The Return Flux Experiment:

A Study of Spacecraft Self-Contamination

Heidi L. K. Manning, Nathan Frank, and Jason Bursack

Abstract

Self-contamination of a spacecraft is a concern facing instrument and spacecraft designers. While on the Earth, gases absorb onto spacecraft surfaces and are later released when placed in the vacuum of space. Models predicting the amount of gas released by a spacecraft that is returned to itself do exist, but these models have had very limited experimental testing. We describe an experiment designed to provide a test of these models and the analysis of the limited data obtained by that experiment.

Introduction

An interesting problem arises when spacecraft such as rockets, satellites and the Space Shuttle are launched into the vacuum of space. These objects emit gases into the surrounding atmosphere during the first few days and weeks after launch. These gases may come from vents, exhaust, cabin leakage, or may desorb from spacecraft surfaces. Contamination problems develop when released gases collide with ambient gases in the upper atmosphere and reflect back towards the spacecraft. These gases can condense or become absorbed onto and damage critical surfaces such as delicate optical instruments, pressure sensitive detectors, or equipment sensitive to gas contamination.

The rate at which the emitted gases are returned to the spacecraft is known as the return flux. Theoretical models predicting the return flux do exist ^1,2,3, however their power of predicting contamination is limited by large uncertainties in the models. Consequently, experiments potentially affected by contamination are often conservatively designed, and valuable time and resources may be unnecessarily wasted. The models have had only limited experimental testing^4, and thus a more extensive evaluation is desirable to improve their accuracy for spacecraft applications.

The only possible way to perform experimental tests of the models is to conduct them in the upper atmosphere. It is very difficult to simulate in the laboratory the conditions experienced while in Earth’s orbit because of the high speeds at which space vehicles encounter ambient gases. Moreover, the presence of atomic oxygen as the major constituent and the rarefied gas environment in the upper atmosphere make tests on the ground impossible.

An experimental test of the return flux models was designed and the experiment flew on a sub-satellite of the Space Shuttle in January 1996. The experiment was called REFLEX, an acronym for return flux experiment. During this mission, a gas mixture of argon and krypton was released into the atmosphere from the satellite. A newly designed mass spectrometer on-board the satellite measured the amount and energy distribution of the returning gas that had undergone collisions and was reflected back towards the satellite. With detailed information about the energies of the returning gases, better tests of the models can be made. In addition to studying the return flux gases, the new mass spectrometer also offered the capability to directly measure the ambient atmosphere.

Experiment Overview

The University of Minnesota developed and calibrated a new mass spectrometer specifically for an in-situ test of the return flux models that incorporates an energy analyzer with a traditional magnetic mass spectrometer⁵. This newly designed instrument determines the composition of a gas sample based on the mass of the molecules and measures the energy of the molecules present. A sketch of the major components of the mass spectrometer is shown in figure 1. The REFLEX mission was carried on a free-flying, sub-satellite of the space shuttle for 40 hours, during which time the experiment was conducted. The spacecraft flew in a circular orbit with a 30-degree inclination. Figure 2 illustrates the path of the spacecraft. A gas mixture of argon and krypton was released from a nozzle located one meter from the mass spectrometer. The gas flow from the nozzle was released at a high and a low flow rate, as well as turned off completely. Also, the experiment was conducted with three different spacecraft orientations with respect to its velocity vector.

Argon and krypton were selected as the released gas for several reasons. Argon and krypton are not virtually non-existent in the Earth’s upper atmosphere. Atoms, not molecules, were chosen so that the collisions could be approximated as hard spheres and energy from the collision was not taken up in vibrational states of a molecule. Also, argon and krypton are noble gases and thus, are not very chemically reactive which simplifies the data analysis. Two different gases were selected to have two data points at two different masses.

