Repositorio Institucional de Documentos

2021 56 Javier Pablo Navarro Development and optimization of 3D advanced functional magnetic nanostructures grown by focused electron beam induced deposition Departamento Director/es Física de la Materia Condensada De Teresa Nogueras, José María Magén Domínguez, César © Universidad de Zaragoza Servicio de Publicaciones ISSN 2254-7606 Javier Pablo Navarro DEVELOPMENT AND OPTIMIZATION OF 3D ADVANCED FUNCTIONAL MAGNETIC NANOSTRUCTURES GROWN BY FOCUSED ELECTRON BEAM INDUCED DEPOSITION Director/es Física de la Materia Condensada De Teresa Nogueras, José María Magén Domínguez, César Tesis Doctoral Autor 2020 Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA Departamento de Física de la Materia Condensada Instituto de Nanociencia de Aragón (INA) Instituto de Ciencia de Materiales de Aragón (ICMA) Universidad de Zaragoza - Consejo Superior de Investigaciones Científicas (CSIC) Doctoral Thesis Development and optimization of 3D advanced functional magnetic nanostructures grown by focused electron beam induced deposition ________________________________________ Javier Pablo Navarro Zaragoza, August 2019 Thesis supervisors: José María de Teresa Nogueras César Magén Domínguez Contents Agradecimientos (Acknowledgements) ........................................................................ 8 Acronyms ...................................................................................................................... 11 Abstract ......................................................................................................................... 13 References .............................................................................................................. 16 Resumen ........................................................................................................................ 17 Referencias ............................................................................................................. 20 1 Introduction ............................................................................................................... 23 1.1 Nanotechnology ................................................................................................ 24 1.2 Nanoelectronics ................................................................................................ 26 1.2.1 Non-volatile resistive memory ............................................................... 28 1.3 Spintronics ........................................................................................................ 29 1.3.1 Magnetic domain walls .......................................................................... 30 1.3.2 Racetrack memory concept .................................................................... 31 1.4 Further applications of magnetic nanowires ..................................................... 32 1.4.1 Spincaloritronics .................................................................................... 33 1.4.2 High frequency devices .......................................................................... 33 1.4.3 Biomedical applications ......................................................................... 34 1.4.4 Magnetic Force Microscopy tips ............................................................ 35 1.4.5 Magnetoplasmonics ............................................................................... 35 1.5 Design and fabrication of magnetic nanowires ................................................. 36 1.5.1 Electrochemical synthesis ...................................................................... 36 1.5.2 Sol-gel technique.................................................................................... 40 1.5.3 Vapor-liquid-solid and vapor-solid processes ........................................ 41 1.5.4 Chemical vapor transport method .......................................................... 42 1.5.5 Quenching and drawing technique ......................................................... 44 1.5.6 Focused Electron Beam Induced Deposition ......................................... 44 1.6 Outline of the thesis .......................................................................................... 46 References ............................................................................................................... 47 2 Experimental techniques ........................................................................................... 57 2.1 Microand nanolithography techniques ........................................................... 58 2.1.1 Dual Beam SEM-FIB system ................................................................. 58 2.1.1.1 Focused Electron Beam Induced Deposition ............................... 64 2.1.2 Optical lithography................................................................................. 66 2.2 Transmission Electron Microscopy .................................................................. 68 2.2.1 Transmission Electron Microscopy techniques ...................................... 73 2.2.1.1 Electron Energy Loss Spectroscopy ............................................ 74 2.2.1.2 Energy-Dispersive X-ray Spectroscopy ...................................... 76 2.2.1.3 Off-Axis Electron Holography .................................................... 77 2.3 Annealing techniques ........................................................................................ 82 2.4 Further magnetic characterization techniques .................................................. 83 2.4.1 Superconducting Quantum Interference Device magnetometry ............ 83 2.4.2 Magneto-Optical Kerr Effect magnetometry ......................................... 84 2.4.3 Magnetic Force Microscopy .................................................................. 85 References ............................................................................................................... 87 3 Tuning the growth of 3D nanowires by FEBID ...................................................... 95 3.1 Principles of FEBID .......................................................................................... 96 3.2 Tailoring the fundamental properties of 3D cobalt nanowires ......................... 99 3.2.1 Experimental details ............................................................................. 101 3.2.2 Linear and radial growth regimes......................................................... 102 3.2.3 Composition as a function of the growth regime ................................. 106 3.2.4 Magnetic induction as a function of the diameter ................................ 112 3.2.5 Discussion of the results ...................................................................... 114 3.3 Electrically-biased patterned metal structure .................................................. 115 3.3.1 Introduction .......................................................................................... 116 3.3.2 Experimental details ............................................................................. 117 3.3.3 Growth on insulating substrates ........................................................... 119 3.3.4 Electric field numerical calculations .................................................... 121 3.3.5 Dimensional modulation as a function of the applied voltage ............. 124 3.3.6 Qualitative discussion of the electric field action ................................ 127 3.3.7 Dimensional modulation as a function of the beam defocus ............... 129 3.4 Conclusions .................................................................................................... 136 References ............................................................................................................ 137 4 Towards properties improvement by thermal annealing .................................... 145 4.1 Introduction .................................................................................................... 146 4.2 Annealing process on 3D cobalt nanowires .................................................... 148 4.2.1 Experimental details ............................................................................. 149 4.2.2 Structural and chemical changes as a function of the annealing temperature .......................................................................................... 154 4.2.3 Magnetic induction dependence with annealing temperature .............. 162 4.2.4 Magnetic characterization by nanoSQUID magnetometry .................. 165 4.3 Annealing process on 3D iron nanowires ....................................................... 172 4.3.1 Experimental details ............................................................................. 172 4.3.1.1 In situ post-growth annealing .................................................... 173 4.3.1.2 Ex situ post-growth annealing ................................................... 174 4.3.2 Morphology and composition of in situ annealed nanowires .............. 176 4.3.3 Morphology and composition of ex situ annealed nanowires .............. 185 4.4 Conclusions .................................................................................................... 190 References ............................................................................................................ 191 12 HAADF: High-Angle Annular Dark Field HCP: Hexagonal-Closest-Packed HRTEM: High Resolution Transmission Electron Microscopy IBE: Ion Beam Etching ICE: Ion-Conversion and Electron ICMA: Instituto de Ciencia de Materiales de Aragón ICMM: Instituto de Ciencia de Materiales de Madrid ICT: Information and Communications Technology INA: Instituto de Nanociencia de Aragón IP: In-Plane LACBED: Large Angle Convergent Beam Electron Diffraction LMA: Laboratorio de Microscopías Avanzadas LMI: Liquid Metal Ion MEMS: Micro-Electro-Mechanical Systems MFM: Magnetic Force Microscopy MOKE: Magneto-Optical Kerr Effect MRAM: Magnetoresistive Random-Access Memory MTJ: Magnetic Tunnel Junction OOMMF: Object Oriented MicroMagnetic Framework OOP: Out-Of-Plane PEEM: PhotoEmission Electron Microscopy ReRAM: Resistive Random-Access Memory SAED: Selective Area Electron Diffraction SE: Secondary Electrons SEM: Scanning Electron Microscopy S-FEG: Schottky Field Emission Gun SQUID: Superconducting Quantum Interference Device STEM: Scanning Transmission Electron Microscopy STO: Strontium Titanium Oxide TEM: Transmission Electron Microscopy TLD: Through Lens Detector TMR: Tunnel Magnetoresistance UV: Ultraviolet VLS: Vapor-Liquid-Solid VS: Vapor-Solid XMCD: X-ray Magnetic Circular Dichroism YBCO: Yttrium Barium Copper Oxide 13 Abstract Magnetic nanostructured materials attract particularly keen interest because of their possibilities to be implemented in future spintronic devices [1]. Specifically, ferromagnetic nanowires and nanotubes are potential candidates for fast and low-power domain wall conduit used for storing and handling information. In this regard, one of the most versatile and promising techniques for the fabrication of these nanostructures is Focused Electron Beam Induced Deposition (FEBID) [2]. As has been customary for more than 35 years, mainly two-dimensional (2D) deposits have been fabricated by this method. However, to meet the new market requirements for the development of more energy-efficient devices, advanced three-dimensional (3D) magnetic nano-objects emerge as extraordinarily promising structures for applications in magnetic data storage, logic and sensing. This document includes different approaches for the tuning of dimensional, compositional and magnetic properties of 3D ferromagnetic nanowires. In this context, the new ARchitectural Adjustment by Grid Overlay Nanotechnology (ARAGON) Chip represents a step forward in the fabrication of tailored functional nanowires by FEBID on insulating substrates, which had been unattainable due to the impossible charge dissipation. This consists of an electrically-biased patterned metal structure which allows the growth and the in situ modulation of the nanowire geometry. Although many efforts have been carried out to improve the properties of 2D deposits [3], scarce investigations have been performed in 3D nanostructures. In this light, the nanofabrication of 3D Co and Fe nanowires by FEBID and the subsequent ex situ and in situ post-growth annealing treatments have been explored for the first time [4][5]. High Resolution Transmission Electron Microscopy (HRTEM) imaging, Electron Energy Loss Spectroscopy (EELS) in Scanning Transmission Electron Microscopy (STEM) mode and Electron Holography (EH) have been used to monitor the structural, 14 chemical and magnetic alterations at each annealing temperature. The metallic composition increases with the temperature up to ~95% at., a recrystallization of the standard nanocrystalline as-deposited structure into large monocrystals, whose size is comparable to the nanowire diameter, is produced and the average net magnetic induction increases dramatically to values very close to the bulk one. This achievement opens new paths for the fabrication of either individual or arrays of 3D nanowires with high purity and crystallinity based on other materials, obtaining nanostructures which could be used for future applications. On the other hand, the combination of more than one material by FEBID can give rise to new functionalities. On this basis, nanoscale heterostructured materials in the form of 3D core-shell nanowires have been developed [6]. This new approach has been applied to synthesize standing nanowires with ferromagnetic cores of Co or Fe coated with a protective Pt-C shell. This architecture aims at minimizing the degradation of magnetic properties caused by the natural surface oxidation of the core to a non-ferromagnetic material. This is a key issue in such thin ferromagnetic objects with a high surface-tovolume ratio. The structure and chemistry of the nanowires have been characterized in Pt-C-coated and uncoated nanostructures, revealing that the surface oxidation is suppressed from the magnetic cores and confined to the Pt-C layer, while keeping the cylindrical shape. The magnetic characterization has demonstrated that the average magnetization of the coated cores is strengthened up to 30% in the thinnest nanowires (~40-nm-thick cores) with respect to the unprotected ones. Based on the method developed to grow 3D core-shell nanostructures by this fabrication technology, the growth and characterization of 3D ferromagnetic Co nanotubes have been performed. The heterostructured materials are composed by a 3D Pt-C nanowire acting as a core, and a Co coating forming the shell and the nanotube architecture. TEM experiments on a cross section have shown cores thinner than 100 nm and shells down to ~11 nm in thickness. The magnetic characterization performed by EH 15 and Magneto-Optical Kerr Effect (MOKE) magnetometry demonstrates their ferromagnetic behaviour, and micromagnetic simulations were carried out to understand the domain wall dynamics. These results prove that these nanostructures provide great functionality with potential application in the performance of magnetic devices. In the past, a direct use of FEBID was the growth of magnetic tips of width around 50 nm with potential application in Magnetic Force Microscopy (MFM) [7]. In order to show that FEBID tips are superior than standard MFM tips and can lead to the nextgeneration of commercial MFM tips, dedicated experiments for the growth of vertical Co and Fe nanowires have been carried out. Here, the optimization of the MFM tips grown by FEBID and a comparison of their behaviour to that of standard MFM tips have been assessed. The tips have been tested in MFM experiments, in ambient conditions as well as in liquid environment, behaving appropriately in terms of mechanical stability, resolution and sensitivity [8]. Indeed, it has been demonstrated that Fe tips with 34 nm in diameter and a 7 nm-wide sharp end can be fabricated to achieve very high resolution as well as relatively low sample-tip magnetic interaction to minimize the influence of the magnetic tip on the magnetic state of small structures such as skyrmions. The Fe content is found to decrease as the tip end is approached, which influences the magnetization value and the magnetic stray fields generated around the tip end. This is the first step towards the research on quantitative MFM measurements. Finally, taking advantage of the versatility of the FEBID technique, 3D Pt-C-coated Co nanowires with a shape-controlled structure have also been fabricated. This architecture consists of forming bends along the height (pinning sites), where domain walls can be located and obtained at remanence after saturating the magnetization with an applied magnetic field in the appropriate direction [9]. Therefore, the growth of complex architectures by FEBID opens new prospects for the development of novel magnetic devices. 16 References [1] A. Fernández-Pacheco, L. Serrano-Ramón, J. M. Michalik, M. R. Ibarra, J. M. De Teresa, L. O’Brien, D. Petit, J. Lee and R. P. Cowburn, “Three dimensional magnetic nanowires grown by focused electron-beam induced deposition”, Sci. Rep. 3, 1492 (2013). [2] J. Pablo-Navarro, D. Sanz-Hernández, C. Magén, A. Fernández-Pacheco and J. M. De Teresa, “Tuning shape, composition and magnetization of 3D cobalt nanowires grown by focused electron beam induced deposition (FEBID)”, J. Phys. D: Appl. Phys. 50, 18LT01 (2017). [3] M. V. Puydinger dos Santos, M. F. Velo, R. D. Domingos, Y. Zhang, X. Maeder, C. Guerra-Nuñez, J. P. Best, F. Béron, K. R. Pirota, S. Moshkalev, J. A. Diniz and I. Utke, “Annealing-Based Electrical Tuning of Cobalt-Carbon Deposits Grown by Focused Electron-Beam-Induced Deposition”, ACS Appl. Mater. Interfaces 8, 32496 (2016). [4] J. Pablo-Navarro, C. Magén and J. M. De Teresa, “Purified and Crystalline ThreeDimensional Electron-Beam-Induced Deposits: The Successful Case of Cobalt for High-Performance Magnetic Nanowires”, ACS Appl. Nano Mater. 1, 38 (2018). [5] J. Pablo-Navarro, R. Winkler, G. Haberfehlner, C. Magén, H. Plank and J. M. De Teresa, “In situ real time annealing of ultrathin vertical Fe nanowires grown by focused electron beam induced deposition”, Acta Materialia 174, 379 (2019). [6] J. Pablo-Navarro, C. Magén and J. M. De Teresa, “Three-dimensional core-shell ferromagnetic nanowires grown by focused electron beam induced deposition”, Nanotechnology 27, 285302 (2016). [7] M. Gavagnin, H. D. Wanzenboeck, S. Wachter, M. M. Shawrav, A. Persson, K. Gunnarsson, P. Svedlindh, M. Stöger-Pollach and E. Bertagnolli, “Free-Standing Magnetic Nanopillars for 3D Nanomagnet Logic”, ACS Appl. Mater. Interfaces 6, 20254 (2014). [8] M. Jaafar, J. M. De Teresa, A. Asenjo, J. Pablo-Navarro, P. Ares, C. Magén and J. Gómez-Herrero, “System for an Atomic Force Microscope”. Spanish OEPM P201731292 (2017) and International patent PCT/ES2018/070709 (2018). [9] A. Wartelle, J. Pablo-Navarro, M. Staňo, S. Bochmann, S. Pairis, M. Rioult, C. Thirion, R. Belkhou, J. M. De Teresa, C. Magén and O. Fruchart, “Transmission XMCD-PEEM imaging of an engineered vertical FEBID cobalt nanowire with a domain wall”, Nanotechnology 29, 045704 (2018). 17 Resumen Los materiales magnéticos nanoestructurados atraen un interés particular debido a la posibilidad de ser implementados en futuros dispositivos espintrónicos [1]. Específicamente, los nanohilos y nanotubos ferromagnéticos son candidatos potenciales para ser utilizados como vehículos excelentes para las paredes de dominio, usadas para almacenamiento y procesamiento de información. En este sentido, una de las técnicas más versátiles y prometedoras para la fabricación de estas nanoestructuras es la deposición inducida por haz de electrones focalizado (FEBID) [2]. Como viene siendo habitual desde hace más de 35 años, mediante esta técnica se han fabricado fundamentalmente depósitos en dos dimensiones (2D). Sin embargo, para dar respuesta a las exigencias del mercado actual con relación al desarrollo de dispositivos más eficientes energéticamente, los nano-objetos magnéticos avanzados en tres dimensiones (3D) se erigen como estructuras sumamente prometedoras para aplicaciones en almacenamiento, detección y lógica magnéticos. Esta tesis incluye diferentes estrategias para controlar las propiedades dimensionales, composicionales y magnéticas de nanohilos 3D ferromagnéticos. En este contexto, un nuevo método cimentado en un chip nanotecnológico basado en una cuadrícula superpuesta para modificaciones estructurales (ARAGON) representa un paso adelante en la fabricación de nanohilos funcionales diseñados por FEBID sobre sustratos aislantes, lo cual había sido inalcanzable debido a la imposibilidad de la disipación de carga eléctrica durante el crecimiento. Esta estrategia consiste en una estructura metálica sometida a una diferencia de potencial que permite la fabricación y la modulación in situ de la geometría del nanohilo. A pesar de que se han llevado a cabo muchos esfuerzos para mejorar las propiedades de los depósitos 2D [3], escasas investigaciones se han realizado con relación a las nanoestructuras 3D. Desde este punto de vista, se ha explorado por primera vez la 18 nanofabricación de nanohilos 3D de Co y Fe crecidos por FEBID y los posteriores tratamientos térmicos ex situ e in situ [4][5]. Para monitorizar los cambios estructurales, químicos y magnéticos en cada temperatura, se ha utilizado microscopía electrónica de transmisión de alta resolución (HRTEM), espectroscopía por pérdida de energía de los electrones (EELS) usando el modo de microscopía electrónica de transmisión por barrido (STEM), y holografía electrónica (EH). En particular, la composición metálica incrementa con la temperatura hasta ~95% at., se produce la recristalización de la estructura nanocristalina inicial en grandes monocristales cuyo tamaño es comparable al diámetro del nanohilo, y la inducción magnética media neta aumenta significativamente hasta valores muy próximos al del material masivo. Estos resultados abren nuevos caminos hacia la fabricación de nanohilos 3D individuales o en serie, con gran pureza y cristalinidad, basados en otro tipo de materiales, obteniendo nanoestructuras que podrían ser utilizadas para futuras aplicaciones. Por otro lado, la combinación de más de un material por FEBID puede dar lugar a nuevas funcionalidades. Por ello, se han desarrollado materiales heteroestructurados en forma de nanohilos 3D con un núcleo y un recubrimiento [6]. Esta nueva estrategia ha sido aplicada para sintetizar nanohilos verticales con núcleos ferromagnéticos de Co o Fe recubiertos de una capa protectora de Pt-C. Esta arquitectura tiene como objetivo minimizar la degradación de las propiedades magnéticas debido a la oxidación superficial natural del núcleo, convirtiendo esta capa externa en un material no ferromagnético. Esto es una cuestión clave en objetos ferromagnéticos estrechos con un cociente superficie/volumen alto. Las propiedades estructurales y químicas de los nanohilos han sido caracterizadas en nanoestructuras sin recubrir y recubiertas con Pt-C, mostrando que la oxidación superficial se suprime de los núcleos magnéticos y es confinada en la capa de Pt-C, al mismo tiempo que se conserva su forma cilíndrica. La caracterización magnética ha demostrado que la inducción magnética media de los nanohilos recubiertos 19 aumenta hasta un 30% en el caso de los nanohilos más estrechos (diámetros de ~40 nm) con respecto a los núcleos sin recubrir. Basándonos en este método para el crecimiento de nanoestructuras 3D con un núcleo y un recubrimiento mediante esta tecnología de fabricación, se ha llevado a cabo la síntesis y caracterización de nanotubos ferromagnéticos 3D de Co. En este caso, la heteroestructura está compuesto por un nanohilo 3D de Pt-C actuando como núcleo, y un revestimiento de Co formando el recubrimiento y la arquitectura del nanotubo. Los experimentos de TEM sobre una sección transversal muy delgada han mostrado núcleos con diámetros menores de 100 nm y recubrimientos que se reducen hasta ~11 nm de espesor. La caracterización magnética realizada mediante EH y magnetometría de efecto Kerr magneto-óptico (MOKE) demuestra su comportamiento ferromagnético. Asimismo, se han llevado a cabo simulaciones micromagnéticas para comprender la dinámica de las paredes de dominio. Estos resultados prueban que estas nanoestructuras ofrecen gran funcionalidad con potenciales aplicaciones en dispositivos magnéticos. En el pasado, un uso directo de la técnica FEBID consistía en el crecimiento de puntas magnéticas con diámetros de alrededor de 50 nm con posibles aplicaciones en microscopía de fuerza magnética (MFM) [7]. Para demostrar que las puntas crecidas por FEBID son superiores a las puntas MFM estándar, y que pueden dar lugar a la siguiente generación de puntas MFM comerciales, se han realizado experimentos destinados al crecimiento de nanohilos verticales de Co y Fe. Se ha evaluado la optimización de las puntas MFM crecidas por FEBID y se ha realizado una comparación de su comportamiento con respecto a las puntas MFM estándar. Las puntas han sido analizadas en experimentos MFM, tanto en condiciones ambientales como en entorno líquido, comportándose apropiadamente en términos de estabilidad mecánica, resolución y sensibilidad [8]. Se ha demostrado que las puntas de Fe de 34 nm de diámetro con un diámetro final de 7 nm en el extremo pueden ser fabricadas para conseguir tanto alta resolución como una interacción magnética muestra-punta relativamente baja que 20 minimice la influencia de la punta magnética sobre el estado magnético de estructuras pequeñas tales como skyrmiones. Además, se ha detectado que el contenido de Fe decrece cuando la distancia a la punta se reduce, lo cual tiene efectos sobre los valores de la inducción magnética y los campos de fuga generados en las proximidades de la punta. Este es el primer paso hacia la investigación destinada a medidas de MFM cuantitativo. Finalmente, aprovechando la versatilidad de la técnica FEBID, también se han fabricado nanohilos 3D de Co recubiertos con Pt-C con una morfología distinta a la de un pilar. Esta arquitectura consiste en formar curvaturas a lo largo de la longitud del nanohilo (sitios de anclaje), donde las paredes de dominio pueden estar localizadas, obteniéndose después de saturar la magnetización con un campo magnético aplicado en la dirección apropiada [9]. Por lo tanto, el crecimiento de arquitecturas 3D complejas por FEBID abre nuevas perspectivas para el desarrollo de novedosos dispositivos magnéticos. Referencias [1] A. Fernández-Pacheco, L. Serrano-Ramón, J. M. Michalik, M. R. Ibarra, J. M. De Teresa, L. O’Brien, D. Petit, J. Lee and R. P. Cowburn, “Three dimensional magnetic nanowires grown by focused electron-beam induced deposition”, Sci. Rep. 3, 1492 (2013). [2] J. Pablo-Navarro, D. Sanz-Hernández, C. Magén, A. Fernández-Pacheco and J. M. De Teresa, “Tuning shape, composition and magnetization of 3D cobalt nanowires grown by focused electron beam induced deposition (FEBID)”, J. Phys. D: Appl. Phys. 50, 18LT01 (2017). [3] M. V. Puydinger dos Santos, M. F. Velo, R. D. Domingos, Y. Zhang, X. Maeder, C. Guerra-Nuñez, J. P. Best, F. Béron, K. R. Pirota, S. Moshkalev, J. A. Diniz and I. Utke, “Annealing-Based Electrical Tuning of Cobalt-Carbon Deposits Grown by Focused Electron-Beam-Induced Deposition”, ACS Appl. Mater. Interfaces 8, 32496 (2016). [4] J. Pablo-Navarro, C. Magén and J. M. De Teresa, “Purified and Crystalline Three- 21 Dimensional Electron-Beam-Induced Deposits: The Successful Case of Cobalt for High-Performance Magnetic Nanowires”, ACS Appl. Nano Mater. 