Cette thèse concerne l'étude des propriétés structurales et magnétiques de nanoparticules magnétiques creuses (HMNPs), coquille et coquille/coquille. Les effets de surface sont exaltés de par la présence des surfaces interne et externe. L'étude expérimentale de HMNPs basée sur des mesures magnétiques et de spectrométrie Mössbauer du 57Fe a montré une structure magnétique complexe. Les HMNPs ayant une épaisseur ultrafine présentent une structure magnétique décrite par 2 sous-réseaux spero-magnétiques opposés, en plus de la présence d’un champ d'échange bias significatif. L'effet de la taille et de l'épaisseur des HMNPs a été également étudié. Les spectres Mössbauer obtenus sous champ magnétique montrent que la structure magnétique est fortement corrélée au rapport surface/volume. Ces résultats expérimentaux ont été confirmés par simulation Monte Carlo. Après optimisation du modèle, l’approche numérique montre d’abord que l'anisotropie de surface Ks gouverne le comportement magnétique des HMNPs et ensuite que la valeur critique de Ks nécessaire pour obtenir une configuration radiale (spike) diminue lorsque la taille des HMNPs augmente. L'étude numérique menée pour différentes tailles et épaisseurs de coquille, a permis de suivre leurs effets sur la structure magnétique des HMNPs. Par ailleurs, l'étude expéri-mentale menée sur des HMNPs shell/shell, montre que le désordre des spins et le champ d'échange bias deviennent plus importants lorsque les HMNPs sont recouvertes d’une coquille antiferromagnétique (NiO). De ces résultats, on peut déduire l'effet du désordre des spins sur les phénomènes d'échange bias dans un tel système. / This thesis concerns the study of structural and magnetic properties of hollow magnetic nanoparticles (HMNPs), shell and shell/shell. These HMNPs present enhanced surface effects resulting from the presence of both inner and outer surface layers. The experimental investigation combining magne-tic measurements and 57Fe Mössbauer spectrometry of such HMNPs has revealed a complex spin magnetic structure. Small HMNPs with ultrathin thickness show highly disordered magnetic structure and the corresponding in-field hyperfine structure can be described by means of 2 speromagnetic antiferromagnetically coupled, in addition to the significant exchange bias phenomenon. The in-field Mössbauer study of the effect of size and thickness of HMNPs shows that the spin disorder is strongly correlated to the surface to volume ratio. Those experimental magnetic behaviors were confirmed using Monte Carlo simulation. Indeed, after improving the numeric model, it is concluded that surface anisotropy Ks has a dominant role in the magnetic behavior of HMNPs and the value of critical Ks necessary to obtain radial (spike) configuration decreases as the size of HMNPs increases, keeping the same thickness. The numeric study for different sizes and shell thicknesses allows the effect of these parameters on the spin structure of HMNPs to be followed. Then, the experi-mental study extended to shell/shell HMNPs indicates that the spin disorder is enhanced in HMNPs with antiferromagnetic shell (NiO) in addition to larger exchange bias field. From those results, one can try to deduce the effect of spin disorder on the exchange bias phenomena in such system.

