Silicon carbide bipolar junction transistors (BJTs) are attractive power switching devices because of the unique material properties of SiC with high breakdown electric field, high thermal conductivity and high saturated drift velocity of electrons. The SiC BJT has potential for very low specific on-resistances and this together with high temperature operation makes it very suitable for applications with high power densities. For SiC BJTs the common emitter current gain (β), the specific on-resistance (RSP_ON), and the breakdown voltage are important to optimize for competition with silicon based power devices. In this thesis, power SiC BJTs with high current gain β ≈ 60 , low on-resistance RSP_ON ≈ 5 mΩcm2, and high breakdown voltage BVCEO ≈ 1200 V have been demonstrated. The 1200 V SiC BJT that has been demonstrated has about 80 % lower on-state power losses compared to a typical 1200 V Si IGBT chip. A continuous epitaxial growth of the base-emitter layers has been used to reduce interface defects and thus improve the current gain. A significant influence of surface recombination on the current gain was identified by comparing the experiments with device simulations. In order to reduce the surface recombination, different passivation layers were investigated in SiC BJTs, and thermal oxidation in N2O ambient was identified as an efficient passivation method to increase the current gain. To obtain a low contact resistance, especially to the p-type base contact, is one critical issue to fabricate SiC power BJTs with low on-resistance. Low temperature anneal (~ 800 oC) of a p-type Ni/Ti/Al contact on 4H-SiC has been demonstrated. The contact resistivity on the ion implanted base region of the BJT was 1.3 × 10-4 Ωcm2 after annealing. The Ni/Ti/Al p-type ohmic contact was adapted to 4H-SiC BJTs fabrication indicating that the base contact plays a role for achieving a low on-resistance of SiC BJTs. To achieve a high breakdown voltage, optimized junction termination is important in a power device. A guard ring assisted Junction Termination Extension (JTE) structure was used to improve the breakdown voltage of the SiC BJTs. The highest breakdown voltage of the fabricated SiC BJTs was obtained for devices with guard ring assisted JTE using the base contact implant step for a simultaneous formation of guard rings. As a new approach to fabricate SiC BJTs, epitaxial regrowth of an extrinsic base layer was demonstrated. SiC BJTs without any ion implantation were successfully demonstrated using epitaxial regrowth of a highly doped p-type region and an etched JTE using the epitaxial base. A maximum current gain of 42 was measured for a 1.8 mm × 1.8 mm BJT with a stable and reproducible open base breakdown voltage of 1800 V. / QC 20100819

Recombination mechanisms in gallium arsenide have been studied using temperature-dependent time-resolved photoluminescence-decay. New analytical methods are presented to improve the accuracy in bulk lifetime measurement, and these have been used to resolve the temperature-dependent lifetime. Fits to temperature-dependent lifetime yield measurement of the radiative-efficiency, revealing that samples grown by the Czochralski and molecular-beam-epitaxy methods are limited by radiative-recombination at 77K, with defect-mediated nonradiative-recombination becoming competitive at 300K and above. In samples grown with both doping types using molecular-beam-epitaxy, a common exponential increase in capture cross-section characterized by a high value of E_infinity=(258 +/- 1)meV was observed from the high-level injection lifetime over a wide temperature range (300-700K). This common signature was also observed from 500-600K in the hole-lifetime observed in n-type Czochralski GaAs where E_infinity=(261 +/- 7)meV was measured, which indicates that this signature parametrizes the exponential increase in hole-capture cross-section. The high E_infinity value rules out all candidate defects except for EL2, by comparison with hole-capture cross-section data previously measured by others using deep-level transient spectroscopy. / Thesis / Doctor of Philosophy (PhD)

