PDS_VERSION_ID = PDS3 RECORD_TYPE = STREAM OBJECT = TEXT INTERCHANGE_FORMAT = ASCII PUBLICATION_DATE = 2001-09-01 NOTE = "N/A" END_OBJECT = TEXT END XCALOVER.TXT Written as part of the NEAR XGRS submission to the Planetary Data System (PDS). INTRODUCTION The X-Ray Spectrometer (XRS) is one of several instruments onboard the NEAR spacecraft. The XRS measures characteristic X-ray emissions induced in the surface of the asteroid by the incident solar flux. The K-alpha lines for the elements Mg, Al, Si, S, Ca, Ti, and Fe are detected with spatial resolution on the order of 3 km when counting statistics are not a limiting factor. The X-ray spectrometer on the NEAR mission is a non-dispersive spectroscopic system. In this approach, the incoming X-ray photon is absorbed by the detector material and a signal prop- ortional to the absorbed energy is measured by the detector as a voltage pulse at the detector output. An analog to digital conversion is then performed and the count is binned by "pulse height" or energy loss and a spectrum is obtained and telemetered to Earth. From an analysis of the pulse height spectrum, elemental composition can be inferred. This document is an overview of the calibration tables provided to the Planetary Data Systems for the NEAR XRS. These tables (except 99) were originally published as figures in "Instrument Calibrations and Data Analysis Procedures for the NEAR X-Ray Spectrometer," by Starr et al. (2000), Icarus Volume 147, pp. 498-519. DETECTOR CALIBRATIONS In order to interpret the data collected by the XRS, a detailed understanding of the detector response is vital. Efficiency and energy resolution as a function of energy, angular response, filter transmission as a function of energy, and rise-time rejection efficiency are most important. Asteroid Pointing Detectors The gas proportional counters on NEAR have a chamber diameter of 42 mm and are filled with P-10 gas, a mixture of 90% argon and 10% methane, to an absolute pressure of 1200 mbar. The anode wire is gold-plated tungsten and has a diameter of 13 micrometers. Figure 1 displays the detector efficiency versus energy calculated for this type of detector. The sharp features in the response of the Mg and Al-filtered detectors at 1.30 and 1.56 keV, respect- ively, are the filter K-edges. The very sharp feature seen in all three detectors at 3.20 keV is the K- edge in argon, the primary constituent of the proportional counter fill gas. This feature complicates the response of the proportional counters for energies above the absorption edge, because the resulting K X-ray, which is 2.97 keV for argon, may escape the detector. A cor- responding escape peak will then appear in the spectrum that lies below the full-energy peak by an amount equal to this charac- teristic energy. This can be seen in Figure 2, which is the response of the Mg filtered proportional counter to an Fe-55 source which emits the Mn K-alpha and K-beta lines at 5.899 and 6.490 keV, respectively. The fit to the measured pulse height spectrum (solid line) requires four peaks (dashed lines); one for the K-alpha line, one for the K-beta line and one for each of the two escape lines. The full energy peak for the K-alpha line is seen in channel 150.9. The corresponding escape peak is in channel 75.7. From the energy calibration provided in the table in the upper left-hand corner of Figure 2, this implies an energy difference of 2.97 keV, as predicted. The energy resolution of gas proportional counters varies in- versely as the square root of the energy, G=2.35*SQRT[W(F+b)/E], where G is the full-width at half-maximum (FWHM) as a fraction of the X-ray energy, W is the energy required to create an ion pair, F is the Fano Factor, b is a parameter that characterizes the avalanche statistics, and E is the energy in keV. This equation represents the statistical limit for a gas proportional counter, which is the best response one can expect. For P-10 these parameters are, W=0.026 keV/ion pair, F=0.17, and b=0.50. The above equation then becomes: G=0.310*SQRT(E), At 5.9 keV this would imply an energy resolution of 12.8% or 0.753 keV. The energy resolution at 5.9 keV for the Mg filtered detector as shown in Figure 2 is 0.824 keV, or about 14.0%, which is typical for all the proportional counters on NEAR. An increase of about 10% over the statistical limit is expected for these detectors due to electronic noise and non-uniformity of the anode wire. The response of a proportional counter to six different X-ray lines is shown in Figure 3. The pulse height spectra were generated by using an alpha particle excitation source on targets of Mg (1.254-keV), Al (1.487-keV), Si (1.740-keV), S (2.307-keV), Ca (3.690-keV), and Ti (4.508-keV). The fractional energy resolution of each of these six X-ray lines, plus the 5.899-keV line from Fe-55 is plotted versus energy in Figure 4. The fitted line is a power-law function in energy with a slope of -0.5 and a constant of proportionality of 0.340. The energy resolution for Ca and Ti is 17.7% and 16.0%, respectively. The energy of the Fe K-alpha line that will be emitted from the asteroid surface is 6.40 keV and the resolution at this energy is 13.4%. While there will be some overlap of these lines in the spectra collected from the asteroid, the energy resolution of the proportional counters will still be sufficient to resolve them. A collimated Fe-55 source, positioned in a plane perpendicular to the detector face, was used to map the response of the detector over the entire window area. This result is shown in Figure 5. The FWHM of the detector response with the collimator in place is 3.2 degrees. The full-width at tenth-maximum (FWTM) is 5.5 degrees. The radial response of the detector