When a permanent magnet is considered for an application, the quantity that quantifies the usability of that material is the magnetic energy product (BH)max. In today’s world, rare earth transition metal permanent magnets like Nd-Fe-B, Sm-Co possesses the maximum magnetic energy product. But still for the industrial application, the ferrite permanent magnets are the primary choice over these rare transition metal magnets. Thus, in the present context, the magnetic energy product of the low cost ferrite system makes it unsuitable for the high magnetic energy application. In this regard, exchange spring magnets which combine the magnetization of the soft phase and coercivity of the hard magnetic phases become important in enhancing the magnetic energy product of the system. In this thesis, the exchange spring behaviour is reported for the first time in hard/soft oxide nanocomposites by microstructural tailoring of hard Barium Ferrite and soft Nickel Zinc Ferrite particles. We have analyzed the magnetization reversal and its correlation with the coercivity mechanism in the Ni0.8Zn0.2Fe2O4/BaFe12O19 exchange spring systems. Using this exchange spring concept, we could enhance the magnetic energy product in Iron Oxide/ Barium Calcium Ferrite nanocomposites compared to the bare hard ferrite by ~13%. The presence of the exchange interaction in this nanocomposite is confirmed by the Henkel plot. Moreover, a detailed Reitveld study, magnetization loop and corresponding variation of the magnetic energy product, Henkel plot analysis and First Order Reversal Curve analysis are performed on nanocomposites of hard Strontium Ferrite and soft Cobalt Ferrite. We have proved the exchange spring behaviour in this composite. In addition, we could successfully tailor the magnetization behaviour of the soft Cobalt Ferrite- hard Strontium Ferrite nanocomposite from non exchange spring behaviour to exchange spring behaviour, by tuning the size of the soft Cobalt Ferrite in the Cobalt Ferrite/Strontium Ferrite nanocomposite. The relative strength of the interaction governing the magnetization process in the composites has been studied using Henkel plot and First Order Reversal Curve method. The FORC method has been utilized to understand the magnetization reversal behaviour as well as the extent of the irreversible magnetization present in both the nanocomposites, having smaller and larger particle size of the Cobalt Ferrite. It has been found that for the all the studied composites, the pinning is the dominant process for magnetization reversal. The detailed structural analysis using thin film XRD, angle dependent magnetic hysteresis and remanent coercivity measurement, coercivity mechanism by micromagnetic analysis and First Order Reversal Curve analysis are performed for thin films of Strontium Ferrite which are grown on c-plane alumina using Pulsed Laser Deposition (PLD) at two different oxygen partial pressures. The magnetic easy directions of both the films lie in the out of plane direction where as the in plane direction corresponds to the magnetic hard direction. Depending on the oxygen partial pressure during deposition, the magnetization reversal changes from S-W type reversal to Kondorsky kind of reversal. Thus, the growth parameter for the Strontium Ferrite single layer which will be used further as a hard layer for realizing oxide exchange spring in oxide multilayer, is optimized. The details of the magnetic and structural properties are analyzed for Nickel Zinc Ferrite thin film grown on (100) MgAl2O4. We have obtained an epitaxial growth of Nickel Zinc Ferrite by tuning the growth parameters of PLD deposition. The ferromagnetic resonance and the angle dependent hysteresis loop suggest that, the magnetic easy direction for the soft Nickel Zinc Ferrite lie in the film plane whereas the out of plane direction is the magnetic hard direction. Using the growth condition of respective Nickel Zinc Ferrite and Strontium Ferrite, we have realized the exchange spring behaviour for the first time in the trilayer structure of SrFe12O19 (20 nm)/Ni0.8Zn0.2Fe2O4(20 nm)/ SrFe12O19 (20 nm) grown on c-plane alumina (Al2O3) using PLD. The FORC distribution for this trilayer structure shows the single switching behaviour, corresponding to the exchange spring behaviour. The reversible ridge measurement shows that the reversible and the irreversible part of the magnetizations are not coupled with each other.

The present thesis describes the synthesis and characterization of magnesium, copper, and iron complexes and their application in the MOCVD (Metal-Organic Chemical Vapor Deposition) process, as well as the synthesis and characterization of yttrium and gadolinium complexes and their use as spin-coating precursors for metal oxide thin layer formation. The objective of this scientific work is the development of the family of bis(β-ketoiminato) magnesium(II) complexes and a series of heteroleptic β-ketoiminato copper(II) precursors for the formation of magnesium oxide and copper/copper oxide layers by using the MOCVD process. Modifications of the ketoiminato ligands affect the physical and chemical properties of the respective complexes. Another central theme of this work is the development of β-diketonato iron(III) complexes for the deposition of carbon-free gamma- and alpha-Fe2O3 layers via MOCVD. The thermal behavior and vapor pressure of the precursors could be influenced by the variation of the β-diketonate ligands. In addition, the synthesis and characterization of yttrium and gadolinium β-diketonates and their use as spin-coating precursors are described. Field-effect transistors were successfully fabricated by the deposition of carbon nanotubes on top of the Y2O3 films.

