This dissertation describes the single-molecule magnet (SMM) behaviour of f-block cyclooctatetraenyl sandwich complexes. Chapter one introduces the concepts that dictate SMM behavior particularly in f-elements. The emphasis is to understand the origin of magnetic behaviour and the properties that make lanthanide elements particularly interesting to explore. Current strategies used to predict such behaviour are discussed and a literature review on the subject is provided. Chapter Two describes the magnetic properties of eight isostructural lanthanide sandwich complexes utilizing 1,4-bis(trimethylsilyl)cyclooctatetraenyl dianion as the ligand, [Li(DME)3][LnIII(COT”)2] (Ln = Ce, Nd, Gd, Tb, Dy, Ho, Er, Tb, COT” = 1,4-bis(trimethylsilyl)cyclooctatetraenyl dianion, DME = dimethoxyethane). The complexes display a wide range of magnetic behaviour. The best performing SMM was the erbium complex, which had a blocking temperature of 8 K. Investigating different lanthanide ions with the same ligand enabled us to evaluate our findings in relation to current models used to predict SMM behaviour in lanthanide complexes. Chapter three extends the discussion of lanthanide sandwich complexes to include higher symmetry cyclooctatetraenyl complexes of ErIII and DyIII, [K(18-C-6)][LnIII(COT)2] (18-C-6 = 1,4,7,10,13,16-hexaoxacyclooctadecane, COT = cyclooctatetraene).The change in symmetry evoked by removing the trimethylsyl- (TMS) groups on the ligand greatly influenced the magnetic properties of both complexes. Ab initio calculations revealed that the magnetic relaxation in the ErIII complex occurs via the second excited state which contributes to the very high blocking temperature of 10 K in this complex. Chapter four presents an organometallic building block approach to create triple decker lanthanide COT” complexes of GdIII, DyIII and ErIII with a molecular formula of LnIII2(COT”)3. Synthetically, we couple together the sandwich complexes discussed in Chapter 2 by oxidatively removing one ligand to produce linear complexes where the two metals are bridged by an aromatic COT” ligand. The magnetic properties of all complexes are compared to their respective mononuclear analogs. Most interesting is the unprecedented 4 K increase in blocking temperature of the triple decker ErIII analog compared to the ErIII mononuclear sandwich complex discussed in Chapter 2. This increase is due to a ferromagnetic dipole-dipole interaction between the ErIII ions through the COT” ring. The aromatic bridging ligand provides a GdIII - GdIII interaction of J = -0.448(1) cm-1. Chapter five extends the discussion of magnetic exchange coupling to include linear K2(THF)4[LnIII2(COT)4] (Ln = Gd, Dy, Er, COT = cyclooctatetraenyl dianion, THF = tetrahydrofuran) complexes of GdIII, DyIII and ErIII. Each complex is composed of two LnCOT2 units bridged linearly by a potassium ion. The magnetic interaction between metal ions is much weaker than in the triple decker complexes discussed in Chapter 4, with a GdIII-GdIII interaction of J = − 0.007(4) cm–1. The magnetic properties of the quadruple decker complexes were compared to their mononuclear equivalents (Chapter 3). Surprisingly, the ErIII complex showed an increase in magnetic blocking temperature over its mononuclear analog despite the large ErIII-ErIII separation of 8.819 Å. Ab initio calculations revealed that this increase is due to single ion effects, most likely an increase in symmetry. Chapter six deviates from lanthanide magnetism to study the magnetic properties of uranium sandwich complexes with multiple ligand systems and oxidation states. Prior to this study the SMM behaviour of uranium sandwich complexes was unknown. We report the synthesis, structure and magnetic properties of both uranium-COT” sandwich complexes and uranium-cycloheptatrienyl complexes with oxidation states spanning (III)-(V). None of the complexes showed zero-field SMM behaviour, indicating a sandwichtype ligand is not appropriate for harnessing the SMM character in uranium. We compared the slow magnetic relaxation of isostructural and valence isoelectronic uranium and neodymium complexes. The improved energy barrier in the uranium complex further motivates the use of uranium in SMM design due to its large spin-orbit coupling.

