This dissertation focuses on an uncooled Mid-Wave Infra-Red (MWIR) detector was developed by doping an n-type 4H-SiC with Ga using the laser doping technique. 4H-SiC is one of the polytypes of crystalline silicon carbide, a wide bandgap semiconductor. The dopant creates an energy level of 0.30 eV, which was confirmed by optical spectroscopy of the doped sample. This energy level corresponds to the MWIR wavelength of 4.21 µm. The detection mechanism is based on the photoexcitation of electrons by the photons of this wavelength absorbed in the semiconductor. This process modifies the electron density, which changes the refraction index and, therefore, the reflectance of the semiconductor is also changed. The change in the reflectance, which is the optical response of the detector, can be measured remotely with a laser beam such as a He-Ne laser. This capability of measuring the detector response remotely makes it a wireless optical detector. The variation of refraction index was calculated as a function of absorbed irradiance based on the reflectance data for the as-received and doped samples. A distinct change was observed for the refraction index of the doped sample, indicating that the detector is suitable for applications at 4.21 µm wavelength. The Ga dopant energy level in the substrate was confirmed by optical absorption spectroscopy. Secondary ion mass spectroscopy (SIMS) of the doped samples revealed an enhancement in the solid solubility of Ga in the substrate when doping is carried out by increasing the number of laser scans. Higher dopant concentration increases the number of holes in the dopant energy level, enabling photoexcitation of more electrons from the valence band by the incident MWIR photons. The detector performance improves as the dopant concentration increases from 1.15×1019 to 6.25×1020 cm-3. The detectivity of the optical photodetector is found to be 1.07×1010 cm·Hz1/2/W for the case of doping with 4 laser passes. The noise mechanisms in the probe laser, silicon carbide MWIR detector and laser power meter affect the performance of the detector such as the responsivity, noise equivalent temperature difference (NETD) and detectivity. For the MWIR wavelength 4.21 and 4.63 µm, the experimental detectivity of the optical photodetector of this study is found to be 1.07×1010 cm·Hz1/2/W, while the theoretical value is 2.39×1010 cm·Hz1/2/W. The values of NETD are found to be 404.03 and 15.48 mK based on experimental data for an MWIR radiation source of temperature 25°C and theoretical calculation respectively. The doped SiC also has a capability of gas detection since gas emission spectra are in infrared range. Similarly, the sensor is based on the semiconductor optics principle, i.e., an energy gap is created in a semiconductor by doping it with an appropriate dopant to ensure that the energy gap matches with an emission spectral line of the gas of interest. Specifically four sensors have been fabricated by laser doping four quadrants of a 6H-SiC substrate with Ga, Al, Sc and P atoms to detect CO2, NO, CO and NO2 gases respectively. The photons, which are emitted by the gas, excite the electrons in the doped sample and consequently change the electron density in various energy states. This phenomenon affects the refraction index of the semiconductor and, therefore, the reflectivity of the semiconductor is altered by the gas. The optical response of this semiconductor sensor is the reflected power of a probe beam, which is a He-Ne laser beam in this study. The CO2, NO, CO and NO2 gases change the refraction indices of Ga-, Al-, Sc- and Al-doped 6H-SiC, respectively, more prominently than the other gases tested in this study. Hence these doped 6H-SiC samples can be used as CO2, NO, CO and NO2 gas sensors respectively.

