Measurement of Depositing and Bombarding Species Involved in the Plasma Production of Amorphous Silicon and Silicon/Germanium Solar Cells

Quarterly Technical Report

Dec. 1, 2003 – Feb. 29, 2004

DOE Interagency Agreement De-AC36-02G010244

Principal investigator: Alan Gallagher

Contributors: Peter Horvath, Damir Kujundcik

DOE Program Officer: Ms. Lizana Pierce

NREL Technical monitor: Dr. Bolko Von Roedern

The technical approach to this project is to utilize mass spectrometry to measure radical and ion species that arrive at the substrate of a plasma-enhanced-chemical-vapor-deposition (PECVD) reactor. Threshold ionization mass spectrometry (TIMS) is utilized for the radical detection. A small-scale reactor mimics those used to produce hydrogenated amorphous (a-Si:H) and microcrystalline (μc-Si) silicon and silicon/germanium (a-Si:Ge:H) solar cells. Radio frequency (RF) and high frequency (HF) discharges will normally be studied, and the reactor utilizes a similar electrode gap (2-3 cm), substrate temperature (20-250 °C), gas pressures (0.1-5 Torr), gas mixtures and discharge power density to that used in an industrial reactor.

As noted in previous reports, the vacuum, gas handling and silane-discharge apparatus is constructed and operating properly. This includes the pyrolyser that breaks up the toxic or flammable effluents (silane and/or germane) into hydrogen gas and solid. The quadrupole mass spectrometer (QMS) is also working well, using our home designed and built control electronics. This includes the ability to operate the QMS at different RF frequencies to allow accurate H atom and SixHn radical detection. The discharge is working as designed, with the discharge RF voltage under computer control. The differential pumping through three chambers, and finally into the mass spectrometer chamber, is working as designed.

A major effort during the last quarter has been testing and improvement of the threshold- ionization electron gun and ion-collection optics. This has been done first by utilizing argon gas, since it does not coat the chamber with a-Si:H or pyrolyse on the hot cathode of the electron gun. Several difficulties were encountered in trying to obtain the designed electron beam behavior. In particular, the beam was severely deflected from the design path. The electron beam is magnetically confined, but the confining field is only ~ 20 Gauss to avoid a fringing field in the discharge region that could influence the discharge. Under these “weak confinement” conditions other sources of magnetic field, although relatively minor, can distort the electron path. A common solution with electron beams is to utilize molybdenum electron-gun elements, but many of internal components other than the electron gun elements are also nearby. The three-chamber differential pumping arrangement required for the experiment requires several large metal vacuum components near the electron gun, and it was most convenient to make these of stainless steel since this is most easily machined and welded. The apparatus parts near the electron gun were made of type 304 stainless steel, which is normally non magnetic. However, machining strains the stainless steel surfaces, and can thereby produce magnetization. Having discovered problems with the electron beam, we evaluated the magnetic properties of these components, and decreased their magnetism with a combination of annealing and “degausing”, where large alternating fields are used to randomize the dipoles within each part. Another change was to increase the confining field, to about 40 Gauss, thereby decreasing the spiraling radius of the electrons. Finally, the cathode-anode region was more fully shielded from nearby grounded surfaces, to minimize off-axis electric fields that increase spiraling. The combination of these mitigation efforts has produced the designed electron beam behavior; in particular 1-5 μA passing about 7 mm below the orifice to the discharge chamber.

Another accomplishment during this quarter is programming of a computer to control the experiment and take data. This involves controlling the discharge power and timing, and recording several types of data. The measurements include the discharge pressure, RF voltage, stable-gas constituents, ionizer electron beam current and energy, the a-Si:H film growth rate, the mass selected for the mass spectrometer, and the ion signal from the mass spectrometer.

A major change in staffing of the project occurred in December. Dr. Karoly Rozsa returned to his laboratory in Budapest, Hungary, after about 5 years of off-and–on work on the NREL and now DOE contracts. He was typically here for about half of the year, returning to guide work at his laboratory in between work in Boulder. We will miss his guidance and council, gained through several decades of research into the physics and applications of gas discharges. The current program is now staffed by two Ph.D. candidate students, and of course the PI. These students are Peter Horvath, a Physics Department student of Budapest University, and Damir Kujundzic, a EE student at the University of Colorado.