abstract: Solar energy, including solar heating, solar architecture, solar thermal electricity and solar photovoltaics, is one of the primary energy sources replacing fossil fuels. Being one of the most important techniques, significant research has been conducted in solar cell efficiency improvement. Simulation of various structures and materials of solar cells provides a deeper understanding of device operation and ways to improve their efficiency. Over the last two decades, polycrystalline thin-film Cadmium-Sulfide and Cadmium-Telluride (CdS/CdTe) solar cells fabricated on glass substrates have been considered as one of the most promising candidate in the photovoltaic technologies, for their similar efficiency and low costs when compared to traditional silicon-based solar cells. In this work a fast one dimensional time-dependent/steady-state drift-diffusion simulator, accelerated by adaptive non-uniform mesh and automatic time-step control, for modeling solar cells has been developed and has been used to simulate a CdS/CdTe solar cell. These models are used to reproduce transients of carrier transport in response to step-function signals of different bias and varied light intensity. The time-step control models are also used to help convergence in steady-state simulations where constrained material constants, such as carrier lifetimes in the order of nanosecond and carrier mobility in the order of 100 cm2/Vs, must be applied. / Dissertation/Thesis / M.S. Electrical Engineering 2013

abstract: Silicon solar cells with heterojunction carrier collectors based on a-Si/c-Si heterojunction (SHJ) have a potential to overcome the limitations of the conventional diffused junction solar cells and become the next industry standard manufacturing technology of solar cells. A brand feature of SHJ technology is ultrapassivated surfaces with already demonstrated 750 mV open circuit voltages (Voc) and 24.7% efficiency on large area solar cell. Despite very good results achieved in research and development, large volume manufacturing of high efficiency SHJ cells remains a fundamental challenge. The main objectives of this work were to develop a SHJ solar cell fabrication flow using industry compatible tools and processes in a pilot production environment, study the interactions between the used fabrication steps, identify the minimum set of optimization parameters and characterization techniques needed to achieve 20% baseline efficiency, and analyze the losses of power in fabricated SHJ cells by numerical and analytical modeling. This manuscript presents a detailed description of a SHJ solar cell fabrication flow developed at ASU Solar Power Laboratory (SPL) which allows large area solar cells with >750 mV Voc. SHJ cells on 135 um thick 153 cm2 area wafers with 19.5% efficiency were fabricated. Passivation quality of (i)a-Si:H film, bulk conductivity of doped a-Si films, bulk conductivity of ITO, transmission of ITO and the thickness of all films were identified as the minimum set of optimization parameters necessary to set up a baseline high efficiency SHJ fabrication flow. The preparation of randomly textured wafers to minimize the concentration of surface impurities and to avoid epitaxial growth of a-Si films was found to be a key challenge in achieving a repeatable and uniform passivation. This work resolved this issue by using a multi-step cleaning process based on sequential oxidation in nitric/acetic acids, Piranha and RCA-b solutions. The developed process allowed state of the art surface passivation with perfect repeatability and negligible reflectance losses. Two additional studies demonstrated 750 mV local Voc on 50 micron thick SHJ solar cell and < 1 cm/s effective surface recombination velocity on n-type wafers passivated by a-Si/SiO2/SiNx stack. / Dissertation/Thesis / Ph.D. Electrical Engineering 2013

