Search

International audience; As part of global efforts to thwart the climate change challenge, the integration of renewable energy systems, including an increasingly dominant solar sector, would contribute by 36% to the reduction of carbon related emissions. Among current photovoltaic technologies, the focus is on silicon-based technologies dominating the photovoltaic market with a share of 95% of the total production in 2017. Despite nearly-zero emissions during the operation, non-negligible environmental impacts of these systems can be associated with the manufacturing stage, which contributes to 80-90 % of the engendered impacts over their lifetime. The electricity mix is a key variable parameter among others affecting the environmental impact in particular manufacturing and recycling phases. In order to tackle this variability, the Life Cycle Assessment method along with comprehensive reviews of latest crystalline-silicon advancements are used in this work to develop a parameterized model. The approach does not only provide a scientific multi-criterion support monitored by more than 25 input variables but also includes updates to current databases for photovoltaic systems, dating from 2005. This parameterization shows the relevance of the model's modular aspect allowing simulations of various scenarios, ensuring up-to-date LCA data, and serving uncertainty and sensitivity analyses in the future.

Silver/aluminum (Ag/Al) paste has been used as metallization for p + emitter of n-type solar cells. Nevertheless, the Ag/Al paste induces junction current leakage or shunting in the solar cells, resulting loss in open circuit voltage (V oc ). However, the details still are not known about how glass frit and aluminum in the paste affect the p + emitter, and result in the electrical losses, respectively. Furthermore, it is not still clear whether and how the aluminum addition induces the electrical losses. In this study, the “floating contact method” proposed by R. Hoenig was applied for the measurement to investigate the respective effect of glass frit and aluminum on the electrical losses. Conductive paste with the glass frit for the p + emitter induces loss in V oc of the cells even if the paste contains no aluminum. The glass frit in the Ag/Al paste induces carrier recombination and significant shunting of p-n junction, but the aluminum in the paste mitigates these electrical losses because of weakening emergence of Ag-crystallites on the p + emitter.

We report an 8 µm-thick silicon solar cell with an efficiency of 9.60%. Nanostructured silicon surface formed via metal assisted etching shows a broadband reflection below 10%. Despite the excellent optical performance, a moderate short-circuit current (J SC ) of 25.44 mA/cm 2 was collected. Relatively low external quantum efficiency (EQE) at short wavelengths was associated with carrier recombination at the enhanced surface and by the Auger process. Moreover, parasitic absorption at the back contact is the main factor resulting a relatively low EQE in long wavelength region of the spectrum. Our optical simulations show that planarization of the rear Si surface and insertion of a low refractive index dielectric spacer between Si and the rear metal can significantly reduce the parasitic absorption in the metal, resulting in J SC values over 35 mA/cm 2 .

35th European Photovoltaic Solar Energy Conference and Exhibition; 653-656, In the last years, our research group has focused on the development of Interdigitated Back Contacted (IBC) solar cells based on transition metal oxides. In particular, a very thin Al2O3/TiO2 stack deposited by ALD has demonstrated excellent results. The point of this work is to try to develop a transparent conductive oxide that could be inserted in between the passivating layers and the metal contact in order to improve rear reflector properties. As a good candidate, we focus on Zinc Oxide (ZnO) deposited from DEZ (Ditehylzinc) and water as gas precursors. A deposition rate of 1.76 A/cycle is found for layers deposited at 125 ºC indicating a self-limited reaction, i.e. atomic layer deposition process. This film has an optical band gap of 3.26 eV demonstrating good optical properties to work as a back reflector. Unfortunately, contact quality with these layers is very poor. In order to investigate the origin of such a poor performance, we measure ZnO workfunction by Scanning Kelvin Probe resulting in 4.85 eV. This value suggests that the developed ZnO films could work better in a hole selective contact. To demonstrate this point, contacts on MoOx layers were fabricated leading to reasonable contact resistance values of 43 mcm. Passivation properties of these contacts will be addressed in the future in order to consider their introduction into the final IBC solar cell.

