This work explores two very different structural systems: n-type hydrogenated amorphous silicon (a-Si:H) and gold nanoparticles (AuNPs) suspended in a matrix of organic ligands. For a-Si:H, examination of the gas-phase concentration of dopant (1-6% PH3/SiH4) and argon diluent effects includes the temperature dependent conductivity, low-frequency electronic noise, and Raman spectroscopy to examine structure. It is found that a-Si:H samples grown with high dopant concentration or with argon dilution exhibit an anomalous hopping conduction mechanism with an exponent of p=0.75. An experimental approach is used to determine correlations between conduction parameters, such as the pre-exponential factor and the characteristic temperature, rather than an analysis of existing models to explain the anomalous conduction. From these results, the anomalous conduction is a result of a change in the shape of the density of states and not a shift of the Fermi level with dopant. Additionally, it is found that argon dilution increases the carrier mobility, reduces the doping efficiency, and causes a degradation of the short-range order. With AuNPs, a comparison of temperature dependent conductivity and low-frequency noise shows that the temperature coefficient of resistance (TCR) is independent of the length of interparticle distance while the noise magnitude decreases.

Στην παρούσα διπλωματική εργασία ασχοληθήκαμε με την επίδραση των πραγματικών συνθηκών στις dc παραμέτρους φωτοβολταϊκών πλαισίων άμορφου πυριτίου (a-Si) και την επενέργειά τους στη διαστασιολόγιση φωτοβολταϊκού συστήματος. Συγκεκριμένα οι μετρήσεις έγιναν στο Πανεπιστήμιο Πατρών και το πλαίσιο είναι άμορφο πυρίτιο τύπου Scott Solar Gmbh ASi-F 32/12. Ειδικότερα, μετρήσαμε την τάση φωτοβολταϊκού πλαισίου αμόρφου πυριτίου, το ρεύμα, την ηλιακή ακτινοβολία, την θερμοκρασία περιβάλλοντος και τις θερμοκρασίες στην πίσω και τη μπροστινή πλευρά του, σε διαφορετικές ώρες της ημέρας, διαφορετικές κλίσεις και καιρικές συνθήκες. Από τις μετρήσεις αυτές εξάγαμε συμπεράσματα και γραφικές παραστάσεις ρεύματος-τάσης (I-V) και ισχύος-τάσης (P-V) για κάθε μέτρηση. Στην συνέχεια υπολογίσαμε το βέλτιστο συντελεστή απόδοσης του πλαισίου και οδηγηθήκαμε σε συμπεράσματα για την ηλιακή ακτινοβολία η οποία όσο μεγαλύτερη είναι έχουμε και καλύτερη απόδοση του πλαισίου μας. Επίσης, μετρήσαμε τις διαστάσεις του Πανεπιστημιακού γυμναστηρίου και βρήκαμε το βέλτιστο αριθμό πλαισίων και τη βέλτιστη κλίση ώστε να τοποθετηθούν στο χώρο αυτό για να έχουμε τη βέλτιστη απόδοση. γίνεται μελέτη εγκατάστασης φωτοβολταϊκών πλαισίων αμόρφου πυριτίου τύπου Scott Solar Gmbh ASi-F 32/12 στο Πανεπιστημιακό Γυμναστήριο Πατρών και η δυνατότητα λειτουργίας αυτού με πλήρη αυτονομία ή με σύνδεσή του στο δίκτυο της ΔΕΗ. Όσον αφορά την πλήρη αυτονομία οδηγηθήκαμε στο συμπέρασμα ότι υπάρχει μεγάλη διαφορά ανάμεσα στην επιφάνεια που έχουμε και στην επιφάνεια που χρειάζεται για να τοποθετηθούν τα Φ/Β πλαίσια έτσι ώστε να καλύπτεται πλήρως το Πανεπιστημιακό Γυμναστήριο. Στη συνέχεια έπειτα από τη μελέτη για το διασυνδεδεμένο σύστημα με το δίκτυο της ΔΕΗ υπολογίστηκε ότι η ετήσια παραγόμενη ηλεκτρική ενέργεια είναι 51,7 MWh/έτος ή 1115 kWh/έτος/kWp για τα συγκεκριμένα Φ/Β πλαίσια και η απαιτούμενη εγκατεστημένη ισχύς PpΣ=46,4 kWp. Με τη χρήση του προγράμματος PVsol προέκυψε ότι η ετήσια παραγόμενη ηλεκτρική ενέργεια είναι ίση με 46,73 kWh/έτος και ότι η εγκατάσταση θα παράγει τελικά 46,73 kWh/έτος *1450 Φ/Β πλαίσια = 67,7 MWh/έτος ή 1.460 kWh/έτος/kWp . Αξίζει να σημειώσουμε ότι οι πειραματικές μετρήσεις της συγκεκριμένης μελέτης αποτελούν ένα μέρος από τις απαιτούμενες μετρήσεις για την ολοκληρωμένη εφαρμογή με την οποία ασχοληθήκαμε στη παρούσα Διπλωματική εργασία. Τέλος, σειρά έχουν τα συμπεράσματα της διπλωματικής αυτής εργασίας και κατόπιν το παράρτημα όπου εκεί βρίσκονται οι μετρήσεις οι οποίες έγιναν για την δημιουργία των γραφικών παραστάσεων που θα χρησιμοποιηθούν παρακάτω / This diploma thesis dealt with the effect of actual conditions in the dc parameters of solar panels amorphous silicon (a-Si) and their effects on diastasiologisi solar system. Specifically, measurements were made at the University of Patras and the framework is amorphous silicon type Scott Solar Gmbh ASi-F 32/12.Specifically, we measured the voltage of amorphous silicon photovoltaic panel, electricity, solar radiation, ambient temperature and temperatures in the back and front side, at different times of day, different slopes and weather conditions.