Conselho Nacional de Desenvolvimento CientÃfico e TecnolÃgico / Os objetivos deste trabalho foram sintetizar, caracterizar e empregar como corantes sensibilizadores os complexos [Ru(dcbpy)(bpy)(bqdi)], [Ru(dcbpy)(bpy)(bqdi-Cl)] e cis-[Ru(H2dcbpy)(bpy)(nic)Cl]PF6, em que H2dcbpy e dcbpy sÃo, respectivamente, o ligante 4,4â-dicarboxi-2,2â-bipiridina e sua forma deprotonada, bpy à 2,2â-bipiridina, bqdi à 1,2-benzoquinonadiimina, bqdi-Cl à 4-cloro-1,2-benzoquinonadiimina e nic à nicotinamida. Os complexos sintetizados foram caracterizados por tÃcnicas espectroscÃpicas e eletroquÃmica. Os espectros UV-vis apresentaram bandas na regiÃo do visÃvel (aproximadamente 520 nm) atribuÃdas a transiÃÃes do tipo MLCT com mÃximos diferentes do complexo de partida devido à modificaÃÃo da esfera de coordenaÃÃo. Os espectros vibracionais na regiÃo do infravermelho mostraram bandas caracterÃsticas do composto de partida e dos ligantes. Os espectros de RMN 1H obtidos para os compostos com bqdi e bqdi-Cl mostraram sinais referentes aos prÃtons destes ligantes coordenados na forma oxidada. A voltametria cÃclica mostrou processos relativos ao par redox RuIII/II positivamente deslocados em relaÃÃo ao composto de partida, apresentando-se irreversÃveis para os complexos contendo os ligantes bqdi e bqdi-Cl e quasi-reversÃvel para o complexo de nicotinamida. Foram montadas DSSCs utilizando os complexos sintetizados como sensibilizadores. Para os complexos [Ru(dcbpy)(bpy)(bqdi)] e [Ru(dcbpy)(bpy)(bqdi-Cl)], observou-se que nÃo houve conversÃo de luz em corrente elÃtrica. Para o complexo cis-[Ru(H2dcbpy)(bpy)(nic)Cl]PF6, verificou-se um potencial de circuito aberto de 610 mV, uma corrente de curto-circuito de 2,4 mA cm-2 e um fator de preenchimento de 0,73. Estes valores sÃo inferiores aos obtidos para o corante N3 nas mesmas condiÃÃes experimentais, mas sÃo indicativos promissores que demonstram que este complexo deve ser melhor investigado como sensibilizador em CÃlulas Solares Sensibilizadas por Corante. / Dye-Sensitized Solar Cells (DSSCs) are devices capable of converting light into electricity. DSSCs are based on the absorption of light by a dye, which injects electrons into the conduction band of a semiconductor. This study aims to synthesize, characterize and employ as sensitizers the complexes [Ru(dcbpy)(bpy)(bqdi)], [Ru(dcbpy)(bpy)(bqdi-Cl)] and cis-[Ru(H2dcbpy)(bpy)(nic)Cl]PF6, where H2dcbpy and dcbpy are, respectively, 4,4â-dicarboxy-2,2â-bipyridine and its deprotonated form, bpy is 2,2â-bipyridine, bqdi is 1,2-benzoquinonediimine, bqdi-Cl is 4-chloro-1,2-benzoquinonediimine and nic is nicotinamide. The complexes were characterized by spectroscopic and electrochemical techniques. These complexes showed broad metal-to-ligand charge transfer (MLCT) absorptions bands at about 520 nm. 1H NMR spectra for the compounds with bqdi and bqdi-Cl showed peaks consistent with the respective oxidized form. Cyclic voltammetry showed a significant positive shift in redox potential for the RuIII/II redox couple in comparison to the starting material. The photovoltaic performances of the solar cells based on these complexes are under investigation. The results for the sensitizers [Ru(dcbpy)(bpy)(bqdi)] and [Ru(dcbpy)(bpy)(bqdi-Cl)] were not promising. On the other hand, the results for cis-[Ru(H2dcbpy)(bpy)(nic)Cl]PF6 revealed a short-circuit current density of 2.4 mA cm-2, an open-circuit voltage of 610 mV and a fill factor of 0.73 under standard AM 1.5 sunlight. These results were lower than that obtained to N3, but it shows that this complex needs further investigation.

