Erten Eser

Measurement of plasma discharge characteristics for sputtering applications

Plasma and characteristics of rf and dc discharges for sputtering were measured by a cylindrical‐type electrostatic probe. Space potentional, floating potential, electron temperature, and plasma density were monitored as a function of discharge voltage, substrate bias, and probe position within the chamber. A description of the probe and supporting circuit used to plot the current–voltage characteristic is given. The electrostatic probe was shown to be a useful tool for characterizing sputtering process and with suitable system calibration, can be used to enhance repeatability of the sputtering process.

Light‐induced degradation and thermal recovery of the photoconductivity in hydrogenated amorphous silicon films

Kinetics of light‐induced degradation and thermal recovery of the photoconductivity in hydrogenated amorphous silicon have been investigated. Rate equations for these processes were derived using the principles of reaction rate theory. Both degradation and recovery were found to be thermally activated with activation energies of 0.1 and 1.35 eV, respectively. Experimental results support the bond breaking model for the degradation. Stabilization of broken bonds is achieved by structural relaxation. Thermal recovery proceeds through hydrogen hopping from one dangling bond to another. Experimental evidence suggests the existence of an upper limit in the density of light‐induced defects. A microscopic model, consistent with experimental observations, is proposed for the light‐induced generation and thermal recovery of defects in a‐Si:H.

Plasma characterization in sputtering processes using the Langmuir probe technique

Abstract The Langmuir probe technique, which has long been used in the field of plasma physics, can be a very powerful tool in characterizing the sputtering process. In this context, theoretical formalism, design requirements and the experimental procedures necessary for the application of this technique to both r.f. and d.c. sputtering are presented. It is shown that the plasma parameters can be determined in a simple way only if the probe operation is in the orbital motion limited regime. This condition is fulfilled for a cylindrical probe when the ratio of the probe radius to the Debye length of the plasma is less than 3. In the case of r.f. discharges erroneous data are obtained if a large r.f. current is flowing to the probe. This current can be blocked by a suitable LC circuit. Using the Langmuir probe technique the variations in the plasma parameters in d.c./r.f. sputtering discharges as a function of target voltage/power and argon pressure are determined.

The effect of bias on d.c. and r.f. sputtered WC-Co coatings

Abstract The structure and chemistry of WC-Co coatings, r.f. and d.c. sputter deposited at various bias levels, were analyzed using techniques such as electron microprobe analysis, Auger electron spectroscopy, scanning electron microscopy and X-ray diffraction. These analyses, together with the measurement of the ion density in the plasma surrounding the substrate, showed that the coating characteristics are controlled by the substrate bias-plasma density combination. An increasing negative bias was shown to inhibit the preferential nucleation defects as well as the columnar growth in r.f. sputter deposited coatings but not in d.c. sputter deposited coatings. The reason for this is that, although the plasma density was only 3 × 10 10 cm -3 in the d.c. discharge, it was 8 × 10 10 cm -3 in the r.f. case. As a result, in r.f. sputtering, not only the substrate temperature but also the re-sputtering from the growing film relative to the sputtering from the target was higher. A higher carbon and cobalt deficiency observed in r.f. deposited coatings also resulted from the higher level of ion bombardment during r.f. sputtering. The carbon deficiency is believed to be associated with the temperature rise, causing re-evaporation of the carbon atoms before they reach equilibrium with the substrate. The decrease in the cobalt content is the result of the increased re-sputtering. A carbon-to-tungsten ratio of less than unity encountered under zero bias conditions, for both r.f. and d.c. sputtering, indicates the reflection of some portion of the carbon atoms. With an increasing level of ion bombardment the crystal structure of the coatings was found to change from f.c.c. β-WC 1− x to h.c.p. α-W 2 C.

