Long lifetime polarized electron beam production from negative electron affinity GaAs activated with Sb-Cs-O: Trade-offs between efficiency, spin polarization, and lifetime

GaAs-based photocathodes are the state-of-the-art in the production of highly spin-polarized electron beams for accelerator and microscopy applications. While various novel structures of GaAs have been shown to increase the degree of polarization and quantum efficiency, all GaAs-based photocathodes require activation to negative electron affinity (NEA) to operate at the photon energies where the highest spin polarization is achieved. In this work, we report on NEA activation of bulk GaAs performed using Sb-Cs-O. We show this activation layer to be more robust with respect to traditional Cs and O layer in terms of charge extraction lifetime. This new activation layer is shown to improve the dark lifetime up to one order of magnitude and the charge extraction lifetime up to a factor of about 60, when compared with the traditional Cs-O activating layer. Trade-offs with other relevant parameters like quantum efficiency and electron spin polarization are also discussed.

Figure 1

Energy band diagram of GaAs with a ${\mathrm{Cs}}_3\mathrm{Sb}$ activation layer. When photon energies lower than 1.6 eV (about 780 nm) illuminate the sample, photoelectrons should be excited only from the GaAs conduction band.

Figure 2

Schematic of the cluster of vacuum chambers in the photocathode laboratory at Cornell University. Samples are transferred from one chamber to another in a vacuum level better than $1\times {10}^{-10}\mathrm{Torr}$. The experimental chambers used in this set of experiments are indicated with dotted red lines.

Figure 3

Typical growth parameters of a Cs-Sb-O coated GaAs sample. Left axis: Photocurrent; the shaded area represents the time interval where the Cs shutter was kept open, the striped area is when both Cs and Sb were codeposited. Right axis: Total pressure during deposition. The base pressure before opening the shutters and the oxygen valve was $2.2\times {10}^{-9}\mathrm{Torr}$.

Figure 4

Spectral response of the 5 samples as measured after growth. While NEA is present in all the samples, demonstrated by sensitivities extending well beyond the 875 nm (corresponding to the nominal value of the GaAs band gap of 1.42 eV), the overall efficiency decreases as the equivalent thickness of the Sb layer increases.

Figure 5

Measured electron spin polarization for the five samples as a function of the illuminating wavelength. The maximum achievable spin polarization (measured at about 780 nm) decreases as the equivalent thickness of the Sb increases beyond 0.12 nm.

Figure 6

QE as a function of the extracted charge as measured for the different samples at 505 nm. The data shows that as the thickness of the activating layer is increased, the initial QE value decreases but the lifetime is increased. The inset shows a plot of the lifetime improvement achieved with respect to the Cs-O activated sample for dark lifetime (${\tau }_{DL}$) and charge lifetime (${\tau }_Q$).

Figure 7

QEs measured at 780 nm before and after the lifetime measurements that are reported in Fig. 6. From the QE measurements at 780 nm we can estimate the lifetimes as function of the extracted charge and relative gains with respect to the Cs-o activates sample. These values are provided alongside the data. Estimated lifetime gains are similar to the ones observed at 505 nm.

Figure 8

Schematic representation of the three different processes leading to the extraction of photoelectrons from GaAs, through the ${\mathrm{Cs}}_3\mathrm{Sb}$ layer into vacuum. Ballistic hot electrons travel trough the activation layer without interaction with lattice or electrons; quasi-ballistic electrons can be still extracted if few scattering events take place so that they have sufficient energy to overcome the barrier to vacuum; hot electrons can exchange energy with low energy electrons and give them sufficient energy to escape into vacuum.

Figure 9

ESP of the samples activated to NEA with ${\mathrm{Cs}}_3\mathrm{Sb}$ estimated using Eq. (5). The used value of ${\mathrm{\Gamma }}_{ee}$ is reported in Table 2.

Figure 10

Spectral response of the samples activated to NEA with ${\mathrm{Cs}}_3\mathrm{Sb}$ and the corresponding prediction of Eq. (10). Barrier thicknesses are given in Table 2.

Figure 11

The fit values of $w_{{\mathrm{Cs}}_3\mathrm{Sb}}$ used in Fig. 10, plotted as a function of the equivalent thickness of the Sb layer. A roughly linear trend is noted, supporting the conclusion that a thicker activating layer weakens the electric dipole at the interface between GaAs and ${\mathrm{Cs}}_3\mathrm{Sb}$.