Les applications d’électronique de spin telles les mémoires à accès aléatoire (MRAM), les capteurs (e.g.les têtes de lecture des disques durs d’ordinateurs) et les éléments de logique magnétique utilisent les interactionsd’échange ferromagnétique/antiferromagnétique (F/AF) dans le but de définir une direction de référence pour lespin des électrons de conduction. Les MRAM à écriture assistée thermiquement (TA-MRAM) utilisent mêmedeux bicouches F/AF : une pour le stockage de l’information et l’autre comme référence. Ces dernièresapplications technologiques impliquent des étapes de nanofabrication des couches minces continues pour formerune matrice de cellules mémoires individuelles. La qualification industrielle du produit final impose de sérieusescontraintes sur la largeur des distributions des propriétés magnétiques - y compris d’échange F/AF - de cellulemémoire à cellule mémoire. Des phases verres de spin réparties de manière aléatoire sur la couche continue, àl’interface F/AF ou dans le coeur de l’AF pourraient contribuer de manière significative à ces distributionsd’échange F/AF dans les dispositifs, après nanofabrication ; comme supposé il y a de cela quelques années. Lebut de cette thèse est d’étudier factuellement le possible lien entre verre de spin répartis dans des couches mincesF/AF et dispersions de propriétés d’échange de cellule mémoire à cellule mémoire dans les dispositifs TAMRAMcorrespondants. Avant cela, l’origine de ces régions verre de spin a été étudiée et une attention plusparticulière a été portée au rôle joué par les diffusions de Mn. Ces dernières ont été directement observées,comprises et l’utilisation de barrières complexes pour les réduire et par là même pour diminuer la quantité dephases verre de spin a été mise en oeuvre avec succès. En guise d’alternative pour varier la quantité de verres despin, l’utilisation d’AFs composites a été également étudiée dans le cadre de cette thèse. Ce type d’AF permet dumême coup de varier la stabilité thermique des grains AF et de répondre à un autre problème identifié pour lesTA-MRAM qui consiste à trouver des matériaux AF avec des propriétés de rétention et d’écriture intermédiairespar rapport aux matériaux actuellement utilisés. Finalement, ces AFs composites ont été utilisés comme moyende varier la quantité de verres de spin dans des dispositifs TA-MRAM réels et de prouver le lien direct avec ladispersion de propriétés de cellule mémoire à cellule mémoire. / Spintronics applications such as magnetic random access memories (MRAM), sensors (e.g.. hard diskdrive read head) and logic devices use ferromagnetic/antiferromagnetic (F/AF) exchange bias (EB) interactionsto set the reference direction required for the spin of conduction electrons. Thermally-assisted (TA-) MRAMapplications even use two F/AF exchange biased bilayers: one for reference and one for storage. Suchtechnological applications involve patterning full sheet wafers into matrix of individual bit-cells. Industrialproducts qualification imposes stringent requirements on the distributions of the magnetic properties from cell tocell, including those related to EB. It was supposed few years ago that randomly spread spin-glass like phases atthe F/AF interface or within the bulk of the AF layer significantly contribute to the distributions of EB propertiesin devices after processing. This thesis aimed at factually studying the link between spin-glasses spread overF/AF thin films and bit-cell dispersions of EB in corresponding TA-MRAM. Prior to that the origin of the spinglasslike regions and more specifically the role of Mn-diffusion was consolidated. Mn-diffusion was directlyobserved, understood and the use of complex barriers to reduce such diffusion and consequently to minimize theamount of spin-glass was successfully studied. Mixing AFs as another way to tune the amount of spin-glass likephases was also evidenced in the framework of this thesis. All at once, this last solution also tuned the AF grainsthermal stability and this solved another issue related to TA-MRAM, i.e. finding AF-materials with intermediateretention and write properties compared to the AFs presently used. Finally, those mixed antiferromagnets werethe mean chosen to tune the amount of spin-glass like phases in real TA-MRAM devices and to factually provetheir link with bit-cell distributions of EB properties.

Ni80Fe20 and Ni80Fe20/NiO films and nano-stripes were characterized magnetically through AC and DC susceptibility measurements, and hysteresis loops as a function of field and temperature. While the near-pattern films were characterized in the in-plane configuration only, the nano-stripes were characterized in parallel, transverse and the perpendicular field configurations. The effects of the constrained geometry on the coercivity, exchange bias field, and the superparamagnetic blocking temperature were studied. It was determined that the coercivity, exchange bias field and the superparamagnetic blocking temperature can be controlled, not only by using a patterned media instead of a plane film, but also by the orientation of that pattern.