Silicon carbide is a promising semiconductor material for high voltage, high frequency and high temperature devices due to its wide bandgap, high breakdown electric field strength, highly saturated drift velocity of electrons and outstanding thermal conductivity. With the aim of overcoming some challenges in metallization and doping during the fabrication of silicon carbide devices, a novel laser-based process is provided to direct metallize the surface of silicon carbide without metal deposition and dope in silicon carbide without high temperature annealing, as an alternative to the conventional ion implantation, and find applications of this laser direct write metallization and doping technique on the fabrication of diodes, endotaxial layer and embedded optical structures on silicon carbide wafers. Mathematical models have been presented for the temperature distributions in the wafer during laser irradiation to optimize laser process parameters and understand the doping and metallization mechanisms in laser irradiation process. Laser irradiation of silicon carbide in a dopant-containing ambient allows to simultaneously heating the silicon carbide surface without melting and incorporating dopant atoms into the silicon carbide lattice. The process that dopant atoms diffuse into the bulk silicon carbide by laser-induced solid phase diffusion (LISPD) can be explained by considering the laser enhanced substitutional and interstitial diffusion mechanisms. Nitrogen and Trimethyaluminum (TMA) are used as dopants to produce n-type and p-type doped silicon carbide, respectively. Two laser doping methods, i.e., internal heating doping and surface heating doping are presented in this dissertation. Deep (800 nm doped junction for internal heating doping) and shallow (200 nm and 450 nm doped junction for surface heating doping) can be fabricated by different doping methods. Two distinct diffusion regions, near-surface and far-surface regions, were identified in the dopant concentration profiles, indicating different diffusion mechanisms in these two regions. The effective diffusion coefficients of nitrogen and aluminum were determined for both regions by fitting the diffusion equation to the measured concentration profiles. The calculated diffusivities are at least 6 orders of magnitude higher than the typical values for nitrogen and aluminum, which indicate that laser doping process enhances the diffusion of dopants in silicon carbide significantly. No amorphization was observed in laser-doped samples eliminating the need for high temperature annealing. Laser direct metallization can be realized on the surface of silicon carbide by generating metal-like conductive phases due to the decomposition of silicon carbide. The ohmic property of the laser direct metallized electrodes can be dramatically improved by fabricating such electrodes on laser heavily doped SiC substrate. This laser-induced solid phase diffusion technique has been utilized to fabricate endolayers in n-type 6H-SiC substrates by carbon incorporation. X-ray energy dispersive spectroscopic analysis shows that the thickness of endolayer is about 100 nm. High resolution transmission electron microscopic images indicate that the laser endotaxy process maintains the crystalline integrity of the substrate without any amorphization. Rutherford backscattering studies also show no amorphization and evident lattice disorder occur during this laser solid phase diffusion process. The resistivity of the endolayer formed in a 1.55 omega•cm silicon carbide wafer segment was found to be 1.1E5 omega•cm which is sufficient for device fabrication and isolation. Annealing at 1000 oC for 10 min to remove hydrogen resulted in a resistivity of 9.4E4 omega•cm. Prototype silicon carbide PIN diodes have been fabricated by doping the endolayer and parent silicon carbide epilayer with aluminum using this laser-induced solid phase diffusion technique to create p-regions on the top surfaces of the substrates. Laser direct metallized contacts were also fabricated on selected PIN diodes to show the effectiveness of these contacts. The results show that the PIN diode fabricated on a 30 nm thick endolayer can block 18 V, and the breakdown voltages and the forward voltages drop at 100 A/cm2 of the diodes fabricated on 4H-SiC with homoepilayer are 420 ~ 500 V and 12.5 ~ 20 V, respectively. The laser direct metallization and doping technique can also be used to synthesize embedded optical structures, which can increase 40% reflectivity compared to the parent wafer, showing potential for the creation of optical, electro-optical, opto-electrical, sensor devices and other integrated structures that are stable in high temperature, high-pressure, corrosive environments and deep space applications.

This work establishes a solid foundation for the use of indirect band gap semiconductors for light emitting application and presents the work on development of white light emitting diodes (LEDs) in silicon carbide (SiC). Novel laser doping has been utilized to fabricate white light emitting diodes in 6H-SiC (n-type N) and 4H-SiC (p-type Al) wafers. The emission of different colors to ultimately generate white light is tailored on the basis of donor acceptor pair (DAP) recombination mechanism for luminescence. A Q-switched Nd:YAG pulse laser (1064 nm wavelength) was used to carry out the doping experiments. The p and n regions of the white SiC LED were fabricated by laser doping an n-type 6H-SiC and p-type 4H-SiC wafer substrates with respective dopants. Cr, B and Al were used as p-type dopants (acceptors) while N and Se were used as n-type dopants (donors). Deep and shallow donor and acceptor impurity level states formed by these dopants tailor the color properties for pure white light emission. The electromagnetic field of lasers and non-equilibrium doping conditions enable laser doping of SiC with increased dopant diffusivity and enhanced solid solubility. A thermal model