Increasing efficiency of solar cells is still a discussed problem. Even if it is well-known that specially formed substrates as Asahi U-type for solar cells are produced, there is still a continuing attention given to the applications of surface roughness to achieve better light trapping and absorptance in solar cells. It was found out the even an exact interface morphology can play an important role in light trapping. In this paper we focused on the issue how final absorptance of a solar cell structure could be affected and possibly increased. The goal of this article is to show which of interfaces has the greatest influence on specular absorptance of the whole structure.

In this work 3D simulations are used to study the impact of technological parameters on device performance of c-Si interdigitated back-contacted IBC solar cells with point-like doped contacts. In these cells, the highly-doped regions are defined in a point-like structure instead of the more classical arrangement of fully doped fingers. Numerical simulations allow us to optimize rear contact geometry, i.e. optimum pitch between contacts, depending on the substrate resistivity and the passivation quality of base contacts. Results show a trade-off between recombination and base resistive losses demonstrating limit efficiencies over 27% for a perfectly passivated structure on p- and n-type substrates. More realistic devices where state-of-the-art surface passivation is considered reach efficiencies beyond 22% on 2 +- 1 ohmcm substrates with optimum pitch values of 200 +- 50 µm.

All process steps in the fabrication of amorphous / crystalline (a-Si:H/c-Si) silicon heterojunction solar cells affect interface recombination and thus have an impact on the open circuit voltage of the final solar cell. Transient photoconductance decay is a convenient method to assess the charge carrier lifetime which is a direct measure for recombination and thus the interface quality. In this contribution we present a step-by-step analysis and evaluation of the charge carrier lifetime throughout the processing of the solar cells starting from the saw-damage etched Si wafer to the complete solar cell device. Particular emphasis is put on the optimization of the random pyramid texture. For this, surface morphologies and a-Si:H/c-Si interface properties were quantitatively evaluated and optimized with respect to low recombination and large carrier lifetimes. Based on this optimization a-Si:H/c-Si heterojunction solar cells with conversion efficiencies exceeding 20% were prepared on c-Si wafers textured in isopropanol(IPA)-free alkaline solution.

33rd European Photovoltaic Solar Energy Conference and Exhibition; 857-860, Black silicon (b-Si) reduces drastically light reflectance in the front side of c-Si solar cells to values near zero for the whole absorbed solar spectrum. In this work, we apply 2D and 3D simulations to explore the efficiency limits of interdigitated back-contacted c-Si(n) solar cells with line or point contacts respectively, using ALD Al2O3 passivated b-Si in the front surface. Realistic physical and technological parameters involved in a conventional oven-based fabrication process are considered in the simulations, especially those related to surface recombination on the b-Si as well as high doped p+/n+ strip regions. One important issue is the temporal stability of surface passivation on b-Si surfaces. In this work experimental long-term b-Si surface passivation data after two years and its impact on cell performance are studied. Simulations demonstrate initial and final photovoltaic efficiencies over 24.6% and 23.2% respectively for an emitter coverage of 80% independently of the cell contact strategy. A photocurrent loss about 1.3 mA/cm2 occurs when surface recombination velocity at the b-Si surfaces degrades from 6 cm/s to a final value of 28 cm/s.

International audience; An in-depth study of metal assisted chemical etching (MACE) of p-type c-Si in HF/H2O2 aqueous solutions using Pt nanoparticles as catalysts is presented. Combination of cyclic voltammetry, open circuit measurements, chronoamperometry, impedance spectroscopy, and 2D band bendingmodeling of the metal/semiconductor/electrolyte interfaces at the nanoscale and under different etching conditions allows gaining physical insights into this system. Additionally, in an attempt to mimic the etching conditions, the modeling has been performed with a positively biased nanoparticle buried in the Si substrate. Following these findings, the application of an external polarization during etching is introduced as a novel efficient approach for achieving straightforward control of the pore morphology by acting upon the band bending at the Si/electrolyte junction. In this way, nanostructures ranging from straight mesopores to cone-shaped macropores are obtained as the Si sample is biased from negative to positive potentials. Remarkably, macroscopic cone-shaped pores in the 1−5 μm size range with a high aspect ratio (L/W ∼ 1.6) are obtained by this method. This morphology leads to a reduction of the surface reflectance below 5% over the entire VIS-NIR domain, which outperforms macrostructures made by state of the art texturization techniques for Si solar cells.