Le recours au plasma pour stériliser des dispositifs médicaux (DM) est un domaine de recherche ne datant véritablement que de la fin des années 1990. Les plasmas permettent, dans les conditions adéquates, de réaliser la stérilisation à basse température (≤ 65°C), tel qu’exigé par la présence de polymères dans les DM et ce contrairement aux procédés par chaleur, et aussi de façon non toxique, contrairement aux procédés chimiques comme, par exemple, l’oxyde d’éthylène (OEt). Les laboratoires du Groupe de physique des plasmas à l’Université de Montréal travaillent à l’élaboration d’un stérilisateur consistant plus particulièrement à employer les effluents d’une décharge N2-%O2 basse pression (2-8 Torrs) en flux, formant ce que l’on appelle une post-décharge en flux. Ce sont les atomes N et O de cette décharge qui viendront, dans les conditions appropriées, entrer en collisions dans la chambre de stérilisation pour y créer des molécules excitées NO*, engendrant ainsi l’émission d’une quantité appréciable de photons UV. Ceux-ci constituent, dans le cas présent, l’agent biocide qui va s’attaquer directement au bagage génétique du micro-organisme (bactéries, virus) que l’on souhaite inactiver. L’utilisation d’une lointaine post-décharge évite du même coup la présence des agents érosifs de la décharge, comme les ions et les métastables. L’un des problèmes de cette méthode de stérilisation est la réduction du nombre de molécules NO* créées par suite de la perte des atomes N et O, qui sont des radicaux connus pour interagir avec les surfaces, sur les parois des matériaux des DM que l’on souhaite stériliser. L’objectif principal de notre travail est de déterminer l’influence d’une telle perte en surface, dite aussi réassociation en surface, par l’introduction de matériaux comme le Téflon, l’acier inoxydable, l’aluminium et le cuivre sur le taux d’inactivation des spores bactériennes. Nous nous attendons à ce que la réassociation en surface de ces atomes occasionne ainsi une diminution de l’intensité UV et subséquemment, une réduction du taux d’inactivation. Par spectroscopie optique d’émission (SOE), nous avons déterminé les concentrations perdues de N et de O par la présence des matériaux dans le stérilisateur, ainsi que la diminution de l’émission UV en découlant. Nous avons observé que cette diminution des concentrations atomiques est d’autant plus importante que les surfaces sont catalytiques. Au cours de l’étude du phénomène de pertes sur les parois pour un mélange N2-%O2 nous avons constaté l’existence d’une compétition en surface entre les atomes N et O, dans laquelle les atomes d’oxygènes semblent dominer largement. Cela implique qu’au-delà d’un certain %O2 ajouté à la décharge N2, seuls les atomes O se réassocient en surface. Par ailleurs, l’analyse des courbes de survie bi-phasiques des micro-organismes a permis d’établir une étroite corrélation, par lien de cause à effet, entre la consommation des atomes N et O en surface et la diminution du taux d’inactivation des spores dans la première phase. En revanche, nous avons constaté que notre principal agent biocide (le rayonnement ultraviolet) est moins efficace dans la deuxième phase et, par conséquent, il n’a pas été possible d’établir un lien entre la diminution des concentrations et le taux d’inactivation de cette phase-là. / The use of plasmas to sterilize medical devices (MDs) is a research field, which really started only at the end of the 90’s. Plasmas under adequate conditions allow achieving low-temperature (≤ 65°C) sterilization, as required by MDs made from polymers, in contrast to heat-driven sterilization methods, and provide a non-toxic method, in contrast to chemical processes such as performed, for example, with ethylene oxide (EtO). The Groupe de physique des plasmas laboratories at Université de Montréal is working on the design and testing of a sterilizer, which has the peculiarity of utilizing the species outflowing from a N2-%O2 discharge at reduced pressure (2-8 Torrs), which is called a plasma flowing-afterglow. It is the N and O atoms of this discharge mixture that, under appropriate conditions, interact in the sterilization chamber to form NO* excited molecules, generating a significant level of UV photons. These are, in the present case, the actual biocidal agent which will create lethal lesions on the genetic material of the microorganisms (bacteria, viruses) that should be inactivated. Using a flowing late afterglow instead of the discharge itself enables us to avoid the presence of the erosive agents of the discharge (ions, metastable-state particles). A major problem of this sterilization method is the reduction in the concentration of NO* molecules resulting from the losses of the N and O atoms on the surfaces of the MD materials that we want to sterilize. These radicals are, in fact, well-known to interact with surfaces and recombine on them. The main aim of our work is to determine the loss level of such atoms following their surface recombination on materials such as Teflon, stainless steel, aluminum and copper and the corresponding influence of such losses on the inactivation rate of bacterial spores. We can expect that surface