The Per Geijer iron oxide-apatite deposits are important potential future resources for Luossavaara-Kiirunavaara Aktiebolag (LKAB), which has been continuously mining magnetite/hematite ores in northern Sweden for almost 130 years. The Per Geijer deposits reveal a high phosphorus content and vary from magnetite-dominated to hematite-dominated ores, respectively. The high phosphorus concentration of these ores results from highly elevated content of apatite as gangue mineral. Reliable, robust, and qualitative characterization of the mineralization is required as these ores inherit complex mineralogical and textural features. The precise mineralogical information obtained by optical microscopy, SEM-MLA and Raman improves the characterization of ore types and will benefit future processing strategies for this complex mineralization. The different approaches demonstrate advantages and disadvantages in classification, imaging, discrimination of iron oxides, and time consumption of measurement and processing. A comprehensive mineral-chemical dataset of magnetite, hematite and apatite obtained by electron microprobe analysis (EPMA) and LA-ICP-MS from representative drill core samples is presented. Magnetite, four different types of hematite and five types of apatite constitute the massive orebodies: Primary and pristine magnetite with moderate to high concentrations of Ti (∼61–2180 ppm), Ni (∼11–480 ppm), Co (∼5–300 ppm) and V (∼553–1831 ppm) indicate a magmatic origin for magnetite. The presence of fluorapatite and associated monazite inclusions and disseminated pyrite enclosed by magnetite with high Co:Ni ratios (> 10) in massive magnetite ores are consistent with a high temperature (∼ 800°C) genesis for the deposit. The different and abundant types of hematite, especially hematite type I, state subsequent hydrothermal events. Chromium, Ni, Co and V in both magnetite and hematite have low concentrations in terms of current product regulations and thus no effect on final products in the future. In terms of a possible future hematite product, titanium seems to be the most critical trace element due to very high concentrations in hematite types I and IV, of which type I is most abundant in zones dominated by hematite. Further interest on other products is generated due to the high variability of hematite and apatite in some of these ores. Information obtained from comminution test works in the laboratory scale can be utilized to characterize ore types and to predict the behavior of ore during comminution circuit in the industrial scale. Comminution tests with a laboratory rod and ball mill of 13 pre-defined ore types from the Per Geijer iron-oxide apatite deposits were conducted in this study. The highest P80 values were obtained by grinding in the rod mill for 10 minutes only (step A). Grinding steps B (25 min ball mill) and C (35 min ball mill) reveal very narrow P80 values. Ore types dominated by hematite have significantly higher P80 values after the primary grinding step (A), which indicates different hardness of the ore types. P80 values are generally lowest after the secondary grinding step C ranging between 26 µm (ore type M1a) and 80 µm (ore type H2a). Generally, Fe content increases in the finer particle size classes while CaO and P contents decrease. The influence of silica or phosphorus seems to be dependent on the dominant iron oxide. Magnetite-dominated ore types are more likely to be affected in their comminution behavior by the presence of the silicates. Contrary, hematite-dominant ore types are rather influenced by the presence of apatite. The difference in the degree of liberation of magnetite and hematite between ore types depends rather on size fractions than the amount of gangue in the ore. Davis tube data indicates that magnetite can be separated from gangue quite efficiently in the magnetite-dominated ore types. Contrary to magnetite ore, hematite-dominated ore types cannot be processed by DT. It is favored to use strong magnetic separation in order to achieve a desirable hematite concentrate. The magnetic material recovered by DT is most efficiently separated at an intensity current of 0.2 A, whereas above 0.5 A the separation process is neglectable. Based on comminution and magnetic separation tests a consolidation to eight ore types is favored which supports possible future mining of the Per Geijer deposits.:Contents ABSTRACT ……………………………………………………………………… I CONTENTS ……………………………………………………………………… II LIST OF FIGURES AND TABLES ……………………………………………… IV LIST OF ABBREVIATIONS ……………………………………………… V 1 INTRODUCTION ……………………………………………………… 1 1.1 Background and motivation of study ……………………………… 2 1.2 Previous and current work on the Per Geijer deposits ……………… 3 1.3 The need for mineral processing and in-situ ore description ……………… 4 1.4 General and generic aspects on iron oxide apatite deposits ……………… 5 Chapter A 2 REGIONAL GEOLOGY ………………………………………………. 