Application of Diamond Films to Electric Propulsion: Low Energy Sputter Yield Measurement and MPD Plasma Assisted Chemical Vapor Deposition

Citation

Blandino, John Joseph (2001) Application of Diamond Films to Electric Propulsion: Low Energy Sputter Yield Measurement and MPD Plasma Assisted Chemical Vapor Deposition. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/270K-TR61. https://resolver.caltech.edu/CaltechETD:etd-03252005-143838

Abstract

Progress made in the area of chemically vapor deposited (CVD) diamond films has led to its consideration for a number of novel applications. One potential application under evaluation at the Jet Propulsion Laboratory involves the use of diamond films as coatings for ion thruster electrodes which are subject to sputter erosion. In order to assess their benefit in mitigating sputter erosion as a failure mode it was necessary to first measure the erosion rate compared with molybdenum and carbon-carbon when subjected to ion bombardment.

Sputter yields were measured for polycrystalline diamond, single crystal diamond, a carbon-carbon composite, and molybdenum subject to xenon ion bombardment. The tests were performed using a 3 cm Kaufman ion source to produce incident ions with energy in the range of 150 - 750 eV and a profilometry-based technique to measure the amount of sputtered material. The yields increased monotonically with energy with values ranging from 0.16 atoms/ion at 150 eV to 0.80 at 750 eV for the molybdenum and 0.06 to 0.14 for the carbon-carbon. At 150 eV the yield for both diamond samples was 0.07 and at 750 eV, 0.19 and 0.17 for the CVD and single crystal diamond respectively. In terms of erosion rate, this translates into a factor of 7 - 12 lower erosion rate for diamond compared to molybdenum and at least a factor of 1.5 compared to carbon-carbon.

With the erosion rates established, the remaining effort concentrated on the experimental and analytical investigation of an electromagnetic (magnetoplasmadynamic or NIPD) plasma source for diamond CVD using a mixture of methane and hydrogen. Specific questions to be addressed included identifying the implications of higher velocity, dissociation, and ionization levels than those expected in an electrothermal (arcjet) plasma source. An experimental facility was used in which the process conditions produced were representative of the high temperature, ionization, and dissociation levels one would expect from an MPD in an actual reactor although not the high velocity. Numerous trials were conducted using methane to hydrogen mixture ratios of 1.5 - 3.5 percent by volume, four different methane injector configurations, and substrate biasing at potentials of 25 - 75 V positive with respect to facility ground. These tests were performed at discharge currents of 700 - 950 A at approximately 18 V (12 - 17 kW).

Crystalline films were produced with predominantly (110) oriented facets. X-ray diffraction spectroscopy was used to identify at least one unambiguous diamond peak in each sample with numerous films containing metal or metallic carbide impurities. Growth rates of 0.8 to 6.3 [micrometers]/hr were measured. The films all exhibited poor Raman spectra with no well defined peak at 1332 cm⁻¹ and a broad background possibly due to high background levels of nitrogen, defects, and contamination.

The question of high velocity and high ionization level was investigated analytically using estimated hydrogen MPD plume data from the literature. For conditions expected with an MPD source, Knudsen numbers in the plume are calculated to be approximately 0.1 in the free stream and less than 0.01 (i.e., transition-continuum boundary) in the stagnation boundary layer. The heavy particle static temperature in the plume is expected to be on the order of 10000 K in the core. For Mach numbers in the range 1.0 - 2.0, the stagnation temperature can be expected to reach peak values of over 20000 K. This temperature far exceeds the range of available thermodynamic and transport databases for hydrocarbon mixtures needed to model stagnation boundary layer chemistry and growth rates, so a scaling relation was used to obtain a relative comparison of the atomic hydrogen mole fraction at the substrate for an electrothermal and electromagnetic accelerated plasma source. Because of its higher jet velocity and lower operating pressure, the electromagnetic: source operates in more of a convection dominated regime resulting in a calculated hydrogen mole fraction at the substrate approximately 40 percent higher than that predicted for an electrothermal accelerator.

An energy balance was used to determine an upper bound on the level of electron heating obtainable for a given bias current density. Results show that even for pressures of a few Torr and ionization fractions of 5 - 25 percent, the required current for a few thousand degree increase in electron temperature over the heavy particle gas temperature is on the order of several tens of A/cm². From this analysis it is concluded that high plasma conductivity in an MPD plume and electron energy losses through inelastic molecular collisions will preclude the effective use of ohmic heating of the electrons as a means of enhancing electron catalyzed dissociation in the boundary layer.

Estimates were made of the optimal residence time for methane decomposition in the plume in order to maximize the flux of methyl radicals and atomic hydrogen to the substrate; two species which have been identified as having a significant role in high rate, high quality diamond synthesis. At pressures of 100 Pa and 333 Pa, which approximate the pressure in the plume and stagnation regions, there is little recombination of the hydrogen even at a temperature of 5000 K which one could expect well into the thermal boundary layer for an MPD stagnation flow. The methyl mole fraction reaches a maximum at a residence time of 1 - 3 [millisec]. Achieving adequate entrainment and mixing of the methane in the hydrogen jet oil such a short time scale is a very difficult challenge.

Two major conclusions of this thesis are: 1) The lower erosion rate measured for CVD diamond as compared with molybdenum makes coating of grid electrodes with CVD diamond a possible option for extending the lifetime of ion thrusters worth further investigation. 2) Based on both the experimental and analytical investigation of the MPD source for diamond deposition, the potential for higher growth rates than those obtainable with more conventional plasma sources is not significant enough to offset the disadvantages associated with contamination of the film due to metal vapor from the high current electrodes and poor entrainment of the carbon precursor gas due to the short, residence time in the plume.

Thesis Files