobalt ions inside the pores was reduced to cobalt nanoparticles as shown in the TEM- micrographs in Figure 4.2.

(a) (b)

50 nm 10 nm

Figure 4.2 TEM micrographs: (a) Mesoporous silica with cobalt nanoparticles inside the pores (b) cobalt nanoparticles on a Cu/carbon grid after silica removal, and (c) the inset, particles size distribution histogram of the obtained nanoparticles and their Gaussian fit.

It was found that after anchoring amino groups with silanol groups on the internal pore surface, the specific surface area and pore volume decreased. Also FTIR results showed that absorption intensity of -OH is decreased even more and also a new absorption band attributed to the -C-N of the primary amine (-CH2-NH2) was observed in the FTIR spectra.

The cobalt metal nanoparticles seen as dark spots in Figure 4.2 a) are dispersed in the silica matrix. No bulk aggregation of the cobalt on the outer surface could be observed, which indicates that the cobalt is confined to the pores. The presence of ultra-fine Co nanoparticles was further supported by EDX spectra. (See Figure 6 in paper 1).

5. Conclusions

In conclusion, the results show that highly dispersed cobalt nanoparticles was obtained via wet chemical process using NaBH4 as the reducing agent and silica SBA-15 with both external and internal functionalized surfaces as the template. The functionalization of the silica walls plays a crucial role on the infiltration and reaction of the reagents inside the silica pores. It is believed that it is possible to use the same procedure for deposition of other metals inside the mesoporous silica.

References

Aguado J., Arsuaga J., Arencibia A., Lindo M., Gascón V., Aqueous heavy metals removal by adsorption on amine-functionalized mesoporous silica, J. Hazard. Mater., 163 (2009) 213- 221.

Attard G., Goltner C., Corker J., Henke S., Templer R., Liquid-Crystal templates for nanostructured metals, Angew. Chem. Inter. Ed., 36 (1997) 1315-1317.

Beck J., Vartuli J., Roth W., Leonowicz M., Kresge C., Schmitt K. D., Chu C. T., Olson D. H., Sheppard E. W., A new family of mesoporous molecular sieves prepared with liquid crystal templates, J. Am. Chem. Soc., 114 (27) (1992) 10834-10843.

Belkacem W., Labidi A., Guérin J., Mliki N., Aguir K., Cobalt nanograins effect on the ozone detection by WO3, Sensors Sens. Actuators, B. 132 (2008) 196-201.

Brunauer S., Emmett PH., Teller E., Adsorption of gases in multimolecular layers, J. Am. Chem. Soc., 60 (1938) 309-319.

Cai H. and Zhao D., A mild method to remove organic templates in periodic mesoporous organosilicas by the oxidation of perchlorates, Microporous Mesoporous Mater., 118 (2009) 513-517.

Fan J., Yu C., Gao F., Cubic mesoporous silica with large controllable entrance sizes and advanced adsorption properties, Angew. Chem. Int. Ed., 42 (2003) 146-150.

Fröba M., Hoffmann F., Cornelius M., Morell J., Silica-Based Mesoporous Organic-Inorganic Hybrid Materials, Angew. Chem. Int., 45 (2006) 3216-3251.

Fuertes A.B., A general and low-cost synthetic route to high surface area metal oxides through asilica xerogel template, J. Phys. Chem. Sol., 66 (2005) 741-747.

Gallis, K.W., Landry C.C., Rapid Calcination of Nanostructured Silicate Composites by Microwave Irradiation, Adv. Mater., 13 (2001) 23-26.

Giri S., Trewyn B., Lin V.Y., Mesoporous silica nanomaterial-based biotechnological and biomedical delivery systems, Nanomedicine, 2 (2007) 99-111.

Hench LL., West J.K., The sol-gel process. Chemical Review, 90 (1990) 33-72.

Hitzky R. and de Juan F., Selective functionalization of mesoporous silica, Adv. Mater., 12 (2000) 430-432.

Ishizaki K., Komarnemi S., Nanko M., Porous materials: process technology and applications, Kluver Academic Publishers, (1998).

Jansen J.C., Shan Z., Marchese L., Zhou W., Puil N., Maschmeyer Th., A new templating method for three-dimensional mesopore networks, Chem. Commun., (2001) 713–714.

Johansson E., Córdoba J., Odén M., The effects on pore size and particle morphology of heptane additions to the synthesis of mesoporous silica SBA-15, Microporous Mesoporous Mater.,133 (2010) 66-74.

Kim M.J., Ryoo R., Synthesis and pore size control of cubic mesoporous silica SBA-1, Chemistry of Materials, 138 (1999) 487-491.

Kresge C.T., Leonowicz M.E., Roth W.J., Vartuli J.C., Ordered mesoporous molecular sieves synthesized by a liquid-crystal template mechanism, Nature, 359 (1992) 710-712.

Lee J., Yoon S., Hyeon T., Oh S.M., Kim K.B., Synthesis of a new mesoporous carbon and its application to electrochemical double-layer capacitors, Chem. Commun., (1999) 2177- 2178.

Lettow J.S., Han Y.J., Schmidt P., Yang P., Zhao D., Stucky G.D., Ying J.Y., Hexagonal to mesocellular foam phase transition in polymer-templated mesoporous silicas, Langmuir, 16 (2000) 8291-8295.

Li D., Guan X., Song J., Di Y., Zhang D., Ge X., Zhao L., Xiao F.S., Highly efficient synthesis of ordered mesoporous silica materials with controllable microporosity using surfactant mixtures as templates, J. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 272 (2006) 194-202.

Lin D.H., Wang Y., Sun S.G., Silver Nanoparticles Confined in SBA-15 Mesoporous Silica and the Application as a Sensor for Detecting Hydrogen Peroxide, J. NANOMATER. 2008 (2008) 1-10.

Liu J., Yang Q., Kapoor M.P., Setoyama N., Inagaki S., Yang J., Zhang L., Structural relation properties of hydrothermally stable functionalized mesoporous organosilicas and catalysis, J. of Physical Chemistry B, 109 (2005) 12250-12256.

Liu X., Wang A., Yang X., Zhang T., Mou C-Y., Su D-S., Synthesis of Thermally Stable and Highly Active Bimetallic Au-Ag Nanoparticles on Inert Supports, Chem. Mater. 21 (2) (2008) 410-418.

Matos J.R., Mercuri L.P., Kruk M., Jaroniec M., Synthesis of large-pore silica with cage-like structure using sodium silicate and triblock copolymer template, Langmuir, 18 (3) (2002) 884-890.

Muth O., Schellbach C., Fröba M., Triblock copolymer assisted synthesis of periodic mesoporous organosilicas (PMOs) with large pores, Chem. Commun., 19 (2001) 2032-2033.

Özkar S., ZahmakIran M., Hydrogen generation from hydrolysis of sodium borohydride using Ru nanoclusters as catalyst, J. Alloys Compd. (2005) 728-731.

Qiao S.Z., Yu C.Z., Hu Q.H., Jin Y.G., Zhou X.F., Zhao X.S., Lu G.Q., Control of ordered structure and morphology of large-pore periodic mesoporous organosilicas by inorganic salt, Microporous Mesoporous Mater., 91 (2006) 59-69.

Rao A.P., Rao A.V., Pajonk G.M., Hydrophobic and physical properties of the ambient pressure dried silica aerogels with sodium silicate precursor using various surface modification agents, Appl. Surf. Sci. 253(14) (2007) 6032-6040.

Rouquerol J., Avnir D., Fairbridge C.W., Recommendations for the characterization of Porous Solids, Pure and Applied Chemistry, 66 (1994) 1739-1758.

Sayari A., Han B., and, Yang Y., Simple Synthesis Route to Monodispersed SBA-15 Silica Rods, J. Am. Chem. Soc., 126 (2004) 14348-14349.

Song S.W., Hidajat K., Kawi S., Functionalized SBA-15 Materials as Carriers for Controlled Drug Delivery: Influence of Surface Properties on Matrix-Drug Interactions, Langmuir, 21 (2005) 9568-9575.

Takagi H., Hatori H., Soneda Y., Yoshizawa N., Yamada Y., Adsorptive hydrogen storage in carbon and porous materials, Materials Science and Engineering: B, 108 (2004) 143-147.

Wu C.W., Yamauchi Y., Ohsuna T., Kuroda K., Structural study of highly ordered mesoporous silica thin films and replicated Pt nanowires by high-resolution scanning electron microscopy (HRSEM), J. of Mater. Chem., 16 (2006) 3091-3098.

Yamada T., Zhou H., Asai K., Honma I., Pore size controlled mesoporous silicate powder prepared by triblock copolymer templates, Materials Letters, 56 (2002) 93.

Yang Ch., Liu P., Ho Y., Chiu Ch., Chao K., Higly dispersed Metal nanoparticles in Functionalized SBA-15, Chem. Mater., 15 (2003) 275-280.

Yang H. and Zhao D., Synthesis of replica mesostructures by the nanocasting strategy, J. Mater. Chem., 15 (2005)1217-1231.

Zdravkov B., Čermák J., Šefara M., Janků J., Pore classification in the characterization of porous materials: A perspective. Central European Journal of Chemistry, 5 (2007) 385-395.

Zhao D., Feng J., Huo Q., Melosh N., Fredrickson G. H., Chmelka B.F., Stucky G. D., Triblock Copolymer Syntheses of Mesoporous Silica with Periodic 50 to 300 Angstrom Pores, Science, 279 (1998) 548-552.

Zhu H., Jones D.J., Zajac J., Dutartre R., Rhomari M., Roziare J., Synthesis of periodic large mesoporous organosilicas and functionalization by incorporation of ligands into the framework wall, Chemistry of Materials,14 (2002) 4886-4894.

PART TWO

CHARACTERIZATION OF BOLIVIAN CLAY MINERALS

Part two of the thesis is about characterization of natural clay minerals from Bolivia and their thermal behaviour.

Scope and objectives of part two Three main kaolinitic-illitic clay deposits are investigated in the present work, in order to assess their application in the ceramic industry as valuable building materials. Two of the deposits are located in the tropical region (IC and EC) and one is located in the valley (U) of Cochabamba-Bolivia.

In this study, emphasis is given to the clay characteristics and phase-microstructural changes of the clay materials taking place during the firing step.

The objectives are to: - Assess the clay characteristics by chemical and mineralogical analysis on representative samples from each deposit. - Study the phase transformation and microstructural evolution during firing. - Investigate microstructural characteristics of the sintered clays.

The results will determine the feasibility of these clays for application in the ceramic industry as useful building materials for this tropical region.

1. Introduction

The general term “ceramic” includes most of the inorganic materials with ionic or covalent bonding. This broad definition of ceramic materials involves most of the elements in the crust of the earth. Clays are one of the most abundant resources that serve as a main raw material to produce traditional ceramics, pottery, white-wares, refractories and technical-engineering including ceramic matrix composites (Murray, 1999).

Traditional ceramics refers to the products commonly used as building materials or internally used in home and industry. Products such as bricks, tiles, sewer pipes, etc., are made mainly of mineral clays that can contain quartz, carbonates, feldspars, iron oxides, etc. (Grim, 2006).

Raw materials used in the traditional ceramic products are varied as well as the wide span of ceramic products available in the market. Some of these clay-based products are shown in Figure 1.1.

Figure 1.1 A variety of white and red clay-based products.

The phases and microstructural composition developed of fired products vary significantly according to the composing minerals in the clay materials. In addition, the physical and mechanical properties of the resulting ceramics will be determined by the microstructure and phase composition developed in the fired products (Lee and Yeh, 2008).

Furthermore, from an application point of view the final product properties also depend on the applied processing techniques, such as shaping technique, firing temperature, thermal cycle, type of kiln, etc. (Baccour et al., 2009).

Figure 1.2 shows the relationship between properties of the final products and the features of raw material and the applied manufacturing processes.

RAW MATERIALS (Clay minerals and oxides)

PROCESSING FINAL PROPERTIES (Shaping and firing) (Physical and mechanical)

Figure 1.2 Relationship between features of raw materials, processing and final properties.

The nature and abundance of other minerals in the clays have significant effect on the thermal behaviour. Some of these common components, that play a fundamental role for optimum processing and hence performances of the structural final products, are feldspars for fluxing, and silica as filler material (McConville and Lee, 2005).

In order to use clay materials in a variety of applications as construction materials, often a proper mixture is prepared by mixing several types of clays to a determined application, such as wall tiles, roof tiles, bricks, etc. Thus, a detailed knowledge of their compositions and thermal behaviour is of fundamental importance to maximize an efficient use of such materials mix to be used for a given application.

It is also very important to understand the mineralogical transformations in the clay minerals, which take place during the firing process. As a result the sintered material might be crystalline or partly crystalline, porous or highly densified.

1.1 Ceramics in Bolivia In Bolivia the ceramic industry has experienced a fast growth during the last 15 years. Especially in those directly related to the construction such as bricks, roof tiles and glazed- tiles.

Brick production in Bolivia is about 645 million pieces per year. Cochabamba department with a production of 226 million of solid bricks is equivalent to 440 458 tons of clay processed by the second largest producer in Bolivia (EELA, 2011).

Nowadays, the internal demand of these building materials is increasing constantly due to the growth rate of the population in some tropical regions such as Ivirgarzama and Entre Rios, municipalities of Cochabamba department.

Actually, some small primitive factories are operating in these regions, producing solid (adobe) bricks from red clay for economic habitat, (Figure 1.3). However, the quality of these products is often poor. They show low mechanical resistance and durability. This is essentially due to the lack of knowledge on the chemical and mineralogical features, the processing like mixing, shaping and drying, and also the use of an adequate firing cycle.

Figure 1.3 Old style artisanal production of red-bricks (left) and furnace (right) in Cochabamba-Bolivia.

It is quite often that the exploitations of the clays are located in the vicinity of the factories as the low value of the raw material does not allow for lengthy transport.

In this context, is important to localize the industries near to the clay deposits. The exploring and characterization of new deposits of clays found in these tropical regions is being considered an important aspect to promote the production of traditional ceramics of good qualities in an efficient way.

2. Theoretical background

This chapter contains a literature review on the characteristics of the natural clays, sequence and reactions of phase transformations during heating and the mechanisms of sintering at high temperatures.

2.1 Clay minerals The term “Clay mineral” refers to phyllosilicate minerals (<2µm), which impart plasticity when wet and harden upon drying and firing. Essentially, the clay minerals are a group of hydrous aluminosilicates that are typically found in the clay fractions of sediments and soils (Guggenheim and Martin, 1995).

Clays are formed as a result of chemical weathering of many different anhydrous aluminosilicate compounds such as igneous rocks including granites, feldspars, rhyolites, etc., (Bétard et al., 2009). In nature, the formation of clays is complex. Most often, the formation of clay is considered in the context of the decomposition of granite, a rock that contains fractions of feldspars, quartz and mica. In addition, associated phases in clays may include materials that do not impart plasticity such as quartz, carbonates, feldspars, pyrites, oxides, hydroxides, organic matter, etc.