abstract: The field of flexible displays and electronics gained a big momentum within the recent years due to their ruggedness, thinness, and flexibility as well as low cost large area manufacturability. Amorphous silicon has been the dominant material used in the thin film transistor industry which could only utilize it as N type thin film transistors (TFT). Amorphous silicon is an unstable material for low temperature manufacturing process and having only one kind of transistor means high power consumption for circuit operations. This thesis covers the three major researches done on flexible TFTs and flexible electronic circuits. First the characterization of both amorphous silicon TFTs and newly emerging mixed oxide TFTs were performed and the stability of these two materials is compared. During the research, both TFTs were stress tested under various biasing conditions and the threshold voltage was extracted to observe the shift in the threshold which shows the degradation of the material. Secondly, the design of the first flexible CMOS TFTs and CMOS gates were covered. The circuits were built using both inorganic and organic components (for nMOS and pMOS transistors respectively) and functionality tests were performed on basic gates like inverter, NAND and NOR gates and the working results are documented. Thirdly, a novel large area sensor structure is demonstrated under the Electronic Textile project section. This project is based on the concept that all the flexible electronics are flexible in only one direction and can not be used for conforming irregular shaped objects or create an electronic cloth for various applications like display or sensing. A laser detector sensor array is designed for proof of concept and is laid in strips that can be cut after manufacturing and weaved to each other to create a real flexible electronic textile. The circuit designed uses a unique architecture that pushes the data in a single line and reads the data from the same line and compares the signal to the original state to determine a sensor excitation. This architecture enables 2 dimensional addressing through an external controller while eliminating the need for 2 dimensional active matrix style electrical connections between the fibers. / Dissertation/Thesis / Ph.D. Electrical Engineering 2012

Neste trabalho investigamos as modificações na morfologia superficial e estrutura de filmes finos de silício amorfo hidrogenado, resultantes da irradiação com pulsos ultracurtos de femtossegundos (150 fs, 775 nm e 1 kHz). Os processos de microfabricação foram conduzidos varrendo, a velocidade constante, um feixe laser com diferentes fluências (1,8 a 6,2 MJ/m2) sobre a amostra. Os espectros de transmissão apresentaram queda para amostras irradiadas, cujas imagens de microscopia eletrônica de varredura mostraram estruturas superficiais condizentes com o fenômeno de LIPSS (Laser Induced Periodic Surface Structures). Uma análise estatística das imagens de microscopia de força atômica foi realizada com um programa que identifica e caracteriza os domínios (picos) produzidos pela microfabricação. O histograma de altura da amostra irradiada com uma fluência de 3,1 MJ/m2 mostrou que a altura média dos picos produzidos é de 15 nm, menor que o centro da distribuição de alturas para uma amostra não irradiada. Porém, para fluências acima de 3,7 MJ/m2 a morfologia é dominada pela formação de agregados. Medidas de espectroscopia Raman revelaram a formação de uma fração de silício cristalino, após a irradiação com pulsos de femtossegundos, de até 77% para 6,2 MJ/m2. Determinamos ainda uma diminuição da dimensão dos nanocristais produzidos com o aumento da fluência do laser de excitação. Portanto, nossos resultados mostram que há um compromisso entre as propriedades obtidas pela microfabricação (transmissão, distribuição de picos, fração de cristalização e tamanho dos nanocristais produzidos) que deve ser levado em