is utilized to determine the laser doping parameters for temperature distribution at various depths of the wafer and a diffusion model is presented including the effects of Fick's diffusion, laser electromagnetic field and thermal stresses due to localized laser heating on the mass flux of dopant atoms. The dopant diffusivity is calculated as a function of temperature at different depths of the wafer based on measured dopant concentration profile. The maximum diffusivities achieved in this study are 4.61x10-10 cm2/s at 2898 K and 6.92x10-12 cm2/s at 3046 K for Cr in 6H-SiC and 4H-SiC respectively. Secondary ion mass spectrometric (SIMS) analysis showed the concentration profile of Cr in SiC having a penetration depth ranging from 80 nm in p-type 4H-SiC to 1.5 [micro]m in n-type 6H-SiC substrates respectively. The SIMS data revealed enhanced solid solubility (2.29x1019 cm-3 in 6H-SiC and 1.42x1919 cm-3 in 4H-SiC) beyond the equilibrium limit (3x1017 cm-3 in 6H-SiC above 2500 [degrees]C) for Cr in SiC. It also revealed similar effects for Al and N. The roughness, surface chemistry and crystalline integrity of the doped sample were examined by optical interferometer, energy dispersive X-ray spectrometry (EDS) and transmission electron microscopy (TEM) respectively. Inspite of the larger atomic size of Cr compared to Si and C, the non-equilibrium conditions during laser doping allow effective incorporation of dopant atoms into the SiC lattice without causing any damage to the surface or crystal lattice. Deep Level Transient Spectroscopy (DLTS) confirmed the deep level acceptor state of Cr with activation energies of Ev+0.80 eV in 4H-SiC and Ev+0.45 eV in 6H-SiC. The Hall Effect measurements showed the hole concentration to be 1.98x1019 cm-3 which is almost twice the average Cr concentration (1x1019 cm-3) obtained from the SIMS data. These data confirmed that almost all of the Cr atoms were completely activated to the double acceptor state by the laser doping process without requiring any subsequent annealing step. Electroluminescence studies showed blue (460-498 nm), blue-green (500-520 nm) green (521-575 nm), and orange (650-690 nm) wavelengths due to radiative recombination transitions between donor-acceptors pairs of N-Al, N-B, N-Cr and Cr-Al respectively, while a prominent violet (408 nm) wavelength was observed due to transitions from the nitrogen level to the valence band level. The red (698-738 nm) luminescence was mainly due to metastable mid-bandgap states, however under high injection current it was due to the quantum mechanical phenomenon pertaining to band broadening and overlapping. This RGB combination produced a broadband white light spectrum extending from 380 to 900 nm. The color space tri-stimulus values for 4H-SiC doped with Cr and N were X = 0.3322, Y = 0.3320 and Z = 0.3358 as per 1931 CIE (International Commission on Illumination) corresponding to a color rendering index of 96.56 and the color temperature of 5510 K. And for 6H-SiC n-type doped with Cr and Al, the color space tri-stimulus values are X = 0.3322, Y = 0.3320 and Z = 0.3358. The CCT was 5338 K, which is very close to the incandescent lamp (or black body) and lies between bright midday sun (5200 K) and average daylight (5500 K) while CRI was 98.32. Similar white LED's were also fabricated using Cr, Al, Se as one set of dopants and B, Al, N as another.

Cette thèse se propose d’étudier les procédés de dopage et de recuit laser comme outils permettant la réalisation de cellules photovoltaïques au silicium cristallin. Des émetteurs dopés ou recuits par laser sont tout d’abord réalisés à l’aide de trois lasers et de différentes sources dopantes. Les lasers utilisés sont un laser vert nanoseconde, un laser excimère et un laser ultraviolet à haute cadence. Comme sources dopantes nous avons utilisé le verre de phosphore, des couches de nitrures de silicium dopées au bore ou au phosphore, ou encore des implantations ioniques de bore ou de phosphore. Des dopages très efficaces sont obtenus avec chaque couple laser/source dopante. En particulier, de faibles valeurs de résistances carrées et de densités de courant de saturation sont obtenues. Ces procédés laser sont ensuite appliqués à la réalisation de cellules à émetteur sélectif et à champ arrière au bore. Les cellules à émetteur sélectif dopé par laser (en utilisant le verre de phosphore comme source dopante) atteignent un rendement de 18,3 %, ce qui représente un gain total de 0,6 %abs comparé aux cellules standard à émetteur homogène. Les cellules à champ arrière au bore recuit par laser (à partir d’une implantation ionique de bore) montrent quant à elles un gain de 0,3 %abs par rapport aux cellules à champ arrière à l’aluminium, offrant ainsi un rendement de 16,7 %. / This study aims at investigating laser doping and laser annealing for crystalline silicon solar cells processing. Laser-processed emitters are firstly realized using three lasers and different dopants sources. The lasers are a nanosecond green laser, an excimer laser and a high-frequency ultraviolet laser. As dopants sources we used either phosphosilicate glass, phosphorus and boron-doped silicon nitrides, or phosphorus and boron ion implantation. Efficient phosphorus and boron doping are obtained using any of these laser/sources couple. In particular, low sheet resistances and low emitter saturation current densities are obtained. These laser processes are then applied to selective emitter and boron back-surface-field solar cells. Laser-doped selective emitter solar cells (using phosphosilicate glass as a dopants source) reach 18.3 % efficiency. This represents an overall gain of 0.6 %abs when compared to standard homogeneous emitter. On the other hand, laserannealed boron back-surface-field solar cells (using implanted boron as a dopants source) feature an overall gain of 0.3 %abs when compared to standard aluminium back-surface-field solar cells, thus yielding an efficiency of 16.7 %.

