FULLTEXT01.pdf

Synthesis, Characterisation and Applications of Iron Oxide Nanoparticles GERMAN SALAZAR-ALVAREZ Doctoral Thesis Stockholm, Sweden 2004 KTH Materialvetenskap ISRN KTH/MSE--04/59--SE+...

Synthesis, Characterisation and Applications of Iron Oxide Nanoparticles GERMAN SALAZAR-ALVAREZ Doctoral Thesis Stockholm, Sweden 2004 KTH Materialvetenskap ISRN KTH/MSE--04/59--SE+CHEM/AVH SE-100 44 Stockholm ISBN 91-7283-884-1 SWEDEN Akademisk avhandling som med tillstånd av Kungl Tekniska högskolan framlägges till offentlig granskning för avläggande av filosofie doktorsexamen fredagen den 17 december 2004 kl 10.00 i D3, Huvudbyggnaden, Kungliga Tekniska högskolan, Lindstedtsvägen 5, Stockholm. c German Salazar-Alvarez, december 2004 Tryck: Universitetsservice US AB iii Abstract This thesis deals with the synthesis, characterisation, and some applications of ferrimagnetic iron oxide nanoparticles. The iron oxide systems presented in this work are Fe3 O4 , γ-Fe2 O3 , Nix Fe3−x O4 , Cox Fe3−x O4. Iron oxide nanoparticles were prepared using chemical synthesis methods. Nanoparticles with a narrow size distribution were synthesised using zone confi- nement methods such as nanoemulsions and the novel flow injection synthesis. The flow injection method consisted of the precipitation of iron oxide nanopar- ticles in a continuous or segmented flow in a capillary reactor under laminar flow. Also, the preparation of particles in a ∼ 40 g/batch scale was carried out in a computer controlled stirred reactor which allowed reproducible synthesis conditions such a ∆pH ∼ 0.1 and ∆T ∼ 1 ◦ C. Several chemical and hydrodyna- mic parameters were optimised to achieve high phase purity and magnetisation. These methods allowed the preparation of particles with tailored mean particle size from 3 up to 12 nm. Spinel iron oxide nanoparticles were also doped with Ni and Co. The influence of the synthesis parameters on the morphological and magnetic pro- perties of nickel- and cobalt-doped magnetite (Nix Fe3−x O4 and Cox Fe3−x O4 ) was studied. Particle size, morphology, and composition of the materials are strongly dependent on reaction conditions, e.g., metal to hydroxide ratio and temperature. Cubic, spherical, and particles of Cox Fe3−x O4 with varying mor- phology could be obtained by varying the metal to hydroxide ratio. Mössbauer studies indicated a disordered structure with a relatively low degree of inver- sion in the spinel structure. The coercivity of the particles increased with the cobalt content. High resolution microscopy and electron diffraction studies showed that cubic particles are monocrystalline. Iron oxide nanoparticles (Fe3 O4 and γ-Fe2 O3 ) with a particle size of about 10 nm were coated with layers of inorganic materials such as silica and gold to render the particles higher functionality and vary the surface properties. Iron oxide nanoparticles were coated with silica in a controlled manner producing either single coated particles, with a mean particle size varying from 12 to 60 nm, or multiparticle-beads with mean particle size in the range of 50-250 nm. The use of iron oxide nanoparticles in two applications was studied, i.e., for DNA purification and as optical power limiting agents. Silica-coated iron oxide particles were investigated for use in magnetic purification of plasmid DNA of the size range 0.1–1.7 kbp. The prepared nanoparticles have shown higher DNA recovery yields when compared to commercial particles. Single silica-coated nanoparticles have DNA recovery yield of about 70%, while the beads of multiple particles showed a recovery of approximately 50%. The optical power limiting application of coated and non-coated iron oxide nanoparticles was investigated. Both aqueous suspensions of pristine iron oxide nanoparticles consisting mainly of γ-Fe2 O3 and silica-coated iron oxide nano- particles had a clamping level of about 3 µJ. The photoinduced nonlinear light scattering is the dominating nonlinear mechanism. Silica-coated nanoparticles have shown a better performance as a result of their higher colloidal stability. iv Sammanfattning I denna avhandlingen diskuteras syntes, karakterisering och tillämpningar för nanopartiklar av järnoxid. De olika järnoxidsystem som presenteras i arbetet består av Fe3 O4 , γ-Fe2 O3 , Nix Fe3−x O4 , Cox Fe3−x O4. Nanopartiklarna av järnoxid har framställts via kemiska syntesmetoder. Nanopartiklar med liten varians i storleken (σ < 20%) och en storleksdistri- bution 3-8 mm framställdes i den zon isolering från inverterade miceller och i ett nyutvecklat metod som bygger på flödesinjektion (flow injection synthesis). Å andra sida, för att få reproducerbar betingelser vid syntes av nanopartik- lar i 40 grams skala har en automatiserad reaktor använts (∆pH ∼ 0.1 och ∆T ∼ 1 ◦ C). Flera kemiska och hydrodynamiska parametrar optimiserades för att högre fasrenhet och bättre magnetiska egenskaper skulle erhållas. Des- sa reaktionsbetingelser tillåter framställning av nanopartikar i relativt stora mängder i storleksområdet 3-12 nm. Järnoxid nanopartiklar med spinell struktur framställdes via doping med nickel och kobolt. Effekterna från syntesbetingelserna på morfologin och magne- tiska egenskaper hos partiklarna har studerats hos nickel och koboltdopad mag- netit (Nix Fe3−x O4 respektive Cox Fe3−x O4 ). Partiklarnas storlek, morfologi och kemiska sammansättning var beroende av koncentrationsförhållandet mel- lan metall-/ hydroxidjon. Via ändringar i detta förhållande kunde Cox Fe3−x O4 partiklarnas utseende styras så att kubiska, sfäriska eller till och med odefinie- rade partiklar bildades. Mössbauersspektroskopi har pekat på en oregelbunden katjondistribution i strukturen. Partiklarnas koercivitet ökade med kobolthal- ten. Högupplösande transmissionselektronmikroskopi och elektrondiffraktion visade att de kubiska partiklarna var monokristaller. För att kunna påverka partiklarnas funktionalitet och ytkemiska egenska- per har nanopartiklar av järnoxid med en partikelstorlek på 10 nm framställts med ett ytskikt av oorganiskt material, såsom kiseloxid och guld. Kiseloxid och guld användes för att de har välkända kemiska egenskaper och har använts un- der en längre tid inom biologiska applikationer och i supraledareindustrin. Ett ytlager av kiseloxid har belagts både på ytan av enskilda nanopartiklarna i stor- leksordningen 12-60 nm och på partikelaggregat, s k beads i storleksordningen 50-250 nm. Kiselbelagda nanopartiklar av järnoxid har studerats för användning inom tillämpningar för rengörning av plasmid DNA i storleksspannet 0,1-1,7 kbp, och som optiska limiter. Återvinning av DNA med enskilda partiklar är upp till 70%, medan beads visat återvinning av ca 50%. Optiska begränsning studerades med suspensioner av coatade och uncoa- tade järnoxid nanopartiklar. Båda suspensionerna har en klämnivå av ca 3 µJ. Ickelinjära spridning av laserljuset är den dominerande mekanism. v Ïðîãóëêà â ìèëþ ñ ëþáèìîé æåíùèíîé ñîñòàâëÿåò ëèøü ñòî øàãîâ. Äàøåíüêå Preface My graduate studies were originally related to studies on the recovery of heavy metals from aqueous media by polymeric membranes. This was an extension of the studies I carried out in Mexico and for which I received my Bachelor in Science in 1999. The work was completed by presenting the Licentiate in Engineering in 2001. During the period 1999-2001, most of my colleagues were working with the synthesis of nanoparticles for various applications ranging from electronic to bio- medical. It was this interdisciplinary environment what caught my interest. After a discussion with my supervisor I started to perform some experiments and read more about the nanoworld immediately after presenting my Licentiate. Within half a year I joined the European project magnanomed, which was oriented towards some biomedical applications of iron oxide-based nanoparticles. During the time of the project we produced nanomaterials to be used for DNA separation and purifi- cation, work which we carried out in collaboration with the Max Planck Institute for Molecular Genetics in Berlin-Dahlem. The initial magnetic characterisation of the materials was undertaken by Liquids Research Ltd. at Bangor, Wales. Shortly after the end of the magnanomed project, in 2003, discussion were held with people from the Department of Polymer Technology at KTH to produce nanopar- ticles of substituted ferrites to be incorporated into polymeric composite structures. The discussions were extended a bit and they turned into two manuscripts -so far. Concomitantly with the main research efforts, during the period 2001-2004 small projects were carried out in collaboration with FOI to investigate the applicability of nanomaterials as optical limiting agents against laser radiation. During this last stage, collaborations were also established with the Department of Physics at Uni- versitat Autònoma de Barcelona and the Department of Physicochemical, Inorganic and Structural Chemistry at Stockholm University. Regarding the present work, I would like to point out certain things. As far as it was possible, SI units and terminology∗ were used throughout the thesis and the manuscripts. The values for various physical and chemical constants were obtai- ned from the Compendium of Chemical Terminology edited by the IUPAC.† Also, ∗ The definition of several terms can be found in http://wikipedia.org/ † For more information about the chemical symbols, values of universal constants, etc., please refer to http://www.iupac.org/ vii viii PREFACE in addition to the regular bibliographic information provided by the references to scientific articles, the document object identifier, DOI, was added whenever it was available.