Thin films of barium and scandium metal were deposited onto W(100) from metal evaporation sources in ultrahigh vacuum. The Ba-Sc-W(100) surface was then exposed to oxygen and heated in stages to several temperatures up to 1400 °C to examine the desorption behavior of these adsorbates. In one case, microsynchrotron radiation x-ray photoelectron spectroscopy (micro-SRXPS) was performed on this surface using a low energy electron microscope. In another experiment, reported here, Auger electron spectroscopy data were collected from the Ba-Sc-O-W(100) surface as a function of heating to temperatures corresponding to the same temperatures used for the micro-SRXPS. Both measurements show that barium desorbs with scandium and oxygen at or below 800 °C.
Scandate thermionic cathodes are used for high power and high frequency vacuum electron devices.1 We have used metal deposition of barium and scandium on the low index tungsten surfaces to model the complex behavior of these metals and their oxides on the surface of commercial thermionic cathodes.2–11 Three distinct behaviors that are significant for the action of barium and scandium on tungsten surfaces have been identified: (1) both barium and scandium dewet from the tungsten surface at temperatures below the operation temperature of commercial thermionic cathodes;3,4 (2) scandium can scavenge oxygen from the tungsten surface and inhibit the formation of tungstates;2 and (3) coadsorbed barium, scandium, and oxygen desorb at temperatures at or below 800 °C.5 The result of these three actions by scandium and barium show that scandium acts as a cleaning agent that prepares the tungsten surface for barium, which lowers the work function of the tungsten surface. Combined, these results indicate that scandium does not directly participate in the lowering of the tungsten surface work function through enhancement of the barium surface dipole.1
Auger electron spectroscopy (AES) was performed at Brookhaven National Laboratory at the Center for Functional Nanomaterials. A microcylindrical mirror analyzer (micro-CMA) from RBD Instruments was used to collect the spectra. The micro-CMA was integrated into the Elmitec Elektronenmikroskopie GmbH LEEM V (low energy electron microscope) microscope, which was utilized for its sample heating and imaging capabilities. The system pressure was maintained below 5 × 10−9 Torr. An electron beam of 2 keV was used with a working distance from the sample of 3.8 mm and a resolution (ΔE/E) of 0.7%.
The sample studied was a 99.999% pure W(100) single crystal from Accumet Materials Co. (Ossining, NY) with a diameter of 9 mm and a thickness of 0.25 mm. The crystal was prepared by heating to 1800 °C in oxygen and then flashed again to 1800 °C to remove adsorbed gases. Cleanliness was confirmed using AES as well as low energy electron diffraction, which gave a sharp 1 × 1 pattern and a clean LEEM image.
Scandium and barium were deposited onto the W(100) surface by evaporation in a preparation chamber attached to the LEEM V that also housed the micro-CMA system. Scandium and barium were deposited on the clean W(100) crystal by thermal evaporation from a tungsten filament. Tungsten wire 0.38 mm in diameter (0.015 in.) in a 4 mm coil of five turns was used for the evaporation. A piece of scandium metal was placed in the coil. A second, identical coil, filled with a piece of barium, was used for barium deposition. Film thicknesses were determined in a separate chamber using a quartz crystal thickness monitor and reproduced in the LEEM by using the same time and deposition current. The same dose was used for all measurements, 1 nm scandium and 20 nm barium. The oxygen dose was 90 L: 5 × 10−7 Torr for 3 min. Depositions were verified by moving the sample from the LEEM chamber into the preparation chamber housing the Auger spectrometer and collecting an AES spectrum. Once depositions were completed, the sample was moved back into the LEEM V for heating to a given temperature, where the temperature was recorded using a type C thermocouple on the sample holder. After each heating cycle, the sample was then moved back into the micro-CMA chamber to measure the Auger spectra. The crystal was flashed to 1800 °C after each deposition–heating cycle was completed.
A series of Auger spectra acquired after the substrate was heated to successively higher temperatures is shown in Fig. 1. The data are initially taken as Auger electron counts [N(E)] versus kinetic energy (E); we show differentiated data [dN(E)/dE] with respect to the kinetic energy, the more commonly used format.12 This increases the noise in the data, which was reduced via exponential smoothing. Characteristic Auger peaks of the W crystal and deposited materials are measured. In the energy range observed, 10–620 eV, all relevant elements have at least one feature. Tungsten has peaks at 48, 169, and 179 eV. Scandium and oxygen have peaks at 340 and 503 eV, respectively. Barium has two primary peaks within this range, located at 57 and 584 eV. The carbon signal at 272 eV is from residual gas that adsorbs during the deposition and heating sequences.