1, 38 (2018). [5] J. Pablo-Navarro, R. Winkler, G. Haberfehlner, C. Magén, H. Plank and J. M. De Teresa, “In situ real time annealing of ultrathin vertical Fe nanowires grown by focused electron beam induced deposition”, Acta Materialia 174, 379 (2019). [6] J. Pablo-Navarro, C. Magén and J. M. De Teresa, “Three-dimensional core-shell ferromagnetic nanowires grown by focused electron beam induced deposition”, Nanotechnology 27, 285302 (2016). [7] M. Gavagnin, H. D. Wanzenboeck, S. Wachter, M. M. Shawrav, A. Persson, K. Gunnarsson, P. Svedlindh, M. Stöger-Pollach and E. Bertagnolli, “Free-Standing Magnetic Nanopillars for 3D Nanomagnet Logic”, ACS Appl. Mater. Interfaces 6, 20254 (2014). [8] M. Jaafar, J. M. De Teresa, A. Asenjo, J. Pablo-Navarro, P. Ares, C. Magén and J. Gómez-Herrero, “System for an Atomic Force Microscope”. Spanish OEPM P201731292 (2017) and International patent PCT/ES2018/070709 (2018). [9] A. Wartelle, J. Pablo-Navarro, M. Staňo, S. Bochmann, S. Pairis, M. Rioult, C. Thirion, R. Belkhou, J. M. De Teresa, C. Magén and O. Fruchart, “Transmission XMCD-PEEM imaging of an engineered vertical FEBID cobalt nanowire with a domain wall”, Nanotechnology 29, 045704 (2018). Chapter 1 28 In overcome these impending limitations, the miniaturization engineering, which has been leading the enhancement of the chips performance to the present, must be driven by new strategies and paradigms [21]. One of the possibilities is the “More Moore” approach, based on better system designs using the same electronic components. The second one, founded on the incorporation of new functionalities and the expansion to 3D architectures of current semiconductor technology, is called the “More than Moore” strategy. In this case, non-digital and non-electronic data (optical, mechanical, thermal, etc.) are combined with digital information in a single device, without the need of following the traditional scale reduction pace of the digital components. As a result, these strategic plans could bring about positive advances in nanoelectronics and lead to the development of a broad diversity of applications [22]. 1.2.1 Non-volatile resistive memory One of the best examples in line with the arguments set out previously is the nonvolatile resistive memory. Within the sphere of nanoelectronics, it should be highlighted that the current flash memory technology is encountering downscaling restrictions, so research on the density and performance increase of the non-volatile memories stands as an urgent challenge to work in [23][24]. One possible alternative is the resistive random access memory (ReRAM), taking advantage of a high-speed resistive switching mechanism in metal/insulator/metal structures [25]. This system is characterized by ultra-small conducting filaments (~1-10 nm in diameter), which appear in the oxide insulating layer due to the electrical breakdown phenomenon when an electric field above a certain threshold is applied between the two metal electrodes. Thus, these filaments are responsible for modifying the transport properties of the capacitor at the nanoscale [26]. Since the work reported on NiO nanowires in 2008 [27], several experiments have been performed in order to improve the resistive switching behaviour by increasing the endurance and stability [28] Introduction 29 or reducing the operating voltage of the device [29]. In addition, novel approaches have been reported for the integration of these nanowires into memory devices such as the crossbar array methodology including metal-oxide core-shell nanowire heterostructures [30]. An impressive application of these nanowires stems from the possibility of fabricating flexible electronic circuits, which can be portable and wearable. As a result, the oxide nanowires are promising candidates to be incorporated in electronic devices for tracking human health parameters where the data could be stored [31]. Figure 1.3. Schematic image of a conventional resistive memory and a nanowire one based on a metal-oxide-metal heterostructure. A Scanning Electron Microscopy (SEM) image of a Au-NiO-Au nanowire is shown to illustrate the real system [32]. 1.3 Spintronics The manipulation of the electron spin as well as the electron charge in solid state systems is the fundamental basis of spintronics. In order to unlock and take full advantage of this potential, the understanding of the interplay between the spin and its environment becomes crucial [33]. Encompassing the knowledge originated from the combination between the spindependent transport phenomena and electronics, several applications have been developed [34]. In particular, the use of ferromagnetic materials, where the electrical resistance depends on the electron current spin polarization, have given rise to remarkable Chapter 1 30 milestones. The first great example is the concept of magnetoresistive read head magnetic recording in 1971, based on the anisotropy magnetoresistance (AMR) [35], and its commercialization some years later [36]. Additionally, the discovery of the tunnel magnetoresistance (TMR) in 1975 brought an unprecedented evolution, establishing the basis of the non-volatile magnetoresistive random-access memory (MRAM) [37]. However, the finding of the giant magnetoresistance (GMR) in 1988 by A. Fert and P. Grünberg is considered as the main achievement in this field. This accomplishment deserves a special mention because it catapulted spintronics and the production of novel read heads for hard disks drives by IBM since 1997. The GMR effect has turned into a practical and real nanoscale device for widespread applications [38][39]. These breakthroughs stimulated new perspectives on the interplay between electron transport and magnetic properties, prompting the investigation of novel effects emanating from the spin-charge currents interaction, including switching mechanisms of the magnetic moments [40], such as the spin transfer torque phenomenon. Specifically, magnetic nanowires are ideal scaffolds where many of these phenomena can take place. 1.3.1 Magnetic domain walls Magnetic domain walls can be defined as nano-objects which separate regions of different magnetization orientation, minimizing the magnetostatic energy. The balance between this energy, the magnetic anisotropy and the exchange energy determines their formation and nature. The study of these nanoscale transition areas as a tool to compute binary information in integrated magnetic circuits is a major step. Domain wall conduit can be achieved by the application of external magnetic fields or currents [41][42], and the control of this process in magnetic nanostructures, such as nanowires, attracts particularly keen interest because of their possible application in storage [43], memory [44] and logic devices [45]. Especially, the high domain wall velocities present in circular magnetic nanowires, the potential suppression of the Walker breakdown phenomenon or Introduction 31 the emergence of new types of domain walls such as curling states or Bloch points are characteristics which make them special architectures. The capability to move magnetic domain walls in magnetic tracks, e.g. nanowires, particularly by spin transfer torque using spin-polarized electrical currents, opens new outlooks considering the storage track memory as the promising candidate [46]. In this case, the information storage unit (the bit) is the domain wall, rather than the magnetic domain itself in the conventional hard disks. The fast processing time of the semiconductor integrated circuits (CMOS) is conserved with no mechanical movement of the components. In addition, the configuration of the domain wall sequence can be very versatile, presenting 2D straight, bent or loop shapes. The ultimate geometry would be 3D magnetic tracks, as introduced by S. S. P. Parkin in 2008, which would definitely boost the storage areal density to compete with existing technologies [44]. 1.3.2 Racetrack memory concept The concept of the racetrack memory, represented in Figure 1.4, is composed by a magnetic track with a sequence of domain walls and two heads devoted to writing and reading data. The write head can use different strategies to generate a domain wall, such as an Oersted line, or the magnetic tunnel junction (MTJ) architecture as the read head does [47]. In this case, an insulating barrier separates two ferromagnetic layers whose magnetization is fixed in one layer and changeable in the other one [48]. This enables two different magnetic configurations (parallel or antiparallel) with different tunnel resistance, building the logic configuration. In practice, a spin-polarized current induces the domain wall motion along the magnetic structure (shifting), the write head switches the magnetization locally, nucleating domain walls (writing) which get pinned (storing) and the read head detects the magnetization direction (reading). Chapter 1 32 Figure 1.4. (a) Composition of the magnetic tunnel junction with the two possible configurations used in the write and read heads. (b) Schematic diagram of a racetrack memory concept [47]. The domain wall displacement is typically driven by spin-transfer-torque as a result of the spin-polarized currents injected into the ferromagnet. In the last decade, wall velocities up to 150 m/s using current densities above 100 MA/cm 2 have been reached by a lateral current injection [49]. The speed was enhanced by introducing a metallic layer with strong spin-orbit coupling under the ferromagnetic track, but keeping the high current densities [50]. Recently, in order to look for low-power-consumption devices, displacement at 500 m/s under 6 MA/cm 2 have been proven by the vertical injection of spin currents [51]. This evidences that continuous improvements are still promoting the refinement of devices within the ICT field. 1.4 Further applications of magnetic nanowires The non-volatile resistive memory and the racetrack memory previously presented are two special examples that show the usefulness of nanowires in nanoelectronics and spintronics. Despite its straightforward geometry, this structure with high aspect and Introduction 33 surface-to-volume ratios has clear potential for a wide range of applications in diverse fields and, particularly in Nanomagnetism [52][53], magnetic nanowires exhibit exciting properties to be implemented in magnetic devices. Hereafter, the discussion will be focused on 3D magnetic nanowires since this specific architecture constitutes the central topic of this thesis. In this section, a general overview of their applications as building block for functional devices is given [32]. 1.4.1 Spincaloritronics The driving of the magnetic moment through spin-polarized currents has been extensively studied in magnetic nanostructures such as wires containing domain walls [54] or vertical multi-layered nanowires [55]. Going one step further, the interaction between spin currents and heat currents can also be investigated, a new research area coined “Spincaloritronics” emerging in recent years. Thermoelectric and thermomagnetic phenomena are delved into the coupling of electron and heat currents [56], demonstrating that the polarization of the spin can be governed not only by electric fields but also by temperature gradients and vice versa. Since the magnetic state of a system can be modified thermally, efforts have been devoted to looking for magnetic devices within the spincaloritronic scope. The thermal manipulation and the interaction between heat and spin currents can lead to different applications such as heat sensors, thermometers, waste heat recyclers, power generators, coolers, etc. 1.4.2 High frequency devices The propagation of electromagnetic waves and spin waves in confined geometries such as magnetic nanowires can give rise to novel microwave devices working up to the THz frequency range. The interaction between an incident electromagnetic wave and a magnetic nanostructure leads to new propagation mechanisms and diffraction phenomena. Specifically, complex architectures such as multi-layered nanowires and Chapter 1 34 nanotubes are still unexploited in this field. Therefore, a large variety of potential applications could arise in the following years in ICT or biomedical goals [57]. 1.4.3 Biomedical applications Magnetic nanowires have a great functionality in applications related to medicine, e.g., drug delivery or magnetic hyperthermia [58][59]. The morphology of these nanoobjects makes them the ideal substitutes to magnetic nanoparticles. In particular, the adjustment of their magnetic anisotropy and coercive fields by modifying the diameter, length and composition, together with the tuning of the magnetocrystalline anisotropy, plays a crucial role in the control of their magnetic properties and behaviour by the application of relatively low external magnetic fields —on the order of 1 T— at a distance. Firstly, the intrinsic magnetic features of these nanostructures allow using them as hybrid magneto-optical systems. Nanowires composed by sections with different nature along the length can be either magnetically or optically responsive to different stimuli, being a great choice to be key components in sensing devices [60]. Also, the cell trapping and separation procedures are benefitted from magnetic nanowires with high magnetic induction and aspect ratio, reducing the required magnetic fields to be used in comparison with other architectures [61]. The selectivity in the manipulation of cells in culture media has been improved as the nanowire length is correlated with the cell diameter. In this case, the anisotropic shape of the magnetic nanowires provides better output than the magnetically isotropic nanoparticles. Regarding disease treatments, the biofunctionalization of magnetic nanowires allows avoiding the damages induced by the traditional cancer therapies. The greater specificity of magnetic hyperthermia enables the death of malignant cells by locating the functionalized nanowires on the tumour and inducing heat by alternating magnetic fields in the range of kHz. In comparison with nanoparticles, Fe nanowires reduce the time Introduction 35 required in the treatments due to their shape and better magnetic performance behaviour [62]. Magnetic nanowires can also be used as magnetic nanoactuators on different biological systems. In particular, Co nanowires were used to apply very accurately controlled forces to living cells through magnetic field-induced torque [63], observing mechanical stress responses which can be very efficient in cell treatments. 1.4.4 Magnetic Force Microscopy tips The high aspect ratio, coercivity and small lateral resolution of magnetic nanowires make them perfect candidates for working as Magnetic Force Microscopy (MFM) tips. Among the several methods used for the fabrication of these nanowires on top of the standard Atomic Force Microscopy (AFM) probes [64][65], Focused Electron Beam Induced Deposition emerges as one of the most promising techniques due to the precise control of the deposition position and tilt angle with respect to the target sample [66]. Further details of this topic will be heavily discussed in Chapter 6. Figure 1.5. Diagram of a 3D Fe nanowire grown by FEBID onto an AFM probe. The inset shows an SEM image of the structure used for MFM measurements [66]. 1.4.5 Magnetoplasmonics Plasmons are collective excitations of the conduction electrons of a metal induced by incoming electromagnetic radiation. The magnetic behaviour of certain materials allows controlling the plasmonic properties of the structures by the application of external magnetic fields. This leads to the development of active magnetoplasmonic devices with applications from telecommunications to biosensing [67]. Chapter 1 36 However, the magnetoplasmonic studies are generally limited to nanoparticles and films, with scarce investigations on ferromagnetic nanowires [68]. Specifically, their combination with plasmonic materials such as Au or Pt could give rise to multifunctional objects for this unexplored field. This research line has a long road ahead with promising applications in optical biochemical sensing devices or nanophotonics [69]. 1.5 Design and fabrication of magnetic nanowires In order to ensure proper operation of the applications presented previously, advanced fabrication techniques are required. In this section, diverse approaches for the crucial nanofabrication mainstay of magnetic nanostructures will be tackled. Different bottom-up and top-down approximations devoted to building magnetic nanowires, particularly in 3D, will be reviewed for a better understanding of their advantages [70]. 1.5.1 Electrochemical synthesis The electrochemical deposition, commonly used in research and industry, is a bottom-up technology which consists of an electrolytic process where, applying a current or voltage, the metallic ions of an electrolyte are reduced, creating a solid deposit on the cathode electrode. Regarding nanowires growth, the concept entails the deposition of the material inside the pores or channels of a membrane, leading to structures with the desired architecture. Although the template method can be employed in combination with other techniques, the electrochemical deposition constitutes a commonly used approach to fill the porous membranes with magnetic materials. Depending on the type of bias applied to the electrolyte, as mention below, different electrodeposition techniques can be explored. On the one hand, it is worth mentioning the potentiostatic electrodeposition, based on a preservation of the potential in the working electrode with respect to the reference one. This type of deposition refers to one of the most common techniques for growing Fe, Co and Ni nanowires and alloys between them [71][72]. On the other hand, the galvanostatic electrodeposition relies on the application of a constant current density and Introduction 37 has been used for the fabrication of magnetic alloy nanowires and multi-layered architectures [73][74]. In comparison with the potentiostatic process, it is a more precise method in terms of the growth rate, implying a better control of the volume of deposited material by governing the deposition time. Additionally, pulse electrodeposition, rooted in the combination of potentiostatic and galvanostatic pulses followed by an intermediate recovering step, could give rise to core-shell Fe@FeO x or Ni@NiO nanowires obtained after an oxidation process once the nanostructures are out of the template [75]. Besides, as shown in Figure 1.6, multisegmented FeCo/Cu nanowires have been fabricated electrochemically by pulse electrodeposition alternating the voltage between -1.8 and -0.7 V for the FeCo and Cu segments, respectively. Moreover, modifying the pulse times, the length of each segment can be tuned [58]. The limitation to use conductive substrates for most of the electrodeposition methods can be overcome by using the alternating current electrodeposition. This strategy allows guaranteeing that the typical passivation oxide layer on top of the substrates will not be charged or polarized. In this way, metallic and semiconductor structures have been commonly fabricated with a direct contact to the substrate [76]. Among the huge number of variables which should be contemplated for a successful growth, e.g., bath pH, temperature or electrolyte composition, it should also be pointed out a correct level of wettability in the pore wall to favour the infiltration of the precursor inside the channel, or the control of the shrinkage array during the solidification of the material. This means that the widely used template-assisted method must be employed under optimized conditions, typically using ordered nanoporous anodic Al 2 O 3 templates [77][78]. To obtain the desired nanostructure geometry, the dimensions of the template (pore diameter and length, and the distance between the pores) must be tuned. All these parameters will determine not only the morphology and structural features, but also the magnetic properties of the nanowires. Chapter 1 44 1.5.5 Quenching and drawing technique The fabrication of amorphous glass-coated microand nanowires has been carried out following the common quenching and drawing method, also called glass-coated melt spinning [107]. This rapid solidification technique is used for the fabrication of structures with diameters ranging from tens of nanometres to millimetres, tailoring the lateral dimension by adjusting the temperature of the precursor alloy and the cooling process [108]. After optimization, the growth of Co 68.15 Fe 4.35 Si 12.5 B 15 and Fe 77.5 Si 7.5 B 15 nanowires with metallic core diameters between 90 and 180 nm were prepared for the first time in 2011 [109]. This technique presents some advantages with respect to other lithography and electrodeposition strategies. It is not only a cheap and straightforward method, with no limitation in terms of the nanowire length, but also allows tuning the composition, magnetization, magnetic anisotropy, switching magnetic field, domain wall motion, etc. All these possibilities should be considered to develop future logic and sensing devices. 1.5.6 Focused Electron Beam Induced Deposition Studies on fundamental properties and technological applications of magnetic nanowires require a reproducible fabrication technique which provides single and isolated objects in targeted positions. These specifications can be fulfilled by Focused Electron Beam Induced Deposition (FEBID) technique, which plays a crucial role in the design and fabrication of many types of architectures allowing great versatility in shape, composition and magnetic features [110][111][112]. Since the nanostructures presented in this thesis have been fabricated by this technology, brief remarks will be done in this subsection and comprehensive explanations will be address throughout the manuscript. The method, introduced for the first time by S. Matsui in 1984 [113], consists of a precursor gas decomposition by an electron beam, eventually producing a solid material only in the areas where the electron beam scans. This single-step nanolithography process Introduction 45 does not require the use of masks, resists or lift-off procedures. It is, however, influenced by a great number of parameters such as the electron beam voltage and current, precursor gas flux, etc. which need to be controlled and optimized. Compositional, electrical and magnetic properties have been widely investigated in 2D Co and Fe nanowires [114][115][116][117], putting forward applications such as the ones based on direct nanomagnet logic devices to compute binary information or nanosensors [116][118]. The next natural step was to move into 3D deposits, where larger versatility in terms of shape can be obtained. In addition, the specific fabrication of 3D nanowires results in higher areal density and novel domain wall configurations. This has led to the FEBID design of magnetomechanical nanoactuators [119] or MFM tips [120], and could give rise to the construction of the 3D racetrack memory. However, scarce investigations have been performed along this line and challenges still remain for this out-of-plane (OOP) architectures [121]. With this motivation in mind, this thesis is devoted to the in-depth study of the growth and characterization of 3D ferromagnetic nanowires grown by FEBID. Figure 1.12. Schematic diagram of some (a) 2D and (b) 3D geometries examples and their magnetic configurations [70]. Chapter 1 46 1.6 Outline of the thesis This thesis is focused on investigating the multitude of aspects related to the development of 3D magnetic nanowires by FEBID and their characterization through advanced techniques. It includes the exploration of the fabrication conditions and growth modes (Chapter 3), the post-growth optimization of their physical properties (Chapter 4), the formation of new architectures and heterostructures (Chapter 5), and applications (Chapter 6). Hitherto, in Chapter 1 the brief historical overview about the current and future applications of magnetic nanowires and the different methods used to fabricate these promising nano-objects have been described. This general perspective has been aimed to serve as a starting point to arouse curiosity into the reader, present a general framework of this manuscript and introduce some of the concepts which will be discuss hereafter. In Chapter 2, descriptions of the fundamental experimental techniques employed for the growth and characterization of 3D nanowires are given. Special emphasis will be placed on FEBID nanofabrication technology, and in compositional and magnetic characterization by Transmission Electron Microscopy (TEM) techniques. Chapter 3 is devoted to explaining how tuning the shape, composition and magnetization of these vertical nano-objects is possible due to the great versatility of FEBID and its combination with other strategies. Particular mention should be given to the new ARchitectural Adjustment by Grid Overlay Nanotechnology (ARAGON) Chip, allowing the fabrication of FEBID nanostructures on insulating substrates —which had been impossible until now—, and presenting an additional degree of freedom to modulate dimensional parameters of the nanostructures. Chapter 4 discusses different approaches for the improvement and optimization of structural, compositional and ferromagnetic properties of the nanowires. Specifically, in situ and ex situ post-growth annealing treatments have been explored to enhance the crystallinity, metallic content and magnetic induction of the nanostructures. Introduction 47 In Chapter 5, the implementation of the core-shell architecture for the growth 3D FEBID Co@Pt and Fe@Pt nanowires is developed. This aims at minimizing the degradation of the ferromagnetic properties of the core caused by its natural surface oxidation to a non-ferromagnetic material. Additionally, this strategy was used to fabricate ferromagnetic Pt@Co nanotubes and investigating their magnetic behaviour. Chapter 6 explores some possible applications of 3D ferromagnetic nanostructures grown by FEBID. Special attention will be paid to optimization of straight vertical nanowires for MFM measurements in different environments, and to the design of more complex 3D architectures based on hook-shaped nanowires aimed for domain wall conduit devices. Finally, Chapter 7 gives a comprehensive overview of the key results of the thesis, summarize the main conclusions and discusses some future perspectives for 3D FEBID growth in the field of Nanomagnetism. References [1] R. P. Feynman, “There’s Plenty of Room at Bottom”, Engineering and Science, 22 (1960). [2] W. Lu and C. M. Lieber, “Semiconductor nanowires”, J. Phys. D: Appl. Phys. 39, R387 (2006). [3] J. Jiu and K. Suganuma, “Metallic Nanowires and Their Applications”, IEEE CMPT 6, 1733 (2016). [4] R. Arpaia, M. Ejrnaes, L. Parlato, F. Tafuri, R. Cristiano, D. Golubev, Roman Sobolewski, T. Bauch, F. Lombardi and G. P. Pepe, “High-temperature superconducting nanowires for photon detection”, Physica C 509, 16 (2015). [5] S. Chaudhury and S. K. Sinha, “Carbon Nanotube and Nanowires for Future Semiconductor Devices Applications”, Nanoelectronics: Devices, Circuits and Systems, Edited by B. K. Kaushik (2018). [6] M. Staňo and O. Fruchart, “Magnetic Nanowires and Nanotubes”, Handbook of Magnetic Materials, Edited by E. Brück (2018). [7] V. Schuler, J. Milano, A. Coati, A. Vlad, M. Sauvage-Simkin, Y. Garreau, D. Demaille, S. Hidki, A. Novikova, E. Fonda, Y. Zheng and F. Vidal, “Growth and Chapter 1 48 magnetic properties of vertically aligned epitaxial CoNi nanowires in (Sr, Ba) TiO 3 with diameters in the 1.8-6nm range”, Nanotechnology 27, 495601 (2016). [8] K. L. Metlov and K. Y. Guslienko, “Stability of magnetic vortex in soft magnetic nano-sized circular cylinder”, J. Magn. Magn. Mater. 