Open Access
Add to BibSonomy
Abstract
In order to replace an uncontrollable gaseous oxygen source and obtain the required computer-controlled evaporation flow, two types of solid oxygen dispensers composed of silver oxide powder and barium peroxide powder are employed to prepare GaAs photocathodes. The experimental results show that the barium peroxide-based dispenser can release oxygen more effectively and deliver better photoemission performance than the silver oxide-based one. For the silver oxide-based dispenser with a long first warm-up time, an improved activation technique is proposed to avoid the Cs over-saturation and achieve the desired symmetry of photocurrent curve shape. This effective activation technique based on current-driven cesium and oxygen sources can facilitate the realization of automatic activation technology.
© 2017 Optical Society of America
1. Introduction
Over the past few decades, GaAs-based photocathodes with negative-electron-affinity (NEA) surface have drawn much attention to some important applications such as low-light-level image intensifiers for optoelectronic detectors and polarized electron sources for high-energy physics [1–3]. Because of many advantages, such as high quantum efficiency, low thermal emittance and narrow energy distribution, GaAs-based photocathodes with the specific structures were utilized to produce high brightness electron beam served for modern light sources on the basis of energy recovery linacs or free electron lasers [4,5]. Recently, NEA GaAs-GaAsP strained superlattice photocathode was employed to the development of spin-polarized transmission electron microscope, which makes it possible to dynamically generate high-contrast magnetic-field images in nano-structural materials [6]. Additionally, NEA GaAs photocathodes excited by the high-speed pulse laser can also meet the requirements of high current density and high response speed, which are prerequisites for developing THz vacuum devices [7]. Nevertheless, it is noteworthy that the performance of GaAs-based photocathodes in practical applications would strongly rely on the surface NEA state, which is closely related to the surface activation technique [8–10].
NEA photocathodes are realized by adsorption of the electropositive alkali-metals and electronegative species on the clean p-type surface to reduce the surface work function [11]. Generally, solid cesium (Cs) dispenser and gaseous oxygen (O2) or nitrogen trifluoride (NF3) are adopted to prepare GaAs-based photocathodes following the specific alternating activation procedure [12–15]. Compared with the toxic NF3 gas [16], O2 is more popular since it does not require any safety precautions. As is well known, the commercial solid Cs dispenser is very convenient to use since the Cs flux is adjusted by the current passing through the dispenser, while the gaseous O2, usually leaked into the vacuum chamber through a thin wall of a heated silver tube or a variable vacuum leak valve [7,17–19], is hard to obtain precisely during the process of Cs/O alternate activation. Considering the importance of easy control and precision control of introduced O2, the current-driven O dispenser just as the Cs dispenser is required. Although the O dispenser containing barium peroxide (BaO2) was used to prepare GaAs photocathodes from Russia [20,21], the detailed usage case regarding this type of O dispenser was not reported. In this paper, two types of O dispensers composed of silver oxide (Ag2O) powder and BaO2 powder respectively are employed to prepare GaAs photocathodes. The relation between gas release and operating current for the two different O dispensers are investigated. Besides, by using the current-driven Cs and O sources, the Cs/O activation technique suitable for GaAs-based photocathodes is optimized to obtain the satisfactory photoemission performance with the aid of the progressive computer-controlled test system.