According to the geological formation, clays can be classified as primary or secondary clays. Primary clays are located at the same site of formation as the igneous rocks. Secondary clay deposits have been moved by erosion and water from their primary location (Prudêncio et al., 2002).

The most common clay minerals for traditional applications such as white and red ceramics are mainly kaolin, mica-illite, smectites, and chlorites. These minerals represent the major phases of clays from sedimentary rocks. Illite, kaolinite and montmorillonite clays are the main clay phases used for the preparation of proper mixtures for this traditional ceramics. Red clays usually have moderate to high plasticity, which facilitates forming of bricks (Bauluz et al., 2003).

With regard to industrial usage, white kaolins are often used for production of porcelain, stoneware tiles, tableware and sanitary ware bodies. Usually, mixtures are used that are mainly composed of kaolin, feldspar and quartz. These materials must have low iron content.

However, red clays are often used for building materials like bricks, roof-, floor- and wall tiles. The compositions of red clays are highly variable. But in general they contain high amounts of alkalis and iron, and are also accompanied by other secondary components (Dondi, 1999).

2.1.1 Main clay minerals groups The majority of clay minerals have sheet silicate structures. The sheet silicates consist of “composite layer” sheets of tetrahedrally coordinated Si, Al and octahedrally coordinated cations (mainly: Fe3+, Fe2+, Al3+, and Mg2+).

According to the sheet type, silicate clay minerals can be classified as clays of two layer or 1:1 type represented by the kaolin and serpentine, and as clays of three layer or 2:1 type represented by the illite-mica, smectite, vermiculite and chlorite groups (Murray, 1999). The corresponding classification of clay minerals is listed in Table 2.1.

Table 2.1 Classification of clay minerals according to the sheet type silicate.

Sheet silicate Group Species (Polytypes) type 1:1 Kaolin and serpentine - Kaolinite, dickite and nacrite. 2:1 Illite-mica - Muscovite, illite, glauconite, celadonite, paragonite. 2:1 Smectite - Montmorillonite, beidellite, nontronite. 2:1 Chlorite - Clinochlore, chamosite, pennantite, etc. 2:1 Vermiculite

Figure 2.1 shows the crystal structures of the main clay minerals, kaolin, illite and montmorillonite and they will be described in the following sections.

Figure 2.1 Crystal structures: (a) Kaolin 2 layer (1:1) type, (b) Illite 3 layer (2:1) type, (c) Montmorillonite 3 layer (2:1) type. (Adapted from Grim R.E. 2006)

Kaolin group To the kaolin group belong the kaolinite, dickite, nacrite and serpentine minerals with two layer 1:1 type structure. Kaolins have one tetrahedral and one octahedral structure in the unit cell (Figure 2.1a), with no net negative charge on the composite layers and consequently no compensating interlayer cations or water layers in the structure.

Kaolinite (Al2Si2O5(OH)4) with interlayer spacing of 7 Å, is the principal clay of its group and it is the most common clay used commercially. Kaolinite can range from well crystallized varieties to poorly crystalline forms (Schwaighofer and Muller, 1987).

The mined kaolinite is generally associated with the presence of various other minerals depending on the geological conditions under which the kaolinite was formed. These associated minerals can modify the physical and chemical properties of a kaolinite and affect its use as an industrial mineral. This type of clay mineral is generally found in acidic tropical soils in areas with high rainfall (Milheiro et al., 2005). 24

Additionally, kaolinite is known as the most refractory clay mineral for its high content of alumina (Al2O3) and low content of fluxing agents. The main product phase after firing of kaolin at high temperatures is mullite phase (3Al2O3·2SiO2) (Ghorbel et al., 2008).

Mica-illite group The name mica can refer to both mica and illite, depending on the potassium content, so the nomenclature can vary. Approximate formulae deduced for illite, can be written as K0.88Al2(Si3.12Al0.88O10(OH)2·nH2O or (Si4)(AlMgFe)2.3O10(OH)2(K,H2O) (Inoue et al., 1987) (Rosenberg, 2002). Illite is a three layer 2:1 clay structure (Figure 2.1b). In the illite structure, the substitutions of Si4+ by Al3+ produce a net negative charge. So the potassium is the principal interlayer cation. Water may also be present in the interlayer sites to fill up empty spaces in the structure (Wilson, 1999).

There are many fields where illitic clays play an important role, for instance illite is one of the components of the red clays often used in the production of cooking pots, plates, tiles and bricks (Ferrari et al., 2006).

Smectite group Smectites are three layer or 2:1 clay minerals, one octahedral sheet sandwiched by two tetrahedral sheets, (Figure 2.1c). They have a charged layer, which is offset by hydrated interlayer cations, mainly Mg2+ and Na+. The hydration of the interlayer cations causes the interlayer crystalline swelling that characterizes the smectites, i.e. water is absorbed into the interlayer sites in the molecular sheets, well known as crystal water (Garrels, 1984).

Montmorillonite is one of the most common of its group, with a nominal composition of (½Ca,Na)0.7(Al,Mg,Fe)4[(Si,Al)8O20](OH)4·xH2O where x is a variable depending on the level of water absorbed. Montmorillonite is a colloidal mineral of very high specific surface area and is a scavenger for cations. So, then montmorillonite clay has high cation exchange/swelling capacity, high porosity and high surface area. Chemical and structural analysis of some smectites revealed a montmorillonitic composition with iron and iron-rich smectites. Some other smectites are richer in Mg and have Fe3+ and Mg2+ in the octahedral position, replacing Al atoms (Prudêncio et al., 2002).

2.1.2 Associated minerals of clays Quartz, (SiO2) is one of the most common non-plastic materials found in variable amounts as accessory in clay minerals. Silicon dioxide is a complex material that can exist in many forms of polymorphism, where the most common stable forms of silicon dioxide are quartz, cristobalite and tridymite. The α-quartz is stable up to 573 °C. As the temperature increases, the stable phase α-quartz is converted to β-quartz. This conversion is quick, reversible and accompanied with slight energy absorption (Heaney P.J. and Veblen, 1991).

The role of silica is that of a filler, used to impart “green” (that is, unfired) strength to the shaped object and to maintain that shape during firing and also reduce distortion and shrinkage. Therefore, the presence of quartz in the clays might enhance the workability in some cases when the clays are highly plastic, especially for brick fabrication by extrusion process, thus providing them a uniform shape that last during the drying step (Vieira et al., 2008).

Feldspars refer to a group of anhydrous aluminosilicate minerals. The composition can vary significantly among the end-members of the system such as K, Na and Ca. These feldspars are orthoclase (KAlSi3O8) albite (NaAlSi3O8) and anortite (CaAl2Si2O8). Feldspar is often left unaltered in certain amounts during the formation of clays.

Feldspars play an important role in ceramic materials, acting as fluxing agents to reduce the melting temperatures of the clays. During firing their fusibility and ability to form eutectics with other components are remarkable, making it possible to reach a high densification even at low temperatures (Das and Dana, 2003). Some examples of products where feldspars play an important role are porcelain pastes, vitreous china, and floor tiles. In fact, the great densification and high mechanical resistance showed by these ceramic materials after firing are due to the action of feldspars (Esposito et al., 2005).

Carbonates in clay minerals are present in various amounts due to differences in the geology formation. Calcite, CaCO3 and dolomite, CaMg(CO3) are the predominant carbonate minerals in clay materials. The presence of CaCO3 in significant proportions in the clays may lead to the formation of undesirable phases such as gehlenite, anorthite and wollastonite after sintering, which impairs the mullite formation (Carretero et al., 2002).

Red clays can be subdivided in terms of their carbonate content, from nil to low, medium or high. Red clays with low carbonated content are usually employed in roof and floor tiles, whereas red clays with a medium to high carbonate content are typically used in porous bricks and wall tiles. A low CaO content in the clays, about less than 6 %, is an indicative of non- calcareous clays (Gonzalez et al., 1990).

Sulphates may be present in clays, the most common sulphates are gypsum, CaSO4·2H2O, anhydrite, CaSO4, and barite BaSO4 as insoluble salts. Some soluble sulphates may also be present in the clays such as sodium and magnesium sulphates.

Sulphates are the most common salts for generating unwanted efflorescence on structural clay products. The sulphates react with the silicates at temperatures above 982 °C. The products of the reaction include sulphur gases. Some of this gaseous sulphur may escape the system, but the rest is retained by the clay body and may subsequently cause some efflorescence in the ceramic material (Brownell, 1949). Thus, a low content of sulphur minerals such as pyrite

(FeS2) and alunite (2KAl3SO4(OH)6) in the clays is always an advantage for the ceramic industry.

Iron is also present in various mineral forms in clay materials. Iron can be present in the forms of hematite (α-Fe2O3), goethite (α-FeO·OH) and limonite which is a mixture of iron oxides and hydroxides of a poorly crystalline nature or simply as Fe3+ ions in the clay structure. Since, Fe3+ can partially substitute the Al3+ ions in the octahedral sites of the clay structure it usually occurs in illite structures (Stepkowska et al., 1992).

The Fe2O3 is formed during sintering in oxidising conditions from the reactions of the iron minerals present in the clays, confers the characteristic red-colour to the ceramic materials. Colour is one of the most important aesthetic aspects for bricks attracting the attention of consumers (Valanciene et al., 2010).

2.2 Decomposition and phase transformations By heating, the clays undergo a series of chemical and physical changes that transform the layered mineral to a combination of crystalline mullite and an amorphous siliceous phase. This is conducted through the intermediate phases, such as metakaolinite and spinel rich in amorphous silica.

Three kinds of processes take place during heating of the clays, decomposition, phase transformations and sintering. The decomposition and phase transformations influence the evolution and intensity of the sintering process (Mota et al., 2008).

2.2.1 Metakaolinite and spinel phases During firing, the crystalline clay structures once they exceed their stability limits, they partially decompose and simultaneously other phases are being formed. There is a basic understanding of the clay decomposition and phase crystallisation sequence.

An overview of these transformations of kaolinite to spinel-amorphous silica (γ-Al2O3-rich silica) with an intermediate phase, namely metakaolinite, is given below:

Al2O3·2SiO2·2H2O  Al2O3·2SiO2 + 2H2O (1) kaolinite metakaolinite

Al2O3·2SiO2  1/2(2Al2O3·3SiO2) + 1/2(SiO2) (2) metakaolinite spinel amorphous

The thermal transformation of kaolinite to metakaolinite (dehydroxylation) as shown in Equation (1) can be divided into two steps. First, the structural water is removed and disruption of the kaolinite sheet structure (delamination) proceeds. Second, it is mostly explained as a kinetically controlled recombination of alumina and silica to form an amorphous metakaolinite structure. Kinetic analysis showed that dehydroxylation of kaolinite is a third-order rate reaction (Ptacek et al., 2010).

Most often, the dehydroxylation process takes place within a temperature interval from 400 to 700 °C and it is highly dependent on experimental factors such as the order in the kaolinite structure (Heide and Földvari, 2006), associated minerals and elements, sample treatment, particle size distribution, heating rate (Castelein et al., 2001) and other structural characteristics.

When the temperature increases at about 980 °C, metakaolinite undergoes a structural transformation, Equation (2). At this temperature, spinel and an amorphous phase is formed. This is an exothermic transformation without weight loss. The spinel phase is similar in structure to the cubic transitional alumina, γ-Al2O3 and the amorphous phase is mainly silica, but it also contains the impurities from the original clay.

In addition, the spinel-silica crystallizes within metakaolinite, which has been reported at various points of temperatures between 900 and 980 ºC (McConville et al., 2005).

2.2.2 Mullite phase When spinel-silica phase is heated about 980 °C, a small fraction of mullite crystals start to form and then they continue to grow at further heating. Mullite growth is accompanied by the disappearance of spinel phase at a slow rate, Equation (3).

1/2(2Al2O3·3SiO2)  1/3(3Al2O3·2SiO2) + 5/6(SiO2) (3) spinel Mullite amorphous

The crystal structure of mullite is orthorhombic. It consists of AlO6 octahedral chains, parallel to the c-axis, which are cross-linked by the (Al,Si)O4 tetrahedral chains. Mullite itself is very stable at high temperatures, i.e. it has refractory properties, low thermal expansion coefficient, low thermal conductivity and a high melting temperature (Ghorbel et al., 2008).

Mullite is described as a solid solution between Al2O3 and SiO2, where the amount of Al2O3 can vary between 55 and 90 mol % depending on the manufacturing conditions (Lee et al., 1994). It can be described by the mole fraction according to the formula: Al4+2xSi2-2xO10-x where x is the number of oxygen vacancies.

The two most stable mullite-type compositions were proposed as (3Al2O3·2SiO2) and (2Al2O3·SiO2) by Aramaki-Roy (1962) and Bowen-Grieg (1924). It is pointed out that 3:2 type mullite (3Al2O3·2SiO2) only forms directly by solid state reaction of oxide precursors while 2:1 type mullite (2Al2O3·SiO2) only forms at very high temperatures by sol-gel processes (Sheneider et al., 1993). Moreover, mullite formed from kaolin clay alone is termed primary mullite (2:1 type-mullite) whereas that formed from reaction with an alkali flux is termed secondary (3:2 type-mullite).

Lundin, (1954), recognized that mullite crystals have different morphologies. For instance, mullite formed in vitreous ceramics from the clay and their interactions with the other components of the microstructure have acicular morphology like needles, and it is believed to have an important effect on mechanical properties due to its interlocking in the ceramic matrix.

The higher the mullite content and the higher the interlocking of the mullite needles, the higher is the strength. Hence the strength of the ceramic material depends on the factors that affect the amount and size of mullite needles, like the firing temperature and composition of alumina and silica in the raw materials. More aluminosilicate liquids fluxed with alkali and iron oxides encourage growth of mullite crystals (Lee et al., 2008).

2.3 Sintering In the processing steps, sintering of the ceramic material is highly fundamental to adjust several desired properties. In this context, the green materials are fired, also termed densified or sintered, at temperatures sufficient to develop the microstructure enough to confer useful properties on the material.

Sintering can be defined as removal of the pores between starting particles accompanied by shrinkage of the ceramic material combined with formation of strong bonds between adjacent particles (Gonzales et al., 1990).

The sintering is governed by different mechanisms according to the temperature. As the temperature increases, sintering is governed by solid state diffusion mechanisms. Stage that in general involves rearrangement of the powder particles and formation of a strong bond between particles. The amount of porosity decreases substantially and particles move closer leading to shrinkage of the material and also grain size increases during this stage.

At temperatures, when some liquid phase starts to form in the ceramic material, sintering is governed by viscous flow mechanisms with high influence on the densification of the green body. The formation of liquid phase plays an important role on the densification in the ceramic products. Most commercial ceramics are densified with a liquid phase present. Thus the presence of components such as low-melting clays is very important (Milheiro et al., 2005).

