Search

Technology innovation and successful market introduction and impact carries the risk that future market conditions may not be as predicted. For PV technologies, the future industry cost of manufacture has been demonstrably hard to foresee, which can result in a poor assessment of the future competitiveness of a technology in development today. Future projections of PV cell and module manufacturing cost will always have significant uncertainty and this is not quantified in standard bottom-up cost-of-ownership calculations. To address this, we have built a Monte-Carlo based cost-of-ownership model to simulate reasonable future scenarios of manufacturing cost for c-Si cells and modules [1]. In this contribution, we analyze the cost/benefit of high-efficiency cell technology as a function of Ag usage, Ag cost, and manufacturing location. We find that at current and most likely future Ag costs, a high efficiency cell technology like silicon heterojunction (SHJ) with higher Ag usage will be more economically competitive in both Europe and China in the future. However, in Europe the SHJ benefit is less sensitive to changes in Ag cost and higher efficiency cells should be more competitive in higher labour cost locations.

A techno-economic analysis method is used to analyse industry standard mono crystalline silicon photovoltaic technologies – Aluminium Back Surface Field (Al-BSF) and Passivated Emitter and Rear Cell (PERC), together with promising process variations – Laser Doped Selective Emitter (LDSE) and three implementations of advanced hydrogenation. Building on a previously reported manufacturing cost and uncertainty analysis method, the impact of uncertainty in module performance and market price is added to estimate the manufacturer's gross margin. Two additional interpretation methods are described – (i) simultaneous Monte Carlo and (ii) contribution to variation – that help distinguish the impact of small differences between sequences, and identify the most important factors (cost, performance or market) affecting commercial viability. Combining these methods allows a rapid commercial viability assessment without requiring exact data on all the inputs. The analysis indicates that PERC is more commercially attractive than Al-BSF, with a median improvement in manufacturer's margin >5%. Al-BSF + LDSE is found to reduce manufacturer's margin. The PERC + LDSE and PERC + advanced hydrogenation sequences are estimated to provide a median margin improvement >2% compared to PERC alone. The advantage of PERC + LDSE depends strongly on delivering the expected 0.9%abs cell efficiency gain with high production yields and receiving a selling price premium for higher power modules; less important is the production cost of the LDSE process steps. The advantage of advanced hydrogenation depends strongly on achieving the expected 0.2%abs as-produced efficiency gain with high production yield, as well as realising a 0.5% selling price premium for being “CID-Free”; these factors outweigh the production costs of the alternative hydrogenation processes.

Water heating is one of the most energy intensive applications in households and domestic electric water heating systems (DEWH) offer large thermal storage for moving electrical load across the day. This study uses a unique dataset from 410 households and presents a comprehensive analysis of electricity consumption and hot water draw of DEWH for the Australian context. Using the real-world data and thermal energy modelling tool TRNSYS, the study analyses the potential of storing and using excess PV generation in DEWH and investigates the impact of different daily hot water draw profiles, PV and DEWH size on the potential for excess PV utilization. The results show that households on average use 6 kWh of energy for DEWH and 142 L of hot water daily. Potential excess PV utilization is highly dependent on the household's daily hot water draw profile and is also affected by seasonality. On average, excess PV generation from a 4.5 kW PV system can provide 48% of daily DEWH energy for a household with a typical working family profile, which corresponds to a 28% increase in PV self-consumption.

Appliance level control and automation is an increasingly promising demand-side management tool with growing installation of advanced metering, monitoring and control infrastructure in both residential and commercial contexts. Successful implementation of appliance control and automation can alleviate network peak demand and improve distributed photovoltaic (D-PV) self-consumption to reduce its network voltage and reverse power flow impacts. Domestic electric water heating (DEWH) systems are widely deployed globally and have one of the highest peak power draw and overall energy consumption of household appliances. DEWH storage tanks offer large thermal energy storage capacity which can be used for shifting demand to lower demand periods. With growing D-PV deployment, they also offer the opportunity to store excess generation that would be otherwise exported to the grid. In this work, an intelligent water heating control tool (IWHC) is developed to store excess D-PV generation in DEWH storage tanks as thermal energy, according to the D-PV generation characteristics, household electricity consumption, hot water draw (HWD) patterns, and real time energy monitoring. The IWHC tool was installed and tested in nine Australian households with D-PV and DEWH systems. For performance comparison, two other commercially available control tools, timer, and diverter, were installed and tested in eleven other households with D-PV and DEWH systems. For each control tool, energy simulation models were developed, and the collected field performance data was used to validate the models. The validated simulation models were extended to a broader set of 380 Australian households with a year of D-PV, household and DEWH electricity consumption data. The results indicate that, on average, households can utilize 2.4 kWh, 1.8 kWh and 3.4 kWh of daily excess D-PV generation for water heating, using the IWHC, timer and diverter, respectively. Financial savings from the control of DEWH are highly dependent on households’ tariffs and daily HWD profiles. Under the most optimal morning dominant HWD profile scenario and with an average tariff, households can, on average, save $100, $80, and $170 per year with the IWHC, timer and diverter, respectively. However, the diverter’s superior field performance comes with higher capital cost, making IWHC the most attractive option.