M.Sc. / The efficiency of commercial polycrystalline silicon solar cells is currently 12% and 15% in the case of single crystalline cells. It is possible to lose about half of the open circuit voltage due to inferior contacts on the cell. It is thus clear that inferior contacts can seriously impede the relative low efficiency and care should be taken to make good ohmic contacts. Experiments were done to evaluate the influence of several factors on the quality and stability of the contacts. 1 C2•cm p-type polycrystalline silicon and 3 52.cm n-type single crystalline silicon were primarily used for these experiments. Results of molybdenum contacts on n-type silicon are also presented and the problems with silver epoxy contacts are discussed. It was found that aluminium contacts on p-type polycrystaline silicon improve with temperature and time, while those on single crystaline n-type degrade with temperature and time. These changes are already present at room temperature and are attributed to solid state diffusion of the aluminium into the silicon. This results in a p + layer. In the case of contacts on p-type, the behaviour is that of a Schottky diode. After the solid state diffusion, it becomes possible for the charges to quantum mechanically tunnel through the p+ layer. This results in an improvement of the contact. The contacts on n-type however, are ohmic just after evaporation. Similar to the p-material, the p+ layer causes a p+-n-junction with the depletion layer primarily in the n-type material. This causes a degradation in the contact quality. It is possible to achieve good quality contacts on polycrystaline p-type material, by annealing the contacts above 500°C for one minute. These contacts however, are non-ideal. SEM photographs show that the silicon surface is crated by pits due to solid state diffusion. It is only at these pits that conduction through the Schottkybarrier is possible. Since the area of the pits constitutes only a portion of the total area, only a portion of the surface will partake in conduction. Contact resistance is always present. For pm sized contacts on integrated circuits, the spesific resistance is of the order of 10 -6 Q.cm2. Contacts on solar cells, however, are of millimetre dimensions and the spesific resistance can be four orders of magnitude larger. The conduction through the surface can be modelled as conduction through a surface that is constituted of a mixture of minute ohmic and diode surfaces.

Ph.D. / A one-dimensional numerical simulation of photovoltaic (PV) cells has been written, and has been designated RAUPV2. An algorithm for determining the optical generation rate profile, taking into account multiple internal reflections in a multilayer cell has been developed. A method which enables realistic boundary values to be calculated, using RAUPV2 itself, has been developed. This method allows all three boundary values (', Fn and Fp) at each surface, to be determined, without the need to specify any additional input parameters. A comprehensive set of input parameters for aSi:H PV cells has been established, in consultation with the literature. Dangling-bond theory has been described and input parameters for dangling-bond defects have been presented. The effect of surface states in the p-layer on the contact potential at the TCO/p interface has been investigated. It was found that there is an intimate relationship between the contact potential and the parameters pertaining to the surface states. A simple method has been demonstrated, which has allowed RAUPV2 to reproduce the J-V curve of an existing aSi:H PV cell. The method requires that only the dangling-bond concentration in the i-layer and the contact potential at the Sn02/P interface needs to be adjusted. Once the J- V curve had been generated, the simulation results were used to characterise the empirical cell, in both thermodynamic- and steady-state equilibrium. This simulated cell was designated the realistic cell. The effect of asymmetries in the input parameters, under carrier band mobility interchange, on the performance of p-i-n cells has been investigated. The results indicate that, while asymmetries in the gap state distributions do give rise to asymmetrical behaviour in the J- V curve, the effect is slight, and it is the positional asymmetry of the optical generation profile that is mostly responsible for the observed asymmetry in the J- V curve under mobility interchange. An investigation of the limiting carrier effect has led to the conclusion that, in a p-i-n aSi:H cell under forward bias, the electron is the limiting carrier. This has been explained by appealing to the form of the optical generation profile, since most electron-hole