conta ao aplicar a técnica de microestruturação com laser de femtossegundos. / In this work we investigated surface morphology and structural modification on hydrogenated amorphous silicon (a-Si:H) thin films, resulting from femtosecond laser irradiation (150 fs, 775 nm and 1 kHz). Microfabrication processes were carried out scanning sample´s surface, at constant speed, with distinct laser fluencies (from 1.8 to 6.2 MJ/m2). A decrease was observed in the transmission spectra of irradiated samples, whose scanning electron microscopy images revealed surface structures compatible with the Laser Induced Periodic Surface Structure (LIPSS) phenomenon. A statistical analyzes of Atomic Force Microcopy images was performed using a specially developed software, that identifies and characterizes the domains (spikes) produced by the laser irradiation. The height histogram for a sample irradiated with 3.1 MJ/m2 reveals that the average height of the produced spikes is at 15 nm, which is smaller than the center of height distribution for non-irradiated sample. For fluencies higher than 3.7 MJ/m2, however, aggregation of the produced spikes dominates the sample morphology. Raman spectroscopy revealed the formation of a crystalline fraction of 77% for laser fluence irradiation of 6.2 MJ/m2, as well as a decrease in size of the produced crystals as a function of fluence. Therefore, our results indicate that there is a compromise of the sample transmission, spikes distribution, crystallization fraction and size of nanocrystals obtained by fs-laser irradiation, which has to be taken into consideration when using this material processing method.

Devido às suas potenciais aplicações tecnológicas (células solares, transistores de filme fino TFT, etc.), o estudo do silício amorfo (a-Si) tem despertado o interesse da comunidade científica desde o final da década de 70. Mais recentemente, este interesse foi renovado com o desenvolvimento da técnica de cristalização induzida por metal (Metal-induced Crystallization MIC), por causa do considerável interesse na obtenção do silício cristalino (c-Si) a baixas temperaturas. Dentre as principais abordagens adotadas para o estudo da MIC, destaca-se aquela realizada em estruturas consistindo de camadas alternadas de silício amorfo e filmes metálicos, por exemplo. Conseqüentemente, concluiu-se que a difusão de átomos do semicondutor para o metal (e/ou vice-versa) é o mecanismo responsável pela cristalização. Esta explicação fenomenológica, entretanto, não considera os mecanismos microscópicos que provocam a cristalização à baixa temperatura. Tendo isto por base, este trabalho diz respeito a uma abordagem diferente e complementar para a investigação do processo de MIC, através da inserção de uma quantidade controlada e homogeneamente distribuída de átomos de metal na rede amorfa. Para este estudo, filmes de silício amorfo dopados com diferentes concentrações de Ni e possuindo diferentes espessuras, depositados em substratos de c-Si, c-Ge, quartzo cristalino e vidro foram preparados pela técnica de sputtering de rádio freqüência, em uma atmosfera controlada de argônio. Com o objetivo de se investigar a influência exercida pela estrutura atômica nos mecanismos de cristalização destes filmes, todos foram submetidos a tratamentos térmicos cumulativos até 1000 oC. Para isto, os filmes foram caracterizados pelas técnicas de espalhamento Raman, transmissão óptica, EDS (energy dispersive spectrometry) e microscopia eletrônica de varredura (SEM). Os resultados experimentais indicam que a quantidade de Ni, a espessura e a natureza do substrato determinam a temperatura para o qual se inicia a cristalização dos filmes de a-Si, e que a espessura e a presença de Ni têm efeito direto sobre as propriedades ópticas dos filmes. Estudos preliminares utilizando-se as técnicas de microscopia de força atômica (AFM) e Extended X-ray Absorption Fine Structure (EXAFS) também foram feitos. / Due to their great technological potential (solar cells, thin film transistors, etc.), the study of amorphous silicon (a-Si) is attracting the attention of the scientific community since the 70s. More recently, such interest was renewed with the development of the Metal-induced