Mecanosíntesis y caracterización de materiales multiferroicos nanoestructurados.

Mecanosíntesis y caracterización de materiales multiferroicos nanoestructurados (Mechanosynthesis and characterization of nanostructured multiferroic materials) Tesis presentada para optar al título de Doctor por: Eva Gil González Director-Tutor: Dr. Luis Allan Pérez Maqueda Co-director: Dr. Antonio Perejón Pazo Instituto de Ciencia de Materiales de Sevilla (ICMSE-CSIC) Departamento de Química Inorgánica (Universidad de Sevilla) Sevilla, 2017 A mis padres y familia i AGRADECIMIENTOS Llega la hora de hacer balance sobre esta etapa de mi vida, que sin duda ha estado marcada por esta tesis doctoral. Por ello, quisiera dedicar unas breves palabras a todas aquellas personas que de una manera u otra han contribuido a la realización de este trabajo y han hecho que el día a día haya sido mucho más agradable y llevadero. En primer lugar quisiera agradecer a Luis la confianza depositada en mí y el haberme dado la oportunidad de llevar a cabo este proyecto. A Antonio, por la transferencia de conocimiento, su ayuda constante desde el primer hasta el último momento y la gran paciencia que ha demostrado tener. A Pedro, por inculcarme su visión crítica al estudiar la bibliografía científica. A José Manuel Criado, por transmitirme parte de su meticuloso orden y limpieza en el trabajo de laboratorio. Al profesor Michael Hayward de la Universidad de Oxford, por aceptarme en su grupo de investigación y por la caracterización magnética de mis materiales. Al profesor Jan Subrt de la Academia de Ciencias Checa, por la cálida acogida y la atención constante que recibí durante mi estancia en su grupo de investigación. Al personal técnico del ICMSE, por la gran ayuda prestada ante múltiples cuestiones. También quisiera dedicarle unas palabras a mis compañeros de batalla durante toda mi etapa como investigadora, puesto que de cada uno de ellos he aprendido algo. Para ello seguiré un orden cronológico. Agradezco a los estudiantes de doctorado que comenzamos nuestras carreras científicas simultáneamente en Abengoa Research, particularmente a Pau y Nicolás, ya que las grandes “glorias” sin vosotros ii no hubieran sido tan llevaderas. A Verónica Carcelén, por todo lo que aprendí de ella. A mis compañeros de laboratorio de la UAM, Marta, Azin y Moha y a aquellas personas que hicieron que mi estancia en Madrid fuera tremendamente positiva. A Alex y Fon, por ser los mejores compañeros de piso que nadie pueda imaginar. De manera muy especial a mis compañeros de laboratorio del ICMSE, puesto que con ellos he compartido mi día a día además de infinitas confidencias. A Cristina, por su increíble disposición a ayudar a todo el mundo desinteresadamente y acabar con la lista de nombres de reconocidos científicos. A Rocío, por haberme acompañado en los inicios de esta tesis. A Alma, por sus contrastes tan peculiares y el cariño que se deja coger. A Bea, por alegrarnos los días de trabajo en el laboratorio. A Mónica, por su sentido del humor tan especial que tanto se parece al mío y por ayudarme a sacar las bolas del molino que se me cayeron por el fregadero. A Ernesto, Carlos Latre no te llega ni a las suelas de los zapatos. A Hamid, el hispano-iraní. A los demás compañeros del ICMSE: Juan, Sara, Antonio, Jaime, César y Mariana. A las personas que hicieron que mis estancias en el extranjero fueran memorables: Midori, Rona, Monika, Eva y los coetáneos del NOOC. A mis amigos: a mi archiconocida Tamara, por nuestra longeva amistad; a Juanma, porque echo de menos salir contigo los fines de semana, fines de año y ferias, ¡vuelve ya!; a Clara, Elena y Gabri, los músicos del Aljarafe; a Jose, por los numerosos planes que hacemos desde que nos conocemos; a Isa, porque el tiempo y la distancia no deteriora nuestra amistad; a Rafa, por sufrir y compartir información sobre los trámites de este programa de Doctorado; a Jaime, por las cervezas en la terracita y las tertulias sobre empleo público. iii Por último, a los verdaderamente responsables de que haya llegado hasta aquí, mi familia. A mis padres Miguel y Carmen, por hacerme la vida tremendamente fácil, sin vosotros sencillamente no hubiera sido posible. A mi hermana Irene, por aguantarme y ser la persona a la que más me gusta chinchar. A mis abuelos, por todo el cariño que me habéis dado y me dais, no hay ni un solo día que no piense en vosotros. A mi tía Juana, por corregirme el resumen y mil cosas más. A Carlos, por dedicarme lo más valioso que tiene y que nunca podré devolverle: su tiempo. A todos vosotros, ¡Gracias de corazón! v RESUMEN El óxido mixto de hierro y bismuto, BiFeO3, con estructura tipo perovskita, es uno de los materiales magnetoeléctricos más estudiados, ya que a temperatura ambiente presenta simultáneamente propiedades ferroeléctricas y antiferromagnéticas. A pesar de su versatilidad y alto potencial de uso como electrocerámica, su implementación se ve frenada por varios factores, tales como la difícil obtención de fases puras y el débil acoplamiento magnetoeléctrico. Concretamente, es extremadamente difícil la obtención del material libre de fases secundarias mediante procedimentos convencionales, como la reacción en estado sólido a partir de óxidos o carbonatos. Así, los tratamientos prolongados a altas temperaturas empleados en los métodos de síntesis y sinterización tradicionales favorecen la aparición de fases secundarias que, inevitablemente, deterioran las propiedades físicas del material. Por lo tanto, es de gran interés buscar métodos alternativos para la síntesis y sinterización de la perovskita BiFeO3. Asimismo, explorar distintas estrategias para mejorar las propiedades físicas, tales como la sustitución parcial del catión Bi3+ por cationes isovalentes de tierras raras (RE3+), también podrían suponer un impulso para la implementación de este material. Esta tesis doctoral abarca la preparación mediante mecanosíntesis de óxidos mixtos de fórmula general Bi1-xRExFeO3 con estructura tipo perovskita, donde el bismuto ha sido sustituido parcialmente por cationes de tierras raras: yterbio y samario. Ambas series (Bi1-xYbxFeO3 y Bi1xSmxFeO3) se han caracterizado mediante un análisis exhaustivo de su xiii CONTENTS AGRADECIMIENTOS ...................................................................................... I RESUMEN ......................................................................................................... V ABSTRACT ....................................................................................................... IX CONTENTS ................................................................................................... XIII 1. INTRODUCTION ...................................................................................... 1 1.1. MULTIFERROIC MATERIALS ......................................................... 1 1.2. THE PEROVSKITE OXIDE BiFeO3 ................................................... 5 1.2.1. BiFeO3 crystal structure .............................................................. 5 1.2.2. Multiferroism in BiFeO3 ............................................................. 7 1.2.3. Potential applications of BiFeO3 ................................................ 9 1.2.4. Challenges in bulk BiFeO3 ........................................................ 12 1.2.5. Bi1-xAxFeO3 systems (A: isovalent cation) ............................... 16 1.2.6. Synthesis methods of BiFeO3 ................................................... 18 1.3. MECHANOCHEMISTRY ................................................................ 23 1.3.1. Process variables ....................................................................... 25 1.3.2. Contamination ........................................................................... 29 1.3.3. Theories and models ................................................................. 30 1.4. SINTERING ....................................................................................... 33 1.4.1. General aspects .......................................................................... 33 1.4.2. Sintering methods ..................................................................... 37 1.4.3. Flash sintering ........................................................................... 41 1.5. REFERENCES .................................................................................... 47 2. OBJECTIVES AND OUTLINE .............................................................. 73 xiv 2.1. OBJECTIVES ....................................................................................... 73 2.2. OUTLINE ............................................................................................ 75 3. MATERIALS AND METHODS ............................................................. 77 3.1. MATERIALS....................................................................................... 77 3.2. SAMPLE PREPARATION ................................................................ 77 3.2.1. Mechanochemical process ........................................................ 77 3.2.2. Sintering ...................................................................................... 82 3.2.3. Flash synthesis ........................................................................... 85 3.2.4. Density measurements .............................................................. 86 3.3. SAMPLE CHARACTERIZATION .................................................. 86 3.3.1. Powder X-ray diffraction and Rietveld refinement ............... 87 3.3.2. Raman spectroscopy .................................................................. 89 3.3.3. Differential scanning calorimetry (DSC) ................................. 90 3.3.4. Scanning electron microscopy (SEM) ...................................... 91 3.3.5. Transmission electron microscopy (TEM) .............................. 92 3.3.6. UV-Visible spectroscopy ........................................................... 94 3.3.7. Dielectric thermal analysis (DEA) ........................................... 95 3.3.8. Impedance spectroscopy ........................................................... 96 3.3.9. Magnetic properties ................................................................. 104 3.4. REFERENCES .................................................................................. 105 4. MECHANOCHEMISTRY OF THE Bi1-XYbXFeO3 SYSTEM ............ 109 4.1. INTRODUCTION ............................................................................ 109 4.2. OBJECTIVES ..................................................................................... 110 4.3. EXPERIMENTAL ............................................................................. 110 4.4. RESULTS AND DISCCUSION ...................................................... 110 4.4.1. Mechanosynthesis and sintering ............................................ 110 4.4.2. Solubility limit determination ................................................ 114 4.4.3. Structural characterization ..................................................... 117 xv 4.4.4. Temperature-dependent behaviour ...................................... 123 4.4.5. Microstructural and chemical characterization ................... 124 4.4.6. Optical properties .................................................................... 126 4.4.7. Electrical properties ................................................................ 128 4.4.8. Magnetic properties ................................................................ 134 4.5. CONCLUSIONS .............................................................................. 136 4.6. REFERENCES .................................................................................. 137 5. MECHANOCHEMISTRY OF THE Bi1-XSmXFeO3 SYSTEM ........... 143 5.1. INTRODUCTION ........................................................................... 143 5.2. OBJECTIVES .................................................................................... 144 5.3. EXPERIMENTAL ............................................................................ 145 5.4. RESULTS AND DISCCUSION ...................................................... 146 5.4.1. Mechanosynthesis and sintering ........................................... 146 5.4.2. Structural characterization ..................................................... 150 5.4.3. Temperature dependant behaviour ...................................... 160 5.4.4. Microstructural and chemical characterization ................... 166 5.4.5. Optical properties .................................................................... 170 5.4.6. Electrical properties ................................................................ 172 5.4.7. Magnetic properties ................................................................ 179 5.5. CONCLUSIONS .............................................................................. 181 5.6. REFERENCES .................................................................................. 183 6. FLASH SINTERING OF BiFeO3 .......................................................... 191 6.1. INTRODUCTION ........................................................................... 191 6.2. OBJECTIVES .................................................................................... 192 6.3. EXPERIMENTAL ............................................................................ 192 6.4. RESULTS AND DISCCUSION ...................................................... 193 6.4.1. Flash sintering of mechanosynthesised BiFeO3 ................... 193 6.4.2. Structural characterization ..................................................... 198 xvi 6.4.3. Microstructural characterization ............................................ 199 6.4.4. Electrical properties ................................................................. 201 6.5. CONCLUSIONS .............................................................................. 203 6.6. REFERENCES .................................................................................. 204 7. FLASH SYNTHESIS OF BiFeO3 .......................................................... 209 7.1. INTRODUCTION ............................................................................ 209 7.2. OBJECTIVES ..................................................................................... 210 7.3. EXPERIMENTAL ............................................................................. 210 7.4. RESULTS AND DISCCUSION ...................................................... 211 7.4.1. Optimization of experimental parameters in the electric fieldassisted solid-state reaction of BiFeO3: electric field and current ...... 211 7.4.2. Reaction mechanism of the electric field-assisted solid-state reaction of BiFeO3 at 50 V cm-1 and 35 mA mm-2 ................................. 217 7.4.3. Structural characterization ..................................................... 222 7.4.4. Microstructural characterization ............................................ 224 7.4.5. Electrical properties ................................................................. 226 7.5. CONCLUSIONS .............................................................................. 228 7.6. REFERENCES .................................................................................. 229 8. CRYSTALLIZATION KINETICS OF NANOCRYSTALLINE BiFeO3 .................................................................................................................... 233 8.1. INTRODUCTION ............................................................................ 233 8.2. OBJECTIVES ..................................................................................... 235 8.3. EXPERIMENTAL ............................................................................. 235 8.4. THEORETICAL ............................................................................... 236 8.4.1. Isothermal methods ................................................................. 237 8.4.2. Non-Isothermal methods ........................................................ 