Abstract

The electron-optical streak and framing cameras which have been used to date for monitoring ultrafast optical processes emitting in the UV-NIR spectral region have incorporated image tubes equipped with conventional semi-transparent, positive electron affinity photocathodes (usually types S1, S11, S20 and S25). Electron emission from such photocathodes consists of only those "hot" electrons that possess sufficient energy to traverse the potential barrier encountered at the cathode/vacuum interface. This in turn implies that any photoelectrons emitted are likely to have been excited very close to the vacuum surface and the photoemission process is therefore expected to be very rapid (probably on a timescale ⪝10^–13 s). In contrast, a heavily p-type doped semitransparent semiconductor layer can be activated in such a way that conduction band electrons have no surface potential barrier to overcome in order to reach the vacuum. For these so-called negative electron affinity (NEA) cathodes, photoelectrons which have been excited into the conduction band and subsequently scattered down into metastable conduction band minima can still be emitted. The associated long thermal electron diffusion lengths attainable in such materials means that the photoemission from NEA surfaces is dominated by thermalised electrons excited comparatively deep within the bulk. Theoretical considerations suggested that although NEA transmission photocathodes have excellent sensitivities as high as 1500μA/lumen (1), their response times (estimated to be ~1ns (2)) were expected to be substantially greater than their positive electron affinity counterparts but no quantitative experimental results on this topic were available. We therefore devised an experimental arrangement involving a UHV-compatible picosecond streak camera by which the temporal response of NEA GaAs transmission-mode photocathodes could be directly measured.

© 1984 Optical Society of America