Crystallization (MIC) technique, because of considerable interest in low-temperature formation of crystalline silicon (c-Si). Amongst the principal approaches to the study of MIC, stand out that performed on structures consisting of alternating layers of amorphous silicon and metal films, for example. Consequently, they conclude that the diffusion of semiconductor atoms into the metal (and/or vice-versa) is the mechanism responsible for the crystallization. This phenomenological explanation, however, does not consider the microscopic mechanisms that provoke the low temperature crystallization. Based on the above ideas, this work refers to a different and complementary approach to investigate the MIC process, by the insertion of a controlled and homogeneously distributed amount of metal atoms in the amorphous network. For this study, amorphous silicon films doped with different Ni concentrations and having different thicknesses, deposited on c-Si, c-Ge, crystalline quartz and glass substrates were prepared by the radio frequency sputtering technique in a controlled atmosphere of argon. In order to investigate the influence exerted by the atomic structure on the crystallization mechanisms of these films, all of them have been submitted to cumulative thermal annealing treatments up to 1000 C. To that aim, the films were investigated by Raman scattering, optical transmission, energy dispersive spectrometry (EDS) and scanning electron microscopy (SEM). The experimental results indicate that the Ni content, the thickness and the nature of the substrate determines the crystallization temperature onset of the a-Si films, and that the thickness and the presence of Ni have direct effect on the optical properties of the films. Preliminary studies using the atomic force microscopy (AFM) and Extended X-ray Absorption Fine Structure (EXAFS) techniques have also been carried out.

Orientador: Francisco das Chagas Marques / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Fisica "Gleb Wataghin" / Made available in DSpace on 2018-08-14T18:36:55Z (GMT). No. of bitstreams: 1 Barbieri_PauloFernando_D.pdf: 3626831 bytes, checksum: 0f15b74219acbe9e82e54ae2ce3d1ed4 (MD5) Previous issue date: 2009 / Resumo: Esta tese mostra um estudo realizado sobre implantação de xenônio (Xe) em uma matriz amorfa de silício (a-Si) utilizando a técnica de Deposição Assistida com Feixe de Íons (IBAD). Para esta finalidade, um feixe de íons de Xe com energia de 1500 eV, obtido de um canhão tipo Kaufman, foi utilizado para a realização de sputtering de um alvo de silício, produzindo um filme de a-Si em um substrato colocado a 15 cm de distância via deposição. Ao mesmo tempo outro feixe de íons de Xe, obtido por um segundo canhão de energia variável (0 a 300 eV), implantava átomos na rede do a-Si que estava em formação. Foi possível implantar concentrações de xenônio de até 5% com energia da ordem de 50 eV. Esta energia de implantação é de várias ordens de grandeza menor do que a energia utilizada no processo convencional de implantação que utiliza valores de energia de ordem de milhares de eV. Para investigar o Xe dentro da matriz amorfa de Si foram utilizadas técnicas de excitações atômicas XAS (X-Ray Absorption Spectroscopy), XPS (X-Ray Photoelectron Spectroscopy) e XAES (X-Ray Auger Excited Spectroscopy) e também simulações teóricas para XAS. Os resultados experimentais de XAS indicam que os átomos de Xe são implantados majoritariamente de forma dispersa pela matriz, isolados dentro de defeitos estruturais. Medidas de XPS mostram que os níveis eletrônicos de caroço dos átomos de xenônio são alterados em função do sítio de aprisionamento, que por sua vez, estão relacionados com a energia de implantação. Um estudo utilizando uma combinação de XPS e XAES permitiu identificar e separar as contribuições relativas às alterações eletrônicas observadas nos átomos de Xe devido a inserção deles dentro do a-Si. A análise destes resultados, XPS/XAES, forneceu indícios sobre os tamanhos das cavidades hospedeiras dos átomos de Xe. Simulações teóricas de XAS foram bastante satisfatórias e corroboram com as observações realizadas por XPS/XAES e indicam timidamente formações de aglomerados para alguns casos, ou