Manufacturing cost analysis is becoming an increasingly important tool in the photovoltaics industry to identify research areas that need attention and enable progress towards cost reduction targets. We describe a method to estimate manufacturing cost that is suitable for use during an early stage of technology development, delivering both the manufacturing cost estimate as well as an uncertainty analysis that quickly highlights the opportunities for greatest cost improvement. We apply the technique to three process sequences for the large-scale production of organic-inorganic hybrid perovskite photovoltaic modules. A process sequence that combines two demonstrated perovskite module sequences is estimated to cost $107/m2 (uncertainty range $87 to 140/m2), comparable with commercial crystalline silicon and cadmium telluride technologies (on a US $/m2 basis). A levelized cost of electricity calculation shows that this perovskite technology would be competitive in 2015 with incumbent photovoltaic technologies if a module power conversion efficiency of 18% and lifetime of 20 years can be achieved. Further analysis shows that even if the cost of the active layers and rear electrode were reduced to zero, a module power conversion efficiency of 18% and lifetime of 20 years would be required to meet the 2020 SunShot levelized cost of electricity targets. Copyright © 2017 John Wiley & Sons, Ltd.

Summary Solar-driven electrolysis of water to generate hydrogen is emerging as a viable strategy to decarbonize the global energy economy. However, this direction is more expensive than traditional fossil fuel generation of hydrogen, and effective pathways to lower this cost need to be identified. Here, we report a Monte Carlo approach to explore a wide range of input assumptions to identify key cost drivers, targets, and localized conditions necessary for competitive stand-alone dedicated PV powered hydrogen electrolysis. We determine the levelized cost of hydrogen (LCOH), considering historical weather data for specific locations to model our PV system, and optimize its size compared to the electrolyzer. This analysis and its methods show the potential for green hydrogen production using off-grid PV, shows the merits of remote systems in areas of high solar resource, and provides cost and performance targets for electrolyzer technologies.

The hydrogen plasma passivation of thin film polycrystalline silicon (pc-Si) was investigated in conjunction with plasma texturing process to make efficient heterojunction solar cells. The pc-Si layers were first treated using direct and remote hydrogen plasma technologies. The heterojunction solar cells were then fabricated by subsequent deposition of i/n+ a-Si:H. Hydrogenation at high temperature (610 °C) results in enhanced dissolution and diffusion of hydrogen in pc-Si by a factor of about 3 and 4, respectively, in comparison with those at low temperature (420 °C). The hydrogen atoms in the pc-Si layer mainly bond to the silicon dangling bonds and form complexes with dopant atoms. In addition, platelets defects are generated by the hydrogen plasma in the sub-surface region of pc-Si hydrogenated at 420 °C and cause higher saturation current in the space charge region whilst they form in the region deeper than 1 μm at 610 °C. Removal of the platelets using SF6/N2O plasma post-texturing after low-temperature hydrogenation not only enhances the short circuit current but also improves the open circuit voltage and the fill factor simultaneously. Combining plasma pre-texturing with high-temperature hydrogenation, the best 2 µm-thick pc-Si heterojunction solar cell reaches an efficiency of 8.54%.

Diode laser crystallization of thin silicon films on the glass has been used to form polycrystalline silicon layers for solar cells. Properties of an intermediate layer stack of sputtered SiOx/SiNx/SiOx between the glass and the silicon have been improved by reactively sputtering the SiNx layer, which result in enhanced optical and electrical performance. Light trapping is further enhanced by texturing the rear surface of the silicon prior to metallization. An initial efficiency of 11.7% with VOC of 585 mV has been achieved using this technique, which are the highest values reported for poly-Si solar cells on glass substrates. Cells suffer a short term, recoverable degradation of VOC, and fill factor. The magnitude of the degradation is reduced via the repeated thermal treatment. A selective p+ metallization scheme has been developed which eliminates the degradation altogether.