7 2.1 Local geology of the Kiruna area ……………………………………… 7 2.2 Geology of the Per Geijer deposits ……………………………………… 9 3 METHODOLOGY ……………………………………………………… 12 3.1 Core sampling and preparation ……………………………………… 12 3.2 SEM – MLA in-situ ore ……………………………………………… 14 3.3 Electron Probe Microanalyses (EPMA) ……………………………… 15 3.3.1 Iron oxide measurements ……………………………………… 15 3.3.2 Apatite measurements ……………………………………… 15 3.4 In-situ LA-ICP-MS ……………………………………………………… 16 3.5 Whole-rock geochemistry ……………………………………………… 18 3.5.1 Exploration drill core assays ……………………………… 18 3.5.2 Chemical assays of rock chips ……………………………… 18 4 RESULTS ……………………………………………………………… 19 4.1 Pre-definition of ore types ………………………………...……………. 19 4.2 Mineralogy of in situ ore ……………………………………………… 21 4.2.1 Ore Type M1a ……………………………………………… 21 4.2.2 Ore Type M1b ……………………………………………… 22 4.2.3 Ore Type M2a ……………………………………………… 23 4.2.4 Ore Type M2b ……………………………………………… 25 4.2.5 Ore Type HM1b ……………………………………………… 26 4.2.6 Ore Type HM2a ……………………………………………… 27 4.2.7 Ore Type HM2b ……………………………………………… 28 4.2.8 Ore Type H1a ……………………………………………… 29 4.2.9 Ore Type H1b ……………………………………………… 30 4.2.10 Ore Type H2a ……………………………………………… 31 4.2.11 Ore Type H2b ……………………………………………… 32 4.2.12 Comparison of ore types ……………………………………… 33 4.3 Geochemistry of in situ ore types ……………………………… 36 4.3.1 Whole-rock chemical assays of drill cores ……………………… 36 4.3.2 Whole-rock geochemistry of rock chips ……………………… 39 4.4 Mineral chemistry of iron oxides ……………………………………… 42 4.4.1 Iron oxides and associated minerals ……………………………… 42 4.4.2 Mineral chemistry of magnetite from Per Geijer ……………… 43 4.4.3 Mineral chemistry of hematite from Per Geijer ……………… 47 4.5 Mineral chemistry of apatite ……………………………………… 51 4.5.1 Apatite and associated minerals ……………………………… 51 4.5.2 Mineral chemistry of apatite from Per Geijer ……………… 53 Chapter B 5 COMMINUTION TESTS ……………………………………………… 58 5.1 Methodology of comminution tests ……………………………………… 59 5.1.1 Sampling for comminution tests ……………………………… 59 5.1.2 Comminution circuit ……………………………………………… 61 5.1.3 Energy consumption calculation ……………………………… 62 5.1.4 SEM – MLA ……………………………………………………… 64 6 MAGNETIC SEPARATION TESTS ……………………………… 65 6.1 Methodology of magnetic separation by Davis magnetic tube ……… 66 6.2 Davis magnetic tube tests for characterization of the Per Geijer ore types 66 6.3 Separation analysis based on the Henry-Reinhard charts .……………... 67 7 RESULTS OF COMMINUTION OF ORE TYPES ……………………… 69 7.1 General characteristics of magnetite-dominated ore types ……………… 69 7.2 General characteristics of hematite-dominated ore types ……………… 72 7.3 General characteristics of magnetite/hematite-mixed ore types ……… 75 7.4 General characteristics of low-grade ore types ……………………… 77 7.5 Mineral liberation characteristics of magnetite-dominated ore types 79 7.6 Mineral liberation characteristics of hematite-dominated ore types 83 7.7 Mineral liberation characteristics of magnetite/hematite-mixed ore types 87 7.8 Mineral liberation characteristics of low-grade ore types ……………… 90 7.9 Total energy consumption of ore types from the Per Geijer deposits 94 8 RESULTS OF MAGNETIC SEPARATION OF ORE TYPES ……… 95 8.1 Magnetic separation of magnetite-dominated ore types ……………… 95 8.2 Magnetic separation of hematite-dominated ore types ……………… 96 8.3 Magnetic separation of magnetite/hematite-mixed ore types ……………… 97 8.4 Magnetic separation of low-grade ore types ……………………………… 98 8.5 Henry-Reinhard charts ……………………………………………… 99 9 DISCUSSION ……………………………………………………… 101 9.1 Mineralogy of the in-situ ore types from the Per Geijer deposits ……… 101 9.2 Geochemistry of the in-situ ore types from the Per Geijer deposits ……… 103 9.3 Mineral chemistry of iron oxides from the Per Geijer deposits ……… 105 9.4 Mineral chemistry of apatite from the Per Geijer deposits ……………… 114 9.5 Comminution of ore types from Per Geijer ……………………… 119 9.6 Magnetic separation of ore types from Per Geijer ……………………… 120 9.7 Issues with process mineralogy of in-situ and grinded ore types ……… 121 10 CONCLUSIONS ……………………………………………………… 128 11 IMPLICATIONS FOR FUTURE WORK ……………………………… 131 12 REFERENCES ……………………………………………………………… 134

This work investigates the initial growth of iron and iron oxides on Ag(001). Surface structure and morphology of both post deposition annealed Fe films (in UHV and O2 atmosphere) as well as reactive grown iron oxide films will be analyzed in detail by low energy electron diffraction (LEED) and scanning tunneling microscopy (STM). The stoichiometry at the surface of the iron oxide films will be determined by X-ray photoelectron spectroscopy (XPS) and Auger electron spectroscopy (AES). The main focus of this work is to shed light on the question whether the growth of iron oxide films on Ag(001) is accompanied by the formation of strain reducing dislocation networks, or superstructures as found for other metal substrates in former studies. Here, we will distinguish between Fe films which were post deposition annealed in a thin O2 atmosphere and reactively grown iron oxide films.

