In principle any good moderator could be used, but it is desirable that the substance chosen should be capable of working at high temperature in order not to lose one of the principal advantages of sodium cooling. A perhaps unexpected difficulty arises from the fact that graphite is a little porous and is therefore capable of absorbing sodium into its structure. Such absorption, if it were allowed to take place, would have two effects; first the moderator would become impregnated with a quantity of neutron-absorbing material, with consequent serious damage to the neutron economy; and second, sodium penetrating between layers of graphite crystals could in time cause mechanical failure of the moderator. Reverting to the coolant, pure sodium does not corrode, and is not affected by, uranium and thorium or graphite, zirconium, and steel. This gives a reasonably wide range from which to choose materials of construction and makes it unlikely that there will be any violent chemical reactions.