pairs (EHPs) are generated near the front of the cell, and it is electrons that must be collected at the back contact. Investigations of the n-i-p aSi:H cell, under forward bias, have shown the hole to be the limiting carrier. It was found that the introduction of positional symmetry into the optical generation rate profile greatly reduced the limiting carrier effect, and it was concluded that the limiting carrier effect arises due to the asymmetries in the material parameters of the cell, particularly the _ positional asymmetry of the optical generation profile. It was observed that the nature of the optical generation profile actually plays an important role in determining the identity of the limiting carrier, in a p-i-n cell. The same effect was not observed in the n-i-p cell. The effective carrier collection length has been defined, and it was seen that the limiting carrier possesses the larger effective collection length. The effect of boron and phosphorous profiling of the i-layer was studied. It was found that boron profiling led to a decrease in cell performance, while phosphorous profiling improved cell performance. It was found that there was a P concentration at which cell performance peaked. The dependence of the spectral response of the realistic cell on device length L, was investigated, showing a general improvement in the spectral response as L was decreased. The spectral response has been interpreted in a novel way. It was assumed that the form of the monochromatic optical generation profiles in the vicinity of the peak in the spectral response represented optimal generation profiles. These profiles were subjected to a linear transformation, such that their form was preserved but that their integrated value was the same as that of the realistic optical generation profile, under global AM1.5 illumination. Using these transformed optical generation profiles, J- V curves were obtained. The maximum power output PM of these J- V curves was seen to exhibit a maximum some 17% greater than that of the realistic cell with a realistic optical generation profile. The spectral response of the phosphorous profiled cell was obtained. In a manner similar to that for the non-P profiled cell, the optimal generation profile was found. The PM for this profile was found to be 7.86mWcm -2 , considerably larger than the 5.60mWcm-2 for the phosphorous profiled cell with a realistic optical generation profile. The effect on the simulation output of variations in numerous dangling-bond defect input parameters has been investigated. It was found that the energy position and concentration of the doped layer defects need not be known to a high degree of precision. On the other hand, it was found that the energy position of the i-layer defects, the standard deviation of the defect distributions, and the defect carrier capture cross-sections, do need to be known with certainty.

M.Sc. / As environmental and energy resource concerns have increased, greater emphasis has been placed on development of renewable energy resources such as photovoltaic electric generators. CuInSe 2/ZnO heterojunction solar cells are currently one of the most promising technologies to produce economically viable, clean electrical energy. The reaction of metallic alloys containing copper and indium to a selenium-containing atmosphere is by far the most promising industrial process. In this study ; copper-indium metallic precursors were prepared by electron-beam evaporation. The selenization process was conducted in vacuum in elemental Se vapour and in the presence of a H 2Se/Ar gas mixture at atmospheric pressure. Attention was given to the optimization of the structural features of the metallic alloys as well as the selenization parameters. Structural analysis revealed that the number of multilayers in , the precursor stack significantly influence the morphological features of the absorber films after selenization. The reaction temperature and reaction periods during the selenization process critically influenced the reaction kinetics of metallic phases. In the case of selenization in elemental Se vapour, temperatures as high as 550°C were required to convert the metallic alloys into fully reacted semiconductor thin films. Selenization in the presence of H2Se gas was more reactive and temperatures around 450°C resulted in the complete formation of CuInSe2. In the majority of cases, traces of CuSe were detected in the bulk of the material by XRD studies. The presence of the Cu-rich binary phases rendered solar cell devices useless. After removal of these detrimental segregated phases by KCN etching, glass/Mo/CuInSe2/CdS/ZnO solar cell devices reached conversion efficiencies around 8%.