238 8.5. RESULTS AND DISCUSSION ....................................................... 240 xvii 8.5.1. Crystallization and crystal growth of nanocrystalline BiFeO3 . ................................................................................................... 240 8.5.2. Crystallization kinetics of nanocrystalline BiFeO3 .............. 248 8.6. CONCLUSIONS .............................................................................. 257 8.7. REFERENCES .................................................................................. 259 9. GENERAL CONCLUSIONS ................................................................ 269 1 1. INTRODUCTION 1.1. MULTIFERROIC MATERIALS Ferroelectricity, ferromagnetism and ferroelasticity are classically known as “ferroic” properties [1]. Ferroelectricity can be described as the property of certain materials to present spontaneous electric polarization, which direction can be switched by an applied field. These materials undergo a phase transition from a high-temperature phase, where the material behaves as an ordinary dielectric, to a low-temperature phase, where the spontaneous polarization appears [2]. Ferromagnetism is analogously described as the property of certain materials to present spontaneous magnetic moment, including any kind of magnetic order (ferromagnetic, antiferromagnetic or ferrimagnetic). Similarly, these materials also present a phase transition from a high-temperate phase that does not have a macroscopic magnetic moment (paramagnetic) to a lowtemperature phase that has a spontaneous magnetization even in the absence of an applied magnetic field [3]. By extension, ferroelasticity can be understood as the phenomena in which a material presents spontaneous strain. In an equivalent way to ferroelectricity and ferromagnetism, ferroelastic materials undergo a phase transition from a high-temperature phase to a low-temperature phase, where spontaneous strain appears [4]. The term multiferroism is used to describe materials that present at least two of these defined “ferroic” properties in the same phase [5]. As Figure 1. 1 illustrates, in a “ferroic” material the electric polarization (P), magnetization (M), or strain (Ԑ) are spontaneously created to produce Chapter 1 INTRODUCTION 2 individually ferroelectricity, ferromagnetism or ferroelasticity, respectively, whereas a multiferroic material combines at least two “ferroic” forms of ordering that leads to additional interactions. Figure 1. 1. Schematic phase control in ferroic and multiferroic materials. These additional interactions give rise to materials with interesting properties for different potential technological applications. For example, it is well known that ferroelectric materials display ferroelasticity as well, such us the PbZr1-xTixO3 system, especially for certain values of composition (x). This means that a change in electric polarization implies a change in shape and vice versa, which can prove useful for several applications, such as high-energy electrical pulse generators. Analogously, there are also well known materials that exhibit simultaneously both ferromagnetic and ferroelastic properties, like some amorphous ferromagnetic alloys, which have been proposed for sensor applications INTRODUCTION Chapter 1 3 [6]. Materials that combine electrical and magnetic properties in the same phase, known as magnetoelectric multiferroics, have been the target of especially intense research, due to interesting functionalities arising from the interactions between the magnetic and electric polarization. Thus, the application of a magnetic field could be used to control the electric polarization (P), or an electric field to control the magnetization (M), as it is illustrated by the black arrows in the scheme of Figure 1. 1 [3, 7-10]. Although the term multiferroic is quite recent [5], research about magnetoelectric multiferroics dates back to the middle of the twentieth century. The first discovered magnetoelectric material was Ni3B7O13I [11]. Later on, intensive research was developed especially in Russia in the second half of the twentieth century looking for magnetoelectric mixed oxides with perovskite structures. However, despite the obvious potential of these kinds of materials and the important effort made, very few materials were actually synthesised. Moreover, those prepared not only exhibited poor coupling between the electrical and magnetic properties but also displayed their multiferroic behaviour only at cryogenic temperatures [12-15]. Those inconveniences made the materials apparently unsuitable for technological applications, which joined to the little understanding of the subjacent nature of the microscopic phenomena, eventually led the scientific community to lose the interest in them. However, since the year 2000 the research about multiferroic materials has seen a sound revival. The number of publications keeps growing by the year as clearly illustrated by Figure 1. 2, which highlights the remarkable growth of publications on the topic within the last 16 years. Chapter 1 INTRODUCTION 10 decomposition under applied voltages and its tendency to fatigue in certain direction [27]. b) Spintronic Due to the multiferroic behaviour of BiFeO3, its introduction in magnetoelectric random access memories (MeRAMs) has been explored [27, 50]. These devices can be read magnetically and written electrically, presenting several advantages over traditional ferroelectric memories. For example, they can be implemented in a solid-state circuit without mobile parts, requiring lower energy [27]. Additionally, the readout process is non-destructive, unlike direct ferroelectric reading, which requires switching the polarization. Nevertheless, more research is required before its implementation, being one of the mayor issues to be overcome that antiferromagnetic domains cannot be easily read [27]. Figure 1. 6 shows a scheme of a feasible MeRAM device based on BiFeO3. Figure 1. 6. MeRAM based on exchange-bias coupling, proposed by Bibes and Barthelemy [50]. INTRODUCTION Chapter 1 11 This arrangement has been proposed by Bibes and Barthelemy and, in principle, it would overcome the problem of reading antiferromagnetic domains [50]. c) Photocatalytic applications BiFeO3 has a small optical band gap (2.1-2.8 eV), which means that can absorb considerable amount of light in the UV-Visible region, in contrast to TiO2-based photocatalyst, the most widely used material, which is effective just in the UV region due to its relatively large band gap energy (3.2 eV) [51]. Therefore, BiFeO3 is an interesting material to be used in photocatalytic applications. It has been widely reported that BiFeO3 nanoparticles can degrade different organic pollutants in water, such as methyl orange, rhodamine B and methylene blue [52-54]. Photocatalytic hydrogen generation (water splitting) with BiFeO3 thin films has been explored as well [55, 56]. d) Photovoltaic applications Solar energy harvesting seems to be other potential application of BiFeO3. Photovoltaic cells based on ferroelectric domains of BiFeO3 have also been a subject of enquiry [57-61]. The maximum photovoltage generated at a single ferroelectric junction was determined to be of the order of 10 mV. Photovoltages up to 15 V have been reported with tens to hundreds of ferroelectric domains connected in series [61]. Nevertheless, there are several intrinsic problems related to its efficiency that make difficult its implementation. Furthermore, most photons from the sun have energies below the band gap of BiFeO3 (2.1-2.8 eV) which makes the process even less efficient [62]. Chapter 1 INTRODUCTION 12 e) Gas sensors Gas sensing properties of materials are generally studied by the comparison of their electrical resistance response in the absence and presence of the target gas. The potential use of BiFeO3 and related materials as gas sensors, for the detection of different organic and inorganic gases, has been investigated. Yu et al. reported in 2009 that BiFeO3 nanoparticles can effectively detect different organic gases, such as acetone and alcohol [63]. Recent studies show good performance of BiFeO3 detecting other gases. For instance, detection of NH3 has been reported using BiFeO3 synthesized by conventional solid state reaction, although it is noteworthy that there is considerable amount of secondary phases in the resulting material [64]. Sensors for the detection of SO2 and O2 based on BiFeO3 and related materials have been recently investigated as well [65, 66]. 1.2.4. Challenges in bulk BiFeO3 As it has been mentioned before, in spite of the promising features of BiFeO3, there are some challenges that are slowing down its development for its implementation in practical applications. As thin films are out of the scope of this thesis, we have just focused on bulk BiFeO3. The main obstacle is probably related to the difficulties in the preparation of phasepure BiFeO3 and related materials. Most employed synthesis methods normally produce impurities, which joined to an intrinsic poor magnetolectric coupling, results in a significant deterioration of the physical properties. This section is devoted to briefly explain the phase diagram and the most common impurities in BiFeO3 as well as how to INTRODUCTION Chapter 1 13 overcome the weak magnetoelectric coupling, caused by the incommensurate spin cycloid. Phase diagram and impurities in BiFeO3 Many works declare the difficulties in preparing phase-pure BiFeO3, without the presence of secondary phases, such as Bi25FeO39 and Bi2Fe4O9 [47, 67-69]. The key for understanding this issue can be found in the most widely accepted compositional phase diagram, which is depicted in Figure 1. 7 [40]. It is noteworthy to point out that other alternative compositional phase diagrams have been reported, containing certain discrepancies [47, 70-73]. These discrepancies may be related to the intrinsic difficulties in preparing phase-pure materials as well as the non-equilibrium conditions at which the experiments were performed [40, 47]. As it can be inferred from the phase diagram (Figure 1. 7), phase-pure BiFeO3 only exists in a narrow compositional window, hence, slight deviations from the stoichiometry towards the Bi2O3 or Fe2O3 rich areas would result in a mixture of BiFeO3 with Bi25FeO39 or Bi2Fe4O9. Additionally, it has been reported that BiFeO3 is actually a metastable phase and that the Gibbs energy differences between BiFeO3 and the decomposition products are so small that in certain conditions BiFeO3 can easily decompose according to the following reaction (1. 2) [68, 69]: 49BiFeO3  Bi25FeO39 + 12Bi2Fe4O9 (1. 2) Whereas there are discrepancies in the range of temperatures at which BiFeO3 decomposes [67, 69, 71], it is well known that Bi2O3 melts at temperatures above 830 ºC. Hence, synthesis methods which rely on temperatures higher than Bi2O3 melting point can lead to materials with uncertain stoichiometry due to bismuth deficiencies with the ensuing Chapter 1 INTRODUCTION 14 presence of secondary phases and the deterioration of the physical properties of BiFeO3, which is normally translated in low remnant polarization (Pr) and high leakage currents. Specifically, high leakage currents have been attributed to Bi3+ deficiencies during the synthesis, which lead to the reduction of Fe3+ to Fe2+ and the formation of anion and cation vacancies [74-76]. Bi3+ ion vacancies can also lead to oxygen vacancy formation which can result in an extra increase of conductivity [77]. Figure 1. 7. Compositional phase diagram of BiFeO3 modified by Palai et al. [40]. Magnetoelectric coupling and spin cycloid In addition to the extremely difficult synthesis, BiFeO3 exhibits weak magnetoelectric coupling. The magnetoelectric coupling in BiFeO3 manifests the relationship between ferroelectric polarization and magnetization such as ferroelectric control of magnetism or magnetic control of ferroelectricity. As it was explained in Section 1.2.2, the spin cycloid averages out the net macroscopic magnetization in crystals bigger INTRODUCTION Chapter 1 15 than its wavelength (62-64 nm). Nevertheless, there are several strategies to improve the physical properties as well as to overcome the issue of the spin cycloid: • Applying high magnetic fields: it has been reported that above certain values of magnetic field (~20 Teslas), also known as “critical field” the spin cycloid is destructed and the magnetoelectric polarization change sings and becomes linear dependent of the magnetic field [7, 78]. • Particle size reduction: enhanced magnetization has been reported in nanoparticles with sizes smaller than 92 nm, synthesized by different wet chemical methods [79-84]. • Chemical substitution: this approach is probably the most widely used to improve the magnetic and electrical behaviour and to favour the synthesis of high quality samples. As this is one of the main scopes of this thesis, it will be discussed with more detail in Section 1.2.5. Different substitution strategies have been employed at either Aor B-site cations of the BiFeO3 perovskite, as well as co-substitutions or the preparation of solid solutions with other oxides [85-90]. The partial substitution of Bi3+ for isovalent cations, such as rare earths (RE) has attracted special attention [91]. It has been widely reported that the different chemical substitution strategies result in an enhancement of the magnetization, depending on the substituent and composition, which can be correlated to an increase of the magnetocrystalline anisotropy, making the cycloidal spin structure energetically unfavourable [30]. All Chapter 1 INTRODUCTION 16 seems to be related to the ionic radius of the substituent and the subsequent distortion of the perovskite structure [91, 92]. 1.2.5. Bi1-xAxFeO3 systems (A: isovalent cation) As it has been introduced, the substitution of Bi3+ for isovalent cations is common not only for allowing the synthesis of phase-pure samples but also for the improvement of its ferroelectric properties, the reduction of leakage currents and the enhancement of the magnetoelectric coupling. Additionally, this also offers the possibility of synthesizing materials which compositions are closed to structural morphotropic phase boundaries, which leads to an additional enhancement of the physical properties of the materials [93]. It has been reported that the structural and physical properties strongly depend on the ionic size of the substituent as well as its concentration [92]. Although it is difficult to generalize and further research is required, Arnold proposed a schematic phase diagram for the lanthanide series that relates the ionic radii of the substituents to the phase transformations that the Bi1-xAxFeO3 systems undergo, as shown in Figure 1. 8. It is concluded that for the largest rare earth ions, an antipolar Pnam or Pbam phase is stabilized between the R3c space group of the parent compound, BiFeO3, and the Pnma space group of the rare earth orthoferrites, REFeO3. Conversely, for the smallest rare earths ions, phase segregation and co-existence of R3c and Pnma phases are observed rather than the transformation to an intermediate phase. INTRODUCTION Chapter 1 17 Figure 1. 8. Schematic representation of Bi1-xRExFeO3 phase diagram [91]. Additionally, Karimi et al. studied the correlation of the ferroelectricparaelectric transition temperature (TC) with the average A-site ionic radii for La3+, Nd3+, Sm3+, and Gd3+-substituted BiFeO3, concluding that as the average A-site ionic radii decreases the TC declines with a reasonably linear relationship, as can be observed in Figure 1. 9 [94]. Figure 1. 9. Variation of Tc for some La3+, Nd3+, Sm3+, and Gd3+-substituted BiFeO3 ceramics with average A-site ionic radii, 〈rA〉 [94]. Chapter 1 INTRODUCTION 18 Despite these common features within different Bi1-xRExFeO3 systems, it is noteworthy to highlight that much controversy can be found in the literature regarding the effects of rare-earth substitution, not only on the physical properties but also on the crystal structure. Thus, many contradictory results have been reported for the same systems. As one of the main scopes of this thesis is to study the effect of Yb3+ and Sm3+ substitution in BiFeO3, a brief survey of the existing literature for these systems can be found in the introduction of Chapters 4 and 5, respectively. 