The way to permanently recover the well-known Light-Induced Degradation (LID) which affects the p-type Cz-Si PERC solar cells represents one of the main challenges of photovoltaic research. In this work we have set up a numerical simulations flow which allows us to reproduce the experimental measured values of figures of merit (FOMs) of four different Cz-PERC solar cells lots subjected to a degradation and two regeneration processes. The recombination centres in bulk and the Boron-Oxygen complexes (B-O) are modeled by means of two trap levels tuned on the basis of experimental data. From simulations we confirm that the FOM degradation levels off after 16hours and the regeneration process characterized by relatively long time process is preferred in terms of performance recovery. In addition, further cells with different passivation films are analyzed by adopting the same methodology.

Two applications of the Green's expression for saturation current vs. bandgap are proposed. The first one aims at assessing the energy bandgap, Eg, in the quasi-neutral regions (QNR) of Si and CIGS solar cells, which is associated to the saturation current in those regions, J0,QNR. Values for J0,QNR are obtained from both fits to pseudo-dark I-V curves and the textbook expression for open-circuit voltage. The second application deals with detecting the recombination mechanism prevailing in the QNR of CIGS solar cells from the order of magnitude of the value that must be assigned to a parameter of the Green's expression to fit measured curves of performance data vs. Eg. The data utilized in the two applications are both original and taken from the literature. © 2015 IEEE.

International audience; We present here an UV (266nm) photoemission setup dedicated to measure properties of conductive materials under DC extraction field as photocathodes. We have successfully tested copper, as reference material, and silicon samples. It allowed us testing photoemission properties of thin CVD nanodiamond films on silicon substrates. We demonstrate a strong influence on silicon doping type on the photoemission yield, pointing out a clear influence of the nanodiamond-silicon interface in the photoemission process. Furthermore, the nanodiamond-silicon structure exhibits one order of magnitude higher photoemission current compared to copper test samples.

1st Africa Photovoltaic Solar Energy Conference and Exhibition Proceedings, AfricaPVSEC 2014

