Two related bodies of work are presented, both of which aim to further the rapid development of next generation concentrating photovoltaic systems using high efficiency multi junction solar cells. They are complementary since the characterization of commercial devices and the systematic application of design principles for future designs must progress in parallel in order to accelerate iterative improvements. First addressed, is the field characterization of state of the art concentrating photovoltaic systems. Performance modeling and root cause analysis of deviations from the modeling results are critical for bringing reliable high value products to the market. Two complementary tools are presented that facilitate acceleration of the development cycle. The “Dynamic real-time I V Curve Measurement System…” provides a live picture of the current-voltage characteristics of a CPV module. This provides the user with an intuitive understanding of how module performance responds under perturbation. The “Shutter technique for noninvasive individual cell characterization in sealed concentrating photovoltaic modules,” allows the user to probe individual cell characteristics within a sealed module. This facilitates non-invasive characterization of modules that are in situ. Together, these tools were used to diagnose the wide spread failure of epoxy connections between the carrier and the emitter of bypass diodes installed in sealed commercial modules. Next, the optimization principals that are used to choose energy yield maximizing bandgap combinations for multi-junction solar cells are investigated. It is well understood that, due to differences in the solar resource in different geographical locations, this is fundamentally a local optimization problem. However, until now, a robust methodology for determining the influences of geography and atmospheric content on the ideal design point has not been developed. This analysis is presented and the influence of changing environment on the representative spectra that are used to optimize bandgap combinations is demonstrated. Calculations are confirmed with ground measurements in Ottawa, Canada and the global trends are refined for this particular location. Further, as cell designers begin to take advantage of more flexible manufacturing processes, it is critical to know if and how optimization criteria must change for solar cells with more junctions. This analysis is expanded to account for the differences between cells with up to 8 subcell bandgaps. A number of software tools were also developed for the Sunlab during this work. A multi-junction solar cell model calibration tool was developed to determine the parameters that describe each subcell. The tool fits a two diode model to temperature dependent measurements of each subcell and provides the fitting parameters so that the performance of multi-junction solar cells composed of those subcells can be modeled for real world conditions before they are put on-sun. A multi-junction bandgap optimization tool was developed to more quickly and robustly determine the ideal bandgap combinations for a set of input spectra. The optimization process outputs the current results during iteration so that they may be visualized. Finally, software tools that compute annual energy yield for input multi-junction cell parameters were developed. Both a brute force tool that computes energy harvested at each time step, and an accelerated tool that first bins time steps into discrete bins were developed. These tools will continue to be used by members of the Sunlab.

L'architecture des cellules solaires multi-jonction permet d'obtenir des records de rendement de conversion photovoltaïque, pouvant aller jusqu'à 46%. Leurs sous-cellules sont chacune conçues pour absorber une partie bien définie et complémentaire du spectre solaire, et sont connectées en série par des jonctions tunnel. La fabrication de cellules solaires tandem InGaP/GaAs d'énergies de bande interdite (" band gap ") 1,87 eV/1,42 eV accordées en maille sur substrat GaAs est bien maîtrisée, et de très hauts rendements peuvent être obtenus en ajoutant une ou deux sous-cellules de plus petit " gap " (1 eV et 0,7eV). Pour cela, les matériaux " petits gaps " fabriqués par Epitaxie par Jets Moléculaires (EJM) doivent être développés ainsi que des jonctions tunnel présentant une faible résistivité électrique, une haute transparence optique et de bonnes propriétés structurales. La croissance EJM et la modélisation de jonctions tunnel GaAs nous a permis d'identifier le mécanisme d'effet tunnel interbande plutôt que le mécanisme d'effet tunnel assisté par les défauts comme mécanisme dominant du transport dans ces structures. Nous avons exploité l'hétérostructure de type II fondée sur le système GaAsSb/InGaAs pour favoriser ce mécanisme d'effet tunnel interbande, et donc obtenir des jonctions tunnel de très faible résistivité tout en limitant la dégradation des propriétés optiques et structurales des composants inhérente à l'utilisation de matériaux " petits gaps " et désaccordés en maille GaAsSb et InGaAs. De plus, nous avons conçu une structure innovante d'hétérojonction tunnel de type II AlGaInAs/AlGaAsSb sous la forme de tampon graduel pour l'incorporation d'une sous-cellule métamorphique à 1 eV. Plusieurs candidats pour le matériau absorbeur à 1 eV à base de nitrure dilué InGaAsN(Bi) ont alors été développés et caractérisés, le contrôle de l'accord de maille étant assuré par un suivi en temps réel de la courbure de l'échantillon pendant la croissance EJM. Des premières cellules solaires