During the firing process of the clays at high temperatures, there is a gradual increase in the amount of liquid phase. At the beginning, the liquid phase formed results when the tetrahedral part of the clay structure combines with any alkalis present to form a viscous liquid phase, due to the fusion of these meltable components with the silica structure. After that, the most refractory components are progressively dissolved by the liquid preformed (Bernardin et al., 2006).

The density increases with the firing temperature up to a maximum, where there is enough liquid phase to block the open porosity. The density decreases as the temperature is increased further, and this is due to the so called bloating phenomenon often occurring on extended heating of the ceramic body (Wattanasiriwech et al., 2009).

Shrinkage is a result of densification, so the amount of shrinkage is dependent on which type of clay is involved. It also depends on the characteristics of the clay such as particle size, and on how many and what types of the secondary components that are present in the clay. This means that highly plastic clays have a very fine particle size and will shrink more, while clays with large particles will shrink less, as well as clays containing non-plastic components such as silt or sand (Jordán et al., 1999).

3. Materials and methods 3.1 Raw materials The representative clays in the present study were collected from the large deposits of Cochabamba-Bolivia. Two of them were collected from different locations of tropical region, termed simply as IC (clay from Ivirgarzama), and EC (clay from Entre Rios). One sample was collected from the valley region, termed U (clay from Uspha-Uspha). This commercial clay (U) was selected for its extensive application in the pottery industry and for bricks and tiles production by local ceramic industry.

The three different clays were investigated in order to compare structural characteristics and thermal behaviour. The deposits are indicated on the map shown in Figure 3.1.

IC: 17º 01´ 18´´S, 64º 57´ 70´´W EC: 17º 09´ 42´´S, 64º 30´ 10´´W U: 17º 46´ 00´´S, 66º 30´ 00´´W

Figure 3.1 Location map and coordinates of the clay deposits in the tropical region and valley of Cochabamba-Bolivia.

The collected and representative samples from each location were crushed, ground and sieved to pass the 315 µm mesh sieve. These samples were then used for all experiments.

Figure 3.2 shows the three places of the collected clays in Cochabamba-Bolivia.

Figure 3.2 Deposits of clays of IC and EC from the tropical region and U clay from the valley of Cochabamba-Bolivia.

As can be seen in Figure 3.2, the clays were collected from deposits with different levels of geological formation, so the mineralogical and chemical characteristics and their thermal behaviour will be a determinant aspect of these clays. Thus, a prior knowledge of the quality and characteristics of the clays will address its application in ceramics.

3.2 Characterization methods 3.2.1 Mineralogical and phase analysis by XRD Powder X-ray diffraction (XRD) was used to identify crystalline clay minerals and other minerals as secondary components in the raw clays. It was also used to study the microstructural evolution on heat treated samples at different temperatures. The clays were heat treated in a chamber furnace (Nabertherm) in air atmosphere at the following temperatures; 450, 650, 950, 1050, 1150 and 1250 °C. The heating rate was 10 °C min-1 up to the final temperature, and the samples were then furnace cooled.

The XRD patterns were recorded using a Siemens D 5000 X-ray diffractometer with CuKα radiation (λ = 1.5418 Å) operating at a tube voltage and current of 40 kV and 20 mA, respectively. Diffraction patterns were recorded in the 2θ range between 5 º and 50 º (2θ), with a step size of 0.02 2θ/degree. The phases were identified from peak positions and intensities using reference data from the JCPDS-ICCD (2004).

3.2.2 Chemical analysis by ICP-AES The chemical analysis was used to quantify the present chemical elements in the clays. The bulk chemical analyses of the collected samples were determined by Inductively Coupled Plasma with Atomic Emission Spectroscopy (ICP-AES). The results of the analyses were expressed in weight % of oxides.

3.2.3 Thermogravimetry and Differential Scanning Calorimetry analysis (TG/DSC) The thermal analysis, TG and DSC, were used to measure mass changes and thermal effects in the clays, due to evaporation, decomposition and interaction with synthetic dry air. These experiments were performed in a Netzsch STA 449C Jupiter instrument, equipped with a Netzsch Aeolos QMS 403C mass spectrometer. The experiments were carried out on powder samples in an alumina crucible between room temperature and 1300 ºC at a heating rate of 10 °C min-1 in synthetic air.

3.2.4 Dilatometry analysis (DIL) The dilatometer was used to study the dimensional changes, transitions and sintering temperature in the compacted samples. The dilatometry experiments were performed in a Netzsch DIL 402C instrument. Prior to measurement, the compacted sample was dried overnight at 110 ºC. The samples were heated from room temperature to 1150 ºC in synthetic air at heating rate of 10 ºC min-1.

3.2.5 Bulk density and open porosity Bulk density and open porosity were measured to correlate the obtained results by dilatometry in the sintering step of the studied clays. Measurements of bulk density and open porosity on samples fired at 950, 1050 and 1150 °C were performed using Archimedes method according to ASTM C20-00(2005) standard.

3.2.6 Scanning Electron Microscopy (SEM) The morphology of the raw clays and the resulting microstructure after sintering in the chamber furnace, at 1150 ºC, and a heating rate of 10 °C min-1, on polished samples both before and after chemically etched with 20 wt.% HF for 10 min, were studied by scanning electron microscopy (SEM) using a Jeol 6460-LV microscope.

4. Results and discussions

This chapter presents results and discussions of mineralogical and chemical analysis of the three studied raw clays and their thermal behaviour at high temperatures, followed by phase evolution analysis by XRD. SEM micrographs of the raw clays and sintered samples at 1150 °C are also included. In addition, results of the measured bulk density and open porosity of samples sintered at various temperatures are also given. The performed work is shown schematically in Figure 4.1.

Raw clays Mineralogical analysis (XRD)

Chemical analysis Characterization (ICP-AES)

Morphology images (SEM)

TG / DSC Thermal analysis

Dilatometry

Microstructural evolution (XRD) Firing/sintering

Microstructure images (SEM)

Figure 4.1 Conceptual framework of the results.

4.1 Mineralogical analysis of raw clays (XRD) The x-ray diffractograms of the three studied raw clays are shown in Figure 4.2.

(IC)  Kaolinite (EC)  Kaolinite      Illite  Illite ..Quartz ..Quartz ^ Orthoclase ^ Orthoclase Montmorillonite



Intensity/a.u.  intensity/a.u. ^                          ^        ^      

10 15 20 25 30 35 40 45 10 15 20 25 30 35 40 45 2/degree 2/degree

(Uspha)   Kaolinite   Illite ..Quartz ^ Orthoclase

 Intensity/a.u.                 ^ 

10 15 20 25 30 35 40 45 2/degree

Figure 4.2 X-ray diffractograms for IC, EC and U raw clays.

The diffractograms show that the clays are composed mainly of kaolinite, illite and α-quartz crystalline phases (JCPDS-ICCD, 2004). Additionally, montmorillonite clay is present in the EC clay.

Minor amounts of feldspars are detected in all clays. This type of component is desirable in clays and it is considered the main fluxing agent of the ceramic materials by its alkali content such as potassium and sodium. Such elements promote formation of liquid phase at low temperatures (Das and Dana, 2003).

According to the intensity of the peaks between illite (8.8 °) and kaolinite (12.2 °), kaolinite is more predominant phase in the IC and EC clays than in the U clay, which is rich in illite phase. Usually, illite favours the sintering of the clays permitting the formation of liquid phase at lower temperatures and it also contributes to a faster densification (Seynou et al., 2011).

Milheiro et al. (2005) pointed out that kaolinitic-illitic- clays are generally found in acidic tropical soils in areas with high rainfall. This characteristic is similar to the IC and EC clays according to the XRD results in this study.

4.2 Chemical analysis of raw clays (ICP-AES) Table 4.1 shows results of the chemical composition, loss on ignition and the calculated ratios of SiO2/Al2O3 of the clay samples compared with standard clays published elsewhere in the literature.

Table 4.1 Chemical composition expressed in oxide form (wt %) for the three studied clays and two standard clays as well as loss on ignition and SiO2/Al2O3 ratios.

Oxides IC EC U Standarda Standardb clay clay clay illite kaolinite SiO2 71.8 65.9 56.4 58,2 43,8 Al2O3 17.5 18.8 26.9 19,8 38,5 Fe2O3 1.67 5.87 2.47 6,2 0,9 K2O 2.32 3.39 3.97 7,7 0,2 Na2O 0.12 0.82 0.74 0,2 0,3 CaO 0.1 0.18 0.10 0,5 trace MgO 0.81 1.76 0.96 2,2 trace TiO2 0.96 0.87 0.98 0,8 0,6 SO3 trace trace 0.85 0,1 trace LOI 6.10 6.75 7.50 4,1 13,8 SiO2/Al2O3 4,1 3,5 2,1 2,94 1,14 ratio a: Silver Hill illite-clay standard (Source Clay mineral Repository at University of Missouri) (Köster, 1996) b: 99 % English China kaolinite standard clay (McConville et al., 2005) LOI: loss on ignition (1000 °C, wt %) Trace: quantity detected below 0,1%

The results show that all clays are constituted mainly of Si and Al. These elements are associated with clay structures, quartz and feldspars.

The quartz amount in the IC and EC clays is higher than in the U clay. In general the SiO2 excess content in the clays based on stoichiometry of standard clays is associated with quartz as accessory mineral present in the clays. It is further confirmed by the SiO2/Al2O3 ratio calculated for the clays (Table 4.1) and the calculated values are well situated within the most frequently published interval from 2.3 to 5 units (Alcantara et al., 2008) (Dondi et al., 2001).

A combination of the XRD results (Figure 4.2) and chemical analysis (Table 4.1) reveal that IC clay contains the highest quartz content and U clay contains the lowest quartz content, so the order in quartz fraction content in the clays is as follows: IC > EC > U.

Quartz is considered an important component in clays for the manufacturing of structural ceramics as it enhances the workability and favours the shaping by extrusion. Furthermore, it diminishes the drying shrinkage and the risk of crack formation in the green bricks, tiles, etc. Quartz can also contribute to the glassy phase formation in presence of feldspars during sintering, wherein mullite phase can be easily crystallized (Seynou et al., 2011) (Dondi et al., 2001).

The potassium content in the three studied clays is in between standard illite and standard kaolinite. The lowest potassium content corresponds to the IC (2.3 %) clay, followed by EC (3.4 %) and U (4 %) clays. It is evident from XRD results that the studied clays are a mixture of kaolinite and illite.

The three studied clays contain iron. The highest iron content corresponds to the EC (6 %) clay, followed by U (2.5 %) and IC (1.7 %) clays. Clays with low iron content, less than 1.5 % is considered suitable material for white ceramic products. The calcium and magnesium content are very low in the studied clays with the exception of EC clay which has a content of 1.8 % MgO.

Table 4.1 shows that pure illite has high content of iron and magnesium. Likely, this suggests that Fe and Mg are forming part of the internal structure (Köster, 1996).

Loss on ignition at 1000 °C around 6-7.5 % is associated with presence of hydroxides and organic matter, and the results are within the interval of standard clays.

Figure 4.3 shows, that the IC clay has low content of oxide fluxes and high content of quartz in comparison with the other clays. The EC and U clays show a similar content of oxide fluxes. The U clay shows a higher content of alumina compared to both IC and EC clays. Alumina is the component in the clays that forms mullite phase in the ceramics.

SiO2 ▲: IC clay 10 90 ■ : EC clay

2 ● : U clay 0 80 ▲ ◊ : Standard illite (a) 30 ■ 70 ◘ : Standard kaolinite (b) 40 ● ◊ 60 50 ◘ 50

0 0 0 0 0 1 2 3 4 5 Al2O3 Oxides (so-called fluxes) (K2O+Na2O+CaO+MgO+Fe2O3)

Figure 4.3 A ternary diagram of SiO2-Al2O3-other oxides. Other oxides are represented by total oxides of (K2O+Na2O+CaO+MgO+Fe2O3) “so-called fluxes”. The symbols are representing the obtained results from chemical analysis previously normalized to 100%.

From the diagram, it is expected that the EC clay will result in a faster sintering and forming higher quantities of glassy phase than IC and U clays, due to its low content of quartz and alumina. Both alumina and quartz confer refractoriness to the ceramic material. This will be discussed later from results obtained by thermal analysis.

Further, iron-rich illite clays along with interlayer cations such as K, Ca and Mg result in large quantities of liquid above 1000 °C (McConville et al., 2005).

4.3 Morphology of clays (SEM) Figure 4.4 shows the morphology of the three studied raw clays.

(IC) (EC)

Flakes-clays

Agglomeration montmorillonite

(U)

Flakes-clay

Figure 4.4 SEM-micrographs showing the microstructure of the studied raw clays.

The SEM-microstructures show the typical characteristics of clay particles like a flaky texture. The IC clay shows a lot of small particles spread around the quartz particles, while the EC clay has some large flakes stacked together with montmorillonite and illite forming agglomerates. Agglomeration can be explained by the fact that montmorillonite absorbs much more water from the ambient air than other clays. The U clay shows some densely-packed layered sheets of aggregate with irregular morphology and forming particles larger than 5 µm.

4.4 Thermal analysis The relationship between thermal behaviour and the phase changes development were examined by DSC/TG and dilatometry.

4.4.1 Thermogravimetry - Differential thermal analysis (TG/DSC) Figure 4.5 shows the TG/DSC and thermodilatometric results of the studied clays. The results were plotted together in order to correlate thermal behaviour of the clays during firing.

Figure 4.5 Thermal analyses results of DSC/TG and linear thermal dilatation curves of the three studied clays. Heating rate 10 °C min-1.

Table 4.2 Summarizes the results of the DSC/TG measurements for the three studied clays.

Mass change (wt%) Maximum temperature of peak (ºC) - DSC at the temperature range (ºC) -TG Clay RT-300 300-700 700-1300 Dehydroxylation α-β quartz Spinel Onset liquid inversion phase formation IC clay 1.31 4.43 0.34 498 574 990 1175 EC clay 2.24 4.07 0.49 497 574 -- 1081 U clay 0.87 5.79 0.88 505 574 990 1170

The DSC curves (dashed-dotted lines) show three characteristic endothermic peaks. First peak at 100 °C is due to evaporation of moisture adsorbed and it is accompanied by weight loss. Second peak around 500 °C is due to dehydroxylation of the hydroxyl groups from the clay structure (reaction described by Equation 1) and it is correlated with the highest weight loss. Third peak at 574 °C corresponds to the α↔β quartz phase transition, but in contrast to the previous ones, it is not accompanied by any change of the weight as expected.

The exothermic unique peak seen for the U clay at 990 °C, is due to the formation of cubic spinel-phase γ-Al2O3 and amorphous silica from metakaolinite, (reaction described by Equation 2) (Chen et al., 2000) (Castelein et al., 2001) (Chen et al., 2004). Also this process is not accompanied by change of the weight of sample.

