THE IMPACT OF LUNAR PROSPECTOR IN A LUNAR POLAR CRATER DID NOT DETECT WATER ICE

UT ENGINEERING NEWS
For more information:
Becky Rische, UT-Austin (512) 471-7272

Embargoed until Oct. 13, 1999

Detailed Discussion

Researchers, led by a team of engineers and astronomers at The University of Texas at Austin, who have been examining the results of the crash of the Lunar Prospector spacecraft on the Moon are now reporting their results. The results are being presented today at the annual meeting of the American Astronomical Society - Division for Planetary Sciences meeting in Padua, Italy. There is first a brief review of the particular end of mission impact concept followed by a discussion of the results.

The Lunar Prospector $63million mission was designed to search the surface of the Moon for specific minerals as well as hydrogen in the lunar soil (the regolith). The hydrogen detected by LP near the lunar poles, in particular, was the best indication ever found for the presence of water ice in permanently shadowed craters near the lunar poles. Such ice, if it existed probably arrived during impacts of comets over the aeons and was possibly stable in cold craters over the age of the solar system.

The spacecraft was in a polar orbit - passing over both of the Moon's poles on each two hour orbit - and it's extended mission was scheduled to end on July 31, 1999. The mission had been a great success, providing ten times more data than anticipated at very low cost. When the mission ended, the spacecraft was originally simply going to crash into the lunar surface at an unknown location when the orbit became sufficiently perturbed. Such perturbations arise mainly due to the gravitational pull of the Earth and Sun and to uneven mass distributions within the Moon itself.

At the end of last year, a team of engineers and astronomers at The University of Texas at Austin suggested that rather than let LP crash randomly, it's final descent should be directed into one of these very cold craters. The energy of the impact was suggested as being possibly sufficient to vaporize perhaps 18kg (40lbs) of water ice and produce a plume of vapor which could be detected from Earth. A spectral detection of water (H2O) or its byproduct OH (hydroxyl radical) shortly after the impact would be definite proof that there was water ice in the crater. However, a lack of such detection would have no meaning at all since there were many possible reasons for a lack of visible vapor (discussed below). It was clearly stated that the chances of a successful detection of water or OH were very small (perhaps only 1 in 10). However, it was felt that considering the negligible cost and the potentially huge payoff, the effort was worthwhile.

The LP team was approached with a written proposal in early Spring and responded favorably to the idea. By late Spring, NASA gave preliminary approval to end the LP mission with a directed impact. A press conference was held at the American Geophysical Union meeting in Boston in June to describe the plans as they were laid out in a short technical paper in "Geophysical Research Letters".

The impact was scheduled for 09:52 GMT (4:52am Texas time) on July 31, 1999 for the following reasons:

  1. All LP objectives would have been met by then.
  2. Funding for the LP mission ran out that day.
  3. No further extensions seemed feasible as the LP battery was weakening and the LP was nearly out of fuel.
  4. On July 31, LP passed directly over a large crater having a permanently shadowed floor and this crater appeared to have the greatest concentration of hydrogen in it.
  5. Directors Discretionary time on Hubble Space Telescope (HST) could be made available.
  6. The impact would take place during the night when the moon was visible over the US so many ground-based telescopes could make observations.
There was a great deal of public interest in the Final Impact related to the novelty of the idea as well as the possibility that something of the impact might be observed from the ground. (A large number of amateur astronomers participated and the Univ. of Texas web site was popular (www.ae.utexas.edu/~cfpl/lunar) with real-time scientific updates before and during the event).

Results

LP impacted the moon as planned at 09:52 GMT (4:52am Texas time) on July 31, 1999. The navigation team feels certain that the impact occurred in the crater expected. The spacecraft certainly did not overshoot the target by much or we would have re-established radio contact as it passed back to the near side of the moon. A figure illustrating the nominal trajectory as well as possible errors is shown.

Figure: Final trajectory of LP spacecraft over local topography. Estimated error band in trajectory is indicated. Note that the estimated error in the topography itself (by Margot et al) was 300m.