The paper presents a review of major features of the crystalline silicon on glass (CSG) technology, its achievements, limitations and challenges, and latest developments. CSG cells are fabricated by solid-state crystallisation (SPC) of 1.5–3.5 µm thick precursor diodes prepared by PECVD or ebeam evaporation followed by thermal annealing, hydrogen passivation and metallisation. The highest efficiency of 10.4% was demonstrated on a PECVD minimodule on textured borosilicate glass. The best performing ebeam-evaporated cells on planar glass reached 8.6% efficiency. CSG cells were also produced on low-cost soda-lime glass with 8.1% and 7.1% efficiencies on PECVD and ebeam material respectively. The performance of SPC CSG cells is limited to below 11% because high defect density in SPC material limits V OC and 1.5–3.5 µm cell thickness limits J SC . A breakthrough came about when thicker poly-Si films with low defect density on glass were prepared by liquid-phase crystallisation (Amkreutz, 2011) leading to development of the next generation, liquid-phase crystallised silicon on glass (LPCSG) solar cells. The best performing LPCSG cells are made by line-focus laser crystallisation of 10 µm thick ebeam silicon films on dielectric layer coated borosilicate glass. High material quality is confirmed by low defect density observed in TEM images, high carrier mobilities, and minority carrier lifetime longer than 260 ns. An intermediate dielectric layer can be SiC x , SiO x , SiN x or their combination and its properties are crucial for cell fabrication and performance. Dopants are introduced into the LPCSG cell absorber either during film deposition or diffused from doped intermediate layer during crystallisation. Light-trapping texture is formed on the exposed silicon surface by wet etching. A cell emitter is created by diffusion from spin-on-dopant source. Cell metallisation is based on point contacts between Al and cell emitter and absorber accessed through vias etched through cell layers to different depths. LPCSG cells outperformed CSG cells, with record V OC of 585 mV and efficiency of 11.7%. Efficiencies above 13% are achievable by improving light-coupling and contacting.

The Australian Centre for Advanced Photovoltaics (ACAP) co-ordinates the activities of the six Australian research institutions and a group of industrial partners in the Australia-US Institute for Advanced Photovoltaics (AUSIAPV) to develop the next generations of photovoltaic device technology and to provide a pipeline of opportunities for performance increase and cost reduction. AUSIAPV links ACAP with US-based partners. These national and international research collaborations provide a pathway for highly visible, structured photovoltaic research collaboration between Australian and US researchers, institutes and agencies with significant joint programs based on the clear synergies between the participating organizations. The research program is organized in five collaborative Program Packages (PPs). PP1 deals with silicon wafer-based cells, focusing on three main areas: cells from solar grade silicon, rear contact and silicon-based tandem cells. PP2 involves research into a range of organic solar cells, organic/inorganic hybrid cells, "earth abundant" thin-film materials and "third generation" approaches. PP3 is concerned with optics and characterization. PP4 will deliver a substantiated methodology for assessing manufacturing costs of the different technologies and PP5 involves education, training and outreach. The main research topics, results and plans for the future are presented.

Solid-phase crystallisation (SPC) of Si films on glass prepared by three deposition methods, plasma enhanced chemical vapour deposition (PECVD), electron-beam evaporation, and a combination of both, are compared for different annealing temperatures. Three independent techniques, optical transmission microscopy, UV reflectance spectroscopy, and X-ray diffraction, are used to characterise the crystallisation kinetics and film properties. The activation energy for the incubation is estimated to be 2.7–3.0 eV. The scanning electron microscopy images of polycrystalline Si films after Secco etching show a gradually decreasing average grain size in each film type for higher SPC temperatures. The crystal quality of all film types degrades at higher crystallisation temperatures. Solar cells fabricated from these polycrystalline Si films were characterised by Suns- V oc and spectral response measurements. According to both the resulting open-circuit voltage and the short-circuit current, the electronic quality of all polycrystalline Si film types and the corresponding cell performance degrade for higher crystallisation temperatures but to a different extent depending on the film deposition method.

Solid-phase crystallisation of Si thin films on glass fabricated by plasma enhanced chemical vapour deposition is compared at different annealing temperatures. Four independent techniques, optical transmission microscopy, Raman and UV reflectance spectroscopy, and X-ray diffraction, are used to characterise the crystallisation kinetics and film properties. The 1.5 μm thick films with the n+/p−/p+ solar cell structure have incubation times of about 300, 53, and 14 min and full crystallisation times of about 855, 128, and 30 min at 600 °C, 640 °C, and 680 °C respectively. Estimated activation energies for incubation and crystal growth are 2.7 and 3.2 eV respectively. The average grain size in the resulting polycrystalline Si films measured from scanning electron microscopy images gradually decreases with a higher annealing temperature and the crystal quality becomes poorer according to the Raman, UV reflection, and X-ray diffraction results. The dopant activation and majority carrier mobilities in heavily doped n+ and p+ layers are similar for all crystallisation temperatures. Both the open-circuit voltage and the spectral response are lower for the cells crystallised at higher temperatures and the minority carrier diffusion lengths are shorter accordingly although they are still longer than the cell thickness for all annealing temperatures. The results indicate that shortening the crystallisation time by merely increasing the crystallisation temperature offers little or no merits for PECVD polycrystalline Si thin-film solar cells on glass.

