1.1

XPS analysis

The GaAs cathode sample after chemical cleaning was sent into the vacuum system, the scanning focused monochromatic X-ray was used to perform micro-area selection analysis. First, the sample was excited by X-ray to generate an SXI image, as shown in Fig. 2. It can be clearly seen that the surface of sample 1 showed an irregularly shaped impurity plaque distribution, while the surface of sample 2 and 3 were distributed in a center-symmetric band, without impurity plaque distribution. With the aid of the scanning focused X-ray imaging positioning technique combined with the SXI image function, the plaque area of interest in the SXI image was selected labeled as the black box as shown in Fig. 2, and the 50 μm X-ray beam spot was used to accurately analyze photoelectron spectroscopy in the box area.

Figure 2.

SXI photos for the three samples after the chemical cleaning process.

After the sample was activated, XPS analysis was performed again in the same area of the activated sample surface. The detected surface chemical composition of the three samples after chemical cleaning and activation are shown in Table 1. It can be seen from Table 1 that compared with the two other samples, the impurity plaques in sample 1 not only contain more C and O impurity pollution, but also Na and Cs pollution, which were not found in the previous XPS test, indicating that if the metal sheet used to fix the cathode sample is not thoroughly cleaned after multiple activations, it will easily contaminate the cleaned cathode sample when used repeatedly. It can be seen that Na contamination existed on the surface of sample 1 after chemical cleaning and activation, which shows that the Na pollution on the cathode surface could not be removed by heating purification at 600°C or Cs/O activation. In Table 1, there is a small amount of Cs 3d signal after sample 1 was chemically cleaned, and no Cs 3d peak signal was detected before samples 2 and 3 was activated. Due to the adsorption of Cs in the process of activation, Cs 3d appeared on the surfaces of samples 1 and 2. The XPS signal of C 1s and O 1s of sample 1 are higher than those of samples 2 and 3. In addition, the elemental contents of Ga 3d and As 3d on the surface of sample 1 is much lower than those of samples 2 and 3, which indicates that the surface Na pollution and other impurities such as C and O will affect the heat purification effect. By comparison of XPS spectra between sample 2 and sample 3, it can be seen that after the chemical cleaning, the C 1s content of sample 3 is less than that of sample 2, but the O 1s content is greater than that of sample 2 due to the UV-O₃ treatment after chemical cleaning. After activation, there is still a little C component on the surface of sample 2, and the O1s content of sample 2 exceeds that of sample 3. However, the content ratio of Cs and O is close to 2:1 for sample 3.

Table 1.

Surface chemical compositions and contents in the selected regions of SXI photos for the three samples.

Figure 3 shows the XPS spectra and peak fitting spectra of Ga 3d on the surface of samples after chemical cleaning. The normal Shirley-style background line and Gauss-Lorentz function are used in the peak fitting. It can be seen from Fig. 3 that the peaks of Ga 3d on the sample has two states, namely Ga₂O₃ with a high binding energy peak (20.4eV) and GaAs bulk with a low binding energy peak (19.1eV). The calculated oxide percentages are shown in Table 2. After chemical cleaning, Ga₂O₃ in samples 1, 2 and 3 accounted for 2.08%, 1.21% and 2.36% of the total amount of Ga 3d, respectively, indicating that the chemical cleaning had a significant removal effect on the surface natural oxide layer. Because of the additional UV-ozone treatment, the oxide content of Ga 3d on sample 3 is the highest. Figure 4 shows the XPS spectra and split peak fitting spectra of Ga 3d after activation. The proportion of Ga₂O₃ in the whole Ga 3d spectrum increased due to the adsorption of O during the activation. From Table 2, it can be seen that Ga2O3 in sample 1 of Fig. 4(a) accounts for 8.11% of the total Ga 3d, Ga₂O₃ in sample 2 of Fig. 4(b) accounts for 4.62% of the total Ga 3d, and that ratio of sample 3 is 1.92%. The change in Ga₂O₃ content indicates that the difference in the proportion of Ga₂O₃ to Ga3d among the three samples was obvious. It seems that the presence of Na and Cs pollution and residual C impurities on the surface can promote the adsorption of O on the activated Ga atom. This may be due to the small electronegativity of Ga, which is more likely to react with the more electronegative adsorbed O to generate oxides^[12].

Figure 3.

XPS spectra of Ga 3d after chemical cleaning for the three samples.

Figure 4.

XPS spectra of Ga 3d after activation for the three samples.

Table 2.

