The apparatus used by Kazda and Dunn for the determination of the critical photoelectric potential of mercury has been completely reconstructed. The stopcocks, the grease of which was the source of contaminating vapors, have been replaced by mercury cut-offs and the high temperature cement used in the construction of the photo-electric cell has been replaced by a graduated quartz-Pyrex seal. Fresh clean mercury was returned to the still. With this new apparatus and under these new working conditions the long wave-length limit for running mercury has been found to be 2735A. The conditions within the apparatus have been found to be practically free from contaminating vapors so that the critical frequency has been found for stationary mercury, namely 2735±10A, which is in entire agreement with Kazda for flowing mercury. This impurity in the old apparatus caused the photo-current to increase four-fold its original value in thirteen to twenty minutes after the still was turned off. It then slowly receded, falling below its initial value in three or four days. This increase in the photo-current was accompanied by a rise in the threshold to 2850A, thence falling in time to a constant value 2680A. In the apparatus as it now stands this same four-fold increase is not reached until some eighty hours after the still is turned off and remains perfectly constant indefinitely thereafter. The long wave-length limit for this maximum sensitivity was found to be 2910A.

Hydrogen, helium, argon, nitrogen, and water vapor in extremely small quantities in contact with the surface and also, except in the case of water vapor, dissolved in the body of the mercury, had no influence whatever upon the photoelectric behavior of the mercury or upon the rate of rise of sensitivity upon turning off the still except that each had a marked cleansing effect in reducing the concentration of the impurity which slowly contaminates the surface. Oxygen had a decided reducing effect upon the threshold value, bringing it down to 2555A in eighteen hours.

• Received 19 July 1928

DOI:https://doi.org/10.1103/PhysRev.32.950