We present the extension of time-dependent R-matrix (TDRM) theory to describe harmonic generation in general multielectron systems. The process of harmonic generation is of current interest because of its use as a probe of ultrafast electron dynamics and as a source of ultrashort (sub-femtosecond) laser pulses. Modelling harmonic generation- the production of high frequency light stimulated by laser-matter interaction- within the highly accurate and extensive TDRM theory allows us to investigate multielectron and multichannel effects as they influence the process. We detail recent advances in the field of attosecond physics, and give an overview of the theoretical methods used. A comprehensive explanation of the TDRM method, and the extensions made by this author, are provided. Calculations of harmonic generation in helium provide a test of the reliability of the method. We compare our results with those from the benchmark HELIUM method, finding striking agreement. We also assess the different methods of calculating the harmonic spectrum, using either the dipole length, velocity or acceleration operators. e We report the first evidence of multielectron interference in harmonic generation in argon. There are two mechanisms by which the fifth harmonic can be produced- excitation of electrons into the continuum, or resonant excitation of electrons into bound states. Interference between the continuum and resonant pathways gives rise to an asymmetric resonance in the fifth harmonic intensity, suggesting that the two pathways are of comparable importance. The final two results chapters explore harmonic generation in singly ionised argon, and uncover multichannel interference (interference between intermediate ion states) as well as multielectron effects in harmonic generation. We find that the dominant contribution to harmonic generation arises from excited states of Ar+. We also investigate the effect of the magnetic quantum number, finding that M=1 Ar+ exhibits a four-fold increase in harmonic generation over Ar+ with M=O.

En raison de la miniaturisation des dispositifs pour semi-conducteurs, le caractère aléatoire de la distribution de dopants dans un dispositif devient un paramètre critique pour les performances de ce dernier. Le but de ce travail est de valider une stratégie de dopage du silicium par un positionnement contrôlé de molécules, alternatif aux implantations, afin de limiter la variabilité de la tension de seuil. Nous avons choisi de contrôler la densité des sites et le positionnement des dopants en combinant le contrôle de la densité des sites d'ancrage et l'utilisation de molécules à fort encombrement stérique. Ceci a été réalisé en étudiant dans un premier temps le greffage de bore sur les silanols de silice amorphe partiellement traitée en température, à partir de molécules porteuses présentant des ligands de différentes tailles et des symétries ; le modèle de greffage a pu être déterminé en utilisant différentes techniques analytiques (IR-DRIFT, multi-core SSRMN et analyses élémentaires). L’élimination des ligands par un traitement thermique a permis de réaliser la fixation du Bore sur la silice avec un rendement supérieur à 96%. Cette méthode a été transférée avec succès à des wafers de silicium recouverts de silice native. Le recuit à haute température permettant la redistribution du bore dans le silicium a été ensuite validée par l’analyse VPD-ICPMS de l’oxyde greffé couplées aux mesures de profil de dopant dans le silicium obtenues par TofSIMS. Ce traitement a conduit à définir un procédé optimal par greffage sur silice mince, donnant des concentrations de dopant dans le silicium équivalentes à celles rapportées par la littérature sur silicium désoxydé, et sans passivation additionnelle de silice pour éviter la volatilisation du Bore greffé. En effet, la taille des ligands permet de contrôler la volatilisation du bore pendant recuit. Les analyses électriques par spectroscopie à effet tunnel ont confirmé l’activation électrique du dopant apporté par greffage et diffusé dans le silicium / With the everlasting shrinking of semiconductor devices, the randomness of dopant distribution within a device becomes more likely to critically impact the performance of the latter. The aim of this work is to validate a silicon doping strategy through a controlled positioning of molecules in place of conventional implantations in order to limit the variability of the threshold tension. In contrast to previous works, doping atoms were directly grafted onto a thin silica layer and not onto a bare silicon surface. Here, we chose to control both site density and positioning by combining the control of site anchoring density and the use of sterically hindered molecules to yield a finely structured doped surface. This was carried out by first optimizing this approach by studying the grafting of boron compounds with ligands of various sizes and symmetries on the surface silanols of non - porous amorphous silica partially treated at high temperatures (700 °C) as a model system. This allowed obtaining a fully characterization of surface species through combined analytical techniques (IR-DRIFT, solidstate multi-core NMR and elemental analyses). The ligands were then eliminated by a thermal treatment, yielding surface boronic acids characterized by IR-DRIFT and NMR with optimal density (> 96%, 6.7*1013 B.cm-²). This technology was then successfully transferred to silicon wafers covered with native silica as evidenced by ICPMS analyses of the grafted oxide layer removed in HF droplet (VPD). Subsequent high temperature annealing step without capping in order to trigger diffusion of boron was then validated on silicon wafers using ICPMS in HF-dipped oxide and in silicon by TofSIMS profile measurements. Such treatment led to a dopant concentration in the silicon matrix equivalent to that reported in the literature (e.g. direct grafting on silicon and cap during annealing). Electrical analyses by tunnel spectroscopy showed the efficiency of the annealing step and confirmed the dopant amount in the surface layer of the silicon wafer