1.2.6. Synthesis methods of BiFeO3 Due to the notorious difficult preparation of phase-pure BiFeO3 ceramics and related materials such as Bi1-xRExFeO3 systems, numerous synthesis methods have been tested. Some of them are listed as follows: a) Conventional solid-state method BiFeO3 was first synthesized by conventional solid-state reaction [24]. This method consists in intimately mixing the starting materials in suitable stoichiometric ratios and the subsequent calcination at high temperature to allow the interdiffusion of the reactants. There are several problems regarding the synthesis of BiFeO3 by conventional solid-state reaction. One of them is related to the different diffusion rates of bismuth and iron [95]. For example, at 700 ºC the tracer diffusion coefficient of Fe3+ in Fe2O3 (D700°C~2.8·10–25 m2 s-1) is five orders of magnitude lower than that of Bi3+ in Bi2O3 (D700°C~6.8·10–20 m2 s-1) [96]. Therefore, below 700 ºC the reaction would be incomplete. Nevertheless, the diffusion rate can be enhanced by increasing the calcination temperature. However, Bi25FeO39, which is a reaction intermediate prior to the formation of BiFeO3 from the starting oxides, melts at approximately 793 ºC. It may cause loss of Bi2O3 INTRODUCTION Chapter 1 19 through segregation of the resulting liquid phase. Thus, the calcination temperature should be limited in principle. Additionally, according to Selbach et al. [69], BiFeO3 is metastable in the temperature region from 447 ºC up to 767 ºC, where Bi25FeO39 and Bi2Fe4O9 have a slightly higher stability than BiFeO3, and it would decompose according to the reaction (1. 2). Therefore, the synthesis of phase-pure BiFeO3 and related materials through solid state reaction seems to be not compatible in terms of temperature range. Other alternative methods have been addressed to overcome this issue. b) Rapid liquid sintering Rapid liquid sintering was first reported by Wang et al. for the synthesis of BiFeO3 [76]. This method is basically a modification of the conventional solid-state reaction method explained above. The differences rely on the modification of the calcination step which is just held for a few seconds (450 s) at a temperature of 880 ºC, with extremely fast heating and cooling rates (~100 ºC s-1). The authors claimed that as the reaction takes place well above the melting point of Bi2O3, the liquid phase together with the fast heating and cooling rates, accelerate the formation of pure BiFeO3, preventing the formation of secondary phases. Nevertheless, the purity of the samples synthesized by this method has been questioned as the reported works just prove the phase purity by means of laboratory XRD, which is not particularly accurate in the detection of small amounts of secondary phases [75, 76, 97, 98]. c) Wet chemical methods The wet chemical routes have also been widely used in an attempt to synthesize phase-pure BiFeO3 as well as related materials. On the one Chapter 1 INTRODUCTION 26 • Milling container: the material and the shape of the milling container are of vital importance to avoid contamination problems and compositional changes in the nominal stoichiometry of the milled products. Due to the milling process some of the container material will be unavoidably incorporated in the resulting products. Some of the most common materials used for milling containers are hardened steel, tool steel, chromium steel, tempered steel, stainless steel, WC-Co, WClined steel, bearing steel, agate, alumina, zirconia and silicon nitride. • Grinding medium (balls): Some of the most common materials used as grinding medium are hardened steel, tool steel, hardened chromium steel, tempered steel, stainless steel, WCCo, bearing steel, agate, alumina, zirconia and silicon nitride. The density must be high enough so that the balls can create enough impact force on the starting materials. The size of the balls is other parameter to take into account and it should be adapted to the size of the milling container. Nevertheless, in general terms bigger diameters produce higher collision energies. • Milling speed: it determines the kinetic energy (), designated according to equation (1. 3):  = 1 2    (1. 3) Where  is the mass of the grinding medium and  the velocity at which it is traveling. Hence, the milling speed is one of the most important parameters to be considered. In general terms, INTRODUCTION Chapter 1 27 the higher the milling speed, the higher energy input into the materials. However, depending on the design of the mill, there is a critical speed value at which the balls will be pinned to the inner wall of the milling container, due to centrifugal forces, and avoiding the balls to take down. Therefore, it is very important to optimize the milling speed in order to maximize the height at which the balls should take down, and which maximizes the collision energy with the material. • Milling time: this parameter should be optimized depending on the type of mill, speed, ball-to-power ratio, temperature, nature of the materials to be milled etc. Nevertheless, it is noteworthy that milling times longer than those strictly necessary to obtain the desired products can lead to the formation of undesired phases and can contaminate the samples from the grinding medium and milling container. • Ball-to-powder weight ratio (BPR): it is the ratio of weight of balls to the weight of powders. It has been widely varied. Nevertheless, the higher the BPR, the shorter the time required to carry out a milling process. This is due to the fact that at high BPR the number of collisions per unit of time increases and consequently more energy is transferred to the powder. The heat generated, which can significantly influence the milling process, also increases at higher BPR. • Extent to filling the vial: around 50% of the grinding container is left empty. It is important to leave enough free space for the balls and powder to move inside the container so that the impact forces can be maximized. Therefore, it is necessary to Chapter 1 INTRODUCTION 28 reach a commitment between the production rate and the left space to optimize the energy impact. • Milling atmosphere: a proper control of the milling atmosphere is important to avoid contamination problems. In most cases, mechanochemical processes are carried out under air. However, it is possible to work under inert atmospheres (Ar or He of high purity) to prevent oxidation or contamination of the resulting materials. Reactive atmospheres are also common, such as nitrogen or ammonium for the synthesis of nitrides and hydrogen for hydrides. • Process control agents (PCAs): PCAs are generally organic substances that act as surface-active agents, lowering the surface tension. PCAs can be useful to minimize cold welding and therefore inhibiting agglomeration. This is translated in lower milling times to achieve a specific particle size. The most common PCAs are ethanol, methanol, hexane and stearic acid. The nature and amount of PCAs used during milling depends on its own nature and chemical stability as well as the nature of the starting materials, purity, grinding medium and the desired final product specification such as particle size and yield. • Temperature of milling: unfortunately, most commercial mills do not allow a direct control of this parameter, although milling temperature can significantly influence the diffusivity, reaction kinetics and microstructure of the resulting products [138]. INTRODUCTION Chapter 1 29 1.3.2. Contamination One of the major drawbacks of milling, and perhaps the most important, is contamination. The large surface area of the milled products as well as the fresh surface created and the continuous abrasion and shear with the grinding media and milling container contribute to the contamination of the products. Therefore, special precaution must be taken into account. The main sources of contamination have been related to the purity of the starting materials, milling atmosphere, milling container, milling media and the nature of the used PCAs. Some considerations regarding each of them are discussed as follows. The starting materials should be of high chemical purity. The particle size is also important because lower particle sizes means higher surface areas and thus, the probability of contaminants adsorption, such as nitrogen, oxygen and water vapor, is also higher. Milling atmosphere plays a major role in the contamination of mechanosynthesized products. Therefore, it is common the use of inert atmospheres such as argon or helium of high purity in hermetic milling containers to avoid air leakage from the exterior. It is necessary to distinguish these processes and mechanosynthesis processes working under reactive atmospheres. Previous coating with the materials to be milled helps to minimize the contamination from the milling container and milling medium. Moreover, the use of harder materials as milling medium than those to be milled also contributes to minimize the contamination. PCAs can decompose during the mechanosynthesis process and as they are normally organic compounds, they are an additional source of Chapter 1 INTRODUCTION 30 impurities, such as carbon, oxygen, nitrogen and hydrogen. Hence, the use of PCAs should be avoided, unless it is strictly necessary. Other additional precautions can be taken into account in order to minimize contamination. For instance, the number of interruptions, normally to pick up sample to monitor the progress of the reaction, should be minimized as well as milling time that should be just enough to reach the stationary state of the chemical reaction. 1.3.3. Theories and models Mechanochemistry is a stochastic process that depends on a considerable amount of experimental parameters and its complexity is more than evident. There are a number of common aspects that have been observed in mechanochemical reactions. For instance, a fine and uniform dispersion of the resulting materials, normally on the nanometric scale, are commonly observed in mechanochemical process [137]. However, further from these aspects, how the mechanical energy is transformed to chemical energy is a question that remains unclear. Several theories and models have been developed to try to understand what is really happening during a mechanochemical process, being rather probably that more than one mechanism act together. A brief summary of some theories and models are reported below. Thermal theory This theory assumes that the kinetic energy is transformed into thermal energy, being this last the responsible of the induction of the chemical reactions [135, 137]. During the amorphization of the reactants increments of temperature can result in the formation of small liquid areas INTRODUCTION Chapter 1 31 which rapidly solidify by contacting the surrounding cold solid. This theory relies more on the formation of “hot-spots”. During the friction process between two sliding solids, temperature peaks of over 1000 K in surfaces of about 1 µm2 can be reached and last for 10-4–10-3 s [135]. These high temperatures can be given near the tip of a propagating crack as well, and this may result in the induction of chemical reactions. Urakaev at al. studied the products emitted from a crack propagation through a NaNO3 crystal, concluding that these products are typical of temperatures above 1000 K [139]. Hence, according to this theory, thermal spikes are the responsible of the mechanochemical reactions. Reactions induced by shear Plastic deformation is essential for the acceleration of a reaction in a mechanochemical process. The importance of shear was pointed out by Larsen et al. when demonstrated that K3Fe(CN)6 transformed into K4Fe(CN)6 not only by the effect of pressure in a Bridgeman anvil but also by applying shear [140]. One of the common characteristic of a mechanochemical process is that the reactants are in intimate contact with each other on the nanometer scale. It means that a large volume of atoms is in the grain boundaries which, in principle, should enhance the diffusion and increase the rate of mechanochemical reaction. Nevertheless, it has been proved that the mass transfer at the first stage of a mechanochemical process is due to rotation of crystal blocks and the formation of defects rather than conventional diffusion process [141]. Figure 1. 11 shows typical defects created during a milling process. The mass transport is accelerated for the increase of defects density during the evolution of the mechanochemical process. Chapter 1 INTRODUCTION 32 Figure 1. 11. Typical defects created during milling [135]. Magma-plasma model This model proposes that a huge amount of energy is set free at the contact spot of colliding particles, creating a special plasmatic state [135, 142]. This causes the emission of excited fragments of solids, electrons and photons for a short period of time, as Figure 1. 12 represents. The fresh surface formed is rather disorder, electrically charged and local temperatures of more than 10000 K can be reached. According to this model all these factors may be the responsibles of the induction of chemical reactions. Figure 1. 12. Magma-plasma model: exo-electrons (E), undeformed solid (N), highly deformed surface layer (D), plasma (P) [142]. INTRODUCTION Chapter 1 33 1.4. SINTERING Most of the methods described in Section 1.2.6 for the synthesis of BiFeO3 ceramics and related materials yield powders which must be sintered into dense and highly pure pellets. One of the aims of this thesis is to explore the newly developed flash sintering technique to sinter BiFeO3 powders prepared by mechanosynthesis. Hence, a brief introduction about sintering and the existing methods of sintering, focusing on flash sintering, is provided in this section. 1.4.1. General aspects Sintering is a processing technique used to consolidate particles, normally of metal or/and ceramic powders, with controlled porosity by applying thermal energy at around 2/3 of its melting point. This temperature is normally enough to allow significant atomic mobility and it is typically taken to establish sintering conditions [143, 144]. Sintering is an irreversible thermodynamic process associated with a lowering of the free energy of the system, where basically the main scope of the process is to decrease the surface area. The sources which provide this lowering of free energy are known as driving forces. The driving forces involved in a sintering process are the curvature of the particle surfaces, an externally applied pressure and chemical reaction. The total interfacial/surface energy can be expressed as γA, where γ is the surface energy and A is the total surface area. Hence, the reduction of the total energy is giving according to equation (1. 4): ∆ ( γ A ) = ∆ γ A + γ ∆  (1. 4) Chapter 1 INTRODUCTION 34 Where the term ∆γA is related to the change in interfacial energy due to densification and γ∆ corresponds to the variation of surface area by grain coarsening [144]. Figure 1. 13 schematically depicts densification and coarsening, the basic phenomena giving during the sintering. Figure 1. 13. Schematic representation of basic phenomena (densification and coarsening) during sintering [144]. Matter transport and mechanism of sintering In the absence of an externally applied pressure and chemical reaction, the main driving force is the curvature of the particles. It is generally well accepted that the surface energy is inversely proportional to the radius of the particles. However, the excess of surface energy is a necessary but not a sufficient condition for sintering. Besides the driving forces that initiate the process, diffusional transport of matter, promoted by thermal energy, also takes place. During sintering, atoms move from their original position, areas with higher chemical potential (source), to the neck region between particles, area with lower chemical potential (sink). Matter transport along different paths determines the mechanism of INTRODUCTION Chapter 1 35 sintering. Six different mechanisms have been proposed [143, 145], as it is shown in Figure 1. 