The TG results of the EC clay shows a higher weight loss (2.24 %) than IC (1.31 %) and U (0.87 %) clays in the range from room temperature to 300 °C. This is related to the loss of absorbed moisture at 100 ºC and the dehydration of crystallized or structural water contained between inter-laminar structures around 250 ºC. Both high moisture absorption and high crystal water interlayer content are typical features of the montmorillonite type clays. The montmorillonite fraction content in the EC clay is well supported by XRD results (Figure 4.2). In addition, around 200-350 °C the weight loss is associated with the decomposition of organic matter from the clays.

The water release of kaolinite in a temperature range of 400 to 700 °C and the formation of metakaolinite is well known since formation of metakaolinite represents the major weight loss detected by TG analysis. The highest weight loss corresponds to U clay (5.8 %), due to its higher clay fraction content and less quartz content than the other two clays. This assertion is well supported by its chemical and mineralogical analysis, Table 4.1 (ratio of SiO2/Al2O3) and X-ray diffraction, Figure 4.2, respectively.

As the temperature increases above 990 ºC, the onset of glassy formation is seen for all the studied clays. The onset temperature is different for all clays. The lowest onset temperature of glassy formation corresponds to the EC clay (1081 °C) followed by U (1170 ºC) and IC (1175 ºC) clays. The onset temperature of glassy formation is varying inversely with the relative total amount of fluxing agents present in the clays (Table 4.1)

4.4.2 Dilatometry analysis (DIL) The dilatometer measurements of the clay samples are shown in Figure 4.5 (dashed lines). From room temperature to 150 °C, a small shrinkage is observed in all clays due to moisture evaporation. From 150 to 560 ºC, a regular expansion is observed which is more pronounced for the IC clay sample. It is related to the dehydroxylation process. As the temperature increases at 573 °C there is a change of expansion due to α→β phase transition of the quartz structure. Also it is observed that the IC clay shows the highest expansion at 573 °C and that is correlated to the quartz content in the clays as supported by chemical analysis and X-ray diffractograms of mineralogical analysis. See Table 4.1 and Figure 4.2.

Between the temperatures 600-950 °C, the slight shrinkage is related to the continuous dehydroxylation of the remaining hydroxyl groups in the illite phase and it also corresponds to the structural reorganization of the metakaolinite. The pronounced shrinkage at 950 ºC for IC clay and 917 ºC for the EC clay and at 1000 ºC for the U clay corresponds to the sintering of the clay samples.

The shrinkage values (dL/Lo, %) as a function of temperature and the total shrinkage at 1150 °C is given in Table 4.3.

Table 4.3 Data collection in the sintering step by dilatometer.

Δ linear shrinkage (dL/Lo, %) Total shrinkage Clay RT- 950 ºC 950-1050 ºC 1050-1150 ºC at 1150 °C

IC -0,24 0,56 1,2 1,52

EC 0,20 2,11 3,92 6,23 U -0,43 1,4 4,09 5,06

The EC clay shows small shrinkage until 950 °C. The onset of sintering starts at 917 °C for the EC clay. The IC and U clays start to shrink in the temperature range of 950 -1150 °C, since the values became positive. This range of temperatures coincides with the total collapse of illite as observed in the XRD analysis (Figure 4.6) and the spinel-amorphous silica recrystallizes to mullite phase at 990 °C.

At higher temperatures, between 1050-1150 °C, shrinkage increases for all clays, being much faster for the EC and U clays than IC clay. The IC clay reached the lowest total shrinkage (1.52 %) up to 1150 °C which indicates a good refractory behaviour as compared to the EC and U clays.

4.4.3 Bulk density and open porosity Table 4.4 shows the calculated bulk density and open porosity of the sintered clays at three different temperatures.

Table 4.4 Firing parameters at three different temperatures.

Bulk density (g cm-3) Open porosity (%) Clay 950 °C 1050 °C 1150 °C 950 °C 1050 °C 1150 °C IC 1,89 1,92 2,10 26,5 24,8 17,2 EC 1,91 2,12 2,52 26,5 16,3 0,31 U 1,81 1,99 2,27 28,1 22,0 7,3

The results indicate that the bulk density increases with the increase of temperature, while the open porosity decreases, following a non-linear relationship. At 950 °C, all studied sintered clays show similar bulk density and open porosity. Above 1050 °C the bulk density of the EC clay increases much faster than the other clays. The highest bulk density (2.52 g cm-3) and lowest porosity (0.3 %) corresponds to the EC clay sintered at 1150 °C. This is well correlated with dilatometric behaviour for the EC clay, showing the highest shrinkage (6.2 %) at this temperature.

4.5 Microstructural evolution analysis of fired samples (XRD) Figure 4.6 shows the evolution of the microstructure at temperatures between 950 °C and 1250 °C.

 1400  1400

950C  Quartz 1200 1050C  Quartz 1200    Hematite  Hematite 1000 1000  Illite   Feldspar  Feldspar 800  800 600 600

  Intensity/a.u. 400 400 Intensity/a.u.

200 Smoothed Y1 200              0     0 Smoothed Y 2   IC clay IC clay                    Smoothed Y 3    EC clay EC clay              Smoothed Y 3 U clay U clay

10 15 20 25 30 35 40 45 10 15 20 25 30 35 40 45 2/degree 2/degree

 1400  350  Mullite 1250C  Mullite 1150C 300  Quartz 1200  Quartz  Hematite  1000 Hematite 250  Feldspar 800 200

 600 150

 Intensity/a.u.

Intensity/a.u. 100  400  50 200                       0       0 Smoothed Y1   IC clay    IC clay                              Smoothed Y 2     EC clay    EC clay                   Smoothed Y 3 U clay U clay 10 15 20 25 30 35 40 45 5 10 15 20 25 30 35 40 45 2/degree 2/degree

Figure 4.6 X-ray diffractograms of clays fired at 950, 1050, 1150 and 1250 °C.

After firing at 450 °C, the intensity of the reflections of kaolinite and illite decrease leading to the formation of amorphous metakaolinite, (reaction described by Equation 1). Metakaolinite is a transition phase resulting of the dehydroxylation processes of kaolinite. A complete dehydroxylation of kaolinite is reached at 650 °C, (XRD results are shown in Figure 5 of paper 2).

Figure 4.6 shows that illite was present at 950 °C. The relict illite structure collapsed between 950 °C and 1050 °C. In this range of temperature, hematite (Fe2O3) began to crystallize only in the EC clay. At 990 °C, it is believed that the octahedral portion (containing aluminium, magnesium and iron) of the illite and metakaolinite structures begin to recrystallize to form spinel-phase (γ-Al2O3) and amorphous silica, reaction described by Equation (2). This transformation is shown by DSC analysis through the exothermic peak (Figure 4.5). In fact, several researchers have confirmed this transformation in well crystallized kaolins with high purity (Ptacek et al., 2010) (Ghorbel et al., 2008).

As the temperature increases above 1050 °C, the hematite peaks become more intense in the EC clay than the U clay. It is correlated with the highest iron content in the EC clay (6 %), as seen in Table 4.1.

At 1050 °C, the EC clay shows increased background intensity between 15° and 30°, 2θ, forming a “hump” which indicates the presence of amorphous phase. This is related to the formation of a large quantity of liquid phase from the breakdown of the clay mineral phases at sintering.

Mullite begins to crystallize somewhere in between 1050 °C and 1150 °C, (reaction described by Equation 3). At 1150 °C, also the quartz peaks are reduced in intensity, particularly for EC and U clays. The quantity of silica-rich liquid phase increases at 1150 °C as some accessory quartz dissolves. It coincides when feldspars start to melt as clearly seen by XRD patterns, and at 1250 °C, mullite and residual quartz are the only crystalline phases present for the IC and U clays, and additionally hematite is formed in the EC clay.

4.6 Microstructure of fired samples (SEM) SEM micrographs of the clays sintered at 1150 °C and heating rate 10 °C min-1, are showed in Figure 4.7.

quartz

microcracks

mullite quartz

hematite quartz

pore mullite

quartz

mullite microcracks

pore

Figure 4.7 SEM micrographs of fired samples at 1150 °C. Images to the left show the microstructure of polished samples. Images to the right show the size and shape of mullite crystals of the samples polished and etched in 20% HF acid for 10 min.

The polished and fired samples (left side) show microcracks around the quartz particles for the IC and EC clays, due to thermal mismatch with the liquid phase during cooling to room temperature.

The polished and etched fired samples (right side) reveal the presence of homogeneous microstructure characterized by needle shaped mullite. Mullite crystals are embedded in a glassy matrix, where single crystals seldom exceed a few µm in size. The mullite formed in these samples is well supported by XRD results at 1150 °C (Figure 4.6).

The EC sample (left side) shows a microstructure of a highly vitrified sample with large pores due to bubble formation inside the structure. The high liquid formation at low temperatures is well correlated with DSC results, which show the lowest onset of liquid formation at 1081 °C for this sample. The chemical composition (Table 4.1) for this sample also shows the highest content of flux elements that would promote liquid phase at lower temperatures than the other studied clays.

The IC and U samples (left side) reveal a more homogeneous structure with still small pores in the structure. Both samples also show a less vitrified appearance than the EC sample. This could be observed by DSC results, since the onset of liquid formation for the IC and EC is about 1175 °C and 1170 °C, respectively, which is 90 °C higher than that observed for the EC sample.

Finally, the EC sample shows iron-rich particles dispersed in the amorphous phase. This iron oxide is present in form of hematite (Fe2O3) as detected by the peaks in the X-ray diffraction results shown in Figure 4.6.

5. Conclusions

The results of the three studied raw clays from Bolivia and the effect of firing temperatures on the developed phases led to the following conclusions:

- Kaolinite and illite are the predominant clay minerals in all clays followed by montmorillonite as a third clay component in the EC clay. - Quartz and feldspars are present in the clays in variable amounts. - Potassium is present in all clays as a main fluxing element. - The highest amount of iron oxide, about 6 %, is detected for the EC clay, and the lowest, about 1.7 %, for IC clay. - Hematite is first detected at 950 °C and increased at higher temperatures for the EC clay being dispersed in the glassy matrix. - Mullite phase starts to form in all clays between 1050-1150 °C, and its content increases with the sintering temperature. - The onset of liquid formation and the onset of sintering start at the lowest temperature for the EC clay. - The EC clay had the largest shrinkage and obtained a low open porosity of 0.3 % after sintering at 1150 °C. - The quartz content in the samples is decreasing at sintering temperatures between 1050 °C and 1150 °C. - At the highest sintering temperature of 1250 °C resulted in formation of mullite, quartz hematite and amorphous phase.

From the results obtained in this study, it can be concluded that the EC and U clays provide excellent characteristics that enable use of these clays in the fabrication of products with red tonality, especially for use in bricks, roofing tiles and rustic floor tiles. While IC clay that exhibits relatively low iron content and some refractoriness, might be used as raw materials to replace both kaolin and silica in the white firing ceramics.

Finally, the clays from Ivirgarzama, Entre Rios and Uspha-Uspha are promising materials and they should be considered as valuable resources for the production of building ceramics.

Future work

Bolivia is a country with large deposits of non-metallic mineral resources, which have not yet been exploited industrially to any significant degree. The use of new technologies has not been implemented with the exception of some small factories of structural ceramics. Poverty is high and there is a need for development. This project aims at developing domestic Bolivian clay minerals for use as building materials. The most important requirements for a building material are that they have high strength and light weight. The future work of this project will involve strengthening of the clays by introduction of reinforcing phases such as sugar cane bagasse ashes, and investigation of the mechanical properties. The effect of the potassium and silica content on sintering behaviour will be studied. The effect of porosity on thermal insulation properties and weight will also be investigated.

References

Alcantara A.C., Beltrao M.S., Oliveira H.A., Gimenez I.F., Barreto L.S., Characterization of ceramic tiles prepared from two clays from Sergipe-Brazil, Appl. Clay Sci., 39 (2008) 160- 165.

Aramaki S. and Roy R., Revised phase diagram for the system alumina-silica, J. Am. Ceram. Soc., 45 (1962) 229-242.

Baccour H., Medhioub M., Jamoussi F., Mhiri T., Influence of firing temperature on the ceramic properties of Triassic clays from Tunisia, J. Mater. Proc. Technol., 209 (2009) 2812- 2817.

Bauluz B., Mayayo M., Constanza F.N., Cultrone G., González J., Assessment of technological of calcareous and non-calcaresous clays used for the brick-making industry of zaragoza (Spain), Appl. Clay Sci., 24 (12) (2003) 121-126.

Bétard F., Caner L., Gunnell Y., Bourgeon G., Illite neoformation in plagioclase during weathering: Evidence from semi-arid Northeast Brazil, Geoderma, 152 (12) (2009) 53-62.

Bowen N.L. and Greig J.F., The system: alumina-silica, J. Am. Ceram. Soc., 7 (1924) 238- 254.

Brownell W. E., Efflorescence Resulting from sulfates in clay raw materials, J. Am. Ceram. Soc., 41 (1949) 310-314.

Carretero, M.I., Dondi M., Fabbri B., Raimondo M., The influence of shaping and firing technology on ceramic properties of calcareous and non-calcareous illitic-chloritic clays, Appl. Clay Sci., 20 (6) (2002) 301-306.

Castelein O., Soulestin B., Bonnet J.P., Blanchart P., The influence of heating rate on the thermal behaviour and mullite formation from a kaolin raw material, Ceram. Int., 27 (2001) 517-522.

Chen C.Y., Lan G.S. Tuan W.H., Microstructural evolution of mullite during the sintering of kaolin powder compacts, Ceram. Int., 26 (2000) 715-720.

Chen Y.F., Wang M.Ch., Hon M.H., Phase transforamtion and growth of mullite in kaolin ceramics, J. Eur. Ceram. Soc., 24 (2004) 2389-2397.

Das S.K. and Dana K., Differences in densification behaviour of K- and Na-feldspar- containing porcelain bodies, Therm. Acta, 406 (1) (2003)199-206.

Dondi M., Clay materials for ceramic tiles from the Sassuolo Distric (Northern Apennines, Italy). Geology, composition and technological properties, Appl. Clay Sci., 15 (1999) 337- 366.

Dondi, M., Guarini G., Ligas P., Palomba M., Raimondo M., Chemical, mineralogical and ceramic properties of kaolinitic materials from the Tresnuraghes mining district (Western Sardinia, Italy), Appl. Clay Sci., 18 (2001) 145-155.

Esposito L., Salem A., Tucci., Gualtieri A., Jazayeri S.H., The use of nepheline-syenite in a body mix for porcelain stoneware tiles, Ceram. Int., 31 (2005) 233-240.

Ferrari S., Gualtieri A.F., The use od illitic in the production of stoneware tile ceramics, Appl. Clay Sci., 32 (2006) 73-81.

Garrels R.M., Montmorillonite/illite stability diagrams, Clays Clay Min., 32 (1984) 161-166.

Ghorbel A., Fourati M., Bouaziz J., Microstructural evolution and phase trasformation of different sintered kaolin powder compacts, J. Mater. Chem. Phys., 112 (2008) 876-885.

