Measurements of the optical, electronic and 1/f noise properties for a series of n-type doped hydrogenated amorphous silicon carbide thin films with varying gas phase concentrations of CH4 are described. The increase in the optical absorption edge of the n-type a-SiCx:H films with the addition of carbon is slower than in p-type films. Studies of the variation in the non-Gaussian statistics which characterize the 1/f noise indicate that the disorder at the mobility edge is greater for films with higher carbon concentrations.

We have incorporated N-atoms into hydrogenated amorphous silicon in the Si-rich alloy regime to ∼12 at.% N, and have observed a transition from n-type doping to alloying as the concentration of N-atoms is increased above about 5 at.%. By analogy with the local bonding arrangements of P-donors in n-doped a-Si:H, we attribute the doping to four-fold coordinated N-atoms with second neighbor H-atoms as in N+-Si-H linkages. The occurrence of these arrangements is supported by (i) IR studies which indicate a non-statistical association of N and H-atoms bonded to the same Si-atom, and (ii) a chemical bonding model in which the large effective electronegativies of four-fold coordinated N+ atoms and neutral O-atoms promote similar bonding properties with respect to their nearest-neighbor arrangements with Si and H atoms such as N+ (O) -Si-H linkages

The key requirements for photovoltaic modules to be accepted for large-scale terrestrial applications are (i) low material cost, (ii) high efficiency with good stability, (iii) low manufacturing cost with good yield and (iv) environmental safety. Thin films of amorphous silicon alloy are inexpensive; the products are also environmentally benign. The challenge has been to improve the stable efficiency of these modules and transfer the R&D results into production. Using a Multijunction, Multi-bandgap approach to capture the solar spectrum more efficiently, we have developed one-square-foot modules with initial efficiency of 11.8%. After 1000 h of one-sun light soaking, a stable efficiency of 10.2% was obtained. Both the efficiency values were confirmed by National Renewable Energy Laboratory. The technology has been transferred to production using an automated roll-to-roll process in which different layers of the cell structure are deposited in a continuous manner onto stainless steel rolls, 14” wide and half a mile long. The rolls are next processed into modules of different sizes. This inexpensive manufacturing process produces high efficiency modules with subcell yields greater than 99%. The key features of the technology transfer and future scope for improvement are discussed.

The chemical reactions at the surface of transparent conductive oxides (SnO2, ITO and ZnO) have been studied in silane and hydrogen plasmas by in-situ ellipsometry and by SIMS as well as XPS depth profiling. SIMS and XPS of the interface reveal an increasing amount of metallic phases upon lowering a-Si:H growth rates (controlled by plasma power), indicating that the ion and radical impact is more than compensated by protecting the surface by a rapidly growing a-Si:H film. Hence, optical transmission of TCO films as well as the efficiency of solar cells can be improved if the first few nanometers of the p-layer are grown at higher rates. Comparing a-Si:H deposition on top of different TCOs, reduction effects on ITO and SnO2 have been detected whereas ZnO appeared to be chemically stable. Therefore an additional shielding of the SnO2 surface by a thin ZnO layer has been investigated in greater detail. Small amounts of H are detected close to the ZnO surface by SIMS after hydrogen plasma treatment, but no significant changes occur to the optical and electrical properties. In-situ ellipsometry indicates that a ZnO layer as thin as 20 nm completely protects SnO2 from being reduced to metallic phases. This provides for shielding of textured TCOs, and hence rising solar cell efficiencies, too. Regarding light trapping efficiency we additionally investigated the smoothing of initial TCO texture when growing a-Si:H on top by combining atomic force microscopy and spectroscopie ellipsometry.

The window layer (p-layer) critically affects the conversion efficiency of a-Si:H based p-i-n type solar cells. This paper surveys possible alternative window materials in comparison with the standard boron doped a-SiC:H p-layer. A novel wide-gap amorphous silicon alloy, a-Si:F:H, will be shown to be a good alternative p-layer. As an example of nontraditional p-layer, boron based alloys (e.g., a-B:H) exhibit notable photovoltaic action in a-Si:H solar cells. We comment on attempts to deposit μc-Si alloy window layers on textured TCO-coated glass substrates.

We have systematically investigated the effect of H2 dilution on the stability of a-Si:H based solar cells. The results clearly show that the device stability against light soaking improves substantially with increasing H2 dilution until a threshold is reached. Beyond this threshold which depends on the substrate temperature, the stability no longer changes with further increase in H2 dilution. On the other hand, at a given ratio of H2 to the reactant gases, the device stability generally improves with increasing substrate temperature. Multi-step light soaking experiments have shown that devices made with H2 dilution saturate much faster (∼100 hours) under one-sun illumination and exhibit little overshoot effect in the recovery process, in sharp contrast to devices made without H2 dilution. Based on the simple two-component model for defect kinetics, these observations and the fact that the apparent saturation time coincides with the time constant of the “fast” defects strongly suggest that negligible amount of “slow” defects exist in materials made with H2 dilution. While H2 dilution generally suppresses the formation of microstructure giving rise to dihydride bonding and microvoids, the differences in the kinetics of light induced degradation cannot always be traced to obvious differences in these structural properties.