242-245, 1015 (2002). [9] S. Jamet, N. Rougemaille, J. C. Toussaint and O. Fruchart, “Head-to-head domain walls in one-dimensional nanostructures: an extended phase diagram ranging from strips to cylindrical wires”, Magnetic Nanoand Microwires: Design, Synthesis, Properties and Applications, Edited by M. Vázquez (2015). [10] N. Biziere, C. Gatel, R. Lassalle-Balier, M. C. Clochard, J. E. Wegrowe and E. Snoeck, “Imaging the Fine Structure of a Magnetic Domain Wall in a Ni Nanocylinder”, Nano Lett. 13, 2053 (2013). [11] S. Da Col, S. Jamet, N. Rougemaille, A. Locatelli, T. O. Mentes, B. Santos Burgos, R. Afid, M. Darques, L. Cagnon, J. C. Toussaint and O. Fruchart, “Observation of Bloch-point domain walls in cylindrical magnetic nanowires”, Phys. Rev. B 89, 180405(R) (2014). [12] A. Wartelle, B. Trapp, M. Staňo, C. Thirion, S. Bochmann, J. Bachmann, M. Foerster, L. Aballe, T. O. Menteș, A. Locatelli, A. Sala, L. Cagnon, J.-C. Toussaint and O. Fruchart, “Bloch-point-mediated topological transformations of magnetic domain walls in cylindrical nanowires”, Phys. Rev. B 99, 024433 (2019). [13] J. Escrig, J. Bachmann, J. Jing, M. Daub, D. Altbir and K. Nielsch, “Crossover between two different magnetization reversal modes in arrays of iron oxide nanotubes”, Phys. Rev. B 77, 214421 (2008). [14] Y. P. Ivanov, A. Chuvilin, S. Lopatin and J. Kosel, “Modulated Magnetic Nanowires for Controlling Domain Wall Motion: Toward 3D Magnetic Memories”, ACS Nano 10, 5326 (2016). [15] A. Ghaddar, F. Gloaguen, J. Gieraltowski and C. Tannous, “Magnetic crossover effect in Nickel nanowire arrays”, Physica B 406, 2046 (2011). [16] S. Bochmann, D. Döhler, B. Trapp, M. Staňo, O. Fruchart and J. Bachmann, “Preparation and physical properties of soft magnetic nickel-cobalt threesegmented nanowires”, J. Appl. Phys. 124, 163907 (2018). [17] A. Thiaville and Y. Nakatani, “Domain-Wall Dynamics in Nanowires and Nanostrips”, Spin Dynamics in Confined Magnetic Structures III, Edited by B. Hillebrands, A. Thiaville (2006). [18] W. Lu and C. M. Lieber, “Nanoelectronics from the bottom up”, Nat. Mater. 6, 841 (2007). [19] N. Spaldin, “Magnetic Materials. Fundamentals and Device Applications”, Introduction 49 Cambridge University Press (2003). [20] G. E. Moore, “Cramming more components onto integrated circuits”, Electronics 38, (1965). [21] R. Chau, B. Doyle, S. Datta, J. Kavalieros and K. Zhang, “Integrated nanoelectronics for the future” Nat. Mater. 6, 810 (2007). [22] European Nanoelectronics Initiative Advisory, “Strategic Research Agenda”, Second Edition (2007). [23] S. H. Park, Y. Kim, W. Kim, J. Y. Seo and B.-G. Park, “Vertical-Channel STacked ARray (VCSTAR) for 3D NAND flash memory”, Solid-State Electronics 78, 34 (2012). [24] G. I. Meijer, “Who Wins the Nonvolatile Memory Race?,” Science 319, 1625 (2008). [25] J. J. Yang, D. B. Strukov and D. R. Stewart, “Memristive devices for computing”, Nat. Nanotechnol. 8, 13 (2013). [26] Y. Yang, P. Gao, S. Gaba, T. Chang, X. Pan and W. Lu, “Observation of conducting filament growth in nanoscale resistive memories”, Nat. Commun. 3, 732 (2012). [27] S. I. Kim, J. H. Lee, Y. W. Chang, S. S. Hwank and K.-H. Yoo, “Reversible resistive switching behaviors in NiO nanowires”, Appl. Phys. Lett. 93, 033503 (2008). [28] K. Oka, T. Yanagida, K. Nagashima, T. Kawai, J.-SKim and B. H. Park, “Resistive-Switching Memory Effects of NiO Nanowire/Metal Junctions”, JACS 132, 6634 (2010). [29] E. D. Herderick, K. M. Reddy, R. N. Sample, T. I. Draskovic and N. P. Padture, “Bipolar resistive switching in individual Au-NiO-Au segmented nanowires”, Appl. Phys. Lett. 95, 203505 (2009). [30] C. Cagli, F. Nardi, B. Harteneck, Z. Tan, Y. Zhang and D. Ielmini, “ResistiveSwitching Crossbar Memory Based on Ni-NiO Core-Shell Nanowires”, Small 7, 2899 (2011). [31] D. Son, J. Lee, S. Qiao, R. Ghaffari, J. Kim, J. E. Lee, C. Song, S. J. Kim, D. J. Lee, S. W. Jun, S. Yang, M. Park, J. Shin, K. Do, M. Lee, K. Kang, C. S. Hwang, N. Lu, T. Hyeon and D.-H. Kim, “Multifunctional wearable devices for diagnosis and therapy of movement disorders”, Nat. Nanotechnol. 9, 397 (2014). [32] M. Vázquez, “Magnetic Nanoand Microwires. Design, Synthesis, Properties and Applications”, Elsevier (2015). [33] I. Žutić, J. Fabian and S. Das Sarma, “Spintronics Fundamentals: and applications”, Rev. Mod. Phys. 76, 323 (2004). Chapter 1 50 [34] S. D. Bader and S. S. P. Parkin, “Spintronics”, Annu. Rev. Condens. Matter Phys. 1, 71 (2010). [35] W. Thompson, “On the Electro-dynamic Qualities of Metals:-Effects of Magnetization on the Electric Conductivity of Nickel and of Iron”, Proc. R. Soc. Lond. 8, 546 (1857). [36] R. P. Hunt, “A Magnetoresistive Readout Transducer”, IEEE Trans. Magn. MAG-7, 150 (1971). [37] M. Julliere, “Tunneling Between Ferromagnetic Films”, Phys. Lett. 54A, 225 (1975). [38] M. N. Baibich, J. M. Broto, A. Fert, F. Nguyen Van Dau and F. Petroff, “Giant Magnetoresistance of (001)Fe/(001)Cr Magnetic Superlattices”, Phys. Rev. Lett. 61, 2472 (1988). [39] S. M. Thompson, “The discovery, development and future of GMR: The Nobel Prize 2007”, J. Phys. D: Appl. Phys. 41, 093001 (2008). [40] A. Hönemann, C. Herschbach, D. V. Fedorov, M. Gradhand and I. Mertig, “Spin and charge currents induced by the spin Hall and anomalous Hall effects upon crossing ferromagnetic/nonmagnetic interfaces,” Phys. Rev. B 99, 024420 (2019). [41] A. Fernández-Pacheco, J. M. De Teresa, R. Córdoba, M. R. Ibarra and D. Petit, “Domain wall conduit behavior in cobalt nanowires grown by focused electron beam induced deposition”, Appl. Phys. Lett. 94, 192509 (2009). [42] D. C. Ralph and M. D. Stiles, “Spin transfer torques”, J. Magn. Magn. Mater. 320, 1190 (2008). [43] M. Weigand, B. Van Waeyenberge, A. Vansteenkiste, M. Curcic, V. Sackmann, H. Stoll, T. Tyliszczak, K. Kaznatcheev, D. Bertwistle, G. Woltersdorf, C. H. Back and Gisela Schütz, “Vortex Core Switching by Coherent Excitation with Single In-Plane Magnetic Field Pulses”, Phys. Rev. Lett. 102, 077201 (2009). [44] S. S. P. Parkin, M. Hayashi and L. Thomas, “Magnetic Domain-Wall Racetrack Memory”, Science 320, 190 (2008). [45] D. A Allwood, G. Xiong, C. C. Faulkner, D. Atkinson, D. Petit and R. P. Cowburn, “Magnetic Domain-Wall Logic”, Science 309, 1688 (2005). [46] A. Brataas, A. D. Kent and H. Ohno, “Current-induced torques in magnetic materials”, Nat. Mater. 11, 372 (2012). [47] W. S. Zhao, Y. Zhang, H.-P. Trinh, J.-O. Klein, C. Chappert, R. Mantovan, A. Lamperti, R. P. Cowburn, T. Trypiniotis, M. Klaui, J. Heinen, B. Ocker, D. Ravelosona, “Magnetic Domain-Wall Racetrack Memory for high density and fast data Storage”, 2012 IEEE 11th Int. Conf. Solid-State Integr. Circuit Technol. (2012). Introduction 51 [48] W. Zhao, D. Ravelosona, J.-O. Klein and C. Chappert, “Domain Wall Shift Register-Based Reconfigurable Logic”, IEEE Trans. Magn. 47, 2966 (2011). [49] L. Heyne, J. Rhensius, D. Ilgaz, A. Bisig, U. Rüdiger, M. Kläui, L. Joly, F. Nolting, L. T. Heyderman, J. U. Thiele, F. Kronast, “Direct Determination of Large Spin-Torque Nonadiabaticity in Vortex Core Dynamics”, Phys. Rev. Lett. 105, 187203 (2010). [50] S. Emori, D. C. Bono and G. S. D. Beach, “Interfacial current-induced torques in Pt/Co/GdOx”, Appl. Phys. Lett. 101, 042405 (2012). [51] P. J. Metaxas, J. Sampaio, A. Chanthbouala, R. Matsumoto, A. Anane, A. Fert, K. A. Zvezdin, K. Yakushiji, H. Kubota, A. Fukushima, S. Yuasa, K. Nishimura, Y. Nagamine, H. Maehara, K. Tsunekawa, V. Cros and J. Grollier, “High domain wall velocities via spin transfer torque using vertical current injection”, Sci. Rep. 3, 1829 (2013). [52] M. Melzer, M. Kaltenbrunner, D. Makarov, D. Karnaushenko, D. Karnaushenko, T. Sekitani, T. Someya and O. G. Schmidt, “Imperceptible magnetoelectronics”, Nat. Commun. 6, 6080 (2015). [53] D. Makarov, M. Melzer, D. Karnaushenko and O. G. Schmidt, “Shapeable magnetoelectronics”, Appl. Phys. Rev. 3, 011101 (2016). [54] G. Tatara and H. Kohno, “Theory of Current-Driven Domain Wall Motion: Spin Transfer versus Momentum Transfer”, Phys. Rev. Lett. 192, 086601 (2004). [55] M. D. Stiles and A. Zangwill, “Anatomy of spin-transfer torque”, Phys. Rev. B 66, 014407 (2002). [56] G. E. W. Bauer, E. Saitoh and B. J. van Wees, “Spin caloritronics”, Nat. Mater. 11, 391 (2012). [57] H. Yang, M. Zeng and R. Yu, “Magnetic properties of the Ni x Co 1-x /Cu multilayer nanowires”, Materials Research Bulletin, 57, 249 (2014). [58] B. Özkale, N. Shamsudhin, G. Chatzipirpiridis, M. Hoop, F. Gramm, X. Chen, X. Martí, J. Sort, E. Pellicer and S. Pané, “Multisegmented FeCo/Cu Nanowires: Electrosynthesis, Characterization, and Magnetic Control of Biomolecule Desorption”, ACS Appl. Mater. Interfaces 7, 7389 (2015). [59] X.-F. Han, S. Shamaila and R. Sharif, “Ferromagnetic Nanowires and Nanotubes”, Electrodeposited Nanowires Their Applications, Edited by N. Lupu (2010). [60] B. He, T. J. Morrow and C. D. Keating, “Nanowire sensors for multiplexed detection of biomolecules”, Curr. Opin. Chem. Biol. 12, 522 (2008). [61] A. Hultgren, M. Tanase, E. J. Felton, K. Bhadriraju, A. K. Salem, C. S. Chen and D. H. Reich, “Optimization of Yield in Magnetic Cell Separations Using Nickel Chapter 1 52 Nanowires of Different Lengths”, Biotechnol. Prog. 21, 509 (2005). [62] W.-S. Lin, H.-M. Lin, H.-H. Chen, Y.-K. Hwu and Y.-JChiou, “Shape Effects of Iron Nanowires on Hyperthermia Treatment”, Journal of Nanomaterials 2013, 237439 (2013). [63] N. J. Sniadecki, A. Anguelouch, M. T. Yang, C. M. Lamb, Z. Liu, S. B. Kirschner, Y. Liu, D. H. Reich and C. S. Chen, “Magnetic microposts as an approach to apply forces to living cells”, PNAS 104, 14553 (2007). [64] G. Yang, J. Tang, S. Kato, Q. Zhang, L. C. Qin, M. Woodson, J. Liu, J. W. Kim, P. T. Littlehei, C. Park and O. Zhou, “Magnetic nanowire based high resolution magnetic force microscope probes”, Appl. Phys. Lett. 87, 123507 (2005). [65] S. Alotaibi, J. Samba, S. Pokharel, Y. Lan, K. Uradu, A. Afolabi, I. Unlu, G. Basnet, K. Aslan, B. N. Flanders, A. Lisfi and B. Ozturk “Individually grown cobalt nanowires as magnetic force microscopy probes”, Appl. Phys. Lett. 102, 092401 (2018). [66] M. Gavagnin, H. D. Wanzenboeck, S. Wachter, M. M. Shawrav, A. Persson, K. Gunnarsson, P. Svedlindh, M. Stöger-Pollach and E. Bertagnolli, “Free-Standing Magnetic Nanopillars for 3D Nanomagnet Logic” ACS Appl. Mater. Interfaces 6, 20254 (2014). [67] G. Armelles, A. Cebollada, A. García-Martín and M. Ujué González, “Magnetoplasmonics: Combining Magnetic and Plasmonic Functionalities”, Adv. Optical Mater. 1, 10 (2013). [68] J. B. González-Díaz, A. García-Martín, G. Amelles, D. Navas, M. Vázquez, K. Nielsch, R. B. Wehrspohn and U. Gösele, “Enhanced Magneto-Optics and Size Effects in Ferromagnetic Nanowire Arrays”, Adv. Mater. 19, 2643 (2007). [69] V. Bonanni, S. Bonetti, T. Pakizeh, Z. Pirzadeh, J. Chen, J. Nogués, P. Vavassori, R. Hillenbrand, J. Åkerman and A. Dmitriev, “Designer Magnetoplasmonics with Nickel Nanoferromagnets”, Nano. Lett. 11, 5333 (2011). [70] A. Fernández-Pacheco, R. Streubel, O. Fruchart, R. Hertel, P. Fischer and R. P. Cowburn, “Three-dimensional nanomagnetism”, Nat. Commun. 8, 15756 (2017). [71] V. Vega, J. García, W. O. Rosa, L. G. Vivas, V. M. Prida, B. Hernando, M. Vázquez, “Magnetic Properties of (Fe, Co)-Pd Nanowire Arrays”, J. Nanosci. Nanotechnol. 12, 7501 (2012). [72] X. Liu, X. Zhao, Z. Gao, Y. Zhu, J. Wang, Z. Dai and G. Wu, “Structurecontrolled Synthesis and Magnetism of CoCu Nanowires”, Chem. Lett. 42, 1108 (2013). [73] V. Vega, W. O. Rosa, J. García, T. Sánchez, J. D. Santos, F. Berón, K. R. Pirota, V. M. Prida and B. Hernando, “Template-assisted CoPd nanowire arrays: Introduction 53 magnetic properties and FORC analysis”, J. Nanosci. Nanotechnol. 12, 4736 (2012). [74] V. Weihnacht, L. Péter, J. Tóth, J. Pádár, Zs. Kerner, C. M. Schneider and I. Bakonyi, “Giant Magnetoresistance in Co-Cu/Cu Multilayers Prepared by Various Electrodeposition Control Modes”, J. Electrochem. Soc. 150, C507 (2003). [75] J. H. Lee, J. H. Wu, J. S. Lee, K. S. Jeon, H. R. Kim, J. H. Lee, Y. D. Suh and Y. K. Kim, “Synthesis and Characterization of Fe - FeO x Core-Shell Nanowires”, IEEE Trans. Magn. 44, 3950 (2008). [76] P. Wang, L. Gao, Z. Qiu, X. Song, L. Wang, S. Yang, R. Murakami, “A multistep ac electrodeposition method to prepare Co nanowires with high coercivity”, J. Appl. Phys. 104, 064304 (2008). [77] M. Vázquez, K. Pirota, J. Torrejón, D. Navas and M. Hernández-Vélez, “Magnetic behaviour of densely packed hexagonal arrays of Ni nanowires: Influence of geometric characteristics”, J. Magn. Magn. Mater. 294, 174 (2005). [78] J. Klammer, J. Bachmann, W. Töllner, D. Bourgault, L. Cagnon, and U. Gösele and K. Nielsch, “Electrochemical route to thermoelectric nanowires via organic electrolytes”, Phys. Status Solidi B 247, 1384 (2010). [79] A. Cortés, R. Lavín, J. C. Denardin, R. E. Marotti, E. A. Dalchiele, P. Valdivia, H. Gómez, “Template Assisted Electrochemical Growth of Cobalt Nanowires : Influence of Deposition Conditions on Structural , Optical and Magnetic Properties”, J. Nanosci. Nanotechnol. 11, 3899 (2011). [80] X. Zianni, “Diameter-modulated nanowires as candidates for high thermoelectric energy conversion efficiency”, Appl. Phys. Lett. 97, 233106 (2010). [81] K. Schwirn, W. Lee, R. Hillebrand, M. Steinhart, K. Nielsch and U. Gösele, “SelfOrdered Anodic Aluminum Oxide Formed by H 2 SO 4 Hard Anodization”, ACS Nano 2, 302 (2008). [82] W. Lee, J.-C. Kim and U. Gösele, “Spontaneous Current Oscillations during Hard Anodization of Aluminum under Potentiostatic Conditions”, Adv. Funct. Mater. 20, 21 (2010). [83] L. Yi, L. Zhiyuan, C. Shuoshuo, H. Xing and H. Xinhua, “Novel AAO films and hollow nanostructures fabricated by ultra-high voltage hard anodization”, Chem. Commun. 46, 309 (2010). [84] W. Lee, K. Schwirn, M. Steinhart, E. Pippel, R. Scholz and U. Gösele, “Structural engineering of nanoporous anodic aluminium oxide by pulse anodization of aluminium”, Nat. Nanotechnol. 3, 234 (2008). [85] W. Lee, R. Ji, U. Gösele and K. Nielsch, “Fast fabrication of long-range ordered Chapter 2 60 Figure 2.1. Schematic diagram of a Dual Beam SEM-FIB system. The piezoelectric platform can be moved in the three spatial directions, rotated 360 degrees and tilted at any angle between -10 and 60 degrees. The vacuum system is comprised by four ion getter pumps working on the columns, and rotatory and turbomolecular pumps acting on the working chamber, reaching pressures below ~10 -7 mbar in the columns (~10 -10 mbar in the gun compartments) and around 1 × 10 -6 mbar in the working chamber. The image acquisition relies on the fundamental basis of a scanning microscope. The primary beam is scanned by the scan coils over a surface area of the sample following a desired pattern called raster, usually a square pattern defined by the scan unit. Normally, this scan entails a series of lines in the horizontal direction of a plane, slightly shifted by the scan coils from one another in the vertical direction using the scan generator. The interaction between the primary beam and the sample generates particles and radiation Experimental techniques 61 which can be collected by the different detectors while the beam scans. After data processing, the final image is displayed in the computer screen. It is useful to place the specimen in the eucentric height, enabling simultaneous imaging using electrons and ions when tilting the stage. The SEM column is generally devoted to imaging. The electron-sample interaction broadly generates secondary electrons (SE), backscattered electrons (BSE), characteristic X-rays, Auger electron and cathodoluminescence (visible light). In our case, SE and BSE are the most used particles for imaging operation. The SE are generated by the inelastic scattering of the primary beam with the valence electrons of the outer shells of the sample atoms. They are emitted with energies lower than ~50 eV, having mean free paths below 50 nm. Thus, since only the surface ones can escape and be detected, they are used to obtain images of topographic contrast. On the other hand, the BSE are generated by the quasi-elastic scattering of the primary beam with the nuclei of the atoms. Thus, the energy lost by these electrons is very little, being the momentum transfer large, allowing the electrons to be scattered at high angles and frequently backscattered. As BSE are highenergy electrons, they also come from the deeper regions of the sample and do not give accurate morphological information. However, they provide images with compositional contrast: the scattering cross section of the BSE increases with the atomic number, ; therefore, the areas containing heavy elements are brighter than that of the lighter ones in the SEM images. Figure 2.2. Basic diagram of the events caused by the beam-sample interaction. Chapter 2 62 In the case of FIB, the use is not limited to imaging. The ion-sample interaction generates SE, secondary ions and ion implantation, among others. When the kinetic energy of the ion beam exceeds the binding energy of the target material, the physical sputtering of the atoms occurs, thus the milling of the material surface. These events can also induce amorphization and volatilization of the irradiated region of the specimen [24]. The detectors installed in the system are usually the Everhart Thornley detector (ETD), the Through Lens detector (TLD) and the Ion-Conversion and Electron detector (ICE). The ETD is a scintillator photo-multiplier detector which collects SE, BSE and secondary ions. The TLD is mainly used for high resolution imaging and collects both SE and BSE. The ICE is a charged particle detector which collects secondary ions, SE and BSE. In our case, mainly ETD and TLD, and in lesser extent ICE, have been used to acquire images collecting SE. In addition, three working modes can be selected to tune the final performance: the field-free mode, the immersion mode and the EnergyDispersive X-ray Spectroscopy (EDS) mode. The field-free mode generally assists navigation at low magnifications and can be used with ETD and ICE detectors. Moreover, an infrared charge-coupled device (CCD) camera allows the inner part of the working chamber to be observed and is used for spatial orientation of the sample. The immersion mode is used for ultra-high-resolution imaging, switching on the immersion lens (applying a magnetic field to collect more electrons) and normally collecting the SE with the TLD detector. Finally, the EDS mode is suitable for X-ray spectroscopy, reducing the power of the immersion lens and improving the X-ray signal. For this purpose, an EDS detector is mounted to collect X-ray photons coming from the sample. Furthermore, the equipment is fitted with a gas injector system (GIS) formed by a set of injectors containing precursor material, which can be delivered locally on the surface substrate thanks to a stainless-steel needle and a very narrow nozzle. A heater regulates the temperature inside the GIS to the tenth of a degree, so an optimum value can be achieved, high enough to sublimate the precursor while preventing its thermal Experimental techniques 63 decomposition inside the crucible. If a constant temperature is reached, a specific steady pressure is obtained. This implies that, if the vacuum in the working chamber is kept constant, the gas flux just depends on the saturated vapor pressure of the precursor and on the resistance due to the needle wall. The precursors used in this thesis and their operation temperatures are: dicobalt octacarbonyl, Co 2 (CO) 8 , at ~27 ºC; diiron nonacarbonyl, Fe 2 (CO) 9 , at ~28 ºC; trimethyl methylcyclopentadienyl platinum, CH 3 C p Pt(CH 3 ) 3 at ~45 ºC; and tungsten hexacarbonyl, W(CO) 6 , at ~55 ºC. The system is also equipped with a nanomanipulator (Omniprobe ® ), used for precise sample mechanical operations in specific procedures such as lamella preparation. Finally, an anti-vibration system to minimize mechanical instabilities is placed in the ground. Figure 2.3. Image of the main components of a Dual Beam system: (A) SEM column, (B) FIB column, (C) process chamber, (D) gas injector system, (E) nanomanipulator, (F) TLD detector, (G) ETD detector, (H) ICE detector, (I) CCD camera and (J) software of the system. Chapter 2 64 Three commercial Dual Beam systems have been used: two FEI Helios NanoLab 600 and 650, installed in the Class 10000 clean room of the Laboratorio de Microscopías Avanzadas (LMA) located in the Instituto de Nanociencia de Aragón (INA) at the Universidad de Zaragoza, and the FEI Nova 200 NanoLab of the Institut für Elektronenmikroskopie und Nanoanalytik - Zentrum für Elektronenmikroskopie (FELMI-ZFE). Even though these instruments present some minor differences, the work is performed similarly in all of them. 2.1.1.1 Focused Electron Beam Induced Deposition FEBID is a single-step nanolithography technique based on the delivery of precursor gas molecules close to the substrate, subsequently adsorbed on the surface, and eventually dissociated by a finely-focused electron beam producing a deposit of solid material [25][26][27][28]. Before the consolidation of FEBID, the potential of electron beam induced processing for patterning was hinted in the 1970s by the local deposition of contaminants in SEM [29]. When the SEM scans a region, a layer of a few nanometres starts to cover the scanned area. This phenomenon is produced by the decomposition of hydrocarbons present in the vacuum chamber as residual gases and adsorbed on the scanned surface. The technique was formally introduced in 1984 by S. Matsui [30][31], injecting different precursor gas molecules on purpose inside the vacuum chamber producing nanostructures with distinct functionalities. The last 30 years have witnessed the increasing interest of the scientific community and industry in the development of FEBID. Nowadays, this method is capable of nanometre-scale resolution for the growth of 2D [32] and 3D [33] structures. This unique capability of FEBID has been exploited in a broad range of applications, such as integrated circuit edit and mask repair [34], creation of electrical contacts [35], growth of magnetic nanowires [36], fabrication of plasmonic nanostructures [37], photodetection [38], gas sensing [39], etc.; thus becoming a key lithographic technique in Nanotechnology [40] and Materials Science [41]. Experimental techniques 65 Once the precursor gas molecules are injected locally near the substrate surface, some of them are dissociated by the electron beam. The non-volatile part of the gas is deposited whereas the volatile one is pumped out of the working chamber. The shape of the deposit is defined by the electron beam scan as well as the complex interactions between electron beam, substrate, precursor molecules and the growing structure [42][43]. For instance, the diffusion, adsorption and desorption phenomena of the molecules over the substrate or the probability of an electron to break the molecule bonds, mainly related to the electron energy, are crucial ingredients to understand the growth processes and determine how the fabrication process takes place. The FEBID is also governed by many different parameters which must be controlled to obtain the desired nanostructures: the electron beam voltage, electron beam current, dwell time, refresh time, overlap, pitch, scan direction, pattern dimensions, precursor gas flux, etc. These variables will be discussed in Chapter 3, giving some details about their main functions and capabilities. Figure 2.4. Diagram of the FEBID process using Co 2 (CO) 8 as gas precursor. Chapter 2 66 2.1.2 Optical lithography The optical lithography is based on the fabrication of microstructures by transferring a pattern into a wafer using masks and photosensitive materials which are subsequently developed with ultraviolet (UV) light. The process consists of several steps in sequence to obtain the final system. In the basic process, a photosensitive chemical photoresist is spread all over the sample, already grown on top of the substrate, by spin coating. The photoresist is a viscous fluid usually composed by a polymer, a photosensitive component and a solvent. The polymer supplies the viscosity, adherence and resilience required for the ulterior chemical etching; the photosensitive component makes the photoresist sensitive to the UV radiation; and the solvent allows the polymer to be in solution and can be subsequently eliminated by soft-baking. Two different types of photoresists can be used, depending on its solubility upon UV irradiation: positive and negative photoresists. In a positive photoresist, the UV light breaks the polymer chains due to chemical reactions and the photosensitive complex increases its solubility, whereas the non-irradiated areas remain insoluble in the developer. The process is quite the contrary for a negative photoresist: it is soluble in the developer, while UV irradiation induces crosslinking of the polymer chains that reduces its solubility. The desired pattern is transferred into the photoresist using a photomask, typically made of quartz with the motifs imprinted in chromium. Quartz is transparent to the UV light and the chromium absorbs this type of light. After the UV exposure, the resist becomes sensitized, the sample is immersed in the developing fluid and the soluble areas are removed. Then, dry or wet etching is performed to eliminate the film areas unprotected by the resist, and so the pattern is transferred to the sample. Finally, the remaining non-sensitized resist preserved on top of the remaining sample is removed by using acetone. Experimental techniques 67 In the lift-off procedure, the growth of the film is carried out only after the motifs have been transferred onto a layer of resist on the substrate. Then, the sample is fabricated covering the substrate unprotected regions and resist areas, ensuring that the thickness of the resist is at least 50% higher than that of the sample. This guarantees that the underneath resist will be removed together with the sample deposited on top, keeping the areas where the sample is in direct contact with the substrate. If a positive mask is used, the sample pattern is supplementary to that of the mask and obtained without the etching step, in contrast to the basic optical lithography process. The system used for optical lithography is the high-precision mask aligner SUSS MicroTec MA6 equipped with a Hg lamp exposure source, achieving a resolution down to ∼2 µm. It is installed in the Class 100 clean room of the LMA-INA at the Universidad of Zaragoza. Besides, a spin coater and a hot plate SUSS MicroTec Delta 20T/200 are available in the same room to produce homogenous photoresist coatings under a speed up to 10 4 rpm and heat it up to 250 ºC. Furthermore, thin film grown in the lift-off process is performed in an electron beam evaporator (E-beam PVD) Edwards 500 installed at the Class 10000 clean room. Thicknesses from 1 nm to 500 nm of metallic materials can be deposited with a resolution of 0.1 nm using a quartz balance for calibration. The base pressure is ~2 × 10 -7 mbar, having four different targets to deposit material. When dry etching is required, the removal of sample material is performed by physical processes using Ion Beam Etching (IBE). In this case, a beam of inert ions sputters the sample surface eliminating material in a uniform and homogeneous way. This has been carried out in the IBE SISTEC 600 equipment, fitted with argon gas, working at 2 × 10 -7 mbar and equipped with a radiofrequency generator at 13.56 Hz and 600 W. Chapter 2 68 Figure 2.5. Images of (a) the mask aligner SUSS MicroTec MA6, and (b) the spin coater and (c) hot plate SUSS MicroTec Delta 20T/200 equipment for optical lithography. Figure 2.6. Images of (a) the Ion Beam Etching and Milling SISTEC 600 and (b) the electron-beam evaporator (E-beam PVD) Edwards 500. 