is isotropic. As described above, it is necessary to use thin filters in front of two of the three proportional counters in order to separate the three low energy lines of Mg, Al, and Si. Figure 6 shows the response of the unfiltered, Mg-filtered and Al-filtered gas proportional counters to these three X-ray lines. The Mg filter significantly reduces the Al and Si line intensities relative to the Mg line. The Al filter reduces the Si line intensity relative to the others. Using this method, the three overlapping lines of Mg, Al, Si can be separated. The details of this technique for line separation are discussed later. The effectiveness of the rise-time discrimination circuitry is displayed in Figure 7. An in-flight spectrum taken with the Fe-55 calibration source in the detector field of view is shown. The rejection efficiency for gamma rays and charged particles was determined to be ~70% by comparing the background counts above channel 51 with and without rise-time discrimination enabled. A comparison of the counts in the Fe-55 peak shows that only about 5% of the Fe-55 X-rays are removed by the rise-time circuitry. Rise-time discrimination is disabled below about 2 keV (channel 51). Below this energy, the rise time of X-rays, gamma rays and cosmic rays is not sufficiently different to allow efficient rejection of the background without loss of a significant amount of valid X-ray events. Solar Monitors The NEAR XRS has two solar monitors. One is a solid state Si-PIN diode, and the other is a gas proportional counter identical to the three asteroid pointing detectors, but with a specially designed graded shield. The gas proportional counter is the primary solar monitor. The PIN detector was added as an engineering test of this new technology. For the PIN detector the efficiency is easily calculated. It is a Si detector 300 microns thick with a 76-micron thick Be window and a 200-nm dead layer. The result of this cal- culation is shown in Figure 8. The impact of the Be window on the efficiency at low energies is evident. The efficiency of the Si-PIN detector at 5.9 keV was measured using an Fe-55 source of known strength, placed on axis at a fixed distance from the detector. The result was 0.95 ± 0.04, which is in good agreement with the calculated value of 0.96. The error reflects the statistical uncertainties, only. Figure 9 is a pulse height spectrum of the Si-PIN with an Fe-55 source in the field of view. The K-alpha and K-beta lines merge, but are separated by the fit. The FWHM of the K-alpha line (5.9 keV) is 0.60 keV or 10.2%. Low-energy electronic noise is seen below channel 42 or about 1.7 keV. As with the proportional counters, the energy resolution of the Si-PIN varies inversely as the square root of the energy. Using the measured resolution of 0.60 keV at 5.9 keV to determine the proportionality constant gives the FWHM as a fraction of the X-ray energy: G=0.247*SQRT(E). The gas proportional counter solar monitor is identical to the three asteroid pointing detectors, but with the addition of a graded shield. This shield is designed to restrict the effective area of the detector to about 1 mm2, and also to attenuate more of the lower energy flux from the Sun in order to enhance the response of the detector at the higher energies. The response of the gas solar monitor at nine angles ranging from 0 to 40 degrees is shown in Figure 99. Gain Correction Before spectra can be summed, the energy scale of each spectrum must be determined and adjusted to a common scale. For the three asteroid pointing detectors this may be accomplished in two dif- ferent ways. First, as described above it is possible to rotate Fe-55 calibration sources into the field of view. Second, the asteroid pointing detectors will, of course, also observe lines of known energy from the surface of the asteroid. Where counting statistics make it possible, these lines too can be used for energy calibration. While it may be necessary to sum several hours of data to obtain well defined spectral features, this is still a useful exercise, because any significant gain changes during these times will be detected as line broadening. The solar monitors cannot be calibrated by either of these methods, and must therefore rely on solar line flux for energy information. While the gain of the proportional counters has been quite stable under most circumstances, gain changes have been observed during two different operational situations: following high voltage turn-on and after large solar particle events. These effects have been simulated in laboratory measurements and are believed to be due to space charging. Following application of high voltage to the proportional counters (typically about 1165 volts) the gain, as determined by measurements of the Fe-55 calibration sources, drops by about 4% and the normally gaussian shaped line peaks develop a low-energy tailing, as shown in Figure 12. The effect is at its worst within about 24 hours after high voltage is applied. The proportional counters take approximately four to six weeks to completely recover. Figure 12 shows the response of the Mg filtered proportional counter to the Fe-55 source over a six week period. The four plots represent a time progression following high voltage turn-on on 17 March 1999. The progressive recovery from the space charging effect can be seen. A similar effect was observed following a large solar particle event such as occurred in November 1997, but the time constants are somewhat different. The degradation develops much more slowly - taking about two weeks to reach its peak levels - and then recovers in about the same amount of time. The level of degradation never reaches that seen following high voltage turn-on.