Vertical cavity surface emitting lasers (VCSELs) are today a commodity on the short wavelength laser market due to the ease with which they are manufactured. Much effort has in the last decade been directed towards making long wavelength VCSELs as successful in the marketplace. This has not been achieved due to the much more difficult fabrication technologies needed for realising high performance long wavelength VCSELs. At one point, GaInNAs quantum wells gain regions grown on GaAs substrates seemed to be the solution as it enabled all-epitaxial VCSELs that could make use of high contrast AlGaAs-based distributed Bragg reflectors (DBRs) as mirrors and lateral selective oxidation for optical and electrical confinement, thereby mimicking the successful design of short wavelength VCSELs. Although very good device results were achieved, reproducible and reliable epitaxial growth of GaInNAs quantum wells proved difficult and the technology has not made its way into high-volume production. Other approaches to the manufacturing and material problems have been to combine mature InP-based gain regions with high contrast AlGaAs-based DBRs by wafer fusion or with high contrast dielectric DBRs. Commonly, a patterned tunnel junction provides the electrical confinement in these VCSELs. Excellent performance has been achieved in this way but the fabrication process is difficult. In this work, we have employed high strain InGaAs quantum wells along with large detuning between the gain peak and the emission wavelength to realize GaAs-based long wavelength VCSELs. All-epitaxial VCSELs with AlGaAs-based DBRs and lateral oxidation confinement were fabricated and evaluated. The efficiency of these VCSELs was limited due to the optical absorption in the doped DBRs. To improve the efficiency and manufacturability, two novel optical and electrical confinement schemes based on epitaxial regrowth of current blocking layers were developed. The first scheme is based on a single regrowth step and requires very precise processing. This scheme was therefore not developed beyond the first generation but single mode power of 0.3 mW at low temperature, -10ºC, was achieved. The second scheme is based on two epitaxial regrowth steps and does not require as precise processing. Several generations of this design were manufactured and resulted in record high power of 8 mW at low temperature, 5ºC, and more than 3 mW at high temperature, 85ºC. Single mode power was more modest with 1.5 mW at low temperature and 0.8 mW at high temperature, comparable to the performance of the single mode lateral oxidation confined VCSELs. The reason for the modest single mode power was found to be a non-optimal cavity shape after the second regrowth that leads to poor lateral overlap between the gain in the quantum wells and the intensity of the optical field. / QC 20100825

<p> The work described in this thesis is devoted to the application of defect engineering in the development of silicon photonic devices. The thesis is divided into simulation and experimental portions, each focusing on a different form of defect engineered silicon: ion implantation induced amorphous silicon and solid-phase epitaxial regrowth suppressed polycrystalline silicon.</p> <p> The simulations are directed at silicon rib waveguide Raman laser applications. It is shown that a uniform, divacancy defect concentration will not enhance Raman gain. The excess optical loss and free carrier lifetime of rib waveguides with remote amorphous silicon volumes were simulated. Net gain was demonstrated depending on the geometry of the structure. For a waveguide structure with rib width, rib height and slab height of W = 1.5, H = 1.5 and h = 0.8 μm respectively, the optimal separation between the edge of the rib and the amorphous region is ~2.0 μm. Surface recombination velocity modification was shown to be an effective means to reduce free carrier lifetime.