2. Experimental
2.1 Solid Cs and O dispensers
In order to realize the effective NEA surface on p-type GaAs-related semiconductors, an important key technical point is how to obtain the high-purity Cs and O gases with the feature of controllable and reproducible evaporation rate. Currently, alkali metal dispensers (AMDs) from SAES Getters are normally used during the preparation of the high-quality photocathodes. As shown in Fig. 1, four Cs dispensers in series and two independent O dispensers, supplied in the thin-walled nickel containers are mounted on a CF63 standard electrical feedthrough. In our activation experiments, the Cs and O dispensers are installed approximately 10 cm away from the cathode sample. Each Cs dispenser is a mixture of a cesium chromate (Cs2CrO4) and a reducing agent Zr 84%-Al 16% (St101) getter material [22]. In addition to the reducing action, the St101 alloy is able to adsorb the chemically active gases produced in the reduction reaction. To introduce the high-purity O2, the current-driven O dispenser just as the Cs dispenser are employed instead of the way usually via leak valve. Two types of O dispensers as shown in Fig. 1(b) and 1(c), which are composed of Ag2O and BaO2 powders respectively are located near the cathode sample, wherein the Ag2O-based O dispenser is also delivered by SAES Getters, which is usually used to effectively prevent lamp blackening caused by carbon deposits on the arc tube [23]. For the solid O dispensers, the high-purity O2 is evaporated into the ultrahigh vacuum (UHV) activation chamber by thermal decomposition of Ag2O or BaO2. Just as the Cs dispenser, the release rate of O2 vapor is simply controlled by adjusting the level of DC electric current passing through the O dispensers. As shown in Fig. 1, it is noted that for the Ag2O-based dispenser from SAES Getters, the gaseous O2 is released from the fine slit in the same way as the Cs dispenser, while for the BaO2-based dispenser, the gaseous O2 is leaked out from several eyelets. The purity and quantity of evaporated gases from the two different O dispensers with the increased operating current can be verified using the quadrupole mass spectrometer (QMS).
2.2 GaAs photocathode preparation
In the modern photocathode formation processes, implementing a computer-controlled activation procedure with rich monitoring tools will be beneficial for activation researches on various photocathodes. For this reason, a computer-controlled test system based on current-driven Cs and O sources for preparing NEA photocathodes was developed to research on activation process optimization of GaAs photocathodes. The schematic diagram of the computer-controlled test system is shown in Fig. 2. Through the RS232 interfaces and DC power supplies, the required Cs and O flux proportional to the operating current can be obtained during the activation procedure. Besides, the vacuum pressure and residual gases in the UHV chamber are monitored by the vacuum gauge and QMS through the serial communication. Most importantly, this self-developed system also realizes the function of monitoring the photocurrent and quantum efficiency, which is an important diagnostic for optimizing the activation procedure and improving photoemission performance. Furthermore, a dedicated man-computer software interface was developed for controlling and monitoring the photocathode activation.
Before the Cs/O activation experiments, the epitaxial p-type GaAs (100)-oriented wafer grown by metal organic chemical vapor deposition was cleaved into 11 × 11 mm2 square plates, and four square GaAs samples were selected to research on activation process of GaAs photocathodes activated by the two different O sources. Each sample has a 2-μm-thick GaAs active-layer with the Zn-doping concentration distributed gradiently from 1 × 1019 cm3 to 1 × 1018 cm3. In the epitaxial structure, a 1-μm-thick AlGaAs buffer-layer with Zn-doping concentration of 1 × 1019 cm3 is sandwiched between the p-type GaAs active-layer and n-type GaAs substrate. The cleaved GaAs cathode samples were first chemically etched in the hydrofluoric acid solution for 5 min. Then the samples mounted on the stainless holder were transferred into the UHV chamber to undergo heat treatment at a temperature of 650 °C for 20 min to ensure the atomically clean surface. When the sample temperature dropped to room temperature, NEA activation for the cleaned samples was performed under the UHV environment with a base pressure of approximately 6 × 10−8 Pa by using the co-deposition technique [10]. During the activation process, the operating current of Cs and O dispensers was adjusted and the photocurrent from the cathode samples irradiated by a tungsten halogen lamp was monitored in real time by the aforementioned computer-controlled test system. When the first photocurrent peak dropped to its 80%, and the operating current of O dispenser was applied and the Cs flux was still continued. In the next several alternate activation cycles, the O flux was suspended when the photocurrent reached its peak again and was introduced once more after the photocurrent dropped to 85% of the peak. In the entire co-deposition process, the Cs source was supplied continuously while the O source was applied periodically. Until no increase in the photocurrent peak, the O flux and Cs flux were stopped in sequence, and then the quantum efficiency curves were measured immediately by the test system. Finally, the photocurrent attenuation cases were measured for the activated GaAs photocathodes under the intensive white light illumination of 100 lx to compare the operational lifetime using the two different O dispensers.