∗ G.S.A. ∗ For more information about the DOI please refer to http://www.doi.org/. LIST OF PUBLICATIONS ix List of publications I G. Salazar-Alvarez, A.A. Zagorodni and M. Muhammed, Magnetite core-inorganic shell nanoparticles for biomedical applications by novel confined-zone synthe- sis, Proceedings of the 2nd IEEE Conference on Nanotechnology 2002, 75-78. II G. Salazar-Alvarez, M. Muhammed and A.A. Zagorodni, Novel flow injection synthesis of iron oxide nanoparticles with narrow size distribution, Submitted to Chemistry of Materials. III R.T. Olsson, G. Salazar-Alvarez, M.S. Hedenqvist, U.W. Gedde, F. Lindberg and S.J. Savage, Controlled synthesis of near-stoichiometric cobalt ferrite na- noparticles, To be submitted to Chemistry of Materials. IV R.T. Olsson, M.S. Hedenqvist, U.W. Gedde, G. Salazar-Alvarez, M. Muham- med, and S.J. Savage, Synthesis and characterisation of cubic cobalt ferrite nanoparticles, J. Am. Ceram. Soc., in press. V G. Salazar-Alvarez, R. Reinhardt, and M. Muhammed, Silica-coated superpa- ramagnetic iron oxide nanoparticles for the purification of plasmid DNA, Submitted to Chemistry of Materials. VI G. Salazar-Alvarez, E. Björkman, C. Lopes, A. Eriksson, S. Svensson and M. Muhammed, Synthesis and nonlinear light scattering of nanoemulsions and na- noparticle suspensions, Submitted to Journal of Materials Chemistry. x PREFACE Contributions of the author I Performing experiments, characterisation of the samples, evaluation of the re- sults, and writing. II Planning of experiments, performing experiments, characterisation of the samples, evaluation of the results, and writing. III Planning of experiments, characterisation of the samples, evaluation of results, and writing. HRTEM analysis done by Stockholms University. IV Design of experiments, characterisation of the samples, evaluation of results, and writing. V Planning of experiments, characterisation of the samples, analysis of results, and writing. DNA purification carried out by Max Planck Institute for Mole- cular Genetics. VI Planning of experiments, characterisation of the samples, analysis of results, and writing of the manuscript. Optical measurements by the Swedish Defence Research Agency. Other work not included 1 G. Salazar-Alvarez, A.N. Bautista-Flores, E. Rodrı́guez de San Miguel, M. Mu- hammed and J. de Gyves, Transport characterisation of a PIM system used for the extraction of Pb(II) using D2EHPA as carrier. Journal of Membrane Science, in press. 2 A. Uheida, G. Salazar-Alvarez, E. Björkman and M. Muhammed, The removal of Co2+ from aqueous solution by γ-Fe2 O3 and Fe3 O4 nanoparticles. To be submitted to Separation Science and Technology. 3 G. Salazar-Alvarez, J. Nogués, F. Lindberg, R.T. Olsson, U.W. Gedde, M. Mu- hammed, On the structure and magnetic properties of disordered Cox Fe3−x O4−δ nanoparticles. In preparation 4 E. Björkman, G. Salazar-Alvarez and Mamoun Muhammed, Measurement of the H2 sorption properties of UO2(s) and fuel alloy particles Project report to Swedish Nuclear Fuel and Waste Management, SKB Symbols and acronyms Latin A Hamaker constant (N·m) B magnetic flux density (A/m, T, or Oe) Bhkl full width at half maximum (rad) C Curie’s constant CZ molar concentration of species Z(m=mol/dm3 ) D particle size (nm) E energy (J) E electrode potential (V) g gravitational acceleration constant (9.81 m/s2 ) H magnetic field strength (A/m, T or Oe) HC coercive field strength (A/m, T or Oe) J exchange energy (J) k Boltzmann’s constant (1.38 × 10−23 J/K) k kinetic formation constant K anisotropy constant (J · m−3 ) m magnetic moment (J/T) M magnetisation (A · m−1 · kg−1 or emu·cm−3 ) n a number (dimensionless) R ratio of occupied octahedral and tetrahedral sites R1 spin-lattice relaxation rate (s−1 ) R2 spin-spin relaxation rate (s−1 ) Re Reynold’s number (dimensionless) S surface area (m2 ) T1 spin-lattice relaxation time (s) T2 spin-spin relaxation time (s) T temperature (K) V particle volume (m3 ) w size polydispersity (dimensionless) Z formula units per cell xi xii SYMBOLS AND ACRONYMS Greeks ζ electrokinetic surface potential or zeta-potential(mV) θ crystallographic measuring angle (deg or ◦ ) λ wavelength (nm) µB Bohr magneton (9.274015 × 10−24 J/T) ρ density (g/cm3 ) σ standard deviation τ relaxation time (s) φ phase χ magnetic susceptibility (dimensionless or emu/Oe·g) Subscripts av average b bulk bk backward reaction B Brown B blocking C coercive C Curie fw forward reaction H hydrodynamic i initial in input out output M magnetic N Néel P physical R remanence s sample S saturation SA surface area T thermal v volume X crystallite Superscripts ◦ measured at standard room temperature and pressure xiii Acronyms AFF anionic ferrofluid CFF cationic ferrofluid DSC differential scanning calorimetry ED electron diffraction fcc face-centred cubic fis flow-injection synthesis HRTEM high resolution transmission electron microscopy or micrograph IMS immumomagnetic separation IUPAC International Union of Pure and Applied Chemistry JCPDS Joint Committee on Powder Diffraction Standards MRI magnetic resonance imaging OPL optical power limiting SAED selected area electron diffraction SI Système International d’Unités SPIO small particles of iron oxides SQUID superconductive quantum interference device magnetometry SFTR segmented flow tubular reactor STP standard temperature and pressure (298 K and 0.101 325 MPa) TEM transmission electron microscopy or micrograph TEOS tetraethyl orthosilicate THPC tetrakis(hydroxymethyl)phosphonium chloride TMAOH tetramethylammonium hydroxide USPIO ultra-small particles of iron oxides XRD X-ray diffraction or diffractogram xiv SYMBOLS AND ACRONYMS Contents Preface vii List of publications............................... ix Symbols and acronyms xi Contents xv 1 Background 1 1.1 Crystal structure............................. 1 1.2 Magnetic properties........................... 4 1.3 Applications of iron oxides........................ 8 1.4 Scope of this work............................ 12 2 Fabrication of iron oxide nanoparticles 15 2.1 Chemical methods............................ 15 3 Materials and methods 21 3.1 Synthesis of nanoparticles........................ 21 3.2 Characterisation............................. 25 3.3 DNA assays................................ 29 4 Results and discussion 31 4.1 Synthesis of various iron oxides..................... 31 4.2 Magnetic characterisation........................ 43 4.3 Applications of magnetic nanoparticles................. 45 5 Conclusions 51 Bibliography 57 xv One Background The iron oxides (generic name for iron oxides, hydroxides, oxyhydroxides, and other related compounds) have been known for millennia. Such minerals were used origi- nally as pigments for paints during the Palaeolithic. Much later, they were used in the magnetic compass when it was invented in China. This was the first application of magnetic iron oxides, also known as lodestones, used by early navigators to locate the magnetic ’north’. The most plentiful deposits of these stones were found in the district of Magnesia in Asia Minor, hence the mineral’s name became magnetite. Magnetite can be found naturally occurring in molluscs, bees, pigeons, magneto- tactic bacteria, and algae. The biogenic formation of magnetite crystals in Magne- tobactericum sp. has received a great deal of attention[Sch99] due the narrow-size distribution of the crystals (magnetosomes) and their magnetic properties.[PM02] Synthetically, magnetite can be formed by precipitation in alkaline aqueous me- dia of a mixture of Fe3+ and Fe2+ , by oxidation of Fe(II) solutions or Fe(OH)2 , by interaction of Fe2+ with ferrihydrite, or decomposition of organic precursors, etc.[CS03, WKC+ 04] 1.1 Crystal structure Several iron oxides share crystal structure with other minerals. For instance, goe- thite is isostructural with diaspore (α-AlOOH), haematite with corundum (Al2 O3 ), and magnetite with spinel (Mg Al2 O4 ). Thus, the denomination spinel ferrites∗ for the latter case. The structure of iron oxides is dominated by the arrangement of the oxygen or hydroxide anions. The cations occupy different positions relative to these layers of anions. Table 1.1 lists crystallographic information regarding some of the iron oxides. In the case of spinel ferrites, the oxygens form a fcc sublattice with the ∗ Ferrites are defined as a material composed of Fe3+ ions as the main cationic component. There are three main families of ferrites: spinel ferrites, garnet ferrites, and hexaferrites. In metallurgy, the term ferrite has a different meaning. 1 2 CHAPTER 1. BACKGROUND c a b b c a Fe+3 Fe Fe+2.5 O O-2 (a) (b) Figure 1.1: Representation of the inverse spinel structure of (a) magnetite and (b) maghemite. cations occupying 16 octahedral (B-sites) and 8 tetrahedral positions (A-sites). Figure 1.1 shows the unit cell corresponding to magnetite and maghemite, typical inverse spinel ferrites. Maghemite has basically the same crystal as magnetite. However maghemite can be considered as an Fe(II)-deficient magnetite with formula (FeIII III 8 )A [Fe40/3 8/3 ]B O32 , where represents a vacancy, A indicates tetrahedral positioning and B octahedral. Table 1.1: Crystallographic information for various iron oxides, from [CS03]. A = Cl− ; 1/2 SO2− 4 , Z is the number of formula units in the unit cell. Mineral Chemical Crystallographic Space Unit cell dimensions (nm) Z Composition system group a b c β◦ Goethite α-Fe O OH Orthorhombic Pnma 0.9956 0.30215 0.4608 4 1.1. CRYSTAL STRUCTURE Lepidocrocite γ-Fe O OH Orthorhombic Bbmm 0.3071 1.2520 0.3873 4 Akaganéite β-Fe O OH Monoclinic I2/m 1.0560 0.3031 1.0483 90.63 8 Schwertmannite Fe16 O16 (OH)y (SO4 )z · nH2 O Tetragonal P4/m 1.066 0.604 δ-Fe O OH Hexagonal P3ml 0.293 0.449 1 Feroxyhyte δ 0 -Fe O OH Hexagonal P3ml 0.293 0.456 2 High pressure Fe OOH Orthorhombic Pn21 m 0.4932 0.4432 0.2994 2 Ferrihydrite Fe5 O8 · 4H2 O Hexagonal P31c; P3 0.2955 0.937 4 Bernalite Fe OH3 Orthorhombic Immm 0.7544 0.7560 0.7558 8 Fe OH2 Hexagonal P3ml 0.3262 0.4596 1 Haematite α-Fe2 O3 Hexagonal R3̄c 0.5034 1.3752 6 Magnetite Fe3 O4 Cubic Fd3m 0.8396 8 Maghemite γ-Fe2 O3 Cubic P43 32 0.83474 8 -Fe2 O3 Orthorhombic Pna21 0.5095 0.879 0.9437 8 β-Fe2 O3 Wüstite Fe O Cubic Fm3m 0.4302 3 4 CHAPTER 1. BACKGROUND (a) (b) (c) (d) Figure 1.2: Different orientations of magnetic dipoles: (a) paramagnetic, (b) fer- romagnetic, (c) antiferromagnetic, and (d) ferrimagnetic. 1.2 Magnetic properties There are various forms of magnetism that arise depending on how the dipoles interact with each other. Figure 1.2 shows a schematic representation of the different types of arrangements of magnetic dipoles. 