2.2 Transmission Electron Microscopy TEM is a nanocharacterization technique based on the formation of an image with the electrons transmitted through a thin specimen irradiated with a high-energy electron beam, of typically 80-300 keV. The spatial resolution of an optical system is limited by the radiation wavelength, . Even though residual aberrations introduced by the electron optics of the microscope, as Experimental techniques 69 well as mechanical, electronic and thermal instabilities, degrades the resolution power, modern microscopes routinely provide atomic resolution, ∼2 Å, and last-generation aberration corrected microscopes improve this value down to the sub-Å range [44]. When the primary electron-specimen interaction takes place, multiple events are originated, such as SE, BSE, characteristic X-rays, Auger electron, cathodoluminescence (visible light), inelastic scattering, elastic scattering, thermal diffuse scattering and Bremsstrahlung X-rays [45]. The microscope is divided in three main parts: the electron gun; the column, which includes the illumination system, the objective lens and the imaging system; and, the camera section. The electron source, located in the upper part of the machine, is composed by an emitter with an electrostatic lens (e.g., Wehnelt electrode in a thermionic gun or gun lens in a FEG). Then, the illumination system, placed in the column, starts to define the way the electrons will irradiate the specimen. It is composed by a set of condenser lenses (usually two or three) and condenser apertures devoted to defining the beam current, size and convergent angle. The microscope can be configurated to provide a broad beam illumination of the specimen, often known as TEM mode, or a convergent beam to form a small (sub-nm) probe. The latter is used in Scanning Transmission Electron Microscopy (STEM) mode, where this convergent probe is scanned over the specimen in a similar fashion as in SEM. To optimize and provide versatility to these types of illumination, a condenser mini lens located before the objective lens is strong excited (TEM mode) or weak excited (STEM mode). Next to the region where the EDS detector are placed, the objective lens is devoted to forming the first image of the sample, thus being determinant for the ultimate resolution. It is disposed after the illumination system and is typically composed by two lenses called objective condenser lens and objective imaging lens, conforming the upper and lower polepieces, respectively. These twin lenses are symmetrically placed, and their Chapter 2 76 electrons of a specific energy loss. As a result, a spectrum is formed by representing the intensity and the energy loss. Many parameters can be controlled such as the spectrometer energy resolution, entrance aperture, collection angle, objective aperture diameter, camera length, etc. Further details and specific values will be given during the next chapters as the experiments are presented. Three different spectrometers have been used: a Tridiem 863 Gatan Energy Filter (GIF) installed in the FEI Tecnai F30 TEM; a Tridiem 866 ERS GIF in the FEI Titan Low Base 60-300; and a GIF Quantum in the FEI Titan Cube G2 60-300 [52]. 2.2.1.2 Energy-Dispersive X-ray Spectroscopy The EDS system is based on the analysis of X-ray photons coming from the sample. This process comes about when the electron beam excites an electron of the inner shell of the atom, generating a hole. Subsequently, electrons from a higher-energy shell occupies the created hole, producing X-rays in cascade with discrete, characteristic energies corresponding to the difference between energy levels of the atom [53]. Since the X-rays have specific energies which depends on the energy difference between two atomic levels and each chemical element has a unique atomic structure, EDS technique provides information about the chemical composition of the sample. In the EDS spectra, where the number of counts as a function of the energy is represented, each peak is associated with one electronic transition of a single element [54]. The EDS system is composed by the detector, the processing electronics and the computer. Firstly, the detector generates a charge pulse which is proportional to the photon energy, converting it to a voltage signal before amplifying it by means of a fieldeffect transistor. Then, the pulse is identified electronically, and a digital signal is stocked in the corresponding channel allocated for that specific energy displaying the spectrum. In this thesis, three different EDS systems have been used at the INA: the APOLLO X detector associated with the EDAX software and mounted in the FEI Helios NanoLab Experimental techniques 77 650; the EDAX 136-5 detector coupled with the Genesis RTEM software embedded in FEI’s TIA software installed in the FEI Tecnai F30 TEM; and, the Oxford INCA 200 EDS setup from Oxford Instruments set in the FEI Helios NanoLab 600. 2.2.1.3 Off-Axis Electron Holography Off-Axis EH is an interferometric technique which measures the amplitude and the phase shift of the electron wave transmitted through the specimen. This phase shift is directly related to the electromagnetic fields produced by the sample, which can be determined in a quantitative way. The behaviour of a relativistic electron wave, , in an electromagnetic field is described by the Dirac equation: 1 2   =  −  ℏ ∇ +           =    ∗ +        (2.1) where   is the rest mass of the electron, ℏ the reduced Planck constant,  the electron charge,  the magnetic potential,  the electric potential, =1+ ∗     ⁄ the relativistic Lorentz factor and  ∗ the relativistic accelerating potential. The solution of the equation is the object wavefunction whose phase shift, !,#, is altered by the Aharonov-Bohn effect [55]. Particularly, the phase shift of an electron wave advancing along the $ axis and passing through a magnetic specimen with neutral charge can be expressed mathematically as:  ! , #  = %   ∗ &  '()  ! , # , $  *$ −  ℏ + , -  ! , # , $  *!*$ (2.2) where  is the electron relativistic wavelength,  '() the mean inner potential and , - the magnetic induction vector orthogonal to the unitary vectors along ! and $ axes [56]. At this point, it should be introduced the object electron wave,  ./0 , defined as the transmitted electron wavefunction in the exit surface of the specimen. If a very thin Chapter 2 78 specimen is considered, the electrons are elastically scattered and the absorption effects can be neglected, giving rise to the following equation:  . /0     = 1     exp        (2.3) where 1 is the amplitude of the exit wavefunction and the phase shift induced by the electric and magnetic potentials of the specimen. Then, the objective lens forms a diffraction pattern and an image in the back focal plane and the image plane, respectively. In this process, the objective lens introduces aberrations to the object electron wave. Thus, the object electron wave in the back focal plane,  5677 , can be expressed as a function of the frequency 8 9  :  5677  8 9   =  ./0  8 9   :  8 9   (2.4) where :8 9   is the phase contrast transfer function, a mathematical expression of the modifications of the electron wave caused by the instrument (not by the specimen): :  8 9   =   8 9   exp ; − <  8 9   = exp ; >  8 9   = (2.5) where 8 9   is the aperture-related frequency cutoff, exp;−<8 9  = the damping introduced by microscope instabilities and >8 9   the phase shift introduced by the aberrations of the objective lens. In a first approximation, if low order aberrations (such as stigmatism, coma) are suppressed: >  8 9   = 2 %  ? Δ A 2   8  + B C 4  E 8 E F (2.6) where Δ A is the defocus and B C is the spherical aberration coefficient. As a result, this phase shift is transferred to the image, ultimately establishing the bases of the phase contrast imaging. Nonetheless, since the image should be in focus, and the magnetic induction emerged from the specimen induces small angular deflections (8≪), >8 9   Experimental techniques 79 can be neglected. In this case, the image electron wave,  6HIJ , and the intensity of the image, K, are:  6HIJ  8 9   = L: M N ;  ./0  8 9   :  8 9   = ≅ 1     exp        =  6HIJ     (2.7) K     =  6HIJ      6HIJ ∗     = | 1     |  (2.8) The image only contains information about amplitude variations (which are very small in thin specimens), and the phase information, dependant of the electric and magnetic fields, is lost. To overcome this issue, EH enables the retrieval of the amplitude and the phase of the electron wave separately [57][58]. Experimentally, EH is based on the interference between two different electron waves; a reference electron wave,  Q7 , which propagates through the vacuum, ideally interacting with no electromagnetic field; and the object electron wave,  ./0 , transmitted through the sample, which experiences a phase shift by interacting with any electromagnetic field inside and around the specimen. The overlapping and interference of  Q7 and  ./0 is originated by the deflection of the electrons when a voltage is applied to the electrostatic Möllenstedt biprism, generating the holographic fringes. The BF image of the specimen, overlapped with the fringe pattern, is the so-called electron hologram [59][60]. From the theoretical point of view, in EH  Q7 and  ./0 are considered plane waves:  Q7     = exp ;  2 % 8 9  ⋅   = (2.9)  ./0     = 1     exp ;  2 % 8 9  ⋅   +      = (2.10) The two waves are deflected by the biprism S 2 ⁄ and −S 2 ⁄ angles along ! axis, respectively. Chapter 2 80 Figure 2.8. Basic scheme of the EH technique, reproduced from [61] with permission. This leads to the  6TU electron wave along ! axis in the superimposing:  6TU     = 1     exp V − % S  ! +      W + exp V % S  ! W (2.11) whose intensity K is: K     = |  6TU     |  = 1 + 1      + 2 1     cos V 2 % S  ! −     W (2.12) where 1+1   contains the intensity of the BF image, and the sinusoidal term represents the interference fringes pattern of the hologram with its argument depending only on the phase of the object electron wave. The calculation of the Fourier Transform (FT) of the intensity leads to: L: | K     | = [  \  + L:  1      + L:  1     exp         ⨂ [ ^ \ 9  + S  ! _ ` + (2.13) + L:  1     exp  −        ⨂ [ ^ \ 9  − S  ! _ ` Experimental techniques 81 where ⨂ indicates the convolution operation. In the reciprocal space, the hologram is formed by three different components: a central band and two sidebands. The first two terms of the right-hand side of the Equation 2.13 represent the central band and contains the BF image related to the elastic and inelastic scattered electrons with the amplitude of the object electron wave. These do not store any detail about the phase, so they have not further interest to obtain magnetic information. On the other hand, each sideband contains redundant information about the amplitude and phase separately; hence, data processing is only done with one of them. The inverse Fourier Transform (FT -1 ) of a centred sideband allows the reconstruction of the object electron wave as the combination of an amplitude image and a phase image: L: M N ; L:  1     exp         ⨂ [  \ 9   = = 1     exp        =  ./0     (2.14) where:     = arctan e Im ;  ./0     = Re ;  ./0     = i (2.15) 1     = V ; Im ;  ./0     ./0 = =  + ; Re ;  ./0     ./0 = =  W N  j (2.16) So far, this process can be applied to any magnetic specimen. However, in order to continue with the process devoted to separating the electrostatic and magnetic contributions, it is illustrative to consider a specific example. Please, refer to Annex A where the data processing is applied for ferromagnetic nanowires. Two different TEM microscopes have been used for the EH experiments: the commercial FEI Titan Cube 60-300 at the INA in Zaragoza and the Hitachi I2TEM at the CEMES-CNRS in Toulouse. Chapter 2 82 2.3 Annealing techniques The fabrication of nanostructures by FEBID technique has some drawbacks which can be overcome by annealing experiments. The metallic purity, the crystallinity and the magnetization of the deposits is susceptible of being improved by this strategy. Two different annealing setups have been used: a heating stage mounted inside the SEM Quanta FEG 250 installed at the INA in Zaragoza and a heating chip Wildfire S3 from DENSsolutions in the FEI Titan Cube G2 60-300 at the FELMI-ZFE in Graz. The first one, devoted to ex situ experiments, is formed by a heating stage base which holds all the components [62]. It includes two disk-shaped insulating elements made of aluminous foam and a heater placed between the insulators consisting of a micro-furnace in which samples are heated from the sides, ensuring temperature uniformly distributed. It is also equipped with a thermocouple, ceramic connectors, a graphite crucible to mount the samples, two ceramics papers on top to reduce heat losses and protect the insulators from damage, a cover plate to fix the components and a heat shield to keep the temperature homogeneous in the sample. In addition, it is fitted with a chamber feedthrough plate, a water chiller, a flow box, water hoses and a microprocessor-controller, devoted to providing the desired temperature in the heating stage. The sensor accuracy is ±1 ºC with an operation range from room temperature to ∼1000 ºC with a maximum heating ramp of 50 ºC/min. The second one is the TEM-heating system Wildfire S3 for in situ experiments. The specimen is placed in a heating chip which includes micro-electro-mechanical systems (MEMS) [63]. The substrate includes oval holes where the suspended nanostructures can be imaged. The chip is specially developed for having small drift (< 1 nm/min at 800 ºC), heating rates of 200 ºC/ms and image resolution of ~0.6 Å at 800 ºC. Then it is mounted into the heating stage using a lock and four pin connections to control the temperature. Experimental techniques 83 Figure 2.9. Images of (a) the heating stage mounted in the SEM Quanta FEG 250, and (b) the Wildfire S3 in situ TEM heating holder from DENSsolutions. 2.4 Further magnetic characterization techniques In this section, three different experimental methods employed for the magnetic characterization of ferromagnetic nanostructures are described. The physical principles of Superconducting QUantum Interference Device (SQUID) magnetometry, MagnetoOptical Kerr Effect (MOKE) magnetometry and MFM will be sketched. 2.4.1 Superconducting Quantum Interference Device magnetometry The SQUID is used for measuring changes in the magnetic field related to the magnetic flux quantization. It is considered one of the most sensitive magnetic flux detectors, achieving a field resolution of 10 -17 T [64]. The principle which underpins the operation of a SQUID is the quantization of the magnetic flux [65][66]: Φ l = ℎ 2  ≅ 2 . 07 × 10 M Nr T ⋅   (2.17) where Φ l is the magnetic flux quantum, ℎ is the Planck constant and  the electron charge. The device is based on a superconducting loop interrupted by either one (RFSQUID) or two parallel Josephson junctions (DC-SQUID) [67]. Explaining the basics in the last case, the bias current enters the loop and is divided into two paths where the Chapter 2 84 Josephson junctions are present by introducing a very thin insulating layer. A magnetic flux, Φ, threads the superconducting loop being an integer number multiple of the elementary magnetic flux quantum, Φ l . Two wavefunctions represent the two superconducting areas separated by the Josephson junctions with phases t N and t  :  N = |  |  6 u v (2.18)   = |  |  6 u w (2.19) Both wavefunctions penetrate the thin insulating layer and overlap, thus Cooper pairs of electrons tunnel through the barrier with a current, K C , which is proportional to the phase difference between the two superconducting parts at the insulating layer [68]: K C = K x sin  t  − t N  (2.20) where K x is the critical current. If a constant bias current is kept, the voltage across the junctions oscillates with the phase change. At the same time, the current flowing through the SQUID is modulated by the magnetic flux passing through the loop. Thus, measuring the voltage, the response to a change of flux can be detected. The SQUID equipment used in this thesis was a home-made system developed by Dr. María José Martínez-Pérez from the Instituto de Ciencia de Materiales de Aragón (ICMA), Universidad de Zaragoza-Consejo Superior de Investigaciones Científicas (CSIC). 2.4.2 Magneto-Optical Kerr Effect magnetometry Magnetization induces changes in the optical properties of the material. In particular, when polarized light is reflected a magnetic surface, its polarization and intensity are modified [69]. These changes of a polarized laser beam after reflecting from a magnetized Experimental techniques 85 sample surface can be detected by MOKE magnetometry, eventually inferring the magnetization of the sample. Firstly, the machine is typically composed by a 3D motion stage to place the sample. Also, a quadrupole electromagnet surrounds the sample and can apply magnetic fields in two orthogonal directions. The magnetic fields are measured by sensors installed into the coils while an electronic control loop adjusts the electron current in the coils in real time to provide the desired field strength. In addition, the optics head unit contains the laser optics allowing the beam to leave the head and come to a polarizer which selects the light polarization. Then, the beam is focused on the sample surface by a set of lenses which allows the laser to be incident either normal to the surface or at 45 degrees to the surface normal [70]. After the beam is reflected from the sample surface, it is collimated by a lens and passes through an analyser. Then, the intensity and polarization are evaluated. The system shows the polarization rotation, also called Kerr signal, as a function of the applied magnetic field, obtaining the hysteresis loop. In general, this technique provides high sensitivity and fast measurement times, being considered an excellent method for experiments on magnetic nanostructures [71], microstructures [72] and thin films [73]. Two different MOKE systems have been used: few experiments were performed in the NanoMOKE ® 3 magneto-optical magnetometer installed in the ICMA, Universidad de Zaragoza-CSIC, and most work has been performed in the MOKE equipment settled in the Cavendish Laboratory at the University of Cambridge. 2.4.3 Magnetic Force Microscopy The MFM technique is a non-contact mode of the scanning force microscopy based on the detection of magnetic tip-sample interactions at the nanoscale. The method lies in understanding and exploiting the long-range forces arising from the magnetic fields in order to investigate magnetic domain structures [74]. Chapter 2 92 studies of ferromagnetic focused electron beam induced nanodeposits”, Magnetic Characterization Techniques for Nanomaterials, Edited by C. S. S. R. Kumar, Springer (2017). [62] FEI Company, “The Quanta FEG 250/450/650 User Operation Manual”, (2010). [63] T. P. Almeida, D. McGrouther, Y. Pivak, H. H. Perez Garza, R. Temple, J. Massey, C. H. Marrows and S. McVitie, “Preparation of high-quality planar FeRh thin films for in situ TEM investigations”, J. Phys.: Conf. Ser. 903, 012022 (2017). [64] R. L. Fagaly, “Superconducting quantum interference device instruments and applications”, Rev. Sci. Instrum. 77, 101101 (2006). [65] B. S. Deaver and W. M. Fairbank, “Experimental evidence for quantized flux in superconducting cylinders”, Phys. Rev. Lett. 7, 43 (1961). [66] R. Doll and M. Näbauer, “Experimental proof of magnetic flux quantization in a superconducting ring”, Phys. Rev. Lett. 7, 51 (1961). [67] B. D. Josephson, “Possible new effects in superconductive tunneling”, Phys. Lett. 1, 251 (1962). [68] R. P. Feynman, R. B. Leighton and M. L. Sands, “The Feynman lectures on physics”, Addison-Wesley Pub. Co. (1963). [69] Z. Q. Qiu and S. D. Bader, “Surface magneto-optic Kerr effect (SMOKE)”, J. Magn. Magn. Mater. 200, 664 (1999). [70] Durham Magneto Optics Ltd, “NanoMOKE ® 3 User Manual”, (2012). [71] S. Pathak and M. Sharma, “Magneto-optical Kerr effect measurements on highly ordered nanomagnet arrays”, J. Appl. Phys. 111, 07E331 (2012). [72] Y. Jingfan, R. Pérez del Real, G. Infante and M. Vázquez, “Local magnetization profile and geometry magnetization effects in microwires as determined by magneto-optical Kerr effect”, J. Appl. Phys. 113, 043904 (2013). [73] A. Berger, S. Knappmann and H. P. Oepen, “Magneto-optical Kerr effect study of ac susceptibilities in ultrathin cobalt films”, J. Appl. Phys. 75, 5598 (1994). [74] U. Hartmann, “Magnetic force microscopy”, Adv. Mater. 2, 550 (1990). [75] L. M. Belova, O. Hellwig, E. Dobisz and E. Dan Dahlberg, “Rapid preparation of electron beam induced deposition Co magnetic force microscopy tips with 10 nm spatial resolution”, Rev. Sci. Instrum. 83, 093711 (2012). [76] M. R. Koblischka, U. Hartmann and T. Sulzbach, “Improving the lateral resolution of the MFM technique to the 10 nm range”, J. Magn. Magn. Mater. 