3. Results and discussion
3.1 Gas release from O dispensers
For the Cs/O activated GaAs photocathodes, the ratio of Cs/O flux is crucial to the photoemission performance. The O flux released from the O dispensers should be carefully matched to the Cs flux, and a slight deviation from the optimal O flux would cause a large decline of the final quantum efficiency [13]. In order to judge the suitable operating current applied to the two different O dispensers, the changes in partial pressure of atomic mass unit (AMU) 32 with the increased current passing through the two different O sources are measured through the QMS, and the results are shown in Fig. 3. It is clearly seen from Fig. 3 that the operating current of the Ag2O-based dispenser is larger than that of the BaO2-based one. The operating currents leading to an increase in O2 pressure for the Ag2O-based and BaO2-based dispensers should be larger than 7.7 A and 1.0 A, respectively. In contrast to the O2 pressure for the Ag2O-based dispenser, the O2 pressure for the BaO2-based dispensers is higher although its operating current is smaller. In addition, it is worth noting that for the Ag2O-based dispenser, the curve of pressure change with current is smooth, while the case for the BaO2-based dispenser is distinctly different and the shape of pressure change curve is stepped. This satisfactory stepped trend of O2 pressure shows that the rate of O2 evaporation can be controlled at will and then guarantees reproducibility of O2 release, which is advantageous for controlling the Cs/O alternate activation process. Besides, through the QMS detection, it is found that in addition to the required O2, the dominant residual gases in the UHV activation chamber are H2, H2O, N2/CO and a trace amount of CO2. The partial pressure of O2 is much higher than that of H2O, CO and CO2, which can minimize the effect of O-containing residual gases on the cathode activation results.
Fig. 3 Variation in O2 pressure with the increase of current passing through (a) Ag2O-based and (b) BaO2-based dispensers.
3.2 Optimization of Cs/O activation technique
By using the computer-controlled test system, the evolution of photocurrent combined with operating currents of Cs and O sources in the entire activation process was on-line recorded. Figure 4(a) is a plot of the evolution of photocurrent for the GaAs sample activated by the Cs and O dispenser from SAES Getters using the usual activation technique. The entire Cs/O activation process on the basis of dipole model [24–26], consists of three phases, named as phase I, phase II and phase III, which correspond to the initial photocurrent growth, the first photocurrent attenuation and the alternate Cs/O peak cycles. In phase I, the Cs flux is first exposed to the GaAs cathode sample, which causes the reduction of surface work function and the growth of photocurrent. By measuring the sudden increase of the photocurrent, the evaporation starting point for the Cs dispenser from SAES Getters is no less than 3.8 A. For the Ag2O-based dispenser, the operating current is set on standby to 6.5 A, since it requires a higher operating current of larger than 7.7 A and needs a longer warm-up time. In phase II, the operating current is regulated up to the designated level and the O flux is first introduced onto sample surface when the photocurrent drops to 80% of its first peak. However, as shown in Fig. 4(a), it is found that after the O flux is introduced, the photocurrent continues to attenuate for a relatively long time before the photocurrent increases again, which is completely different from the activation process using the BaO2-based dispenser. The long-time attenuation of photocurrent indicates that in fact the O flux is not released from the dispenser immediately, and the surface of cathode sample has experienced a serious excess of Cs, which is adverse to subsequent activation. This phenomenon shows that as same as the case in Fig. 3, the Ag2O-based dispenser is poorly controlled and the first warm-up time is too long and will delay the normal activation, which is the largest difference from the BaO2-based dispenser. In phase III, there exists another adjacent photocurrent peak when the O flux is closed at the point the photocurrent begins to drop, which is consistent with the activation rule using the BaO2-based dispenser.