14 and Table 1. 1. In the initial stages of sintering, the necks are considered to be the sinks for atoms. Hence, according to these mechanisms matter can be transported from the surface, along the surface (1), through the lattice (2) or by vapour transport (3) to the neck regions. Matter can be also transported from grain boundaries along grain boundary (4) or through the lattice (5) to the necks. Alternatively, matter can be transported from dislocations to the necks though plastic flow (6). The contribution of each sintering mechanism to the overall process depends on the material composition, density of grain boundaries and dislocations and temperature. Moreover, densification only happens when the centres of the particles are moved closer together. This is achieved through mechanisms (4), (5) and (6), as the matter is removed from grain boundaries and bulk and it is deposited in the necks. Conversely, mechanisms (1), (2) and (3) yield to grain coarsening through neck growth without pore removal. Figure 1. 14. Sintering mechanisms represented by numbers (Table 1. 1) in a three particles array [143]. Chapter 1 INTRODUCTION 42 In Figure 1. 17, two regions can be clearly distinguished: at fields lower than 40 V cm-1, densification occurs gradually, while at higher fields densification is obtained almost instantaneously. These two regions are known as field assisted sintering (FAST) and flash sintering (FS), respectively. In the flash sintering region, the onset of sintering temperature decreases with the applied electric field. The interest of flash sintering within the scientific community is more than evident. In the last years, it has been shown that a small dc current produces sintering at low furnace temperatures in a number of materials: yttria [159], 8YSZ [160], Al2O3 [161], Co2MnO4 [162, 163], SrTiO3 [164], La0.6Sr0.4Co0.2Fe0.8O3 [165], TiO2 [166], Ca5(PO4)3(OH) [167], BaTiO3 [168], CaCu3Ti4O12 [169], CeO2 [170], ZnO [171], SiC [172, 173], KNbO3 [174], Gd-BaCeO2 [175], SnO2 [176], SOFC [177]. Figure 1. 18 shows the onset of flash sintering in several ceramics in constant heating rate experiments. Interestingly, the flash event takes place in a small power density range, from 10 to 100 mW mm-3 [178]. Figure 1. 18. The onset of flash in several ceramics in constant heating rate experiments [178]. INTRODUCTION Chapter 1 43 Flash sintering stages Flash sintering experiments can be carried out in two different ways: running the furnace at a constant heating rate under an applied electric field, where the flash is activated at a threshold temperature, or working the furnace under isothermal conditions and them apply the electric field. In this latter case, the flash is accompanied by an incubation time. For the sake of clarity, Figure 1. 19 shows the time dependence of the electric field, current, power dissipation and relative density under isothermal conditions for 3YSZ. Jha et al. summarized the events taking place in three stages of a flash experiment based on previous research with 3YSZ [179]: • Stage I: in this region the power supply is under voltage controlled mode. The applied electric field is maintained constant, while the furnace temperature is either increasing or stable. The sample is mainly heated by joule heating. • Stage II: after an incubation time (stage I), a non-linear rise of the current, due to a sharp increase in the conductivity of the sample, is produced. If the current is not limited, the sample would experiment very high joule heating and it would eventually melt. Thus, the power supply is switched from voltage to current controlled mode, in which the applied voltage, determined by the conductivity of the sample, is adjusted to maintain the current set point. As power dissipation is the product of voltage and current, a spike is observed in the power density when the power supply switches from voltage controlled to current controlled mode. Sintering occurs within 1–5 seconds. Beside the sharp increase in the conductivity of the Chapter 1 INTRODUCTION 44 sample, electroluminescence and a dramatic increase of the densification rate are also observed in this region. Grain growth may be observed as well. • Stage III: during this stage, power supply is still under current controlled mode. The flash state can be maintained within the sample. Grain growth occurs rapidly while electroluminescence is observed as well. The furnace can be turned off and the sample cooled during this stage. Figure 1. 19. Time dependence of electrical parameters in a flash experiment carried out at isothermal furnace temperature [179]. INTRODUCTION Chapter 1 45 Mechanisms and experimental parameters Although flash sintering mechanism is still under debate, in general terms, during flash sintering experiments there is a sharp increase in conductivity and electroluminiscence that, normally, but not necessarily, is accompanied by sintering. Joule heating, which consists in the increment of the specimen temperature from electrical energy dissipation, can be related to the dramatic increment in the densification rate [180]. Thus, much effort has been devoted to estimate the temperature of the specimen during the process, although it is particularly rough in stage II given its short duration. Good results have been obtained through the theory of black-body radiation, assuming that the effects of convention and conduction are negligible versus radiation. The temperature of the specimen can be estimated according to equation (1. 5) [180]: ∆    = 4A "  # (1. 5) Where ∆ is the increase in specimen temperature, is the electrical power dissipated,  is the total surface area of the sample, " is the Stefan Boltzmann constant (5.67·10-8 Wm2 K-4) and  is the furnace temperature. Reasonable good agreement between measured temperatures by a pyrometer focused on the samples and the calculated temperatures through this theory was found, particularly in stage I, where the measured temperature was the same than that of the furnace, and in the steady part of stage III. However, no spike on the temperature measured by the pyrometer was detected [180]. The predicted temperatures obtained from the theory of the black body radiation have been further corroborated by synchrotron experiments for 3YSZ samples [181]. In those experiments, the temperature of the specimens could be indirectly calculated due to the Chapter 1 INTRODUCTION 46 relationship between the thermal expansion and the lattice parameters, which were in-situ measured from the shift in the diffraction peaks while the flash sintering experiments were carried out. This method allows the detection of any difference in the anisotropic lattice expansion between materials sintered under conventional and flash sintering conditions. Nevertheless, the mechanism of flash sintering cannot be explained just by joule heating and recent studies have shown that the electric field has an effect that goes beyond the joule heating of the sample [180]. Another characteristic event during flash sintering is photoluminescence. It has been observed that the photoluminescence increases with the dissipated power or the electrical conductivity of the samples [181-183]. Particularly, Terauds et al. carried out an exhaustive analysis by measuring the optical spectrum of the emission during flash sintering experiment for 3YSZ and compared the results with the expected values if the emission arises from black body radiation [181]. It was found that peaks of a particular wavelength under flash conditions did not change with temperature, unlike for black body radiation, where peak shifts towards lower wavelengths, as temperature increases. This indicates that the source of optical emission cannot be just temperature [181]. Although the mechanisms of flash sintering remain unclear, it is generally well accepted the tendency of certain experimental parameters during flash sintering. It has been widely reported that flash sintering temperature decreases with the applied electric field [184, 185]. Additionally, in isothermal experiments, higher electric fields decrease the incubation time [186]. For a certain applied electric field, higher current densities lead to higher densifications [186]. Likewise conventional sintering, the initial particle size of the green body also influences the sintering kinetics. Hence, bigger particle size decreases the driving force INTRODUCTION Chapter 1 47 (curvature of the particle surface), leading to lower densification rates [187]. It has been reported that an externally applied pressure during flash (sinter-forging experiment), also contributes to higher densification rates at the expense of grains growth and it also declines the flash threshold temperature [184]. Although flash sintering is still in its initial stage of development and more research should be carried out in order to clarify certain aspects, particularly related to the mechanisms, the high interest in flash sintering is not surprising due to the advantages that it presents in comparison with other sintering methods. Besides the obvious energy saving and higher sintering rate, in contrast to conventional sintering, it enables a greater control over ceramic processing which may produce materials with tailored microstructures and novel properties. Additionally, the shape of the specimens is not limited to those of graphite dies, such as in hot pressing or SPS. Either vacuum or inert atmospheres are not essential requirements and relatively cheap experimental setups can be used in a laboratory scale. Thus, it is not surprising that industrial applications of flash sintering are being developed since 2012. It has been recently demonstrated that commercial floor tiles can be produced at furnace temperatures well below that used in conventional procedures and at higher production rate [188]. 1.5. REFERENCES [1] Glinchuk M.D., Ragulya A.V. and Stephanovich V.A., Ferroics, in Nanoferroics. 2013, Springer Netherlands: Dordrecht. p. 1-32. Chapter 1 INTRODUCTION 48 [2] Kao K.C., 4 - Ferroelectrics, Piezoelectrics, and Pyroelectrics, in Dielectric Phenomena in Solids. 2004, Academic Press: San Diego. p. 213-282. [3] Hill N.A., Why are there so few magnetic ferroelectrics?. Journal of Physical Chemistry B, 2000, 104 (29), p. 6694-6709. [4] Abrahams S.C., Ferroelasticity. Materials Research Bulletin, 1971, 6 (10), p. 881-890. [5] Schmid H., Multi-ferroic magnetoelectrics. Ferroelectrics, 1994, 162 (1), p. 317-338. [6] Nan C.-W., Bichurin M.I., Dong S., Viehland D. and Srinivasan G., Multiferroic magnetoelectric composites: Historical perspective, status, and future directions. Journal of Applied Physics, 2008, 103 (3), p. 031101. [7] Eerenstein W., Mathur N.D. and Scott J.F., Multiferroic and magnetoelectric materials. Nature, 2006, 442 (7104), p. 759-765. [8] Spaldin N.A. and Fiebig M., The renaissance of magnetoelectric multiferroics. Science, 2005, 309 (5733), p. 391-392. [9] Cheong S.-W. and Mostovoy M., Multiferroics: a magnetic twist for ferroelectricity. Nature Materials, 2007, 6 (1), p. 13-20. [10] Wang K.F., Liu J.M. and Ren Z.F., Multiferroicity: the coupling between magnetic and polarization orders. Advances in Physics, 2009, 58 (4), p. 321-448. [11] Ascher E., Rieder H., Schmid H. and Stossel H., Some properties of ferromagnetoelectric nickel-iodine boracite Ni3B7O13I. Journal of Applied Physics, 1966, 37 (3), p. 1404-1405. INTRODUCTION Chapter 1 49 [12] Smolenskii G.A., Isupov V.A. and Agranovskaya A.I., A new group of ferroelectrics - (with layered structure). Soviet Physics-Solid State, 1959, 1 (1), p. 149-150. [13] Brixel W., Rivera J.P., Steiner A. and Schmid H., Magnetic FieldInduced Magnetoelectric Effects, (ME)H, in the Perovskites Pb2CoWO6 and Pb2FeTaO6. Ferroelectrics, 1988, 79, p. 495-498. [14] Drobyshe.La, Alshin B.I., Tomashpo.Yy and Venevtse.Yn, Growing, structure, and magnetic properties of Pb1/2Mn1/2Nb1/2O3 Single Crystals. Soviet Physics Crystallography, Ussr, 1970, 14 (4), p. 634. [15] Astrov D.N., Alshin B.I., Zorin R.V. and Drobyshe.La, Spontaneous magnetoelectric effect. Soviet Physics Jetp-Ussr, 1969, 28 (6), p. 3. [16] Fiebig M., Revival of the magnetoelectric effect. Journal of Physics D: Applied Physics, 2005, 38 (8), p. 123-152. [17] Filipev V.S., Smolyninov N.P., Fesenko E.G. and I.N. B., Synthesis of BiFeO3 and determination of the unit cell. Soviet Physics Crystallography, Ussr, 1960, 5 (6), p. 913-914. [18] Zaslavskii A.I. and Tutov A.G., The structure of a new antiferromagnetic, BiFeO3. Doklady Akademii Nauk Sssr, 1960, 135 (4), p. 815-819. [19] Fedulov S.A., Determinaton of Curie Temperaure for BiFeO3 ferroelectric. Doklady Akademii Nauk Sssr, 1961, 139 (6), p. 13451346. [20] Kiselev S.V., Zhdanov G.S. and Ozerov R.P., Detection of magnetic arrangement on BiFeO3 ferroelectric by means of neutron diffraction study. Doklady Akademii Nauk Sssr, 1962, 145 (6), p. 1255-1258. Chapter 1 INTRODUCTION 50 [21] Tomashpolskii I.J., Venevtse.In and Zhdanov G.S., Electron diffraction study of BiFeO3 crystals. Doklady Akademii Nauk Sssr, 1963, 153 (6), p. 1313-1313. [22] Venevcev J.N., Roginskaya J.E., Ivanova V.V., Fedulov S.A. and Zdanov G.S., Investigation of some solid solutions based on ferrroelectric-antiferromagnetic BiFeO3. Acta Crystallographica, 1963, 16 (13), p. 188. [23] Smolenskii G.A. and Bokov V.A., Coexistence of magnetic+electric ordering in crystals. Journal of Applied Physics, 1964, 35 (3P2), p. 915-918. [24] Smolenskii G.A., Agranovskaia A.I., Popov S.N. and Isupov V.A., New ferroelectrics of complex composition .2. Pb2Fe3+NbO6 and Pb2YbNbO6. Soviet Physics-Technical Physics, 1958, 3 (10), p. 19811982. [25] Smolenskii G.A., Isupov V.A. and Agranovskaya A.I., New Ferroelectrics of Complex Composition of the Type A22+(BiI3+BiII5+)O-6. Soviet Physics-Solid State, 1959, 1 (1), p. 150-151. [26] Wang J., Neaton J.B., Zheng H., Nagarajan V., Ogale S.B., Liu B., Viehland D., Vaithyanathan V., Schlom D.G., Waghmare U.V., Spaldin N.A., Rabe K.M., Wuttig M. and Ramesh R., Epitaxial BiFeO3 Multiferroic Thin Film Heterostructures. Science, 2003, 299 (5613), p. 1719-1722. [27] Catalan G. and Scott J.F., Physics and Applications of Bismuth Ferrite. Advanced Materials, 2009, 21 (24), p. 2463-2485. [28] Goldschmidt V.M., Die Gesetze der Krystallochemie. Naturwissenschaften, 1926, 14 (21), p. 477-485. INTRODUCTION Chapter 1 51 [29] Wang W.R., Xu D.P. and Su W.H., Raman shift of RMnO3 (R = La, Pr, Nd, Sm) manganites. Chinese Physics Letters, 2005, 22 (3), p. 705-707. [30] Zhang X., Sui Y., Wang X., Wang Y. and Wang Z., Effect of Eu substitution on the crystal structure and multiferroic properties of BiFeO3. Journal of Alloys and Compounds, 2010, 507 (1), p. 157161. [31] Li J.B., Rao G.H., Xiao Y., Liang J.K., Luo J., Liu G.Y. and Chen J.R., Structural evolution and physical properties of Bi1−xGdxFeO3 ceramics. Acta Materialia, 2010, 58 (10), p. 3701-3708. [32] Moreau J.M., Michel C., Gerson R. and James W.J., Ferroelectric BiFeO3 X-ray and neutron diffraction study. Journal of Physics and Chemistry of Solids, 1971, 32 (6), p. 1315-1320. [33] Kubel F. and Schmid H., Structure of a ferroelectric and ferroelastic monodomain crystal of the perovskite BiFeO3. Acta Crystallographica Section B, 1990, 46 (6), p. 698-702. [34] Bucci J.D., Robertson B.K. and James W.J., The precision determination of the lattice parameters and the coefficients of thermal expansion of BiFeO3. Journal of Applied Crystallography, 1972, 5 (3), p. 187-191. [35] Shvartsman V.V., Kleemann W., Haumont R. and Kreisel J., Large bulk polarization and regular domain structure in ceramic BiFeO3. Applied Physics Letters, 2007, 90 (17), p. 172115. [36] Khomskii D., Trend: Classifying multiferroics: Mechanisms and effects. Physics, 2009, 2 (20). Chapter 1 INTRODUCTION 58 multiferroic BiFeO3 nanoparticles. Nano Letters, 2007, 7 (3), p. 766772. [83] Selbach S.M., Tybell T., Einarsrud M.