III-V à base de GaAs, de nitrure dilué à 1 eV et de GaInAs métamorphique ont été fabriquées afin de valider les architectures développées de jonctions tunnel. Ce travail a permis de démontrer le potentiel de l'hétérostructure de type II GaAsSb/InGaAs pour répondre aux principaux défis de conception et de fabrication des cellules solaires multi-jonction sur substrat GaAs, que ce soit au niveau de la jonction tunnel ou au niveau de l'incorporation des sous-cellules de gap 1 eV. / Multi-Jonction Solar Cells (MJSCs) are leading the way of high efficiency photovoltaic devices, with conversion efficiency up to 46%. Their subcells are designed to absorb in a specific and complementary range of the solar spectrum, and are connected in series with tunnel junctions. The tandem architecture InGaP/GaAs - with bandgaps of 1.87 eV and 1.42 eV respectively - is mature and its efficiency could be enhanced by incorporating subcell(s) with bandgaps of 1 eV and/or 0.7 eV. The Molecular Beam Epitaxy (MBE) growth of such low bandgap materials has thus to be developed, as well as low-resistive tunnel junctions with good structural and optical properties. Based on the MBE growth and the simulation of GaAs tunnel junctions, we have identified interband tunneling as the predominant transport mechanism in such devices rather than trap-assisted-tunneling. The interband tunneling mechanism could be enhanced with the type II GaAsSb/InGaAs heterostructure. Using this material system, we have then demonstrated tunnel junctions with very low electrical resistivity with a limited degradation of the optical and structural properties inherently induced by the use of low band-gap and lattice-mismatched GaAsSb and InGaAs materials. Moreover, we fabricated an innovative AlInGaAs/AlGaAsSb tunnel junction as a graded buffer architecture that could be used for the incorporation of a 1 eV metamorphic subcell. We then developed and characterized InGaAsN(Bi) materials with band-gaps of ~1eV, taking advantage of in-situ wafer curvature measurements during the MBE growth to control the lattice-mismatch. Preliminary solar cells based on GaAs, 1 eV dilute nitride and metamorphic InGaAs have been fabricated and characterized validating the developed tunnel junction architectures. This work has enabled to demonstrate the potential of the type II GaAsSb/InGaAs heterostructure to meet the challenges posed by the conception and the fabrication of GaAs-based MJSCs, both for the tunnel junction and the 1 eV subcell.

No description available.

Le travail de thèse effectué a pour objectif d'amener vers une meilleure compréhension des comportements électrique et optique des modules CPV, dans des conditions environnantes variables. La première partie est consacrée à l’étude de la performance des modules en conditions réelles de fonctionnement. Quatre technologies de module, toutes équipées de cellules triple-jonctions, mais de concentrateurs optiques différents, ont été testées en extérieur sur des périodes de un mois à deux ans. Les résultats montrent que la sensibilité à la température de lentille, la température de cellule et au spectre incident varie selon le type d'architecture optique. La sensibilité la plus importante à la température de lentille a été obtenue pour un dispositif sans optique secondaire. Le coefficient en température de la tension Voc a été calculé et varie entre les technologies. Enfin, les variations importantes de facteur de forme avec le spectre incident observées pour une technologie, mettent en évidence la nécessité d'étudier les phénomènes de non-uniformités d'irradiance sur la cellule. Dans une deuxième partie, le développement d’un banc de test en intérieur permettant de mesurer les performances électriques et optiques est présenté. Ce banc a pour objectif de permettre la reproduction des conditions réelles de fonctionnement des modules de façon contrôlée en intérieur. Un système d’imagerie est utilisé pour déterminer la distribution spatiale et spectrale d’irradiance sur la cellule. Associé à un traceur de courbes IV, il vise à caractériser les effets de flux non-uniformes sur la cellule. Le banc de mesure a pour avantage de découpler les paramètres d’études, telles que la température de la lentille et la température de la cellule, et permet ainsi de décorréler leurs effets respectifs sur l'ensemble optique-cellule, ce qui n’est que difficilement possible sur des mesures en extérieur. Le procédé de calibration et la validation du dispositif sont détaillés dans le manuscrit. Enfin, dans une dernière partie, le banc développé est utilisé pour caractériser trois différents dispositifs CPV : un sans optique secondaire, et deux avec des optiques secondaires différentes. Les impacts de la distance lentille-cellule et de la température de lentille sur les performances de la cellule sont quantifiés optiquement et électriquement. Les résultats montrent comment ces paramètres modifient la distribution de densités de courant des sous-cellules, et donc le comportement électrique du dispositif. Ils soulignent plus spécifiquement comment les non-uniformités spectrales et spatiales affectent les performances de la cellule pour les différents concentrateurs. Le dispositif sans optique secondaire montre une sensibilité importante à la température de la lentille et la distance optique primaire-cellule, qui se traduit par une perte de production d'énergie dans des conditions réelles de fonctionnement. / The goal of this doctoral thesis is to bring answers to a better understanding of the electrical and optical behavior of CPV modules, under different