Technical Comments

Technical Comment on “Hydrogen Mapping of the Lunar South Pole Using the LRO Neutron Detector Experiment LEND”

See allHide authors and affiliations

Science  25 Nov 2011:
Vol. 334, Issue 6059, pp. 1058
DOI: 10.1126/science.1203341


Based on a study of high-energy epithermal (HEE) neutrons in data from the Lunar Exploration Neutron Detector (LEND) on NASA’s Lunar Reconnaissance Orbiter (LRO), the background from HEE neutrons is larger than initially estimated. Claims by Mitrofanov et al. (Reports, 22 October 2010, p. 483) of enhanced hydrogen abundance in sunlit portions of the lunar south pole and quantitative hydrogen concentration values in south pole permanently shaded regions are therefore insufficiently supported.

The paper by Mitrofanov et al. (1) provides new interpretations about the distribution of lunar polar hydrogen (H) abundances using data from the Lunar Exploration Neutron Detector (LEND) on NASA’s Lunar Reconnaissance Orbiter (LRO). Based on the presence of a significant uncollimated background from high-energy epithermal (HEE) neutrons and consistent observations of the global neutron emission as measured using the neutron spectrometers on Lunar Prospector (LP) and LRO missions, we conclude that claims of enhanced H abundance in sunlit portions of the lunar south pole and quantitative H concentration values in south pole permanently shaded regions (PSR) are insufficiently supported.

Standard neutron transport physics (2) shows that the inner 10B shield of the LEND collimator will become transparent to neutrons with sufficiently high energy, where the energy threshold for transparency is defined as Etrans (3). Low-energy epithermal (LEE) neutrons have energies less than Etrans and are efficiently absorbed by the 10B collimator. Because the LEND 3He sensors detect the presence of a neutron but not its incident angle or energy, HEE neutrons are indistinguishable from LEE neutrons. The spatial distribution and magnitude of the HEE background are driven by instrument parameters, such as Etrans, which has not been published for the LEND instrument. We have consequently estimated Etrans using instrument specifications (4) and the particle transport code MCNPX (5). We find Etrans to be in the range of 1 to 10 keV (6). We obtain qualitative information about the HEE neutron background by studying its compositional dependencies, and we estimate its magnitude using measured LEND data.

LEE neutrons, which dominate the neutron count rate in the uncollimated LEND epithermal sensor and uncollimated neutron spectrometer on NASA’s LP mission (8, 9), show a strong anticorrelation with the abundance of neutron absorbing elements Fe, Ti, Gd, and Sm for equatorial soils with minimal H. This anticorrelation is clearly observed in both measured data (8) and simulations of neutron count rate from lunar regolith (Fig. 1A). In contrast, HEE neutrons follow a positive correlation with average atomic mass (Fig. 1B), which is similar to the positive correlation between LP-measured fast neutrons and the average atomic mass of the lunar regolith (8, 10). Figure 1B shows a 10% variation of HEE emission between mare and highlands soils (11). For soils enriched with H, as observed in the lunar polar regions (poleward of ~70° latitude), HEE neutron emission is almost identical to LEE neutron emission (Fig. 1C) and exhibits a strong dependence on H abundance. Thus, the relative magnitude of the uncollimated LEND HEE neutron background should correlate with fast neutrons in nonpolar regions and correlate with LEE neutrons in H-enriched polar regions.

Fig. 1

(A) Relative count-rate dependence of LEE neutrons versus Fe concentration for lunar returned sample soils with no H [adapted from (9)] as modeled for LEND-type 3He neutron sensors [5 cm diameter, 7.7 cm length, 20 atmosphere pressure (4)]. The strong similarity between these results and those for LP-type 3He sensors [5.7 cm diameter, 20 cm length, 10 atmosphere pressure (9)], indicates that the observed and validated compositional dependencies are relatively insensitive to specific 3He sensor parameters. Labels on data points specify individual soil types: A11, Apollo 11; L16, Luna 16; FAN, ferroan anorthosite. (B) Relative count-rate dependence of HEE neutrons (Etrans = 1 keV) versus average atomic mass for returned sample soils, as modeled for LEND-type 3He neutron sensors. (C) Relative count-rate dependence of HEE neutrons versus water-equivalent hydrogen abundance in a FAN soil for LEND-type 3He neutron sensors. Data points show the HEE neutron calculation; the solid line shows the analytic LEE neutron dependence taken from (9).

Figure 2 shows a summed count-rate map of all LEND collimated sensors with data from the NASA Planetary Data System (PDS). The enhancements of neutron flux in the nearside maria (90°W to 45°E, 45°S to 45°N) are spatially correlated with LP measurements of fast neutrons (8, 10), which suggests that background HEE neutrons are responsible for this feature. The difference between the maximum maria count rate and mean highlands count rate is ~0.25 counts per second (cps) (5.0 to 4.75). If this count-rate difference represents the ~10% HEE dynamic range shown in Fig. 1B, then the highlands HEE neutron count rate is 2.25 cps (12). This estimated uncollimated HEE neutron count rate is almost an order of magnitude larger than the 0.3 cps estimated in (1) for the nonspacecraft, uncollimated background; is similar in magnitude to the independent background of 2.8 cps induced by cosmic ray interactions with the spacecraft (1); and is a significant portion of the total count rate of 5 cps (1). The sum of the two independent backgrounds imply that the collimated count rate is significantly smaller than the 1.9 cps stated by (1) and may be closer to the value of <0.2 cps estimated by (2).