recombination of these atoms leads to a reduction in the UV emission intensity and, as a result, in a reduction in the inactivation rate. Using optical emission spectroscopy (OES), we have determined the loss of N and O concentrations resulting from the presence of various materials in the sterilizer chamber as well as the corresponding decrease in UV emission intensity. We have observed that this reduction in atomic concentrations increases with the catalytic properties (recombination coefficient) of these materials. While examining the surface recombination phenomenon on these various materials, we have noticed a competition between the surface recombination of N and O atoms where the latter appear to play the main role. This implies that above a certain percentage of O2 added to N2, only the O atoms do recombine on these surfaces. On the other hand, the analysis of the bi-phasic survivor curves has enable us to show a strong correlation between the consumption of N and O atoms on surfaces and the reduction in the inactivation rate coefficient in the first phase of the survivor curve. We have also observed that our main biocidal agent is less efficient in the second phase of the survivor curve and, as a result, it was not possible to make a connection between the reduction in N and O atom concentration and the inactivation rate of the second phase.

This thesis deals with the analysis of nano-structured layers deposited by high-frequency magnetron sputtering on the monocrystalline silicon surface. The content of the work focuses on the magnetron sputtering application as an alternative method for passivation and antireflection layers deposition of silicon solar cells. The procedure of pre-deposite silicon surface cleaning by plasma etching in the Ar/H2 gas mixture atmosphere is suggested. In the next step the silicon nitride and aluminum nitride layers with hydrogen content in Ar/N2/H2 gas mixture by magnetron sputtering are deposited. One part of the thesis describes an experimental pseudo-carbide films deposition from a silicon target in the atmosphere of acetylene (C2H2). An emphasis is placed on the research of sputtered layers properties and on the conditions on the silicon-layer interface with the help of the standard as well as special measurement methods. Sputtered layers structure is analyzed by modern X-ray spectroscopy (XPS) and by Fourier infrared spectroscopy (FTIR). Optical ellipsometry and spectrophotometry is used for the diagnostic of the layers optical properties depending upon the wavelength of incident light. A special method of determining the surface lay-out of the charge´s carrier life in the volume and on the surface of silicon is employed to investigate the passivating effects of the sputtered layers.

Le recours au plasma pour stériliser des dispositifs médicaux (DM) est un domaine de recherche ne datant véritablement que de la fin des années 1990. Les plasmas permettent, dans les conditions adéquates, de réaliser la stérilisation à basse température (≤ 65°C), tel qu’exigé par la présence de polymères dans les DM et ce contrairement aux procédés par chaleur, et aussi de façon non toxique, contrairement aux procédés chimiques comme, par exemple, l’oxyde d’éthylène (OEt). Les laboratoires du Groupe de physique des plasmas à l’Université de Montréal travaillent à l’élaboration d’un stérilisateur consistant plus particulièrement à employer les effluents d’une décharge N2-%O2 basse pression (2-8 Torrs) en flux, formant ce que l’on appelle une post-décharge en flux. Ce sont les atomes N et O de cette décharge qui viendront, dans les conditions appropriées, entrer en collisions dans la chambre de stérilisation pour y créer des molécules excitées NO*, engendrant ainsi l’émission d’une quantité appréciable de photons UV. Ceux-ci constituent, dans le cas présent, l’agent biocide qui va s’attaquer directement au bagage génétique du micro-organisme (bactéries, virus) que l’on souhaite inactiver. L’utilisation d’une lointaine post-décharge évite du même coup la présence des agents érosifs de la décharge, comme les ions et les métastables. L’un des problèmes de cette méthode de stérilisation est la réduction du nombre de molécules NO* créées par suite de la perte des atomes N et O, qui sont des radicaux connus pour interagir avec les surfaces, sur les parois des matériaux des DM que l’on souhaite stériliser. L’objectif principal de notre travail est de déterminer l’influence d’une telle perte en surface, dite aussi réassociation en surface, par l’introduction de matériaux comme le Téflon, l’acier inoxydable, l’aluminium et le cuivre sur le taux d’inactivation des spores