-A. and Grande T., Sizedependent properties of multiferroic BiFeO3 nanoparticles. Chemistry of Materials, 2007, 19 (26), p. 6478-6484. [84] Shetty S., Palkar V.R. and Pinto R., Size effect study in magnetoelectric BiFeO3 system. Pramana-Journal of Physics, 2002, 58 (5-6), p. 10271030. [85] Masó N. and West A.R., Electrical Properties of Ca-Doped BiFeO3 Ceramics: From p-Type Semiconduction to Oxide-Ion Conduction. Chemistry of Materials, 2012, 24 (11), p. 2127-2132. [86] Gu Y.H., Wang Y., Chen F., Chan H.L.W. and Chen W.P., Nonstoichiometric BiFe0.9Ti0.05O3 multiferroic ceramics with ultrahigh electrical resistivity. Journal of Applied Physics, 2010, 108 (9), p. 094112. [87] Ahmmad B., Kanomata K., Koike K., Kubota S., Kato H., Hirose F., Billah A., Jalil M.A. and Basith M.A., Large difference between the magnetic properties of Ba and Ti co-doped BiFeO3 bulk materials and their corresponding nanoparticles prepared by ultrasonication. Journal of Physics D-Applied Physics, 2016, 49 (26). [88] Biswal M.R., Nanda J., Mishra N.C. and Anwar S., Electrical and Magnetoelectric Properties of Ni-Co Doped BiFeO3 Ceramics. Advanced Science Letters, 2016, 22 (2), p. 363-366. [89] Woodward D.I., Reaney I.M., Eitel R.E. and Randall C.A., Crystal and domain structure of the BiFeO3-PbTiO3 solid solution. Journal of Applied Physics, 2003, 94 (5), p. 3313-3318. INTRODUCTION Chapter 1 59 [90] Khomchenko V.A., Kiselev D.A., Vieira J.M., Jian L., Kholkin A.L., Lopes A.M.L., Pogorelov Y.G., Araujo J.P. and Maglione M., Effect of diamagnetic Ca, Sr, Pb, and Ba substitution on the crystal structure and multiferroic properties of the BiFeO3 perovskite. Journal of Applied Physics, 2008, 103 (2). [91] Arnold D.C., Composition-Driven Structural Phase Transitions in RareEarth-Doped BiFeO3 Ceramics: A Review. Ieee Transactions on Ultrasonics Ferroelectrics and Frequency Control, 2015, 62 (1), p. 62-82. [92] Bielecki J., Svedlindh P., Tibebu D.T., Cai S., Eriksson S.-G., Börjesson L. and Knee C.S., Structural and magnetic properties of isovalently substituted multiferroic BiFeO3: Insights from Raman spectroscopy. Physical Review B, 2012, 86 (18), p. 184422. [93] Khomchenko V.A., Troyanchuk I.O., Bushinsky M.V., Mantytskaya O.S., Sikolenko V. and Paixão J.A., Structural phase evolution in Bi7/8Ln1/8FeO3 (Ln = La–Dy) series. Materials Letters, 2011, 65 (12), p. 1970-1972. [94] Karimi S., Reaney I.M., Han Y., Pokorny J. and Sterianou I., Crystal chemistry and domain structure of rare-earth doped BiFeO3 ceramics. Journal of Materials Science, 2009, 44 (19), p. 5102-5112. [95] Bernardo M.S., Jardiel T., Peiteado M., Caballero A.C. and Villegas M., Reaction pathways in the solid state synthesis of multiferroic BiFeO3. Journal of the European Ceramic Society, 2011, 31 (16), p. 30473053. Chapter 1 INTRODUCTION 60 [96] Mukherjee J.L. and Wang F.F.Y., Kinetics of solid-state reaction of Bi2O3 and Fe2O3. Journal of the American Ceramic Society, 1971, 54 (1), p. 31-34. [97] Pradhan A.K., Zhang K., Hunter D., Dadson J.B., Loiutts G.B., Bhattacharya P., Katiyar R., Zhang J., Sellmyer D.J., Roy U.N., Cui Y. and Burger A., Magnetic and electrical properties of single-phase multiferroic BiFeO3. Journal of Applied Physics, 2005, 97 (9), p. 093903. [98] Singh H. and Yadav K.L., Structural, dielectric, vibrational and magnetic properties of Sm doped BiFeO3 multiferroic ceramics prepared by a rapid liquid phase sintering method. Ceramics International, 2015, 41 (8), p. 9285-9295. [99] Wei J. and Xue D., Low-temperature synthesis of BiFeO3 nanoparticles by ethylenediaminetetraacetic acid complexing sol–gel process. Materials Research Bulletin, 2008, 43 (12), p. 3368-3373. [100] Hasan M., Islam M.F., Mahbub R., Hossain M.S. and Hakim M.A., A soft chemical route to the synthesis of BiFeO3 nanoparticles with enhanced magnetization. Materials Research Bulletin, 2016, 73, p. 179-186. [101] Selbach S.M., Einarsrud M.-A., Tybell T. and Grande T., Synthesis of BiFeO3 by wet chemical methods. Journal of the American Ceramic Society, 2007, 90 (11), p. 3430-3434. [102] Masoudpanah S.M., Mirkazemi S.M., Shabani S. and Dolat Abadi P.T., The effect of the ethylene glycol to metal nitrate molar ratio on the phase evolution, morphology and magnetic properties of single phase BiFeO3 nanoparticles. Ceramics International, 2015, 41 (8), p. 96429646. INTRODUCTION Chapter 1 61 [103] Kumar M., Yadav K.L. and Varma G.D., Large magnetization and weak polarization in sol–gel derived BiFeO3 ceramics. Materials Letters, 2008, 62 (8–9), p. 1159-1161. [104] Mukherjee A., Basu S., Chakraborty G. and Pal M., Effect of Ydoping on the electrical transport properties of nanocrystalline BiFeO3. Journal of Applied Physics, 2012, 112 (1). [105] Schiemer J., Withers R.L., Carpenter M.A., Liu Y., Wang J.L., Norén L., Li Q. and Hutchison W., Temperature-dependent electrical, elastic and magnetic properties of sol–gel synthesized Bi0.9Ln0.1FeO3 (Ln = Nd, Sm). Journal of Physics: Condensed Matter, 2012, 24 (12), p. 125901. [106] Vijayanand S., Potdar H.S. and Joy P.A., Origin of high room temperature ferromagnetic moment of nanocrystalline multiferroic BiFeO3. Applied Physics Letters, 2009, 94 (18), p. 182507. [107] Paraschiv C., Jurca B., Ianculescu A. and Carp O., Synthesis of nanosized bismuth ferrite (BiFeO3) by a combustion method starting from Fe(NO3)3·9H2O-Bi(NO3)3·9H2O-glycine or urea systems. Journal of Thermal Analysis and Calorimetry, 2008, 94 (2), p. 411-416. [108] Iorgu A.I., Maxim F., Matei C., Ferreira L.P., Ferreira P., Cruz M.M. and Berger D., Fast synthesis of rare-earth (Pr3+, Sm3+, Eu3+ and Gd3+) doped bismuth ferrite powders with enhanced magnetic properties. Journal of Alloys and Compounds, 2015, 629, p. 62-68. [109] Ke H., Wang W., Wang Y., Xu J., Jia D., Lu Z. and Zhou Y., Factors controlling pure-phase multiferroic BiFeO3 powders synthesized by chemical co-precipitation. Journal of Alloys and Compounds, 2011, 509 (5), p. 2192-2197. Chapter 1 INTRODUCTION 62 [110] Kothai V. and Ranjan R., Synthesis of BiFeO3 by carbonate precipitation. Bulletin of Materials Science, 2012, 35 (2), p. 157-161. [111] Hu Z.-T., Lua S.K., Yan X. and Lim T.-T., Nanostructured hexahedron of bismuth ferrite clusters: delicate synthesis processes and an efficient multiplex catalyst for organic pollutant degradation. RSC Advances, 2015, 5 (106), p. 86891-86900. [112] Liu Z., Qi Y. and Lu C., High efficient ultraviolet photocatalytic activity of BiFeO3 nanoparticles synthesized by a chemical coprecipitation process. Journal of Materials Science: Materials in Electronics, 2010, 21 (4), p. 380-384. [113] Shami M.Y., Awan M.S. and Anis-ur-Rehman M., Phase pure synthesis of BiFeO3 nanopowders using diverse precursor via coprecipitation method. Journal of Alloys and Compounds, 2011, 509 (41), p. 10139-10144. [114] Medina L.M.S., Jorge G.A. and Martín Negri R., Structural, dielectric and magnetic properties of Bi1−xYxFeO3 (0 < x < 0.2) obtained by acid– base co-precipitation. Journal of Alloys and Compounds, 2014, 592, p. 306-312. [115] Yoshimura M. and Byrappa K., Hydrothermal processing of materials: past, present and future. Journal of Materials Science, 2008, 43 (7), p. 2085-2103. [116] Han J.T., Huang Y.H., Wu X.J., Wu C.L., Wei W., Peng B., Huang W. and Goodenough J.B., Tunable Synthesis of Bismuth Ferrites with Various Morphologies. Advanced Materials, 2006, 18 (16), p. 21452148. INTRODUCTION Chapter 1 63 [117] Reddy Vanga P., Mangalaraja R.V., Giridharan N.V. and Ashok M., PTCR behavior of BiFeO3 synthesized by the solvothermal method. Materials Letters, 2015, 143, p. 230-232. [118] Chen C., Cheng J., Yu S., Che L. and Meng Z., Hydrothermal synthesis of perovskite bismuth ferrite crystallites. Journal of Crystal Growth, 2006, 291 (1), p. 135-139. [119] Han S.H., Kim K.S., Kim H.G., Lee H.-G., Kang H.-W., Kim J.S. and Cheon C.I., Synthesis and characterization of multiferroic BiFeO3 powders fabricated by hydrothermal method. Ceramics International, 2010, 36 (4), p. 1365-1372. [120] Wu C.F., Wei J. and Kong F.S., Effect of rare earth dopants on the morphologies and photocatalytic activities of BiFeO3 microcrystallites. Journal of Materials Science-Materials in Electronics, 2013, 24 (5), p. 1530-1535. [121] Gajović A., Šturm S., Jančar B., Šantić A., Žagar K. and Čeh M., The Synthesis of Pure-Phase Bismuth Ferrite in the Bi–Fe–O System Under Hydrothermal Conditions without a Mineralizer. Journal of the American Ceramic Society, 2010, 93 (10), p. 3173-3179. [122] Chen X.-Z., Qiu Z.-C., Zhou J.-P., Zhu G., Bian X.-B. and Liu P., Large-scale growth and shape evolution of bismuth ferrite particles with a hydrothermal method. Materials Chemistry and Physics, 2011, 126 (3), p. 560-567. [123] Prado-Gonjal J., Villafuerte-Castrejón M.E., Fuentes L. and Morán E., Microwave–hydrothermal synthesis of the multiferroic BiFeO3. Materials Research Bulletin, 2009, 44 (8), p. 1734-1737. Chapter 1 INTRODUCTION 64 [124] Li S., Nechache R., Davalos I.A.V., Goupil G., Nikolova L., Nicklaus M., Laverdiere J., Ruediger A. and Rosei F., Ultrafast Microwave Hydrothermal Synthesis of BiFeO3 Nanoplates. Journal of the American Ceramic Society, 2013, 96 (10), p. 3155-3162. [125] Dutta D.P., Jayakumar O.D., Tyagi A.K., Girija K.G., Pillai C.G.S. and Sharma G., Effect of doping on the morphology and multiferroic properties of BiFeO3 nanorods. Nanoscale, 2010, 2 (7), p. 1149-1154. [126] Szafraniak I., Połomska M., Hilczer B., Pietraszko A. and Kępiński L., Characterization of BiFeO3 nanopowder obtained by mechanochemical synthesis. Journal of the European Ceramic Society, 2007, 27 (13– 15), p. 4399-4402. [127] Da Silva K.L., Menzel D., Feldhoff A., Kübel C., Bruns M., Paesano A., Düvel A., Wilkening M., Ghafari M., Hahn H., Litterst F.J., Heitjans P., Becker K.D. and Šepelák V., Mechanosynthesized BiFeO3 Nanoparticles with Highly Reactive Surface and Enhanced Magnetization. The Journal of Physical Chemistry C, 2011, 115 (15), p. 7209-7217. [128] Perejon A., Murafa N., Sanchez-Jimenez P.E., Criado J.M., Subrt J., Dianez M.J. and Perez-Maqueda L.A., Direct mechanosynthesis of pure BiFeO3 perovskite nanoparticles: reaction mechanism. Journal of Materials Chemistry C, 2013, 1 (22), p. 3551-3562. [129] Perejon A., Sanchez-Jimenez P.E., Perez-Maqueda L.A., Criado J.M., Romero de Paz J., Saez-Puche R., Maso N. and West A.R., Single phase, electrically insulating, multiferroic La-substituted BiFeO3 prepared by mechanosynthesis. Journal of Materials Chemistry C, 2014, 2 (39), p. 8398-8411. INTRODUCTION Chapter 1 65 [130] Perejon A., Gil-Gonzalez E., Sanchez-Jimenez P.E., Criado J.M. and Perez-Maqueda L.A., Structural, Optical, and Electrical Characterization of Yttrium-Substituted BiFeO3 Ceramics Prepared by Mechanical Activation. Inorganic Chemistry, 2015, 54 (20), p. 98769884. [131] Pattanayak S., Parida B.N., Das P.R. and Choudhary R.N.P., Impedance spectroscopy of Gd-doped BiFeO3 multiferroics. Applied Physics A, 2013, 112 (2), p. 387-395. [132] Layek S. and Verma H., Magnetic and dielectric properties of multiferroic BiFeO3 nanoparticles synthesized by a novel citrate combustion method. Advanced Materials Letters, 2012, 3 (6), p. 6. [133] Heinicke G., Tribochemistry. 1984, Berlin: Akademie Verlag. [134] James M. Grinding Chemicals Together in an Effort to be Greener. http://www.nytimes.com/2016/07/19/science/green-chemistrymechanochemistry.html?_r=0 2016. [135] Balaz P., Achimovicova M., Balaz M., Billik P., CherkezovaZheleva Z., Criado J.M., Delogu F., Dutkova E., Gaffet E., Gotor F.J., Kumar R., Mitov I., Rojac T., Senna M., Streletskii A. and Wieczorek-Ciurowa K., Hallmarks of mechanochemistry: from nanoparticles to technology. Chemical Society Reviews, 2013, 42 (18), p. 7571-7637. [136] Boldyrev V.V., Mechanochemistry and mechanical activation of solids. Solid State Ionics, 1993, 63, p. 537-543. [137] Suryanarayana C., Mechanical Alloying and Milling, ed. Dekker Marcel. 2004, New York. Chapter 1 INTRODUCTION 66 [138] Humphry-Baker S.A., Garroni S., Delogu F. and Schuh C.A., Meltdriven mechanochemical phase transformations in moderately exothermic powder mixtures. Nature Materials, 2016, 15 (12), p. 1280-1286. [139] Urakaev F.K., Boldyrev V.V., Pozdnyakov O.F. and Regel V.R., Study of mechanism of mechanochemical decomposition of solid inorganic-compounds. Kinetics and Catalysis, 1977, 18 (2), p. 292299. [140] Larsen H.A. and Drickamer H.G., Chemical Effects of Plastic Deformation at High Pressure. The Journal of Physical Chemistry, 1957, 61 (9), p. 1249-1252. [141] Bokhonov B., Konstanchuk I., Ivanov E. and Boldyrev V., Stage formation of quasi-crystals during mechanical treatment of the cubic Frank-Kaspes phase Mg32(Zn, Al)49. Journal of Alloys and Compounds, 1992, 187 (1), p. 207-214. [142] Thiessen P.A., Meyer K. and Heinicke G., Grundlagen der Tribochemie. 1967, Berlin: Akademie Verlag. [143] Castro R. and van Bentherm K., eds. Sintering: Mechanisms of Convention Nanodensification and Field Assisted Processes. Vol. 35. 2013, Springer: Heidelberg, New York, Dordrecht, London. [144] Kang S.-J.L., Sintering process, in Sintering: densification, grain growth and microstructure. 2005, Butterworth-Heinemann: Oxford. p. 3-8. [145] Swinkels F.B. and Ashby M.F., A second report on sintering diagrams. Acta Metallurgica, 1981, 29 (2), p. 259-281. [146] Coble R.L., Sintering Crystalline Solids. I. Intermediate and Final State Diffusion Models. Journal of Applied Physics, 1961, 32 (5), p. 787792. INTRODUCTION Chapter 1 67 [147] Coble R.L., Sintering Crystalline Solids. II. Experimental Test of Diffusion Models in Powder Compacts. Journal of Applied Physics, 1961, 32 (5), p. 793-799. [148] Kang S.-J.L., 4 Initial stage sintering, in Sintering: Densification, Grain Growth, and Microstructure. 2005, Butterworth-Heinemann: Oxford. p. 39-55. [149] Kingery W.D., Uhlmann D.R. and Bowen H.K., eds. Introduction to ceramic. 2nd ed. 1976, John Wiley & Sons: New York [150] Venkatachari K.R. and Raj R., Enhancement of Strength through Sinter Forging. Journal of the American Ceramic Society, 1987, 70 (7), p. 514-520. [151] Rahaman M.N. and De Jonghe L.C., Sintering of CdO Under Low Applied Stress. Journal of the American Ceramic Society, 1984, 67 (10), p. 205-207. [152] Lakshmanan A., ed. Sintering of Ceramics – New Emerging Techniques. 2012, InTech: Rijeka. [153] Munir Z.A., Anselmi-Tamburini U. and Ohyanagi M., The effect of electric field and pressure on the synthesis and consolidation of materials: A review of the spark plasma sintering method. Journal of Materials Science, 2006, 41 (3), p. 763-777. [154] Garay J.E., Current-Activated, Pressure-Assisted Densification of Materials. Annual Review of Materials Research, 2010, 40, p. 445468. [155] Guillon O., Gonzalez-Julian J., Dargatz B., Kessel T., Schierning G., Raethel J. and Herrmann M., Field-Assisted Sintering Technology/Spark Plasma Sintering: Mechanisms, Materials, and Chapter 2 OBJECTIVES AND OUTLINE 74 dependent behaviour, microstructure and the study of its optical, electrical and magnetic properties. • Synthesis by direct mechanochemistry of the samarium partially-substituted BiFeO3 system (Bi1-xSmxFeO3) in a wide range of composition and its characterization by an exhaustive analysis of their crystal structure, microstructure, temperature dependant behaviour and the evaluation of their optical, electrical and magnetic properties. • Employment of the novel flash sintering approach for the densification of mechanosynthesized BiFeO3 powders as well as the evaluation of the effects of the experimental parameters on the purity, microstructure and electrical properties of the densified specimens. • Inspired by the flash sintering procedure, exploring an alternative synthetic method based on the electric field-assisted solid-state reaction of the pristine oxides, Fe2O3 and Bi2O3, to prepare BiFeO3. Additionally, the optimization of the synthesis conditions and the study of the reaction mechanism. • Finally, the study of the thermal crystallization event in nanocrystalline BiFeO3 powders by the combination of the information obtained from different characterization techniques. The crystallization is one key aspect in the resulting properties of functional materials, but to the best of our knowledge, studies of the thermal crystallization kinetics in BiFeO3 are scarce and OBJECTIVES AND OUTLINE Chapter 2 75 none of them have properly addressed its study by a combination of several characterization techniques. 