The 25th International Conference on Amorphous and Nano-crystalline Semiconductors
August 18–23, 2013 Toronto, Ontario Canada
a-Si:H/c-Si Interface III
Chair: Stephen O'Leary, University of British Columbia-Kelowna
Stabilizing Amorphous Silicon Against Photodegradation using Nanocrystalline Silicon
1. Department of Physics, Indian Institute of Technology, Kanpur India
2. Materials Science Program, IIT Kanpur, India
Amorphous silicon (a-Si:H) thin films containing varying amounts of nano-crystalline silicon (nc-Si:H) are prepared by the Plasma Enhanced CVD method. The dilution of the silane gas with hydrogen is used as the main parameter to vary the crystalline fraction in (χ) the films. High Resolution Transmission Electron Microscopy shows the presence of nano crystals. The films containing nc-Si:H are found to be more stable against light soaking (LS), with the stability improving with increasing χ. Raman spectrum shows a peak at 490 cm–1, which overlaps the amorphous peak at 480 cm–1 and has been attributed to the Intermediate Range Order (IRO). The IRO peak grows with increasing χ. This suggests that the IRO is probably responsible for the improved stability. We also measured the electrical conductivity and thermopower as functions of temperature to obtain the width of the long range potential fluctuations (LRPF) arising from the inhomogeneities. We find that the films with higher χ degrade less and LRPF are also smaller for them. Thus, the films are most stable when χ is large. However, these are not the best suited for making the solar cells. Other material parameters (e.g. band gap, photosensitivity etc.) favor a-Si:H, without any crystalline fraction. The best compromise, therefore, is the material which is on the verge of crystallization, as observed in practice.
Work supported by CSIR, New Delhi
Keywords: amorphous silicon, nanocrystalline silicon, Staebler-Wronski effect, heterogeneities, photodegradation
High Current Density, Hybrid Nanocrystalline/Amorphous Silicon Schottky Diodes
Department of Electrical Engineering, Princeton University, Princeton, New Jersey, 08544, USA and Princeton Institute for the Science and Technology of Materials (PRISM)
We are developing a manufacturable process for large-area systems, based on laminating multiple sheets of thin-film electronics and CMOS ICs, to form "sensing wallpaper" . In order to transmit signals and power from one sheet to another in a robust manner, which is tolerant to misalignment, reliable, and flexible, we use non-contact inductive links. When a sheet of thin-film electronics receives an AC-modulated signal it has to down-convert it to the envelope signal, leading to the need for a thin-film diode for AC-to-DC rectification. In this talk we present a hybrid nanocrystalline silicon / amorphous silicon Schottky diode developed for this application. The diode has a current density of 5 A/cm2 at 1 V, one of the highest current density thin-film diodes (of any technology processed at < 200°C) to date.
The diode is deposited at 180°C by plasma enhanced chemical vapor deposition at 70 MHz frequency. The structure consist of a Schottky chrome contact, 750 nm n- nc-Si, 150 nm n+ nc-Si, and a chrome ohmic contact. The Schottky interface is formed between the chrome contact and an amorphous silicon incubation layer, deposited before the n- nc-Si film has nucleated. The n- film is doped by background impurities and has an ionized impurity density of 1016 cm–3 (from CV measurements). The high current density of the device (5 A/cm2 at 1V) can be attributed to the high conductivity of the nc-Si, which means that unlike a Schottky diode formed entirely out of a-Si, it is not affected by space charge limited current  . The device also has a high ON-to-OFF current ratio (>1000 V for +1 V/ –8 V). A Schottky device has the advantages over a p-n junction of reduced turn-on voltage and the need for only a single doping source.
For our application, it is desirable to operate at high frequencies, since inductive links are more power efficient at high frequencies. A 0.01 mm2 hybrid diode has a series resistance of 120 Ω and a capacitance of 7 pF, leading to a calculated critical frequency (fc=1/(2πRSCJ)) of 190 MHz. Diodes have been tested in a half-wave rectifier configuration with a parallel 100 kΩ resistive and 10 nF capacitive load. At 10 MHz for an AC signal with a 4V amplitude, we obtain a DC rectified voltage of 3.4 V. Testing at higher frequencies is ongoing.
 Y. Hu et al., VLSI Symposium, June 2013 (In press)
 J. Sanz-Robinson et al., Device Research Conference (DRC), June 2012
Keywords: Schottky diode, inductive link, large-area system, rectifier, nanocrystalline silicon
A Comprehensive Model for Injection-dependent Charge Carrier Lifetime Curves
Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, Lise-Meitner-Campus, Hahn-Meitner-Platz 1, 14109 Berlin, Germany
Nanolayers of different materials, such a-Si:H, SiO2 and Al2O3, for the passivation of recombination-active electronic defects at crystalline silicon surfaces and interfaces have been intensively characterized by injection-dependent charge carrier lifetime curves (CCL curves) in the past. Therefore the quantitative analysis of such CCL curves has gained rising interest during the last decade and elaborate models have been developed for extracting material parameters such as interface defect density and interface defect charge by inverse modeling. However, the precise interpretation of different features of such curves is still a matter of debate and different physical processes have been suggested as explanation.
Therefore, we have developed an easy-to-use computer program that integrates the most important physics models such as: (1) recombination pathways at the interface in the bulk and in space charge region as function of the injection-level and the surface band bending, (2) defect density and capture cross sections as function of the band gap energy, (3) the amphoteric nature of dangling bond defects, (4) the defect-pool model for the calculation of Fermi level-dependent defect distributions and (5) injection-dependent recharging of interface defects. Furthermore our tool comprises a robust minimization algorithm such that it can be applied for automatic inverse modeling of experimental CCL curves to extract material parameters.
Theoretical studies will be presented to highlight the underlying physics of different features of the CCL curve. Furthermore the uniqueness of solution and error estimation for inverse modeling of experimental CCL curves will be discussed. The applicability of the model to determine the passivation mechanisms of nanolayers of different material will be shown. Material systems that will be investigated include a-Si:H, SiO2 and Al2O3 single layers as well as layer stack systems with combination of materials and additional TCO capping layers. The dominant passivation mechanisms of those layer systems, as deduced by inverse modeling, will be discussed and compared to results of alternative characterization methods such as surface photovoltage.
Furthermore the model, originally developed for the analysis of CCL curves, represents a zero-dimensional device simulator and can be applied for convenient and fast modeling of solar cells. It will be shown that not only predictions for the open circuit voltages as function of the nanolayer and absorber properties can be obtained but also for the influence of the field-effect passivation on the fill factor of solar cells can be made. Using the example of a heterojunction solar cell, a sensitivity analysis of implied open circuit voltage and fill factor as function of interface defect density and interface charge will be presented to serve as a guideline for solar cell optimization.
Finally, the functionality of the computer program for analysis of CCL curves and for zero-dimensional device simulation will be shortly presented; it will be made freely available to the scientific community.
Keywords: lifetime, modeling, interface passivation, photovoltaics