Nous avons observé une augmentation ‘’transient’’du taux de cristallisation interfacique de l’a-Si lorsqu’on réimplante du Si à proximité de l’interface amorphe/cristal. Après amorphisation et traitement thermique à 650°C pendant 5s de la couche a-Si crée par implantation ionique, une partie a été réimplantée. Les défauts produits par auto-réimplantation à 0.7MeV se trouvent à (302±9) nm de l’interface initiale. Cela nous a permis d’étudier d’avantage la variation initiale de la vitesse SPE (Épitaxie en phase solide). Avec des recuit identiques de 4h à 500°C, nous avons déterminé les positions successives des interfaces et en déduit les taux de cristallisation SPE. La cristallisation débute à l’interface et continue graduellement vers la surface. Après le premier recuit, (252±11) nm s’est recristallisé dans la zone réimplantée soit un avancement SPE de 1.26x10^18at./cm2. Cette valeur est environ 1.50 fois plus importante que celle dans l’état relaxé. Nous suggérons que la présence de défauts à proximité de l’interface a stimulé la vitesse initiale. Avec le nombre de recuit, l’écart entre les vitesses diminue, les deux régions se cristallisent presque à la même vitesse. Les mesures Raman prises avant le SPE et après chaque recuit ont permis de quantifier l’état de relaxation de l’a-Si et le transfert de l’état dé-relaxé à relaxé., We observed a ‘’transient’’ increase of planar crystallization rate of a-Si when one reimplanted Si near the interface amorphous / crystal. After amorphization and heat treatment at 650°C for 5s, one part has been re-implanted. The defects produced at 0.7 MeV by self-re-implantation are located at (302±9) nm of the initial interface. This allows us to better study the initial variation of SPE speed (solid phase epitaxy). With recrystallisation anneals at 500±4°C for 4h, we have determined the successive positions of the interfaces and have deduced the SPE recrystallization rate. Crystallization began at the interface and continues gradually to the surface. After the first annealing, (252±11)nm was recrystallized in the re-implanted state. That means 1.26x10^18at./cm2 SPE enhancement. This value is approximately 1.50 times greater than that in the relaxed state. We suggest that the presence of defects near the interface stimulate the speed. Raman measurements taken after each annealing allowed us to know the transfer of the un-relaxed state to the relaxed state. After the number of anneals treatments, both areas progress almost at the same speed, Dans ce travail nous avons étudié le phénomène de diffusion du cuivre et de l’argent dans a-Si en présence de l’hydrogène à la température de la pièce et de recuit. Une couche amorphe de 0.9μm d’épaisseur a été produite par implantation de 28Si+ à 500 keV sur le c-Si (100). Après celle-ci, on procède à l’implantation du Cu et de l’Ag. Un traitement thermique a produit une distribution uniforme des impuretés dans la couche amorphe et la relaxation de défauts substantiels. Certains défauts dans a-Si sont de type lacune peuvent agir comme des pièges pour la mobilité du Cu et de l’Ag. L’hydrogène implanté après traitement thermique sert à dé-piéger les impuretés métalliques dans certaines conditions. Nous n’avons détecté aucune diffusion à la température de la pièce au bout d’un an, par contre un an après à la température de recuit (1h à 450°C) on observe la diffusion de ces métaux. Ce qui impliquerait qu’à la température de la pièce, même si l’hydrogène a dé-piégé les métaux mais ces derniers n’ont pas pu franchir une barrière d’énergie nécessaire pour migrer dans le réseau., In this work we studied the diffusion phenomenon of copper and silver in a-Si in the presence of hydrogen at room temperature and annealing temperature. The 0.9 μm -thick a-Si layers were formed by ion implantation 28Si + at 500 keV on c-Si (100). After this Cu ions and Ag ions were implanted at 90keV.The heat treatment produces a uniform distribution of impurities in the amorphous layer and the relaxation of substantial defects. Vacancies defects in a-Si can act as traps for the mobility of Cu and Ag. Hydrogen implanted is used to de-trap metal impurities such as Cu and Ag. However we did not detect any diffusion at room temperature during 1 year, but after one year at the annealing temperature (450°C for 1h) we observe the distribution of these metals. Implying that the room at temperature, although the hydrogen de-trapping metals but they could not crossed an energy barrier required to migrate in the network.

28th European Photovoltaic Solar Energy Conference and Exhibition; 1607-1610, The back-contact back-junction BC-BJ solar cell concept is a promising photovoltaic structure for both laboratory and industrial c-Si solar cells. High efficiency devices based on this concept have been reported in the past using either diffused regions or applying the heterojunction with intrinsic thin layer HIT concept to perform both base and emitter contacts. In this work we use 3D numerical simulations to study the impact of technological parameters on device performance of c-Si(p) BC-BJ solar cells with point-like doped contacts. Numerical simulations allow us to optimize rear contact geometry as a trade-off between recombination and base resistive losses, leading to photovoltaic efficiencies higher than 18.3% and up to 22.3% on 2.2 cm FZ substrates depending on the back contact pattern and the passivation quality of base contacts.

