Brian S. Felker

Electrical optimization of plasma-enhanced chemical vapor deposition chamber cleaning plasmas

Fluorinated gas discharges are widely used by the semiconductor industry in etching and chamber cleaning applications, but the performance of these discharges varies in unpredictable ways, for unknown reasons believed to be electrical in origin. To investigate possible mechanisms for this behavior, we have measured the electrical characteristics of NF3/Ar, CF4/O2/Ar, and C2F6/O2/Ar chamber cleaning plasmas at 6.7–267 Pa in a 13.56 MHz, capacitively coupled, parallel-plate reactor, using radio-frequency current and voltage probes and optical emission spectroscopy. From the measurements, power losses in the external circuitry surrounding the discharge were determined. Furthermore, using the measurements and equivalent circuit models, the mechanisms by which power was absorbed within the discharge itself were investigated. Power was absorbed most efficiently at particular values of the discharge impedance. These optimal impedances occur in the middle of a transition from capacitive impedances at low pressure...

THE ROLE OF DILUENTS IN ELECTRONEGATIVE FLUORINATED GAS DISCHARGES

To study the role of diluents in NF3plasma processing we have correlated SiO2 and plasma chemical vapor depositionsilicon nitride (SiN) etch rate measurements with rf electrical impedanceanalysis. A series of rare gas (He, Ar) and molecular (N2, O2, N2O) mixing gases were added to NF3plasmas at different pressures to understand the effect of diluents on the chemical and physical properties of NF3discharges. The etch rate experiments show that for NF3plasmas the choice of mixing gas can have a profound effect on the etch rates of SiO2 and SiN with 25 mol % NF3 in Ar yielding the highest rates and 25 mol % NF3 in N2O the lowest. The electrical measurements revealed that the diluents have a profound effect on the plasma impedance and actual power dissipated in the discharge. NF3plasmas diluted with Ar exhibited the lowest impedances and highest real power dissipation at higher pressures while N2O diluted plasmas had the highest impedances and lowest power dissipation levels. These results indicate that the diluents which result in the highest power dissipation in the discharge, at high pressures, result in the highest etch rates. We propose that the dominant role of the diluent in NF3plasmas is to control the electronegativity of the discharge, and thus to control real power dissipation. This function is in contrast to the role of diluents in plasmas based on other fluorinated gases, where the diluents are seen as primarily affecting the concentrations of reactive species which deposit or remove materials from the surface of the thin film being processed.

Ion energy distributions and optical emission spectra in NF3-based process chamber cleaning plasmas

To minimize ion bombardment induced damage in NF3-based chamber cleaning plasmas, we have studied the effects of diluent gases and reactor pressure on ion energy distribution functions in NF3 plasmas. We have utilized plasma ion mass spectrometry, ion energy analysis, and optical emission spectroscopy in 25 mol % NF3 plasmas with argon, helium, and oxygen diluents. We have also compared the NF3-based plasma measurements to those of 50 mol % C2F6/O2 plasmas. We have demonstrated that diluting with helium and operating at higher pressures will reduce ion energies in NF3 plasmas while maintaining superior chamber cleaning performance. In addition, we have correlated the intensity ratio of specific argon emission lines to average ion energies at the grounded electrode. This correlation provides a practical diagnostics tool for further optimization work.

Optimizing utilization efficiencies in electronegative discharges: The importance of the impedance phase angle

We have investigated the operating conditions that result in the greatest utilization efficiencies (UEs) of NF3, CF4, and C2F6 in a capacitively coupled GEC reference cell. We have also independently measured the rf electrical characteristics and optical emission spectra of the plasmas. To avoid inadvertently attributing changes in the UE, discharge impedance, rf currents, or atomic emission intensities to parasitic losses in the matching network or rf delivery system, the rf generator was adjusted to ensure that the same amount of power was dissipated within each discharge. For the NF3 plasmas, argon was used as a diluent and both the NF3 concentration and reactor pressure were varied. For the CF4 and C2F6 based plasmas, the gas compositions were fixed (86 mol % CF4/O2 and 50 mol % C2F6/O2) and the reactor pressure was varied. The greatest NF3 UEs occurred within a narrow range of NF3 partial pressures. The greatest CF4 and C2F6 UEs occurred within a narrow range of reactor pressures. For all mixtures, o...

Materials selection for HBr service

Abstract This study was undertaken to determine the compatibility of hydrogen bromide (HBr) with common materials of construction used for specialty gas delivery systems. Reactions between reactive gases and materials of construction can result in the formation of particles and volatile metal complexes as well as the creation of corrosion products that can retain water. We found that when moisture is below 1 ppm v (designated as anhydrous ), bromine from HBr is not incorporated beyond the native oxide of electropolished 316L stainless steel (EP316L) and no macroscopic degradation of the metal occurs. Also, if adequate purging and evacuation procedures are followed to remove the HBr, this material can be exposed to moist air without diminishing the initial surface quality. However, if adequate precautions are not followed to eliminate water in the presence of HBr, iron-and bromine-rich crystalline deposits form on the surface. Purge and evacuation procedures are inadequate for removal of the reactive species on this surface and corrosion proceeds upon subsequent exposure to air. EP316L exposed to HBr containing 1700 ppm v H 2 O appears visually unaltered, but close inspection by SEM reveals the onset of corrosion. Of the materials examined in this study, Nickel-200 and Hastelloy C-22 are more resistant to HBr corrosive environments. In contrast, deleterious reactions occur between anhydrous HBr and elemental iron and the iron-rich surface of oxygen-passivated 316L.

Study of Silicon Surface Roughness by Atomic Force Microscopy

Progress is reported in developing reliable methodology for imaging silicon surfaces with the atomic force microscope (AFM). A new form of AFM, known as tapping mode AFM, has been found to provide the best quality data for surface roughness determinations. Commercially available colloidal gold spheres have been used to fabricate tip characterization standards and are used to report tip size with roughness data. Power spectral density calculations are shown to provide a useful roughness calculation based on lateral wavelength.

The Effects of Chemical Vapor Cleaning Chemistries on Silicon Surfaces

This study explores the effects of two chemical vapor cleaning chemistries on silicon surfaces. The silicon surfaces are not significantly roughened by exposure to either process. Trace amounts of fluorine are found on the surfaces exposed to 1,1,1,5,5,5-hexafluoro-2,4-pentanedione (HFAC). A thin silicon nitride film forms on the silicon surface as a result of exposure to the HMDS process and is attributed to the H 2 /N 2 plasma treatment used in the first step of the process.

Chemical Vapor Cleaning of Sodium from SiO 2

Chemical vapor cleaning (CVC) is an emerging technology which has been used to remove transition metal contamination from wafer surfaces. CVC is a gas phase/surface reaction which does not incorporate any wet steps nor condensation of reagents onto the wafer surface. In previous work, we reported the effective removal of trace iron and copper from native oxide surfaces using 1,1,1,5,5,5-hexafluoro-2,4-pentanedione (HFAC) and other chelating or coordination compounds. In general, surface metal contaminants form volatile reaction products with these CVC reagents at relatively low temperatures (