operating conditions. In the first part, a study on module performance under real conditions is presented. Using an outdoor automated test bench, the sensitivity of four different CPV module technologies to most operating conditions relevant to CPV systems has been studied, namely DNI, spectrum, cell and lens temperature and clearness of the sky. In order to isolate the influence of a single operation parameter, the analysis of outdoor monitoring data from one month to two years is performed. The results show how the optical design influences the sensitivity of the electrical parameters to the mentionned operating conditions. The effect of lens temperature on cell current has been found to be maximum for the CPV module without Secondary Optical Element. Also the $V_{oc}$ thermal coefficient was found to vary between module technologies. Finally, the important variations of the fill factor for one technology underlines the need of studying non-uniformities effects on the cell performance. According to the results observed outdoors, an indoor tool was developed in order to uncorrelate outdoor parameters. A test bench that measures multi-spectral irradiance profiles, through CMOS imaging and bandpass filters in conjunction with electrical $IV$ curves, is used as a mean to visualize and characterize the effects of chromatic aberrations and nonuniform flux profiles under controllable testing conditions. The bench allows decoupling the temperatures of the Primary Optical Element and cell allowing the analyze of their respective effects on optical and electrical performance. In varying the temperature of the Primary Optical Element, the effects on electrical efficiency, focal distance, spectral sensitivity, acceptance angle, or multi-junction current matching profiles can be quantified. Calibration procedures and validation process are detailed. Finally, the developed testbench is used for analyzing the behvaior of three different CPV devices : one without Secondary Optical Element, and two with different Secondary Optical Elements. The impacts of cell position and lens temperature on the cell performance are quantified optically and electrically. The results show how these parameters modify the current density distribution of the subcells, and hence the electrical behavior of the device. They underline more specifically how spectral and spatial non-uniformities affect the cell performance for the different devices. The device without SOE shows a strong sensitivity to lens temperature and POE-cell distance, that will correspond to a decrease of energy production under real conditions of operation.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / This work deals with the analysis applied to the main methodologies found in literature for estimating the properties related to the physical phenomena in photovoltaic devices (parametrization), as well as the most important mathematical models used in the calculation of operating electrical characteristics of these devices (characterization). These devices are related to the mono and multi-junction technologies, when they are exposed to a condition where the temperature and solar radiation vary. Therefore, four parametrization methods were shown, including three analytical, and five models of electrical characterization, where two of them are specifically for multi-junction devices. Thus, several case studies were proposed which defined different situations for comparing the performance of the methods evaluated. In this way, the procedures that best fit to each type of photovoltaic technology were identified. Finally, according to the results obtained in the parameterization, the method based on the Generalized Reduced Gradient (GRG) nonlinear algorithm showed greater accuracy for all case studies and for all photovoltaic devices. As for the characterization, the main advantages and disadvantages of all models were determined, highlighting Domínguez, et al. (2010) model, due to the highest robustness and wide application range. / Esta dissertação apresenta uma análise aplicada às principais metodologias encontradas na literatura que permitem determinar as propriedades físicas relativas aos fenômenos que ocorrem nos dispositivos fotovoltaicos (etapa de parametrização), assim como dos modelos matemáticos de maior importância utilizados no cálculo das características elétricas operacionais destes dispositivos (etapa de caraterização). Tais dispositivos referem-se às tecnologias mono e multi junção quando submetidos à variações de temperatura e radiação solar. Portanto, foram apresentados quatro métodos de parametrização, entre eles três analíticos e cinco modelos de caracterização elétrica, sendo dois especificamente para dispositivos multi junção. Assim, estabeleceram-se vários estudos de caso para os quais foram definidas diferentes situações que permitiram comparar o desempenho de cada um dos métodos avaliados. Em consequência, foram identificados os procedimentos que melhor se ajustaram a cada tipo de tecnologia fotovoltaica. Dessa forma, de acordo com os resultados obtidos na parametrização, a metodologia baseada na aplicação do algoritmo de Gradiente Reduzido Generalizado (GRG) não linear, demonstrou maior exatidão para todos os estudos de caso e para todos os dispositivos fotovoltaicos. Já para a caraterização, foram determinadas as principais vantagens e desvantagens entre os modelos aplicados, destacando o modelo de Domínguez, et al. (2010), que apresentou maior robustez e ampla faixa de aplicação. / Mestre em Engenharia Mecânica

No description available.