L'énergie photovoltaïque est promise à une forte croissance dans les prochaines années. Propre et renouvelable, elle possède en effet de sérieux atouts pour répondre aux grands enjeux posés par le réchauffement climatique et l'appauvrissement des ressources en énergie fossile. Elle reste néanmoins une énergie chère en comparaison des autres formes de production électrique. D'importants efforts de R&D doivent être mis en œuvre pour abaisser son coût et la rendre plus compétitive. Il existe d'ores et déjà dans les laboratoires des technologies permettant d'augmenter significativement le rendement des cellules solaires en silicium (qui représentent aujourd’hui l'essentiel du marché). Mais elles font appel le plus souvent à des procédés, comme la photolithographie, qui restent chères pour l'industrie photovoltaïque. Les technologies laser sont une voie envisagée pour répondre à ce problème. Sélectifs, sans contact et autorisant de hautes cadences, les procédés laser permettent de réaliser des structures avancées de cellules à moindre coût. Il existe ainsi une forte dynamique de recherche autour de ces procédés. Les traitements laser permettent d’usiner ou de modifier la matière, de façon rapide et fiable. Il est ainsi possible d’ablater sélectivement certains matériaux, de réaliser des tranchées ou des trous, ou encore de modifier des profils de dopage. Des architectures complexes deviennent ainsi accessibles sans recourir aux couteuses technologies de la microélectronique. C'est dans ce contexte que se déroule ce travail de thèse, financé par l'ADEME et la société SEMCO Eng., et s'inscrivant également dans le projet de l'Agence National pour la Recherche PROTERRA. Deux objectifs principaux ont motivé sa mise en place : développer un savoir-faire au laboratoire INL sur les technologies laser avec l'ambition de rejoindre les instituts leaders sur ces thématiques et transférer les procédés développés à l'équipementier SEMCO Eng. pour lui donner accès à une technologie aujourd'hui inédite dans l'industrie photovoltaïque française. Ces recherches ont porté sur les cellules photovoltaïques en silicium, dites de première génération, et se sont articulées autour de trois axes principaux : la modélisation de l'interaction laser matière, l'ablation sélective de diélectriques (notamment de la couche antireflet afin de permettre de nouvelles techniques de métallisation) et la réalisation de dopages localisés. Des cellules de grandes dimensions fabriquées en collaboration avec SEMCO Eng. et tirant parti de ces procédés ont permis d’obtenir des rendements en accord avec l’état de l’art (proche de 18 %). / Silicon solar cells still require cost reduction and improved efficiency to become more competitive. New architectures can provide a significant increase in efficiency, but today most of the approaches need additional fabrication steps. In this context, laser processing offers a unique way to replace technological steps like photolithography that is not compatible with the requirements of the photovoltaic industry. This PhD thesis will present two promising laser processes for silicon solar cells: selective laser doping and selective laser ablation. Laser-assisted diffusion of dopants is a promising way to produce at low cost advanced silicon solar cells with high efficiency. Indeed, selective emitters, which rely on high dopant concentration localized under the front electrical contacts are an effective way to reduce power losses at the front surface of silicon solar cells. Several laser-based techniques are competing to optimise the emitter geometry. One of the main approaches is to take advantage of the doping glass (usually P2O5 for p-type silicon solar cells) that is formed during the standard diffusion process. Selective laser ablation is an effective way to open the antireflection layer (SiNx) in order to perform alternative front side metallization. Indeed, in the industrial production of standard silicon solar cells, the front side metallization is made by screen printing of metal paste. This process scheme is very cost efficient but it leads to serious limitations of the solar cell efficiency. Electrochemical metallization avoids these issues but requires a selective opening of SiNx, which is usually done by photolithography. Direct laser ablation allows to consider this approach at an industrial level. These processes are presented illustrated by research conducted during this PhD at INL in laser technologies for photovoltaics. An innovative and potentially self-aligned process is also discussed, where the laser is used to open locally the antireflection and passivation coating, and at the same time, achieve local phosphorus diffusion. Moreover solar cells results above 18% have been obtained thanks to a selective emitter structure achieved with selective laser doping.