27th European Photovoltaic Solar Energy Conference and Exhibition; 1733-1738, In the process of finding alternatives to conventional silver thick-film metallization for silicon solar cells the interest in chemical deposition of nickel layers is rising. Electroless plating of nickel/phosphorus layers is a common surface coating method to protect metallic surfaces from corrosion and is currently examined as an option for solar cell metallization. However, some chemical compositions for this process contain alkaline substances (e.g. sodium hydroxide or ammonium), which are known to chemically react with silicon. This work examines the effect of Ni/P plating on crystalline silicon substrates with varying n-type emitters (20 Ω/sq, 50 Ω/sq, 80 Ω/sq and 120 Ω/sq). Two different chemical bath compositions, both with an alkaline electrolyte, are tested at different temperatures and varying plating durations. By comparing the mass of Cz-Si wafers in different stages of the Ni/P plating process (before plating, after plating with the Ni/P layer intact, after plating with the Ni/P layer removed) one can detect an average silicon etching depth between 40 nm and 80 nm ± 13 nm. ECV measurements suggest an even higher damage to the emitter of >250 nm up to a complete consumption of a shallow emitter. SEM imaging reveals a morphological change of the wafer surface.

Low “Euros per Watt-peak” and ease of industrialization are the main drivers towards successful introduction on the market. In this regard, back-contact solar cells on n-type silicon offer significant benefits. The efficiency of back contact cells, such as Metal Wrap Through (MWT) cells, compared to the traditional H-pattern cells is higher at cell level, thanks to the reduced shading losses, and is higher at module level, thanks to the reduced interconnection resistance losses. N-type silicon benefits from improved electrical properties of n-type silicon compared to p-type (higher minority carrier diffusion lengths, lower sensitivity to many impurities). Furthermore, the availability of an industrial cell process designed by ECN, resulting in bifacial cells (good rear surface passivation and light trapping), makes n-type silicon a perfect candidate for high efficiency solar cells and requires only modest changes to the current wafer and cell production processes. In order to reduce processing costs and increase module efficiencies, we have started two years ago the development of the Metal-Wrap-Through (MWT) solar cell technology on n-type mono-crystalline silicon wafers. Within the last year, efficiency of our MWT silicon solar cells manufactured from ntype Cz silicon wafers has been improved by 1% absolute. Based on common industrial cell processing steps such as diffusion, screen-printing metallization and firing through, we have obtained efficiencies up to 19.70% (in-house measurements) on large area wafers (239 cm2, 5 Ωcm), with clear potential for further improvement. In this article, we present a first direct comparison experiment between n-type bifacial MWT and “conventional” n-type bifacial Hpattern technologies, in which an efficiency gain of 0.30% absolute for MWT is demonstrated. At the moment, series resistance and, as a result, fill factor are still sub-optimal. Nevertheless, with current density (Jsc) values approaching 40 mA/cm2 and open circuit voltage of 644 mV, n-type MWT solar cells already outperform n-type H-pattern cells manufactured with a comparable process.

In a recent previous work, we proposed a rotating polarizer-analyzer ellipsometer (RPAE) in which the two elements are rotating synchronously in the same direction with a speed ratio 1: 3. We applied this technique to bulk samples. In this work, we present theoretically the characterization of 100 nm SiO2 thin film using this spectroscopic RPAE. We assume a structure consisting of air (ambient)/SiO2 (thin film)/c-Si (substrate). The ellipsometric parameters ψ and Δ are calculated when a clean signal is received by the detector and when a hypothetical noise is imposed on this signal. The film thickness and the optical constants of the film are calculated for the noisy signal in the spectrum range 200-800 nm. The results are compared with the proposed thickness and with the accepted values for SiO2 optical constants.