seja, minoritariamente. Dependendo da energia de implantação, a dinâmica de crescimento do filme é alterada possibilitando estruturas diferentes na matriz. Foi também constatado que o aumento da densidade do a-Si por efeito de compactação pode ser mascarado pela introdução de elementos com massas relativamente grandes, como por exemplo, Xe / Abstract: A study of xenon (Xe) implantation in amorphous silicon (a-Si) by Ion Beam Assisted Deposition (IBAD) technique is investigated in this thesis. A Xe ion beam with energy of 1500 eV, obtained from a Kaufman gun, was used for sputtering a Si target, depositing an a-Si film on the substrate holder placed 15 cm apart. Simultaneously, another Xe beam with variable energy (0 ¿ 300 eV), was used to bombard the a-Si for the implantation of Xe atoms. It was obtained Xenon concentration up to 5%, using energy as low as 50 eV, which is orders of magnitude smaller than those usually used by conventional implantation, that requires thousands of eV. For the investigation of the Xe atoms in the a-Si matrix, XAS (X-Ray Absorption Spectroscopy), XPS (X-Ray Photoelectron Spectroscopy) and XAES (X-Ray Auger Excited Spectroscopy) techniques and theoretical simulations of XAS were used. The experimental results of XAS showed that Xe atoms are dispersed in the a-Si matrix as isolated element majoritively into structural defects. XPS measurement shows that the electronic core levels of the Xe atoms are shifted as function of the trapping site, which in turn, are related to the energy implantation. The use of XPS and XAES allowed the identification and the separation of the contribution to the shift in the binding energy due to the relaxation energy. The analysis of this results, XPS/XAES, provided indications as to size of the cavity containing the trapped Xe. XAS theoretical simulations gave satisfactory results and corroborate with XPS/XAES observations and tentativelly indicated small clusters, for a minority of cases. Depending on the implantation energy, the growth dynamics of the films is changed, allowing different structures in the matrix. It was also verified that the increase in the density of the film by the packing effect is obscured by the introduction of heavier elements, for example, Xe / Doutorado / Física da Matéria Condensada / Doutor em Ciências

Ce travail a pour but d'implémenter des procédés d’implantation ionique pour des cellules solaires à hétérojonctions de silicium (SHJ) afin d'en simplifier le procédé de fabrication ou d’en augmenter les performances.Nous avons d'abord étudié le procédé pour réaliser le dopage des couches de silicium amorphe hydrogéné (a-Si:H). Par ce nouveau procédé, il est possible de réaliser des dopages localisés de manière simple, à travers des masques, ce qui peut permettre une diminution des coûts de fabrication de certains types de cellules SHJ comme les cellules à contacts interdigités à l'arrière. Les implantations de phosphore et de bore ont été étudiées, pour la réalisation de dopage respectivement de type n et p. Les comportements et les conclusions sont très différents pour ces deux types de dopage. Le phosphore étant plus lourd que le bore, il est possible de l'implanter dans des couches très minces sans endommager fortement l'interface avec le silicium cristallin, mais la création très importante de défauts dans le a-Si:H, résistant à des recuits post-implantation, conduit à de fortes dégradations des propriétés électriques du a-Si:H, et il n'a pas été possible d'atteindre des niveaux de conductivité suffisants. Au contraire, pour le bore, conformément à des résultats de la littérature, les atomes sont activés plus facilement par un recuit post-implantation grâce à la forte diminution de la concentration de défauts localisés. Cependant, le bore, implanté plus profondément, atteint e plus facilement l'interface, ce qui nécessite des recuits à plus haute température pour guérir les défauts d'interface. Néanmoins, pour des couches de a-Si:H de l'ordre de 25 nm, nous avons pu trouver des conditions technologiques permettant d'obtenir des propriétés comparables à celles obtenues par le procédé classique de dépôt de (p) a Si:H assisté par plasma, à savoir des valeurs élevées