Systematic optimization of Cs-Sb-O activated NEA GaAs photocathodes for long-lived accelerator electron sources

• Systematic optimization of Cs-Sb-O activation on p-doped GaAs (100) was performed for temperature and Sb deposition parameters. • Optimal Sb thickness of 0.3-0.5 nm yielded QE >2% at 780 nm and a 15-fold longer dark lifetime than conventional Cs-O activation. • Excess Sb (>0.6 nm) suppressed QE due to increased electron scattering in the NEA activation layer. • Temperature dependence differed between Sb-CsO-GaAs and CsO-Sb-GaAs, reflecting stability of Cs-Sb surface phases. Negative Electron Affinity (NEA)-activated GaAs-based superlattice photocathodes remain the only source capable of generating spin-polarized electron beams with polarization exceeding 90%, making them indispensable for applications such as particle accelerators and spin-resolved spectroscopy. However, conventional Cs-O activation suffers from chemical instability, resulting in rapid degradation of quantum efficiency (QE) due to gas adsorption and ion back bombardment (IBB). To maintain high performance, operation under ultra-high vacuum conditions (<10 −9 Pa) is required; however, the photocathode’s lifetime remains short. In our system, for example, multialkali photocathodes (e.g., K 2 CsSb) typically exhibit an operational lifetime of approximately 4000 h, while NEA-GaAs photocathodes degrade significantly faster, with a lifetime of only about 20 h. To overcome these limitations, NEA activation using heterojunctions with alkali metals and antimony (Sb) or tellurium (Te) has been proposed. In particular, a Cs-Sb-O activation scheme shows promise in improving both QE and lifetime. In this study, we systematically investigate the effects of substrate temperature, Sb thickness, and timing of Sb deposition on performance. We find that a 0.5 nm Sb layer yields a QE of approximately 2% at 780 nm-comparable to Cs-O activation-while extending the dark lifetime by a factor of 10. These results demonstrate the viability of the Cs-Sb-O scheme for stable, long-lived spin-polarized electron sources. These findings provide practical guidelines for the development of robust electron sources in accelerators.