Gonzalez F., Romero V., Garcia G., Gonzalez M., Firing transformations of mixtures of clays containing illite, kaolinite and calcium carbonate used by ornamental tile industries, Appl. Clay Sci., 5 (1990) 361-375.

Grim R.E., Relation of the composition to the properties of clays, J. Am. Ceram. Soc., 22, (2006) 141-151.

Guggenheim D. and Martin R.T., Definition of clay mineral: Joint report of the AIPEA nomenclature and CMS nomenclature committes, Clays Clay Min., 43 (1995) 255-256.

Heaney P.J. and Veblen D.R., observation os the alpha-beta phase transition in quartz: A review of imaging and diffraction studies and some new results, Am. Mineral., 76 (1991) 1018-1032.

Heide K. and Földvari M., High temperature mass spectrometric gas-release studies of kaolinite Al2[Si2O5(OH)4] decomposition. Therm. Acta, (2006) 106-112.

Jordán M., Boix A., Sanfeliu T., de la Fuente C., Firing transformations clays used in the manufacturing of ceramic tiles, Appl. Clay Sci., 14 (1999) 225-234.

Köster H.M., Mineralogicasl and chemical heterogeneity of three standard clay mineral samples, Clay minerals, 31 (1996) 417-422.

Lee V. G., Yeh T.H., Sintering effects on the development of mechanical properties of fired clay ceramics, Mater. Sci. Eng., A,485 (2008) 5-13.

Lee, W.E., Souza G.P., McConville C.J., Tarvornpanich T., Iqbal Y., Mullite formation in clays and clay-derived vitreous ceramics, J. Eur. Ceram. Soc., 28 (2008) 465-471.

Lundin S., Electron microscopy of whiteware bodies, Trans. Ceram. Soc., 4 (1954) 383-390.

McConville C.J. and Lee W.E., Microstructural Development on Firing Illite and Smectite Clays Compared with that in Kaolinite, J. Am. Ceram. Soc., 88 (2005) 2267-2276.

Milheiro F.A., Freire M.N., Silva A.G., Holanda J.N., Densification behaviour of red firing Brazilian kaolinitic clay, Ceram. Int., 31 (2005) 757-763.

Mota L., Toledo R., Machado A.L., Holanda J., Vargas H., Faria R., Thermal characterisation of red clay from the Northern Region of Rio de Janeiro State, Brazil using an open photoacoustic cell, in relation to structural changes on firing, Appl. Clay Sci., 42 (2008) 168- 174.

Murray H., Traditional and new applications for kaolin smectite, and palygorskite: a general overview, Appl. Clay Sci., 17 (2000) 207-221.

Prudêncio M.I., Sequeira M.A., Paquet H., Waerenborgh J.C., Pereira L.C.J., Gouveia M.A., Clay mineral assemblages in weathered basalt profiles from Central and Southern Portugal: climatic significance, CATENA, 49 (2002) 77-89.

Ptacek P., Kubatova D, Havlika J., Brandstetr J., Soukal F., Opravil T., The non-isothermal kintic analysis of the thermal decomposition of kaolinite by thermogravimetric analysis, Powder Technol., 204 (2010) 222-227.

Schwaighofer B. and Muller H., Mineralogy and genesis of the Pugu Hill kaolin deposit, tanzania, Clay Minerals, 22 (1987) 401-409.

Stepkowska E. and Jefferis S., Influence of microstructure on firing colour of clays, Appl. Clay Sci., 6 (1992) 319-342.

Seynou M., Millogo Y., Ouedraogo R., Traore K., Tirlocq J., Firing transformations and properties of tiles from a clay from Burkina Faso, Appl. Clay Sci., 51 (2011) 499-502.

Valanciene V., Siauciunas R., Baltusnikaite J., The mineralogical composition on the colour of clay body, J. Eur. Ceram. Soc., 30 (2010) 1609-1617.

Vieira C., Sanchez R., Monteiro S.N., Characteristics of clays and properties of building ceramics in the state of Rio de Janeiro, Brazil, Constr. Build. Mater., 22 (2008) 781-787.

Wattanasiriwech D., Srijan K., Wattanasiriwech S., Vitrification of illitic clay from Malaysia, Appl. Clay Sci., 43 (2009) 57-62.

Wilson M. J., The origin and formation of clay minerals in soils" past, present and future perspectives, Clay Minerals, 34 (1999) 7-25.

Paper I

Powder Technology 221 (2012) 359–364

Contents lists available at SciVerse ScienceDirect

Powder Technology

journal homepage: www.elsevier.com/locate/powtec

Synthesis of homogeneously dispersed cobalt nanoparticles in the pores of functionalized SBA-15 silica

Edwin Escalera a,1, Mohamed A. Ballem b,⁎,1, José M. Córdoba b, Marta-Lena Antti a, Magnus Odén b a Division of Materials Science, Department of Applied Physics and Mechanical Engineering, Luleå University of Technology, SE-971 87 Luleå, Sweden b Nanostructured Materials, Department of Physics, Chemistry and Biology, Linköping University, 581 83, Linköping, Sweden article info abstract

Article history: Cobalt nanoparticles were prepared at room temperature by reducing cobalt sulfate heptahydrate with Received 18 July 2011 sodium borohydride using functionalized SBA-15 mesoporous silica as a hard template. It was found that Received in revised form 3 November 2011 both external and internal fuctionalization of silica walls play a crucial role on the inﬁltration and reaction Accepted 16 January 2012 of the reagents in the silica framework. Subsequent heat treatment of the impregnated silica at 500 °C in Available online 24 January 2012 air or nitrogen atmospheres leads to growth of crystals of the deposited cobalt and formation of cobalt oxide and cobalt nanoparticles, respectively. Dissolution of the silica template by NaOH resulted in well Keywords: Functionalization dispersed Co and Co3O4 nanoparticles ranging in size between 2 and 4 nm. The functionalization of the silica Cobalt was studied by FTIR, N2-physisorption, and thermogravimetric techniques and the obtained nanoparticles Cobalt oxide were characterized by XRD, TEM and EDX analysis. Nanoparticles © 2012 Elsevier B.V. All rights reserved. SBA-15 silica Sodium borohydride

1. Introduction reducing agent is needed. Recently, Bahadur and co-workers [15] have successfully synthesized iron-cobalt alloy nanoparticles by reduction In recent years the synthesis of metals and metal oxides in nano- from cobalt and iron chlorides using sodium borohydride, which is a particle form with well-deﬁned sizes, shapes, and crystallinity has widely used and powerful reducing agent in wet-chemical processes. received increasing attention because of their unique physicochemi- In this study, we present an effective approach to synthesize cal properties, which make them desirable in many technological monodisperse cobalt nanoparticles with a narrow size distribution applications such as sensors [1], catalysis [2], and hydrogenation pro- by modifying internal and external surfaces of SBA-15 mesoporous cesses [3]. The unique characteristics of cobalt metal, especially the materials with two different functional groups (see Scheme 1). magnetic properties, make it attractive in a wide range of applications, e.g. data information storage [4], recording devices, and mag- 2. Experimental details netic refrigeration including biomedical systems [5]. Cobalt nanoparticles can be synthesized using various techniques, 2.1. Materials such as thermal decomposition [6], pressure drop-induced decomposition [7], microemulsion methods [8], chemical reduction of cobalt Pluronic P123 (EO20PO70EO20, Aldrich), tetraethyl orthosilicate salts by borohydride derivates and hydrazine hydrate as the reducing (TEOS) (reagent grade, 98%, Aldrich), hydrochloric acid (purity≥37%, agent [9]. Depending on the synthesis method, cobalt particles puriss. p.a., Fluka, ACS Reagent, fuming), toluene (anhydrous 99,9%), with different morphologies and sizes can be obtained, e.g. dendritic Trimethylchlorosilane (TMCS≥99%, Aldrich), 3-Aminopropyl- structures [10], nanorods [11], and microspheres [12]. However, trimethoxysilane (APTMS 97%, Aldrich), sodium hydroxide pellets size and dispersion control of synthesized Co-nanoparticles remain (purity≥97%, purum. p.a., Fluka), cobalt (II) sulfate heptahydrate subjects to explore. (CoSO4·7H2O, 99%, Aldrich ), sodium borohydride (NaBH4, 99%, Selective functionalization of the silica surface has been used as a Aldrich), were used as received. method to synthesize Ag nanoparticles in the channels of SBA-15 by using formaldehyde as the reducing agent [13,14]. However, to 2.2. Synthesis reduce other metals such as cobalt from its salt solution, a strong The steps involved in the synthesis are illustrated in Scheme 1. Monodispersed mesoporous silica SBA-15 with a hexagonal pore ⁎ Corresponding author. Tel.: +46 13 28 2996; fax: +46 13 13 7568. E-mail address: [email protected] (M.A. Ballem). arrangement (Scheme 1a) was synthesized and used as molds for 1 Both authors contributed equally to this work. templating of cobalt and cobalt oxide nanoparticles. The detailed

Scheme 1. Steps involved in the synthesis of cobalt nanoparticles.

description of the synthesis route was reported by Sayari et al. [16], or N2 atmospheres to obtain cobalt oxide and cobalt metal, respec- where the triblock copolymer Pluronic P123 was used as a structure tively (Scheme 1f). The temperature was increased with a heating directing agent, and tetraethyl orthosilicate TEOS as silica source in rate of 10 °C /min to 500 °C and held for 5 h. Finally, to remove acidic condition. The obtained silica sample is labeled as (as-SBA-15). the silica template, a solution of NaOH (0.25 M) was mixed with The external surface of the as-SBA-15 sample was functionalized the sample and sonicated at 55 °C for 5 h; the synthesized nanopar- by trimethylchlorosilane (TMCS) according to the earlier procedure ticles (Scheme 1g) were collected using centrifugation (4100 rpm/ developed by Zhang et al. [14]. Typically, 1.0 g of as-synthesized silica 15 min). (as-SBA-15) was stirred into 100 mL of toluene in a round bottom flask for 30 min at room temperature. Subsequently, 15 mL of tri- 2.3. Characterization methylchlorosilane was added and heated at 80 °C for 8 h under stirring. The filtrated solid was washed twice with toluene and dried at The crystalline structure was determined by powder X-ray diffrac- room temperature (Scheme 1b). The obtained sample was labeled tometry (XRD) using a Siemens D 5000 diffractometer and Cu Kα ra- (TMCS-SBA-15). Subsequently, to remove P123 from the pores, the diation. Transmission electron microscopy (TEM) was performed functionalized silica was calcinated at 300 °C for 5 h (Scheme 1c). with an FEI Tecnai G2 microscope operated at 200 kV, and the chem- This sample is labeled (300-TMCS-SBA-15). ical composition determined by energy dispersive X-ray spectroscopy The modification of the internal walls of the silica was carried (EDX) in the TEM. For preparing TEM samples, the product of interest out using 3-aminopropyl-trimethoxisilane (APTMS), where – was dispersed in acetone and then deposited onto carbon copper

Si(CH2)3NH2 groups are grafted into the channels of silica through grids and allowed to dry before analysis. the reaction between silanol groups and 3-aminopropyl- Nitrogen adsorption–desorption measurements were performed trimethoxysilane (Scheme 1d). The procedure applied is a modifica- at 77 K using a Micromeritics ASAP 2020 surface area and porosity an- tion to what has been previously reported [13,17]. Typically, 0.5 g of alyzer. The samples were degassed at 373 K for 9 h before the mea- 300-TMCS-SBA-15 sample was suspended for 30 min in 80 mL of surement. The specific surface area was determined by the toluene, and then 5.0 mL of APTMS was added and stirred at room Brunauer–Emmett–Teller (BET) model [18] over a relative pressure fl temperature for 24 h. After that the solid was re uxed for 5 h by sox- (P/P0) range of 0.08–0.2; the pore size distribution was derived from let extraction in 80 mL of toluene in order to remove the physical the adsorption isotherm branch using Kruk–Jaroniec–Sayari (KJS) sorption of APTMS. The obtained sample was labeled (TMCS- method [19]. Finally the total pore volume was calculated from the

APTMS-SBA-15). amount of adsorbed N2 at P/P0 =0.975. The method to reduce cobalt salt sources inside channels of the The study of functional groups on the surfaces of silica by silica (Scheme 1e) was performed as follows: 0.5 g of dried TMCS- Fourier transformed infrared spectroscopy (FTIR) was performed APTMS-SBA-15 sample was transferred to a round bottom flask and using a Shimadzu FTIR-8400S spectrophotometer, using pressed mixed with 30 mL of cobalt (II) sulfate solution (0.02 M), and sonicat- KBr pellets. To prepare the pellets, 0.8 mg of sample and 120 mg ed for 15 min at room temperature. In order to remove the excess of of KBr powder were finely ground to remove scattering effects, cobalt precursor the sample was filtered and dried, and subsequently and the powder mixture pressed to a pellet size suitable for the

30 mL of NaBH4 aqueous solution (0.1 M) was added drop wise instrument. during magnetic stirring for 20 min to reduce the cobalt ions; the Thermogravimetric (TG) and differential scanning calorimetric mixture was then kept in a sonication bath for 15 min at room (DSC) analyses were performed using a Netzsch STA 449C Jupiter in- temperature. The obtained sample was labeled (TMCS-APTMS-Co- strument. The measurements were carried out in air and nitrogen. SBA-15). Approximately 20 mg of material was placed in a sintered alumina Growth of the cobalt and cobalt oxide nanocrystals was achieved crucible and the temperature was increased from room temperature by heat treatment of the sample TMCS-APTMS-Co-SBA-15 under air to 700 °C at a heating rate of 10 °C/min. E. Escalera et al. / Powder Technology 221 (2012) 359–364 361

Fig. 1. TEM micrograph of the silica template (SBA-15); the inset represents the pore structure.

3. Results and discussions

Monodispersed mesoporous silica SBA-15 has been used as a template for confined growth of cobalt and cobalt oxide nanoparticles. The overall morphology and pore structure of this template are shown in Fig. 1. The particles are rodlike shaped with a diameter of 0.4–0.5 μm and a length of 1–1.5 μm. Hexagonally ordered mesoporous channels run along the length of the particles, with a pore size of about 10 nm (see the inset). The functionalization process used is a well known method based on slow hydrolysis of an alkyl trimethylsilane and condensation of a compound with free OH-binding sites at the silica surface [20].In the present work the external surface of the synthesized SBA-15 silica was first modified with trialkylsilane (–Si(CH3)3). Thereby a highly hydrophobic surface was achieved, which proved to be sufficient to Fig. 3. TG (dashed line) and DTA (solid line) responses for (a) as-SBA-15, and (b) 300- avoid formation of large cobalt particles outside the silica channels. TMCS-SBA-15.