As seen in the figure, even with reasonable worst case scenarios, the LP almost certainly struck the surface within the crater.

The professional astronomers have now mostly completed their initial analysis of their telescopic data. Unfortunately, there has been no clear detection of any evidence of impact: no detection of water vapor, no OH (from HST, Keck 1, or McDonald), no enhanced sodium emission, no HCN, no dust, no C2, nothing.

A discussion of the OH observations by the University of Texas team follows:

The observations with HST were difficult in that pointing a moving telescope in Earth orbit at a moving target (the limb of the Moon) is hard. Some attempted spectral images were lost due to pointing errors but others were acquired successfully. In particular, we obtained a few good spectral images radially above the lunar limb. We were searching for the spectral signature of OH, the photo-dissociation by product of H2O. The OH molecule fluoresces in the ultraviolet in a band at 3085 angstroms. After standard CCD data reduction techniques and subtraction of the scattered moonlight (please see detailed description in the appendix), no emission lines in the 3085A band were detected. Upper limits, based on the noise levels in these spectra, indicate a brightness of about 5-30 Rayleighs, a level comparable to what we had hoped to have produced with the impact itself. That is, if LP had vaporized the assumed 18kg of water, and the vapor had behaved as predicted, we could have expected an OH brightness of 15-20Ra. The noise level indicates that we should have been able to detect an OH brightness if it had been in excess of 30Ra, and we didn't. Our observations, however, are by far the lowest noise level observations ever done in search for OH molecules near the moon. Such observations can thus provide useful constraints as to the density of water/OH vapor naturally occurring about the Moon due to the the continuous flux of meteorites or mini-comets. That is, we suggest that the natural atmosphere of the Moon had less than several tens of kilograms of water vapor at the time of our observations.

What does our lack of a positive detection of OH mean? What are the possible reasons that there was no detection of water?

  1. The approach angle was very shallow due to limited fuel constraints. LP might have missed the target, say hitting the crater near rim.
  2. We might have hit a rock or dry regolith at the target site.
  3. The water molecules may have been firmly bound up in rocks as hydrated minerals (e.g., concrete) as opposed to existing as free ice crystals. The impact had sufficient energy to have vaporized some ice crystals but certainly did not have enough energy to separate water from hydrated minerals.
  4. Perhaps there really was no water there in the crater after all and the hydrogen detection by LP was simply pure hydrogen somehow implanted by the solar wind.
  5. It may be that the impact studies and assumptions were inadequate, particularly the modeling related to a poor energy coupling to the soil caused by a ricocheting impact.
  6. The modeling parameters used in the plume development and exosphere models may have been inappropriate.
  7. There may have been unlucky pointing of the telescope detectors which have a very small field of view.
  8. There may have been an unfortunate asymmetric "jetting" of material so the plume didn't rise above the crater walls of rose away from the telescope field of view.
The point is that the controlled crash of NASA's Lunar Prospector mission into a crater near the south pole of the Moon in late July produced no observable dust plume or water vapor cloud, according to researchers from around the world who observed the crash on July 31. The lack of positive evidence from this ambitious but low probability attempt to confirm indirect evidence of ice from science instruments aboard the small NASA spacecraft as it orbited the Moon leaves open the question of whether ancient comets have left their mark in shadowed regions of Earth's closest planetary neighbor.

The experiment, though, was not a failure. We established a remarkable coordinated observing program among many observers, established useful upper limits on the Moon's natural atmosphere, and established a possible means of lunar prospecting via direct impacts. This last point, in particular, may be important if we can identify any other spacecraft whose useful life is over but which may have sufficient fuel and controllability to repeat the LP impact experiment. It is presently unclear if such spacecraft exist. But if we can find one, we'd like to do it again. Similar opportunities may also be provided by natural impact events associated with meteor showers such as the Leonids in November.