The effects of electrostatic fluctuations due to charged extended defects and strain-induced bandgap fluctuations are examined in polycrystalline silicon on glass solar cells. The analysis is based on models previously applied to Cu(In,Ga)Se2 solar cells, but with a new interpretation of the local ideality factor associated with electrostatic fluctuations. It is shown that electrostatic fluctuations become influential to the cell voltage properties as the absorber dopant concentration falls below a certain threshold (a few 10 15 cm � 3 ), and the degradations to the open circuit voltage and fill factor are expected to increase with further lowering of dopant density. It is equally plausible that the electrostatic fluctuations originate from charged dislocations or grain boundaries. Bandgap fluctuations on the other hand can be detrimental to the open circuit voltage of cells of any absorber dopant density. However, this voltage degrading effect is seen only in the cells deposited by electronbeam evaporation, and not amongst those made by plasma enhanced chemical vapour deposition. Copyright # 2011 John Wiley & Sons, Ltd.

Crystalline silicon on glass (CSG) solar cell technology was developed to address the difficulty that silicon wafer-based technology has in reaching the very low costs required for large-scale photovoltaic applications as well as the perceived fundamental difficulties with other thin-film technologies. The aim was to combine the advantages of standard silicon wafer-based technology, namely ruggedness, durability, good electronic properties and environmental soundness with the advantages of thin-films, specifically low material use, large monolithic construction and a desirable glass superstrate configuration. The challenge has been to match the different preferred processing temperatures of silicon and glass and to obtain strong solar absorption in notoriously weakly-absorbing silicon of only 1.4 μm thickness, the thinnest active layer of the key thin-film contenders. A rugged, durable silicon thin-film technology has been developed arguably with the lowest likely manufacturing cost of these contenders and confirmed efficiency for small pilot line modules already in the 8–9% energy conversion efficiency range, on the path to 12–13%.

Strong contact adhesion is an important requirement for durable, manufacturable solar cells. Advanced contacting technologies require new methods to measure adhesion. We describe a scratch test for measuring contact adhesion that involves scanning a weighted stylus across the cell while measuring the horizontal force F D required to dislodge the contacts. F D is characteristic of the adhesive bond but independent of the contact height, stylus weight and scan speed. We observe that contact peeling depends also on the tensile strength of the metal finger. The tests provide a valuable way to assess and optimize the adhesion of metal contacts.

Polycrystalline silicon thin-film solar cells with large grains up to 1 mm wide and 10 mm long created by diode laser crystallization require an enhanced high performance plasma hydrogenation process for better cell performance. Higher process temperature, longer process time and higher plasma power (800°C, 16 minutes and 3.9 kW, respectively) than those used for smaller grain polycrystalline silicon films were applied and their effects on the cell performance were evaluated. It is found that open circuit voltage increases from 417 mV to 451 mV for 0.1-suns and from 518 mV to 529 mV for 1-sun due to lower the diode saturation currents from 3.1e-5 A/cm2 to 1.2e-5 A/cm2 after such a hydrogenation process. This result shows that enhanced hydrogenation has a potential for application to various polycrystalline silicon films with wide range of grain size and defect density.

Line-focus diode laser is applied to advance crystalline silicon thin-film solar cell technology. Three new processes have been developed: 1) defect annealing/dopant activation; 2) dopant diffusion; 3) liquid phase crystallisation of thin films. The former two processes are applied to either create a solar cell device from pre-crystallised films or improve its performance while reducing the maximum temperature experienced by substrate. The later process is applied to amorphous silicon films to obtain high crystal and electronic quality material for thin-film solar cells with higher efficiency potential. Defect annealing/dopant activation and dopant diffusion in a few micron thick poly-Si films are achieved by scanning with line-focus 808 nm diode laser beam at 15-24 kW/cm2 laser power and 2~6 ms exposure. Temperature profile in the film during the treatment is independent from laser power and exposure but determined by beam shape. Solar cell open-circuit voltages of about 500 mV after such laser treatments is similar or even higher than voltages after standard rapid-thermal treatments while the highest temperature experienced by glass is 300C lower. Amorphous silicon films can be melted and subsequently liquid-phase crystallised by a single scan of line laser beam at about 20 kW/cm2 power and 10-15 ms exposure. Solar cells made of laser-crystallised material achieve 557 mV opencircuit voltage and 8.4% efficiency. Electronic quality of such cells is consistent with efficiencies exceeding 13% and it is currently limited by research-level simplified cell metallisation.