1.2.1 Paramagnetism Due to the magnetic field generated by unpaired electrons, atoms may behave as small magnets under the influence of an external magnetic field. However, when the applied magnetic field is removed the thermal fluctuations would make the magnetic moment of the paramagnetic material to move randomly. Under relatively low magnetic field saturation this effect can be described by Curie’s law: B M =C (1.1) T where M is the resulting magnetisation, B is the magnetic flux density of the applied field, T is the absolute temperature, and C is a material-specific Curie constant. The expected behaviour is that paramagnetic materials would increase their ma- gnetisation with the applied magnetic field but decrease with the temperature. In general paramagnetic effects are small and the response of the material to the ma- gnetic field, i.e., the magnetic susceptibility, χ, is in the order of 10−3 to 10−5 , which is several orders of magnitude smaller than that of ferromagnetic materials. Saturation magnetisation,MS , is the point at which all the spins are aligned with the magnetic field; the coercive field, HC , is the internal magnetic field of the material; and the remanent magnetisation, MR , is the magnetisation retained by the material. Figure 1.3 depicts the position of the aforementioned points in a diagram so-called ’hysteresis loop’. 1.2. MAGNETIC PROPERTIES 5 M MS Mr χi HC H Figure 1.3: Magnetisation (M ) vs. applied field (H) for ferromagnetic (solid line), paramagnetic (broken line), and diamagnetic materials (dotted line). HC represents the coercive field of the material, MS the saturation magnetisation, Mr the remanent magnetisation, and χi the initial susceptibility. 1.2.2 Ferrimagnetism In many ionic crystals the exchange energy,∗ J, between the spins of two neigh- bouring atoms is negative. This occurs due to the fact that the spin of the shared electron from the metal ions with the oxygen follows Hund’s rule and Pauli’s prin- ciple. Thus, if the metal ion have an outer shell less than half-occupied all the spins will be aligned parallel inside the shell and antiparallel with respect to the other metal ion. This means that antiparallel arrangement is the one that provides with the lowest energy. As mentioned earlier, in the case of spinel ferrites, the metal ions in the unit cell exist in 8 tetrahedral and 16 octahedral positions, where the metal ions are arranged into two sublattices A and B corresponding to tetrahedral and octahedral places. Neutron diffraction studies showed that there are two types of spinel fer- ∗ The difference in energy of two electrons in a system with antiparallel and parallel spins is called the exchange energy. 6 CHAPTER 1. BACKGROUND E1 E0 Jeff M1 O M2 Figure 1.4: Ground state and excited state of two metal ions, M1 and M2. The spin configurations as drawn become intermingled, causing an extra reduction of ground state energy; this does not occur if M1 and M2 have parallel spins. Adapted from [SW59]. rites: normal and inverse. Studies demonstrated that the normal spinel ferrites are paramagnetic whereas the inverse spinel ferrites behaved as ferromagnetic. In the case of normal spinel ferrites, the divalent metal ions are located in the sublattice A and the iron(III) ions are located in the sublattice B; whereas in the case of in- verse spinel ferrites one half of the iron(III) ions occupy the sublattice A with the sublattice B occupied by the other half of iron(III) ions and the divalent metal ions. In spinel iron oxides, the metallic ions are separated by an oxygen(II) ion and since the distance between the metallic ions is too large, it was proposed that there is a superexchange between the metallic ions through the oxygen ion.[Née48, Née52] Figure 1.4 shows a schematic representation of the superexchange interactions bet- ween two metal ions through an oxygen. Half-century ago Louis Néel explained why in the case of inverse spinel ferrites the metal ions in the sublattice A are antiparallel with respect to the metal ions in the sublattice B.[Née49] The net magnetic moment of the material is the difference in magnetic moments of sublattices A and B, which explained why the magnetic moment per formula unit was lower than the otherwise expected value. Thus, the higher the magnetic moment of the divalent cation, the higher the magnetisation.∗ For instance, manganese ferrite having Mn(II) with 5 unpaired electrons, i.e., 5 µB has a larger saturation magnetisation than that of nickel ferrite which is 2 µB. ∗ This rule does not apply for compositions containing both paramagnetic ions and diamagnetic ions. 1.2. MAGNETIC PROPERTIES 7 1.2.3 Superparamagnetism Superparamagnetic relaxation[BL59] can be explained using spherical particles – with uniaxial anisotropy– as a model system. If we consider spherical particles, the magnetic anisotropy can be approximated to be proportional to the particle volume, V. The energy barrier that separates easy magnetisation axes is the magnetic anisotropy energy, Kv V , with Kv as the volume anisotropy constant. For a small particle, the particle volume (size) is small and the magnetic energy Kv V may become comparable to (or lower than) the thermal energy (see Eq. (1.2)). Thus, the magnetic moment of the particle may fluctuate behaving as a paramagnetic atom but with a total magnetic moment that can be up to 1000 times larger than that of a single atom. Hence the term superparamagnetism. Kv V ≤ 25kT (1.2) Nevertheless, the moments of the particles follow a Boltzmann distribution when subjected to external applied magnetic field and thermal equilibrium. The orien- tation of the magnetic moment of the (non-interacting) paramagnetic particles can then be described by the Langevin function given by Eq. (1.3).[AK22] 1 mH mav = m coth α − ; α= (1.3) α kT where H is the applied magnetic field, k is Boltzmann constant, T is the absolute temperature, and m is the magnetic moment. Beside the dependence of the magnetisation on the particle size and composition, the magnetic properties depend greatly on the temperature. From Eq. (1.2) it is possible to obtain a critical temperature defined as the blocking temperature, TB , above which ferro- or ferrimagnetic particle will behave as superparamagnetic (Eq. (1.4)). Kv V TB = (1.4) 25k 1.2.3.1 Magnetic relaxation Considering particles under the influence of an external magnetic field, it can be shown that when the applied magnetic field is removed the magnetisation of the particles is reversed. There are two main mechanisms for the magnetisation reversal: spin rotation or particle rotation, which describe whether the change in direction of the magnetic moment of the particle is due to the reversal of the magnetic spin of the particle or the actual physical rotation of the particle. The time required for the reversal of the magnetic moment of the particle (spin rotation) is related to the magnetic anisotropy of the material, Néel relaxation time, τN. The mathematical expression of Néel relaxation time is shown by Eq. (1.5). 8 CHAPTER 1. BACKGROUND Kv VM τN = τ0 exp ; τ0 ∼ 10−9 s (1.5) kT The characteristic time for the randomisation of the magnetisation due mainly to the thermal motion of the particle is known as Brown relaxation time, τB , expressed by Eq. (1.6). 3VH η τB = (1.6) kT For a superparamagnetic material τN τB , which means that the magnetisation reversal occurs by rotation of the magnetic axis. Alternatively, when the magnetic moment of the particle is locked to a given crystallographic direction (easy axis), the particle is thermally blocked and τB τN. Previously it was mentioned that, under certain circumstances of particle size and temperature, ferro- or ferrimagnetic particles may be considered superparama- gnetic (see Eq. (1.2) and (1.4)). The criterion of superparamagetism is regularly assumed considering particles with a relaxation time lower than 100 s. The experimental techniques used for the measurement of magnetic properties have different measuring times and thus a particle may appear as superparamagne- tic using a slow technique such as SQUID or VSM (characteristic measuring time tens of seconds) whereas when measured with a fast technique such as Mössbauer spectroscopy (characteristic measuring time 10−7 s), it may appear as ferri- or ferro- magnetic. In the case of Fe3 O4 with an anisotropy constant, Kv = 4.4 · 104 J · m−3 , the critical superparamagnetic particle size at room temperature, T = 290 K, is between 9 and 17 nm. 1.3 Applications of iron oxides In the recent years there has been an increased interest in the application of nano- technology to biotechnology.[WH00] Thus, a number of different uses of nanoobjects has been investigated. In the case of iron oxide-based materials, the focus has been on separations and diagnostics,[WKJ83, Shi02] DNA analysis,[Uhl89] and recently, on the use of iron oxides as contrast enhancers for magnetic resonance imaging (MRI)[JLJ+ 88, SWE+ 88], among others. 1.3.1 Biomedical applications 1.3.1.1 Cell separations Micrometre-sized magnetic particles have been used as commercial applications in biology, biomedicine and biotechnology for the separation of cells. The separation of cells, or immumomagnetic separation (IMS), is achieved by covalently binding 1.3. APPLICATIONS OF IRON OXIDES 9 magnetic nanoparticles –or magnetised beads∗ – to the cells. The bead carrying a specific antibody -usually monoclonal antibodies- on the surface binds selectively the target cell which can then be removed from the suspension with the help of a magnet. The IMS has been used for the purging and isolation of cancer cells, studies of HIV and AIDS, isolation of granulocytes, isolation of cells from various tissues, stem cells, etc.[PBMPS97] 1.3.1.2 DNA analysis The separation and purification of nucleic acids has become more and more impor- tant for the accurate mapping of the genome. Nucleic acids are usually extracted from other cell components after cytolysis and repetitive washing procedures. The use of superparamagnetic particles for the analysis of plasmids and nucleic acids was introduced nearly 20 years ago by Uhlén et al.[SHO+ 88] Separation and purification of DNA with magnetisable microspheres (with a ty- pical size range of 200–2000 nm) has been reported by numerous groups.[SHO+ 88, Uhl89, HOMRS94, DTSB98] The DNA molecule is able to wrap around nanoob- jects in a similar way it does around the histone core in the nucleosomes.