</p> <p> Experimental work was devoted to the characterization of a novel form of polycrystalline silicon created by amorphizing the entire silicon overlayer of a silicon-on-insulator wafer. Solid-phase epitaxial regrowth of the amorphous silicon is suppressed upon annealing due to the lack of a crystal seed and results in polycrystalline silicon. This material was characterized with ellipsometry, positron annihilation spectroscopy and x-ray diffraction. The material properties are shown to be heavily dependent on the annealing conditions. Ellipsometry showed that the refractive index at 1550 nm is comparable to crystalline silicon. Positron annihilation spectroscopy showed that the polycrystalline material exhibits a high concentration of vacancy-type defects while vertically regrown crystalline silicon does not. X-ray diffraction showed that the polycrystalline silicon is non-textured, strained in tension and is characterized by grain sizes less than 300 nm.</p> <p> Defect etching and optical measurements using a waveguide geometry were performed in order to characterize the lateral regrowth and the optical loss of the polycrystalline material. Lateral regrowth in the [011] direction was 1.53 and 0.96 μm for 10 minute anneals at 750 and 900 °C respectively, and at least 2.5 μm at 650 °C. Waveguide optical loss measurements with adjacent polycrystalline regions separated from the rib by at least 5.5 μm showed no separation dependence. The intrinsic optical loss of the polycrystalline material was estimated to be 1.05 and 1.57 dB/cm for TM and TE polarizations after a 900 °C anneal. Vertically regrown c-Si was shown to exhibit less than 3.0 dB/cm optical loss after annealing at 550 °C .</p> / Thesis / Master of Applied Science (MASc)

La réduction des dimensions des dispositifs CMOS (Complementary Metal Oxide Semiconductor) implique de nombreux défis dans la formation de jonctions. La recroissance par épitaxie en phase solide (SPER) à des températures inférieures à 600 °C est une technique attrayante dans la mesure où elle permet de réaliser des jonctions abruptes avec une forte concentration de dopants actifs et qui sont nécessaires pour les nœuds avancés tels que le 20 nm et au-delà. Dans ce manuscrit, on présente un modèle atomistique basé sur la méthode Monte-Carlo cinétique sur réseau (LKMC) afin de simuler la cinétique de SPER dans le silicium. Le modèle s'appuie sur la description phénoménologique des mécanismes microscopiques de recristallisation proposé par Drosd et Washburn dans [J. Appl. Phys. 53, 397 (1982)] en distinguant des événements {100}, {110} et {111} selon le plan local de recroissance et a été implémenté dans le simulateur MMonCa [Appl. Phys. Lett. 98, 233109 (2011)]. Il s'agit de la même base que le modèle de Martín-Bragado et Moroz [Appl. Phys. Lett. 95, 123123 (2009)] qui a été implémenté dans le simulateur commercial Synopsys SProcess KMC. Néanmoins, dans notre travail, la formation de macles lors des évènements {111} a été introduite ce qui a nécessité des changements importants dans l'implémentation. Le modèle a été calibré sur des résultats expérimentaux et permet de prédire l'anisotropie et la dépendance en température. En particulier, il a été utilisé afin d'expliquer la formation de zones défectueuses dans les dispositifs FDSOI à l'issue de la SPER à une température réduite. Le modèle LKMC a, en outre, été étendu dans le but d'inclure l'influence d'une contrainte non-hydrostatique et la recroissance accélérée du fait de la présence de dopants actifs. Les effets d'une contrainte non-hydrostatique ont été introduits en utilisant le concept de tenseur d'activation proposé par Aziz, Sabin et Lu dans [Phys. Rev. B 44, 9812 (1991)] et seulement quatre paramètres indépendants sont nécessaires. La présence de dopants ionisés cause une accélération de la vitesse de recroissance qui est attribué à un effet lié à la position du niveau de Fermi à l'interface amorphe/cristal. Un solveur 3D auto-cohérent de l'équation de Poisson avec le modèle de Thomas-Fermi a été implémenté et couplé avec le modèle LKMC afin de