272-276, 2138 (2004). [77] G. Binnig, C. F. Quate and C. Gerber, “Atomic Force Microscope”, Phys. Rev. Lett. 56, 930 (1986). Experimental techniques 93 [78] D. Rugar and P. Hansma, “Atomic Force Microscopy”, Physics Today 43, 23 (1990). [79] O. Kazakova, R. Puttock, C. Barton, H. Corte-León, M. Jaafar, V. Neu and A. Asenjo, “Frontiers of magnetic force microscopy”, J. Appl. Phys. 125, 060901 (2019). Chapter 3: Tuning the growth of 3D nanowires by FEBID This chapter covers the capability of Focused Electron Beam Induced Deposition to tailor shape, composition and magnetization of 3D nanowires. Particularly, a comprehensive study about the complex mechanisms which govern the growth process and strategies undertaken to tailor their physical properties are tackled. Further, a special section about electrically-biased metal structure patterned on insulators is addressed, which allows fabricating nanostructures on insulating substrates and tuning the lateral dimension of 3D nano-objects by electric biasing. Chapter 3 96 3.1 Principles of FEBID FEBID relies on the decomposition of the precursor gas molecules delivered close to the substrate by a finely-focused electron beam, producing a deposit [1][2][3][4]. The shape of the deposit is determined by the electron beam scan as well as complex interactions between the electron beam, substrate, precursor gas molecules and the growing structure [5][6]. Understanding the influence of the FEBID parameters is a key point to develop proper approaches for the fabrication of nanostructures with optimum properties. Simultaneously, it is important to understand the mechanisms governing the deposition to provide a complete view of this technique. The concurrent control of the various parameters that play a role in the growth process of nanostructures by FEBID is a daunting task [1]. The energy of the primary electrons, which is directly related to beam acceleration voltage, and the electron beam current, defined as the electron flow reaching the sample surface, are two of the most important variables. However, one should also point out other parameters related to the electron beam scan such as the dwell time, which identifies the lapse for which the electron beam is held still on a particular point; the refresh time, which is the period of time between the start of two consecutive loops and allows precursor gas replenishment during the time where the beam is paused in between the loops; the pitch, which is the distance between two neighbouring dwell points; and, the scan direction, associated with the movement of the beam within the pattern. In addition, the number of loops, pattern dimensions and geometry are included. Finally, parameters related to the environmental conditions such as the base pressure, the type of gas precursor and flux, the position of the GIS nozzle, the type of substrate, the temperature or the residence time of the precursor molecules are also decisive variables for the fabrication of nano-objects. With the goal of learning how these parameters affect the growth operation and handle them, the dissociation process caused by the electron beam deserves special attention. Among the many distinct electron-molecule interaction processes [7], the Tuning the growth of 3D nanowires by FEBID 97 elastic scattering, vibrational and electronic excitation, dissociative electron attachment, neutral and bipolar dissociation, and dissociative ionization can be mentioned. In light of this situation, it may be thought that the beam-molecule interaction is too complex to have a thorough knowledge and comprehension of the mechanism at molecular scale. In fact, few experimental [8] and theoretical studies [9] in this regard are reported in literature. However, unsophisticated simulations of the fabrication process by continuum models and approaches of growth geometries by Monte Carlo method have been already considered to clarify the intricacies of the FEBID growth process [10]. To shed light on this topic, the single precursor species continuum model of FEBID should be contemplated. This model considers a weak precursor-substrate interaction and neglects the interplay between the adsorbed precursor gas molecules. Firstly, Langmuir adsorption where the surface coverage is described as a function of the precursor adsorbate density, , is assumed. Secondly, an average residence time, , of the precursor gas molecules on the substrate surface is taken into account. Thirdly, the surface diffusion phenomenon of the precursor gas molecules is also considered. Finally, the dissociation induced by the electron beam is modelled by a depletion of  proportional to the dissociation cross-section,   , and the electron flux distribution per unit time and area, . Thus, the radially symmetric rate equation is written as follows [11]:    ,    =   1 −    ,      −    ,    +        ,      + 1      ,     − (3.1) −          ,    1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1  where the adsorption, desorption, diffusion and dissociation terms can be sequentially identified in the right-hand side of the equation. In the adsorption term,  is the sticking coefficient,  the precursor flux,   the complete area density which corresponds to that of a monolayer, and 1−   ⁄ the fraction of surface sites available for adsorption. The desorption term defines the adsorbate density per  period before desorption. The Chapter 3 98 diffusion term denotes the spreading of the precursor gas molecules on the surface and is proportional to the diffusion constant, . Finally, the dissociation term indicates the reduction of  assuming an energy-integrated   and a radially symmetric  with the Gaussian shape:     =   ⁄ 2   exp $ −   2  % (3.2) where  is the electron beam current,  the electron charge, the standard deviation and  the radial parameter. Additionally, the local deposition growth rate ℛ can be defined as: ℛ    = '        ( )    ,   * + , - (3.3) where ' is the non-volatile deposited product arisen from a precursor gas molecule and  ( the dwell time. Neglecting the diffusion term —which can be a good approach for a wide range of  and  ( values [1]—, the analytical solution of Equation 3.3 can be obtained: ℛ = '      . 1 / 0 − 1 /  1 1 − exp  − /   (  /   ( +  /   (3.4) where /  is the depletion rate and / 0 the replenishment rate: /  =    + 1  +    (3.5) / 0 =    + 1  (3.6) A rough interpretation of the growth rate can be obtained as a function of  ( . On the one hand, for relatively small  ( , — typically on the order of nanoseconds —, ℛ increases Tuning the growth of 3D nanowires by FEBID 99 as a function of . This depositing scenario is the electron-limited regime because the growth is restricted by inability of the electron beam to decompose all precursor molecules. On the other hand, in the case of large  ( , the dissociation rate exceeds the gas precursor replenishment rate, giving rise to the precursor-limited regime. The balance between the availability of precursor molecules on the growth area and the electron beam current is very important because it will determine whether the growth occurs in the precursor-limited regime or the electron-limited regime, which will affect not only the growth rate but also the composition of the nanowire [12]. Furthermore, when thermal heating of the growing deposit occurs, as previously found in FEBID [13][14][15][16][17], the decomposition of the precursor gas molecules will be more efficient if temperatures close to the thermal decomposition of the precursor are reached. These concepts are of utmost importance for determining the appropriate conditions, i.e., the ideal value of each parameter, to ensure an efficient decomposition process. The general notion of the theory presented here can be applied for the fabrication of any type of structure and offers a starting point to begin working. However, although this brings a global overview, given the multitude of nanostructures shapes and materials, the optimization process of the specific variables is essential to find out the best results. 3.2 Tailoring the fundamental properties of 3D cobalt nanowires Many types of materials raise great interest for their study at the nanoscale. For instance, thin-film layers and multilayers based on magnetic materials have nowadays various applications in the fields of data storage and sensing, one example of this being the hard disks [18][19]. Even so, individual magnetic nano-objects are also being investigated and engage huge attention due to their great potential in applications such as sensors [20], memories [21] and logics [22]. In this case, most of the approaches for their fabrication rely on standard lithography procedures, most adequate for patterning 2D structures onto magnetic thin films and multilayers. However, there is an increasing Chapter 3 100 interest on the fabrication of 3D magnetic nanostructures. In this regard, FEBID is one of the techniques that allow addressing the growth of such 3D structures [23][24][25][26][27], particularly those based on magnetic materials [28][29][30][31] [32][33][34][35][36][37]. The use of precursor gas molecules containing magnetic elements such as Co, Fe and Ni permits the growth of magnetic deposits [28][38][39][40][41][42][43][44]. A large development has been made towards the growth of magnetic deposits with high metal content, high magnetization, high resolution and complex shapes, as recently reviewed [45][46]. Such development has been focused on the optimization of thin inplane magnetic layers, whereas limited work has been done in the case of 3D magnetic deposits. However, there are many promising applications of 3D magnetic deposits in scanning probe techniques, such as MFM [34] and Ferromagnetic Resonance Force Microscopy [47], racetrack-type magnetic memories [31], Hall sensors [48][49], nanomagnetic logic circuits [34][50], superconducting vortex lattice pinning [51], remote magneto-mechanical actuation [37], etc. 3D fabrication implies the understanding of specific growth phenomena which do not happen in the case of 2D deposits, and as a result new interpretation emerges from the nanofabrication process. Previous work on the growth of 3D nanowires by FEBID has shown the relevance of several parameters that should be taken into account. For example, the use of sub-nA electron beam currents produced by field-emission guns is mandatory for the growth of narrow nanowires (<100 nm in diameter) [31][52]. Moreover, thermal effects can be of tremendous importance in 3D nanostructures given that precursor replenishment in the area of growth occurs at a lower rate compared to in-plane deposits because the diffusion mechanism of precursor gas molecules from the substrate will be weakened as the deposit grows in height. When growing a 3D nanowire, the substrate surface close to the growth point is small and the number of precursor gas molecules adsorbed to be decomposed is lower than in 2D deposits. In addition, since heat dissipation is difficult, the temperature Tuning the growth of 3D nanowires by FEBID 101 increases in the growth point and can give rise to interesting effects such as growth regime transitions, discussed in the next subsections. In this context, the impact of the most relevant growth parameters on the specific properties of 3D ferromagnetic nanostructures and their final performance will be explored in this section. The focus is put on the characterization of the obtained nanowire’s diameter, composition and magnetization, with the aim of growing narrow nanowires (<100 nm in diameter), with high Co content (>80% at.) and magnetization approaching the bulk value. 3.2.1 Experimental details The nanowires were fabricated in the commercial Helios Nanolab 600 and 650 Dual Beam equipment using Co 2 (CO) 8 as a gas precursor. The substrates were TEM Cu grids. Co deposits were grown with low electron beam currents (≤100 pA). The voltage was fixed to 5 kV given that preliminary experiments did not lead to significant changes in the composition from 5 kV to 30 kV. The nanowires were grown in spot mode, where the electron beam is continuously irradiating a single point. A base chamber pressure of ~1 × 10 -6 mbar was achieved before the injection of the precursor, with the GIS needle position at ~50 µm in 2, 3 and 4 directions. The precursor gas flux was tuned via a manual valve, which permits to vary the chamber pressure up to ~4 × 10 -5 mbar. Even though the precursor gas flux cannot be measured directly, given its linear relationship with the chamber pressure increase during gas injection, Δ6, and the precursor gas flux, , stated as ∝Δ6 [53], monitorization of the chamber pressure during growth allows to establish relative correlations with the physical properties of the 3D nanowires. Some of the EDS experiments were performed in the Helios Nanolab 650 Dual Beam, using an excitation electron beam voltage of 5 kV and beam current of 800 pA. Other EDS experiments were carried out in an FEI Tecnai F30 TEM operated at 300 kV. EELS experiments were performed in the FEI Tecnai F30 TEM and in a probe-corrected Chapter 3 108 Although specific experiments and/or simulations could shed more light on the origin of this change in composition, from general arguments it can be stated that at lower precursor gas flux the Co content diminishes due to decomposition of residual contaminant species in the working chamber, mainly C and O. Also, the behaviour of the Co content as a function of Δ6 resembles that observed in in-plane deposits [49]: an optimum precursor flux window (1 × 10 -5 mbar < Δ6 < 1.5 × 10 -5 mbar) exists, where the Co content is relatively high. On the other hand, at higher precursor gas flux the Co content is reduced because incompletely decomposed precursor gas molecules are incorporated to the deposit. The number of precursor gas molecules is so high that the electron beam cannot properly dissociate the gas. The different origin of the decreased Co content at low and high precursor flux can be also noted in the C/O ratio, which is smaller than 1 at high precursor gas flux and larger than 1 at low precursor gas flux. From Figure 3.4, under this growth conditions, it is clear that optimum Co content (>85% at.) can be only achieved in the radial regime, where the diameter is at least ~120 nm. The inefficient decomposition phenomenon has also been found in the experiment represented in Figure 3.5. To evidence clearly this effect, STEM-EELS chemical maps and profiles were acquired in a nanowire grown in the radial regime under the appropriate conditions to induce the phenomenon. As illustrated, a Co content decrease is exhibited in the core centre. This entails a relative C composition increase in that area, so the uniform composition disappears completely. It could be thought that the central part of the nanowire is so thick that EELS signals are not quantified properly due to multiple scattering [55]. However, the decrease of intensity in the core region observed in the HAADF-STEM image of the cross section, shown in the rounded inset of Figure 3.5, can only be explained by a reduction of the average atomic number in this region, thus a reduction of the Co content. Tuning the growth of 3D nanowires by FEBID 109 Figure 3.5. Relative composition profiles as a function of the radius of the nanowire determined by STEM-EELS. The inset depicts the HAADF-STEM image of a cross-section in grey scale, and the chemical maps showing the spatial distribution of Co, C and O in green, blue and red, respectively. Undefined scale bars are 25 nm in the STEM image and 100 nm in the STEM-EELS maps. This is in good agreement with previous studies suggestions, where the Co signal decrease could be induced by a lack of precursor gas molecules due to a temperature increase, caused by the constant impact of the primary electrons [56][57]. A higher temperature leads to a decrease in  and the precursor desorbs faster. By this, the material deposited in the centre of the nanowire are residual components, basically C atoms with a slight contribution of O ones. On the contrary, the volume around the core centre mainly grows thanks to SE emission, and it contains a higher metallic content due to its lower temperature. This is reminiscent of the basics related to a recent publication for the Chapter 3 110 growth of vertical hollow nanowires by He + Focused Ion Beam Induced Deposition (FIBID) [58]. Since the decomposition of the precursor gas molecules is closely linked to the electron beam properties, the electron beam current appears to be a critical growth factor. As a first approximation, let us focus the attention on nanowires with just one single growth mode. As previously mentioned, a low electron beam current is a pre-requisite for the growth of small-diameter nanowires. This is typically produced by using small apertures which limit the electron beam size for the fabrication process and, in this case, because of the inherent Co 2 (CO) 8 properties. For example, the results obtained for the Fe 2 (CO) 9 precursor gas are completely different, where the metallic composition is constant for all the available electron beam currents. In this scenario, it is important to assess the evolution of all the compositional elements in the resulting 3D Co-FEBID nanostructures as a function of the electron beam current. Figure 3.6 shows Co, C and O contents evaluated as a function of the electron beam current. An increase of the electron beam current provokes an increment of the metallic composition. This fact reinforces the idea that a compromise in the electron beam current regarding the lateral dimensions and the metallic composition must be made. Figure 3.6. Relative composition as a function of the electron beam current for Co nanowires. Tuning the growth of 3D nanowires by FEBID 111 In this line, dedicated experiments have been carried out using the best growth conditions to obtain the highest possible metallic content at each particular value of the nanowire diameter. The results are summarized in Figure 3.7, where the Co content is represented as a function of the diameter for optimum growth conditions. The specific growth parameters of each nanowire are displayed in Table 3.2. Figure 3.7 indicates that a high Co content (>85% at.) can be achieved in nanowires with diameter larger than ~120 nm, which correspond to the radial regime. However, the Co content in the nanowires with linear regime, whose diameter is smaller than ~80 nm, starts from ~75% at. for diameters of ~80 nm and diminishes quickly as the diameter is reduced. For instance, for diameters of ~60 nm the Co content is only ~45% at. Given that the nanowires present typical oxidized shells of around 5 nm [52][59], the measured average Co content will be lower as the wire diameter decreases. This means that in the core of the nanowire the Co content is expected to be higher than the average value, this effect being more significant for the narrowest nanowires. Figure 3.7. Co composition as a function of the nanowire diameter for optimized growth conditions at each particular value of the diameter. Chapter 3 112 Diameter (nm) Co composition (%, at.) Technique Growth current (pA) 8 P (10 -6 mbar) 56.9 40.6 EDS 50 2.8 58.8 46.9 EDS 25 10.3 60.4 43.1 EDS 50 10.2 65.0 58.3 EDS 100 10.5 65.6 50.0 EDS 100 9.3 73.7 67.3 EDS 100 8.1 79.1 81.2 EDS 50 6.9 80.9 67.5 EDS 100 8.1 81.2 83.8 EDS 100 8.4 82.6 84.0 EDS 50 7.8 119.2 83.1 EELS 50 2.8 123.9 87.4 EDS 50 5.2 148.5 83.2 EDS 100 6.9 149.0 87.4 EELS 100 6.1 Table 3.2. Data associated with the information represented in Figure 3.7: diameter, Co content and technique used to obtain it, electron beam current and Δ6 during growth. A typical error of ~2% at. in composition is considered assuming uniform distribution. 3.2.4 Magnetic induction as a function of the diameter In order to correlate the Co content of the nanowires with their magnetization, EH experiments have been carried out. Three nanowires which contain different metallic compositions and growth modes have been analysed to obtain quantitative values of the magnetic induction. By using Equation A.3, the average magnetic induction inside the nanowire along its long axis, denoted simply as 9 hereafter, can be calculated. In Figure 3.8, the results corresponding to three nanowires, representative of the three different regimes found, are shown. Tuning the growth of 3D nanowires by FEBID 113 Figure 3.8. (a) Magnetic induction flux lines of nanowires with Co content of (a) 87.4% at. (NW1), (b) 67.5% at. (NW2), and (c) 40.6% at. (NW3), all obtained from the magnetic phase image after normalizing by the maximum thickness and performing the cosine of 700 times the magnetic phase shift image. (d) Transversal profiles of 9 of NW1, NW2 and NW3. The values obtained for 9 close to the nanowires’ surfaces are not reliable due to the uncertainties in the sample thickness at those positions and edge effects at the oxidized nanowire surface. For this reason, the values of 9 obtained at the edges of the nanowires are masked with a semi-transparent band, whereas only the values obtained in the central part of the nanowires are trustworthy. The nanowire with the largest diameter, 123.9 nm, corresponding to the radial regime, presents a 9 (~1.33 T) not far from the bulk value, Chapter 3 114 1.76 T. This high value of the magnetization correlates well with the high Co content in the nanowire, 87.4% at. A second nanowire, corresponding to the intermediate linear and radial regime, has been analysed at the base, in the section grown under linear regime. It presents a 9 of 0.78 T, around 50% of the bulk magnetization of Co. This reduction is expected given the reduced Co content (67.5% at.) in this nanowire. A third nanowire, corresponding to the linear regime, presents a lower 9 (~0.41 T), which can be expected given its even lower Co content (40.6% at.). 3.2.5 Discussion of the results It can be argued that FEBID growth of functional magnetic nanostructures requires exhaustive control of a high number of growth parameters. Their precise tuning can be crucial in particular cases, such as the growth of 3D Co nanowires discussed in this section. In the process of growth optimization, a number of interesting phenomena that should be taken into account for their practical application have been encountered. The first important finding regards the existence of two growth modes with different physical properties, denoted linear and radial regimes due to certain similarities with reported growth of 3D Fe nanowires [13]. In the radial regime, the nanowires feature diameters >120 nm with very high Co contents (>85% at.), showing a high magnetization not far from the bulk value (1.76 T). In the linear regime, the diameters are typically lower than 80 nm with the Co content strongly diminishing for decreasing diameter. For instance, 9 is around 1/2 of the bulk value for nanowires with ~80 nm in diameter and ~75% at. Co, reducing 9 down to 1/4 of the bulk value for nanowires with ~60 nm in diameter and ~45% at. Co. It cannot be discarded that the nanowires of low Co content have areas with inhomogeneous composition, with the richer Co regions contributing more to the magnetization of the nanowire. Interestingly, inside the same nanowire, a transition between both growth modes can be observed in a certain range of growth parameters. This effect seems to indicate that thermal desorption and diffusion effects Tuning the growth of 3D nanowires by FEBID 115 during the growth may play a key role. The capacity to dissipate the heat originated by the electron beam is reduced as the nanowire grows in height, being the tip growing progressively further away from the substrate. At a certain height, there is an overheating which could result in a change of the growth mode. Recent results, published after our work had finished, highlight this effect and provide a quantitative description [17]. The existence of such single nanowires with two diameters seems useful for studies of magnetic-domain-wall propagation, given the tendency of domain walls to get pinned at the location of the transition between both diameters [60]. The correlation found between the diameter of the nanowire and its composition is important given the relationship observed between the Co content and the magnetization of the nanowire. If a nanowire with 9 close to the bulk value is required, the best option is to grow a nanowire with diameter of at least 120 nm. However, in many practical situations, narrow nanowires (<100 nm) are requested, in which case a maximum Co content of ~80% at. can be achieved. In such situation, 9 is observed to decrease with respect to the bulk value. There are many potential applications of these nanowires such as magnetic functionalization of cantilevers [61][28][34][62][30], 3D logic structures [34][50], cylindrical conduits for domain-wall propagation [31], etc. where lateral resolution is more important than the absolute value of the magnetization. In those cases, the type of nanowires grown here in the linear regime meet the required physical properties. Just as a particular example, the Fe magnetic rods used in the past by Franken et al. had 9 of 0.13 T and were able to pin domain walls in a domain-wall conduit [33]. 3.3 Electrically-biased patterned metal structure In spite of what has been presented in the previous section, FEBID still presents some limitations with respect to the precise control of the dimensions of nano-objects as well as its use on insulating substrates. To overcome these limitations, in this section a novel approach based on the use of electrically-biased metal structures patterned on the Chapter 3 116 surface of the substrate is explored, called the ARAGON-Chip (acronym derived from ARchitectural Adjustment by Grid Overlay Nanotechnology-Chip) [63]. Hereafter, this refers to any type of insulating substrate top-covered with a periodically patterned metal layer. 3.3.1 Introduction One of the prominent properties searched by any lithography technique is the capability to be used on an arbitrary substrate. In the case of lithography techniques based on charged particles, such as EBL and FIB lithography, a limiting factor is the difficulty of working on insulating substrates [64][65]. The building of electric fields in the vicinity of an insulating substrate (due to charging effects) produces unwanted beam deflections that ruin the resolution or even impede their practical application [66][67][68]. In the case of EBL, the problem can be circumvented by working under critical-energy conditions [69], using variable-pressure EBL [70], or including additional steps in the process, like the evaporation of a metal layer on top of the resist that is subsequently etched away [71]. In the case of FIB, it has been found that using electron irradiation simultaneously is helpful towards charge minimization [72]. FEBID is another charged-particle nanolithography technique suffering from the same problem when applied on insulating substrates. In general, FEBID applications are developed using substrates allowing the charge dissipation and the avoidance of detrimental electric potentials in the proximity of the substrate surface. In fact, it has recently shown that the use of an insulating substrate (such as polycarbonate) hampers the growth of high-resolution nanostructures by FEBID unless pads for charge dissipation exist on the substrate, close to the area of growth [73]. In this thesis, the strategy followed to overcome the charging problem of FEBID consists on the use of patterned metal structures fabricated by optical lithography, which could also be produced by any other lithography technique. Moreover, if the patterned Tuning the growth of 3D nanowires by FEBID 117 metal structure on an insulating substrate is electrically biased, tunable electric fields can be created on the substrate surface, producing a tremendous impact on the trajectories of the primary electrons and the generated SE. Hereafter, it will be shown that this new degree of freedom in FEBID is useful to tune the dimensions of 3D nanostructures grown by this technique. A previous proof-of-concept work has shown that an electric field produced by a continuous metallic substrate can modify the dimensions of FEBID deposits grown on its surface [74]. Also, local electric fields have been previously used to tailor the growth of nanowires using techniques such as VLS processes [75], resulting in a flexible method to tune their dimensions. However, in this case the interest is focused on patterned metal structures with micrometric holes, which allows the local application of tunable high electric fields inside the holes. Finally, electron beam defocus will be used as a new tool to modulate the diameter and length of 3D nanowires. Although the dimension broadening effect of beam defocus has already been reported in FEBID [76], in situ control of this parameter to obtain 3D structures with tailored varying diameters represents a new avenue in the use of FEBID for the growth of functional magnetic, superconducting or photonic materials. 