Fig. 4 Activation process of GaAs samples activated by (a) usual and (b) improved co-deposition activation technique using Ag2O-based dispenser.
When using the Ag2O-based dispenser, an effective technical approach is proposed to avoid the Cs over-saturation on the GaAs cathode surface during the initial Cs deposition. We think the O flux for the Ag2O-based dispenser should be introduced in advance before the photocurrent drops. Accordingly, aiming to the problem of the slow release of O2 gas from the Ag2O-based dispenser for the first time, the operating current of the O dispenser is regulated up to the designated level when the photocurrent reaches its first peak. Furthermore, the operating current of Cs dispenser can be regulated down to a slightly lower lever to avoid the Cs over-saturation in phase II. In this way, the attenuation rate of photocurrent can be slowed down. When the photocurrent increases again and maintains at a steady rate of growth, the operating current of Cs dispenser is reverted to its original level. At this moment, the rate of O evaporation for the Ag2O-based dispenser is considered acceptable. The distinct improvement in the evolution of photocurrent by using the proposed activation approach is shown in Fig. 4(b). Activation experiments with different Cs/O current ratios were performed on the three GaAs cathode samples using the Ag2O-based dispenser. The photocurrent curves during the period of Cs/O alternate activation cycles are shown in Fig. 5, and the detailed activation process parameters are listed in Table 1. For the three GaAs samples, the Cs/O current ratios are 3.9A/8.0A, 3.8A/7.8A and 4.0A/8.0A, respectively, in which sample 1 are activated by using the usual co-deposition activation technique, while samples 2 and 3 are activated by using the improved co-deposition activation technique as mentioned above. It is seen from Fig. 5 and Table 1 that when the operating currents for the Cs and O dispensers get smaller, the elapsed time of the entire Cs/O activation process would get longer. Obviously, among the three samples, sample 2 exposed by the smallest Cs and O flux consumes the longest activation time.
Fig. 5 Comparison of photocurrent curves of GaAs samples with different Cs/O current ratios using Ag2O-based dispenser.
Table 1. Activation process parameters of GaAs samples with different Cs/O current ratios.
In phase III, how to judge whether the Cs/O current ratio is appropriate or not is crucial to the formation of regular surface activation-layer and the final photoemission capability. In our daily Cs/O activation experiments, the activation rule for reference is as follows: in each Cs/O alternate activation cycle, the peaks and valleys of the photocurrent curve should be symmetrical and the half width of the peaks and valleys should be narrow as possible. As shown in Fig. 5, when the photocurrent reaches the local O peak and it begins to drop. At this time, the O flux is turned off and the Cs flux is continued. The photocurrent continues to increase until it reaches another higher Cs peak and begins to drop again. After that, the O flux is introduced again when the photocurrent drops to a certain proportion of its Cs peak. After the photocurrent drops for a while, it begins to increase once more in turn. It is clearly seen from Figs. 4 and 5 that the photocurrents at latter peak and valley are higher than those at previous peak and valley. After several alternate cycles, the activation process is completed when the peak value increases no further. It is seen from Fig. 5 that, when the Cs/O current ratio is 4.0A/8.0A for the dispensers from SAES Getters, the increase rate of photocurrent reaching O peak is almost as same as that of photocurrent reaching Cs peak, meanwhile the decline rate of photocurrent reaching the valley is approximately equal to the increase rate of photocurrent reaching O peak. As a result, the photocurrent curve during the Cs/O alternate activation process exhibits a perfect symmetry. In this case, the operating currents for Cs and O sources, and the Cs/O current ratio are considered suitable. As for sample 3, the shapes of peaks and valleys of the photocurrent are more symmetrical and the half width is narrower, which indicates that the Cs/O current ratio of 4.0A/8.0A is appropriate to the activation of GaAs photocathodes when using the Ag2O-based dispenser.