We report on the effects of hydrogenation of InSb GaAs epilayers. The as-grown MOCVD InSb is p-type with an acceptor activation energy of 16 meV. The 16 meV impurity, however, is fully passivated during protonation in a hydrogen microwave plasma, revealing a shallower, 2 meV acceptor which appears unaffected by further processing. The material remains p-type after hydrogenation and the Fermi level calculations reveal that the majority carriers over the entire temperature range are holes. The electron to hole mobility ratio is sufficiently large however, that the Hall data shows evidence for an apparent type conversion.

Photoreflectance measurements of multilayered Al x Ga 1−x As GaAs semiconductor structure in the regime of high modulating laser powers are reported. The main photoreflectance feature observed at low powers at about 1.4 eV is related to Franz-Keldysh oscillations caused by electric field at the Al x Ga 1−x As GaAs interface. The value of electric field deduced from these oscillations is in agreement with that determined from calculation using the accepted value of the Al x Ga 1−x As GaAs band offset and the doping level in the GaAs layer. On increasing the power of the modulating laser, the spectrum changes its character and after decomposition reveals a feature at the energy corresponding to GaAs bandgap. The intensity of this line increases with the modulating laser power, moreover the line is in phase with the Franz-Keldysh oscillations. These observations are consistent with the assignment of the line to the photoreflectance at the underlying GaAs buffer/GaAs substrate interface. This finding underlines the altered character of the photoreflectance spectra at high light intensities, compared to the standard low intensity regime.

We have studied both the molecular-beam epitaxy (MBE) and metal-organic chemical vapour deposition (MOCVD) grown AlGaAsAlAsGaAs double barrier multiple quantum well structures for 3–5 μm infrared (IR) photodetector application. The intersubband absorptions of these samples were measured at the Brewster angle as well as through a multipass 45° wedge waveguide. In the former case, we have also studied the polarisation dependence of the IR transmission. The MBE grown sample has an absorption at about 3.1 μm, while the absorption peak for the MOCVD sample is shifted slightly, depending on the wafer location. The low temperature photoluminescence peaks show a significant shift when the probe position is moved across the MOCVD wafer along the gas flow direction, owing to thickness non-uniformity. This variation was also confirmed by double crystal X-ray diffraction data. Theoretically, we have carried out a bound and quasibound state energy calculation for these structures as a function of well width and related them to the experimental results. In addition, the IR photoresponse for the MBE grown sample at about 80 K has been measured for IR photodetector application.

We report the use of monoethyl arsine, in conjunction with trimethyl indium, for the low temperature growth of InAs, and demonstrate its viability as a source of arsenic for growth by metalorganic chemical vapour deposition. Good quality InAs was relatively insensitive to both growth temperature, in the range 420–490°C, and V III ratio, with ratios between 1.5 and 5 resulting in epilayers of good morphology. At a temperature of 470°C and trimethyl indium molar flow of 3.25 μmol min−1, the growth rate of InAs was 5 μm h−1. The growth rate was found to be thermally activated, with an activation energy of 37.5 kcal mol−1 with no evidence for parasitic reactions. The V III ratio is lower than for the growth of InAs from arsine, while the temperature dependence of morphology is much less critical than that reported for growth of InAs from trimethyl indium and tertiary-butyl arsine.

p-type InSb was prepared from trimethyl indium and trimethyl antimony at temperatures between 420 and 490 degrees C. A narrow window in the growth conditions of temperature and V/III ratio was found for InSb, with non-uniform, inhomogeneous epilayer growth for conditions other than optimum. Optimum morphologies were achieved for V/III ratios between 2.2 and 2.5 and at a temperature of 450 degrees C and the use of low pressures in the growth of InSb was observed to result in an enhanced uniformity. The InSb prepared displayed type conversion over the range 70-300 K; analysis of the transport properties is then complicated by the contributions from two carriers. Calculations using Fermi-Dirac statistics were carried out, identifying the majority carrier to be holes over the entire temperature range considered. Calculations also identify an acceptor level with activation energy of 16+or-0.5 meV, an acceptor concentration of (2.65+or-0.02)*1017 cm-3, in good agreement with conductivity measurements, and a low-temperature hole mobility of 450 cm2 V-1 s-1. The detailed analysis employing Fermi-Dirac statistics is then extended to determine the temperature dependence of the Fermi level in InSb and InAs for various impurity compositions.