La limite théorique de rendement des cellules photovoltaïques simple-jonction est de l’ordre de 33% d’après le modèle de Shockley-Queisser, ce qui reste éloigné de la limite de Carnot, prédisant une limite maximale de conversion énergie solaire → électricité de 93%. L’écart important entre ces deux limites découle des pertes intrinsèques, essentiellement liées à la conversion inefficace du spectre solaire et à la disparité entre les angles solides d’absorption et d’émission. Pour surmonter ces pertes et se rapprocher de la limite de Carnot, trois stratégies sont envisagées dans cette thèse : les cellules multi-jonction àconcentration, la combinaison de la concentration et de la restriction angulaire et les systèmes hybrides PV/CSP. Chacune de ces stratégies est limitée par des mécanismes qui dégradent leur performance.L’objectif de cette thèse est donc de comprendre dans quelle mesure les différents mécanismes limitants sont susceptibles d’affecter les performances des différentes stratégies étudiées, et d’optimiser l’architecture des cellules dans le but d’accroitre leur efficacité de conversion. Dans ce but, un modèle détaillé de cellule solaire tenant compte des principaux mécanismes limitant a été développé. Un outil d’optimisation par algorithme génétique a également été mis au point, afin d’explorer l’espace des différents paramètres étudiés pour identifier les conditions d’opération optimales. Nous démontrons l’importance majeure que revêt l’adaptation des propriétés optoélectroniques des matériaux utilisés aux conditions opératoires, que ce soit dans le cas des cellules solaires à concentration endurant des pertes résistives significatives, ou encore dans le cas de cellules solaires fonctionnant à des niveaux de températures très supérieurs à l’ambiante. Enfin, nous avons déterminé l’effet des principaux facteurs limitant que constituent les pertes résistives et les recombinaisons non-radiatives sur les cellules solairessimultanément soumises au flux solaire concentré et à la restriction angulaire du rayonnement émis. / The maximum efficiency limit attainable with a single-junction PV cell is ~ 33% according to the detailed balance formalism (also known as Shockley-Queisser model), which remains far from the Carnot limit, predicting a solar to electricity efficiency upper value of 93%. The large gap between both limits is due to intrinsic loss mechanisms, including the inefficient conversion of the solar spectrum and the large discrepancy between the solid angles of absorption and emission. To overcome these losses and get closer to the Carnot limit, three different strategies are considered in this thesis: concentrated multi-junction solarcells, the combination of solar concentration and angular confinement, and hybrid PV/CSP systems. Each strategy is inherently limited by several loss mechanisms that degrade their performances. The objective of this thesis is, hence, to better understand the extent to which these strategies are likely to be penalized by these losses, and to tailor the cell properties toward maximizing their efficiencies. To address these questions, a detailed-balance model of PV cell accounting for the main loss mechanisms was developed. A genetic-algorithm optimization tool was also implemented, aiming at exploring the parameter space and identifying the optimal operation conditions. We demonstrate the uttermost importance of tailoring the electronic properties of the materials used with both multi-junction solar cells undergoing significant series resistance losses, and PV cells operating at temperature levels exceeding ambient temperature. We also investigate the extent to which series resistances losses and non-radiative recombination are likely to affect the ability of PV cells simultaneously submitted to concentrated sunlight and angular restriction of the light emitted by band-to-band recombination.