prendre en compte la courbure des bandes à l'interface amorphe/cristal. La correction phénoménologique de décalage du niveau de Fermi généralisé (GFLS) proposée par Williams et Elliman dans [Phys. Rev. Lett. 51, 1069 (1983)] a été utilisée pour modifier les fréquences de recristallisation des évènements microscopiques. Des simulations de la vitesse de recroissance en fonction de la température pour différentes concentrations de dopants ont montré un bon accord avec les données expérimentales. En résumé, dans ce manuscrit, un modèle unifié de SPER basé sur une approche LKMC est présentée. Il prend en compte l'influence de différents paramètres sur la cinétique de recroissance et ayant un intérêt technologique tels que la température, l'orientation cristalline, la contrainte et la présence de dopants. Le modèle est, en soi, tridimensionnel et permet donc d'explorer les phénomènes de recroissance impliquant plusieurs fronts de recristallisation et qui ont lieu lors du procédé de fabrication de dispositifs électroniques réels. / Complementary metal oxide semiconductor (CMOS) device scaling involves many technologicalchallenges in terms of junction formation. Solid phase epitaxial regrowth (SPER) at temperaturesbelow 600 ˝C is an attractive technique since it enables to form highly–activated andabrupt junctions that are required for advanced technology nodes such as 20 nm and beyond.In this manuscript, we present a comprehensive atomistic model relying on the lattice KineticMonte Carlo (LKMC) method to simulate SPER kinetics in silicon. The model is based onthe phenomenological description of the microscopic recrystallization mechanisms proposedby Drosd and Washburn in [J. Appl. Phys. 53, 397 (1982)] by distinguishing among {100},{110} and {111} events depending on the local regrowth plane and has been implemented inthe MMonCa simulator [Appl. Phys. Lett. 98, 233109 (2011)]. This is the same basis than theatomistic model of Martín–Bragado and Moroz proposed in [Appl. Phys. Lett. 95, 123123(2009)] and available in the Synopsys SProcess KMC commercial tool. Nevertheless, in ourwork the formation of twin configurations during {111} events has been incorporated givingrise to significant changes in the implementation. The model has been calibrated on single–directional SPER experiments and allows predicting the regrowth anisotropy and temperaturedependence. In particular, it has been used to explain the formation of defective regions inFDSOI devices annealed with a low processing temperature. In this work, the LKMC modelhas also been extended in order to include the influence of non–hystrostatic stress and dopant–enhanced regrowth that are technologically relevant. Non–hydrostatic stress effects have beenincorporated using the concept of activation strain tensor introduced by Aziz, Sabin and Luin [Phys. Rev. B 44, 9812 (1991)] and only four independent parameters are required. Thepresence of ionized dopants has been shown to cause an enhancement of the regrowth velocitywhich has been attributed to a Fermi level effect. A three–dimensional Thomas–Fermi–Poisson solver has been implemented and coupled with the LKMC model allowing to takeinto account the band bending at amorphous/crystalline interface. The phenomenological generalizedFermi level shifting (GFLS) correction proposed by Williams and Elliman in [Phys.Rev. Lett. 51, 1069 (1983)] has been used to modify the microscopic recrystallization rates.Simulations of the regrowth velocity as a function of temperature for different dopant concentrationshave shown a reasonable agreement with experimental data. In summary, in thismanuscript a unified SPER model relying on the LKMC approach is presented. It takes intoaccount various technologically relevant parameters influencing the regrowth kinetics such astemperature, crystalline orientation, stress and dopants. The model is per se three-dimensionaland can therefore be used to explore multi–directional regrowth phenomena that take place inreal electronic devices.