2.2. OUTLINE This thesis is structured in 9 chapters, where the most relevant results are presented in chapters, 4, 5, 6, 7 and 8. These chapters include a brief introduction and specific conclusions to the topic addressed. The outline of this thesis is described as follows: • Chapter 1 includes a literature review, which aim is to provide a general overview about BiFeO3 and related materials, mechanochemistry and flash sintering. • Chapter 2 addresses the motivation, objectives and outline of this thesis. • Chapter 3 describes the materials, experimental setups and characterization techniques employed in this research. • Chapter 4 is devoted to the preparation by direct mechanosynthesis of the Bi1-xYbxFeO3 system. This chapter also includes the characterization carried out to the mechanosynthesised system in terms of the crystal structure, microstructure, temperature dependant behaviour and the evaluation of its optical, electrical and magnetic properties. • Chapter 5 addresses the preparation by direct mechanosynthesis of the Bi1-xSmxFeO3 system. The characterization of the crystal Chapter 2 OBJECTIVES AND OUTLINE 76 structure, microstructure, temperature dependent behaviour and physical properties are also presented and discussed. • Chapter 6 reports the densification of mechanosynthesised BiFeO3 powders by the flash sintering technique. The effects of the applied electric field and current density on the purity, microstructure and electrical properties of the sintered materials are evaluated. • Chapter 7 presents the preparation of BiFeO3 by electric fieldassisted solid-state reaction, also denominated flash synthesis, for the first time in literature. The reaction mechanism is also discussed and the quality of the prepared samples is tested by Rietveld refinement and impedance spectroscopy. • Chapter 8 studies the thermal crystallization event in nanocrystalline BiFeO3 powders by the combination of the information obtained from: X-ray thermodiffraction, transmission electron microscopy and differential scanning calorimetry. • Chapter 9 summarizes the conclusions derived from this thesis. 77 3. MATERIALS AND METHODS 3.1. MATERIALS The high purity commercially available oxides, listed in Table 3. 1, were used for the preparation of the multiferroic materials, without any further purification treatment. Table 3. 1. Commercially available oxides employed for the preparation of the multiferroic materials. Oxide Catalogue Number CAS Number Purity (%) Company Bi2O3 223891 1304-76-3 ≥99% Sigma-Aldrich Fe2O3 310050 1309-37-1 ≥99% Sigma-Aldrich Yb2O3 246999 1314-37-0 ≥99% Sigma-Aldrich Sm2O3 AB255271 12060-58-1 99.9% ABCR GmbH 3.2. SAMPLE PREPARATION 3.2.1. Mechanochemical process The different materials studied in this thesis were prepared by mechanosynthesis in a high-energy planetary ball mill. A planetary ball mill consists of at least one grinding jar which is arranged eccentrically on a rotating support disk. Figure 3. 1 illustrates the scheme of a planetary ball mill. A special drive mechanism causes the jars to rotate around their own axes (ωv) while the supporting disk is rotating at the same time (ωp), similarly to a planet-like movement. The jars and the supporting disk rotate in opposite directions and, therefore, the balls in the grinding jars Chapter 3 MATERIALS AND METHODS 78 are subjected to superimposed rotational movements (Coriolis forces), that alternately act in like and opposite directions. This causes the grinding balls to run down the inside wall of the jar producing friction effect, followed by impact effect due to the collisions of the balls between them and with the opposite inside wall of the jars [1]. This allows the transmission of a high amount of energy to the reactants causing different physical and chemical phenomenon, such as particle size reduction, defects and the induction of chemical reactions, within others. Figure 3. 1. Scheme of a planetary ball mill (modified from [2]). A mechanochemical process is quite simple. It basically consists in mixing the stoichiometric amounts of the staring materials with the grinding media (balls) inside the jars, under a proper atmosphere and grinding for a certain period of time to induce the chemical reactions. However, as it was pointed out in Section 1.3, it is a stochastic process that depends on a considerable amount of experimental parameters. MATERIALS AND METHODS Chapter 3 79 Two different models of the planetary Micro Mills PULVERISETTE 7 (Fritsch, Idar-Oberstein, Germany) were used for the preparation of the samples in this thesis: the Premium Line (Figures 3.2 a) and b)) and the Classic Line (Figures 3.2 c) and d)). Figure 3. 2. Planetary Micro Mills PULVERISETTE 7: a), b) Premium Line model and c), d) Classic Line model. Both mills have been modified to work under controlled atmosphere of gases. For this purpose, a rotary valve that allows the connection of the gas cylinder with the stainless steel jar during the whole milling experiment by means of a 5 mm diameter polyamide tube, has been incorporated to the mills (Figure 3. 2) [3]. These rotary valves are able to Chapter 3 MATERIALS AND METHODS 80 work up to a maximum pressure of 70000 bar and a rotation speed of 25000 rpm. In order to maintain the pressure constant inside the jar, even if the gas is consumed during the reaction, the jars are equipped with a male taper straight adaptor and are sealed with a Viton O-ring. In all experiments, the jars were purged several times with the selected gas before starting the milling and then the desired pressure was selected and maintained during the whole milling treatment [3-5]. The experimental conditions for the mechanosynthesis of the BiFeO3 related samples were optimized in previous research carried out by the group [6]. The most relevant parameters used in both mills are detailed as follows. a) Planetary Micro Mill PULVERISETTE 7 Premium Line: This mill was used for the preparation of samples in Chapters 4, 5 and 6. • Type of jar and volume: due to the high resistance to abrasion, jars of tempered steel were used for the mechanosynthesis of all the samples. The jars have a diameter of 23.2 mm and a volume of 80 cm3. • Milling speed: the spinning rate of the supporting disk and the superimposed rotation in the opposite direction of the jars was set at 700 rpm. • Grinding medium: 9 hardened stainless steel balls with a diameter of 15 mm (total mass of ~120 g). MATERIALS AND METHODS Chapter 3 81 • Extent of filling the jar: 1/5 approximately of the total volume of the jar. • Ball-to-powder weight ratio (BPR): 1:20, which means that ~6 g of sample per jar can be obtained in each mechanosynthesis process. • Atmosphere: 7 bar of oxygen in order to avoid bismuth reduction with the iron from the grinding media [3]. • Time: the milling time depends on the type of sample and its composition. All the mechanosynthesis reactions were monitored by powder X-ray diffraction, taking small amounts of the samples at different reaction times. b) Planetary Micro Mill PULVERISETTE 7 Classic Line: This mill was used for the preparation of samples in Chapter 8 (nanocrystalline BiFeO3) [3, 6]. • Type of jar and volume: jars of tempered steel with a diameter of 19.8 mm and a volume of 45 cm3 were used. • Milling speed: the spinning rate of the supporting disk and the superimposed rotation in the opposite direction of the jars was also set at 700 rpm. • Grinding medium: 6 hardened stainless steel balls with a diameter of 15 mm (total mass of ~81 g). Chapter 3 MATERIALS AND METHODS 82 • Extent of filling the jar: 1/5 approximately of the total volume of the jar. • Ball-to-powder weight ratio (BPR): 1:20. Approximately 4 g of sample per jar can be obtained in each mechanosynthesis process. • Atmosphere: 7 bar of oxygen in order to avoid bismuth reduction with the iron from the grinding media. • Time: 25 h of milling were required for the preparation of pure nanocrystalline BiFeO3 powders. 3.2.2. Sintering The densification of the powders obtained by mechanosynthesis was achieved by conformation using uniaxial pressing and sintering. As it has been introduced in Section 1.4, sintering is a complex process where solidstate mass transport occurs, reducing the particle interface/surface energy in order to form a dense body [7]. In this thesis, conventional sintering and flash sintering has been used in order to obtain dense specimens, particularly for the study of their electrical properties. In conventional sintering, for the preparation of the green pellets, the collected mechanosynthesized powders were pressed into cylindrical pellets, by applying an uniaxial pressure of 0.93 GPa for 5 minutes in a stainless steel die 6.35 mm of diameter. The mass of the pellets was 400 mg approximately with a thickness of ~2 mm. The density of the green pellets, which was estimated geometrically, was above 75% of the theoretical density in every sample. Subsequently, the pellets were sintered in air in a Carbolite MTF 12/25 tube furnace (Carbolite, Bamford, England) at MATERIALS AND METHODS Chapter 3 83 temperatures ranging from 825 ºC to 900 ºC, and using different dwell steps, depending on the type of sample and its composition. In every case, the cooling and heating rates were both set at 10 ºC min-1. As described in Section 1.4.3, flash sintering is a relatively new electric field-assisted sintering procedure developed in 2010 by Cologna et al. [8]. It basically consists in applying an electric field directly to the specimen, allowing the current to flow through the sample [8, 9]. The mechanosynthesized powders were pressed into dog-bone shaped specimens, using the die depicted in Figures 3.3 a) and b) and by applying a uniaxial pressure of 250 MPa [10]. Details about the dog-bone sample dimensions are shown in Figure 3. 3 c), which thickness was of 1 mm approximately, when 800 mg of powders were used. Platinum paste was applied in the holes of the green-body samples, in order to improve the electrical contact with the electrodes. Figure 3. 3. a), b) Dog-bone die and c) Dog-bone sample dimensions. Chapter 3 MATERIALS AND METHODS 90 In this thesis, Raman spectra of the samples were recorded at room temperature by a dispersive Horiva Jobin Yvon LabRam HR800 microscope equipped with a 20 mW green laser (532.14 nm) and a 100X objective with a confocal pinhole of 10 µm. 3.3.3. Differential scanning calorimetry (DSC) Differential scanning calorimetry (DSC) is a thermoanalytic technique that measures the difference of heat flow between a sample and an inert reference as a function of temperature. The heat flow should be the same until some thermal event such as melting, phase transitions or decomposition takes place. It is translated to extra heat input to the sample or reference depending on whether the process is endothermic or exothermic, respectively [16]. This themoanalytic characterization technique is particularly useful for the study of the multiferroic behaviour of BiFeO3 and related materials. In this thesis, the DSC curves were measured in a wide range of temperatures, from -50 ºC to 850 ºC. Two different equipments were used depending on the temperature range. From -50 ºC to 400 ºC a DSC Instrument (Q200, TA Instruments, Crawley, UK) was used, whereas from 400 ºC to 850 ºC a simultaneous TG/DSC Instrument (Q600 SDT, TA Instruments, Crawley, UK) was employed. Approximately 40 mg of the samples were placed in open alumina pans. The experiments were registered at the heating rate of 10 ºC min-1 and under airflow of 100 cm3 min-1. The Néel and Curie transition temperatures where considered at the minimum of the endothermic peak. MATERIALS AND METHODS Chapter 3 91 3.3.4. Scanning electron microscopy (SEM) The scanning electron microscope (SEM) uses a focused beam of high-energy electrons to generate a variety of signals at the surface of solid specimens. These signals, derived from electron-sample interactions, reveal information about the sample, including chemical composition, external surface features, texture and topography. This is due to the depth of focus of SEM instruments that allow generating pictures of threedimensional quality. SEM instruments cover a wide range of magnification normally from 10-2 to 102 µm, approximately [11]. Samples require specific preparation before measuring in SEM, as they should be totally dried, due to the vacuum conditions required for the electrons to form the images and, in addition, it is normally necessary to coat the electrically insulating samples with a thin layer of conductive material, such as gold or graphite, to prevent the build-up of charge in the surface of the solid. Within the signals generated by the interaction between the bean of high-energy electrons and the sample, secondary electrons (SE), backscattered electrons (BSE) and X-ray are of particular relevance in scanning electron microscopy. In this thesis just SE and X-ray have been used for the characterization of the samples by SEM. SE are electrons of low energy coming from the atoms in the sample which provide topographic information of the surface of the sample in grayscale. Additionally, the X-ray generation is characteristic of each element presented in the sample, hence, qualitative and semiquantitative compositional information of the sample can be provided as well, by either scanning the wavelength dispersed (WD) or energy dispersed (ED) [17]. Chapter 3 MATERIALS AND METHODS 92 In this thesis, the microstructure, morphologic and chemical composition of the samples was studied using a Hitachi S-4800 SEM-FEG, equipped with an energy dispersive X-ray spectrometer (EDX), Quantax Bruker. The field-emission gun operated at 2 kV for obtaining the micrographs, while for EDX analysis, it operated at 20 kV. The samples in the shape of pellets required previous surface preparation including polishing and thermal etching for 30 minutes at temperatures of 90% of those used for their sintering, in order to reveal grain boundaries. Moreover, some of the samples were Au sputter-coated due to its insulating nature at room temperature, using an Emitech K550 Sputter Telstar (Barcelona, Spain). 