Since many years, the importance of thin layers increases for lots of technical uses. Beginning in the field of microelectronics, the use of thin layers spread increasingly to other areas. Coatings for surface refining and optimisation of the mechanical properties for material engineering, customisation of the surface chemistry in catalysts, as well influencing of the transmission and reflection characteristics of surfaces in optics are only some examples of the high scientific and economic weight of the thin layer technology. Thin magnetic layers are the basis of many known storage media ranging from the tape recorder to the hard disk up to the credit card. Nowadays, these thin layers again gain interest in the research field of nanoelectronics as ultrathin layers. So-called spinvalve-read/write heads being already installed in actual hard disks use the Tunnel Magneto Resistance effect for a significant rise in memory density synonymous capacity. Such read/writeheads consist of a magnetic layersystem. This use of the magnetic as well as the electric characteristics of the electrons is called spintronics. The iron oxide magnetite exhibits a high iron portion, is strong antiferrimagnetic and has a high Curie-temperature. Since many years, it is used as a magnetic pigment on already mentioned magnetic tapes. Literature [1, 2, 3, 4] considers ultrathin epitaxial layers of magnetite on magnesium oxide for uses in the spintronics as a most promising candidate, because it inheres a complete spin polarisation at Fermi-level. Moreover, thin magnetite layers serve in the chemical industry as a catalyst in the Haber- Bosch-procedure and to the dehydration of ethylbenzene to styrene. Being already used and considered to be of ongoing interest, ultrathin magnetite layers offer a wide range of technological applications in many modern industrial and scientific fields. Because there is, nevertheless, a variety of other iron oxide (cf. chapter 4), it is a matter to determine the special growth conditions of magnetite. These ultrathin iron oxide layers were grown reactively on the (001)-surfaces of the magnesium oxide substrate by molecular beam epitaxy. Besides, the surface is examined by the diffraction of low-energy electrons concerning its crystalline structure. X-ray photo electron spectroscopy approaching the stochiometry completes these first characterisations. Other investigations are carried out at HASYLAB / DESY in Hamburg by X-ray reflectivity and X-ray diffraction. The exact thickness of the layers, its crystal properties in bulk, as well as the thickness of the crystalline portion of the layers can be determined among other features of the system. The evaluation of XRR-and XRD-investigations is done via simulations with in chapter 5 introduced software packages. The reader finds the theoretical backgrounds to the used techniques in chapter 3. The experimental setups in Osnabr¨uck and Hamburg as well as the backgrounds to the preparation are presented in chapter 5. Because the formation of the different iron oxides is described in literature [5, 6, 7, 8] as mostly depending on annealing temperatures, the experimental results in chapter 6 are graded accordingly. The dependence on temperature, layer thickness and annealing time should be examined for the iron oxides possible on this substrate. The aim of this work is the preparation of ultrathin epitaxial iron oxide layers with thicknesses up to few nanometers. The main goal is to find the growth parameters for ultrathin crystalline magnetite layers.