Fig. 2. FTIR spectra of (a) as-SBA-15, (b) TMCS-SBA-15, (c) 300-TMCS-SBA-15, and (d) TMCS-APTMS-SBA-15. 362 E. Escalera et al. / Powder Technology 221 (2012) 359–364

C–H asymmetric stretching vibration around 3092–2900 cm− 1 from the surfactant P123, and silanol groups Si–OH stretching vibration at 3496 cm− 1. The absorption intensity of silanol (Si–OH) groups decreases for functionalized SBA-15 with trialkylsilane groups (Fig. 2b). The C–H stretching vibration intensity increases because of the presence of a

higher amount of trimethyl (–CH3) groups anchored to silica. The absorption band at 830 cm− 1 is attributed to new bonds formed between Si–C atoms. Based on these observations, we conclude that the external silica surface has been successfully modiﬁed with hydrophobic groups. Calcination at 300 °C has previously been shown to successfully remove P123 from SBA-15 pore channels, [21] which is also shown to be the case here (Fig. 3). Fig. 2c shows the FTIR spectrum for this sample. The absorption intensity for C–H groups (3092–2900 cm− 1) decreased, while the intensity for silanol (Si–OH) groups at 3496 cm− 1 increased. These ﬁndings are in accordance with P123 loss and perhaps the formation of new silanol groups during calcination. The observed remaining absorption bands after calcination of C–H and Si–C are expected in the presence of the external

hydrophobic functional groups, i.e. trimethyl (Si–CH3)3 groups at 3092–2900 cm− 1 and Si–C at 830 cm− 1, respectively. The reaction between APTMS and Si–OH groups of the internal surface of 300-TMCS-SBA-15 sample is shown below:

≡ – þ ðÞ– ðÞ– →≡ – – ðÞ– þ Si OH CH3O 3 Si CH2 3 NH2 Si O Si CH2 3 NH2 3CH3OH Internal surface Hydrophilic agent Internal functionalization ð2Þ

Stretching vibrations from the amino groups (–NH2) normally ap- pear at 3380 and 3310 cm− 1 [22]. Here, they are only weakly seen because of an overlap with the broad band between 3500 and 3400 cm− 1 originating from hydroxyl groups present in the silica walls (Fig. 2d) A new band at 1530–1588 cm− 1 attributed to –C–N– stretching

vibration of primary amine (–CH2NH2) is observed when comparing Fig. 4. Nitrogen adsorption–desorption curves. (a) Isotherms for the samples at differ- the spectra before (Fig. 2c) and after (Fig. 2d) internal functionaliza- ent synthesis steps and (b) their corresponding pore size distributions. tion. Moreover, the intensity ratio of Si–OH/C–H decreased because

of the new C–H bonds present as –CH2– groups included in the at- tachment of the amino groups (–NH ) to the silica SBA-15, according During the external functionalization process, the silanol (–OH) 2 to the equation below: groups of the silica react with TMCS according to Eq. (1) to form a hydrophobic external surface. ≡ – − ðÞ− þ ðÞþ2 → ≡ – – ðÞ− −1− ðÞþ1 Si O Si CH2 3 NH2 Co H2O 6 Si O Si CH2 3 HN Co H2O 5 ≡ – þ – ðÞ→ ≡ – – ðÞþ ð Þ Internal functionalized surface Hydratedcobalt ion Stable complex aminecobalt ion Si OH Cl Si CH3 3 Si O Si CH3 3 HCl 1 External surface Hydrophobic agent External functionalization ð3Þ

The incorporation of these groups on the SBA-15 surface has been When the cobalt salt was dissolved in water the cations are hy- +2 qualitatively conﬁrmed by FTIR analysis. As can be seen in the Fig. 2a, drated (Co(H2O)6 ). Therefore, hydrated cobalt ions have formed a the spectrum of as-prepared SBA-15 sample shows a characteristic stable complex with amino groups. The anchored amine groups pro- absorption band of Si–O–Si groups in the range 1150–1062 cm−1, vide a stronger negative charge than the silanol groups [23], and

Fig. 5. TEM micrographs of (a) mesoporous silica with cobalt nanoparticles inside the pores (b) cobalt nanoparticles on a Cu/carbon grid after silica removal, and (c) particles size distribution histogram of the obtained nanoparticles and corresponding Gaussian ﬁt (solid line). E. Escalera et al. / Powder Technology 221 (2012) 359–364 363

Table 1 signals from Si, O, and Co, is consistent with silica SBA-15 being loaded N2 physisorption data of the samples at different synthesis steps. with cobalt nanoparticles. After dissolving the silica with NaOH solu- 2 3 Sample SBET [m /g] Vt [cm /g] Dp [nm] tion (Fig. 6b), the Co peaks became more pronounced. Cu is also detected and has its origin in the grid supporting the particles. As-SBA-15 167 0.20 8.75 300-TMCS-SBA-15 610 0.65 9.96 Fig. 7 shows XRD measurements for the sample (TMCS-APTMS- TMCS-APTMS-SBA-15 349 0.37 8.97 Co-SBA-15) after heat treatment at 500 °C under inert condition (N2 300-SBA-15 (calcinated silica) 781 0.70 9.97 gas), and in air, and the hexagonal crystal structures of ε-Co metal

SBET: BET surface area; Vt: total pore volume calculated at P/P0 =0.975; Dp: mesopore [26] and Co3O4 [27] are observed, respectively. Previously reported diameter calculated from the adsorption isotherm branch using the KJS-method. experimental data show that ε–Co is the most frequently found crystal structure of Co nanoparticles prepared by wet chemistry [28]. enhance the attraction of cobalt ions to the silica pores. The complete The formation of cobalt particles inside the silica (i.e. in sample modification route is illustrated in Scheme 1. TMCS-APTMS-Co-SBA-15) is a result of reduction of cobalt ions, As-SBA-15 and 300-TMCS-SBA-15 samples were analyzed by TG- which are attracted by APTMS groups; the reduction reaction occurs fi DTA analysis, in order to con rm that P123 was completely removed by generation of hydrogen gas from the hydrolysis NaBH4 + and that TMCS groups remain attached on the silica walls after calcina- 2H2O→NaBO2 +4H2 [29,30]. Depending on heat treatment, the crys- tion at 300 °C. The P123 template decomposes in the temperature range tal structure can be affected in the (TMCS-APTMS-Co-SBA-15) in two 180–300 °C, seen as an exothermic peak in Fig. 3a, which is ways. It has previously been shown that a temperature of about accompanied by weight loss. In the case of TMCS groups (Fig. 3b), de- 200 °C is sufficient for surface diffusion of cobalt atoms [28].Hence, composition and weight loss begin at about 430 °C. Hence, a distinct dif- above this temperature cobalt coalescence and crystal growth is ference between the decomposition temperatures of TMCS and P123 is expected. At approximately 700 °C a phase transformation between established. The obtained result confirms that P123 was removed hexagonal and face centered cubic structures occurs [31]. Our observa- completely from the inside of the channels at 300 °C, without affecting tions are in agreement with these results. However, despite the rela- the hydrophobic property of TMCS groups on the external surface. tively high heat treatment temperature of 500 °C only limited crystal Nitrogen physisorption isotherms and pore size distribution for growth occurred. This suggests a cobalt mobility that is affected by the samples as-SBA-15, 300-TMCS-SBA-15, TMCS-APTMS-SBA-15, the functionalized pore framework, especially the APTMS groups. and for the calcinated silica at 300 °C (300-SBA-15) are shown in Fig. 4. The isotherms for all the samples (Fig. 4a) are type IV isotherm curves with H1 hysteresis loops. Such hysteresis loops indicate narrow pore size distributions (Fig. 3b) and an intact porous framework during the functionalization steps, which is also apparent from the TEM observations (Fig. 5). The N2 physisorption data of all the samples are listed in Table 1.

Starting with the as-SBA-15 sample, the speciﬁc surface area (SBET) is 167 m2/g. After functionalization of the external surface and removal of the template by calcination at 300 °C, SBET increased to 2 610 m /g. The total pore volume (Vt), and the average pore size (Dp) also increase. The measured surface area is smaller compared with calcinated unfunctionalized silica (300-SBA-15) and the differ- ence is attributed to blocking of some pores with TMCS groups during the functionalization step [24]. After grafting of 3–aminopropyl- trimethoxysilane with silanols groups on the internal pore surface

(sample TMCS-APTMS-SBA-15), SBET and Vt decrease from 610 to 349 m2/g and from 0.6 to 0.37 cm3/g, respectively. The average pore diameter decreases by about 1 nm at the same time. Transmission electron microscopy (TEM) provides direct observation of the formation of cobalt nanoparticles inside the silica template. Fig. 5a shows that the mesoporous channels maintain their well-ordered structure after nanoparticle formation. The cobalt nanoparticles seen as dark spots are dispersed in the silica matrix. No bulk aggregation of the cobalt on the outer surface could be observed, which indicates that the Co is confined to the pores. It is also clear from Fig. 5a that the channels are only partially occupied by cobalt nanoparticles, and no formation of nanowires could be found. Fig. 5b shows ultra-fine cobalt nanoparticles with an irregular shape recovered after mild dissolution of the silica template in a NaOH solution. The size of the nanoparticles was in the range 2–4 nm. The average particle size was determined by image analysis of TEM micrographs by measuring the diameter of about 200 particles using DigitalMicrograph ™ 3.9.3 software for GMS 1.4.3 by Gatan Software Team. The result is plotted in Fig. 5c as a size distribution histogram. From the fit of a Gaussian function to the histogram the average diameter of the particles was determined to be 2.8±0.4 nm. The size distribution is narrow and the size of the obtained nanoparticles is small in comparison with that obtained using other techniques [25]. fi The presence of ultra- ne Co nanoparticles is further supported by Fig. 6. EDX elemental analysis profiles of (a) cobalt inside SBA-15 silica and (b) cobalt the EDX spectra presented in Fig. 6. The spectrum in Fig. 6a, with strong nanoparticles after dissolving the silica. 364 E. Escalera et al. / Powder Technology 221 (2012) 359–364

Fig. 7. Powder X-ray diffractograms of SBA-15 silica inﬁltrated with cobalt (sample TMCS-APTMS-Co-SBA-15 after heat treatment at 500 °C in N2 (a) and in air (b).

4. Conclusion [12] X. Yang, Q.-W. Chen, J.-Z. Zhang, Shape-controlled synthesis and self-assembly of hexagonal cobalt ultrathin nanoflakes, Materials Chemistry and Physics 113 (2–3) (2009) 675–679. A wet chemical process was successfully performed to synthesize [13] D.-H. Lin, Y.-X. Jiang, Y. Wang, S.-G. Sun, Silver Nanoparticles Confined in SBA-15 cobalt nanoparticles using NaBH4 as the reducing agent, and silica Mesoporous Silica and the Application as a Sensor for Detecting Hydrogen Peroxide, SBA-15 with both external and internal functionalized surfaces as Journal of Nanomaterials 10 (2008) 2008. the template. This route has proven to be a simple and efficient way [14] Z. Zhang, S. Dai, D.A. Blom, J. Shen, Synthesis of ordered metallic nanowires inside ordered mesoporous materials through electroless deposition, Chemistry of to synthesize ultra fine cobalt particles, with the possibility to obtain Materials 14 (3) (2002) 965–968. cobalt oxide nanoparticles, depending on the atmosphere used during [15] S.K. Pal, D. Bahadur, Shape controlled synthesis of iron-cobalt alloy magnetic nano- – the annealing step. particles using soft template method, Materials Letters 64 (10) (2010) 1127 1129. [16] A. Sayari, B.-H. Han, Y. Yang, Simple synthesis route to monodispersed SBA-15 silica rods, Journal of the American Chemical Society 126 (44) (2004) 14348–14349. Acknowledgement [17] J. Aguado, J.M. Arsuaga, A. Arencibia, M. Lindo, V. Gascón, Aqueous heavy metals removal by adsorption on amine-functionalized mesoporous silica, Journal of Hazardous Materials 163 (1) (2009) 213–221. This work was partially financially supported by the Swedish In- [18] S. Brunauer, P.H. Emmett, E. Teller, Adsorption of gases in multimolecular layers, ternational Development Cooperation Agency, Sida. Journal of the American Chemical Society 60 (1938) 309–319. [19] M. Kruk, M. Jaroniec, A. Sayari, Application of large pore MCM-41 molecular sieves to improve pore size analysis using nitrogen adsorption measurements, References Langmuir 13 (23) (1997) 6267–6273. [20] A.P. Rao, A.V. Rao, G.M. Pajonk, Hydrophobic and physical properties of the ambient [1] W. Belkacem, A. Labidi, J. Guérin, N. Mliki, K. Aguir, Cobalt nanograins effect on pressure dried silica aerogels with sodium silicate precursor using various surface the ozone detection by WO3 sensors, Sensors and Actuators B: Chemical 132 modification agents, Applied Surface Science 253 (14) (2007) 6032–6040. (1) (2008) 196–201. [21] E.M. Johansson, J.M. Córdoba, M. Odén, The effects on pore size and particle [2] E.-S.M. Duraia, K.A. Abdullin, Ferromagnetic resonance of cobalt nanoparticles morphology of heptane additions to the synthesis of mesoporous silica SBA-15, used as a catalyst for the carbon nanotubes synthesis, Journal of Magnetism and Microporous Mesoporous Materials 133 (1–3) (2010) 66–74. Magnetic Materials 321 (24) (2009) L69–L72. [22] S.W. Song, K. Hidajat, S. Kawi, Functionalized SBA-15 materials as carriers for [3] B.M. Vogelaar, A.D. van Langeveld, P.J. Kooyman, C.M. Lok, R.L.C. Bonné, J.A. Moulijn, controlled drug delivery: influence of surface properties on matrix−drug interac- Stability of metal nanoparticles formed during reduction of alumina supported tions, Langmuir 21 (21) (2005) 9568–9575. nickel and cobalt catalysts, Catalysis Today 163 (1) (2011) 20–26. [23] X. Liu, A. Wang, X. Yang, T. Zhang, C.-Y. Mou, D.-S. Su, et al., Synthesis of thermally [4] X.H. Liu, W. Liu, W.J. Hu, S. Guo, X.K. Lv, W.B. Cui, et al., Giant reversible stable and highly active bimetallic Au−Ag nanoparticles on inert supports, magnetocaloric effect in cobalt hydroxide nanoparticles, Applied Physics Letters Chemistry of Materials 21 (2) (2008) 410–418. 93 (20) (2008) 202502. [24] F. Hoffmann, M. Cornelius, J. Morell, M. Fröba, Silica-based mesoporous organic– [5] J.P. Wilcoxon, B.L. Abrams, Synthesis, structure and properties of metal nanoclusters, inorganic hybrid materials, Angewandte Chemie International Edition 45 (20) Chemical Society Reviews 35 (11) (2006) 1162–1194. (2006) 3216–3251. [6] M. Salavati-Niasari, F. Davar, M. Mazaheri, M. Shaterian, Preparation of cobalt nano- [25] H.-X. Wu, C.-X. Zhang, L. Jin, H. Yang, S.-P. Yang, Preparation and magnetic properties particles from [bis(salicylidene)cobalt(II)]-oleylamine complex by thermal decom- of cobalt nanoparticles with dendrimers as templates, Materials Chemistry and position, Journal of Magnetism and Magnetic Materials 320 (3–4) (2008) 575–578. Physics 121 (1–2) (2010) 342–348. [7] M. Huuppola, N. Doan, K. Kontturi, C. Johans, Evolution of size distribution in [26] Powder Diffraction File JCPDS 01-089-7094, International Center for Diffraction pressure drop induced decomposition synthesis of cobalt nanoparticles, Journal Data, Swarthmore, PA, 1992. of Colloid and Interface Science 344 (2) (2010) 292–297. [27] Powder Diffraction File JCPDS 01-073-1701, International Center for Diffraction [8] M. Wen, Q.-Y. Liu, Y.-F. Wang, Y.-Z. Zhu, Q.-S. Wu, Positive microemulsion Data, Swarthmore, PA, 1992. synthesis and magnetic property of amorphous multicomponent Co-, Ni- and [28] V.F. Puntes, K.M. Krishnan, A.P. Alivisatos, Colloidal nanocrystal shape and size Cu-based alloy nanoparticles, Colloids and Surfaces A: Physicochemical and control: the case of cobalt, Science 291 (5511) (2001) 2115–2117. Engineering Aspects 318 (1-3) (2008) 238–244. [29] S. Özkar, M. ZahmakIran, Hydrogen generation from hydrolysis of sodium [9] V.I. Simagina, P.A. Storozhenko, O.V. Netskina, O.V. Komova, G.V. Odegova, Y.V. borohydride using Ru(0) nanoclusters as catalyst, Journal of Alloys and Compounds Larichev, et al., Development of catalysts for hydrogen generation from hydride 404–406 (2005) 728–731. compounds, Catalysis Today 138 (3–4) (2008) 253–259. [30] R.E. Davis, E. Bromels, C.L. Kibby, Boron hydrides. III. Hydrolysis of sodium [10] Y. Li, J. Zhao, Y. Zhu, D. Ma, Y. Zhao, S. Hou, et al., Controlled synthesis of rice ear- borohydride in aqueous solution, Journal of the American Chemical Society 84 like cobalt microcrystals at room temperature, Colloids and Surfaces A: (6) (1962) 885–892. Physicochemical and Engineering Aspects 356 (1-3) (2010) 156–161. [31] J. Ahmed, S. Sharma, K.V. Ramanujachary, S.E. Lofland, A.K. Ganguli, Microemulsion- [11] W. Liu, W. Zhong, X. Wu, N. Tang, Y. Du, Hydrothermal microemulsion synthesis mediated synthesis of cobalt (pure fcc and hexagonal phases) and cobalt-nickel of cobalt nanorods and self-assembly into square-shaped nanostructures, Journal alloy nanoparticles, Journal of Colloid and Interface Science 336 (2) (2009) of Crystal Growth 284 (3–4) (2005) 446–452. 814–819.