bactériennes. Nous nous attendons à ce que la réassociation en surface de ces atomes occasionne ainsi une diminution de l’intensité UV et subséquemment, une réduction du taux d’inactivation. Par spectroscopie optique d’émission (SOE), nous avons déterminé les concentrations perdues de N et de O par la présence des matériaux dans le stérilisateur, ainsi que la diminution de l’émission UV en découlant. Nous avons observé que cette diminution des concentrations atomiques est d’autant plus importante que les surfaces sont catalytiques. Au cours de l’étude du phénomène de pertes sur les parois pour un mélange N2-%O2 nous avons constaté l’existence d’une compétition en surface entre les atomes N et O, dans laquelle les atomes d’oxygènes semblent dominer largement. Cela implique qu’au-delà d’un certain %O2 ajouté à la décharge N2, seuls les atomes O se réassocient en surface. Par ailleurs, l’analyse des courbes de survie bi-phasiques des micro-organismes a permis d’établir une étroite corrélation, par lien de cause à effet, entre la consommation des atomes N et O en surface et la diminution du taux d’inactivation des spores dans la première phase. En revanche, nous avons constaté que notre principal agent biocide (le rayonnement ultraviolet) est moins efficace dans la deuxième phase et, par conséquent, il n’a pas été possible d’établir un lien entre la diminution des concentrations et le taux d’inactivation de cette phase-là. / The use of plasmas to sterilize medical devices (MDs) is a research field, which really started only at the end of the 90’s. Plasmas under adequate conditions allow achieving low-temperature (≤ 65°C) sterilization, as required by MDs made from polymers, in contrast to heat-driven sterilization methods, and provide a non-toxic method, in contrast to chemical processes such as performed, for example, with ethylene oxide (EtO). The Groupe de physique des plasmas laboratories at Université de Montréal is working on the design and testing of a sterilizer, which has the peculiarity of utilizing the species outflowing from a N2-%O2 discharge at reduced pressure (2-8 Torrs), which is called a plasma flowing-afterglow. It is the N and O atoms of this discharge mixture that, under appropriate conditions, interact in the sterilization chamber to form NO* excited molecules, generating a significant level of UV photons. These are, in the present case, the actual biocidal agent which will create lethal lesions on the genetic material of the microorganisms (bacteria, viruses) that should be inactivated. Using a flowing late afterglow instead of the discharge itself enables us to avoid the presence of the erosive agents of the discharge (ions, metastable-state particles). A major problem of this sterilization method is the reduction in the concentration of NO* molecules resulting from the losses of the N and O atoms on the surfaces of the MD materials that we want to sterilize. These radicals are, in fact, well-known to interact with surfaces and recombine on them. The main aim of our work is to determine the loss level of such atoms following their surface recombination on materials such as Teflon, stainless steel, aluminum and copper and the corresponding influence of such losses on the inactivation rate of bacterial spores. We can expect that surface recombination of these atoms leads to a reduction in the UV emission intensity and, as a result, in a reduction in the inactivation rate. Using optical emission spectroscopy (OES), we have determined the loss of N and O concentrations resulting from the presence of various materials in the sterilizer chamber as well as the corresponding decrease in UV emission intensity. We have observed that this reduction in atomic concentrations increases with the catalytic properties (recombination coefficient) of these materials. While examining the surface recombination phenomenon on these various materials, we have noticed a competition between the surface recombination of N and O atoms where the latter appear to play the main role. This implies that above a certain percentage of O2 added to N2, only the O atoms do recombine on these surfaces. On the other hand, the analysis of the bi-phasic survivor curves has enable us to show a strong correlation between the consumption of N and O atoms on surfaces and the reduction in the inactivation rate coefficient in the first phase of the survivor curve. We have also observed that our main biocidal agent is less efficient in the second phase of the survivor curve and, as a result, it was not possible to make a connection between the reduction in N and O atom concentration and the inactivation rate of the second phase.