Les siliciures métalliques constituent un élément crucial des contacts électriques des transistors que l'on retrouve au coeur des circuits intégrés modernes. À mesure qu'on réduit les dimensions de ces derniers apparaissent de graves problèmes de formation, liés par exemple à la limitation des processus par la faible densité de sites de germination. L'objectif de ce projet est d'étudier les mécanismes de synthèse de siliciures métalliques à très petite échelle, en particulier le NiSi, et de déterminer l’effet de l’endommagement du Si par implantation ionique sur la séquence de phase. Nous avons déterminé la séquence de formation des différentes phases du système Ni-Si d’échantillons possédant une couche de Si amorphe sur lesquels étaient déposés 10 nm de Ni. Celle-ci a été obtenue à partir de mesures de diffraction des rayons X résolue en temps et, pour des échantillons trempés à des températures critiques du processus, l’identité des phases et la composition et la microstructure ont été déterminées par mesures de figures de pôle, spectrométrie par rétrodiffusion Rutherford et microscopie électronique en transmission (TEM). Nous avons constaté que pour environ la moitié des échantillons, une réaction survenait spontanément avant le début du recuit thermique, le produit de la réaction étant du Ni2Si hexagonal, une phase instable à température de la pièce, mélangée à du NiSi. Dans de tels échantillons, la température de formation du NiSi, la phase d’intérêt pour la microélectronique, était significativement abaissée., Currently metal silicide constitutes a crucial component in the formation of electrical contacts for transistors that forms the heart of modern day integrated circuits. As we reduce the dimensions of the latter, we are faced with serious problems of formation, related for example to the process limitation due to the weak density of germination sites. The objective of this project is to study at small scale the synthesis mechanisms of metal silicide, in particular NiSi, and to determine the effect of Si implantation damage on the phase sequence. We have determined the different phase sequences of the Ni-Si system for samples composed of a 10 nm Ni surface layer deposited on a-Si. These were obtained by time resolved x-ray diffraction (TR-XRD) measurements. As for samples quenched at critical temperatures we identified the different phases, their composition and their microstructure were determined by pole figures, Rutherford back scattering (RBS) spectrometry and transmission electron microscopy (TEM). We noted that for approximately half the samples, a spontaneous reaction happened before annealing. The result of the reaction was hexagonal Ni2Si, a phase unstable at room temperature, mixed with NiSi. In theses samples, the temperature of formation for the phase of interest, the NiSi, was lower.

25th European Photovoltaic Solar Energy Conference and Exhibition / 5th World Conference on Photovoltaic Energy Conversion, 6-10 September 2010, Valencia, Spain; 1770-1773, In this work the influence of varied diffusion parameters for an industrial open tube POCl3 diffusion furnace upon the emitter saturation current density on monocrystalline silicon is investigated. Further on, the effect of phosphorus gettering on multicrystalline silicon and on the sheet resistance on both mono- and multicrystalline silicon is under investigation. In addition, diffusion profiles are determined using the ECV (Electrochemical Capacitance Voltage) technique. Aim of this work is to enhance the performance of lowly doped emitters (80- 140 Ω/sq) applied in a photolithography based high efficiency solar cell process with special respect to defect-rich block cast multicrystalline silicon material. Understanding the influence of temperature, time and gas flow variations during the diffusion process is very important to enhance solar cell performance especially for mc silicon. For such materials the POCl3 gettering effect and the defect kinetics during the diffusion and the cool down phase after the diffusion are of major interest besides the reliable contact formation and low emitter saturation currents resulting in a good blue response of solar cells. The experiments performed in this work demonstrate that different mono- and multicrystalline silicon materials can benefit from adapted diffusion recipes in terms of significantly reduced emitter saturation currents and increased bulk lifetimes resulting in enhanced solar cell efficiencies.