Figure 6 shows the difference in photocurrent evolution during the Cs/O activation process between GaAs samples activated by the Ag2O-based and BaO2-based dispensers, wherein sample 4 is activated by the usual co-deposition technique as same as sample 1. As shown in Fig. 6, the sample activated by the BaO2-based dispenser can obtain a higher photocurrent peak and a shorter activation time compared with that activated by the Ag2O-based dispenser, although sample 3 is activated by the improved co-deposition technique as mentioned above. The peaks and valleys of the photocurrent curve for sample 4 are symmetrical as well, and the half width for sample 4 is narrower than that for sample 3. It is seen from Fig. 6 and Table 1 that the photocurrent in phase II for sample 4 begins to increase again prior to that for sample 3, which fully proves that the BaO2-based dispenser needs a shorter warm-up time and can release the effective O2 gas more rapidly.
3.3 Quantum efficiency and operational lifetime
The measured quantum efficiency curves for the four GaAs cathode samples are shown in Fig. 7, and the detailed spectral parameter corresponding to each sample were listed in Table 2. Obviously, for the three GaAs cathode samples activated by the Ag2O-based dispenser, the sample 1 activated by the usual co-deposition technique achieves the lowest photocurrent peak and consequently obtains the worst quantum efficiency over the region of 400-900 nm. Whereas, the sample 4 activated by the improved co-deposition technique obtains the highest quantum efficiency in the entire response waveband. Compared with sample 2 exposed by smaller Cs and O flux, sample 3 with the Cs/O current ratio of 4.0A/8.0A can obtain higher quantum efficiency in the long-wavelength response region, which is beneficial to improve the cathode integral sensitivity and the response capability to night sky light. Therefore, when using the Ag2O-based dispenser to activate GaAs cathode samples, the improved co-deposition technique combined with the symmetrical peaks and valleys and the narrow half width would be indeed effective for enhancing the photoemission performance of GaAs photocathodes. However, the GaAs sample activated by using the BaO2-based dispenser can obtain higher quantum efficiency in the entire response waveband than that activated by using the Ag2O-based dispenser, which could be ascribed to the more rapid and more sufficient release of O2 from the current-driven O dispenser conveniently.
Table 2. Spectral parameters of GaAs samples with different Cs/O current ratios.
Meanwhile, the operational lifetime is another pivotal parameter for NEA GaAs-based photocathodes under working condition, which is because that the Cs/O activation layer on the activated cathode surface is extremely sensitive to the light illumination and the contamination of O-containing residual gas in the environment. The photocurrent attenuation with time reflects the degradation of activated photocathodes. Following a common definition, we say that the cathode lifetime is the time interval during which the photocurrent decreases by a factor of 1/e. Figure 8 and Table 3 show the difference in operational lifetime between the GaAs cathode samples activated by the two different O dispensers. The two GaAs samples were irradiated with continuous intensive white light of 100 lx and the photoelectrons were collected with a bias voltage of 200 V. Obviously, compared with sample 3 activated by the Ag2O-based dispenser, sample 4 activated by the BaO2-based dispenser can obtain a longer operational lifetime along with a higher photocurrent. The differences in photocurrent and lifetime indicate that the Cs/O activation layer on the surface of sample 4 has a more ordered atomic arrangement and thus a more stable structure, which could be caused by the more effective O2 release from the BaO2-based dispenser. Therefore, the BaO2-based dispenser is an ideal O source that can replace the gaseous O source usually via the uncontrollable leak valve.
Table 3. Operational lifetime of GaAs cathode samples activated by different O sources.
4. Conclusions
In summary, two types of current-driven solid O sources consisting of the Ag2O-based and BaO2-based dispensers have been employed to prepare GaAs photocathodes, which can replace the gaseous O source usually via the uncontrollable leak valve and greatly improve the portability of O2 release into UHV chamber. By comparison of the activation experiments, it is found that the GaAs sample activated by the BaO2-based dispenser with more effective O2 release can obtain better photoemission performance than that activated by the Ag2O-based dispenser. The unfavorable activation result using the Ag2O-based dispenser can be ameliorated by the improved co-deposition activation technique, which helps to avoid the Cs over-saturation and achieve the symmetric peaks and valleys during the Cs/O alternate activation process. In future, with the aid of current-driven Cs and O dispensers, the automatic activation technology completely based on computer control would be a major improvement in reproducibility of high performance photocathodes.