de conductivités du a-Si:H (10-4 Ω-1cm-1) et de passivation d’interface (i VOC > 700 mV).Une deuxième partie de ce travail est consacrée à l’étude d’une nouvelle cellule, dite à homo hétérojonction de silicium (HHJ) comprenant un homo-émetteur additionnel (p+) c-Si à l’hétéro-interface côté émetteur. Le but est d’améliorer la passivation de l’interface afin d’augmenter le rendement de la cellule. Des simulations numériques ont mis en évidence une augmentation de FF de la cellule HHJ, que nous avons pu attribuer à une meilleure passivation par effet de champ et à une diminution de la résistance globale du a-Si:H due à des modifications des courbures de bandes. Elles ont aussi montré la nécessité d’un homo-émetteur suffisamment mince et fortement dopé (5×1018 cm-3). De ce fait, nous avons utilisé le procédé d’implantation ionique pour développer des profils de bore adéquats et avons pu vérifier expérimentalement que l'incorporation de la couche de (p+) c-Si permet la diminution de la résistance de contact et l'amélioration de la passivation de l'interface (i) a-Si:H/(p+) c-Si par effet de champ lorsque la concentration de bore en surface n'est pas trop importante. Ces deux améliorations ont pu être concrétisées dans la réalisation de cellules présentant une amélioration du facteur de forme et de meilleurs rendements de conversion par rapport à des cellules SHJ de référence. Cette réalisation constitue la première preuve de concept pour les cellules de type HHJ. / This work aims at investigating the use of ion implantation to process silicon heterojunction solar cells (SHJ) in order to improve the ratio of cost to produced power (€/Wp) of the cells either by cost reduction due to manufacturing simplification or by increase of the cell performance.A first part of the work consists in doping hydrogenated amorphous silicon (a-Si:H) layers by ion implantation. Using hard masks, doping of localized regions required in cell architectures like interdigitated back contact cells can thus be easily achieved at lower cost. Both boron and phosphorus implantation have been studied for p- and n-type doping, respectively. These two types behave very differently. Phosphorous being heavier than boron, very shallow implantation can be achieved on thin a-Si:H layers onto crystalline wafers without damaging the interface. However very high defect densities are created in a-Si:H which cannot be annealed out by post-implantation annealing treatments. Therefore it was not possible to reach conductivity values suitable for solar cell applications. For B implantation, consistently with previous work, the activation of B atoms has been achieved upon annealing thanks to a decrease of localized bandgap states. Also, boron can penetrate deeper and reach high concentration at the a-Si:H/c-Si interface, which requires higher temperature annealing compared to P implantation to recover a good interface passivation quality. Nevertheless, for a-Si:H layers of about 25 nm process conditions allowing similar properties to PECVD-doped (p) a-Si:H deposition (i.e. conductivity of 10-4 Ω-1cm-1 and interface passivation allowing i-VOC > 700 mV) have been obtained.A second study is dedicated to the study of a new cell concept, named silicon homo-heterojunction (HHJ) which comprise an additional homo-emitter (p+) c-Si at the emitter interface. The goal is to improve the interface passivation in order to increase the cell efficiency. Numerical simulations have evidenced an improved fill factor in this cell that is attributed to a field effect passivation improvement and a decrease in series resistance related to band bending changes in the a-Si:H layers. The need of sufficiently shallow and strongly doped (> 5×1018 cm-3) emitter has also been evidenced. Therefore, ion implantation has been used to develop suitable boron profiles and both the increase in fill factor and the decrease in contact resistances have been obtained when the boron surface concentration is not too high. These improvements have been validated by processing HHJ solar cells that exhibit a fill factor improvement and an improved efficiency compared to SHJ cells. This achievement is a first proof of concept of the HHJ architecture.