[LMR+ 97] This finding indicates that DNA purification can be achieved even with simple non- functionalised nanoparticles such as silica nanoparticles. However, to improve the efficiency and selectivity of the process the surface of the particles is functiona- lised with specific molecules, e.g. streptavidin, which can bind non-covalently to biotinylated-DNA.[SHO+ 88] 1.3.1.3 MRI contrast agents Magnetic resonance imaging (MRI) is based on the nuclear magnetic resonance of protons in the molecules, mainly water, that exist in a given tissue. Since the local environment of a given tissue varies depends on its position in the body it is possible to use MRI to identify various types of tissues. The characteristic measure in MRI is the proton relaxation rate or its inverse, the proton relaxation time, R1 ≡ 1/T1 and R2 ≡ 1/T2. In this technique there are two different types of proton relaxation times: the longitudinal relaxation time, T1 , (spin-lattice relaxation) and the transverse relaxa- tion time, T2 (spin-spin relaxation). Hence, there are two types of contrast agents: positive agents, which act on T1 providing a positive enhancement of the signal, appearing bright on the MRI scan; and the negative agents, which reduce the signal and give a negative enhancement, appearing as dark spots in the scan.[Pro22] Typical positive contrast agents are paramagnetic compounds based on rare earth ions, typically gadolinium-containing chelates. Nevertheless, other alterna- ∗ Beads is the term hereby used for particles or group of particles dispersed in an inorganic or organic matrix in a second step after the synthesis of the particle. The beads usually have a size from 100 nm to several micrometres and spheroidal shape. 10 CHAPTER 1. BACKGROUND tives are sought after due to the high toxicity of the heavy metals and the low contrast enhancement. However, negative contrast enhancers (T2 agents) as iron oxide nanoparticles have been used due to a much higher magnetisation per concen- tration unit,[JLJ+ 88, SWE+ 88] reported to be at least one order of magnitude higher than gadolinium-containing molecules.[Bjø02] For a more complete survey of the relaxation effects induced by superparamagnetic particles refer to Roch et al.[RMG99] Passive targeting Insofar there have been two different types of superparama- gnetic contrast agents based on iron oxides that are clinically approved: the small particles of iron oxides (SPIO, 40 < DH < 200 nm) and the ultra-small particles or iron oxide (USPIO, DH < 40 nm). Due to their large size, the SPIO agents are rapidly cleared (several minutes) from blood by the reticuloendothelial system organs such as liver and spleen, thus enabling the imaging of lesions in such organs. On the other hand, the smaller USPIO agents have longer residence times (hours), allowing for imaging of the lymphatic system.[MVGD04] Furthermore, it has been reported that the use of USPIO permits imaging for more than 24 h, and sometimes as long as five days.[NVB+ 04] USPIO have lower T1 /T2 ratio than SPIO, which leads to a higher contrast on T2 -weighted images and enables the use of relatively weak magnetic fields (< 0.5 T).[KFG+ 99] Active targeting The surface of USPIO and SPIO agents is functionalised with specific biomolecules to actively target certain tissues to be visualised by MRI tech- nique. Several attempts have been performed to improve their surface properties, e.g. coating with different types of biocompatible materials, such as protein, po- lysaccharide, DNA fragments, etc.[MKCB+ 04, PST+ 04, ICB+ 01] A schematic re- presentation of the degree of functionalisation required for such particles is given by Fig. 1.5. The visualisation of certain organs needs special functional proper- ties of coating layer.[AMOB03] Thus, for instance, Kim et al. reported the use of polysaccharide-coated superparamagnetic iron oxide nanoparticles (SPION) for the MR imaging of brain tissue.[KZK+ 01, Kim02]. Also, in case of lesions detection in the brain, e.g., multiple sclerosis, Alzheimer’s disease, etc.; The MRI contrast agent has to pass across the blood brain barrier (BBB).[RPB+ 03, PWC+ 02] However, there are limited reports on the penetration of the BBB by nanoparticles.[NH02] Moreover, after passing the BBB those agents have to be recognised and up-taken by cells. Receptor mediated endocytosis has been proposed as possible way to intro- duce contrast agent in vivo and several studies have been performed where various types of cells were labelled by SPIO and further visualised by MRI.[AFHH+ 03] 1.3. APPLICATIONS OF IRON OXIDES 11 Figure 1.5: Representation of functional nanoparticles with a protective coating layer enhanced with specific ligands. 1.3.2 Other technological applications 1.3.2.1 Optical power limiting agents Advances in the development new communication methods and new sources of ra- diation, particularly optical power sources such as lasers, had resulted in research oriented towards the protection from exposure to such sources. In this context, the development of optical power limiting (OPL) has been focused on the design of novel materials having nonlinear optical phenomena upon irradiation with high energy sources.[Hol99] Figure 1.6 shows linear and nonlinear optical phenomena as well as the clamping level that can be achieved with a nonlinear mechanism. The clamping level demonstrates that a very pronounced optical limiting can be achieved even at high energy fluences. Among the processes that produce optical limiting are two-photon absorption, reverse saturable absorption, nonlinear scattering, nonlinear beam fanning, beam steering, photochroism, etc.[PHJ+ 99] Several types of laser blocking compounds and being extensively studied for pro- tection from laser radiation, including liquid crystals,[UCS86, THD+ 97, KWG+ 98] dyes,[WHS+ 92, MRL98] and suspensions of carbon nanotubes,[VAR+ 99, LZQ+ 03] fullerenes[NW95, BS02] and different nanoparticles.[MSS92, JBHF98] 1.3.2.2 Ferrofluids Ferrofluids are made of ferro- or ferrimagnetic single-domain particles suspended in a continuous medium with no long-range order between the particles. Some appli- cation areas of ferrofluids are: sealing, vibration damping, heat transfer, bearing, sensing/detection, etc.[RM90] 12 CHAPTER 1. BACKGROUND E out linear absorption clamping level nonlinear absorption E in Figure 1.6: OPL curves: the solid line represents the linear limiting mechanism whereas the broken line shows the nonlinear mechanism. The dotted line shows the possible clamping level that can be achieved. There are certain stability requirements for the preparation of ferrofluids.[Ros85] These are influence by the thermal energy ET = kT , the magnetic energy EM = µ0 M HV , and the gravitational energy EG = δρV gDP. It is then possible to es- tablish dimensionless relationships that provide information on the stability of a ferrofluid under the influence of a particular field. The maximum particle size that would be stable against settling under the influence of a small hand-held magnet can be determined by Eq. (1.7) ET kT = ≥1 (1.7) EM µ0 M HV Rearrangin Eq. (1.7) for the particle diameter DP = (6V /π)1/3 and assuming that H = 8 × 10−4 A/m, M = 4.46 × 105 A/m, and T = 298 K, gives a DP ∼ 8 nm. At room temperature the normal gravitational field have little influence on the stability of a ferrofluid. Also, the attractive van der Waals forces must be taken into consideration. These fluctuating dipole-dipole interactions are always present and are of great importance in neutral particles. 1.4 Scope of this work This work is concerned with the synthesis of iron oxide nanoparticles with inverse spinel structure using different chemical methods with controlled particle size, par- ticle size distribution, composition, and morphology. The surface modification of the particles was undertaken with different inorganic materials, namely silica and gold. Also, the particles were investigated to evaluate their application for the pu- rification of DNA, as a contrast agent for magnetic resonance imaging, and as novel optical power limiting agents. 1.4. SCOPE OF THIS WORK 13 Papers I and II deal with the use of a novel method for the synthesis of these particles by flow-injection technique for the manufacturing of iron oxide nanopar- ticles in the size range of 1-10 nm with a narrow size distribution. In this study we investigate the hydrodynamic effects of different mixing schemes on the preparation of magnetite nanoparticles, as well as the factors influencing the size distribution of the synthesised particles. Papers III and IV are related to studies on the synthesis of cobalt ferrite par- ticles. These papers present a detailed study of the effect of the chemical synthesis parameters (concentration and concentration ratios, temperature, digestion time and addition rate) on the particle characteristics: composition, particle sizes, par- ticle size distribution, crystal structure, degree of crystallinity and morphology. The studies on the application of iron oxide nanoparticles are presented in pa- pers V and VI. In paper V, the work is focused on the preparation of magnetic nanoparticles of magnetite (Fe3 O4 ) and maghemite (γ-Fe2 O3 ) with high saturation magnetisation, their coating with silica and their application in the magnetic pu- rification of DNA. The suspensions of particles were characterised with traditional vibrating sample magnetometry and the novel nuclear magnetic relaxation. Finally, paper VI is concerned with studies on the effectiveness of nanoemulsions and sus- pensions of pristine and silica-coated iron oxide nanoparticles as novel optical power limiting agents against laser radiation. Two Fabrication of iron oxide nanoparticles 2.1 Chemical methods In the last decade, the chemical methods of preparation of nanoparticles have been used for the fabrication of several types of nanoparticles where a greater control on the size, composition, etc., could be achieved. Iron oxide nanopar- ticles have been prepared by a variety of methods[WKC+ 04] including sonoche- mical reactions,[SKG+ 97] mechanochemical synthesis,[JZRM95] hydrolysis[KKA96, SSI98] and thermolysis[HCP+ 02] of precursors, polymeric matrices,[CGLQRR97, HY02] and co-precipitation in bulk,[TM82, PK89] flow injection synthesis,[SAMZ04] and in the confined zones of microemulsions.