We have studied the saturation behavior and thermal annealing recovery of light-induced degradation in amorphous silicon-germanium (a-SiGe:H) alloy solar cells with various Ge contents. In contrast to amorphous silicon (a-Si:H) alloy cells, a-SiGe:H alloy cells with high microvoid density show similar thermal annealing behavior to those with lower microvoid density. Infrared studies show that unlike a-Si:H alloys, a-SiGe:H alloys with higher microvoid density do not show higher hydrogen or dihydride content. Moreover, even with a Ge-content of 41 %, there is saturation in degradation after light soaking for 1000 hours.

Our previously reported 7.5% stabilized efficiency single junction a-Si devices had relatively low Voc, around 0.80 volts. Devices with higher Voc tended to exhibit a decrease in Voc upon light soaking. This effect was found to be independent of the properties of the i-layer. Further optimizing the p layer and p-i interface has enabled us to produce single junction devices with a stable Voc of up to 0.86 volts and a 8.5% stabilized device efficiency has been established.

We have fabricated hydrogenated amorphous silicon alloy solar cells using hydrogen dilutions at 175 °C and 300 °C, and obtained improved photovoltaic characteristics in both the initial and degraded states for the highly diluted cells; both the fill factor and the open-circuit voltage exhibit higher values before and after light soaking. Infrared analyses reveal that for a given deposition temperature the amount of bonded hydrogen has similar concentrations between the high and low hydrogen diluted samples. Optical Modelling shows a 20 MeV difference in their optical bandgap. Defect densities obtained from constant photocurrent measurements give similar values for a given deposition temperature both before and after light soaking, inconsistent with solar cell performance.

We light-soaked a-Si:H pin solar cells with blue laser light through the front glass substrate or electron irradiated them through the back Al layer. By using short wavelength light and energy resolved electron beams we localize the damage preferentially near the p/i or n/i interface and thus study their effect on cell performance. I-V characteristics and quantum efficiency spectra were measured before and after the damage was introduced. We Modeled the changes in the quantum efficiency using the numerical device Model AMPS. Comparing changes in the measured and modeled quantum efficiencies (QE) we find that cells light-soaked with the blue laser have damage localized near the front of the device, and localized near the back for cells irradiated with electrons. This indicates that quantum efficiency measurements are sensitive to the location Of defects in the i-layer of the solar cell. However after long light-soaking and electron irradiation times the difference between the two types of damage becomes less visible in the QE Measurements.

We report results of stability tests of 4 ft2 triple-junction a-Si alloy photovoltaic (PV) Modules. These Modules were produced in ECD's 2 Megawatt (MW) continuous, roll-to-roll PV Manufacturing line during the early stage of optimization. The stable module efficiency after 600 hours of 1 sun light soaking at approximately 50°C under load, is 8%. This is the highest stable efficiency for large area (≥4 ft2) a-Si alloy PV Modules Made in a production line.

While a-Si:H tandem solar cells deposited on textured TCO substrates have already shown high efficiency and good stability [1], the light absorption profile in such cells is still one of the important quantities to know. An accurate absorption profile is required when we do solar cell analysis, design, and optimization, or when we do computer simulations.

With the computer program GENPRO2D [2] which was developed in our laboratory, and was based on the incoherent Multi-rough-interfaces (scattering) Model, we are able to calculate the light absorption profile in a tandem a-Si:H solar cell for different light spectra and different light incident angles, provided that the scattering of light at every interface in the cell is known. In this paper, we will present our calculated results with comparison to the measured results of a real tandem cell.

The introduction into a traditional p.i.n structure of two defective buffer layers near the p/i and i/n interfaces can improve the device stability and efficiency through an enhancement of the electric field profile at the interfaces and a reduction of the available recombination bulk centers. The defectous layer (“i'-layer”), grown at a higher power density, present a high density of defects and acts as “gettering centers” able to tailor light induced defects under degradation conditions. If the i-layer density of states remains below 1016 eV−1 cm−3 and assuming a Gaussian distribution of defect states, the gettering center distribution will not affect significantly the carrier population but only its spatial distribution.

We report here about a device numerical simulation that allows us to analyse the influence of the “i'-layer” position, thickness and density of states on the a-Si:H solar cells performances. Results of some systematic simulation from the ASCA program (AMorphous Solar Cell Analysis), and for different configurations will be presented.

This paper analyzes the charge transport in “tunnel junctions” of amorphous Si Material based multijunction solar cells and proposes some guidelines for making good “tunnel junctions” based on the analysis. The Mechanism of the current flow in these “tunnel” junctions is recombination. However, the recombination mechanism is not the usual localized recombination but is non-localized recombination. The energy analysis shows that the usual localized recombination will cause a large energy loss and result in much lower energy conversion efficiency and Voc than the experimentally measured values. In non-local recombination, opposite charge carriers located at different locations can recombine by tunneling into a defect state. Based on this Mechanism, a good “tunnel junction” should be thin, with a large defect state density in the middle region of the “tunnel junction”. Broad tail states material and a thin layer small band gap material in tunnel junctions may improve the non-local recombination by providing more intermediate states for charge to tunnel through.