Les cellules à hétérojonctions de silicium (HET-Si) sont basées sur un substrat de silicium cristallin (c-Si) dopé n (p), une couche très fine de passivation (en général du silicium amorphe (a-Si:H) non dopé), et une couche d’une dizaine de nanomètres de silicium amorphe dopé p (n). Ces cellules atteignent aujourd’hui des rendements de l’ordre de 26% (record de 26,6% par l’entreprise Kaneka en 2017). Un des axes importants de recherche sur les cellules HET-Si porte sur l’étude de l’interface c-Si/a-Si:H qui est un élément clé dans le rendement des cellules. Ce rendement dépend en particulier de la présence d’états recombinants à l’interface c-Si/a-Si:H. Nous nous sommes donc tout particulièrement intéressés aux défauts d’interface en développant un calcul basé sur le modèle du réservoir de défauts (Defect-Pool Model ou DPM) dans le silicium amorphe et en corrélant nos résultats de modélisation avec des résultats expérimentaux de mesure de durée de vie. Afin de déterminer les caractéristiques de l’interface c-Si/a-Si:H, nous avons procédé comme suit : (1) Calcul de la densité d’états (DOS) volumique dans les couches de a-Si:H (dopé et non dopé), en nous appuyant sur le DPM. Dans ce modèle, la DOS varie en fonction notamment de la position du niveau de Fermi par rapport au bord de bande. La courbure des bandes de la jonction a-Si:H/c-Si implique ainsi une variation spatiale de la DOS dans le a-Si:H. (2) Calcul de la DOS surfacique à l'interface par projection des états volumiques présents à l’interface dans le a-Si:H. (3) Calcul des taux de recombinaison puis de la durée de vie effective sur des structures symétriques a-Si:H/c-Si/a-Si:H et comparaison avec des résultats expérimentaux. Nous avons ainsi pu étudier l’impact des paramètres matériaux du a-Si:H sur la durée de vie effective des porteurs minoritaires. L’évolution de la durée de vie avec les paramètres du a-Si:H est parfois contre-intuitive car deux phénomènes de passivation liés à la position du niveau de Fermi à l’interface s’opposent : passivation par la diminution de la densité d’états à l’interface et passivation par effet de champ. Le seul calcul de la DOS à l’interface ne suffit pas toujours à expliquer les variations de durées de vie, un calcul complet sous lumière est nécessaire. Nous avons montré que l’impact de certains paramètres du DPM peut-être grand sur la DOS mais faible sur la durée de vie effective à cause de cette compensation entre les phénomènes de passivation. Nous avons également étudié des structures correspondant aux faces avant : (p)a-Si:H/(i)a-Si:H/(n)c-Si(PIn) et arrière : (n)a-Si:H/(i)a-Si:H/(n)c-Si(NIn) des cellules HET-Si. Nos simulations permettent de montrer que les interfaces NIn sont moins critiques en terme de recombinaisons que les interfaces de type PIn. Nous montrons que la recombinaison aux interfaces PIn est dominée par la capture des électrons par les liaisons brisées de silicium chargées positivement. Nous montrons également que l’énergie d’Urbach est un paramètre qui joue de manière importante dans le calcul de la durée de vie effective et que l’utilisation de valeurs fixes de cette énergie d’Urbach dans la couche de passivation ne permet pas de reproduire les tendances expérimentales dans les structures avec des interfaces PIn. Nous proposons un modèle de variation de l’énergie d’Urbach avec l’épaisseur de la couche de passivation, qui permet de reproduire les tendances expérimentales pour les faibles épaisseurs de la couche de passivation mais qui demande à être complété pour de plus grandes épaisseurs. / Silicon heterojunction (Si- HET) solar cells are based on an n-doped (p-doped) crystalline silicon (c-Si) substrate, a very thin (a few