[MP96] The previous methods were used to prepare particles with homogeneous com- position and narrow size distribution. However, the co-precipitation method in bulk has shown to be an economic and versatile technique used to synthesise large amounts of materials with different compositions and particle sizes.[Mat98] It has been found that the particles obtained in batch operation of the co-precipitation method have a broad size distribution and therefore a higher control of the synthesis conditions are required. 2.1.1 Synthesis of magnetite and maghemite Magnetite can be formed by the addition of alkali solution to an aqueous so- lution containing Fe3+ and Fe2+ in a molar ratio of 2, which leads to the for- mation of green rust complexes and thence to a dark red complex with formula 2+ FeII FeIII 2 Ox (OH)2(3−x) · xH2 O from which magnetite particles precipitates.[CS03] Also, it has been suggested that the formation of magnetite involves the interac- tion of Fe2+ with precipitated ferrihydrite. In all cases the solubility product of magnetite has to be exceeded to be able to precipitate. The hydrolysis of iron(III) species proceeds through the formation of low mole- cular weight species (Eq. (2.1)) and that above OH/Fe ∼ 1 these species interact to produce polynuclear species (Eq. (2.2)). 15 16 CHAPTER 2. FABRICATION OF IRON OXIDE NANOPARTICLES Fe3+ + H2 O Fe OH2+ + H+ (2.1a) Fe OH 2+ + H2 O Fe(OH)+ 2 +H + (2.1b) kf w 2Fe OH2+ ) −− −− Fe2 (OH)4+ * 2 (2.2) kbk The formation of the dimer is a fast reaction with a kf w = 630 m/s whereas the dissolution is very slow in the absence of protons, which suggest that further polymerisation may be very fast.[CS03] In the particular case of the spinel ferrites, the formation of magnetite, for instance, can be considered to proceed according to the overall reaction expressed by Eq. (2.3). Fe2+ + 2Fe3+ + 8OH− → Fe3 O4 + 4H2 O (2.3) whereas maghemite is formed primarily by the topotactic∗ oxidation of magnetite (Eq. (2.4)). The oxidation of synthetic magnetite to maghemite occurs via the formation of a solid solution of γ-Fe2 O3 in Fe3 O4 [CGLS65, GFM68] with the ca- tions migrating to the surface and layers of oxygen being added to the structure to preserve the crystal structure.[GFM68]      2H2 O2   2H2 O  4Fe3 O4 + 2H2 S O4 → 6γ − Fe2 O3 + 2H2 SO3 (2.4) O2 −     From Eq. (2.3) it is clearly seen that in order to obtain Fe3 O4 the spontaneous oxidation of Fe2+ (E◦Fe3+ /Fe2+ = 0.77 V) by O2 (E◦O2 /H2 O = 1.23 V) has to be prevented. Otherwise, the most common by-products of the hydrolysis of Fe3+ , i.e., goethite and haematite, are formed. Fe3+ + 3OH− → α − Fe OOH + H2 O(at around 25 ◦ C) (2.5a) 3+ − ◦ C) 2Fe + 6OH → α − Fe2 O3 + 3H2 O(at around 100 (2.5b) It has been reported that the end product of the oxidation (STP and below 600 ◦ C) of small crystals (< 3000 Å) of pure magnetite is γ-Fe2 O3 , whereas the end product of the oxidation of large crystals (> 3000 Å) is α-Fe2 O3.[FG70] Neverthe- less, it has been also reported that samples of magnetite containing impurities of α-Fe2 O3 transforms completely into haematite via epitaxial growth† on the ∗ A topotactic transformation is characterised by internal atomic displacements, which may include either loss or gain of material, so that the initial and final lattices are in coherence. † Epitaxial growth is the growth of crystals of one mineral on the crystal face of another mineral, such that the crystalline substrates of both minerals have the same structural orientation. 2.1. CHEMICAL METHODS 17 Figure 2.1: Schematic representations of reverse micelles and nanoemulsions (a) o/w nanoemulsion, (b) w/o nanoemulsion, (c) bi-continuous disper- sion, (d) isolated and aggregated o/w dispersion, and (e) isolated and aggregated w/o dispersion. faces of the spinel crystal whereas large crystals form nuclei of α-Fe2 O3 , and that it is the formation of nuclei what determines the behaviour of the particles when exposed to oxygen.[CFG+ 68] 2.1.2 Confined-zone processing The fabrication of nanoparticles with a given composition and structure requires a high degree of control on the reproducibility of the synthetic procedures. Some of the known problems are an uncontrolled mixing, low production (small units), lack of control of the residence time, etc. Zone confinement has been used to produce monodispersed particles, i.e., par- ticles with very narrow size distributions, and to tailor the particle’s morphology. One of the techniques applied recently, is the confinement of the reaction to ca- vities with pre-defined size and shape, so-called templating, e.g., synthesis of par- ticles in micro-emulsions,[OKC+ 01, MP96, PP97] sol-gel porous matrices,[MZM+ 02] aerogels,[CRM+ 02] ion exchange resins,[SFIB01, ZGW+ 92] and in cellulose.[SFIB01] 2.1.2.1 Synthesis in nanoemulsions Emulsions consist of two (or more) immiscible liquids, such as oil (organic solvent) and water (aqueous solution), in which a discontinuous phase is formed as fine droplets dispersed in a continuous phase. The dispersion is achieved by mechanical mixing to produce turbid, thermodynamically unstable solution, where the droplet size of the dispersed phase is in the order of 1 – 50 µm. 18 CHAPTER 2. FABRICATION OF IRON OXIDE NANOPARTICLES The addition of a surfactant with a suitable concentration stabilises the emulsion as it reduces the interfacial tension thus producing thermodynamically stable and transparent solutions, known as microemulsions. The droplet size in microemulsion systems is generally below 150 nm and therefore transparent to visible light. Na- noemulsion is actually an updated term of microemulsion as its dimension of the droplets is limited to the nanometre scale. The formation of the dispersed phase in an oil-surfactant-water (o/s/w) na- noemulsion system can be controlled at will; i.e. to form oil-in-water (o/w) or water-in-oil (w/o) nanoemulsions (see Figure 2.1). The oil and water phases often contain several dissolved components and therefore the selection of the surfactant (and co-surfactant) employed depends on the physicochemical characteristics of the system. Thus, a nanoemulsion requires to be designed for specific solutes in either the organic or aqueous phases by selecting an appropriate polar head (hydrophilic- lipophilic balance). Several types of surfactants as cationic, anionic, or non-ionic can be used. The specific composition of a nanoemulsion can be determined according to a ternary phase diagram of the o/s/w system. Therefore, it is essential to determine such phase diagrams for different organic solvents (oils) and surfactants, which determine the structures and properties of a nanoemulsion to be concerned. 2.1.2.2 Concept of flowing confined-zone synthesis The flow-injection synthesis (fis) technique consists of continuous or periodic mi- croinjection of reactants into a carrier solution. The fis technique takes advantage of the flow-injection systems introduced by Růžička and Hansen[RH82, RH00] and have such characteristics[VdC84]: – Injection of reactant directly into the carrier stream, achieving always condi- tions of laminar flow (Reynolds number, Re 2000) which provides with high mixing homogeneity. – Possibility of segmenting the flow with inert fluids, thus providing the system with pool zones confined to volumes of few microlitres. – Control over the dispersion of the injected reactants by manipulation over the geometric and hydrodynamic conditions of the system. – High degree of automation by including a computerised detector in the mani- fold array. – High reproducibility in injection volumes and residence times. The fis integrates several components, such as the injection system, the propul- sion unit, the capillary reactor, and the detector or container. The injection system consists, generally, of a T-shaped injector where the reactants mix head-on with high precision in time and volume. The injector can be replaced by a computer- controlled manifold that enables the system with segmentation capabilities. The injection system can be tailored so as to provide the chemical reaction with the 2.1. CHEMICAL METHODS 19 required limitations such as inert atmosphere, temperature control, etc. The pro- pulsion unit comprehends the reactant, carrier, inert media solutions, and a mean for the transportation of sample such as a peristaltic pump. The capillary reactor consists of a capillary system where the solutions are injected and where the reac- tion takes place. The residence time of the reactants and products in the reactor which is the extent of the reaction, is controlled by the length of the capillary and the pumping rate of the solutions. The fis capillary reactor is constructed of Teflon tubing and has an internal diameter of 0.1 to 1 mm and a varied length so as to provide reaction pools of few microlitres without the need of using microsystems technology. One of the main characteristics of the system, along with the reduced volumes is that the reaction point can be adjusted. For a continuous injection where the carrier participates as reactant, the reaction takes place at the point of mixing, i.e, the injection point, which as opposite to the earlier reported[JDB+ 03] segmented- flow tubular reactor (SFTR) where the reaction takes place in a micromixer outside of the capillary system. In both cases the reaction evolves under laminar flow conditions, whose advantages in mixing have been recently reported.[HBBR03] The fis methodology can be used for the production of nanosized materials, whereas the SFTR has been applied to the synthesis of micron-sized ceramic powders. Three Materials and methods 3.1 Synthesis of nanoparticles Two different approaches were used for the synthesis of the magnetic nanoparticles: bulk solution and confined-zone processing. The bulk solution processing was car- ried out in a stirred batch reactor with volumes of 100 to 1000 cm3. The pH, temperature and stirring sped were carefully adjusted to minimise variations throu- ghout the reaction time. In the second method, the reaction was carried out in confined zones provided by nanoemulsions and our earlier developed method of flow injection synthesis with constant hydrodynamic conditions.