Analysis

The analysis of the data obtained from this mission involved several different procedures. Some of the analysis involved monitoring the instrument status while other analysis involved the measurements of the ambient atmosphere and the return flux test results. A preliminary study of the data involved checking the functioning of the instrument’s detectors. Periodic checks of the detectors were conducted during the mission to ensure they were operating at full potential. These check were necessary because with extremely high-count rates the detectors will begin to deteriorate and would require higher voltages to be applied in order to operate properly⁶. The test of the detectors involved varying the applied voltage and monitoring the detector output. Figure 3 is a graph of the detector signal versus applied detector voltage. The first data point is the nominal applied voltage on the detector prior to the test. Then the applied voltage is reduced and the signal drops dramatically as expected. As the applied voltage is increased, the signal strength increases until it reaches a maximum value and levels off. At this point, increasing the voltage does not increase the signal. The proper operating voltage of a detector is one that is on the plateau of the curve; here, the maximum possible signal is achieved. Figure 3 indicates that the nominal applied voltage on the detector prior to the test (the first data point) produces a signal strength that is on the plateau of the curve. Thus, we can conclude the detectors were operating within their specifications. During the mission 20 tests of the three different detectors were performed. The analysis of all tests indicates that all detectors were properly functioning throughout the entire mission.

The analysis of the data set also involved examining the calibration and stability of the instrument’s energy scale. However, because of several complications with this mission, very limited data was available, and the flight data obtained needed many corrections and careful inspection before any detailed analysis could be done. One of the problems that occurred during the mission was an unexpected drift in the zero of the energy scale. During the mission, data was taken in small steps across the energy peak. The center of this peak corresponds to the nominal ‘zero’ on the energy scale. Each energy scan performed during the mission was fit to a spline and the center of the peak was determined. Energy scans were performed for a variety of masses throughout the entire mission. A plot of the energy peak center for various masses versus mission elapsed time (MET) is presented in Figure 4. The graph clearly shows that the center for all masses drifts at about the same rate, and when a sudden abrupt change occurs in the energy center, all masses change their energy center simultaneously. From this, we conclude that the drift in the zero energy is due to electrical charging within the instrument, and the sudden changes in the peak center occur because of an electrostatic discharge (a spark). This conclusion points to ways in which the instrument could be re-designed for improved measurements on future flights.

Knowing that the zero point on the energy scale drifted during the mission, we were able to use this information to correct the data taken during the return flux test. To test the return flux models, an experiment was conducted that involved the spacecraft releasing argon and krypton. The instrument then measured the amount and energy of the argon and krypton returning to the spacecraft. Due to improper spacecraft orientation, very little high-energy argon and krypton data was obtained. To try to account for the low signals, the data was integrated for an entire orbit. Figures 5 and 6 were created from this integrated data. As indicated on the vertical axis, very low signals were obtained. It must be understood that these very low signals have a correspondingly high statistical uncertainty associated with them. Figures 5 and 6 show that some high-energy data was obtained, but peak shape and energy distribution is virtually impossible to determine from the data. Consequently, the direct testing of the return flux models was not possible.

Despite these complications, some information about the return flux can be gathered by carefully observing and applying data correction to low energy thermal gas signals. Figure 7 is a graph of the amount of released argon, ambient oxygen, and ambient nitrogen having thermal energies as measured by the instrument over time. As seen on the graph, the ambient oxygen and nitrogen vary throughout the mission. This oscillation is expected as the ambient atmospheric density increases with daylight and decreases at nighttime (Figure 8). Careful inspection of the data indicates that the returning thermal argon also shows an oscillation that correlates to the oscillations in the ambient atmospheric density. This observation indicates that the argon returning to the spacecraft is due primarily to collisions with the ambient atmosphere and not to collisions with other released argon atoms. This is an important result as it helps to distinguish somewhat between the different return flux models.

It can also be seen on figure 7 that the strength of the argon signal abruptly changes periodically. This is explained by figure 9 in which the vertical lines that overlay the data in figure 7 indicate a change in the gas flow rate. As mentioned earlier, the flow rate of the released gas varied from a high flow, a low flow and completely off. Figure 10 also is an overlay of vertical lines on the data from figure 7. In this graph the vertical lines indicate different spacecraft orientations. The spacecraft was supposed to be oriented directly into the ram (ram is defined as “head-on” with the instrument opening pointing parallel to the velocity vector), 45 degrees to the ram and 90 degrees to the ram. However, there were difficulties with the spacecraft orientation and these exact angles were never obtained. Spacecraft orientation angles of approximately –11 degrees, +25 degrees and +71 degrees were obtained. Figure 10 indicates that the amount of returning argon varied with the angle of attack. This additional information of the amount of returning gas, as attack angle varies, places additional restrictions on the return flux models.