Structural and compositional characterization of sputter‐deposited WC+Co films

WC+Co films were sputter deposited by both rf and dc techniques on the electrically floating flat beryllium substrates. Film thickness varied from 1.0 to 1.68 μm. Structural and compositional analysis of the coatings were done by x‐ray diffraction, SEM, EMA, and AES. Columnar growth and preferential nucleations present in the coatings by dc sputtering were not found in the coating by rf sputtering. X‐ray diffraction showed that dc sputter‐deposited coatings are beta tungsten carbide. Compositional analysis showed that the films were deficient in both carbon and cobalt and this deficiency was more pronounced in the films deposited by rf sputtering. Oxygen concentrations were found to be lower and argon concentrations higher for films deposited by rf sputtering. Results show the effect of ion bombardment on the substrates during deposition by the negative self‐bias obtained by electrically floating the substrates. Carbon deficiency is associated with the low sticking coefficient of carbon due to its low mass.

Friction and wear results from WC+Co coatings by −dc‐biased rf sputtering in a helium atmosphere

WC+Co films were deposited on flat polished beryllium substrates by rf sputtering with selected values of −dc bias applied to the substrates. Topographical, structural, and compositional analysis of the coatings were done by SEM, x‐ray diffraction, and EMA. X‐ray diffraction results show the deposits to vary from amorphous to αW2C. Compositional analysis showed variations in carbon and cobalt in relation to the bias level. Friction and wear tests carried out in a helium atmosphere by rotating one specimen in contact with another at constant speed and load for a specified time showed coefficient of friction and wear effects to be related to substrate bias. WC+Co coatings have been shown to be effective wear‐resistant surfaces when applied by predetermined deposition parameters.

Reversibility of (Ag,Cu)(In,Ga)Se2 electrical properties with the addition and removal of Na: Role of grain boundaries

The Na content of (Ag,Cu)(In,Ga)Se2 films was cyclically adjusted using a novel method involving cycles of water rinsing at 60 °C followed by heating in air at 200 °C to remove Na and evaporation of NaF to re-introduce Na back into the film. The low temperatures and short heating times ensure that Na is removed only from grain boundaries while leaving grain interiors unaffected. Cross-grain conductivity and Seebeck coefficient were measured during this removal procedure and both measurements decreased when Na was removed and both recovered upon the re-addition of Na, consistent with an increase in compensating donor defects in the absence of Na. These results demonstrate that Na reversibly affects the electrical properties of grain boundaries. We propose that Na reversibly passivates donor-like defects such as InCu double donors at grain boundaries.

Pilot-scale manufacture of Cu(InGa)Se/sub 2/ films on a flexible polymer substrate

Cu(InGa)Se/sub 2/ films were deposited in a multi-source inline evaporation system onto a 6-inch wide polyimide/Mo web. The materials properties of the Cu(InGa)Se/sub 2/ films were characterized along and across the web over a five-foot length. Thickness and composition are uniform within measurement error and larger variations were found from deposition to deposition. Average device performance over a five-foot length was between 8 to 9% and the best cell had an efficiency of 12.1%. A Ga-In-Se precursor layer deposited on the polyimide/Mo web improved adhesion of the Cu(InGa)Se/sub 2/ film.

Grain structure of CdS and Cd1−xZnxS films

The grain structure of vapor-deposited CdS and Cd1−xZnxS films was analyzed using standard metallographic techniques and found to be columnar. Columnar grains of submicron diameter grow initially normal to the substrate surface. During further film growth they tilt towards the source and grow in size as a result of the elimination of less favorably oriented grains. An appreciable number of grain boundaries were low energy subgrain boundaries. Quantitative analysis of the grain size on the film surface showed that the average grain diameter is thermally activated and that at low temperatures the grain structure is “bimodal”. The bimodal structure was more common in Cd1−xZnxS films. It is suggested that grain boundary motion plays a role in the development of the grain structure in these films. Possible effects of the observed microstructure on the performance of Cu2S/Cd1−xZnxS solar cells are discussed.

Manufacturing process for interconnected submodules of hydrogenated amorphous silicon photovoltaic panels

Large scale deposition of hydrogenated amorphous silicon (a‐Si) for photovoltaic panels is discussed. The ‘‘box carrier’’ concept is introduced for hydrogenated a‐Si deposition. The spectral response studies prove that cross contamination is not present. Using this new concept, the average of 72 panels in one run is 5.6%. An I‐V curve for a square foot panel with an efficiency of 7.2% is shown. The utilization efficiency of silane is much higher (61%) than what has previously been reported (8%).