We report a study of a-Si:H/a-SiGe:H dual-bandgap p-i-n type tandem solar cells and Modules. Refined bandgap grading in the a-SiGe:H bottom cell results in higher Jsc (higher red response) without loss in FF or Voc. IMprovements in the tunnel junction consisting of doped widegap a-SiC:H and μC-Si lead to higher FF and lower optical loss. Undoped widegap a-SiC:H is used as interface layers in the top junction. Initial conversion efficiencies of 11.4% and 10.5% have been obtained for small-area cells and one-square-foot modules, respectively. The double junction cells and modules exhibit good stability and saturation of light-induced degradation, showing only ∼15% decrease in conversion efficiency after more than 1000 hours of simulated AM1.5 illumination. The stabilized efficiencies are 9.6% and 9.1%, respectively, for small area cell and large area Module. We also present some findings from the computer simulations of multijunction solar cell performance using a simple lump-circuit device Model.

Dark current-voltage characteristics of one hundred twenty single-junction a-Si:H alloy solar cells were studied. Parametric Modelling of dark I-V Measurements was used to determine some of the cell parameters. Average shunt resistances were 31 to 1200 kΩ for intrinsic layer thicknesses of 200 to 800 nm, respectively. Current switching was observed during dark I-V Measurements; the current in the reverse-bias region varies approximately as the square of the voltage following the onset of switching. The I-V characteristics in the switched mode are not understood. Voltage history and scan rate play a role in dark I-V characteristics.

To explain realistic circumstances with regard to energy band profiles at the TCO/a-Si:C:H heterojunction, the ASPIN computer simulation has been used. Numerical calculations indicate that the increased interface defect densities result in a steep potential drop inside the interface region, while the rest of the work function difference extends into the p-layer. The detrimental effect of a-Si:C:H partial oxidation has been simulated by additionally increased density of states at a-Si:C:H surface, and its influence on the potential barrier has been analyzed. The impact of both TCO/a-Si:C:H interface states and a-Si:C:H surface states on the photoelectric properties of p-i-n a-Si:H solar cell is discussed, and a possible improvement of Voc is envisaged.

In this work we report experimental results on light induced metastability of a-Si:H p.i.n. devices with different Microscopic/Macroscopic structures and we discuss them in terms of improved stability through spatial control of charged defects grown during light exposure. By placing a thin (few Å) intrinsic layer (i') between both p/i and i/n a-Si:H interfaces we are able to reduce the effective degradation rate through spatial modification of the electric field profile in the device. The electronic transport and the stability changes that accompany the change in Microstructure (R) and hydrogen content (CH) of the i- and i'-layer, were monitored throughout the entire light induced degradation process and compared with the corresponding μτ product (for both carriers) inferred through steady state photoconductivity and Flying Spot Technique (FST) Measurements. Results show that the degradation rate is a function of CH and R of both layers and can be correlated with the density of microvoids and di-hydride bonding. Since the i'- layers have a higher Cfj bonded mainly as SiH1 radicals (R≈0.4), they act as an hindrance to the growth of the defect, in the active region, generating “gettering centers” whose localisation and density are tailored in such a way that they will control spatially the electric field profile during light exposure.

Preliminary results show improvements in film's stability when the interfacial layer is included. So future progress toward more stable and efficient a-Si:H solar cells will depend on a careful engineering design of the devices.

A technical trend for a-Si TFTs is their application to large-size, high-pixel density AM-LCDs such as XGA, EWS, and HDTV. In order to realize these LCDs, the TFT device characteristics must be improved. Future technologies, which will be necessary to fabricate TFTs with improved characteristics are as follows

(1) Fully self-aligned TFT technology: A SA-TFT structure reduces the feedthrough voltage caused by parasitic capacitance due to gate/source overlap. This results in an improved picture quality and a higher aperture ratio. Fabrication of such a structure would require ion doping technology.

(2) Ion doping technology: This non-Mass-separated implantation technique has large area doping capability and much higher doping speed compared to conventional ion implantation technique. The Major problems with the ion doping technique is the implantation of unwanted species which deteriorate the quality of source/drain and channel regions of TFTs.

Recently we described a new computer program which used a defect pool model of the density of states in a thin film transistor. With this model we modelled the effects of thermal bias annealing on the field effect characteristics of amorphous silicon thin film transistors, and used these results to determine key defect pool parameters. Here we use this program to provide new insight into other features of thin film transistor properties, such as activation energies of conduction as a function of gate bias, details of the subthreshold slopes, Movement of the apparent flat band voltages, and contact effects including current crowding. These effects are all successfully modelled and explained with our first principles defect pool Model, and compared with experimental results on a variety of material systems. In particular we show how measurements of the apparent flat band voltage produce potentially misleading results, and explain the inconsistency between observations of much less current crowding in contacts compared to predictions based on observed bulk defect densities and conventional Models.