Paper II

High temperature phase transformation in Bolivian kaolinitic-illitic clays E. Escalera1, R. Tegman1, M.-L. Antti1 and M. Odén2 1Division of Materials Science, Luleå University of Technology, 97187 Luleå, Sweden 2Nanostructured Materials, Department of Physics, Chemistry, and Biology, Linköping University, 581 83 Linköping, Sweden E-mail: [email protected]

Abstract. The thermal behaviour of two clays extracted from different locations in Bolivia is studied. The clays contain kaolinite, illite, quartz and small amounts of feldspars. Elements such as K, Na, Fe, and Mg are present in the clays in different amounts. The phase transformations are characterized from room temperature to 1300 °C. For both clays, kaolinite was completely transformed into metakaolinite when heated up to 650 °C, while illite remained unchanged. During further heating to 1050 °C, illite undergoes total dehydroxylation. Mullite was formed in the temperature interval of 1050-1150 °C and its formation rate was dependent on the amount of K and Fe present in the clays. The clay with higher amounts of K and Fe had an onset temperature for sintering at about 900 °C and an onset temperature for liquid formation at 1080 °C. This is a 90 °C lower onset temperature for sintering and 50 °C lower onset temperature for liquid formation than the clay with less amounts of K and Fe.

1. Introduction

The prospect of using clays as building materials is attractive for non-developed countries as it may lead to the development of new domestic industries when adding value to natural resources. Clay materials are the results of gradual chemical weathering over long periods of time from feldspars, and others silicate minerals, [1]. Clay deposits are considered as a complex mixture of different types of minerals. They can contain quartz, feldspars, micas, smectites, chlorites, as well as various exchangeable ions located either in the interlayers or in the internal structures such as iron- alkaline- and alkaline earth- ions, [2, 3].

Clay materials from tropical regions of Bolivia, located in the eastern lowlands including large sections of Amazonian rainforest and the Chaco plain, have traditionally been used to manufacture building materials. Building ceramics involves products such as roof-, floor- and wall tiles, and bricks. There are some required properties of clays to be suitable for such products. The most important properties of building ceramics are high strength and low porosity. Alkaline and alkaline earth ions are known as fluxing components and can promote the formation of a liquid phase at low temperatures during firing, which will increase the rate of densification and the strength, [4, 5]. Quartz improves the workability and increases the strength. The workability is especially important in bricks because they are produced mainly by extrusion in contrast to tiles that are mainly produced by semi-dry pressing, [6]. In addition, aesthetic aspects such as the colour are also important. For example, a red or red- brown ceramics can be achieved if the raw material contains Fe, which can be transformed into hematite (Fe2O3) during firing under oxidizing conditions, [7].

Knowledge of the quality and characteristics of the raw material is a main aspect prior to any clay- based process. In this context many studies have been conducted to ascertain the quality of raw clay based materials in countries with tile and brick production, such as Italy, Spain, Brazil, India, etc., [8, 9, 10, 11]. In particular, illite-kaolinite-rich clays have shown the best behaviour for building ceramics [12, 13]. Thus, the suitability of a clay in these applications is strongly dependent on its mineralogical, and chemical composition, [14, 15].

1 Corresponding author: [email protected] 1

The secondary components, such as alkaline, alkaline earth ions and iron affect the thermal behaviour and phase formation during heat treatment and consequently influence the final mechanical properties of the building ceramic, [16, 17]. Therefore the thermal behaviour and possible phase formations need to be considered in detail. There are some studies presenting results from thermal treatment, and just a few studies address these issues at a detailed level, [18].

The aim of this work was to characterize and study the phase formation and microstructural changes of two different natural clays from Bolivia, via a combination of several techniques such as DSC/TG, dilatometer, XRD and SEM, in order to determine the suitability of these clays for building applications.

2. Materials and methods

The Chapare region hosts the largest deposits of clays found in Bolivia. Ivirgarzama clay (IC) and Entre Rios clay (EC) deposits are located 220 and 290 km, respectively, from the city of Cochabamba. The coordinates and locations of the two main deposits are given in Figure 1.

In order to ensure representative samples, the clays were collected from several different parts of each deposit. The samples coming from the same deposit were mixed, sun-dried and then quartered into small fractions of about 1 kg. The small fractions were dried at 110 °C and then ground to a fine powder in agate mortar until all powder passed through a 315 mesh size sieve. The ground and sieved powder was used as representative samples for all experiments.

IC: S 17º 01´ 18.´´ W 64º 57´ 7´´ EC: S 17º 09´ 42´´ W 64º 30´ 10´´

Figure 1. Location map and coordinates of the clay deposits in the tropical region and valley of Cochabamba-Bolivia (Google maps, 2011).

2.1. Characterization of as received clay samples The material was characterised by x-ray powder diffraction (XRD). The XRD patterns were obtained with a Siemens D 5000 x-ray diffractometer with CuKα radiation (λ = 1.5418 Å) operating at tube voltage and current of 40 kV and 20 mA respectively. Diffraction patterns were recorded in the 2θ range between 5º and 50º with a step size of 0.02 degrees. The phases were identified from peak positions and intensities using reference data from the JCPDS-ICCD (2004).

Chemical analyses of the collected clays were performed by a certified analytic laboratory (ALS Scandinavia, Sweden) by inductively coupled plasma with atomic emission spectroscopy (ICP-AES).

2.2. Thermal analysis Differential scanning calorimetry and thermal gravimetry (DSC/TG) were performed to investigate the phase transformations and mass changes during heating. The experiments were performed in a Netzsch STA 449C Jupiter instrument, equipped with a Netzsch Aeolos QMS 403C mass spectrometer. The DSC/TG experiments were carried out on powder samples in an alumina crucible between room temperature and 1300 ºC at a heating rate of 10 °C min-1 in an air atmosphere.

The dilatometry experiments were performed in a Netzsch DIL 402C instrument. Prior to measurement, the compacted sample was dried overnight at 110 ºC. The samples were heated from room temperature to 1150 ºC in air atmosphere at a heating rate of 10 ºC min-1. The density and open porosity of the samples were determined using Archimedes principle.

2.3. Characterization of heat treated material The clays were heat treated in a chamber furnace in air atmosphere at the following temperatures; 450, 650, 950, 1050, 1150 and 1250 ºC. The heating rate was 10 °C min-1 up to the final temperature, and the samples were then furnace cooled. The selected temperatures were chosen based on the results of the thermal analysis. The phase identification of the sintered specimens was performed by x-ray diffraction as described above.

The morphology of the as received clays and heat treated specimens at 1150 ⁰C (before and after chemical etching with 20 wt % HF for 10 min) was studied by scanning electron microscopy (SEM) using a Jeol 6460-LV microscope.

3. Results and discussions

3.1. Characterization of as received clay samples XRD patterns from the two materials (IC and EC) are shown in Figure 2.

 Kaolinite  Kaolinite (IC)   (EC)  Illite  Illite ..Quartz ..Quartz ^ Feldspar ^ Feldspar Montmorillonite



Intensity/a.u.  intensity/a.u. ^                                ^ ^       

10 15 20 25 30 35 40 45 10 15 20 25 30 35 40 45 2θ/degree 2θ/degree

Figure 2. X-ray powder diffractograms of as received IC and EC clays.

The main clay minerals identified in both samples are kaolinite (2θ: 12.3º, 20º, and 24.9º), and illite (2θ: 8.8º, 17.6º, and 19.8º). Small amounts of feldspar mineral are also detected in both samples. Additionally, the EC sample also contains small amounts of montmorillonite mineral (2θ: 5.8º and 29.5º). The non-plastic component present in the clays is quartz (2θ: 20.8º, 26.7º, and 36.5º).

Quartz is the dominating crystalline phase in both clays. The relative quartz content in the IC clay was estimated to be about twice as high compared with the EC clay. The presence of quartz in clays has been suggested to enhance the workability and diminish shrinkage and the risk of crack formation during drying. Moreover, quartz has proved to be important to increase strength and toughness in some ceramic based materials, [19]. For example, Vieira, [10] concluded that an addition of quartz

3 was necessary to improve workability of some clays from Rio de Janeiro (Brazil), especially for brick fabrication. Likewise, some clay from Burkina Faso needed an addition of quartz sand, feldspar, and talc to obtain the required tile performance, [20].

Table 1 shows the chemical composition of the clays expressed as oxides. Both clays show SiO2 and Al2O3 to be the predominant components. The relatively high SiO2 content is a result of the presence of free quartz in the clays, and correlate well with XRD analysis (Figure 2). The SiO2/Al2O3 ratio is higher for the IC clay (4.1) compared with the EC clay (3.5), which indicates a larger amount of free silica in the IC clay. The IC clay has low levels of iron oxide (1.67 %), in comparison with the EC clay (5.87 %). An iron oxide content of almost 6 wt % in the EC clay was sufficient to give a brick red colour after firing. The IC clay, with lower amount of iron oxide showed an off-white appearance after firing. Clays with low iron oxide content (<1.5wt %) are considered suitable material for white ceramic products, [21].

Table 1. Chemical composition in oxide form (wt %) of the IC and EC clays.

SiO2 Al2O3 Fe2O3 K2O Na2O CaO MgO TiO2 SO3 LOI* IC clay 71.8 17.5 1.67 2.32 0.12 0.10 0.81 0.96 0.02 6.10 EC clay 65.9 18.8 5.87 3.39 0.82 0.18 1.76 0.87 <0.01 6.75 *LOI = loss on ignition

The iron content in the EC clay is probably accommodated in the structure of illite and montmorillonite phases, since no iron-mineral was detected by XRD in the raw samples. This is in good agreement with results from Mössbauer spectroscopy and IR spectroscopy on clays, demonstrating that most of the total Fe amount is located as a structural element in the clay minerals, mainly in the illite and montmorillonite phases, [3, 22].

Table 1 also shows that the EC clay contains a higher amount of potassium, sodium and magnesium than the IC clay. Most of these elements are present as interlayer cations mainly in the illite and montmorillonite phases. These components contribute to glass phase formation in ceramics during the sintering step, [23, 24]. And also, the alkali components are able to form low temperature melting eutectics with other components, making it possible to reach a higher densification at lower temperatures, [25, 26].

There are very low levels of calcium (Table 1), which indicates minimal CaCO3 content in the clays. This is beneficial for densification because a higher content could lead to porosity, hinder mullite formation and instead lead to formation of gehlenite and wollastonite, [9, 13], A study on clays from Sergipe (Brazil) with up to 10 wt % of CaO, led to high porosity in the sintered specimens making them unsuitable for roof tiles manufacturing, [27].

It is worth pointing out that sulphur containing minerals such as pyrite (FeS2) and alunite (2KAl3SO4(OH)6) have not been detected. This is also in accordance with the chemical analysis (Table 1), which shows very low sulphur contents (expressed as SO3). Sulphur containing minerals have been found in other studies, for example in some clays from Rio de Janeiro (Brazil), [10] and from Western Sardinia (Italy), [8]. Both studies showed that sulphur minerals are undesirable because they may lead to efflorescence after drying and release of acid gases such as SO2 and SO3 during the firing stage.

3.2. Thermal analysis The results obtained by DSC and TG are shown in Figure 3. Phase changes and chemical reaction resulting in thermal exchange are recorded as endothermic and exothermic peaks, and the temperatures corresponding to these peaks are summarized in Table 2. Mass losses as well as suggested types of reactions during heating are also given in the table.

Figure 3. Thermograms for IC and EC clays from DSC and TG.

The DSC curves for both samples (Figure 3) show three endothermic peaks. The first peak at 100 ºC represents evaporation of absorbed moisture. The second peak at 498 ºC and 497 ºC for IC and EC samples respectively, is due to the elimination of water molecules through dehydroxylation of the kaolinite and illite, [28]. The third peak at 574 ºC corresponds to the α↔β quartz inversion and it occurs at the same temperature for both samples.

The weight losses between room temperature and 300 ºC are 1.31 % (IC) and 2.24 % (EC). The higher weight loss for the EC clay is due to the larger presence of montmorillonite fraction. Montmorillonite is known to have high water absorption capacity. There is a main weight loss between 300-700 ºC for both samples, which is due to the dehydroxylation. After the dehydroxylation, metakaolinite and illite anhydride structures are formed, [22].