3.3.2 Experimental details Patterned metal structures were fabricated on SiO 2 , MgO and quartz substrates by optical lithography. Firstly, the electron-beam-evaporator equipment was used to deposit two consecutive metallic layers: a Cr layer with ~10 nm in thickness to improve the adherence and a Cu layer of ~500 nm. The chamber base pressure was 4 × 10 -7 mbar, increasing up to 1 × 10 -6 mbar in the Cr deposition, using a current of 20 mA for 105 seconds. The pressure for the Cu deposition was 1.5 × 10 -6 mbar at 25 mA for 18 minutes. After that, a standard lithography procedure was followed using a positive resist and a mask with a periodic array of holes, so the metallic film was perforated with circular holes with a diameter of ~4 µm, as shown in Figure 3.9. Chapter 3 124 The vector map of the electric field in the 23 plane of Figure 3.11 is displayed in Figure 3.13(a), which illustrates that the application of a negative (positive) DC voltage in such structure produces large and highly directional electric fields pointing towards (or from) the centre of the hole. In the centre of the hole, the electric field vanishes. The total electric field in the plane perpendicular to the hole, 24 plane in Figure 3.11, corresponds to the vector map shown in Figure 3.13(b). The electric field is very small along the symmetry axis of the circular hole and grows rapidly as the distance from this axis increases. The electric field numerical calculations and vector maps representations were performed by Dr. Soraya Sangiao. 3.3.5 Dimensional modulation as a function of the applied voltage It has been observed that the growth of the nanowires under different voltage values applied to the ARAGON-Chip, whilst keeping the other growth conditions constant, affects dramatically the dimensions of the nanowires. In these experiments, Pt-C and WC nanowires are grown at the centre of the holes with amorphous SiO 2 substrate under different voltages in the range of -200 V to +200 V. On the one hand, as shown in Figure 3.14(a), the nanowire diameter is found to change from a large value around 100 nm under -200 V to a minimum value of 50 nm at -25 V, and then increasing up to 90 nm under +200 V. This figure illustrates the tuning of the value of the nanowire diameter by a factor of 2 by application of the created local electric fields. On the other hand, the aspect ratio of the nanowire is also represented in Figure 3.14(a). It shows a maximum value at around -25 V, in good correspondence with the minimum value of the diameter at such voltage, and indicates that the aspect ratio increases by a factor of 4 working under such voltage in comparison to -200 V. SEM images of Pt-C nanowires grown respectively under -100 V, -35 V and +100 V are shown in Figure 3.14(b-d). At -100 V the diameter at a height of 1 µm is 73 nm, whereas it Tuning the growth of 3D nanowires by FEBID 125 decreases to 50 nm under -35 V and increases again to 83 nm under +100 V. Moreover, the nanowire length is, respectively to those voltages, 2.6 µm (-100 V), 2.9 µm (-35 V) and 2.2 µm (+100 V), corresponding to linear growth rates of 43.3 nm/s, 48.3 nm/s and 36.7 nm/s, respectively. Thus, the results are conclusive with respect to an optimized nanowire growth in terms of resolution under negative voltages around -25 V. Figure 3.14. Dimensional modulation of Pt-C nanowires grown for 60 seconds on amorphous SiO 2 substrate: (a) dependence of the diameter and the aspect ratio with the voltage applied to the ARAGON-Chip. SEM artificially coloured images of selected nanowires grown under (b) -100 V, (c) -35 V and (d) +100 V, matching the points marked with green arrows in (a). Chapter 3 126 Similar experiments have been carried out to investigate the influence of local electric fields on the dimensions of W-C nanowires using the ARAGON-Chip on the same amorphous SiO 2 substrate. As shown in Figure 3.15, the minimum nanowire diameter (~65 nm) occurs in the -25 V to -50 V range and the aspect ratio is maximized in the same voltage range. In this case, the degree of tunability of the diameter and the aspect ratio compared to the Pt-C case is lower but substantial (factors of 1.3 and 1.7, respectively). In view of the set of results obtained for Pt-C and W-C nanowires, spatiallydependent electric fields can be established as a new degree of freedom to tune the dimensions of nanowires grown by the FEBID technique using this novel approach. The origin of the effect will be discussed hereafter in the light of the physical phenomena involved in the FEBID growth and the electric field distribution in the ARAGON-Chip. Figure 3.15. Dimensional modulation of W-C nanowires grown on amorphous SiO 2 substrate illustrating the dependence of their diameter and aspect ratio with the voltage applied to the ARAGON-Chip. Tuning the growth of 3D nanowires by FEBID 127 3.3.6 Qualitative discussion of the electric field action Regarding the physical phenomena involved in the growth of vertical structures by FEBID, Fowlkes et al. noted that the most significant contribution to the dissociation of the precursor gas molecules arises from the primary beam electrons and their generated secondary electrons (SE-I), with lower relevance from BSE and their generated secondary electrons (SE-II) [80]. Later, Smith et al. stressed the important role played by forward-scattered electrons (FSE) in the growth of 3D nanowires as well as the dependence of the effect with the primary beam energy and the resulting growth regime [81]. Moreover, Arnold et al. showed that SE-II are responsible for the halo deposit around the main deposit in 2D structures, affecting the final deposit resolution [82]. A similar study by Schmied et al. in the case of 3D deposits has shown various competing effects, with significant broadening of the deposit width under intermediate or low electron beam energy [83]. The FEBID process is thus complex and depends on many parameters, but in general the precursor dissociation process is determined by highenergy electrons (mainly arising from primary beam electrons and FSE) as well as by low-energy electrons (mainly SE-I and SE-II). Therefore, our experimental finding of a considerable effect of local electric fields on the dimensions of FEBID deposits should be correlated with the effect produced by such electric fields on the trajectories of these electrons responsible for the precursor gas molecules dissociation. In consideration of the previously reported results, a qualitative explanation of the observed effect can be provided. From the electrostatic point of view, the electric field distribution generated at the perforated Cu film is in fact acting as a small (de)focusing lens. The electrostatic lenses can be found in electron and ion microscopes [84], and other devices based on charged particles [85]. In the ARAGON-Chip, as the primary electron beam approaches the perforated charged film, it suffers a deviation towards the optical axis (under negative voltage) or off the optical axis (under positive voltage), as sketched in Figure 3.16. In both cases, this effectively leads to an electron beam defocus. Chapter 3 128 Figure 3.16. Sketch illustrating the relevant phenomena involved in the growth of 3D nanowires by FEBID when an external positive or negative voltage is applied to the ARAGON-Chip. On the one hand, the primary beam becomes slightly defocused, giving rise to the diameter broadening. On the other hand, the SE are strongly attracted to (' Z[ \0) or repelled from (' Z[ ]0) the optical axis, contributing to the nanowire narrowing or broadening, respectively. As reported by Plank et al. [76], the effect of defocusing the primary electron beam triggers an increase of the nanowire diameter. On the other hand, the high electric field will strongly modify the trajectory of the SE produced in the substrate and the growing structure, which is also crucial for the FEBID growth. Due to the low energy of SE-I and SE-II (<50 eV) [86], they will undergo strong attraction to or repulsion from the optical axis for negative or positive voltage, Tuning the growth of 3D nanowires by FEBID 129 respectively. In fact, similar voltage values in the range of a few tens or hundreds of volts are actually used in electron microscopes to optimize the imaging contrast by SE, where biasing voltage can be applied to the sample itself [87][88] or to the electron detector [89][90]. As a result, combining the growth by FEBID and the ARAGON-Chip strategy, the use of a negative voltage will push SE-I and SE-II to the hole centre contributing to decrease the nanowire diameter, whereas the use of a positive voltage makes SE-I and SE-II to be pulled out from the hole central axis leading to a diameter enlargement. Summarizing, when a negative voltage is applied to the perforated Cu film, two effects will compete: nanowire narrowing due to the effect of the electric field on the SEI and SE-II, and nanowire broadening due to the defocusing effect of the electric field on the primary beam electrons. This agrees with the experimental results shown in Figure 3.14 and Figure 3.15. Starting from 0 V, the nanowire diameter first decreases down to an applied voltage of around -25 V, then increasing the thickness as the voltage is reduced to -200 V. As the sketch in Figure 3.16 also illustrates, this is the expected evolution given the rapid focusing of the SE towards the hole central axis at low voltages and the smoother increase of the primary beam diameter with the voltage. Finally, at positive voltage, the diameter is always increasing with voltage given that both effects contribute to the nanowire broadening. 3.3.7 Dimensional modulation as a function of the beam defocus In order to verify the hypothesis described above, additional experiments were performed where the electron beam focus conditions were controllably changed before and during the nanowire growth. For that, the focus height, 4, was varied with respect to the optimum focus condition, (4=0), to investigate the corresponding changes in the nanowire diameter and height with the defocus, ∆4. The first experiments consisted of growing Pt-C nanowires for different 4 values fixed during the whole nanowire growth. From the optimum focus condition, both Chapter 3 130 overfocus (4\0) and underfocus (4>0) conditions were considered. In the first case the beam is focused below the hole 23 plane, whereas in the second one the beam is focused above it. The obtained results for the values of the diameter and aspect ratio as a function of 4 are represented in Figure 3.17. In order to check that the broadening effect is independent of the voltage applied to the ARAGON-Chip, the experiments were performed at -25 V and -200 V, represented in Figure 3.17(a) and 3.17(b), respectively. The 4 value has been swept a maximum of 40 µm, leading to enormous differences in the nanowire diameter and aspect ratio. These variations reach a factor of 3 for the diameter and a factor of 12 for the aspect ratio with ∆4 of ~20 µm. On the one hand, these results highlight the importance of a good primary beam focus in order to achieve the highest resolution in the growth of 3D nanostructures. In order to show the differences which appear in the SEM images when the 4 value varies, an example is shown in Figure 3.18. It can be appreciated the defocusing effect of a ∆4 of ~20 µm in comparison with the optimum focus position for imaging acquisition. On the other hand, the results interestingly show an additional ingredient to play with during the growth of 3D nanostructures by FEBID. To exploit this additional parameter, the 4 value was varied during the growth of single nanowires in order to create continuous nanostructures with modulated diameter. First, as shown in Figure 3.19(a), a Pt-C nanowire with large diameter (146 nm) starts to grow due to the primary beam defocus (∆4 of 20 µm). At 72 seconds, by bringing 4 into the optimal position with a quick (~1 second) change in the 4 position, a narrow section (64 nm in diameter) starts to grow. After 72 seconds, the 4 position is modified 20 µm, producing a primary beam defocus and the corresponding increase of the diameter to 119 nm. After another 72 seconds in these conditions, the growth is finished, obtaining a diameter-modulated nanowire with wide-narrow-wide sections. A second diameter-modulated nanowire has been grown following a similar strategy but with narrow-wide-narrow sections, as shown in 3.19(b). Tuning the growth of 3D nanowires by FEBID 131 Figure 3.17. Dimensional modulation of Pt-C nanowires grown for 60 seconds on amorphous SiO 2 substrate: dependence of the diameter at half nanowire’s height and aspect ratio as a function of the primary beam focus height. The external voltage applied to the ARAGONChip was (a) -25 V and (b) -200 V. Chapter 3 132 Figure 3.18. SEM top view images of a ∼4µ m-diameter hole with a 3D Pt-FEBID nanowire grown on the amorphous SiO 2 substrate, with the electron beam focused (a) 20 µm above the height of the Cu surface, (b) at the height of the Cu surface and (c) 20 µm under the height of the Cu surface. In all cases there is no voltage applied to the ARAGON-Chip. Figure 3.19. Diameter-modulated Pt-C artificially coloured nanowires grown on SiO 2 substrate by means of the in situ variation of the primary electron beam focus condition. Each thin and thick diameter region was fabricated for 72 seconds. Tuning the growth of 3D nanowires by FEBID 133 It is interesting to mention that, despite every section in the nanowires of Figure 3.19 has been grown for 72 seconds, their respective length is different. This is expected for sections with dissimilar diameter, where it would be expected the conservation of the volume growth rate instead of the linear growth rate. However, this length difference also occurs for sections with the same diameter, as it can be clearly observed in Figure 3.19(b) for the narrow sections. In this case, the top narrow section is shorter than the bottom one, which can be explained by the diminishment of the growth rate with an increasing distance to the substrate caused by precursor gas diffusion effects [91]. It has been checked that the nanowire diameter increases as 4 moves away from its optimum position. In the light of this finding, inducing a ∆4 keeping the substrate still has the same effect as moving the substrate in the vertical axis when keeping 4 constant. However, mechanical movements of the stage make this latter approach less refined, thus the diameter changes have been always obtained introducing a ∆4. So far, for a single nanowire grown under a constant 4, slight changes in diameter along the length were detected. Nevertheless, additional experiments were carried out growing W-C nanowires and inducing diameter changes keeping the same 4 value and substrate position during the whole growth. For this purpose, a further step towards dimensional modulation of 3D nanowires can be taken by using the immersion imaging mode during growth. In this mode, a magnetic field is applied to collect a higher number of electrons than in standard field-free mode, obtaining ultra-high-resolution images. As shown in Figure 3.20, this mode will be used for the fabrication of nanowires. The nanostructures were grown at different positions of a still sloping substrate under the same focus condition, where just the substrate height changes for each case. The red dotted line indicates the constant 4 value used for the growth of the nanowires. Figure 3.20(a) illustrates the nanowire with the optimum focus height at the growth starting point. A smooth continuous beam defocus ∆4 is introduced per se during the growth of the nanostructure, increasing the diameter from 39 nm to 52 nm as the focus is gradually driven away. Chapter 3 140 Nanotechnology 16, 1321 (2005). [40] R. Lavrijsen, R. Córdoba, F. J. Schoenaker, T. H. Ellis, B. Barcones, J. T. Kohlhepp, H. J. M. Swagten, B. Koopmans, J. M. De Teresa, C. Magén, M. R. Ibarra, P. Trompenaars and J. J. L. Mulders, “Fe:O:C grown by focused-electronbeam-induced deposition: magnetic and electric properties”, Nanotechnology 22, 025302 (2011). [41] A. Perentes, G. Sinicco, G. Boero, B. Dwir and P. Hoffmann, “Focused electron beam induced deposition of nickel”, J. Vac. Sci. Technol. B 25, 2228 (2007). [42] R. Córdoba, B. Barcones, E. Roelfsema, M. A. Verheijen, J. J. L. Mulders, P. H. F. Trompenaars and B. Koopmans, “Functional nickel-based deposits synthesized by focused beam induced processing”, Nanotechnology 27, 065303 (2016). [43] E. Nikulina, O. Idigoras, J. M. Porro, P. Vavassori, A. Chuvilin and A. Berger, “Origin and control of magnetic exchange coupling in between focused electron beam deposited cobalt nanostructures”, Appl. Phys. Lett. 103, 123112 (2013). [44] M. Gavagnin, H. D. Wanzenboeck, D. Belic, M. M. Shawrav, A. Persson, K. Gunnarsson, P. Svedlindh and E. Bertagnolli, “Magnetic force microscopy study of shape engineered FEBID iron nanostructures”, Phys. Status Solidi A 211, 368 (2014). [45] J. M. De Teresa and A. Fernández-Pacheco, “Present and future applications of magnetic nanostructures grown by FEBID”, Appl. Phys. A 117, 1645 (2014). [46] J. M. De Teresa, A. Fernández-Pacheco, R. Córdoba, L. Serrano-Ramón, S. Sangiao and M. R. Ibarra, “Review of magnetic nanostructures grown by focused electron beam induced deposition (FEBID)”, J. Phys. D: Appl. Phys. 49, 243003 (2016). [47] F. Guo, L. M. Belova and R. D. McMichael, “Spectroscopy and imaging of edge modes in permalloy nanodisks”, Phys. Rev. Lett. 110, 017601 (2013). [48] M. Gabureac, L. Bernau, I. Utke, and G. Boero, “Granular Co-C nano-Hall sensors by focused-beam-induced deposition”, Nanotechnology 21, 115503 (2010). [49] L. Serrano-Ramón, R. Córdoba, L. A. Rodríguez, C. Magén, E. Snoeck, C. Gatel, I. Serrano, M. R. Ibarra and J. M. De Teresa, “Ultrasmall functional ferromagnetic nanostructures grown by focused electron-beam-induced deposition”, ACS Nano 5, 7781 (2011). [50] N. Sharma, R. A. van Mourik, Y. Yin, B. Koopmans and S. S. P. Parkin, “Focused-electron-beam-induced-deposited cobalt nanopillars for nanomagnetic logic”, Nanotechnology 27, 165301 (2016). [51] O. V. Dobrovolskiy, E. Begun, M. Huth, V. A. Shklovskij and M. I. Tsindlekht, Tuning the growth of 3D nanowires by FEBID 141 “Vortex lattice matching effects in a washboard pinning potential induced by Co nanostripe arrays”, Physica C 471, 449 (2011). [52] J. Pablo-Navarro, C. Magén and J. M. de Teresa, “Three-dimensional core-shell ferromagnetic nanowires grown by focused electron beam induced deposition”, Nanotechnology 27, 285302 (2016). [53] I. Serrano-Esparza, R. Córdoba, J. J. L. Mulders, M. R. Ibarra and J. M. De Teresa, “Precursor competition in focused-ion-beam-induced co-deposition from W(CO) 6 and C 10 H 8 ”, Sci. Lett. 4, 127 (2015). [54] A. Szkudlarek, M. Gabureac and I. Utke, “Determination of the Surface Diffusion Coefficient and the Residence Time of Adsorbates via Local Focused Electron Beam Induced Chemical Vapour Deposition”, J. Nanosci. Nanotechnol. 11, 8074 (2011). [55] R. F. Egerton, “Electron Energy-Loss Spectroscopy in the Electron Microscope”, Springer (2011). [56] A. Fernández-Pacheco, J. M. De Teresa, A. Szkudlarek, R. Córdoba, M. R. Ibarra, D. Petit, L. O’Brien, H. T. Zeng, E. R. Lewis, D. E. Read and R. P. Cowburn, “Magnetization reversal in individual cobalt microand nanowires grown by focused-electron-beam-induced-deposition”, Nanotechnology 20, 475704 (2009). [57] F. J. Schoenaker, R. Córdoba, R. Fernández-Pacheco, C. Magén, O. Stéphan, C. Zuriaga-Monroy, M. R. Ibarra and J. M. De Teresa, “Focused electron beam induced etching of titanium with XeF 2 ”, Nanotechnology 22, 265304 (2011). [58] R. Córdoba, A. Ibarra, D. Mailly and J. M. De Teresa, “Vertical Growth of Superconducting Crystalline Hollow Nanowires by He + Focused Ion Beam Induced Deposition”, Nano Lett. 18, 1379 (2018). [59] S. Sangiao, C. Magén, D. Mofakhami, G. De Loubens and J. M. De Teresa, “Magnetic properties of optimized cobalt nanospheres grown by focused electron beam induced deposition (FEBID) on cantilever tips”, Beilstein J. Nanotechnol. 8, 2106 (2017). [60] E. Berganza, C. Bran, M. Jaafar, M. Vázquez and A. Asenjo, “Domain wall pinning in FeCoCu bamboo-like nanowires”, Sci. Rep. 6, 29702 (2016). [61] Y. M. Lau, P. C. Chee, J. T. L. Thong, and V. Ng, “Properties and applications of cobalt-based material produced by electron-beam-induced deposition”, J. Vac. Sci. Technol. A 20, 1295 (2002). [62] G. Tosolini, J. M. Michalik, R. Córdoba, J. M. de Teresa, F. Pérez-Murano and J. Bausells, “Magnetic properties of cobalt microwires measured by piezoresistive cantilever magnetometry”, Nanofabrication 1, 80 (2014). [63] J. Pablo-Navarro, Soraya Sangiao, C. Magén and J. M. De Teresa, “Diameter Chapter 3 142 modulation of 3D nanostructures in focused electron beam induced deposition using local electric fields and beam defocus”, Nanotechnology (2019). [64] K. M. Satyalakshmi, A. Olkhovets, M. G. Metzler, C. K. Harnett, D. M. Tanenbaum and H. G. Craighead, “Charge induced pattern distortion in low energy electron beam lithography”, J. Vac. Sci. Technol. B 18, 3122 (2000). [65] L. D. Menard and J. M. Ramsey, “Fabrication of sub-5 nm nanochannels in insulating substrates using focused ion beam milling”, Nano Lett. 11, 512 (2011). [66] J. Cazaux, “Some considerations on the electric field induced in insulators by electron bombardment”, J. Appl. Phys. 59, 1418 (1986). [67] M. Toth, W. R. Knowles and B. L. Thiel, “Secondary electron imaging of nonconductors with nanometer resolution”, Appl. Phys. Lett. 88, 023105 (2006). [68] W.-Q. Li, K. Mu, and R.-H. Xia, “Self-consistent charging in dielectric films under defocused electron beam irradiation”, Micron 42, 443 (2011). [69] J. Joo, B. Y. Chow and J. M. Jacobson, “Nanoscale patterning on insulating substrates by critical energy electron beam lithography”, Nano Lett. 6, 2021 (2006). [70] B. D. Myers and V. P. Dravid, “Variable pressure electron beam lithography (VPeBL): A new tool for direct patterning of nanometer-scale features on substrates with low electrical conductivity”, Nano Lett. 6, 963 (2006). [71] D. R. S. Cumming, I. I. Khandaker, S. Thoms and B. G. Casey, “Efficient diffractive optics made by single-step electron beam lithography in solid PMMA”, J. Vac. Sci. Technol. B 15, 2859 (1997). [72] D. J. Stokes, T. Vystavel and F. Morrissey, “Focused ion beam (FIB) milling of electrically insulating specimens using simultaneous primary electron and ion beam irradiation”, J. Phys. D: Appl. Phys. 40, 874 (2007). [73] P. Peinado, S. Sangiao and J. M. De Teresa, “Focused Electron and Ion Beam Induced Deposition on Flexible and Transparent Polycarbonate Substrates”, ACS Nano 9, 6139 (2015). [74] Y. R. Choi, P. D. Rack, B. Frost and D. C. Joy, “Effect of electron beam-induced deposition and etching under bias”, Scanning 29, 171 (2007). [75] F. Panciera, M. M. Norton, S. B. Alam, S. Hofmann, K. Mølhave and F. M. Ross, “Controlling nanowire growth through electric field-induced deformation of the catalyst droplet”, Nat. Commun. 7, 12271 (2016). [76] H. Plank, C. Gspan, M. Dienstleder, G. Kothleitner and F. Hofer, “The influence of beam defocus on volume growth rates for electron beam induced platinum deposition”, Nanotechnology 19, 485302 (2008). [77] J. M. De Teresa, R. Córdoba, A. Fernández-Pacheco, O. Montero, P. Strichovanec Tuning the growth of 3D nanowires by FEBID 143 and M. R. Ibarra, “Origin of the Difference in the Resistivity of As-Grown Focused-Ionand Focused-Electron-Beam-Induced Pt Nanodeposits”, J. Nanomater. 2009, 936863 (2009). [78] M. Huth, D. Klingenberger, C. Grimm, F. Porrati and R. Sachser, “Conductance regimes of W-based granular metals prepared by electron beam induced deposition”, New J. Phys. 11, 033032 (2009). [79] K. A. G. Rao, N. Nadasen and V. M. Naik, “Theoretical and numerical analysis of non-conventional systems for generating static electric fields: A monograph”, Strategic Book Publishing (2013). [80] J. D. Fowlkes, S. J. Randolph and P. D. Rack, “Growth and Simulation of HighAspect Ratio Nanopillars by Primary and Secondary Electron-Induced Deposition”, J. Vac. Sci. Technol. B 23, 2825 (2005). [81] D. A. Smith, J. D. Fowlkes and P. D. Rack, “A nanoscale three-dimensional Monte Carlo simulation of electron-beam-induced deposition with gas dynamics”, Nanotechnology 18, 265308 (2007). [82] G. Arnold, R. Timilsina, J. Fowlkes, A. Orthacker, G. Kothleitner, P. D. Rack and H. Plank, “Fundamental resolution limits during electron-induced direct-write synthesis”, ACS Appl. Mater. Interfaces 6, 7380 (2014). [83] R. Schmied, J. D. Fowlkes, R. Winkler, P. D. Rack and H. Plank, “Fundamental edge broadening effects during focused electron beam induced nanosynthesis”, Beilstein J. Nanotechnol. 6, 462 (2015). [84] J. Orloff and L. W. Swanson, “An asymmetric electrostatic lens for field-emission microprobe applications”, J. Appl. Phys. 50, 2494 (1979). [85] A. T. J. B. Eppink and D. H. Parker, “Velocity map imaging of ions and electrons using electrostatic lenses: Application in photoelectron and photofragment ion imaging of molecular oxygen”, Rev. Sci. Instrum. 68, 3477 (1997). [86] Y. Lin and D. C. Joy, “A new examination of secondary electron yield data”, Surf. Interface Anal. 37, 895 (2005). [87] M. Futamoto, M. Hanbücken, C. J. Harland, G. W. Jones and J. A. Venables, “Visualization of submonolayer and surface topography by biassed secondary electron imaging: Application to Ag layers on Si and W surfaces”, Surf. Sci. 150, 430 (1985). [88] C. A. Walsh, “Effect of specimen bias on secondary electron images in the STEM”, Ultramicroscopy 45, 85 (1992). [89] T. E. Everhart and R. F. M. Thornley, “Wide-band detector for micromicroampere low-energy electron currents”, J. Sci. Instrum. 37, 246 (1960). [90] B. L. Thiel, M. Toth, R. P. M. Schroemges, J. J. Scholtz, G. Van Veen and W. R. Chapter 3 144 Knowles, “Two-stage gas amplifier for ultrahigh resolution low vacuum scanning electron microscopy”, Rev. Sci. Instrum. 