J. Vac. Sci. Technol., A, 9, 682-687 (1991).
Nguyen, P. L., Turkot Jr., R. B., Ruzic, D. N.
A compact and inexpensive cubical energy analyzer has been developed to measure ion and electron energy distributions at the boundaries of a rf plasma. It consists of two grids parallel to the plasma boundary behind which lies a 1 cm×1 cm×1 cm cube. The absorbed current and voltage on the grids and on each face of the cube can be monitored and biased independently. A Helmholtz coil on the outside of the cube allows a 0–15 G field to be produced perpendicular to the path through the gridded apertures. The advantages over standard gridded energy analyzers include direct measurements of ion-induced secondary electrons and separation of electrons, negative ions, and heavy and light mass positive ions to different collection surfaces. Measurements of the ion and electron energy distributions at the grounded electrode of a parallel plate 13.56 MHz plasma discharges clearly show a time-varying rf sheath and a non-Maxwellian electron energy distribution. Data is presented for Ar, CH4, and H2 plasmas.
J. Vac. Sci. Technol., A, 9, 614-618, (1991).
Myers, A. M., Doyle, J. R., Abelson, J. R., Ruzic, D. N.
Monte Carlo simulations of the particle transport process during dc magnetron sputter deposition were performed to determine the energy and angular distributions of the energetic deposition species. The model itself is quite general, and here we present the specific example of hydrogenated amorphous silicon film growth. This process involves the sputtering of a silicon target in an argon-plus-hydrogen plasma. The three-dimensional model incorporates fractal trim data for the distribution of Si energies and emission angles sputtered from the target surface. Modified “universal” interatomic potentials are used to determine the scattering processes during gas phase transport. Energy and angular distributions of the deposition flux reaching the substrate are calculated as a function of pressure from 0.01 to 5.5 mTorr. As the pressure increases we find that the average energy per deposited atom remains essentially constant, but the energy and angular distributions of the arrival flux change dramatically.
J. Vac. Sci. Technol., A, 9, 702-706 (1991).
Rossnagel, S. M., Schatz, K., Whitehair, S. J., Guarnieri, C. R., Ruzic, D. N., Cuomo, J. J.
In a conventional, axial electron cyclotron resonance (ECR) plasma source, the substrate is typically in the diverging magnetic field region immediately downstream from the ECR region. Due to the high electron mobility along magnetic field lines, the substrate potential may influence the properties of the plasma. A substrate electrode consisting of seven, individually biasable, concentric rings was positioned below the ECR source, such that the potential at the termination of the field lines could be controlled. The ECR source used was a commercial, 1.5-kW device operating at 2.45 gHz and utilized two large electromagnets to provide magnetic fields in the range of 200–1000 G. The plasmas were diagnosed by means of a microwave interferometer in the source region, an automated Langmuir probe, and an array of ground potential metal rings in both the source and sample regions of the chamber. The results suggest that ions in this source are poorly confined (by the magnetic field) while electrons are well confined. The electron density could be increased slightly by electron reflection from the substrate region, or could be depleted strongly by substrate potentials exceeding the floating potential. The net efficiency of these sources is low: only about 14% of the ion flux from the plasma is incident on the sample position. The rest of the ions are lost at the source walls.