Heteroepitaxial InAsSb is routinely prepared on InAs, InSb, or GaAs substrates under conditions favorable to dislocation formation, since the binary components are lattice mismatched by about 7.4%, and the ternary are mismatched to GaAs by between 7.2% and 14.5%, depending on composition. We here extend the description of electron scattering in InAsSb to include the effects of grain boundaries and dislocations. Comparison with experiment confirms that dislocation scattering has a strong effect on transport, while alloy scattering limits mobility in ternary samples grown with a minimum of defects.

High temperature, 640 °C or 680 °C, and 8 to 60 min long treatment was applied to amorphous Si films on glass prior to solid-phase crystallisation (SPC) at 600 °C to selectively enhance and shorten the “incubation-only” period in the SPC process. Effects of such enhanced-incubation anneal and its conditions on SPC kinetics and resulting properties of polycrystalline silicon thin-film solar cells are studied. The incubation period at 600 °C can be shortened by 50–80% (by 200–400 min) and the crystal growth rate can be doubled after only a few minute long enhanced-incubation anneal at 640 or 680 °C. However, cell performance parameters degrade for all such pre-treatments in a direct correlation with treatment intensity: the higher enhanced-incubation temperature and the longer its time, the worse the cell performance. Both open circuit voltage and short circuit current gradually decrease with longer and hotter enhanced-incubation annealing. It is the cell blue response that is mostly affected by the treatment, while the red response stays similar. Accordingly, the modeled minority carrier diffusion length is shorter in the cell emitter and it does not change in the absorber after the enhanced-incubation anneal. The increasing diode ideality factor means more depletion region recombination, which also indicates that the applied treatment predominantly affects the front, glass-side layers of the cell, where the junction and the emitter are located. The observed results are unexpected as the cell performance worsens for treatments conducted within the incubation period only, i.e. before nucleation and crystal growth start to take place and the Si film is still fully amorphous. The results draw attention to the importance of the incubation period to the final electronic quality of the polycrystalline Si.

Nuclear radiation detectors constructed from GaAs Schottky barrier structures require control of barrier height if high resolution is to be reproducibly obtained. The authors have therefore studied the variation of device barrier height after treatment with common etching solutions. The heights of Au and Al Schottky barriers on respective samples of n- and p-type liquid-phase epitaxial GaAs were found from reverse bias I-V measurements subsequent to etching with aqua regia and 3H2SO4:H2O2:H2O (3:1:1). An increase in barrier height observed after the 3:1:1 etch was found to be a consequence of sulphur contamination. This result is consistent with those reported for deliberate sulphur passivation of GaAs. Low diode ideality factors, from 1.007 to 1.064, were measured for layers with carrier concentrations in the range 4*1014 cm-3 to 2.5*1016 cm-3. A comparison was made between samples etched in the 3:1:1 solution at 5 degrees C and at 100 degrees C. C-V measurements gave anomalously large barrier heights for those samples etched at 5 degrees C, indicating the formation of a significant interfacial layer prior to Schottky barrier metallization, while X-ray photoelectron spectroscopy measurements confirmed the presence of a large oxide layer. No such anomalies were observed for the samples etched in 3:1:1 at 100 degrees C.

We review the experimental evidence for the origin and location of the four native point defects in the wide gap semiconducting indium and gallium nitrides and compare then with experimental predictions. The donor triplets associated with nitrogen vacancies and the deep compensating centres ascribed to the antisite substitutional defects appear to have the greatest effect on macroscopic properties, apparently including the four luminescent bands in GaN. Calculated mobilities in InN and GaN depend principally on ionised impurity and polar-mode phonon scattering. We reconcile these results with experimental data and point out the consequences for improvements in material growth.