77, 033705 (2006). [91] J. Pablo-Navarro, D. Sanz-Hernández, C. Magén, A. Fernández-Pacheco and J. M. De Teresa, “Tuning shape, composition and magnetization of 3D cobalt nanowires grown by focused electron beam induced deposition (FEBID)”, J. Phys. D: Appl. Phys. 50, 18LT01 (2017). Chapter 4: Towards properties improvement by thermal annealing In this chapter, thermal annealing strategies are followed to overcome the Focused Electron Beam Induced Deposition deficiencies in terms of purity and crystallinity. Different ex situ and in situ post-growth treatments were performed in 3D Co and Fe nanowires to increase the metallic composition, induce a crystallization of the pseudoamorphous as-grown structures and enhance the net magnetic induction values. Dedicated in situ TEM experiments were carried out to monitor in real time changes in the nanowires properties as a function of temperature and time, shedding light on the nanoscale processes involved during the annealing procedure. Chapter 4 146 4.1 Introduction The FEBID fabrication process implies the electron-beam-induced decomposition of a metal-organic precursor gas adsorbed on the growth surface, fundamentally driven by the SE emitted by the substrate or the nascent deposit material [1]. One of the main issues of this synthetic technique is the existence of chemical impurities due to incompletely dissociated precursor gas molecules incorporated into the deposit. In addition, the structures also contain different levels of contaminants originating from the dissociation of residual gases in the working chamber [2]. Despite this, as described in Chapter 3, purity can be improved by precise control of primary electron beam parameters, chamber base pressure, precursor gas flux, etc. [3], which in some cases enables the fabrication of deposits of high metallic content [4][5][6][7][8]. However, in many cases the as-grown metallic content often remains moderate [9]. Thus, the lack of purity and the poor crystallinity will in general be limiting factors for capitalizing the full potential of FEBID. To face this limitation, diverse approaches have been followed to increase the purity of FEBID deposits, as reviewed by Botman et al. in 2009 [2], and further explored since. This includes in situ [10][11] and ex situ [12][13][14] post-annealing treatments at high vacuum and under controlled reactive gas atmospheres [15][16], use of substrates at high temperatures during growth [17][18], electron beam irradiation of the deposits [19][20], laser-assisted heating during deposition [21][22], post-growth Joule heating upon injection of high electric currents [23], supersonic jet delivery of precursor gas [24], use of carbon-free precursor gases [25], and combinations of all these methods [16][26][27][28][29]. For example, astounding success has been achieved in growing virtually-pure Pt deposits by post-growth electron irradiation in O atmosphere [15], or functional Au plasmonic nanostructures by electron beam irradiation in water vapor atmosphere [30]. However, such widely-used strategies have been generally applied to 2D deposits, whereas purification of 3D objects have been scarcely attempted [31]. The Towards properties improvement by thermal annealing 147 latter presents specific problems: for instance, the architecture stability can be compromised due to the high-volume shrinkage occurring in low metal content deposits, eventually leading to a substantial modification of the object shape or even the collapse of the 3D structures [32]. In our case, a special emphasis should be given to magnetic materials grown by FEBID. Specifically, Co- [33], Fe- [34] and Ni-based [35] precursor gases have been used, in which low metallic content results in degraded magnetic properties. As a result, the functionality of FEBID-grown magnetic devices may be compromised. Nevertheless, under optimum growth conditions, Co and Fe deposits are amongst the FEBID materials with highest metallic content levels (~95% at.) [5][36][37][38]. In both cases, as-grown high purity 2D deposits have been achieved with limited lateral resolution and small crystalline size. In the case of 3D magnetic deposits, the dispersion of composition values found in the literature is high, given the strong dependence of the growth mode to small changes in the numerous growth parameters [37][39]. Nonetheless, 3D nanostructures with a high aspect ratio, such as vertical nanowires, evidence a drastic reduction of the metallic content for decreasing diameter. Moreover, the natural oxidation of the surface upon air exposure (typically 5-nm-thick) becomes critical as the surface-to-volume ratio significantly increases and impairs the ferromagnetic properties of the nanostructures due to the formation of non-ferromagnetic species [40][41]. It is particularly important to emphasize that the purity and crystallinity will have great impact in the electrical transport [10] and magnetic properties (saturation magnetization, magnetocrystalline anisotropy, coercive field, etc.) [39][42], both key for the potential application of FEBID magnetic deposits. For all these reasons, purification of 3D FEBID magnetic objects is an exciting challenge and different methods have been explored over the years. The first attempts to purify Fe deposits were carried out by Shimojo et al. through ex situ thermal annealing and electron beam irradiation in ultra-high vacuum, demonstrating the formation of highly-magnetic α-Fe deposits, in some cases coexisting Chapter 4 148 with Fe carbides [13]. The same group subjected 3D nanostructures to post annealing up to 600 ºC, achieving Fe contents near 60% at. [43]. Studies on FEBID Co include the work by Belova et al., who analyzed C-seeded micrometric deposits grown on substrates at 70 ºC, on the verge of the thermal decomposition of the precursor gas, which presented a metal content >95% at. Co [44]. Begun et al. explored the catalytic activity of Co in a H 2 reactive atmosphere upon electron beam irradiation, observing the formation of metallic Co only in a 20-nm-thick surface region of the deposit [45]. The effect of intense electric current on the properties of suspended Co nanowires has also been reported to induce purification from 80% to 89% at. Co and crystallization into large face-centeredcubic (fcc) crystals caused by local Joule heating and electromigration [23]. Recently, post-growth annealing in high vacuum of thin Co stripes increased its composition from 67% to 84% at. Co, with a conductivity of metallic behavior and three orders of magnitude higher than that of the as-grown material [10]. Such previous work has provided hints on how to improve the purity and crystallinity of magnetic deposits grown by FEBID. This chapter will summarize the high-vacuum thermal annealing procedures followed in this thesis to obtain purified and crystalline 3D Co magnetic nanowires from FEBID deposits, maintaining the original shape. In the case of 3D Fe-FEBID nanostructures, the evolution of the morphology and compositional distribution was monitored in real time [31][46]. Such successful postprocessing methods may facilitate the development of future applications based on 3D magnetic nanostructures. 4.2 Annealing process on 3D cobalt nanowires In this section, the attention is drawn to the synthesis of pure and fully-crystalline 3D Co nanowires. The aim is to increase the crystallinity and metallic content under ex situ annealing, simultaneously improving the saturation magnetization value. Moreover, another key objective is to minimize the volume shrinkage, ensuring good mechanical Towards properties improvement by thermal annealing 149 stability of the 3D objects. Thanks to the purification processes, the attention of as-grown FEBID magnetic nanostructures can be focused on tailoring the lateral dimensions and increasing architectural accuracy to produce individual or arrays of customized 3D nanostructures. These can be later ex situ annealed to obtain devices of pure material with optimum magnetic response for the numerous applications foreseen, i.e., magnetic data storage and logic systems, Hall-effect nanosensing, catalytic growth of nanostructures, cantilever functionalization or magnetic nanoactuators [38]. 4.2.1 Experimental details The nanostructures were grown in the commercial Helios Nanolab 650 Dual Beam system using Co 2 (CO) 8 as precursor gas. After the optimization of the parameters, the deposits were fabricated selecting an electron beam voltage of 5 kV, an electron beam current of 100 pA and a chamber growth pressure of 2.6 × 10 -6 mbar (base pressure of 1.3 × 10 -6 mbar). Two different batches were grown under the same main conditions: batch 1 for structural, chemical and magnetic characterization, and batch 2 for nanoSQUID magnetometry. In batch 1, the growth pattern was a single point scanned in spot mode by the electron beam for 45 seconds in standard TEM Cu grids [31][39], whereas in batch 2 the growth was performed for 40-60 seconds onto a 500-nm-thick Si 3 N 4 membrane covered by 10 nm of Al avoiding charging effects for the nanowire which was not annealed and in thinned (few µm-thick) TEM Cu grids for the nanowires devoted to annealing [47]. The ex situ post-growth annealing in high vacuum took place in an SEM Quanta FEG 250 system. The SEM chamber was initially evacuated until the base pressure decreased below 4 × 10 -6 mbar. A heating ramp of 50 ºC/min was programmed, corresponding to the maximum allowed by the equipment, until the target annealing temperature for each sample was reached. Then the samples were annealed at 150 ºC, 300 ºC, 450 ºC and 600 ºC for 100 minutes. The use of metallic grids is paramount to Chapter 6 252 Figure 6.10. Comparison of the MFM images acquired in air and water environments for the commercial Nanosensors PP-MFMR and Team Nanotec tips, and for the functionalized Olympus BioLever mini with an Fe-FEBID nanowire. Given the performance constraints of the commercial probes, experiments with FEBID functionalized tips were performed using dedicated cantilevers for working in liquid media. In particular, the Olympus BioLever mini AFM probe with a FEBID nanowire ensures very good performance, as shown in Figure 6.10 and 6.11, evidencing remarkable improvement of contrast with respect to the standard commercial MFM tips in liquid. In fact, the image quality and sensibility are approximately the same in both air and water. Also, this tip has shown an excellent performance after one year in storage. Current and future applications based on FEBID 253 Figure 6.11. MFM signal profiles obtained from the green lines depicted in Figure 6.10 for each tip case. The FEBID tips allow selecting the most appropriate cantilever to optimize the MFM acquisition, avoiding the lack of adhesion of the magnetic layer to the probe when working in liquid media, controlling the stray field and fabricating deposits with high magnetic induction, which is impossible using the typical sputtering or evaporation methods due to the cantilever geometry. Additionally, an increase of the signal-to-noise ratio can be accomplished by minimizing the van der Waals interaction thanks to a closer approach of the FEBID tip to the sample than the commercial ones. So far, there is no magnetic commercial probe which satisfies these requirements. Chapter 6 254 6.1.5 Magnetic skyrmions observation The magnetic skyrmions are spin textures of nanometric scale which present outstanding potential properties for spintronic applications. These topologicallyprotected quasiparticles were predicted theoretically in the 1960s [21] and are typically stabilized in systems presenting Dzyaloshinskii-Moriya interaction (DMI) and uniaxial magnetic anisotropy [22][23][24][25]. The resulting Bloch and Néel skyrmions can be found in bulk non-centrosymmetric materials [26] or ultrathin films with strong spinorbit coupling in the interface [27][28][29]. To date, the stabilization of either Bloch or Néel skyrmions was limited to systems with intrinsic magnetic anisotropy. However, an investigation conducted in collaboration with the group of Dr. Agustina Asenjo provides an evidence of the stabilization of Néel skyrmions in confined systems with neither DMI nor perpendicular magnetic anisotropy. In particular, the detection of the non-chiral magnetic Néel hedgehog skyrmions was performed at room temperature in soft magnetic sub-100 nm diameter polycrystalline nanodots made of permalloy where, however, either a vortex or a single domain state with IP magnetization was expected due to low magnetocrystalline anisotropy [30][31]. The permalloy nanodots have proven to serve as architectures to form metastable magnetic Néel skyrmions, consisting of rotating the out-of plane (OOP) magnetization component from the core to the boundaries by the spin curling in radial planes [32][33]. This has been demonstrated by analytical calculations, micromagnetic simulations and experiments based on applying external magnetic fields while performing MFM measurements [11]. Despite this remarkable result, the aim of this subsection is not to present a comprehensive study of the hedgehog skyrmions, but to introduce a novel potential in the performance of FEBID MFM tips. A great emphasis is to be given in the detection, observation and stabilization of the skyrmions configuration by the magnetic field arisen from these functionalized tips. Current and future applications based on FEBID 255 Whereas in permalloy nanodisks with 140 nm in diameter MFM images reveal dark (attractive interaction) or bright (repulsive interaction) contrast at the centre, corresponding to the magnetization parallel or antiparallel to the tip polarization; in sub100 nm nanodots the contrast is always bright, evidencing an antiparallel tip-core configuration regardless of the MFM tip polarization, as illustrated in Figure 6.12(a). In a system with diameters below 30 nm, a non-skyrmionic behaviour is exhibited, establishing a radius boundary for the skyrmions stabilization. On the other hand, as shown in Figure 6.13, under IP applied magnetic field the skyrmions core moves parallel or antiparallel to the field until reaching a critical value when the magnetization is completely aligned with the field direction. These experiments proved the existence of a radial IP magnetization component in the nanodots, discarding the vortex system and strongly demonstrating the Néel skyrmionic configuration. In addition, analytical calculations suggest that Néel skyrmions are highly metastable states which can be stabilized in permalloy nanodots in absence of external fields and destroyed by small perturbations. In order to evaluate the stabilization of this singular nanostructures considering the tip-sample interaction, MFM experiments with different types of probe have been carried out as a function of the IP applied magnetic field. Figure 6.12. MFM images performed with an Fe-FEBID tip of (a) nanodots with 70 nm in diameter presenting 100% of skyrmion configuration and (b) nanodots with 30 nm in diameter presenting 22% of skyrmion configuration, the rest being single domain and other different configurations. Chapter 6 256 The experimental results are displayed in Figure 6.13, certifying that the stray field coming from the MFM tip contributes to the stabilization of the skyrmions [34]. The commercial Nanosensors probe presents the highest stray field, the Co-coated tip by sputtering offers an intermediate value and the Fe-FEBID nanowire tip exhibits the lowest one. As can be noted, the saturating field decreases as the stray field produced is reduced. As a result, the stray field of the tip enables the control of the skyrmion stability. In the latter case, a nanowire of 1 µm in length, 30 nm in diameter and a very sharp apex with just 7 nm allows imaging the skyrmions with better resolution and under the lowest external tip invasiveness, maximizing the OOP/IP stray field ratio. Since the application of OOP stray fields serves to tune the stability of the skyrmions, FEBID nanolithography technique can modulate the stabilization providing the stray field customization. Therefore, the use of FEBID functionalized probes unveils a new ability to analyse magnetic skyrmions without perturbating their magnetic state and exploring their magnetization dynamics. Moreover, these tips are very useful to study magnetic textures which are very much sensitive to external perturbations. Figure 6.13. MFM images performed with a commercial Nanosensors probe, a Co-coated tip and an Fe-FEBID nanowire tip. The field sequences show different IP fields depending on the type of tip. All image sizes are 250 × 250 nm 2 . Current and future applications based on FEBID 257 6.2 Engineered 3D cobalt nanowires Shape customization by FEBID entails a great advantage for the fabrication of 3D samples with several features. Apart from the straight structural shape, this technical benefit enables a wide range of architectural configurations despite its difficulty. In this section, this asset will be further exploited to grow and investigate curved ferromagnetic FEBID nanostructures, pursuing the development of devices based on the domain wall motion such as high-density non-volatile memories. 6.2.1 Introduction Although a large variety of shapes and materials can be produced by distinct techniques [35][36][37][38], the architecture of the nano-objects is usually restricted to straight cylindrical or tubular designs [39][40]. The possibility to fabricate 3D nanostructures with many different geometries by FEBID offers a great versatility in terms of shape, areal density and novel magnetic domain configurations [41]. Based on these fundaments and harnessing the polyvalence of this single-step nanolithography technology, engineered 3D Co and Co@Pt nanowires have been designed containing bent segments which can promote the formation of magnetic domain walls. The strategy of curved nanostructures seems reasonable and well-adapted to the objective because the pinning of domain walls thanks to bending sections has already been reported in experimental [42] and theoretical [43] studies, where an increase in pinning situations was detected with the curvature and the angle of the bends [44][45]. The following study aims to grow 3D ferromagnetic nanowires with single or multiple very well-defined bends acting as pinning sites along the length of the nanostructure. The magnetic state will be characterized exploring the capability of such singular shapes to generate specific sites where magnetic domains pointing in different directions meet to form a domain wall. As shown in Figure 6.14, the purpose of this geometry is to trigger the presence of domain walls at remanence after saturating the sample with a magnetic Chapter 6 258 field applied parallel to the substrate and in the plane of the bends. Although this approach was tested in the past in many different systems, e.g., 2D cylindrical curved permalloy nanowires [46], its application in 3D nano-objects still constitutes a major challenge. Regarding the detection of such magnetic configurations, no conventional magnetooptical and MFM methods are fully appropriate for 3D structures. For this reason, OffAxis EH and X-ray magnetic circular dichroism in combination with photoemission electron microscopy (XMCD-PEEM) techniques have been selected [47][48][49][50]. In fact, although previous studies on 3D samples of diverse shapes were already reported using both techniques [51][52][53], shadow XCMD-PEEM had not been performed in vertical nanostructures with such high aspect ratio. Figure 6.14. SEM images of a 3D Co@Pt nanowire (a) before and (b) after the Pt-C coating. (c) Scheme of the IP magnetic field application favouring the domain walls formation at remanence in the areas indicated by red-rimmed circles, where the alternating magnetic charge are denoted. 6.2.2 Experimental details The nanostructures were grown in the commercial Helios Nanolab 650 Dual Beam system using Co 2 (CO) 8 and CH 3 CpPt(CH 3 ) 3 precursor gases. The substrates were TEM Cu grids and Si wafers. The Co nanowires were fabricated with a 5 kV electron beam voltage, a 100 pA electron beam current and a chamber growth pressure of 3.3 × 10 -5 Current and future applications based on FEBID 259 mbar (base pressure of ~4.7 × 10 -6 mbar). Although each segment of the nanowire could be fabricated separately tilting the stage at every stage, here the whole Co nanostructure was grown in one single deposition with the stage remaining in the horizontal position. The Co pattern is composed of 77 points separated 14 nm in a straight line parallel to the flat edge of the Co GIS, keeping constant the precursor molecules flux guaranteeing no shadowing effect [54]. As illustrated in Figure 6.14, the nanowire is formed by seven segments —numbered from bottom to top—, each one having its particular growth strategy. Vertical segments are obtained by scanning a single pattern point while the electron beam stands still. By contrast, bent segments are fabricated by scanning a sequence of 18 points while shifting the electron beam position. For a fixed total horizontal shift and dwell time during the sequence, the angle of the segment with respect to the substrate depends on the number of points. The higher that number, the shorter the distance between two subsequent points, and therefore the overlap between them will be higher and the angle with respect to the substrate will increase. Each bent segment corresponds to 18 pattern points with a scanning time of 97.2 ms. To form the ∼90 degrees bends, the joint between the bent segments is fabricated by a single point scanned for 581.2 ms. Then, the bend is completed by reversing the electron beam shift direction. The first segment was grown by depositing on the first point for 2903.8 ms. Then, the second and the third ones were fabricated scanning the sequence of 18 points for each one, as described above, taking into account that the points of the third one are exactly over those of the second segment but scanned in the reverse direction, thus forming the first bend of the nanostructure. The fourth segment was completed scanning a single point for 1549.1 ms. Then, the fifth and the sixth segments, as well as the joint between them were carried out just as the first bend. The top segment was fabricated scanning the last point during 2419.8 ms. On the other hand, following the same procedure, 3D Co nanowires with only one bend have also been fabricated. Chapter 6 260 For the Co@Pt nanowires, the Pt-C shell was grown immediately after the ferromagnetic core (~65%–70% at. Co) following the process described in Chapter 5 in order to avoid its oxidation [22]. An electron beam voltage of 5 kV and an electron beam current of 100 pA were used, with a chamber growth pressure of 2.4 × 10 -5 mbar. In this case, a polygonal Pt-C pattern was set following the shape of the nanowire core viewed from the perspective shown in Figure 6.14(a). A Pt-C deposition of ~2 s in each side increases the diameter by ~13 nm, as seen in Figure 6.14(b). In order to characterize magnetically the ferromagnetic nanowires, Off-Axis EH and XMCD-PEEM imaging experiments have been performed. In the first one, experiments were performed in the FEI Titan Cube, operated at 300 kV. The excitation of the biprism was adjusted according to the nanowires shape, with a fringe contrast ranging from 20% to 25% and an acquisition time of 5 s. The second technique was carried out in the XPEEM branch of the HERMES beamline (Synchrotron SOLEIL-France) [35] by the group of Dr. Olivier Fruchart. 6.2.3 Magnetic state characterization Firstly, Off-Axis EH experiments were performed on single-bend 3D Co nanowires to check the pinning of domain walls at remanence in the bends after applying a magnetic field in the appropriate direction, i.e., in the plane of the bends and approximately perpendicular to the substrate surface. As illustrated in Figure 6.15, two almost identical nanostructures were used to explore the reproducibility of the experiment. The results show that two different domain walls were found in the first and second bends starting from the bottom, which is evidenced by the contrast change of the magnetic flux lines in the bend and the presence of stray fields in the surrounding vacuum. In fact, the two domain walls can be appreciated in the same places in each nanowire, guaranteeing the replicability of its formation. Current and future applications based on FEBID 261 Figure 6.15. (a,c) SEM images of two different 3D Co nanowires and (b,d) their associated magnetic flux lines distribution. In XMCD-PEEM experiments, the shadow of the nanostructure generated when the beam passes through the specimen provides information for recovering the magnetic configuration of the sample. On the one hand, the magnetic imaging with synchrotron soft X-rays is based on the magnetic circular dichroism, revealing the difference in resonant absorption of left and right circularly-polarized light and obtaining the projection of the magnetization. On the other hand, the collection of photoelectrons coming from the nanostructure during X-ray absorption allows imaging. Specifically, the image contrast, which is proportional to the cosine of the angle between the beam wave vector and the magnetization direction, is used for the reconstruction of the magnetic configuration. As already suggested in Figure 6.14, after applying the magnetic field —parallel to the substrate and in the bends plane—, three domain walls are expected to be nucleated on double-bend 3D Co nanowires. Figure 6.16 shows the correlation between the shadow features and the magnetic state of the segments. Whereas the strong field emission at the nanowire tip leads to a reduction of the signal-to-noise ratio, the top segments (4, 5 and 7) are clearly identified in Figure 6.16(b), far away from the nanowire position. The lack Chapter 6 268 5, 7781 (2011). [48] J. Kimling, F. Kronast, S. Martens, T. Böhnert, M. Martens, J. Herrero-Albillos, L. Tati-Bismaths, U. Merkt, K. Nielsch and G. Meier, “Photoemission electron microscopy of three-dimensional magnetization configurations in core-shell nanostructures”, Phys. Rev. B 84, 174406 (2011). [49] R. Streubel, V. P. Kravchuk, D. D. Sheka, D. Makarov, F. Kronast, O. G. Schmidt and Y. Gaididei, “Equilibrium magnetic states in individual hemispherical permalloy caps”, Appl. Phys. Lett. 101, 132419 (2012). [50] S. Jamet, S. Da Col, N. Rougemaille, A. Wartelle, A. Locatelli, T. O. Menteş, B. Santos Burgos, R. Afid, L. Cagnon, S. Bochmann, J. Bachmann, O. Fruchart and J. C. Toussaint, “Quantitative analysis of shadow x-ray magnetic circular dichroism photoemission electron microscopy”, Phys. Rev. B 92, 144428 (2015). [51] D. Shindo, “Electron Holography of Nanocrystalline Magnetic Materials”, Mater. Trans. 44, 2025 (2003). [52] R. Streubel, P. Fischer, F. Kronast, V. P. Kravchuk, D. D. Sheka, Y. Gaididei, O. G. Schmidt and D. Makarov, “Magnetism in curved geometries”, J. Phys. D: Appl. Phys. 49, 363001 (2016). [53] V. Rouco, R. Córdoba, J. M. De Teresa, L. A. Rodríguez, C. Navau, N. Del-Valle, G. Via, A. Sánchez, C. Monton, F. Kronast, X. Obradors, T. Puig and A. Palau, “Competition between Superconductor-Ferromagnetic stray magnetic fields in YBa 2 Cu 3 O 7-x films pierced with Co nano-rods”, Sci. Rep. 7, 5663 (2017). [54] L. Keller and M. Huth, “Pattern generation for direct-write three-dimensional nanoscale structures via focused electron beam induced deposition”, Beilstein J. Nanotechnol. 