Ph.D. / The reaction of direct current (DC) magnetron sputtered metallic CuIn and CuInGa alloys to a reactive H2Se/Ar/H2S gaseous atmosphere is an attractive industrial production process to produce Cu-based chalcopyrite absorber films for applications in high efficiency photovoltaic modules. This deposition process is generally referred to as a two-step deposition technology. However, the obvious technological advantages of this deposition technology are overshadowed by growth-related anomalies, such as the separation or at least partial separation of the ternary phases (i.e. CuInSe2, CuGaSe2 and CuInS2) during the high temperature chalcogenization. This in turn prevents the effective band-gap widening of the semiconductor alloys in order to achieve open-circuit voltages in excess of 600mV, which is a critical prerequisite for the optimal performance of thin film solar modules. Against this background, a detailed study was undertaken in order to understand the formation kinetics of quaternary CuIn(Se,S)2 and pentenary Cu(In,Ga)(Se,S)2 alloys deposited with a reproducible two-step growth technology. The main objective of this study was to optimize a complex set of experimental parameters in order to deposit homogenous alloys in which the band-gap value of the resulting semiconductor film could be modified in order to maximize the operating parameters of photovoltaic devices. This was achieved by the homogenous incorporation of S and/or Ga into the chalcopyrite lattice, resulting in shrinkage of the lattice parameters and hence increase in band-gap value Eg. However, the substitution of In with Ga and Se with S proved to be a complex process. It was, for example, observed that separation or at least partial separation of the ternary phases already occurs during the chemical reaction between the hydrogen selenides (H2Se) gas and the metallic precursors. Detailed studies indicated that this phenomenon was strongly related to the selenization parameters (e.g. reactive gas concentration, and reaction temperature and time) as well as the Cu/(In + Ga) atomic ratio. In optimized processes, the metallic precursor films were partially selenized in order to produce at least one partially reacted Cu-III-VI2 ternary alloy and group Cu-VI and III-VI binary phases. The partially selenized alloys were subsequently sulphurized under optimal thermal conditions in a H2S:Ar gas mixture to produce homogeneous single-phase quaternary and pentenary chalcopyrite alloys. X-ray diffraction (XRD) studies revealed that the lattice parameters of the chalcopyrite lattice decreased linearly with the incorporation of S and/or Ga, according to the predictions of Vegard’s law. Gracing incidence x-ray diffraction (GIXRD) studies on the compound semiconductors revealed that the lattice parameters remained virtually constant through the entire depth of the layer. Optical studies revealed a shift in the band-gap value of the absorber films as function of the S concentration. The band-gap of the absorber films could be varied between 0.99 and 1.35eV by controlling the S/Se anion ratio during the diffusion process, while maintaining the Ga/III atomic ratio constant at 0.25. Solar cells were completed by chemical bath deposition (CBD) of CdS and radio frequency (RF) sputtered intrinsic and highly conductive ZnO films onto the absorber films. The cells were evaluated under standard A.M. 1.5 conditions. Devices manufactured from CuIn(Se,S)2 and Cu(In,Ga)(Se,S)2 based alloys demonstrated average open-circuit voltages (Voc) and short-circuit current densities (Jsc) values of 470 and 650 mV and 20 and 33 mA.cm-2, respectively. A plot of the open-circuit voltage as function of the band-gap revealed an experimental relationship of: Voc = (Eg/q – 0.6) mV for Eg < 1.3 eV. The fill factor (FF) values varied between 35 and 56% and device efficiencies () between 4 and 13%, depending on the S/Se anion ratio and Ga incorporation. The findings from the studies clearly indicated that a better understanding of the CuIn(Se,S)2 and Cu(In,Ga)(Se,S)2 formation process led to absorber material with improved material properties. It was also demonstrated that it is possible to produce a homogenous CuIn(Se,S)2 and Cu(In,Ga)(Se,S)2 absorber films with the scalable two-step deposition process.