XPS fitted peak ratios of Ga 3d and As 3d after chemical cleaning and activation for the three samples.

Figure 5 and Figure 6 shows the XPS spectra and split peak fitting spectra of As 3d. The As 3d peaks in the sample can be fitted to the spectrum peaks in three states, from the high binding energy end to the low binding energy end are As₂O₃ (44.4ev), As (41.6ev) and GaAs bulk (40.8ev) respectively. It can be seen from Table 2 that the ratio of the peaks of sample 1 in the three states of As₂O₃, As and GaAs bulk after chemical cleaning is 0.93%, 34.30% and 64.77%, respectively, while the ratio of the peaks after activation is 1.75%, 31.62% and 66.63%, which shows that the three peaks of sample 1 in As 3d spectrum were not changed very much. For sample 2, the ratio of As₂O₃, As and GaAs after chemical cleaning are 0.97%, 43.21%, and 55.82%, and the peaks after activation are 0% and 33.27% and 66.73%. While for sample 3, the ratio of As₂O₃, As and GaAs after chemical cleaning are 0.99%, 43.01%, and 56%, and the peaks after activation are 0% and 29.93% and 70.07%. It can be clearly seen that the bulk GaAs peak proportion in samples 2 and 3 increased, while the As spectral peak proportion decreased significantly. Besides, the As₂O₃ component disappeared after activation for samples 2 and 3, which is because that during heating purification, As₂O₃ can react with GaAs to generate volatile As^[13,14]. Furthermore, it is found that the proportion of As peak and As₂O₃ peak in sample 1 has little change before and after activation, which is mainly because Na and other impurities hindered the thermal volatilization of As and conversely enhanced the reaction between As and adsorbed O. From the results in Table 2, it can be inferred that for the original GaAs photocathode without Na and Cs pollutants, in addition to the Cs-O dipoles formed on the surface of the active cathode, there exist extra Ga oxides. Therefore, the surface of the GaAs photocathode presented the GaAs-O-[Cs]:[Cs+]-O^-2-Cs+ dipole model proposed by Su et al^[15]. More interestingly, it can also be seen from Table 1 that the proportion of O content in the activated surface of sample 3 is minimum. Whether the impurity pollutions such as Na, Cs and C affect the removal of original oxides during heat treatment and the following O adsorption during activation needs to be further studied.

Figure 5.

XPS spectra of As 3d after chemical cleaning for the three samples.

Figure 6.

XPS spectra of As 3d after activation for the three samples.

1.2

Activation analysis

In addition to the XPS analysis of the surface chemical composition of the cleaned and activated GaAs photocathode samples, it is also necessary to directly determine the effect of metal contamination and different cleaning methods on cathode activation performance through activation experiments. Using the multi-information on-line measurement and control system, the photocurrent curves of the GaAs cathode samples activated by Cs and O co-deposition recorded in real time are shown in Fig. 7. It can be seen from the comparison that the first Cs peak of sample 3 is the earliest and the final photocurrent is the largest. Although the rise time of photocurrent curve of sample 2 is slightly slower than that of sample 3, the increase extent of photocurrent is not so large, and the final photocurrent peak is smaller than that of sample 3. The photocurrent appears in sample 1 after activation for about 15 minutes, and the first Cs peak rise slowly, and the final photocurrent peak is the smallest. Therefore, the atomically clean surface of the GaAs photocathode is very important for the Cs and O adsorption, which is conducive to generating a larger photocurrent, which is well proved by the activation phenomena.

Figure 7.

Photocurrent curves of the three samples.

Through the on-line spectral response test system, the quantum efficiency curves of the activated samples were also measured, as shown in Fig. 8. It can be seen from the figure that the spectral response of sample 3 is significantly greater than that of sample 1 and sample 2 at the same wavelength, and sample 3 has higher quantum efficiency in the whole band range. Since the three samples were all from the same GaAs photocathode single crystal wafer, they had the same structure and photoelectric physical properties. Therefore, one of the important factors that cause the difference of the emission performance between the three is the difference of the surface electron affinity, which is determined by the surface potential barrier. The photoelectrons of the GaAs cathode are excited by the incident photons, and the photoelectrons will be scattered by the surface barrier during the escape to the surface. The impurity contamination on the surface is obviously not conducive to the well-ordered adsorption of Cs and O and the formation of double dipole layers, which will cause a larger surface barrier and result in the overall low quantum efficiency of the whole band.

Figure 8.

Spectral response curves of the three samples.