Funding
National Natural Science Foundation of China (NSFC) (61771245, 61301023, 61405025); Science and Technology on Low-Light-Level Night Vision Laboratory Foundation of China (J20150702); Specialized Research Fund for the Doctoral Program of Higher Education of China (20133219120008).
References and links
1. K. Chrzanowski, “Review of night vision technology,” Opto-Electron. Rev. 21(2), 153–181 (2013). [CrossRef]
2. W. Liu, Y. Chen, W. Lu, A. Moy, M. Poelker, M. Stutzman, and S. Zhang, “Record-level quantum efficiency from a high polarization strained GaAs GaAsP superlattice photocathode with distributed Bragg reflector,” Appl. Phys. Lett. 109(25), 252104 (2016). [CrossRef]
3. X. Jin, S. Ohki, T. Ishikawa, A. Tackeuchi, and Y. Honda, “Analysis of quantum efficiency improvement in spin-polarized photocathode,” J. Appl. Phys. 120(16), 164501 (2016). [CrossRef]
4. L. B. Jones, S. A. Rozhkov, V. V. Bakin, S. N. Kosolobov, B. L. Militsyn, H. E. Scheibler, S. L. Smith, A. S. Terekhov, D. G. Crabb, Y. Prok, M. Poelker, S. Liuti, D. B. Day, and X. Zheng, “Cooled transmission-mode NEA-photocathode with a band-graded active layer for high brightness electron source,” AIP Conf. Proc. 1149(1), 1057–1061 (2009). [CrossRef]
5. S. Karkare, L. Boulet, L. Cultrera, B. Dunham, X. Liu, W. Schaff, and I. Bazarov, “Ultrabright and ultrafast III-V semiconductor photocathodes,” Phys. Rev. Lett. 112(9), 097601 (2014). [CrossRef] [PubMed]
6. M. Kuwahara, S. Kusunoki, Y. Nambo, K. Saitoh, X. Jin, T. Ujihara, H. Asano, Y. Takeda, and N. Tanaka, “Coherence of a spin-polarized electron beam emitted from a semiconductor photocathode in a transmission electron microscope,” Appl. Phys. Lett. 105(19), 193101 (2014). [CrossRef]
7. K. Mitsuno, T. Masuzawa, Y. Hatanaka, Y. Neo, and H. Mimura, “Activation process of GaAs NEA photocathode and its spectral sensitivity,” in 3rd International Conference on Nanotechnologies and Biomedical Engineering (Springer, 2016) pp. 163–166. [CrossRef]
8. X. Jin, A. A. C. Cotta, G. Chen, A. T. N’Diaye, A. K. Schmid, and N. Yamamoto, “Low energy electron microscopy and Auger electron spectroscopy studies of Cs-O activation layer on p-type GaAs photocathode,” J. Appl. Phys. 116(17), 174509 (2014). [CrossRef]
9. S. Moré, S. Tanaka, S. Tanaka, Y. Fujii, and M. Kamada, “Interaction of Cs and O with GaAs (100) at the overlayer–substrate interface during negative electron affinity type activations,” Surf. Sci. 527(1–3), 41–50 (2003). [CrossRef]
10. D. T. Pierce, R. J. Celotta, G. C. Wang, W. N. Unertl, A. Galejs, C. E. Kuyatt, and S. R. Mielczarek, “The GaAs spin polarized electron source,” Rev. Sci. Instrum. 51(4), 478–499 (1980). [CrossRef]
11. W. E. Spicer, “Negative affinity 3-5 photocathodes: Their physics and technology,” Appl. Phys. (Berl.) 12(2), 115–130 (1977). [CrossRef]