A final analysis done with the data was to compare the ambient atmospheric density measured by the REFLEX instrument to the well-tested and very robust Mass Spectrometer and Incoherent Scatter Model (MSIS)⁷. This model was developed in 1980 and has been refined three times each time incorporating additional experimental data ^8,9,10,11,12. Figure 11 is a graph of the ambient oxygen densities measured with the REFLEX mass spectrometer and the oxygen densities predicted by the MSIS model. Both lines indicated the same period of oscillation. It is also seen that the density measured by the REFLEX mass spectrometer is significantly lower than the density predicted by the MSIS model. As previously mentioned, there were several problems that occurred during the mission. One of the problems occurred when an unexpectedly large amount of a contaminant from the spacecraft caused the initial in-flight calibrations to be incorrect. This improper calibration caused the density measurements made by the REFLEX instrument to be much lower than they actually were.

Conclusion

Some reasonable data was obtained from this experiment despite the many problems occurring with both the instrument and with the spacecraft. We determined the detectors remained fully operational during the entire mission. We also gained knowledge as to the cause of the drift in the zero of the energy scale. This information will assist in the re-design of this and similar instruments. In addition, we gained some limited information on the return flux models that spacecraft designers may use to improve and refine their spacecraft contamination models.

Acknowledgments

This work was funded in part by the Centennial Scholars Research at Concordia College. The authors would also like to thank Willis Wilson for his assistance in providing IDL code used in the analysis of this data and to Code 724 at NASA/GSFC for providing financial assistance in support of this work.

Literature Cited

¹Scialdone, J. J., Self-Contamination and Environment of an Orbiting Satellite, Jour. of Vac. Sci. and Tech., 9, 2, 1007, (1972).
²Robertson, S. J., Bhatnagar-Gross-Krook Model Solution for Back-Scattering of Outgas Flow from Spherical Spacecraft, 10th International Symposium on Rarified Gas Dynamics, Aspen, CO, 51, 1, 479, (1976).
³Naumann, R. J., Contamination Assessment and Control in Scientific Satellites, NASA TN D-7433, (1973).
⁴Scalidone, J. J., Hedin, A. E., and Rice, C. J., Comparison of Satellite Self-Contamination Experiments and Scattering Return Flux Calculations, Jour. Geophys. Res., 83, A1, 195, (1978).
⁵Mattauch, J. and Herzog, R., Z Phys., 89, 796, (1934).
⁶Potter, W. E., and Mauserberger, K., Spiral Electron Multiplier Operation Characteristics Using Positive Ions, Rev. of Sci. Instru., 43, 9, 1327, (1972).
⁷Hedin, A. E., Mayr, H. G., Reber, C. A., Spencer, N. W., and Carignan, G. R., Empirical Model of Global Thermospheric Temperature and Composition Based on Data from the OGO 6 Quadruple Mass Spectrometer, Jour. Geophys, Res., 79, 215, (1974).
⁸Hedin, A. E., Salah, J. E., Evan, J. V., Reber, C. A., Newton, G. P., Spencer, N. W., Kayser, D. C., Alcayde, D., Bauer, P., Cogger, L., and McClure, J. P., A Global Thermospheric Model Based on Mass Spectrometer and Incoherent Scatter Data MSIS 1. N2 Density and Temperature, Jour of Geophys. Res., 82, 16, 2139, (1983).
⁹Hedin, A. E., Reber, C. A., Spencer, N. W., Briton, H. C., and Kayser, D. C., Global Model of Longitude/UT Variations in Thermospheric Composition and Temperature Based on Mass Spectrometer Data, Jour. Geophys. Res., 84, A1, 1, (1979).
¹⁰Hedin, A. E., A Revised Thermospheric Model Based on Mass Spectrometer and Ion Scattering Data: MSIS-83, Jour. Geophys. Res., 88, A12, 10, 170, (1983).
¹¹Hedin, A. E., MSIS-86 Thermospheric Model, Rev., Jour. Geophys. Res., 92, A5, 4649, (1987).
¹²Hedin, A. E., and Carigan, G. R., Morphology of Thermospheric Composition Variations in the Quiet Polar Thermosphere from Dynamic Explorer Measurements, Jour. Geophys. Res., 90, 5269, (1985).