La fabrication de jonctions implique de nombreux défis technologiques à mesure que les dispositifs se rétrécissent. Afin de mitiger les problèmes liés à la diminution agressive des dimensions des transistors, des substrats SOI ainsi que du silicium-germanium (SiGe) contraint ont été introduits dans les nœuds avancés. Ces nœuds nécessitent toutefois une jonction abrupte fortement activée, qui est réalisable avec la recristallisation en phase solide (SPER) et un faible budget thermique (500°C-5h).Dans ce manuscrit, la SPER du silicium, germanium et d’alliages SiGe est étudiée avec des méthodes atomistiques telles que le Monte Carlo Cinétique (KMC) et la dynamique moléculaire (MD). Le modèle KMC de SPER se base sur une équation d'Arrhenius et distingue des configurations locales à l'interface amorphe-cristal pour simuler la dépendance de la vitesse de SPER par rapport à l’orientation de substrat. Les simulations en dynamique moléculaire montrent que la vitesse de SPER sur les orientations de {111} est fortement dépendante de la taille de la cellule ainsi que de la température et du temps de recuit.Le modèle KMC est de plus étendu afin de considérer l'effet du bore pendant la SPER. Le bore peut en effet créer des complexes à la fois dans l’amorphe et le cristal et augmenter la vitesse de SPER. Cette augmentation est toutefois saturée lorsque le bore atteint de trop fortes concentrations. Un modèle de réaction de défauts traitant les complexes a été adjoint au modèle de SPER afin de correctement simuler la vitesse de SPER pour toutes les concentrations de bore. Dans les alliages (100)SiGe relaxés, l'énergie d'activation de la SPER possède un maximum à 40% de concentration de Ge.Le modèle KMC doit introduire en plus des liaisons Si-Si et Ge-Ge, la liaison Si-Ge pour simuler correctement la recristallisation des alliages. Le modèle est également utilisé pour émettre des hypothèses sur la vitesse de SPER sur d'autres orientations. Les simulations en dynamique moléculaire confirment également le comportement de l’énergie d'activation dans les alliages SiGe.Des expériences de diffractions par rayons-X suivant en temps réel la recristallisation d’alliages de SiGe contraints ont été réalisées avec un rayonnement synchrotron. La contrainte est perdue dans les alliages riches en Ge et la température de recuit semble avoir un rôle sur la relaxation. La rugosité de l'interface pourrait être le lien entre la relaxation de la contrainte et la température, du fait que des simulations en dynamique moléculaires révèlent l’influence de la température de recuit sur la rugosité de l'interface et que les défauts relaxant la contrainte ont été associés à une interface rugueuse.En résumé, le SPER et ses diverses dépendances ont été étudiées dans ce manuscrit par des approches atomistiques. Les conclusions tirées améliorent la compréhension actuelle de la SPER, permettant ainsi une meilleure optimisation de la fabrication des jonctions. / The junction fabrication involve numerous technological challenges as the devices shrink. To alleviate issues brought by the aggressive device scaling, Fully Depleted SOI substrates as well as strained silicon-germanium (SiGe) have been introduced in advanced nodes. They however require a highly-activated abrupt junction achievable with solid phase epitaxial regrowth (SPER) and a low thermal budget (500$^circ$C-5h).In this manuscript, the SPER of silicon, germanium and SiGe alloys is investigated using Kinetic Monte Carlo (KMC) and Molecular Dynamics (MD) methods. The KMC model of SPER uses an Arrhenius equation and distinguishes local configurations at the amorphous-crystalline interface to simulate the SPER rate dependence on substrate orientations. In MD simulations, the SPER rate on {111} orientations is found to heavily depends on the cell size, anneal temperature and time.The KMC model is furthermore refined to consider the effect of boron during SPER. Boron is known to create complexes in both amorphous and crystalline phases and increase the SPER rate. This increase however saturates at high boron concentrations. A defect reaction model handling the complexes has been conjoined to the SPER model to correctly simulate the SPER rate behaviour for all boron concentrations.In relaxed (100)SiGe alloys, the SPER activation energy possesses a maximum at 40% of Ge concentration. The KMC model introduces in addition to Si-Si and Ge-Ge bonds, the Si-Ge bond to correctly simulate alloy recrystallisation. The model is also used to hypothesise the rates on other orientations. MD simulations also confirm the activation energy behaviour in SiGe alloys.Finally, X-ray diffractions following in real-time the recrystallisation of strained SiGe alloys are performed with synchrotron radiations. The strain is lost in Ge-rich alloys. The strain relaxation can be related to the anneal temperature. The interface roughness could be the link between the strain relaxation and the temperature, as MD simulations exhibit an influence of the anneal temperature on the interface roughness and strain relaxing defects are associated to a rough interface.In summary, the SPER and its several dependencies are investigated in this manuscript with atomistic approaches. The drawn conclusions increase the current understanding of SPER, allowing a better optimisation of junction fabrication.

