A numerical study of neutralization and sputtering processes in the NSTAR thruster

A computational model is used to calculate the neutralization lengths upstream and downstream of a single aperture of the optics in the NSTAR ion thruster. These data are compared to theoretical results, showing that the theoretical results are not accurate. The computed data are then analyzed to find appropriate curve fits giving the neutralization lengths for a range of operating conditions. The resulting relations compare very well to the modeled data. Molybdenum sputtered from the optics is also simulated. The amount of molybdenum depositing on the optics is calculated, as is the amount which flows upstream and downstream of the optics. It is found that nearly 25% of the sputtered material deposits on the accel grid barrel. Results are provided to illustrate the density and velocity of the sputtered molybdenum. Introduction The success of the Deep Space One mission and the NSTAR ion thruster have led the way to further research in the operation of ion thrusters. As such, computer simulation of ion thruster operations is more important than ever. Experiments on ion thrusters, especially life-tests, are costly and time-consuming. Computer simulations attempt to reduce both of these factors by predicting behavior for ion thrusters relatively quickly and cheaply. Ion thrusters are not simple to model, so continual improvements and modifications must be made to ensure that the best possible results are obtained. Recent advancements in ion thruster simulation include fully three-dimensional codes developed by Wang et al and Nakayama and Wilbur. Wang’s code focuses on modeling erosion behavior in the NSTAR thruster, while the code developed by Nakayama has been applied to high specific impulse thrusters. The focus of this paper is on the improvement of a DSMC-PIC code developed to simulate ion thruster optics. This code has previously been verified to accurately model ion thrusters by Boyd and Crofton, and it is an excellent basis for providing further modeling of the details of ion thruster operation. In its present state, the code assumes that the plasma flow is neutralized at both the upstream and downstream edges of the computational domain. Copyright © 2002 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved. Lengths must be chosen upstream and downstream of the optics corresponding to the neutralization lengths of the domain. If these distances are too short, then results from the simulation are unreliable. If the distances are too long, the simulation takes longer to run than necessary. The neutralization lengths are far from constant, as they vary with any change in the domain state variables such as beamlet current or discharge potential. A study is therefore undertaken to find a general relation involving some of the domain variables, so that any given simulation can be run with appropriate neutralization lengths. Aperture widening caused by erosion of the optics is the primary mode of failure in an ion thruster. An accurate method for predicting erosion is therefore necessary to the modeling of ion thrusters. Some progress has been made in this areathe model in Ref. 1 predicts erosion on the downstream face of the accel grid very well compared to experimental data, and Ref. 2 compares very well to experimental measurement of erosion of the accel grid barrel. However, neither model has incorporated simulation of the sputtered optics material, although it has the potential to be important to the operation of the thruster. Sputtered molybdenum which flows upstream of the optics coats the walls of the discharge chamber, and material which flows downstream can contaminate the spacecraft. If the molybdenum deposits on the optics it is possible that the total amount of erosion is reduced, but this behavior is very difficult to detect experimentally. Simulation of sputtered molybdenum is added to the model to address these concerns. Presented in the following is the method used to determine the neutralization lengths as a function of input beamlet current, discharge potential, and accel grid voltage. Comparisons are given between the computed data, derived relations, and theoretical results. Data is given for the density and velocity of the simulated molybdenum. Also computed are the proportions in which the molybdenum flows upstream, flows downstream, and deposits on the various parts of the optics. Model Operation The computational code is a 2-D axisymmetric simulation of a single aperture in an ion thruster. The grid used is composed of evenly spaced rectangular