[WM99] 3.1.1 Batch synthesis in stirred reactor 3.1.1.1 Synthesis of Fe3 O4 and γ − Fe2 O3 Iron(II) and iron(III) chlorides (CFe(III) /CFe(II) = 2) were dissolved under argon at- mosphere in a deoxygenated aqueous solution of HCl 0.2 m to a total iron concen- tration of 1.5 m(iron source solution). The iron source solution was added to a deoxygenated aqueous solution of N H4 OH at 70 ◦ C under mechanical stirring. The reaction was allowed to pro- ceed for 45 min under an atmosphere of argon followed by decantation using a magnet to retain the black gel containing the magnetite nanoparticles. Magnetite particles were obtained by washing the gel several times with deoxyge- nated water, adjusting the pH to neutral, and drying it under vacuum. Maghemite particles were obtained by washing the gel several times with an aqueous solution of H N O3 0.2 m and finally drying the particles in air. 3.1.1.2 Synthesis of Nix Fe3−x O4 An aqueous solution of nickel(II) chloride was mixed with a freshly prepared aqueous solution of iron(II) chloride. The metal solution was then added to an ammonium hydroxide solution under vigorous mechanical stirring. The pH of the reaction was 21 22 CHAPTER 3. MATERIALS AND METHODS kept at 11.5-12.5 with additions of ammonium hydroxide solution and the reaction was allowed to proceed 20 min after a dark colouration appeared. The gel was then washed several times with water and dried. 3.1.1.3 Synthesis of Cox Fe3−x O4 An aqueous solution containing iron(II) sulfate and cobalt(II) chloride (CFe2+ /CCo2+ = 2) was heated up to 90 ◦ C and added to an aqueous solution of sodium hydroxide and potassium nitrate under mechanical stirring. The total mass balance in the finally mixed system was constant. The potassium nitrate was used as a mild oxidation agent in order to oxidise ferrous ions to ferric ions.[TM82] During all experiments the concentration ratio of nitrates to iron(II) was kept constant and equal to 0,66. A black precipitate was formed after mixing of the two solutions. The reaction mixture was allowed to proceed for a given period of time, after which the superna- tant was decanted by sedimenting the particles using a magnet. The particles were washed with water, dried, and milled manually. 3.1.1.4 Controlled stirred batch reactor In order to exert higher degree of control over the synthesis conditions, a fully- automated computer-controlled batch reactor was constructed. The reactor has a production capacity of ca. 40 g per batch of iron oxide. Among the controllable conditions are pH, temperature, stirring speed, and rate of addition of reactants. As schematically depicted in Figure 3.1, the reactor consists of iron source solu- tions that are pumped to the reaction vessel by peristaltic pumps, whose speed is controlled by the main computer. With a computerised pH meter in the reaction vessel the addition of acidic or basic solutions are regulated as to keep a constant pH throughout the synthesis with a σ = 0.1 pH units. The feedback obtained from the temperature probe provides information used to regulate the synthesis temperature within 1 ◦ C variation. 3.1.2 Synthesis in confined zones 3.1.2.1 Synthesis of iron oxides by Flow Injection The design of the flow injection system is described elsewhere.[WM99] The home- built system included a multi-channel peristaltic pump (Gilson Minipuls 2, France), 0.5 mm Teflon tubing lines, and a T-shaped connector or a manifold injector. The valve of a Flow Injection Analysis apparatus Fia05 (Biofak, Sweden) was used as the manifold valve gave periodic and reproducible microinjections when controlled by a computer system. The Reynolds number was calculated for the different manifold configurations and flow rates and it was always below 200, indicating that conditions of laminar flow were always attained. Typically, an acidic solution of iron with an Fe3+ : Fe2+ ratio of 2 was mixed with a stream of Na OH. The synthesis temperature was controlled with a thermostat 3.1. SYNTHESIS OF NANOPARTICLES 23 Figure 3.1: Schematic representation of the automated batch reactor and its per- ipheral components. and set to 80 ◦ C. The particles were then collected into a separation vessel where an external magnet was used to collect the product that was later washed thoroughly. Samples were dried in air and under vacuum. 3.1.2.2 Synthesis of iron oxides in nanoemulsions Nanoemulsions composed of reverse micelles were used for the synthesis of iron oxide nanoparticles. The nanoemulsion system consisted of AOT-BuOH/cHex/H2 O with a molar ratio of surfactant to water of 2,85 and a surfactant to co-surfactant molar ratio of 1. A sequential synthetic procedure was used in order to prepare the nano- particles. One nanoemulsion containing the iron source and another containing a solution of sodium hydroxide, were mixed to form the magnetite nanoparticles. The nanoemulsion was lysed with acetone to remove the particles from the surfactant and washed several times with ethanol. 24 CHAPTER 3. MATERIALS AND METHODS 3.1.3 Surface modification of nanoparticles 3.1.3.1 Preparation of ferrofluids Cationic and anionic ferrofluids were prepared modifying the procedures described by Massart.[Mas81] Briefly, the black gel obtained from the synthesis of the particles was peptised with nitric acid solution at pH=2, followed by rinsing with degassed water, the last step was repeated four times. A cationic ferrofluid (CFF), consisting mainly of maghemite, was obtained by rinsing one more time the gel with HNO3 solution, whereas an anionic ferrofluid (AFF) consisting mainly of magnetite was prepared by rinsing the gel with a TMAOH solution. Finally, the ferrofluids were dialysed against the respective acidic or alkaline solution for 24 h refreshing the solutions every 8 h. This procedure would yield stable ferrofluids with a wide concentration range, from 1 mg/cm3 to 40 mg/cm3. 3.1.3.2 Silica coating of particles Materials Besides the reagents already reported for the preparation of iron oxide, other chemicals used are tetraethyl orthosilicate (TEOS, 99%, Aldrich) and sodium trisilicate (27% Si O2 in 14% Na OH, Aldrich). Procedure Iron oxide particles from solutions AFF and CFF were coated∗ with a dense silica layer. A solution of sodium trisilicate was passed through an cationic exchange and the pH of the eluate was kept at 9. The thickness of the silica shell was varied by mixing the acidified silicate solution with the AFF or CFF at different silica/iron oxide ratios ranging from 0.05 to 5.4. To further increase the thickness of the coating shell, the silica layer on the par- ticles was grown using the Stöber method. [SFB68] Ammonia and various amounts of TEOS were dissolved in pure ethanol, to which a suspension of particles was added, the vessel was capped and shaken for 24 h. The suspension was neutrali- sed with a H Cl solution, dialysed against water and concentrated by evaporation. Multi-particle beads were prepared by varying the initial mass ratio of iron oxide to sodium trisilicate and growing the shell with the Stöber method. To check the quality of the silica coating, the coated particles were added to a 0.1 m HCl solution and the iron concentration which could eventually be leached was measured after one day. 3.1.3.3 Gold coating of particles Materials Besides the reagents already reported for the preparation of iron oxide, other chemicals used are hydroxylamine (NH2 OH, 98%, Aldrich), tetrakis(hydr- oxymethyl)phosphonium chloride (THPC, 80% in water, Aldrich) and gold pellets (99.9999%). ∗ A coated nanoparticle is defined as a fully surrounded particle by a well defined close layer of another material. 3.2. CHARACTERISATION 25 Procedure Iron oxide nanoparticles were prepared by dissolving iron(III) chloride and iron(II) sulfate with a molar ratio of CFe3+ /CFe2+ = 2 in a deoxygenated aqueous solution containing 0.1 m H Cl. The iron source solution was added to an aqueous solution of Na OH. The reaction was allowed to proceed for 1.5 h at 80 ◦ C under mechanical stirring. The gel containing the particles was rinsed with deoxygenated water several times and collected with a magnet. 2.5 cm3 of a suspension of the iron oxide nanoparticles with a concentration of about 20 mg/cm3 were diluted to 100 cm3 with water. To this, 50 cm3 of 1.87 mm N H2 OH ·H Cl were added under mechanical stirring. The pH after the addition of hydroxylamine was around 11. After 10 min 100 cm3 of a solution containing 0.44 µm Au(OH)3 was added.∗ The colour of the solution turned from dark brown to dark purple after 10-20 min and the reaction was allowed to proceed for 40 more min. After the last step the particles were rinsed with water several times. 3.2 Characterisation 3.2.1 Crystal structure Phase identification of the precipitated powders was performed by comparing X-ray powder diffraction (XRD) data obtained with a PW 1830 diffractometer (Philips, λ = 0.154021 nm) against the JCPDS standards database. The obtained spec- tra were corrected for instrumental line broadening and refined with the program celref.[LB22] 3.2.2 Electron microscopy Particle imaging was undertaken with a JEM-2000EX (JEOL) transmission electron microscope operating at an accelerating voltage of 200 kV. Specimens were dispersed in ethanol under sonication and deposited onto 200 mesh carbon-coated copper grids (Agar Scientific). The excess of solvent was blown-dried with a gentle stream of N2. High resolution images and single-particle electron diffraction patterns were ac- quired with a JEM-3010 (JEOL) high resolution transmission electron microscope operating at an accelerating voltage of 400 kV using a camera distance of 1 m. Speci- mens were dispersed in butanol under sonication and deposited onto carbon-coated grids. The excess of solvent was evaporated with light irradiation. 3.2.3 Surface characterisation 3.2.3.1 ζ-potential The electrokinetic surface potential of the particles was determined by measuring the electrokinetic sonic amplitude (ESA) with a ESA-9800 (Matec Sciences). ∗ The Au(OH)3 solution was prepared by adding 2.49 g of K2 C O3 to 1 dm3 of an aqueous solution containing 0.44 µm Au3+ 26 CHAPTER 3. MATERIALS AND METHODS 3.2.3.2 Surface area Specific surface area (SSA ) of the powders was determined by measuring the nitrogen desorption using the BET method in a Flowsorb II 2300 (Micromeritics). 