25th European Photovoltaic Solar Energy Conference and Exhibition / 5th World Conference on Photovoltaic Energy Conversion, 6-10 September 2010, Valencia, Spain; 1091-1096, Current industrial monocrystalline Cz Si solar cells based on screen-printing technology for contact formation and homogeneous emitter have an efficiency potential of around 18.4%. Apart from limitations at the rear side by the fully covering Al-BSF the front side is limiting e.g. by relatively high j0E values. This can be changed by selective emitter designs allowing a decoupling and separate optimization of the metalized and non-metalized areas. Several selective emitter concepts that are already in industrial mass production or close to it are presented, and their specialties and status concerning cell performance are demonstrated. Key issues to be considered are costeffectiveness, added complexity, additional benefits, reliability, and efficiency potential. The efficiency increase for best cells is around 0.5-0.6%abs and the current efficiency potential already demonstrated for all technologies is around 19.0%. Average efficiencies in industrial mass production for selected technologies are 18.5-18.6% for Cz and 17.1% for mc Si. The efficiency increase by selective emitter formation is higher for inline emitters, but selective emitters based on POCl3 show the highest absolute efficiency. By decreasing the phosphorous surface concentration, selective emitters are more sensitive to surface passivation and the use of a SiNx:H layer with a higher refractive index increases implied Voc values even further. Encapsulation under module glass and EVA reduces part of the gained jsc caused by a better blue response on cell level, but a calculation reveals that this extra loss for selective emitter solar cells is

Both mono- and multicrystalline p-type silicon wafers were used for the phosphorous ion implantation. After ion implantation the silicon is strongly disordered or amorphous within the ion range. Therefore subsequent annealing is required to remove the implantation damage and to activate the doping element. During FLA, only the wafer surface is heated homogeneously to very high temperatures at ms time scales, resulting in the annealing of the implantation damage and an electrical activation of the phosphorous. It is demonstrated that FLA at 800 oC for 20 ms even without preheating is sufficient to recrystallized implanted silicon. The highest recrystallization and the lowest resistivity were obtained after annealing at 1200 oC for 20 ms both for mono- and multicrystalline silicon wafers. Photoluminescence results point towards P - cluster formation at high annealing temperatures which effects on metal impurity gettering within the emitter.

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal.

24th European Photovoltaic Solar Energy Conference, 21-25 September 2009, Hamburg, Germany; 845-856, CrystalClear was an EU co-financed Integrated Project aimed at developing technology for waferbased silicon solar modules at 1 € per watt-peak manufacturing costs and a strongly improved environmental profile. The project consortium has selected a number of technologies that potentially comply with these aims as well as research that serves as basis for further developments (beyond the project aims). These technologies have been demonstrated in the form of demonstrators, i.e. full-size modules featuring all innovations necessary to comply with the aims. The project has shown that wafer-based multicrystalline-silicon solar modules can be produced at 1 € per watt-peak at a world-record efficiency of 16% and an energy pay-back time of less than 2 years in Southern Europe.

24th European Photovoltaic Solar Energy Conference, 21-25 September 2009, Hamburg, Germany; 1400-1405, Wafers for the PV industry are mainly sawn with a multi-wire slurry saw. This process is slow (it takes almost half a day to complete a cut) and generates a lot of waste: around half the silicon is sawn away and contaminating the slurry, and the wire is worn and has lost strength. After each cut, the slurry has to be cleaned from the silicon debris and the wire has to be exchanged. In contrast, sawing the wafers with a diamond-plated wire is faster, requires only a cooling liquid that is easy to filter from silicon debris and uses a wire that can be kept for several cut. But this new sawing technique only has a chance to develop if the solar cell production lines developed for slurry sawn wafers is capable of processing these diamond-plated wire sawn wafers efficiently. This study focused on the differences of surface properties of wafers cut via a slurry wire-saw and via a diamond-plated wire-saw. From these surface differences, it is possible to explain the differences in cell processing behaviour and to update the cell production line. Finally, it is shown that wafers sawn with a diamond-plated wire can give cells that are as efficient as the slurry sawn wafers, which validates this new diamond-plated wire wafering method for the production of solar cells.

