Researching this answer led to ³He in permanently shadowed lunar polar surfaces published in Icarus. The abstract is tantalizing but terse:
Abstract
Because of their cryogenic temperatures, analysis indicates that permanently shadowed polar lunar craters may have substantially higher levels of ³He than sunlit lunar surfaces and are conservatively estimated to contain as much as 50 ppb or more.
Wikipedia's Helium-3; Solar nebula (primordial) abundance says:
One early estimate of the primordial ratio of ³He to ⁴He in the solar nebula has been the measurement of their ratio in the atmosphere of Jupiter, measured by the mass spectrometer of the Galileo atmospheric entry probe. This ratio is about 1:10,000, or 100 parts of ³He per million parts of ⁴He. This is roughly the same ratio of the isotopes as in lunar regolith, which contains 28 ppm helium-4 and 2.8 ppb helium-3 (which is at the lower end of actual sample measurements, which vary from about 1.4 to 15 ppb).
Question: Why exactly would "permanently shadowed polar lunar craters... have substantially higher levels of ³He than sunlit lunar surfaces?" What is it exactly about permanently shadowing surfaces from the Sun that is thought to allow them to accumulate up to 50 ppb of helium-3 compared to a lunar average of only 2.8 ppb?
Is it the far lower temperature, or the shielding from the solar wind, or something else?
Possibly relevant factoid, the boiling point of helium-3 is only about 3.2 Kelvin, much lower than the 4.2 Kelvin of helium-4.