12. N. Chanlek, J. D. Herbert, R. M. Jones, L. B. Jones, K. J. Middleman, and B. L. Militsyn, “High stability of negative electron affinity gallium arsenide photocathodes activated with Cs and NF3,” J. Phys. D Appl. Phys. 48(37), 375102 (2015). [CrossRef]
13. Z. Liu, Y. Sun, S. Peterson, and P. Pianetta, “Photoemission study of Cs-NF3 activated GaAs (100) negative electron affinity photocathodes,” Appl. Phys. Lett. 92(24), 241107 (2008). [CrossRef]
14. S. Pastuszka, A. S. Terekhov, and A. Wolf, “‘Stable to unstable’ transition in the (Cs, O) activation layer on GaAs (100) surfaces with negative electron affinity in extremely high vacuum,” Appl. Surf. Sci. 99(4), 361–365 (1996). [CrossRef]
15. F. Ciccacci and G. Chiaia, “Comparative study of the preparation of negative electron affinity GaAs photocathodes with O2 and with NF3,” J. Vac. Sci. Technol. A 9(6), 2991–2995 (1991). [CrossRef]
16. W. T. Tsai, “Environmental and health risk analysis of nitrogen trifluoride (NF3), a toxic and potent greenhouse gas,” J. Hazard. Mater. 159(2-3), 257–263 (2008). [CrossRef] [PubMed]
17. F. C. Tang, M. S. Lubell, K. Rubin, A. Vasilakis, M. Eminyan, and J. Slevin, “Operating experience with a GaAs photoemission electron source,” Rev. Sci. Instrum. 57(12), 3004–3011 (1986). [CrossRef]
18. M. Kuriki, C. Shonaka, H. Iijima, D. Kubo, H. Okamoto, H. Higaki, K. Ito, M. Yamamoto, T. Konomi, S. Okumi, M. Kuwahara, and T. Nakanishi, “Dark-lifetime degradation of GaAs photo-cathode at higher temperature,” Nucl. Instrum. Methodes Phys. Res., Sect. A 637(1), S87–S90 (2011).
19. T. Guo and H. R. Gao, “Photoemission stability of negative-electron-affinity GaAs photocathodes,” Proc. SPIE 1982, 127–132 (1993). [CrossRef]
20. L. I. Antonova and V. P. Denissov, “High-efficiency photocathodes on the NEA-GaAs basis,” Appl. Surf. Sci. 111, 237–240 (1997). [CrossRef]
21. V. L. Alperovich, A. G. Paulish, and A. S. Terekhov, “Unpinned behavior of the electronic properties of a p-GaAs (Cs,O) surface at room temperature,” Surf. Sci. 331–333, 1250–1255 (1995). [CrossRef]
22. M. Succi, R. Canino, and B. Ferrario, “Atomic absorption evaporation flow rate measurements of alkali metal dispensers,” Vacuum 35(12), 579–582 (1985). [CrossRef]
23. Seehttps://www.saesgetters.com/products-functions/products/dispensers/oxygen-dispensers
24. C. Y. Su, W. E. Spicer, and I. Lindau, “Photoelectron spectroscopic determination of the structure of (Cs,O) activated GaAs(110) surfaces,” J. Appl. Phys. 54(3), 1413–1422 (1983). [CrossRef]
25. J. Zou, B. Chang, Z. Yang, Y. Zhang, and J. Qiao, “Evolution of surface potential barrier for negative-electron-affinity GaAs photocathodes,” J. Appl. Phys. 105(1), 013714 (2009). [CrossRef]
26. V. V. Bakin, K. V. Toropetsky, H. E. Scheibler, A. S. Terekhov, L. B. Jones, B. L. Militsyn, and T. C. Q. Noakes, “p-GaAs(Cs,O)-photocathodes: Demarcation of domains of validity for practical models of the activation layer,” Appl. Phys. Lett. 106(18), 183501 (2015). [CrossRef]