3.3.5. Transmission electron microscopy (TEM) In transmission electron microscopy (TEM), a beam of high energy electrons (100-800 kV) is transmitted through an ultra-thin specimen, generating an image from the interaction of the transmitted electrons with the sample, which is magnified and focused onto an imaging device, such as a fluorescent screen or a charge-coupled device (CCD camera) [18]. Particularly, high-resolution electron microscopy (HR-TEM) is a powerful tool in imaging defects, as it is capable of giving information on an atomic scale, by direct lattice imaging. TEM can be operated in several modes. Bright field image is the most common mode of operation, where the image is directly formed from the direct beam by its interaction with the sample. The contrast depends on mass-thickness, crystallinity and the atomic mass of the atoms in the sample. Thicker and heavy atoms enriched areas appear dark, while no sample regions appear bright. Dark field image is another mode of MATERIALS AND METHODS Chapter 3 93 operation, where the direct beam is blocked by the aperture and only the beams diffracted from the particles of interest are allowed to recombine to form the image. Hence, only regions of the samples that are diffracting the beams to the selected area appear bright, whereas no sample regions or regions that transmit the beam appear dark. This mode is particularly useful to distinguish between different crystalline features and regions [19]. Electron diffraction patterns can be also generated by TEM. This technique is known as selected aperture electron diffraction (SAED), as electrons are dispersed elastically, satisfying diffraction condition of the sample's crystal structure and indicating a crystallographic direction. Generally, SAED patterns are formed by a number of well-defined spots that depends on the orientation of the specimen for monocrystalline samples, or a serial of rings for polycrystalline and amorphous samples. SAED is a powerful tool that allows a detailed study of the symmetry, orientation and crystallographic planes of the selected area [18, 19]. The instruments employed for the TEM characterization of the samples were: • A 200 kV TEM Philips CM 200 microscope with B6La filament, located in the Instituto de Ciencia de Materiales de Sevilla (ICMSE-CSIC). • A high resolution TEM with field mission gun (FEG-HRTEM) from FEI Company, USA, (Model TECNAI G2 F30 S-twin), with a Fischione high angle annular dark-field detector (HAADF) Instruments, USA (0.16 nm point resolution) to work in STEM mode and one INBCA ZX-max 80 silicon drift detector for EDS analysis. The experiments were performed at 300 kV with a Chapter 3 MATERIALS AND METHODS 94 resolution of 0.2 nm. This instrument is located in the Instituto de Ciencia de Materiales de Sevilla (ICMSE-CSIC). • A 300 kV JEOL JEM 3010 UHR electron microscope with a LaB6 electron source and equipped with a Semi STEM, located in the Institute of Inorganic Chemistry of the Czech Academy of Science in Řež, Czech Republic. The samples for TEM measurements were prepared by dispersing the powders in isopropanol for 1 hour, picking up a drop of the diluted supernatant and the subsequent deposition in a glass sample holder. Immediately, the carbon coated grids were smoothly rubbed in the deposited drops. Then, the grids were dried slowly at room temperature for the solvent evaporation. The analysis of high resolution micrographs and the first Fourier transform for phase interpretation were performed with Digital Micrograph software (Gatan Inc., USA) and the Java version of JEM Software. 3.3.6. UV-Visible spectroscopy The optical properties of the samples were determined by ultraviolet-visible spectroscopy (UV-Visible spectroscopy). This characterization technique relies on the principle that under certain conditions, materials absorb or emit energy. Particularly, this technique uses light in the visible, near-UV and near-infrared region, which translated to an energy scale correspond approximately from 102 to 104 kJ mol-1, and are often associated with colour. Moreover, this is typically the same range of the energy required for transitions of electrons between the outermost energy levels. Thus, various types of electronic transitions can occur, providing valuable information about the studied materials and MATERIALS AND METHODS Chapter 3 95 specifically local structure, as the positions of the absorption bands are sensitive to coordination environment and bond character [11]. In this thesis, UV-Visible spectroscopy was used to determine the energy band gap of the mechanosynthesized materials, as the promotion of electrons from the valence band to the conduction band in semiconductors lies between the visible and UV-region. UV-Vis measurements were recorded in a UV-Vis spectrophotometer (Perkin-Elmer Lambda 35, Massachusetts, USA) equipped with an integrating sphere and were taken over a wavelength range from 400 nm to 800 nm. Barium sulphate was used as coated standard pattern. The energy band gap (Eg) values were calculated by means of Tauc’s plot:  ( ℎ 5 ) =  ( ℎ 5 −  7 ) 8 /  (3. 2) Where  is the absorption coefficient, ℎ: is the photon energy,  is a constant and ; depends on the type of optical transition [20]. For BiFeO3 and related materials the band gap is direct and, therefore, ; equals 1 [21, 22]. Thus, the band gap energy, 7, can be estimated from the plot of (ℎ5)2 versus photon energy (ℎ5), through the extrapolation of a tangent line from the point of inflection to (ℎ5)2 = 0. 3.3.7. Dielectric thermal analysis (DEA) Dielectric thermal analysis (DEA) is commonly used to characterize the response of materials by analysing separately the dielectric and resistive contribution, normally in terms of dielectric constant (also known as relative permittivity), Ԑ,, and dielectric loss, =;>. An electric field at fixed frequency is applied to a sample as a function of temperature [23]. Chapter 3 MATERIALS AND METHODS 96 Besides the dielectric response, DEA allows the identification of phase transition temperatures. Particularly, in ferroelectric materials, the dielectric constant experiment a dramatic increment in the vicinity of the Curie temperature. In this thesis, DEA was used to study the nature of the phase transitions of the prepared samples. Relative permittivity, Ԑ,, and =; > measurements were carried out on sintered pellets from 50 ºC to 450 ºC, using a 4263B LCR meter (Agilent Technologies, California, USA) at a fixed frequency of 105 Hz. All the data were corrected taking into account the geometry of the pellets, whose opposite faces were previously Au sputter-coated using an Emitech K550 Sputter Telstar (Barcelona, Spain). 3.3.8. Impedance spectroscopy Impedance spectroscopy is a powerful technique for the analysis of the electrical properties of electroceramics, because it enables characterizing separately the resistive (R) and reactive (C) contribution of the different electro-active region in the sample. In other words, the macroscopic dielectric constant in polycrystalline materials commonly consists in several contributions, such as interfacial phenomena (sampleelectrode), grains and grain boundaries. The deconvolution of these different contributions is totally essential for a proper understating of the functionality of the material [24-27]. Additionally, and besides the electrical microstructure characterization to the different electro-active regions presented in the sample, additional information can be provided, such as the assessment of the electrical homogeneity as well as information related to the charge transport mechanism, by performing experiments in MATERIALS AND METHODS Chapter 3 97 different oxygen partial pressure atmospheres, variable temperature and dc bias [27]. Impedance measurements are carried out over a wide range of frequencies, commonly from 10-2 to 107 Hz. Impedance, ?, can be defined as the ratio of voltage, @, and current, A: ? = @ A (3. 3) Experimentally, it is measured using a low voltage applied across a sample, which can be expressed according to equation (3. 4) as a function of time, : @ = @  sin ( E ) (3. 4) Where @ is the amplitude and E is the angular frequency: E = 2 F f (3. 5) Being H the frequency. The resulting current, A, can be determined as: A = A  sin ( E + I ) (3. 6) Where A is the amplitude and I is the phase difference between the applied voltage, @, and the resulting current, A. This phase difference, I, is directly related to the reactive component, C, of the system due to the different dielectric relaxations that take place in the sample, whereas the total magnitude of impedance is related to the resistive component, R [25]. The data are normally expressed in the complex impedance notation: ? ∗ = ( ? ’ + L? ’’ ) (3. 7) Chapter 3 MATERIALS AND METHODS 98 Where the real part, ?’, and the imaginary part, ?’’, are measured separately, being L the imaginary number. The different electro-active regions of a sample are characterized according to their dielectric relaxation times or time constants. These regions are described by a resistance and a capacitance, usually placed in parallel, commonly known as RC elements [24, 25]. The characteristic relaxation time or time constant, M, of each RC element is given by the product of R and C: M = OP (3. 8) In the frequency domain, each RC element can be separated due to the following relationship: 2 F H &%Q OP = 1 (3. 9) Where H&%Q is the frequency of maximum loss in the impedance spectrum [25]. The separation of different dielectric relaxations depends on the resolution of their associated time constants. Impedance data are typically represented in the form of imaginary (Z’’) versus real (Z’) impedances, which is known as the complex impedance plot. The parallel RC elements normally give different semicircles in the complex impedance plot, from which the R and C values can be obtained. The values of R for each electro-active regions can be directly obtained from the intercept of each semicircle on the real impedance axis, Z’, of the complex impedance plot (Figure 3. 6 a)). Ideally, the values of C can be obtained by applying equation (3. 9) [24]. The origin of each dielectric relaxation can be identified by means of the capacitance value of its associated RC element, according to the classification proposed by Irvine et al., as shown in Table 3. 2. [24]. MATERIALS AND METHODS Chapter 3 99 Table 3. 2. Capacitance value and their possible interpretation [24]. Capacitance (F) Phenomenon Responsible 10-12 Bulk 10-11 Minor, secondary phase 10-11-10-8 Grain boundary 10-10-10-9 Bulk ferroelectric 10-9-10-7 Surface layer 10-7-10-5 Sample-electro interface 10-4 Electrochemical reacions Besides complex impedance, ?∗, and in order to help with the data interpretation, other formalisms are used, such as admittance, R∗, electric modulus, S∗, and complex permittivity, Ԑ∗ [25]. They are defined by the following equations: R ∗ = 1 ? ∗ (3. 10) S ∗ = LE P  ? ∗ (3. 11) Ԑ ∗ = 1 S ∗ = R ∗ LE P  (3. 12) Being P the vacuum capacitance of the empty measuring cell, which is defined as: P  = Ԑ   T (3. 13) Where Ԑ is the permittivity of free space (8.854·10-14 F cm-1) and  and T are the electrode area and separation, respectively. In order to carry out a detailed electrical characterization, these formalisms are useful because they highlight different characteristics of a sample. Due to their Chapter 3 MATERIALS AND METHODS 106 [8] Cologna M., Rashkova B. and Raj R., Flash Sintering of Nanograin Zirconia in <5 s at 850°C. Journal of the American Ceramic Society, 2010, 93 (11), p. 3556-3559. [9] Yu M., Grasso S., McKinnon R., Saunders T. and Reece M.J., Review of flash sintering: materials, mechanisms and modelling. Advances in Applied Ceramics, 2017, 116 (1), p. 24-60. [10] Francis J.S.C., A study on the phenomena of flash-sintering with tetragonal zirconia in Department of Mechanical Engineering. 2013, University of Boulder: Boulder. p. 189. [11] West A.R., ed. Basic Solid State Chemistry. 2nd ed. 1999, John Wiley & Sons, LTD: Chichester, New York. [12] Rietveld H., A profile refinement method for nuclear and magnetic structures. Journal of Applied Crystallography, 1969, 2 (2), p. 65-71. [13] Rodriguez-Carvajal J., In Fullprof: A Program for Rietveld Refinement and Pattern Matching Analysis in Satellite Meeting on Powder diffraction of the XV Congress of the IUCr. 1990. [14] Toby B.H., R factors in Rietveld analysis: How good is good enough?. Powder Diffraction, 2006, 21 (1), p. 67-70. [15] Ferraro J.R., Nakamoto K. and Brown C.W., Introductory Raman Spectroscopy. 2nd ed. 2003, San Diego: Academic Press. [16] Interpretating DSC curves. Part 1: Dynamic measurements. UserCom, 2000, 11. [17] Vernon-Parry K.D., Scanning electron microscopy: an introduction. IIIVs Review, 2000, 13 (4), p. 40-44. MATERIALS AND METHODS Chapter 3 107 [18] Williams D.B. and Carter C.B., Transmission Electron Microscopy: A Textbook for Materials Science. 2009: Springer. [19] Egerton R.F., Physical Principles of Electron Microscopy. An Introduction to TEM, SEM, and AEM. 2005: Springer. [20] Tauc J., Grigorov R. and Vancu A., Optical properties and electronic structure of amorphous germanium. Phys. Status Solidi, 1966, 15 (2), p. 627-637. [21] Hauser A.J., Zhang J., Mier L., Ricciardo R.A., Woodward P.M., Gustafson T.L., Brillson L.J. and Yang F.Y., Characterization of electronic structure and defect states of thin epitaxial BiFeO3 films by UV-visible absorption and cathodoluminescence spectroscopies. Applied Physics Letters, 2008, 92 (22), p. 222901. [22] Kim Y.I., Atherton S.J., Brigham E.S. and Mallouk T.E., Sensitized layered metal-oxide semiconductor particles for photochemical hydrogen evolution from nonsacrifical electron-donors. Journal of Physical Chemistry, 1993, 97 (45), p. 11802-11810. [23] Wu Y.-J., Chen X.-K., Zhang J., Liu J., Xiao W.-S., Wu Z. and Chen X.-J., Pressure effect on structural and vibrational properties of Smsubstituted BiFeO3. Journal of Applied Physics, 2013, 114 (15). [24] Irvine J.T.S., Sinclair D.C. and West A.R., Electroceramics: Characterization by Impedance Spectroscopy. Advanced Materials, 1990, 2 (3), p. 132-138. [25] Sinclair D.C., Characterization of electro-materials using ac impedance spectroscopy. Boletín de la Sociedad Española de Cerámica y Vidrio, 1995, 34 (2), p. 10. Chapter 3 MATERIALS AND METHODS 108 [26] West A.R., Sinclair D.C. and Hirose N., Characterization of Electrical Materials, Especially Ferroelectrics, by Impedance Spectroscopy. Journal of Electroceramics, 1997, 1 (1), p. 65-71. [27] Barsoukov E. and Macdonald J.R., Impedance Spectroscopy: Theory, Experiment, and Applications, ed. Sons John Wiley &. 2005. [28] Fleig J. and Maier J., The impedance of ceramics with highly resistive grain boundaries: Validity and limits of the brick layer model. Journal of the European Ceramic Society, 1999, 19 (6-7), p. 693-696. [29] Clarke J., Squids. Scientific American, 1994, 271 (2), p. 46-53. [30] Kleiner R., Koelle D., Ludwig F. and Clarke J., Superconducting quantum interference devices: State of the art and applications. Proceedings of the IEEE, 2004, 92 (10), p. 1534-1548. 109 4. MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM 4.1. INTRODUCTION As stated in Section 1.2.5, the partial substitution of Bi3+ for isovalent cations has been widely employed, not only for obtaining phase-pure BiFeO3 related materials but also for improving the ferroelectric properties, reduction of leakage currents and enhancement of the magnetoelectric coupling. The partial substitution of Bi3+ with small rare earth cations, such as Yb3+, is attractive because it may lead to an enhanced lattice distortion and consequently an increase in the ferroelectric polarization of the materials [1]. Nevertheless, the existing literature about the Bi1xYbxFeO3 system is quite limited and significant discrepancies can be found [1-4]. For instance, there is no clear consensus about the Yb content that induces phase transitions and the crystal structures that the system adopts. In addition, a full characterization in terms of structural analysis, electrical and magnetic properties has not been reported [1-4]. The mechanochemical synthesis of BiFeO3 and related materials has been shown to be an effective preparation method, from which truly phase-pure and highly insulating compounds can be obtained [5-7]. To the best of our knowledge, the preparation of the Bi1-xYbxFeO3 system has just been addressed by hydrothermal synthesis, rapid liquid phase sintering, a modified sol-gel method and conventional solid-state reaction [1-4]. Thus, Chapter 4 MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM 110 exploring the preparation of the partial substitution of BiFeO3 with ytterbium by mechanochemistry is of the most interest. 4.2. OBJECTIVES The objective of this chapter is the preparation of the Bi1-xYbxFeO3 system by mechanosynthesis, as an alternative route for the synthesis of these materials, as well as to carry out a complete characterization of the samples by means of XRD, DSC, SEM, UV-Vis spectroscopy, impedance spectroscopy and magnetic measurements. 4.3. EXPERIMENTAL Samples of the Bi1-xYbxFeO3 system in the nominal compositional range x = 0.02, 0.05 and 0.07, were prepared by mechanochemistry followed by sintering, according to the procedure detailed in Section 3.2.1 (Chapter 3). All the pellets were conventionally sintered at 825 ºC for 6 min in air. 4.4. RESULTS AND DISCCUSION 4.4.1. Mechanosynthesis and sintering Samples of the Bi1-xYbxFeO3 system in the nominal compositional range x = 0.02, 0.05 and 0.07, were prepared by mechanosynthesis followed by sintering. Figure 4. 