L’activation à basse température est prometteuse pour l’intégration 3D séquentielle où lebudget thermique du transistor supérieur est limité (<650 ºC) pour ne pas dégrader letransistor inférieur, mais aussi dans le cas d’une intégration planaire afin d’atteindre des EOTultra fines et de contrôler le travail de sortie de la grille sans recourir à une intégration de type« gate-last ». Dans ce travail, l’activation par recroissance en phase solide (SPER) a étéétudiée afin de réduire le budget thermique de l’activation des dopants.L’activation à basse température présente plusieurs inconvénients. Les travauxprécédents montrent que les fuites de jonctions sont plus importantes dans ces dispositifs.Ensuite, des fortes désactivations de dopants ont été observées. Troisièmement, la faiblediffusion des dopants rend difficile la connexion des jonctions source et drain avec le canal.Dans ce travail, il est montré que dans un transistor FDSOI, l’augmentation des fuites dejonctions et la désactivation du Bore peuvent être évités grâce à la présence de l’oxyde enterré.De plus les conditions d’implantation ont été optimisées et les transistors activés à650 ºC atteignent les performances des transistors de référence. / Low temperature (LT) process is gaining interest in the frame of 3D sequentialintegration where limited thermal budget (<650 ºC) is needed for top FET to preserve bottomFET from any degradation and also in the standard planar integration for achieving ultra-thinEOT and work function control with high-k metal gate without gate-last integration scheme.In this work, LT Solid Phase Epitaxial Regrowth (SPER) has been investigated for reducingthe most critical thermal budget which is dopant activation.From previous works, LT activated devices face several challenges: First, higher junctionleakage limits their application to high performance devices. Secondly, strong deactivation ofthe metastable activated dopants was observed with post anneals. Thirdly, the dopant weakdiffusion makes it difficult to connect the channel with S/D.In this work, it is shown that the use of FDSOI enables to overcome junction leakage andBoron deactivation issues thanks to the defect cutting off and sinking effect of buried oxide.As a consequence, dopant deactivation in FDSOI devices is no longer an issue. Finally,implants conditions of LT transistors have been optimized to reach similar performance thanits standard high temperature counterparts.

Supersaturated substitutional alloys formed by ion implantation and rapid liquid-phase epitaxial regrowth induced by pulsed-laser annealing have been studied using Rutherford-backscattering and ion-channeling analysis. A series of impurities (As, Sb, Bi, Ga, In, Fe, Sn, Cu) have been implanted into single-crystal (001) orientation silicon at doses ranging from 1 x 10^15/cm2 to 1 x 10^17/cm2. The samples were subsequently annealed with a Ω-switched ruby laser (energy density ~1.5 J/cm2, pulse duration 15 x 10-9 sec). Ion-channeling analysis shows that laser annealing incorporates the Group III (Ga, In) and Group V (As, Sb, Bi) impurities into substitutional lattice sites at concentrations far in excess of the equilibrium solid solubility. Channeling measurements indicate the silicon crystal is essentially defect free after laser annealing. The maximum Group III and Group V dopant concentrations that can be incorporated into substitutional lattice sites are determined for the present laser-annealing conditions. Dopant profiles have been measured before and after annealing using Rutherford backscattering. These experimental profiles are compared to theoretical model calculations which incorporate both dopant diffusion in liquid silicon and a distribution coefficient (k') from the liquid. It is seen that a distribution coefficient (k') far greater than the equilibrium value (k0) is required for the calculation to fit the experimental data. In the cases of Fe, Zn, and Cu, laser annealing causes the impurities to segregate toward the surface. After annealing, none of these impurities are observed to be substitutional in detectable concentrations. The systematics of these alloys systems are discussed.