The Photovoltaics Group at the Universitat Politècnica de Catalunya is investigating silicon carbide (SiC) for the electronic passivation of the surface of crystalline silicon solar cells. The doping of SiC passivation layers with phosphorus resulted in a clear improvement of the minority carrier lifetime compared to undoped SiC films applied on p-type doped substrate wafers. In a further approach, the passivation layers simultaneously served as a doping source in order to create an emitter in the underlying p-type doped silicon substrate. The aim of this Master Thesis is to establish a process for growing silicon nitride (SiN) on p-type doped crystalline silicon substrates from Plasma Enhanced Chemical Vapor Deposition (PECVD). Photoconductance measurements on the samples analyse the passivating behaviour of the SiN films. The passivation layers are doped by adding phosphine to the process gas, and further measurements investigate the effect of doping on the carrier lifetime. Furthermore, a thermal treatment initiates the diffusion of phosphorus in order to create an emitter in the substrate underneath. Sheet resistance (Rsq) and the lower limit of the emitter saturation current density (J0e) deduced from lifetime measurements, characterise the quality of the emitters. Eventually, a cell prototype finishes this work integrating both concepts of using SiN as a passivation layer, and of diffusing an emitter from the very same SiN layer.

23rd European Photovoltaic Solar Energy Conference and Exhibition, 1-5 September 2008, Valencia, Spain; 1388-1392, c-Si samples whose surface passivation is based on highly inverted surfaces experimentally show a decrease in effective lifetime at low illumination levels while a flat dependence is expected from the current theoretical understanding. As an explanation for this discrepancy we propose a defect-rich bulk region close to c-Si surface. From an analytical approach, we deduce that the final lifetime dependence is a combination between an emitter-like term and a quasi-neutral bulk recombination term. Next, we use numerical modeling to simulate eff(n) curves depending on damage depth, lifetime in the damaged region and the doping density. Simulations show that inverted surfaces in highly-doped substrates are more sensitive to shallow defect regions. On the other hand, very lowly doped c-Si substrates show a lifetime reduction at high injection and the presence of a small “hump” in the lifetime when overall recombination becomes Auger-dominated. Finally, we show that experimental eff(n) curves of a-SiNx:H and Al2O3 c-Si passivated c-Si samples demonstrate the predicted characteristics which could not properly be accounted for previously.

23rd European Photovoltaic Solar Energy Conference and Exhibition, 1-5 September 2008, Valencia, Spain; 1393-1396, In this work we study aluminium laser-fired contacts for intrinsic amorphous silicon layers deposited by Hot-Wire CVD. This structure could be used as an alternative low temperature back contact for rear passivated heterojunction solar cells. An infrared Nd:YAG laser (1064 nm) has been used to locally fire the aluminium through thin amorphous silicon passivating layers. We discuss here the optimization of the laser firing parameters, very low specific contact resistances (up to rc= 4,6 mW.cm2) have been obtained on 2.8 W·cm p-type c-Si wafers but with a strong damage of the sample. While for a set of optimized laser conditions we obtain a low specific contact resistance (rc= 12,2 mW.cm2) with a low damage rate (19,96 %). Moreover this investigation focuses on maintaining the passivation quality of the interface

En este trabajo se ha fabricado un equipo de medida que combina la alta temperatura con transmisión óptica, utilizando el espectro de emisión de un LED comercial de bajo coste de adquisición.A partir de las medidas de transmisión óptica se ha obtenido el coeficiente de absorción y a través de este conocer la temperatura del substrato de c-Si:Así como para conocer los mecanismos fundamentales que intervienen en el proceso de la absorción óptica (Banda a banda y Cargas libres). Se ha ampliado el modelo teórico de calculo del coeficiente de absorción para ajustar las medidas echas con el espectro de emisión de un LED.El equipo construido se ha empleado para monitorizar in-situ la fase de cristalización de las capas de a-SiCx:H intrínsecas y dopadas con fósforo depositadas sobre substrato de c-Si. Los resultados obtenidos se han comparando con otras medidas ópticas que involucran técnicas diferentes: SEM, X-RD, FTIR, Monocromador, que confirman los resultados encontrados.