3.2.4 Magnetic characterisation 3.2.4.1 Vibrating Sample Magnetometry The principle of VSM is the measurement of the electromotive force induced by a ferromagnetic sample when it is vibrating at a constant frequency, under the presence of a static and uniform magnetic field. Magnetic characterisation of the materials was undertaken by analysing approximately 40 mg of powder in fields up to 11 kOe at a vibration frequency of 67 Hz. The magnetic field error was estimated to be 2 Oe. The measurements were carried out at room temperature in an Oxford Instru- ments 1.2 VSM, at the Laboratori de Propietats Magnètiques i tèrmiques of the Universitat Autònoma de Barcelona. The powders were weighted, pressed into a small quartz container and finally mounted at the end of a rigid ceramic rod. To avoid the movement of the powders inside the tube they were compacted using quartz wool. To get optimum signal the samples are previously centred in angle and height, applying a field of 2 kOe. Sub- sequently they are demagnetised using a sequence of alternating decreasing fields. The VSM was calibrated in magnetic moment and sample temperature. The magnetic moment calibration was carried out using a Ni standard (sphere) with known saturation magnetization (MS = 54.4 emu/g). The temperature calibration was performed using ferromagnetic alumel and Ni standards, with Curie tempera- tures of 438 and 627 K, respectively. 3.2.4.2 Mössbauer spectroscopy The Mössbauer effect is based on the absorption of γ-radiation by 57 Fe. The source are 57 Co atoms in a matrix of Rh. The 57 Co nuclei decay to 57 Fe in the excited state with spin 5/2 (half-life time approx 10−9 s) and these transform fast to the state 3/2 (half-life time approx 10−7 s) and then to the state 1/2. The radiation emitted during this last step is the one that allows for the excitation of the 57 Fe atoms present in the sample. The abundance of 57 Fe is 2.2 %, which permits an efficient and statistically large enough population of atoms for a quantitative measurement. Mössbauer spectra for different samples were recorded at room temperature using a source of 57 Co atoms in a matrix of Rh. The working speed range was from -8 mm/s to 8 mm/s. A thin foil of α − Fe was used as standard to determine the baseline of the speed. 3.2. CHARACTERISATION 27 a b DP DH c d DM DX Figure 3.2: Schematic representation of particle sizes as determined with various techniques. 3.2.5 Thermal analysis Thermochemical analysis of the produced particles was performed with a differential scanning calorimeter DSC 2920 (TA Instruments). Measurements of the heat flow in the temperature range of 25-600 ◦ C were recorded for few milligrams of sample placed in aluminium pans and heated at a rate of 10 ◦ C/min. 3.2.6 Particle size measurements The developments in the preparation of particles have led to a wide number of particulate materials with different characterisation techniques providing different information on the materials. Thus, there are a number of ’particle sizes’ that arise depending on the measuring technique, as shown schematically in Figure 3.2. a) Physical size, DP : This refers to the “true” size of the particle and is usually obtained with electron microscopy. b) Hydrodynamic size, DH : The hydrodynamic size or diameter of a particle is usually determined using photon correlation spectroscopy (PCS). This length indi- cates the diameter of a particle including the fluid molecules around the electrostatic double layer. c) Magnetic size, DM : Means the size of the magnetic core of a particle. The size is usually determined from magnetic hysteresis loops. The magnetic size is the crystalline core and it does not include the non-stoichiometric surface layer. Therefore this is generally smaller than the physical size. 28 CHAPTER 3. MATERIALS AND METHODS d) Crystallite size, DX : The size corresponds to the mean value of the crystalline domain size of the particles, calculated from Scherrer’s equation which uses the line broadening of X-ray diffraction measurements. e) Surface areal size, DS : Calculated from the gas adsorption isotherms assuming a specific shape of the particles. 3.2.6.1 Physical size measurements The diameters of >500 particles were measured on digitised TEM micrographs with the image analysis program ImageJ.[Ras01] Mean diameters were thus obtained by fitting the experimental data with a lognormal distribution function, as suggested by O’Grady and Bradbury,[OB83] i.e., 0 )2 ((ln DP ) − ln DP 1 f (DP ) = √ exp − 2 (3.1) 2πwP DP 2wP with mean diameter < DP >= DP 0 exp(w 2 /2) and w as the standard deviation P P 0. The standard deviation of the mean diameter σ is around ln DP P 0 2 2 1/2 σP = DP [exp (2wP ) − exp (wP )] (3.2) 3.2.6.2 Hydrodynamic size measurements All samples were measured by dynamic light scattering (DLS) to quantify the degree of agglomeration. DLS measurements were carried out with a BI-90 particle sizer (Brookhaven Instruments Corp.). 3.2.6.3 Magnetic size measurements Measurements were carried out at room temperature in particle suspensions. The size of the magnetic cores DM were determined by assuming a lognormal distribution of particle volumes for the Langevin function, i.e., r 1/3 18kT χi DM = (3.3) πMsb 3Ms · H0 which has a standard deviation wVSM around the magnetic core of 1/2 1 3χi σM = ln (3.4) 3 Ms · 1/H0 k represents Boltzmann’s constant, T is the temperature, Msb is the saturation magnetisation of the bulk material and Ms is that of the sample; χi is the initial susceptibility (χi = dM/dHH→0 ), and 1/H0 is the intercept on the 1/H for the high field extrapolation of the magnetisation versus 1/H. 3.3. DNA ASSAYS 29 3.2.6.4 Crystallite size measurements The crystallite size DX was determined from XRD measurements using the Scher- rer’s equation 0.9λ DX = (3.5) Bhkl · cos θ where Bhkl represents the full width at half maximum of the main peak expressed in radians and θ is the position of the peak. 3.2.6.5 Surface area measurements Particle sizes (DSA ) were calculated from the SSA and the specific density of the particles measured assuming spherically shaped particles according to: 6000 DSA = (3.6) ρ · SSA 3.3 DNA assays Pristine (uncoated) iron oxide nanoparticles and silica-coated particles were used to purify plasmid DNA fragments of various sizes in the range 0,1–4 kbp (1 kb=1000 base pairs). Purifications were carried out both manually and using 96- and 384- well experimental set-ups. Initial manual DNA purification with the magnetic beads was carried out by binding the DNA to the beads, removing excess of materials by washing the beads, and eluting the DNA from the particles. Four Results and discussion In this chapter are presented the results of the investigation of the different factors that influence the synthesis of superparamagnetic and ferrimagnetic iron oxide na- noparticles , i.e., iron, cobalt and nickel spinel ferrites with narrow size distribution and different morphologies. The incorporation of the materials into stable ferro- fluids and the coating with inorganic compounds, such as silica and gold, as well as organic surfactants is also reported. In the last section of the chapter some of the investigated applications of iron spinel ferrites are presented. 4.1 Synthesis of various iron oxides 4.1.1 Preparation of magnetite and maghemite 4.1.1.1 Control of crystallographic phase The synthesis of particles was accomplished using the information provided by che- mical equilibrium diagrams. Several experiments were performed in order to inves- tigate the effect of the synthesis temperature on the final crystallographic phase of the particles. Figure 4.1 shows the DSC analysis of the prepared particles, where it is possible to observe two transitions, one endothermic at ∼100 ◦ C, associated to the desorption of surface water; and an exothermic at ∼450 ◦ C, related to the oxidation of magnetite to maghemite, Fe3 O4 → γ-Fe2 O3. Particles were then synthesised at temperatures higher than 60 ◦ C to minimise the presence of adsorbed water. Experiments were also performed using different iron(III) to iron (II) ratios to determine the boundaries at which spinel iron oxides are formed (Figure 4.2(a)). According to the results, it is possible to see that goethite is formed when there is no Fe(II) present and that even a small amount of iron(II), i.e., a high Fe(III) to Fe(II) ratio, is enough to induce the formation of the inverse spinel structure. However, due to the peak broadening associated with the small size of the particles, it is not possible to make a differentiation between magnetite and maghemite based solely on the XRD results. The particles were characterised by differential scanning 31 32 CHAPTER 4. RESULTS AND DISCUSSION o 80 C o 70 o C 60 oC 50 C Exo o 40 C 0 100 200 300 400 500 600 Temperature (oC) Figure 4.1: Calorimetric analysis of the particles synthesised at different tempe- ratures. calorimetry (Figure 4.2(b)) showing an endothermic peak centred at ∼ 270 ◦ C, which was associated to the dehydroxylation of hydroxide species present in goethite, 2α-FeO OH → α-Fe2 O3 +H2 O. It is therefore important to avoid the excess of Fe2+ during the wet preparation of Fe3 O4. Magnetite particles were synthesised and the digestion time was investigated. The size of the particles, as determined by XRD, did not change significantly. On the other hand, the most drastic change was found in the density of the materials. In order to eliminate the possibility of interference by water or incomplete dehydration, the as-synthesised powders were dried for 6 h at 120 ◦ C. Figure 4.3 shows the density variations with the digestion time, which is a result of the variation in the formation of different phases. In section 2.1.1 it was indicated that the formation of magnetite (or maghemite) proceeds through the formation of polynuclear hydroxide complexes. Hence, the gradual dehydroxylation of these complexes yields Fe3 O4 (or γ-Fe2 O3 ). The formation of different (hydr)oxide with the time result in large variations in the macroscopic density of the materials. Table 4.1 shows the density of different iron(II) and iron(III) (hydr)oxide compounds. However, the XRD diffractograms did not indicate the presence of such secondary phases, although their absence of such secondary phases may be due to the high noise-to-signal ratio arising from the low crystallinity of the samples. 