1 shows the XRD patterns of the powders obtained for each composition after different milling times. MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM Chapter 4 111 Figure 4. 1. XRD patterns of powders obtained after milling 15 min, 3 h, 6 h and 12 h stoichiometric amounts of the single oxides in oxygen (7 bar) for different nominal compositions: a) x = 0.02, b) x = 0.05, c) x = 0.07. Chapter 4 MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM 112 The starting oxides suffer an initial amorphization during the milling treatment, as inferred from the comparison of the XRD patterns registered after 15 min and 3 h of milling of Figure 4. 1. The amorphization is followed by the crystallization of the new phases as grinding proceeds up to 6 h. In terms of milling time, the same behaviour was observed for each composition prepared. Further mechanical treatments did not produce modifications in the powders, as it is observed from XRD measurements after 12 h of milling (Figure 4. 1). In order to avoid contamination from the grinding media, the optimum milling time was taken at 6 h for every composition. The evolution of the microstructure of the milled powders as a function of time was followed by SEM. SEM micrographs for nominal x = 0.02, x = 0.05 and x = 0.07 compositions after different milling times are shown in Figure 4. 2. Similar features are observed for every nominal composition as milling proceeds. After short milling times (15 min, Figures 4. 2 a), d) and g)), the morphology of the powders is plate-like, due to delamination of bismuth oxide, mixed with ytterbium oxide and iron oxide particles. As milling continues up to 3 h (Figures 4. 2 b), e) and h)), the morphology changes to highly aggregated powders. Finally, after 6 h of mechanical treatment the obtained powders are composed of aggregated nanometric subunits, as it can be observed in Figures 4. 2 c), f) and i). MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM Chapter 4 113 Figure 4. 2. SEM micrographs of powders of the Bi1-xYbxFeO3 samples with nominal x = 0.02, x = 0.05 and x = 0.07 compositions, at different milling times, 15 min (a, d, g), 3 h (b, e, h) and 6 h (c, f, i). Due to the nanometric size of the crystallites obtained after milling, the XRD peaks are quite broad (Figure 4. 1), which made difficult to obtain detailed crystallographic information. In order to increase the crystallinity, the milled powders of each nominal composition were pressed into pellets and conventionally sintered at 825 ºC for 6 min. These sintering conditions were considered optimal after studying their influence on the purity and density of the ceramics. The resulting pellets density was 7.3 g cm-3 (89%), 6.9 g cm-3 (83%) and 6.8 g cm-3 (82%) for the samples with nominal x = 0.02, 0.05 and 0.07 compositions, respectively. Figure 4. 3 shows the XRD patterns of the sintered samples. Although the main phase can be indexed in the R3c space group, typical from the parent BiFeO3 compound, small reflections corresponding to other phases are observed in the samples with nominal x = 0.05 and x = 0.07 compositions. Moreover, the intensity of Chapter 4 MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM 114 these reflections increases with the ytterbium content of the samples, which suggests that the solubility limit of ytterbium in the BiFeO3 lattice has been reached. 20 25 30 35 40 45 x = 0.02 x = 0.07 * secondary phase * * * * * ** * ** x = 0.05 Intensity (square root) 2θ (º) Figure 4. 3. XRD patterns of the samples obtained after milling for 6 h and sintering Bi1-xYbxFeO3 powders of nominal compositions: a) x = 0.02, b) x = 0.05 and c) x = 0.07. The intensity is represented in square root in order to make the secondary phase more noticeable. 4.4.2. Solubility limit determination DSC resulted an useful characterization technique for the estimation of the solubility limit of ytterbium in BiFeO3 (and, therefore, the stoichiometry of the main phase for the nominal x = 0.05 and x = 0.07 compositions) since a linear relationship between the Curie temperature (TC) and the real composition (x) of bismuth substituted samples has been reported [8]. Figure 4. 4 depicts the DSC curves registered from 300 ºC to 850 ºC for the nominal x = 0.02, 0.05 and 0.07 compositions. MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM Chapter 4 115 300 320 340 360 380 400 500 550 600 650 700 750 800 b) a) TN x=0.07 x=0.05 x=0.02 Heat Flow (a.u.) T(ºC) x=0.07 x=0.05 x=0.02 Tc Tc Tc Heat Flow (a.u.) T(ºC) Figure 4. 4. DSC traces for the Bi1-xYbxFeO3 samples with x = 0.02 and nominal 0.05 and 0.07 compositions, obtained after milling for 6 h and sintering at 825 ºC: a) temperature range from 300 ºC to 400 ºC, b) temperature range from 500 ºC to 825 ºC. The dashed line in a) is a guide to the eye. Two endothermic peaks are observed in the DSC traces for all samples. The first one appears at approximately 370 ºC and corresponds to the Néel temperature (TN), (Figure 4. 4 a)), associated with the second order phase transition (antiferromagnetic-paramagnetic) and remaining constant for all compositions. The second peak (Figure 4b) is more intense and corresponds to the TC (ferroelectric-paraelectric transition) [9, 10]. The sample with composition x = 0.02 presents the TC at 796 ºC while for Chapter 4 MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM 122 12 bands: four of them correspond to A1 vibrational modes (at 157, 173, 222 and 411 cm-1), whereas the others can be attributed to E vibrational modes (at 263, 279, 290, 347, 370, 470, 522, 615 cm-1). Additionally, some differences are observed between the samples. As the amount of ytterbium increases, there is a small shift of the bands to higher wavelengths and two of them disappear (279 and 347 cm-1). Moreover, the bands associated with the A1 vibrational modes bellow 200 cm-1 (159 and 170 cm-1), decrease their relative intensity. It may be attributed to the partial substitution of bismuth by ytterbium in the structure, as this behaviour is similar to that observed for the introduction of other substituents in the A position of the BiFeO3 perovskite [7, 17-19]. 100 200 300 400 500 600 700 Intensity (a.u.) Raman Shift (cm-1) a) b) c) A (157) A (173) A(222) E(263) E(279) E(290) E(347)E(370) A (411)E(470) E(522) E(615) Figure 4. 10. Raman spectra of the samples obtained after milling for 6 h and sintering at 825 ºC: a) x = 0.02 composition, b) nominal x = 0.05 and c) nominal x = 0.07 compositions. MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM Chapter 4 123 4.4.4. Temperature-dependent behaviour The temperature dependant behaviour of the Bi1-xYbxFeO3 system was also studied by means X-ray thermodiffraction. Figure 4. 11 presents the temperature-dependent XRD patterns, recorded from 766 ºC to 800 ºC in the 2θ range from 30º to 34º, for the samples obtained after milling for 6 h and sintering at 825 ºC. As temperature increases the samples suffer a phase transformation from R3c to Pnma (ferroelectric-paraelectric transition), which is reflected in the double peak, at approximately 31.5º32º, that becomes to three overlapping peaks very close together. Additionally, the temperature ranges at which the phase transformations take place are consistent with the obtained transitions in the DSC data (Figure 4. 4 b)), confirming the decrease of TC as the concentration of ytterbium increases. A similar behaviour has been observed for other substituents [7, 8]. It is worth noting that the TC of the nominal x = 0.05 and x = 0.07 compositions, observed in the diffraction data, are very close to each other, in agreement with the DSC data, suggesting that these samples are two-phase mixtures of an x = 0.03 phase and an ytterbium rich phase. 30 31 32 33 34 30 31 32 33 34 30 31 32 33 34 c) b) 788 ºC 782 ºC 800 ºC 794 ºC Intensity (a.u.) 2θ (º) a) 790 ºC 784 ºC 778 ºC 772 ºC Intensity (a.u.) 2θ(º) 796 ºC 784 ºC 778 ºC 772 ºC 769 ºC 766 ºC Intensity (a.u.) 2θ (º) Figure 4. 11. Temperature-dependent XRD patterns of the Bi1-xYbxFeO3 samples obtained after milling for 6 h and sintering at 825 ºC: a) x = 0.02, b) nominal x = 0.05 composition, c) nominal x = 0.07 composition. Chapter 4 MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM 124 4.4.5. Microstructural and chemical characterization SEM micrographs of the different samples, milled for 6 h and then sintered at 825 ºC, are shown in Figures 4. 12 a), c) and e). Pellets were thermally etched for 30 min at 90% of the sintering temperature to reveal the grain boundaries. The microstructure is granular in all the samples. The grain size distribution is more homogenous in the sample of x = 0.02 composition (Figure 4. 12 a)), with typical values between approximately 1 and 3 µm. When the amount of ytterbium increases the grain size decreases to values smaller than 2 µm. Moreover, the porosity also increases with the amount of substituent, in agreement with the calculated densities (reported in Section 4.4.1). The EDX spectrums of the samples are shown in Figures 4. 12 b), d) and f) for x = 0.02 composition and nominal x = 0.05 and x = 0.07 compositions, respectively. The elemental composition has been determined by a semiquantitative analysis of the EDX spectra, as it is presented in Table 4. 2. The experimental results are mostly coincident with the expected values, being an indication that the initial stoichiometry has not been modified during the preparation of the samples. However, slight differences, which can be attributed to the intrinsic errors of the EDX, are observed. MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM Chapter 4 125 Figure 4. 12. SEM micrographs and EDX spectrum of the Bi1-xYbxFeO3 samples obtained after milling for 6 h and sintering at 825 ºC. Table 4. 2. Elemental atomic composition of the Bi1-xYbxFeO3 samples obtained after milling for 6 h and sintering at 825 ºC, determined by a semiquantitative analysis of EDX spectra. Experimental atomic composition (%) Sample (Nominal compositions) Theoretical atomic composition Fe/Yb/Bi (%) Fe Yb Bi Bi0.98Yb0.02FeO3 50/1/49 53.62 ± 3.04 1.02 ± 0.28 45.36 ± 3.52 Bi0.95Yb0.05FeO3 50/2.5/47.5 53.94 ± 3.07 2.45 ± 0.34 43.61 ± 3.73 Bi0.93Yb0.07FeO3 50/3.5/46.5 53.56 ± 3.01 2.78 ± 0.77 43.66 ± 1.41 Chapter 4 MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM 126 4.4.6. Optical properties The optical properties of the samples were studied by measuring their UV-Vis diffuse reflectance absorption spectra. Figure 4. 13 shows the absorption spectra of the samples of the different compositions obtained after milling for 6 h and sintering at 825 ºC. It can be clearly observed that the samples absorb considerable amount of light in the visible region, which suggests their potential use as visible-light photocatalysts. As BiFeO3 and related materials are considered direct band gap semiconductors, n is equal to 1 in Tauc’s plot (equation (3.2), Chapter 3). Thus, the band gap energy can be estimated from the plot of (ℎ5)2 versus photon energy (ℎ5), through the extrapolation of a tangent line from the point of inflection to (ℎ5)2 = 0 [20, 21], as it is shown in the insets of Figure 4. 13. The band gap energy values are 2.06 eV for the sample of x = 0.02 composition and 2.04 eV for both samples with nominal x = 0.05 and 0.07 compositions. This is again another indication that these two last samples contain the same Bi1-xYbxFeO3 phase. These values of band gap energy are smaller than those previously reported for pure BiFeO3 [9, 22, 23], or those reported for ytterbium substituted BiFeO3 [2]. Hence, this suggests that ytterbium substituted samples prepared by mechanosynthesis may be good photocatalytic materials. MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM Chapter 4 127 Figure 4. 13. UV-Vis diffuse reflectance spectra of the Bi1-xYbxFeO3 samples obtained after milling for 6 h and sintering at 825 ºC: a) x = 0.02 composition, b) nominal x = 0.05 and c) nominal x = 0.07 compositions. The insets represent (ℎ5)2 versus photon energy for the calculation of the corresponding band gap energy. Chapter 4 MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM 128 4.4.7. Electrical properties The electrical properties of the samples were studied by impedance spectroscopy. Similar results were obtained from all the samples. Thus, they resulted to be highly insulating at room temperature, presenting modest levels of conductivity from 300 ºC. For this reason, the impedance measurements were carried out from this temperature. The impedance complex plane plots of the samples at different temperatures are constituted by single slightly distorted arcs, as shown in Figures 4. 14 a), 4. 15 a) and 4. 16 a) for x = 0.02, nominal x = 0.05 and x = 0.07 compositions, respectively. For every sample, the associated Z’’/M’’ spectroscopic plots at 380 ºC show single peaks with small displacements between them for the maxima frequencies (Figures 4. 14 b), 4. 15 b) and 4. 16 b)). These results suggest that the samples are electrically homogenous due to the absence of any additional peaks at lower frequencies in the Z’’ spectrum, despite the presence of the small amount of ytterbium enriched secondary phase for the samples with nominal x = 0.05 and 0.07 compositions. The capacitance values are presented in Figures 4.14 c), 4. 15 c) and 4. 16 c). The capacitance remains approximately constant in the entire frequency range for the three samples. Hence, the electrical homogeneity of the samples is again evidenced. MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM Chapter 4 129 0 200 400 0 -100 -200 -300 -400 102103104105106 0 -5 -10 -15 -20 -25 102103104105106 10-11 10-10 10-9 1.55 1.60 1.65 1.70 1.75 -6.0 -5.8 -5.6 -5.4 -5.2 -5.0 -4.8 380 ºC 360 ºC 340 ºC 320 ºC 300 ºC 380 ºC 360 ºC 340 ºC 320 ºC 300 ºC Z''(kΩ cm) Z'(kΩ cm) Z''(kΩ cm) f(Hz) 380 ºC 0 1 2 3 4 5 10-10M''ε0 -1 (F-1 cm) C'(F cm-1) f (Hz) Ea=1.12 0.06 eV log (σ/Scm-1) 1000 K/T ± c ) d ) a)b) c)d) Figure 4. 14. a) Impedance complex plane plots, b) Z″/M″ spectroscopic plots at 380 ºC, c) C′ versus frequency and d) bulk Arrhenius plot for the sample with x = 0.02 composition obtained after milling for 6 h and sintering at 825 ºC. Chapter 4 MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM 130 0 100 200 0 -50 -100 -150 -200 102103104105106107 0 -2 -4 -6 -8 -10 -12 -14 102103104105106 10-12 10-11 10-10 10-9 1.55 1.60 1.65 1.70 1.75 -5.6 -5.4 -5.2 -5.0 -4.8 -4.6 -4.4 380 ºC 360 ºC 340 ºC 320 ºC 300 ºC Z''(kΩ cm) Z'(kΩ cm) 380 ºC 360 ºC 340 ºC 320 ºC 300 ºC 380 ºC Z''(kΩ cm) f (Hz) 0 1 2 3 4 5 10-10M''ε0 -1 (F-1 cm) C' (F cm-1) f (Hz) log (σ/Scm-1) 1000 K/T ± Ea=1.03 0.05 eV c ) d ) a)b) c)d) Figure 4. 15. a) Impedance complex plane plots, b) Z″/M″ spectroscopic plots at 380 ºC, c) C′ versus frequency and d) bulk Arrhenius plot for the sample with nominal x = 0.05 composition obtained after milling for 6 h and sintering at 825 ºC. MECHANOCHEMISTRY OF THE Bi1-xYbxFeO3 SYSTEM Chapter 4 131 0 40 80 120 160 0 -40 -80 -120 102103104105106107 0 -2 -4 -6 -8 -10 -12 102103104105106 10-12 10-11 10-10 10-9 1.55 1.60 1.65 1.70 1.75 -5.6 -5.4 -5.2 -5.0 -4.8 -4.6 -4.4 380 ºC 360 ºC 340 ºC 320 ºC 300 ºC 380 ºC 360 ºC 340 ºC 320 ºC 300 ºC Z''(kΩ cm) Z'(kΩ cm) Z''(kΩ cm) f(Hz) 0 1 2 3 4 5 10-10M''ε0 -1 (F-1 cm) 380 ºC C'(F cm-1) f(Hz) log (σ/Scm-1) 1000 K/T ± Ea=1.03 0.01 eV a)b) c)d) Figure 4. 16. a) Impedance complex plane plots, b) Z″/M″ spectroscopic plots at 380 ºC, c) C′ versus frequency and d) bulk Arrhenius plot for the sample with nominal x = 0.07 composition obtained after milling for 6 h and sintering at 825 ºC. The resistivity values of the samples, obtained from the intercept on the real Z’ axes of the complex plane plots at different temperatures, in conventional Arrhenius format (equation (3.16)) are presented individually in Figures 4. 14 d), 4. 15 d) and 4. 16 d) for the three samples. A linear relationship between the conductivity and the inverse of