Three independent spacecraft teams reported the presence of infrared absorptions consistent with molecular water (10-1000 ppm) and hydroxyl (OH) on the sunlit lunar surface at latitudes too warm for water-ice to be stable [1-3]; a possible mechanism for retaining molecular water at these latitudes is molecular chemisorption. The LCROSS impact experiment at Cabeus crater showed the presence of up to 6% concentration (by mass) of water and ice in the ejecta plume . Understanding where the water came from and where else it is expected to be trapped requires sophisticated computer modeling. Monte Carlo models  track water molecules as they migrate about the lunar surface, however, existing models neglect molecular chemisorption and may be significantly underestimating the time a migrating water molecule spends on the surface between each migration step. Similarly, the accuracy of models examining migration of water vertically within the regolith would be improved by including molecular chemisorption.
Temperature program desorption (TPD) measurements of water adsorbed on lunar materials provide values for the energy required to desorb chemisorbed water molecules from the regolith surface, which can be related back to thermal stability and minimum water lifetimes for a given temperature profile. TPD experiments were performed under Ultra-High Vacuum (UHV) conditions. The general procedure involves mounting the sample in UHV, cooling to the desired dosing temperature, dosing the adsorbate, and heating the sample at a constant temperature ramp while recording the rate of desorption versus temperature .
Proof-of-concept TPD experiments were conducted with the mineral albite and the lunar simulant JSC-1A . While these simulants are mineralogically similar to lunar samples, the space-weathering environment of the lunar surface leads to significantly modified surface structure compared to terrestrial materials. Thus, an examination of five lunar soil samples is underway to better address the thermal stability of water on lunar materials. The lunar samples are low and high-Ti mare soils, KREEPy soil, highland soil, and a mixed lot. Preliminary data from the low-Ti mare soil show molecular chemisorption on the order of a few 10’s of ppm by mass, depending on the assumptions used in the calculation. While this is higher than was reported for the simulants, the highest binding energy sites are weaker than those seen on the simulants, resulting in lowered ability of the regolith to retain water through the lunar noon and a net decrease in persisting water molecules compared to the simulants. TPD experiments on the lunar samples are currently underway and the latest results and interpretations will be presented at the Forum.
 Clark, R. N. (2009) Science, 326, 562.
 Pieters, C. M., et al. (2009) Science, 326, 568.
 Sunshine, J. M., et al. (2009) Science, 326, 565.
 Colaprete, A., et al. (2010) Science, 330, 463.
 Crider, D. H. and R. R. Vondrak (2000) JGR, 105 (E11), 26773.
 Poston, M. J., et al. (2013) JGR, 118(105), doi: 10.1002/jgre.20025.