Electron-beam (e-beam) evaporation provides both exciting opportunities and challenges for the preparation of poly-crystalline silicon (poly-Si) thin film solar cells. A conversion efficiency of 6.7% was recently achieved for solid phase crystallized poly-Si mini-modules on planar SiN-coated glass deposited at a deposition rate of 600 nm/min, demonstrating the excellent electronic quality of e-beam evaporated silicon. Even at significantly increased background pressures of 5×10−6 mbar, the photovoltaic performance of the mini-modules was considerably high, showing a decline in open circuit voltage of 17 mV per cell. The implementation of light trapping structures into the device led to an efficiency increase of 1.1%, yielding module efficiencies of 7.8%. By systematically studying the implementation of ZnO:Al as a front contact layer into the poly-Si solar cell device structure, we unraveled novel features that prove the supreme suitability of ZnO:Al for poly-Si thin film solar cells. Not only can etched ZnO:Al be utilized as a front side texture, but its electrical properties can also improve during the crystallization process of the Si layer, showing a record charge carrier mobility of 67 cm2/Vs after thermal annealing. In addition, ZnO:Al drastically modifies the crystallization kinetics of the Si on ZnO:Al, enabling us to control the crystallization process by adjusting the deposition temperature. The nucleation process of Si on ZnO:Al was found to be influenced by a variation of the deposition temperature of the amorphous Si in a critical temperature regime of 200 °C to 300 °C. The nucleation rate decreased significantly with decreasing deposition temperature, while the activation energy for nucleation increased from 2.9 eV at a deposition temperature of 300 °C to 5.1 eV at 200 °C, resulting in poly-Si which comprised grains with features sizes of several µm.

Solid-phase crystallization (SPC) of Si thin films on glass deposited by electron-beam evaporation is compared at different annealing temperatures. Four independent film characterization methods, optical transmission microscopy, Raman and UV reflectance spectroscopy (UV-R), and X-ray diffraction (XRD), reveal that 1.4 micron thick films with n+/p-/p+ solar cell structures have the incubation times of about 285, 32, and 10 min and the full crystallization times of about 825, 94, and 24 min at 600°C, 640°C, and 680°C respectively. The estimated activation energies for the incubation and crystal growth are 3.0 and 3.2 eV respectively. The scanning electron microscopy (SEM) images of the resulting polycrystalline Si (poly-Si) films after Secco etching show that the average grain size gradually decreases with a higher SPC temperature. Similarly, according to the Raman, UV-R, and XRD spectra the crystal quality of the poly-Si films is poorer for the higher SPC temperatures. The sheet resistivities, which are dominated by the properties of the two thin but heavily doped n+ and p+ layers, are similar for all SPC temperatures. The open-circuit voltages (Voc) of about 450 mV measured by Suns-Voc method are also similar for all SPC temperatures and comparable to previously reported values for the best evaporated poly-Si cells. However, the minority carrier diffusion length of about 0.5 micron in the absorber layers for all SPC temperatures extracted from the external quantum efficiency (EQE) data is much shorter than the cell thickness and than previously reported, which can be due to relatively high boron concentration in the absorber found by means of the capacitance-voltage (C-V) analysis.

The electron mobility in InAs1−xSbx is calculated for an ionized impurity density between 5× 1014 cm−3 and 1 × 1017 cm−3 at 77 K. The various methods of calculating electron mobility are discussed and their respective merits and limitations summarized. In general, alloy scattering is the mobility limiting process at low carrier density (n ≤ 1015 cm−3) while ionized impurity scattering is important at higher carrier density. The effect of compensation (NA/ND) on low field drift mobility is also calculated for a range of carrier concentrations at x=0.1, 0.6, 0.9, and 1.0 compositions of particular technological significance. Compensation is found to degrade electron mobility quite significantly in all cases. The calculations at x=1.0 (InSb) are also found to be in good agreement with available experimental data.

By using a contactless double crystal x‐ray diffraction technique with either photoluminescence or infrared intersubband absorption and theoretical calculations, it is possible to determine the dimensions and composition of a three layered multiple quantum well (MQW) structure accurately. A strained AlGaAs/AlAs/InGaAs double barrier (DB) three layered MQW structure was used to demonstrate this. Moreover, it is shown that this structure is well suited for infrared photodetection in the 3 μm wavelength region, based on intersubband absorption. The compositions and thicknesses evaluated are in good agreement, and transmission electron microscopy is utilized to confirm the thicknesses.

A few microns thick silicon films on glass coated with a dielectric intermediate layer can be crystallised by a single pass of a line-focused diode laser beam. Under favorable process conditions relatively large linear grains with low defect density are formed. Most grain boundaries are defect-free low-energy twin-boundaries. Boron-doped laser crystallised films are processed into solar cells by diffusing an emitter from a phosphorous spin-on-dopant source, measuring up to 539 mV open-circuit voltage prior to metallisation. After applying a point-contact metallisation the best cell achieves 7.8% energy conversion efficiency, open-circuit voltage of 526 mV and short-circuit current of 26 mA/cm2 . The efficiency is significantly limited by a low fill-factor of 56% due to the simplified metallisation approach. The internal quantum efficiency of laser crystallised cells is consistent with low front surface recombination. By improving cell metallisation and enhancing light-trapping the efficiencies of above 13% can be achieved.