Ph.D. / Thin film photovoltaic modules based on Cu(In,Ga)Se2 (CIGS) thin films possess attributes that enable them to compete effectively with silicon-based modules. These attributes are stability, high efficiency, and low material cost. A very promising industrial related process to produce the chalcopyrite absorber layers involves the selenization of metallic precursors. However, recent literature suggests that it is extremely difficult to incorporate an appreciable amount of gallium into the active region of the CIGS thin film. Regardless of its location in the precursor stack, gallium has been observed to segregate to the back of the film during the high temperature selenization step. Consequently, the resulting films are phase-segregated with CuGaSe2 near the Mo electrode and CuInSe2 at the film surface. In this study, the incorporation of gallium and sulfur into CuInSe2 thin films was systematically investigated to establish a scientific and engineering base for the fabrication of homogeneous CuIn(Se,S)2 and Cu(In,Ga)Se2 quaternary alloys with optimum band gap values between 1.1 and 1.2 eV. The selenization of seleniumcontaining (i.e. Cu/InSe, InSe/Cu and InSe/Cu/InSe) precursors in elemental Se vapour at temperatures around 550°C resulted in CuInSe2 thin films with superior structural properties. In an attempt to increase the band gap of these films, the selenium species were replaced by sulfur species during a solid-state diffusion process. Alternatively, gallium was introduced into the structure by replacing the InSe/Cu/InSe precursors with InSe/Cu/GaSe precursors. Important process parameters such as the deposition temperature of precursor elements, the selenization temperature in elemental Se vapour, as well as the concentration of gallium in the alloys were optimized during subsequent studies. From these systematic studies optimum experimental conditions were determined for the deposition of homogeneous Cu(In,Ga)Se2 thin films. The monophasic nature of the quaternary alloys was confirmed by XRD studies, revealing a shift in the lattice spacing due to the homogeneous incorporation of gallium into the chalcopyrite lattice. Completed solar cell devices revealed open-circuit voltages above 500mV, which confirmed the increase in the band gap value of the absorber films. / Professor V. Alberts

M.Sc. / Thin film solar cell devices based on chalcopyrite absorber layers have reached a high performance level over the last few years, especially on laboratory scale. Despite this progress, there is still an urgent need to develop an industrial easily scalable deposition technology for depositing chalcopyrite thin films on a large scale. In this study, homogeneous single-phase quaternary Cu(In1-xGax)Se2 thin films were prepared with a reproducible two-step growth technique. The growth process is based on the controlled selenization of sputtered metallic CuIn0.75Ga0.25 alloys in a H2Se/Ar gas mixture at atmospheric pressure. Attention was mainly focused on the optimization of the reaction parameters such as the temperature profiles, gas concentrations and reaction periods. In an optimal reaction process, the reaction velocities of the binary selenide phases were carefully controlled to prevent the formation of stable group I-III-VI2 ternary alloys during the initial selenization step. The composite alloys were subsequently annealed in an inert atmosphere, followed by a second selenization step to promote the homogeneous alloying of gallium with partially formed CuInSe2. Glancing incident angle x-ray diffraction (GIXRD) at incident angles between 0.2º and 10º revealed virtually no shift in d-spacing with sample depth, which confirmed the monophasic nature of the quaternary alloys. Optical measurements revealed an increase in the band gap value of the chalcopyrite alloy due to the homogeneous incorporation of gallium into the CuInSe2 structure. Solar cell devices were fabricated by depositing cadmium sulphide (CdS) buffer layers and zinc oxide (ZnO) window layers onto the CuIn0.75Ga0.25Se2 absorber films. These devices were measure under standard A.M. 1.5 conditions and favorable conversion efficiencies were demonstrated. / Prof. V. Alberts

M.Sc. / As environmental and energy resource concerns have increased, greater stress has been placed on development of renewable energy resources such as photovoltaic electric generators. CuInSe2/ZnO heterojunction solar cells are currently one of the most promising technologies. CuInSe2 and its related alloys such as Cu(In,Ga)Se2 have been deposited by a number of techniques, including methods which have been demonstrated to be scalable to mass production volumes. In this study, attention was focused on (i) developing a relatively simple deposition technology for the production of chalcopyrite absorber films, (ii) detailed characterization of the semiconductor thin films in terms of the experimental parameters and (iii) production of completed CuInSe2/CdS/ZnO solar cell devices. A new two-stage growth process was developed which involved a low temperature precursor formation step and a subsequent high temperature selenization step. Selenium containing Cu-In-Ga and Cu-In-Ga-Se precursors were deposited by a thermal process in which the constituent elements were evaporated from a single graphite crucible onto heated substrates in presence of a selenium overpressure. These precursors were subsequently reacted in vacuum to elemental selenium vapour or to H2Se/Ar at atmospheric pressure in a separate diffusion reactor. In order to investigate the growth kinetics of the respective processes, the precursors were reacted to the Se in the temperature range between 300„aC and 600„aC. The structural features (morphology, presence of crystalline phases, bulk and in-depth compositional uniformity) of the respective films were compared and correlated against the growth parameters. From this systematic study, optimum growth parameters were determined for the production of completed solar cell devices. / Prof. V. Alberts