Safety concerns arising from cytotoxic behavior of nanoparticles (NPs) in complex biological environment remain the main problem limiting NPs application in biomedicine. In this study, we have investigated cytotoxicity of NPs with different composition, shape and size, namely SiO2 NPs (SiNPs, 7-14 nm), superparamagnetic iron oxide NPs (SPIONs, 8 nm) and carboxylated multiwalled carbon nanotubes (CNTCOOHs, diameter: 60-100 nm, length: 1-2 μm). Cytotoxicity was evaluated with newly designed screening assay capable to simultaneously assess activity of cell dehydrogenases, activity of lactate dehydrogenase (LDH) released from cells into environment and number of intact cell nuclei and apoptotic bodies in human umbilical vein endothelial cell (HUVEC) culture growing in the very same well of the 96-well plate. Aforementioned attributes were subsequently utilized to obtain information about cell viability and necrotic and apoptotic aspects of cell death. Results from this "three-in-one" cell death screening (CDS) assay showed that SiNPs and CNTCOOHs evoked pronounced cytotoxic effect demonstrated as decrease of cell viability and development of apoptotic bodies formation. In contrast to this, SPIONs induced only mild cytotoxicity. Moreover, SiNPs impaired cell membrane leading to increased LDH release...

Cette thèse présente les travaux de développement de formulations injectables capables de se solidifier in situ, formant ainsi un implant piégeant des microparticules magnétiques en vue du traitement de tumeurs par induction magnétique d'une hyperthermie locale modérée. Nous exposons tout d'abord le contexte physique, biologique et clinique de l'hyperthermie comme traitement anticancéreux, particulièrement des modalités électromagnétiques. Les performances in vitro et in vivo des matériaux et formulations sont alors présentées. L'objet du chapitre suivant est la caractérisation des propriétés physicochimiques, magnétiques, et chauffantes, dans un champ magnétique alternatif (115 kHz, 9 - 12 mT), des microparticules de silice renfermant des nanoparticules d'oxyde de fer superparamagnétiques (SPIONs) et de deux de leurs formulations: un hydrogel d'alginate de sodium et un organogel de poly(éthylène-co-alcool vinylique) dans le diméthylsulfoxide. Finalement, nous présentons le potentiel thérapeutique de 20 minutes d'hyperthermie locale induite après injection de l'organogel superparamagnétique dans un modèle murin sous-cutané de tumeurs nécrosantes de colocarcinome humain.

The atomic metal anions Rb-, Cs-, Cu-, Ag- and Fe- have been generated in the gas phase and reacted with various neutral reactants in a triple quadrupole mass spectrometer. The metal anions were formed via electrospray ionization of the metal-oxalate solutions and form in gas phase between the capillary and the first quadrupole. Neutral gas phase reactants investigated include NO, NO2, SO2, C6F5OH, C6F5NH2, C6F6, E-octafluoro-butene and 1,2,3/1,2,4/1,3,5 trifluoro-benzene. When possible, CBS-4M methods were used to suggest the lowest energy products based on relative energy. Observed reactions of atomic metal anions with the aforementioned neutral species include electron transfer and dissociative electron transfer to the neutral gas phase reactants. In addition, hydrogen abstraction and fluorine abstraction forming a neutral metal hydride or fluoride as well as the formation of multiply substituted metal-oxide/fluoride anions was also observed. Metal-complex anions observed from the gas phase reactions include CuF-,CuF2-,CuO-,CuO2-, FeO-, FeO2-, FeO3-, FeF-, FeF2-, FeF3-, CsF- and CsF2-.

The atomic metal anions Rb-, Cs-, Cu-, Ag- and Fe- have been generated in the gas phase and reacted with various neutral reactants in a triple quadrupole mass spectrometer. The metal anions were formed via electrospray ionization of the metal-oxalate solutions and form in gas phase between the capillary and the first quadrupole. Neutral gas phase reactants investigated include NO, NO2, SO2, C6F5OH, C6F5NH2, C6F6, E-octafluoro-butene and 1,2,3/1,2,4/1,3,5 trifluoro-benzene. When possible, CBS-4M methods were used to suggest the lowest energy products based on relative energy. Observed reactions of atomic metal anions with the aforementioned neutral species include electron transfer and dissociative electron transfer to the neutral gas phase reactants. In addition, hydrogen abstraction and fluorine abstraction forming a neutral metal hydride or fluoride as well as the formation of multiply substituted metal-oxide/fluoride anions was also observed. Metal-complex anions observed from the gas phase reactions include CuF-,CuF2-,CuO-,CuO2-, FeO-, FeO2-, FeO3-, FeF-, FeF2-, FeF3-, CsF- and CsF2-.