Appendix

We are looking for a characteristic signature of the OH molecules in the ultraviolet region of the spectrum. OH is the photo-dissociated byproduct of water vapor.

A spectrum is an array of the electromagnetic radiation arranged in order of wavelength. We are most familiar with the spectrum of visible light, which we see in rainbows, for example, but the visible spectrum (4000A-7000A) is merely one part of a much larger spectrum of light.

Instruments

Modern astronomers seldom look directly through a telescope. They utilize special instruments which are more sensitive than the human eye to analyze the light the telescope gathers. The most common imaging system now used in astronomy is a light sensitive silicon wafer called a CCD (Charge-Coupled Device). The CCD is used like a photographic plate, but has the advantage that it can detect both bright and faint objects in a single exposure, CCDs are much more sensitive than photographic plates, have a linear response to light energy, and their digital output is read directly into computer memory for later analysis. Also we need a spectrograph to separate the light gathered by a telescope according to wavelength to produce a spectrum. All modern spectrographs use reflection gratings to spread the light into its rainbow of colors. A grating is a piece of glass with thousands of microscopic parallel lines scribed onto its surface. Different wavelengths (colors) of light reflect from the grating at slightly different angles, so incoming light is spread into a spectrum.

Because we were observing just off the very bright limb of the moon, we had to take special precautions to reduce the scattered light in the telescope system. In particular, the mirrors on the 107'' and 82'' telescopes were cleaned shortly before the event.

The McDonald 82'' telescope was also used to image the event. Special narrow band-pass filters, prepared especially to observe OH emission associated with the recent Hale-Bopp comet apparition, were used. 100 second exposure images were taken of the south polar region shortly before the Lunar prospector struck the surface and were again taken shortly afterward. The idea was to subsequently subtract the images and look for evidence of an expanding vapor plume.

Figure: Image of the impact area taken on July 30 with the 82" telescope at McDonald Observatory. The image is approx. 180 arc seconds across and was taken in the OH. North is up. The impact site is marked with an arrow.

Data Reduction

Standard CCD Reduction: First we correct the CCD images with zero and flat-field frames. Zero frames are obtained through a series of short exposures with the camera shutter closed. We need zero frames to obtain the signal introduced by the readout electronics of each pixel and remove those values from raw images. Flat-field frames are obtained through exposures focused to a uniform background such as the telescope dome and/or the sky just before sunrise or after sunset. Flat field frames give us information about the relative sensitivity of each pixel. The flat field processing also removes the effects of dust or scratches in the optical train. There are other processes like bad pixel removal, correction for sky background, removal of pixels struck by cosmic rays, etc. After the image processing, a wavelength calibration is done. A raw spectral image does not have the information about wavelength, so we use a comparison lamp ( we used mercury light) to get an accurate wavelength fiducial for the raw spectrum. We took CCD frames of the mercury (Hg) spectrum during the observation to provide wavelength calibration. Also to get an absolute flux level, we need images of a standard star for which we already know the flux level. After we finish the wavelength and flux calibration, each spectrum is extracted at different heights above the lunar surface. The ratio of the signal level to the random noise level was small so several spectra were obtained at adjacent altitudes above the surface and then combined to reduce the noise level.

Figure: Lunar disk spectrum scaled to the level of the continuum at 3105A and then subtracted from each offset spectrum produces the residual spectrum seen in the lower part of the figure. Note the absence of spectral lines at the indicated locations within the OH band.

We next subtract a solar spectrum from each observation to get the pure signal of OH. The solar spectra used were either standard spectra measured by various satellites or were taken as scattered light off the lunar surface itself in regions well away form the impact site. Even though we were searching for the OH signal above the lunar limb (against the black background of space), atmospheric scattering or stray reflections within the telescopes themselves scatter light from the bright lunar disk onto our CCD array. HST is obviously not subject to atmospheric scattering and therefore has the least noise and should be best at discriminating between any OH signal and scattered solar light. Observing conditions at Keck and McDonald were, however, very good at the time of the impact.