nanometers) passivation layer of undoped hydrogenated amorphous silicon (a-Si:H) and a layer of p-doped (n-doped) a-Si:H, approximately 10 nanometer- thick. These cells currently lead the performance of silicon solar cells with conversion efficiencies in the order of 26% (with a record of 26.6% being achieved by the Kaneka company in 2017). One of the major focal points of research in Si- HET cells is the study of the c-Si/a-Si:H interface, which is a key factor in the cells' efficiency. In particular, this efficiency is strongly dependent on the recombination states at the interface between c-Si and a-Si:H. We therefore focused on developing a model of recombination through interface defects, which were evaluated based on the Defect-Pool Model (DPM) in a-Si:H. We calculated the effective lifetime vs excess carrier density curves and their dependence on the undoped a-Si:H passivation layer thickness and compared them to experimental results.In order to determine the characteristics of the c-Si/a-Si:H interface, we proceeded as follows: (1) Calculation of the volumic density of states (DOS) in a-Si:H layers (doped and undoped) using the DPM. In this model, the DOS varies as a function of the position of the Fermi level in relation to the band edge. The band bending at the a-Si:H/c-Si interface thus implies a spatial variation of the DOS in a-Si:H. (2) Calculation of the surface DOS at the interface by projection from the volumic states present in a-Si:H at the interface. (3) Calculation of the recombination rates and of the effective lifetime curves for symmetrical a-Si:H/c-Si/a-Si:H structures and comparison with experimental results. Thus we were able to study the impact of material parameters of a-Si:H on the effective lifetime curves. The change in lifetime as a function of a-Si:H parameters is sometimes counter-intuitive because two passivation mechanisms, namely passivation by field-effect or by the reduction of the DOS at the a-Si:H/c-Si interface, have opposed behavior in relation to the position of the Fermi level at the interface. A simple calculation of the DOS at the interface is not, therefore, sufficient to explain variations in lifetime, and a complete calculation of effective lifetime under illumination is required and has been performed. We demonstrate that the impact of certain DPM parameters may have a significant effect on the DOS but only a minor effect on the effective lifetime due to the compensation by the field-effect passivation. Moreover we have studied both types of silicon heterojunctions, (p)a-Si:H/(i)a-Si:H/(n)c-Si(PIn), and (n)a-Si:H/(i)a-Si:H/(n)c-Si(NIn) that are used as front emitter and back surface field junctions, respectively, in double-side contacted silicon Si-HET solar cells. Our simulations allowed us to emphasize that NIn interfaces are less critical in terms of recombination than PIn interfaces. We demonstrate that recombination at PIn interfaces is dominated by the capture of electrons by positively charged silicon dangling bonds. We further show that the Urbach energy is the major a-Si:H parameter that determines the effective lifetime in Si-HET solar cells and that the use of fixed values for this Urbach energy in the passivation layer whatever the layer thickness does not permit the experimental trends of PIn interfaces to be reproduced. Instead, we propose a model featuring that the Urbach energy decreases with the thickness of the passivation layer, which does allow experimental trends to be reproduced for very thin passivation layers (< 10 nm), but which requires further elaboration for larger thicknesses, for instance with a combined bandgap variation.

No description available.

No description available.