9, 2581 (2018). [55] C. Schwalb, GETec Microscopy, Private Communication (2019). Chapter 7: General conclusions and outlook In this chapter, a global overview of the key results presented in this thesis and the main conclusions emanated from them are presented, together with the promising prospects for the future research on 3D FEBID magnetic nanostructures in the framework of Nanomagnetism. Chapter 7 270 Since the emergence of Nanotechnology and the multitude of applications derived from it, intensive efforts have been devoted to fabricating functional nanostructures. Specifically, magnetic nanostructured materials attract particularly keen interest because of their would-be implementation in key information and communications technologies, such as data storage, logic and sensing devices [1]. Accordingly, the development of novel nanofabrication techniques or the refinement of the existing methods is one of the essential cornerstones for the growth of advanced nano-objects, allowing the subsequent study of physical phenomena at nanometric scale [2]. In this light, Focused Electron Beam Induced Deposition (FEBID) arises as one of the most versatile nanofabrication technologies which could play a crucial role in the production of several types of architectures and materials in the nanoscale [3]. During the last decades, this single-step lithographic method has been widely used for the growth of two-dimensional deposits, creating magnetic nanostructures considered as promising candidates for the development of forthcoming spintronic applications. However, the increasing demand for high-density and low-power nanodevices naturally entails the expansion to three-dimensional (3D) deposits [4]. As a result, vertical magnetic nanostructures are currently a central topic in nanomagnetism. The functionality of these potential building blocks for 3D magnetic devices relies on the precise control of domain wall motion with spin-polarized currents or magnetic fields which, combined with the high areal density allowed by 3D architectures, is bound to boost their operational performance. The overarching objective of this thesis involves the synthesis and characterization of 3D ferromagnetic nanowires, intending to shed light upon novel and advanced optimization processes which help to improve the operational behaviour of these nanostructures. The main achievements presented in this manuscript are described below. General conclusions and outlook 271 7.1 Tailoring of 3D nanowires grown by FEBID Firstly, the FEBID versatility has been exploited to adjust the thickness, composition and magnetic induction of 3D Co nanowires. In this regard, nanostructures with doublesection diameters, different shapes and chemical contents have been investigated, providing insight on the linear and radial growth regimes. The interpretation of the results indicates that thermal and diffusion effects are ultimately responsible for these growth modes, helping in the progress towards future upgraded nanofabrication strategies. As a practical example, diameter changes or bent shapes are natural locations for magnetic domain-wall pinning with applications in magnetic storage and logics. Standing by these arguments, simultaneous high metallic content (~80% at.), small diameters (<100 nm) and high magnetic induction (~1 T) have been obtained for nanowires grown under optimized conditions. These results are a forward step to the goal of driving FEBID towards a remarkably practical nanolithography technique to fabricate 3D functional nanostructures with unique lateral resolution. Along this line, Co and Fe nanowires with very sharp apex have been analysed and found to be ideal architectures for Magnetic Force Microscopy measurements. Secondly, the difficulty of FEBID to fabricate nanostructures on insulating substrates due to severe charging effects has been overcome by using the ARchitectural Adjustment by Grid Overlay Nanotechnology (ARAGON) Chip, an electrically-biased patterned metal structure which enables charge evacuation during the growth of nanoobjects. In addition, this novel approach has been proven to serve as a new knob for in situ modulation of the nanowire geometry, by dint of the application of spatiallydependent electric fields, acting as an electrostatic lens on the trajectories of the primary and secondary electrons. Therefore, the in situ modification of the applied voltage or the focus height offers a new route to create 3D functional complex nanostructures with tailored lateral dimensions. These findings have enabled not only a better understanding of the growth mechanisms by FEBID, but also hinting new approaches for the Chapter 7 272 implementation of the ARAGON-Chip. For instance, instead of patterning the holed metal plate onto the sample, the ARAGON-Chip could be redesigned to be inserted automatically in the vicinity of the substrate as an aperture in the Scanning Electron Microscopes, with different hole geometries to produce custom distributions of electric fields. 7.2 Annealing treatments to optimize nanowire properties The crystallinity, composition and magnetic induction of 3D ferromagnetic Co and Fe nanowires differs greatly from the bulk materials mainly due to the presence of impurities coming from the incompletely dissociation of the precursor gas molecules. In order to remove these undesired byproducts, post-growth high-vacuum annealing processes have been performed. In the case of Co, ex situ thermal annealing at 600 ºC has been found to produce purified and crystalline nanowires with diameters below 90 nm, a metallic content above 95% at., and a net magnetic induction up to 1.6 T, near the bulk Co. The combined effect of contaminants migration to the surface and recrystallization of the as-grown nanocrystalline structure gives rise to nanowires with physical properties close to bulk ones. Besides, given the relatively high metal content of the as-grown deposits (~70% at.), the changes in shape of the deposits after purification is minimal, facilitating their functional implementation in 3D devices. In the case of Fe, real-time monitoring of the chemical purification and structural crystallization processes of ultra-narrow nanowires (<50 nm in diameter) has been carried out by in situ annealing in a Transmission Electron Microscope. The heating up to 700 ºC reveals local increases of the metallic content along the nanowire length concomitant with the growth of large Fe single crystals from initially pseudo-amorphous compounds with just ~40% at. of Fe. Besides, a reduction of the diameter down to ~30 nm has been achieved in the highest metallic regions. This evolution tracking has provided insight into General conclusions and outlook 273 the nanoscale processes involved during the annealing treatment, serving as a future ideal method for determining the minimum requirements of purity and geometry for as-grown nanowires and the appropriate thermal conditions to be successfully annealed. Along this line of thought, future advances in the purification strategies could point towards exploring further crystallization upon longer and precise annealing times with the aim of routinely producing high-quality single-crystalline nanowires or the use of reactive atmosphere to etch the remaining contaminants at the surface [5]. The same route could be used to explore the possibility of annealing bimetallic or heterogeneous systems and the potential for temperature-induced alloying. 7.3 Dual purpose of the core-shell architectural approach On the one hand, the natural surface oxidation of the ferromagnetic nanostructures to a non-ferromagnetic material (~5 nm in thickness) implies the degradation of the magnetic properties of 3D Co and Fe nanowires, which becomes critical at the smallest diameters. This negative impact can be palliated by growing a 10-20 nm-thick Pt-C protective shell, retaining the original net magnetic induction. This improvement with respect to the uncoated nanowires can reach 35% for the thinnest nanowires (<40 nm core diameters), when the surface oxidized layer greatly contributes to the total nanowire diameter. This strategy demonstrates that the operability of nanostructured objects often relies on the combination of more than one material, enhancing the performance of the device or conferring the desired functionality. On the other hand, 3D ultra-thin Co nanotubes have also been synthesized on Pt-C templates adapting the core-shell approach to produce Pt@Co nanowires. Dimensional, compositional and magnetic characterization has proved the ferromagnetic behavior of the Co nanotubes and allowed studying their magnetization state and dynamics. The switching mechanism is governed by the domain-wall formation and propagation, having a strong and potential impact in the functionality of devices composed by these Chapter 7 274 nanostructures. In particular, ferromagnetic nanotubes are of tremendous interest for investigations of fast-propagating magnetic domain walls, which is a too-demanding requirement for the development of applications within Nanomagnetism. References [1] S. S. P. Parkin, M. Hayashi and L. Thomas, “Magnetic domain-wall racetrack memory”, Science 320, 190 (2008). [2] M. Vázquez, “Magnetic Nanoand Microwires. Design, Synthesis, Properties and Applications”, Elsevier (2015). [3] M. Huth, F. Porrati, C. Schwalb, M. Winhold, R. Sachser, M. Dukic, J. Adams and G. Fantner, “Focused electron beam induced deposition: A perspective”, Beilstein J. Nanotechnol. 3, 597 (2012). [4] A. Fernández-Pacheco, R. Streubel, O. Fruchart, R. Hertel, P. Fischer and R. P. Cowburn, “Three-dimensional nanomagnetism”, Nat. Commun. 8, 15756 (2017). [5] E. Begun, O. V Dobrovolskiy, M. Kompaniiets, R. Sachser, C. Gspan, H. Plank and M. Huth, “Post-growth purification of Co nanostructures prepared by focused electron beam induced deposition”, Nanotechnology 26, 075301 (2015). Conclusiones generales y perspectivas 275 Conclusiones generales y perspectivas En este capítulo se presenta un resumen de los principales resultados obtenidos e incluidos en este documento. Asimismo, de forma global se recogen las conclusiones generales surgidas de esta tesis doctoral, acompañadas de unas perspectivas prometedoras y con gran potencial para futuras investigaciones con estructuras 3D fabricadas por deposición inducida por haz focalizado de electrones dentro del marco del Nanomagnetismo. Chapter 7 276 Desde el surgimiento de la Nanotecnología y de las múltiples aplicaciones derivadas de este campo, se han dedicado numerosos esfuerzos para la fabricación de nanoestructuras funcionales. Específicamente, los materiales magnéticos nanoestructurados atraen un gran interés por sus potenciales implementaciones en tecnologías de la información y la comunicación, tales como los dispositivos dedicados al almacenamiento, lógica y detección magnéticos [1]. En este sentido, el desarrollo de nuevas técnicas de nanofabricación o el perfeccionamiento de los métodos ya existentes es una de las piedras angulares para el crecimiento de nano-objetos avanzados, permitiendo el consiguiente estudio de los fenómenos físicos a escala nanométrica [2]. En este contexto, la deposición inducida por haz focalizado de electrones (FEBID) surge como una de las técnicas de nanofabricación más versátiles, la cual puede desarrollar un papel crucial en la producción de estructuras con diversas formas y materiales en la nanoescala [3]. Durante las últimas décadas, este método de nanolitografía se ha utilizado ampliamente para el crecimiento de depósitos en dos dimensiones, creando nanoestructuras magnéticas consideradas como candidatas potenciales para el desarrollo de los futuros dispositivos espintrónicos. Sin embargo, la creciente demanda para generar métodos de almacenamiento de información en nanodispositivos de alta densidad y con menor consumo de energía conlleva de forma natural a investigar depósitos en tres dimensiones (3D) [4]. Como resultado, las nanoestructuras magnéticas verticales se erigen actualmente como unos de los temas centrales dentro del Nanomagnetismo. La funcionalidad de estas piezas básicas para los futuros dispositivos magnéticos en 3D recae en la posibilidad de controlar de forma precisa el movimiento de las paredes de dominio, bajo la aplicación de corrientes de espín polarizadas o de campos magnéticos que, combinados con la mayor densidad de área ofrecida por las estructuras 3D, permite aumentar su rendimiento [2]. El objetivo principal de esta tesis doctoral integra la síntesis y caracterización de nanohilos ferromagnéticos en 3D, tratando de arrojar luz sobre nuevos y avanzados Conclusiones generales y perspectivas 277 procesos de optimización que ayuden a mejorar el comportamiento operacional de estas nanoestructuras. A continuación, se describen los principales resultados presentados en este documento. 7.1 Diseño de nanohilos 3D crecidos por FEBID En primer lugar, la versatilidad de la técnica FEBID ha sido aprovechada para el ajuste del espesor, composición e inducción magnética de nanohilos de Co en 3D. En este sentido, se han investigado nanoestructuras de diámetro variable, con diferentes formas y contenido metálico, proporcionando un conocimiento y comprensión más profundos sobre los regímenes de crecimiento lineal y radial. La interpretación de los resultados indica que los efectos térmicos y de difusión son los responsables de los diferentes modos de crecimiento en último término, suponiendo un impulso para la mejora y actualización de las futuras estrategias de crecimiento mediante esta técnica. Como ejemplo práctico, cabe destacar que las regiones donde se modifica el diámetro son sitios naturales de anclaje para las paredes de dominio, con aplicaciones en almacenamiento y lógica magnéticos. En base a estos argumentos, bajo condiciones optimizadas, se han obtenido nanohilos con alto contenido metálico (~80% at.), reducido diámetro (<100 nm) y altos valores de inducción magnética (~1 T). Estos resultados suponen un paso adelante para el objetivo de situar a la técnica FEBID como un método de litografía extraordinario para la fabricación de nanoestructuras funcionales en 3D con resoluciones laterales únicas. En esta línea, se ha comprobado que nanohilos de Co y Fe con puntas muy afiladas son estructuras ideales para medidas en Microscopía de Fuerza Magnética. En segundo lugar, la dificultad de la técnica FEBID para la fabricación de nanoestructuras sobre sustratos aislantes debido a los severos efectos de carga, ha sido vencida mediante el uso del ARchitectural Adjustment by Grid Overlay Nanotechnology (ARAGON) Chip, es decir, un chip formado por una estructura metálica con forma de Annex A 284 Figure A.1. Examples of electron holograms of different 3D Co nanowires: (a) as-deposited and annealed at (b) 150 ºC, (c) 300 ºC, (d) 450 ºC and (e) 600 ºC. Whereas both holograms include the same electrostatic phase shift contribution, the magnetic contribution changes its sign. After the addition and subtraction of the holograms, the two solutions obtained always cancelled one of the phase contributions giving rise to an image with two times the other one. As a result, the electrostatic and magnetic phase shift images can be plotted separately, as shown in Figure A.2. The gradient of the phase shift can be calculated as: ∇    ,   = ∇     ,   + ∇     ,   = (A.2) _ _ _ _ _ _ _ _ _ _  =        ,   −  ℏ     ,      ,   where   =   ∗ ⁄ is a constant which depends on the energy of the incident electron beam with a value of   = 6.53 × 10 6 rad·V -1 ·m -1 at 300 kV,   is the mean inner potential and has a value of 26 V for pure Co [7] and 17.4 V for Fe [8], t is the magnetic thickness of the specimen, e is the electron charge, ћ is the reduced Planck constant, x and y are orthogonal equivalent directions in the plane of the sample and   the magnetic induction component perpendicular to one of the in-plane components and the electron beam direction, z. Electron Holography data processing 285 Figure A.2. (a) Electrostatic and (b) magnetic phase shift images extracted from a 3D Co nanowire. (c) Profiles of both types of phase shifts along the green arrows averaged across a length of 58 nm. Since the nanostructures are not usually composed by a pure material,   value does not correspond to that of the bulk. Thus, considering the first term of the right-hand side of Equation A.2, the electrostatic phase shift image can be used to calculate   , knowing   value and the nanowire thickness, assuming that the nano-object is cylindrical and chemically homogeneous. Then, attending to the second term of the righthand side of Equation A.2, the magnetic induction can be calculated by straightforward Annex A 286 mathematical operations using the gradient of the magnetic phase shift image considering one of the two equivalent directions, x and y. For instance, in the case of the x axis: |     ,   | = ℏ  ·  #     ,   # (A.3) Figure A.2 simultaneously demonstrates the cylindrical shape of the nanostructure (  profile) and shows the change of the   which proves the ferromagnetic behavior of the nanowire. Indeed, the steeper the slope, the higher the magnetic induction. Finally, the magnetic flux representation can be illustrated normalizing the magnetic phase contribution to the maximum thickness for a better comparison and calculating the cosine of n times the magnetic phase shift. Figure A.3. Magnetic induction flux representations of different 3D Co nanowires: (a) asdeposited and annealed at (b) 150 ºC, (c) 300 ºC, (d) 450 ºC and (e) 600 ºC. References [1] M. J. Hÿtch, E. Snoeck and R. Kilaas, “Quantitative measurement of displacement and strain fields from HREM micrographs”, Ultramicroscopy 74, 131 (1998). [2] L. A. Rodríguez, C. Magén, E. Snoeck, C. Gatel, L. Marín, L. Serrano-Ramón, J. L. Prieto, M. Muñoz, P. A. Algarabel, L. Morellon, J. M. De Teresa and M. R. Ibarra, “Quantitative in situ magnetization reversal studies in Lorentz microscopy and electron holography”, Ultramicroscopy 134, 144 (2013). [3] H. Lichte and M. Lehmann, “Electron holography-basics and applications”, Rep. Prog. Phys. 71, 016102 (2008). [4] A. Tonomura, T. Matsuda, J. Endo, T. Arii and K. Mihama, “Holographic interference electron microscopy for determining specimen magnetic structure and thickness distribution”, Phys. Rev. B 34, 3397 (1986). Electron Holography data processing 287 [5] J. C. Loudon, N. D. Mathur and P. A. Midgley, “Charge-ordered ferromagnetic phase in La 0.5 Ca 0.5 MnO 3 ,” Nature 420, 797 (2002). [6] R. E. Dunin-Borkowski, M. R. McCartney, B. Kardynal and D. J. Smith, “Magnetic interactions within patterned cobalt nanostructures using off-axis electron holography”, J. Appl. Phys. 84, 374 (1998). [7] M. De Graef, N. T. Nuhfer, and M. R. McCartney, “Phase contrast of spherical magnetic particles”, J. Microsc. 194, 84 (1999). [8] T. Fujita, M. Chen, X. Wang, B. Xu, K. Inoke and K. Yamamoto, “Electron holography of single-crystal iron nanorods encapsulated in carbon nanotubes”, J. Appl. Phys., 101, 014323 (2007). Annex B: Mechanical properties of 3D cobalt nanowires B.1 Introduction Nanowires have also a high interest because of their potential applications in electromechanical devices [1][2]. Although the mechanical characterization of these nanostructures is not straightforward, it is critically important to understand their performance behaviour and determine their mechanical properties. For this purpose, the fabrication and characterization of suspended 3D Co nanowires grown by FEBID have been carried out. Several approaches are used for the mechanical analysis such as AFM nanoindentation [3], contact resonance AFM [4], AFM bending [5], in situ SEM resonance [6] and tension [7] or in situ TEM resonance [8] and tension [9]. Here, the three-point bending method by AFM is proposed [10][11], consisting of the bending of a double-clamped suspended nanostructure. On this basis, horizontally-suspended doubleclamped Co nanowires have been synthesized with diameters between ~55 and ~75 nm and a suspended length of 1.2 µm. Then, the Young modulus and the yield strength have been investigated. B.2 Experimental details The nanowires were fabricated in the commercial Helios Nanolab 650 Dual Beam system using Co 2 (CO) 8 as precursor gas. The substrate consists of arrays of microtrenches between Si pads separated 1.2 µm, patterned by optical photolithography and reactive ion etching. The suspended nanowires were grown acting as a bridge between the pads, as shown in Figure B.1. Annex B 290 The deposits were fabricated with an electron beam voltage of 5 kV, an electron beam current of 100 pA and a chamber growth pressure of ~2.3 × 10 -5 mbar (base pressure of ~1.5 × 10 -6 mbar). The pattern was formed by an array of 120 points, separated 15 nm between each other, connecting two Si pads. The electron beam scanned only once from one edge to the other with a dwell time of 1 µs. To compensate the natural inclination of the nanowire with respect to the horizonal axis due to the overlap of the points sequence, the substrate was tilted ~35-45 degrees with respect to the horizontal axis to make the connection possible. EDS experiments after the natural oxidation of the nanostructures revealed chemical compositions of ~53% at. Co, ~20% at. C and ~27% at. O. The mechanical characterization of the nanowires was performed using a Dimension ICON AFM from Bruker. The bending tests were carried out applying a force in the midpoint of the nanowire and measuring its displacement. The spring constants of the cantilevers were 3 and 20 N·m -1 , with a tip approaching speed from 100 to 500 nm·s -1 . Figure B.1. Schematic diagram illustrating the FEBID process for the fabrication of 3D horizontally-suspended Co nanowires. Noting that during the process the stage was tilted ~3545 degrees with respect to the horizontal position. The inset shows an SEM image of a 3D Co nanowire. Mechanical properties of 3D cobalt nanowires 291 B.3 Results and discussion The bending experiment was performed with an AFM tip at the midpoint of the nanowire, undertaking a load-unload process to obtain the relationship between the applied force and the displacement of the central part of the nanostructure. Figure B.2(a) shows the typical curve acquired below the fracture load point. The overlap between the load and unload data reveals the absence of hysteresis. In addition, AFM and SEM inspections performed after these bending experiments did not indicate any failure of the clamping sites. To determine the Young modulus, , and the yield strength,   , the three-point bending method was carried out until the nanowires fracture. Figure B.2(b) plots the mechanical behaviour for a Co nanowire, where a linear trend with pure elastic behaviour can be identified in the small deflection range and tends to grow when increasing deflections [12] before the fracture, evidenced by a decrease of the applied force. Figure B.2. (a) AFM image of a 3D Co nanowire, representing the bending force as a function of the displacement, where the load (red) and unload (blue) curves are shown. (b) Plot of the bending experiment where the linear region and the fracture are highlighted. The inset shows an AFM image of the Co nanowire after the fracture. Undefined colour scale range and scale bars in all images are 100 and 200 nm, respectively. Annex B 292 Firstly, from the linear region, considering the apparent elastic constant,   , i.e., the slope of the experimental curves,  can be calculated according to the following equation, corresponding to a double-clamped nanowire with diameter, , and suspended length,  [13][14]:  =     3    (B.1) where   is calculated from the linear fit of the bending curves. The experimental value of  ranges from (58 ± 5) GPa to (209 ± 24) GPa for nanowires between ~55 nm and ~75 nm in diameter, noticing a slight association of the higher  values with the smaller diameters. Thus, these result shows that the  values for the narrower nanowires are very close to the bulk one (~209 GPa), and the values obtained for the wider nanowires are comparable to the ones reported for polycrystalline Co oxide nanowires [10]. Secondly, the subsequent non-linear behaviour is comparable to that reported for double-clamped Si [14], Au [15] and ZnO [16]. When the displacement begins to not be comparable to the nanowire radius, the axial tension due to stretching dominates over the radial one and the curve departs from the linear behaviour. Thirdly, the applied force reduction evidences the nanowire’s plastic deformation and, finally, the fracture. Along this line, the yield strength,   , defined as the stress at which the material starts to be deformed plastically, can be estimated using the following expression [15]:   = 4       (B.2) where   is the yield force, corresponding to the maximum load immediately before the fracture. An average   of (6.1 ± 2.3) GPa was found, not far from the theoretical one (  ~0.1) [17], but significantly greater than the bulk Co value (  ~345-485 MPa). Mechanical properties of 3D cobalt nanowires 293 The fracture occurs typically by the emergence of fissures originated at defects [18]. Since the probability of producing a fissure decreases with reducing the material size [15], nanostructures are expected to undergo higher stress than bulk materials. Therefore, the higher   values obtained for nanowires can be explained by the limited number of defects. Additionally, differences in the microstructure and composition of the Co nanowires fabricated by FEBID with respect to the bulk can also contribute towards improving   , e.g., noticing that the small grain sizes present in the nano-objects hampers the formation of defects. Similar results have been obtained by us on the mechanical properties of suspended W-C nanowires grown by FIBID and investigated using the same experimental technique [19]. B.4 Conclusions A novel nanofabrication method for the growth of horizontally-suspended 3D Co nanowires by FEBID has been presented. The mechanical characterization performed by the three-point bending method shows good mechanical performance of the nanostructures with Young modulus values comparable to the bulk material ones, and exhibiting a yield strength 15 times higher than the bulk one. The small lateral resolution and the microstructure of these 3D nanostructures confers the nanowires outstanding mechanical properties which provide robustness, potentially constituting the basis for the design of future advanced nano-mechanical devices. References [1] X. L. Feng, R. He, P. Yang and M. L. Roukes, “Very High Frequency Silicon Nanowire Electromechanical Resonators”, Nano Lett. 7, 1953 (2007). [2] P. Vavassori, M. Pancaldi, M. J. Perez-Roldan, A. Chuvilin and A. Berger, “Remote Magnetomechanical Nanoactuation”, Small 12, 1013 (2016). [3] X. Li, H. Gao, C. J. Murphy and K. K. Caswell, “Nanoindentation of Silver Nanowires”, Nano Lett. 3, 1495 (2003).