Fig. 2

Map of LEND collimated neutron count rate from the first 456 days of science mapping, summed for all LEND collimated sensors and given in units of cps scaled to the nominal 5 cps of (1). Data have been smoothed with a kappa function with width 36 km using the algorithm of (14). The contours outline the fast neutron data from (8) and are spatially correlated with the LEND nearside count-rate enhancements.

In the polar regions, the LEND count-rate decrease is well correlated with a similar map of LP epithermal neutron count rates (8, 9), which is the characteristic signature of enhanced H abundance. Based on the compositional dependency of HEE neutron emission and the likely observation of a significant HEE background in the nonpolar LEND data, the polar regions may also have a significant, uncollimated HEE background that varies with H abundance. The presence of an H-dependent, uncollimated background from HEE neutrons, if sufficiently large compared with the collimated signal, results in an effective LEND spatial resolution approaching that of the uncollimated LP measurements (i.e., ~50 to 75 km rather than ~10 to 20 km based on the physical geometry of the LEND collimator). With such a broad spatial resolution, two central results from (1) are not supported. First, because adjacent sunlit regions and PSRs cannot be resolved with a spatial resolution of >30 km (1), the conclusion that broad expanses of H-enriched soil are located in both sunlit regions and PSRs is not justified. Second, because a significant fraction of all counts are detected from regions outside of the geometrical field of view (FOV) of the collimator, H concentrations cannot be derived for specific south pole PSRs because measured neutrons from these regions are contaminated by neutrons originating from nearby sunlit regions.

Fully characterizing and quantifying the uncollimated HEE background beyond the estimates given here requires detailed modeling and documentation of the LEND energy and angle response. Analyses of Earth-to-Moon cruise and global elliptical orbit data (13), which have large altitude variations, can provide independent information about the background signals. This information can be obtained because each of the signal components (FOV foreground and various backgrounds) depend differently on spacecraft altitude. Such altitude dependencies should be identifiable within these data, which can enable the magnitude of each component to be independently quantified.

References and Notes

  1. In practice, Etrans is not an abrupt energy transition between LEE and HEE neutrons and depends on the neutron energy spectrum, collimator material, collimator geometry, and energy moderation of high-energy neutrons from the polyethylene portion of the LEND collimator. Nevertheless, we approximate it as a single energy, and the results presented here do not strongly depend on its specific value. See (11).
  2. An alternate way to determine Etrans uses instrument parameters and nuclear cross-section data. The neutron-absorbing portion of the LEND collimator contains 97% enriched 10B with a density of 1 g/cm3 (3). Based on 10B cross sections (7), the mean free path (MFP) (which represents the approximate maximum distance for full neutron absorption) for 1 to 10 keV neutrons is 0.8 to 2.3 cm; the MFP for 100 keV neutrons is ~4 cm. The thickness range of the LEND collimator is 0.4 to 2.4 cm (4). Thus, 1 to 10 keV is a reasonable range for Etrans.
  3. For Etrans = 1 keV, the count-rate variation is ~10%, as seen in Fig. 1B, and is roughly the same (10.5%) for Etrans = 10 keV. For Etrans = 100 keV, the magnitude increases to ~16%. The linear relation of HEE neutron count rate versus average atomic mass is preserved for all Etrans values. The HEE count rate behavior with H is almost identical for the range Etrans = 1 to 10 keV. For Etrans = 100 keV, the neutron variation with H is somewhat muted compared with the lower Etrans values.
  4. Based on Fig. 1B, the maria/highlands count-rate difference, ΔC = Cmaria – Chigh, will be ~10% of the total mare count rate, Cmaria, if the measured variation is due to HEE neutrons. Since ΔC = 0.25 cps, then Cmaria = 2.5 cps and Chigh = 2.25 cps. These values hold for any Etrans from 1 to 10 keV. For Etrans = 100 keV, which is likely higher than the effective value for LEND, the maria/highlands fractional variation is 0.16, which implies Chigh ~ 1.3 cps. This is still a factor of 4 larger than the 0.3 cps uncollimated background given by (1).
  5. Earth-to-Moon cruise data and early mission elliptical orbit data are not currently available in the NASA PDS but were used by (1) to make background estimates. Due to their importance in understanding the LEND background, it is hoped that these data will soon be made available to the scientific community.
  6. Acknowledgment:: Work at the Johns Hopkins University Applied Physics Laboratory was supported by the NASA Lunar Science Institute. Work at Los Alamos was performed under the auspices of the U.S. Department of Energy.
View Abstract

Navigate This Article