Indium phosphide negative electron affinity photocathodes: Surface cleaning and activation

InP(100) is a very important semi-conductor for many applications. When activated by Cs and oxygen, the InP surface achieves the state of Negative Electron Affinity (NEA) making the Cs+O/InP system a very efficient electron source. Despite many years of study, the chemical cleaning and activation of InP are still not well understood. In our work, we have established an understanding of the basic physics and chemistry for the chemical cleaning and activation of the InP(100) surface. Synchrotron Radiation Photoelectron Spectroscopy is the main technique used in this study because of its high surface sensitivity and ability to identify chemical species present on the surface at each stage of our process.; A clean, stoichiometric InP(100) surface is crucial for obtaining high performance of NEA photocathodes. Therefore, the first part of our study focused on the chemical cleaning of InP(100). We found that hydrogen peroxide based solutions alone, originally developed to clean GaAs(100) surfaces and widely used for InP(100), do not result in clean InP(I00) surfaces because oxide is left on the surface. A second cleaning step, which uses acid solutions like HCl or H₂SO₄, can remove all the oxide and leave a 0.4 ML protective layer of elemental phosphorous on the surface. The elemental phosphorous can be removed by annealing at 330°C and a clean InP(100) surface can be obtained. Cs deposition on InP(100) surface shows clear charge transfer from the Cs ad-atoms to the substrate. When the Cs/InP(100) surface is dosed with oxygen, the charge transfer from the Cs to substrate is reduced and substrate is oxidized. The activation of InP as a NEA photocathode is carried out by an alternating series of steps consisting of Cs deposition and Cs+O co-deposition. Two types of oxygen are found after activation. The first is dissociated oxygen and the other is a di-oxygen species (peroxide or superoxide). The decay of quantum-yield with time and with annealing is studied and changes in both the Cs oxide and oxidation of substrate are observed, which we conclude are the causes for the decrease of the quantum yield.