While a photoelectric cell constructed in this way gives a quali tative measure of the light entering it, it is certainly somewhat insensitive, and great improvements have been effected by at tention to two points. The first is the correct preparation of the light-sensitive surface—potassium in the case already described. It appears that the sensitivity can be greatly improved by the following process. The potassium is carefully purified by dis tillation, and is finally deposited by the same process inside the cell. A little oxygen is then admitted, and a glow discharge passed for a few moments. The oxygen is then pumped out and the potassium is left in a very sensitive colloidal state. The second improvement is to admit a small amount of helium into the cell instead of leaving it completely evacuated. The electrons leaving the potassium surface and accelerated by the applied potential now cause ionization by collision, and a greatly increased current is obtained. The one objection is that the current now depends on the applied potential, as can be seen from fig. 5. It is usually safe to apply a potential of zoo volts, and at this point a change of 1 volt produces a change of about 1% in the current.
A matter of great practical im portance in the use of photo electric cells is that they are not equally sensitive to all wave lengths. This can be seen from fig. 6, which shows in arbitrary units the photoelectric effects in potassium, rubidium and caesium cells for equal amounts of inci dent energy of different wave lengths. The vertical scale is arbitrary in each case, and the curves do not show the relative sensitivities of the three types of cell. The ratio of the total sensitivities of potassium, rubidium
and caesium cells as here constructed is about as 4:2:1, but the relative sensitivity for any particular colour depends on the wave length; e.g., for light of 6,000 A.U., the order would be reversed, the caesium cell being most sensitive. A sodium cell can often be used with advantage for light of a shorter wave-length since its maximum occurs at about 3,400 A.U., For ultra-violet work the en velope is usually of quartz, as glass begins to absorb strongly. It will be realized from these examples that a photoelectric cell can not be used for comparing directly the intensities of two beams of light unless they are of the same colour, or, more accurately, of the same spectral distribution.
The uses of photoelectric cells are so manifold that a catalogue of them would be monotonous. Briefly they replace the human eye whenever quantitative measurements are needed; e.g., they are frequently used to measure or compare the light arriving at the earth from stars; also, many photometers (see PHOTOMETRY) depend on their use. The density of a photographic plate may be found either by measuring the diminution in intensity of a beam of light passing through it, or alternatively may be put equal to the density of some standard plate when each produce the same weakening of the beam. In either case the essential is to have a quantitative measure of the intensity of a light beam, and for this purpose photoelectric cells are usually used.