A LOW temperature working space has normally to be protected from thermal radiation by plaqing a cold trap between it and the surroundings. The simplest trap consists of a bath of liquid refrigerant, but when space is very limited such as between the poles of a magnet, the trap can be a metal cylinder cooled from one end. This type of trap is known as a radiation shield, and in certain circumstances can also act as the enclosing wall of the working space. To minimize temperature gradients the shield can be made of a material of high thermal conductivity such as copper, and the surrounding space evacuated. Nevertheless, temperature gradients will exist, and the problem for the designer is to estimate whether the enclosure even approximates to an isothermal one.
Unfortunately it is extremely difficult to calculate the magnitude of such gradients, both because of the complexity of the theoretical analysis, and also because of great uncertainty in the values of the emissivities of the two facing surfaces. It therefore seemed worth while to measure the temperature gradient down a copper radiation shield held at 78 ยฐ K at its top end, and to compare the results with a rather crude 'first order' theory.