Ni80Fe20 and Ni80Fe20/NiO films and nano-stripes were characterized magnetically through AC and DC susceptibility measurements, and hysteresis loops as a function of field and temperature. While the near-pattern films were characterized in the in-plane configuration only, the nano-stripes were characterized in parallel, transverse and the perpendicular field configurations. The effects of the constrained geometry on the coercivity, exchange bias field, and the superparamagnetic blocking temperature were studied. It was determined that the coercivity, exchange bias field and the superparamagnetic blocking temperature can be controlled, not only by using a patterned media instead of a plane film, but also by the orientation of that pattern.

Les applications d’électronique de spin telles les mémoires à accès aléatoire (MRAM), les capteurs (e.g.les têtes de lecture des disques durs d’ordinateurs) et les éléments de logique magnétique utilisent les interactionsd’échange ferromagnétique/antiferromagnétique (F/AF) dans le but de définir une direction de référence pour lespin des électrons de conduction. Les MRAM à écriture assistée thermiquement (TA-MRAM) utilisent mêmedeux bicouches F/AF : une pour le stockage de l’information et l’autre comme référence. Ces dernièresapplications technologiques impliquent des étapes de nanofabrication des couches minces continues pour formerune matrice de cellules mémoires individuelles. La qualification industrielle du produit final impose de sérieusescontraintes sur la largeur des distributions des propriétés magnétiques - y compris d’échange F/AF - de cellulemémoire à cellule mémoire. Des phases verres de spin réparties de manière aléatoire sur la couche continue, àl’interface F/AF ou dans le coeur de l’AF pourraient contribuer de manière significative à ces distributionsd’échange F/AF dans les dispositifs, après nanofabrication ; comme supposé il y a de cela quelques années. Lebut de cette thèse est d’étudier factuellement le possible lien entre verre de spin répartis dans des couches mincesF/AF et dispersions de propriétés d’échange de cellule mémoire à cellule mémoire dans les dispositifs TAMRAMcorrespondants. Avant cela, l’origine de ces régions verre de spin a été étudiée et une attention plusparticulière a été portée au rôle joué par les diffusions de Mn. Ces dernières ont été directement observées,comprises et l’utilisation de barrières complexes pour les réduire et par là même pour diminuer la quantité dephases verre de spin a été mise en oeuvre avec succès. En guise d’alternative pour varier la quantité de verres despin, l’utilisation d’AFs composites a été également étudiée dans le cadre de cette thèse. Ce type d’AF permet dumême coup de varier la stabilité thermique des grains AF et de répondre à un autre problème identifié pour lesTA-MRAM qui consiste à trouver des matériaux AF avec des propriétés de rétention et d’écriture intermédiairespar rapport aux matériaux actuellement utilisés. Finalement, ces AFs composites ont été utilisés comme moyende varier la quantité de verres de spin dans des dispositifs TA-MRAM réels et de prouver le lien direct avec ladispersion de propriétés de cellule mémoire à cellule mémoire. / Spintronics applications such as magnetic random access memories (MRAM), sensors (e.g.. hard diskdrive read head) and logic devices use ferromagnetic/antiferromagnetic (F/AF) exchange bias (EB) interactionsto set the reference direction required for the spin of conduction electrons. Thermally-assisted (TA-) MRAMapplications even use two F/AF exchange biased bilayers: one for reference and one for storage. Suchtechnological applications involve patterning full sheet wafers into matrix of individual bit-cells. Industrialproducts qualification imposes stringent requirements on the distributions of the magnetic properties from cell tocell, including those related to EB. It was supposed few years ago that randomly spread spin-glass like phases atthe F/AF interface or within the bulk of the AF layer significantly contribute to the distributions of EB propertiesin devices after processing. This thesis aimed at factually studying the link between spin-glasses spread overF/AF thin films and bit-cell dispersions of EB in corresponding TA-MRAM. Prior to that the origin of the spinglasslike regions and more specifically the role of Mn-diffusion was consolidated. Mn-diffusion was directlyobserved, understood and the use of complex barriers to reduce such diffusion and consequently to minimize theamount of spin-glass was successfully studied. Mixing AFs as another way to tune the amount of spin-glass likephases was also evidenced in the framework of this thesis. All at once, this last solution also tuned the AF grainsthermal stability and this solved another issue related to TA-MRAM, i.e. finding AF-materials with intermediateretention and write properties compared to the AFs presently used. Finally, those mixed antiferromagnets werethe mean chosen to tune the amount of spin-glass like phases in real TA-MRAM devices and to factually provetheir link with bit-cell distributions of EB properties.