Table 2. Summary of DSC and TG data of the studied clays. Mass loss (wt %) Peak at max. temperature (ºC) at temp. range (ºC) Sample RT-300 300-700 700-1300 Dehydroxylation α-β quartz Si-Al Onset of liquid inversion spinel formation IC clay 1.31 4.43 0.35 498 574 990 1175 EC clay 2.24 4.07 0.49 497 574 -- 1081

The inflexion point around 990 ºC for the IC clay is suggested to be related to the formation of cubic spinel γ-Al2O3 formed from metakaolinite, [23]. The transformations in the temperature range 700- 1300 ºC occurred with no or only minor weight changes. The onset of liquid formation is seen at 1175 ºC for the IC clay and 1081 ºC for the EC clay respectively. The lower onset temperature for liquid formation of the EC clay is due to its higher alkali and iron (K2O+Na2O+Fe2O3) content (10.1 %) compared to the IC clay (4.1%) (Table 1).

Dilatometer curves for both clays are shown in Figure 4. The shapes of the curves are similar. A small shrinkage after moisture evaporation is seen up to 150 ºC followed by a continuous expansion up to 600 ºC, related to dehydroxylation of kaolinite and illite. This expansion is more pronounced for the IC sample (Figure 4), which is likely related to its higher content of kaolinite. An expansion at 574 °C resulting from the α↔β quartz inversion is seen in both samples.

dL/Lo α⇔ β (IC) 0,005 574°C (EC)

0,000

-0,005

-0,010 C C C C C ° 600 ° 950 ° 917 ° 150 ° -0,015 1100 ∆ -0,020 shrinkage

-0,025

-0,030

100 200 300 400 500 600 700 800 900 1000 1100 Temperature/oC

Figure 4. Dilatometer measurements for IC and EC clays up to 1150 °C.

Above the quartz inversion the samples shrink slightly between 600 and 900 ºC. This behaviour is associated with the gradual dehydroxylation of the illite, giving rise to the formation of Al-Si spinel. The pronounced shrinkage between 917 - 950 ºC for both clays indicate onset of sintering in the samples. The sintering accelerates above 950 ºC due to the formation of a liquid phase. This liquid phase results in better densification of the clays.

The shrinkage at 1100 °C for the EC sample is 2.7 % larger compared with the IC sample, which indicates that the EC sample is able to form more liquid phase at lower temperature and sinter faster in comparison with the IC sample. This is well correlated with the DSC results showing onset of melting 80 °C earlier for this clay (Figure 3).

The density of the IC and EC clays after dilatometer runs up to 1150 ºC were 2.1 and 2.5 g cm-3, respectively. The open porosity was 17 % for the IC clay and 0.3 % for the EC clay.

The densification behaviour in the EC sample is influenced by the amount of alkaline oxides (K2O and Na2O) that act as flux materials. The amount of alkaline oxides is higher in the EC sample (4.21 %) than in the IC sample (2.44 %), see Table 1. This is due to the relatively larger amount of illite and feldspar in the EC sample, see Figure 2. Moreover, the amorphous and mullite phases play an important role in the densification process.

3.3 Characterization of heat treated material XRD diffractograms of fired samples are shown in Figure 5. They show that kaolinite is completely transforming into amorphous metakaolinite up to 650 ºC in both samples. Dehydroxylation of the kaolinite is consistent with TG and DSC analysis (Figure 3) which show a rapid weight loss at temperature range 400-700 °C which correspond well with the exothermic reaction around 500 °C. Consistent with dehydroxylation of kaolinite its diffraction peak disappears, see Figure 5.

At 950 ºC peaks belonging to illite are less intense but still detectable in both samples, while above 1050 ºC illite is not detected. This is suggested to be due to the collapse of the illite structure in this temperature range, [15, 23]. However, the feldspars are still detected at 1050 ºC.

(IC)  Mullite (EC)  Mullite Quartz   Quartz    Kaolinite  Kaolinite  Illite  Illite  Hematite  Hematite  ^ Felsdpar  ^ Felsdpar                                 o 1250oC  1250 C              o      o   1150 C ^ ^      1150 C   1050oC     ^ ^ 1050oC ^        950oC   ^    ^   ^   950oC    

  

o    ^ 650 C  ^      ^      650oC       ^      o     ^  450 C    o      ^ ^   450 C     

10 20 30 40 10 20 30 40 2θ/degree 2θ/degree

Figure 5. X-ray diffractograms of samples fired at different temperatures for IC and EC clay.

Hematite (Fe2O3) begins to crystallize at 950 °C in the EC clay and at 1150 ºC in the IC clay. For the EC samples at 1250 ºC the intensity of the hematite peaks was significantly higher compared with the lower temperatures. This is in good agreement with the gradually increasing red colour with temperature of the EC clay above 950 °C. Hematite has been demonstrated to be the main colorant responsible for the red or red-brown colour in ceramic materials, [7].

Mullite was detected in both samples at 1150 °C and the amount increased at 1250 °C (Figure 5). Several investigations of the evolution of mullite have shown mullite formation at 1100 ºC from kaolinite and at higher temperatures needle-shaped mullite grains may form, [29, 30, 31]. Variables affecting the mullite formation include type of fluxing agents, such as K, Na and Fe elements in the clay structure, [4]. Mullite is a desired phase in building ceramics, as it increases the strength of the final product.

The presence of an amorphous glassy phase formed during cooling would result in a broad diffraction hump, which is weakly seen between 15º- 30º/2θ for the EC clay at temperatures above 1150 ºC (Figure 5). In the IC clay the expected amount of glassy phase is much smaller due to its smaller amount of K and Mg, (Table 1). Hence, for this clay the amorphous hump is even weaker. The presence of a glassy phase was also observed in illite containing clays from Malaysia with the same K content, which formed a liquid phase at lower temperatures (1100 °C) and favoured the mullite- crystals growth during cooling, [15]. At 1250 °C mullite, quartz, and hematite are the only existing crystalline phases in both clays. This is an advantage from an application point of view compared with other clays, which have shown residues of undesirable phases, such as cristobalite and sillimanite at these temperatures, [32].

3.4. Morphology and microstructure The SEM micrographs of the raw clay powder samples (Figure 6a and b) show the morphology of the particles. Some of the detected phases are identified based on local chemical analysis by EDS. The kaolinite is identified to have an irregular sheet-like morphology. Figure 6(b) shows large flakes stacked together with montmorillonite and illite forming agglomerates. Montmorillonite has ability to absorb water from the moisture in air.

Micrographs of fired samples at 1150 ºC are shown in Fig 6c and d. Here quartz and hematite are identified. The hematite is evenly dispersed in the glassy matrix. A series of microcracks around the quartz particles within the glass matrix is also observed. The microcracks are likely generated by the thermal mismatch between the crystalline quartz particles and the glass matrix.

Micrographs of HF-etched samples (6e and f) show the morphology of mullite phase formed in both samples. It is clearly seen that mullite is mainly made up of thin needles (length 6-10 µm), which is expected, [18, 29].

Illite-smectite kaolinite

quartz hematite quartz microcracks pore

quartz quartz

mullite mullite

Figure 6. SEM micrographs showing morphology of the raw material (a) IC clay, (b) EC clay and microstructure of the sintered clays at 1150 °C (c) IC clay and (d) EC clay. Mullite crystals at 1150 °C (e) IC clay and (f) EC clay.

4. Conclusions

The feasibility of naturally occurring clays from Bolivia to be used as raw material for building ceramics has been investigated. In this context the thermal behaviour and the microstructure evolution during sintering have been investigated. Mullite is formed in the clays during heating and the formation was favoured by a liquid phase formed. Efflorescence formation is not seen since the amounts of sulfur and carbonates are low. Given the beneficial microstructure evolution during firing

8 we conclude that these clays have high potential as suitable raw material for ceramic building applications.

Furthermore, the characteristics of two types of clay from Bolivia and the effect of firing temperatures on different phases obtained led to the following conclusions:

• Kaolinite and illite are the main clay components along with quartz for both samples. • Montmorillonite is detected as a third clay component in the EC sample. • In both clays, kaolinite phase dehydroxylated into metakaolinite during heating up to 650 ºC. • At 950 °C quartz and feldspars remain unchanged, and illite is gradually decomposed at 1050 °C. • The onset of liquid formation is about 1081 °C for the EC clay and about 1175 °C for the IC clay. • Mullite phase is formed in both samples between 1050 and 1150 °C and the mullite content increases with temperature. • The onset of sintering is about 917 °C for the EC clay and about 950 ºC for the IC clay. • K and Fe facilitate the mullite phase transformation and onset of liquid formation, but have no influence on the dehydroxylation. • When sintered at temperatures up to 1150 °C the open porosity in the EC clay is very low about 0.3 %.

From the results obtained in this study, the clays from Ivirgarzama and Entre Rios deposits are promising materials and they should be considered as valuable resources for the production of building ceramics.

5. Acknowledgments This work was financially supported by the Swedish International Development Cooperation Agency, SIDA.

6. References

1. J.C, H., Crystallinity of kaolin minerals and their weathering sequence in some soils from Nigeria, Brazil and Colombia. Geoderma, (1980) 24(4): p. 317-325. 2. Betard, F.O., et al., Illite neoformation in plagioclase during weathering: Evidence from semi- arid Northeast Brazil. Geoderma, (2009) 152(12): p. 53-62. 3. Prudêncio, M.I., et al., Clay mineral assemblages in weathered basalt profiles from central and southern Portugal: climatic significance. CATENA, (2002) 49(12): p. 77-89. 4. Li, J., et al., Effects of different potassium salts on the formation of mullite as the only crystal phase in kaolinite. Journal of the European Ceramic Society, (2009) 29(14): p. 2929-2936. 5. Das, S.K. and K. Dana, Differences in densification behaviour of K- and Na-feldspar- containing porcelain bodies. Thermochimica Acta, (2003) 406(1): p. 199-206. 6. Carretero, M.I., et al., The influence of shaping and firing technology on ceramic properties of calcareous and non-calcareous illitic-chloritic clays. Applied Clay Science, (2002) 20(6): p. 301-306. 7. Stepkowska, E.T. and S.A. Jefferis, Influence of microstructure on firing colour of clays. Applied Clay Science, (1992) 6(4): p. 319-342. 8. Dondi, M., et al., Chemical, mineralogical and ceramic properties of kaolinitic materials from the Tresnuraghes mining district (Western Sardinia, Italy). Applied Clay Science, (2001) 18(34): p. 145-155. 9. Bauluz, B., et al., Assessment of technological properties of calcareous and non-calcareous clays used for the brick-making industry of Zaragoza (Spain). Applied Clay Science, (2003) 24(12): p. 121-126.

10. Vieira, C.M.F., R. Sanchez, and S.N. Monteiro, Characteristics of clays and properties of building ceramics in the state of Rio de Janeiro, Brazil. Construction and Building Materials, (2008) 22(5): p. 781-787. 11. Manoharan, et al., Analysis of temperature effect on ceramic brick production from alluvial deposits Tamilnadu, India. Applied Clay Science, (2011) (54): p. 20-25. 12. Aydin Aras, The change of phase of composition in kaolinite- and illite- rich clay-based ceramic bodies. Applied Clay Science, (2004) 24: p. 257-269. 13. Gonzalez-Garcia, F., et al., Firing transformations of mixtures of clays containing illite, kaolinite and calcium carbonate used by ornamental tile industries. Applied Clay Science, (1990) 5(4): p. 361-375. 14. Michele, D., Clay materials for ceramic tiles from the Sassuolo District (Northern Apennines, Italy). Geology, composition and technological properties. Applied Clay Science, (1999) 15(34): p. 337-366. 15. Wattanasiriwech, D., K. Srijan, and S. Wattanasiriwech, Vitrification of illitic clay from Malaysia. Applied Clay Science, (2009) 43(1): p. 57-62. 16. Mota, L., et al., Thermal characterization of red clay from the Northern Region of Rio de Janeiro State, Brazil using an open photoacoustic cell, in relation to structural changes on firing. Applied Clay Science, (2008) 42(2): p. 168-174. 17. Gualtieri, M.L., et al., Thermal conductivity of fired clays: Effects of mineralogical and physical properties of the raw materials. Applied Clay Science, 49(3): p. 269-275. 18. Lee, W.E., et al., Mullite formation in clays and clay-derived vitreous ceramics. Journal of the European Ceramic Society, (2008) 28(2): p. 465-471. 19. R.W, S., Toughening mechanisms for ceramic materials. Journal of the European Ceramic Society, (1992) 10(3): p. 131-142. 20. Seynou, M., et al., Firing transformations and properties of tiles from a clay from Burkina Faso. Applied Clay Science, 51(4): p. 499-502. 21. Njoya, A., et al., Genesis of Mayouom kaolin deposit (western Cameroon). Applied Clay Science, (2006) 32(1-2): p. 125-140. 22. Castelein, O., et al., 57Fe Massbauer study of iron distribution in a kaolin raw material: influence of the temperature and the heating rate. Journal of the European Ceramic Society, (2002) 22(11): p. 1767-1773. 23 McConville, C.J. and W.E. Lee, Microstructural Development on Firing Illite and Smectite Clays Compared with that in Kaolinite. Journal of the American Ceramic Society, (2005) 88(8): p. 2267-2276. 24. Esposito, L., et al., The use of nepheline-syenite in a body mix for porcelain stoneware tiles. Ceramics International, (2005) 31(2): p. 233-240. 25. Osborn E. and Muan A., Phase equilibrium diagrams of oxide systems, American Ceramic Society and the Edward Orton Jr., (1960). 26. Stjernberg, J., et al., Microstructural characterization of alkali metal mediated high temperature reactions in mullite based refractories. Ceramics international, (2010) 36: p. 733- 740. 27. Alcantara, A.C.S., et al., Characterization of ceramic tiles prepared from two clays from Sergipe Brazil. Applied Clay Science, (2008) 39(34): p. 160-165. 28. Heide, et al., High temperature mass spectrometric gas-release studies of kaolinite Al2[Si2O5(OH)4] decomposition. Thermochimica Acta, (2006): p. 106-112. 29. Castelein, O., et al., Shape, size and composition of mullite nanocrystals from a rapidly sintered kaolin. Journal of the European Ceramic Society, (2001) 21(13): p. 2369- 2376. 30. Chen, Y.-F., M.-C. Wang, and M.-H. Hon, Phase transformation and growth of mullite in kaolin ceramics. Journal of the European Ceramic Society, (2004) 24(8): p. 2389-2397. 31. Stjernberg, J., et al., Extended studies of degradation mechanisms in the refractory linig of a rotary kiln for iron ore pellet production. Journal of the European Ceramic Society, (2012). 32. Mahmoudi, S., E. Srasra, and F. Zargouni, Firing behaviour of the lower cretaceous clays of Tunisia. Journal of African Earth Sciences, 58(2): p. 235-241.