Quantitative analysis of Auger spectra typically involves taking ratios of differentiated peak heights of two elements.12 The normalized Ba/Sc ratio plotted against temperature is shown in Fig. 2. Ba has two peaks at 57 and 584 eV and both were used to generate Ba/Sc ratios. The peak at 57 eV also partially overlaps with the tungsten peak at 48 eV. In order to get the height of the Ba peak at that energy, a baseline W spectrum was subtracted. Errors in these ratio measurements stem from the initial noise in the data, especially at the lower energies. The temperature measurements have a broadly estimated ±100 °C uncertainty due to variations in the quality of the thermocouple attachment to a given sample holder.
In earlier work,5 we used microsynchrotron radiation x-ray photoelectron spectroscopy (micro-SRXPS) to examine the same adsorption–desorption process. Nearly identical adsorbate thicknesses and oxygen exposures were used. The micro-SRXPS was performed in an Elmitec LEEM III, which has a spherical energy analyzer in the imaging path. The electrons emitted from the surface and collected by the objective lens can be imaged in three ways: (1) without energy analysis or (2) an energy filtered image can be observed. In the third method, (3) the dispersive plane of the energy analyzer is projected onto the detector, and a spectrum can be acquired. The electrons for each of these imaging or spectroscopy modes of operation come from the same sample area. A 2 μm field aperture is used to restrict the area from which the spectroscopic information is obtained.
In the micro-SRXPS spectra, shown in Fig. 3, the desorption of Ba-Sc-O-W(100) was monitored using the Ba 4d peaks.5 The barium 4d peaks with scandium and oxygen present are shown in the left panel of Fig. 3, I. The barium peaks are absent after heating to 800 °C. In the right-hand panel, Fig. 3, II, scandium was no longer present, and the barium redose of 20 nm and exposure to 90 L oxygen resulted in the barium 4d peaks persisting up to 1200 °C. The signal to noise ratio in these spectra is excellent. No smoothing was necessary. Given the data in Fig. 3, we can confidently describe the function of scandium on the W(100) surface as a cleaning agent, which scavenges oxygen from the W(100) surface and desorbs with barium.
An illustrative example that involves the dewetting of the metal adsorbates is shown in Fig. 1. In Fig. 1, the tungsten peak at 169 eV is observed to appear after the sample is heated to 800 °C. This change could be interpreted as a decrease in the relative amount of adsorbate coverage on the surface. However, Fig. 2 shows that the relative amount of barium to scandium is largely unchanged before 800 °C. Because barium and scandium have different desorption temperatures, another mechanism, dewetting, is responsible.5,13 The dewetting of the Ba/Sc film at these temperatures, where the metals retract into droplets when heated to temperatures around ½Tmelting, exposes the underlying tungsten crystal.3,4 On the temperature scale studied, dewetting of the Ba/Sc film at low temperatures (<800 °C) reintroduces the tungsten signal into the spectra. At mid- to high-temperatures (≥800 °C), desorption is the significant process.
A commercial thermionic cathode is a complex device. Often, AES or energy dispersive x-ray analysis is used to investigate the surface properties of a cathode.1 Dewetting is a process that produces an inhomogeneous surface. Pores in the porous tungsten cathode body can also cause inhomogeneities that complicate the interpretation of spectra with a strong dependence on the electron escape depth. Although applied to a smooth model surface, W(100), the micro-SRXPS data are simpler to interpret in this case because the barium peaks do not overlap any of the other elements of interest. Synchrotron radiation delivers an intense beam so that data have excellent signal to noise ratios without further smoothing. With the knowledge of scandium and barium dewetting, the conclusions reached in Ref. 5 are clearly supported by the AES data.
We have compared two nearly identical model surfaces of adsorbed Ba-Sc-O-W(100) in order to understand the action of scandium in thermionic “scandate” cathodes. One model surface was investigated with micro-SRXPS and the other with AES. The two results agree. Knowledge of the dewetting of scandium and barium on W(100) is essential for the understanding of the AES and micro-SRXPS data.
This work was supported by the DARPA INVEST (Grant No. N66001-16-1-4040). This research used resources of the Center for Functional Nanomaterials and National Synchrotron Light Source II, which are U.S. Department of Energy (DOE) Office of Science Facilities, at Brookhaven National Laboratory under Contract No. DE-SC0012704.