4.1. SYNTHESIS OF VARIOUS IRON OXIDES 33 Ratio = 2 Ratio = 1 * * Ratio = oo * ** * ** * * * * * * * 20 30 40 50 60 70 o Diffraction angle ( 2θ) (a) Exo 0 100 200 300 400 500 600 o Temperature ( C) (b) Figure 4.2: (a) XRD (b) and DSC analysis of particles synthesised with Fe(III):Fe(II) ratios of 1, 2, ∞ in a batch reactor. The symbols indi- cate peaks corresponding to (∗) α-FeOOH, to () inverse spinel, and to () α-Fe2 O3. 34 CHAPTER 4. RESULTS AND DISCUSSION 5,0 4,8 Density (g·cm3) 4,6 4,4 4,2 10 100 1000 time (min) Figure 4.3: Variation in the density of iron oxide nanoparticles as a function of reaction time. Table 4.1: List of properties of some iron (hydr)oxide compounds. Values obtai- ned from [Lid04]. Compound Density (kg/m3 ) Fe(OH)2 3200 Fe(OH)3 3120 Fe O OH 4260 α-Fe2 O3 5250 Fe3 O4 5170 Co Fe2 O4 5240a a Theoretical value. 4.1. SYNTHESIS OF VARIOUS IRON OXIDES 35 (a) Flow-injection synthesis (b) Batch synthesis Figure 4.4: TEM micrograph and respective particle size distribution analysis for iron oxide nanoparticles prepared by fis (a) and batch methods (b). 4.1.1.2 Control of particle size and size distribution Particle analysis by TEM revealed spheroidal particles with sizes in the nanometre scale. Figure 4.4 shows the TEM for particles obtained with the continuous injection scheme (Scheme A) under standard conditions and those obtained in a small batch reactor (∼ 50 cm3 ) in identical chemical conditions. From the respective particle size distribution analysis it can be seen that the particles synthesised using a fis-based method have a lower particle size and lower polydispersity than those obtained in a batch reactor. The values determined for particles prepared in a fis reactor were 3,3 nm with a polydispersity of 0,23 compared to 4,3 nm with polydispersity of 0,33 determined for particles prepared in a batch reactor. The differences in size and distribution width (polydispersity), provides indi- cation that the mixing in the capillaries is much faster and homogeneous than in batch reactor. The Reynolds number was calculated for the fis conditions hereby studied and it was always close to zero, indicating that conditions of laminar flow were always attained. 4.1.2 Synthesis of nickel and cobalt-doped magnetite 4.1.2.1 Synthesis of Nix Fe3−x O4 Spinel ferrite nanoparticles were initially prepared using only iron salts as starting material. However, the properties of the ferrites vary dramatically as their compo- sition changes. The synthesis of nanoparticles with more complex composition was then studied. The synthesis of ferrite nanoparticles based on nickel with a chemical composition Nix Fe3−x O4 was carried out. The as-precipitated particles were cha- racterised for the composition with XRD. The obtained diffractograms are shown in Figure 4.5. 36 CHAPTER 4. RESULTS AND DISCUSSION * (d) * * * * * * 20 30 40 50 60 70 (c) o o o o 20 30 40 50 60 70 (b) 20 30 40 50 60 70 (a) 20 30 40 50 60 70 o Scattering angle ( 2θ) Figure 4.5: XRD diffractograms of Nix Fe3−x O4 particles prepared with an ini- tial nickel content of (a) 0, (b) 5, (c) 10, and(d) 30 at%. The sym- bols represent the identified peaks for spinel ferrite (∗) and haematite phases(◦ ). The results shown in Figure 4.5 indicate the formation of the spinel ferrite phase in all samples. The presence of haematite was detected only in the samples prepa- red with low amount of nickel (5 and 10 at%). Chemical analysis of the powders revealed that the nickel content in the samples that developed haematite is nearly negligible, whereas the sample prepared with a nickel content (30 at%) close to the stoichiometric one had an actual nickel content of about 20 at%. The crystallite size, as determined with Scherrer’s equation, was about 25 nm for all samples. The latter suggests that the nickel has practically no influence on the particle formation and that it might incorporate into the spinel lattice at an intermediate stage. The development of haematite in Nix Fe3−x O4 and Fe3−x O4 nanoparticles has been observed whenever there was an excess of Fe2+ in solution. As discussed earlier, haematite can grow epitaxially on magnetite or maghemite and therefore the surface area influences the kinetics of transformation. 4.1.2.2 Synthesis of Cox Fe3−x O4 Figure 4.6 shows that the cobalt content in the analysed solids can be adjusted in a controlled manner by varying the metal ion to hydroxide ion concentration ratio 4.1. SYNTHESIS OF VARIOUS IRON OXIDES 37 5,2 5,1 5,0 Density (g·cm-3 ) 4,9 4,8 4,7 4,6 4,5 0,80 0,85 0,90 0,95 1,00 1,05 1,10 1,15 1,20 x (Cox Fe3-x O4 ) Figure 4.6: Iron/cobalt molar ratio of nanoparticles as a function of the metal ion- hydroxide ion molar ratio of the mixed suspension, [Me2+ ]M /[OH− ]M. The digestion time was 3 h. in the mixed solution, [Me2+ ]M /[OH − ]M. The control of the particle composition can also be achieved by varying the initial iron to cobalt ratio.[TFM90, LK03] In the present study, particles with a cobalt composition corresponding to x values in the range 0,86 to 1,14 were synthesized. Although the initial iron to cobalt ratio prior to mixing was constant (stoichiometric ratio) the resulting phase has varying composition. Figure 4.7 shows the variation of density as a result of varying the metal to hydroxide ratio [Me2+ ]M /[OH− ]M. The pronounced decrease in density can be attributed to the gradual increase of secondary phases. Table 4.1 shows the density of various metal (hydro)oxides that may form during the precipitation of cobalt ferrite. The density values for magnetite, cobalt ferrite, and haematite are quite similar. However, the density of the hydroxide species is much lower. The redox potential in the system imposed by nitrate and iron(II) ions discard the existence of Fe(OH)2. Hence, the secondary phases that may lower the density are Fe(OH)3 and Fe O OH. The presence of these phases was detected by XRD measurements carried out on particles withdrawn from the reaction liquors at regular time intervals and at high [Me2+ ]M /[OH− ]M. 4.1.2.3 Variation of particle morphology Variation of the metal to hydroxide ratio [Me2+ ]/[OH− ] resulted in particles with different morphology. The particles were mainly cubic and spherical, although irregularly shaped particles were also observed. Fig. 4.8 shows electron micro- 38 CHAPTER 4. RESULTS AND DISCUSSION 5200 5100 5000 Density (kg·m-3 ) 4900 4800 4700 4600 4500 0,30 0,35 0,40 0,45 0,50 0,55 0,60 - [Me2+]M /[OH ]M Figure 4.7: Variation of the density of the particles with the metal to hydroxide ratio. The amount of metal ions was varied while keeping a constant amount of sodium hydroxide, nOH− = 0, 528 mol. The volume of the suspension was 1,6 dm3 and all the samples were digested for 3 h. graphs of particles observed under different conditions. Spherical and cubic par- ticles were equal in amounts when the precipitation conditions were adjusted to generate the compositions close to the empirical stoichiometric value for cobalt fer- rite, Cox Fe3−x O4 , x = 1. Several physical properties of particles are dependent on the surface composition. In this work we also studied the crystallinity and the presence of amorphous phases in the particles with HRTEM. Figure 4.9 shows the ED pattern and HRTEM of a cubic particle taken from a zone axis. The measured lattice fringes correspond to the unit cell of inverse spinel ferrite. Also, the high crystallinity of the particles can be observed, for which we can assume that each cubic particle is a single crystal with no noticeable amorphous surface layers. 4.1.3 Surface modification According to the double layer theory, most of the transition metal oxides have a point of zero charge (PZC) when suspended in aqueous media. At the PZC there is a minimum in the surface charge density and due to the positive Hamaker constant that exist between two surfaces of the same composition, there would be a strong attraction between the particles. Therefore, at neutral pH values iron oxide nanoparticles (PZC∼7) tend to aggregate into large clusters, causing the loss of properties related to superparamagnetic relaxation. The surface properties can be varied by coating the particles with a layer of organic surfactant. 4.1. SYNTHESIS OF VARIOUS IRON OXIDES 39 Figure 4.8: TEM of particles prepared with a [M2+ ]/[OH− ] ratio of (top) 0,2, (middle) 0,3, and (bottom) 0,5. Note the different scale bars used in the figures. 40 CHAPTER 4. RESULTS AND DISCUSSION Figure 4.9: ED pattern and HRTEM of a cubic particle of cobalt ferrite taken from the zone axis. The lattice fringes correspond to the unit cell of inverse spinel ferrite. 4.1.3.1 Coating of iron oxide nanoparticles by silica In this work we studied the coating of iron oxide nanoparticles with a shell of silica since it has been previously reported that silica does not affect the magnetic properties of the coated magnetic particles.[dGBMP99] The thickness of the coating layer was further increased by growing the silica shell via slow hydrolysis of tetraethoxy silane. It was observed that the thickness of the shell could be controlled by adjusting the total amount of added silane. From the graph, it can be seen that the thickness of the shell increases with the TEOS added and has a maximum at 0.15 cm3 of TEOS, then it decreases abruptly and increases at a slow rate. From the electron micrographs (4.10) it was observed that at this concentration of TEOS there is a homogeneous nucleation of silica which increases steadily with the TEOS added to the suspension.The reason why the nuclei do not increase in size is due to the fact that the number of nuclei increases instead. 4.1.3.2 Coating of iron oxide nanoparticles with gold Experiments were carried out in order to coat iron oxide nanoparticles with a layer of gold. The first test were oriented towards the preparation and characterisation of gold nanoparticles. The citrate reduction method following the

Document Details

Tags

Related

Full Transcript

Upgrade to continue