Experimental data on the nature and energetic location of levels associated with native point defects in the group-III metal nitrides are critically reviewed and compared with theoretical estimates. All three show strong evidence of the existence of a triplet of donorlike states associated with the nitrogen vacancy. Ground states are at about 150, 400, and 900 meV from the conduction-band edge in InN, GaN, and AlN, respectively, with their charged derivatives lying closer to the band edge. These values agree with both modified-hydrogenic and deep-level calculations, surprisingly well in view of the inherent approximations in each in this depth range. The InN donor ground state is both optically active and usually occupied, showing a distinctive absorption band which is very well described by quantum-defect analysis. Variation of threshold with electron concentration shows a Moss-Burstein shift commensurate with that observed in band-to-band absorption. In both GaN and AlN, levels have been identified at about 1/4${\mathit{E}}_{\mathit{G}}$ and about 3/4${\mathit{E}}_{\mathit{G}}$, which correlate well with predictions for the antisite defects ${\mathrm{N}}_{\mathit{M}}$ and ${\mathit{M}}_{\mathrm{N}}$, respectively, while similar behavior in InN is at odds with theory. The metal-vacancy defect appears to generate a level somewhat below midgap in AlN and close to the valence-band edge in GaN, but has not been located experimentally in InN, where it is predicted to lie very close to the valence-band edge. A tentative scheme for the participation of two of the native defects in GaN, namely ${\mathit{V}}_{\mathrm{N}}$ and ${\mathrm{N}}_{\mathrm{Ga}}$, in the four broad emission bands found in Zn-compensated and undoped GaN is offered.

The nitrides of Al, Ga and In are III-V compound semiconductors with properties more closely akin to those of the II-VI system and applications problems of similar type. All three have wide, direct band gaps and relatively light, therefore mobile, electrons. Less encouragingly, native point defects appear to play a significant role in both optical and electronic properties.Both experiment and theory point to a triplet of donor states associated with the nitrogen vacancy, with deep compensating centres deriving from antisite defects. The ionic radius of the metal then seems to determine the conductivity of as-grown material, indium is reluctant to occupy nitrogen sites while aluminium does so readily and gallium is equivocal. Thus the upper donor level of InN is not depleted and n-type behaviour is always observed, the equivalent level in AIN is always overcompensated and the remaining donor levels are too deep to contribute free electrons at normal temperatures so that the material is consistently insulating. GaN may be n-type or semi-insulating since compensation ratios either side of unity appear to be possible, depending on the method of growth.In this paper we review the evidence, both optical and electrical, for the existence, nature and energetic location of the four basic point defects in each nitride, noting in particular that all four broad luminescence bands in GaN:Zn can be accounted for by the presence of nitrogen on gallium sites and of nitrogen vacancies.

The performance of a surface normal modulator operating at visible wavelengths is described. Modulation at these wavelengths is achieved by incorporating aluminium in the well material and increasing the aluminium content in the barrier. Maximum contrast is achieved at 660 nm with a low insertion loss (1.5 dB) and a differential reflectivity of 13% for 13 V.

X‐ray photoelectron spectroscopy with argon ion depth profiling was employed to investigate sulfur deposition from a standard 3H2SO4:H2O2:H2O etchant onto (100) GaAs. From these experiments it appears that etching with 3H2SO4:H2O2:H2O leaves a mixed sulfide/oxide layer, with free sulfur at the top surface of this layer. The sulfide within the layer was found to be bound to gallium, and recognition of the presence of sulfur allowed the clarification of the Ga(3d) photoelectron peak position for Ga2O3. No evidence of sulfate formation was found. Removal of the sulfur contamination using a solution of HCl:H2O was investigated. Also, atmospheric contamination of samples exposed to room conditions for six months was observed.

A comparison is reported of the properties of AlN, GaN and InN films reactively r.f. sputtered from pre-nitrided targets. All three have a densely packed (00.2) polycrystallite orientation, irrespective of the substrate used. Hydrogenic donors associated with nitrogen vacancies are found at 50 meV, 110 meV and 220 meV in InN, GaN, and AlN respectively. Compensating acceptor levels at depths between 200 and 250 meV seem to derive from M N antisite defects. Their densities, which depend on the metal ion radii, are such that they partially compensate InN, sometimes fully compensate GaN and always ovecompensate AlN.

Summary form only given. We have demonstrated that gratings of different strengths in unhydrogenated fiber display different stabilities at elevated temperatures, however, gratings written in hydrogenated fiber did not show this trend. These observations should provide further insight into the mechanism of photo-induced index change during both writing and annealing.