Work function lowering of LaB₆ by monolayer hexagonal boron nitride coating for improved photo- and thermionic-cathodes

Free electrons are vital components for a wide range of systems from photodetectors and microscopes to accelerators. Multiple cathode types are capable of generating such free electrons (photo, thermionic, and field emission), and an intrinsic material property that determines their performance is the work function or the minimum energy required for electrons to escape from a material surface into a vacuum. For photocathodes, Spicer's three-step model¹ identifies the work function as the main property that determines the escape rate of electrons into a vacuum when a material is excited with photons. Therefore, the work function is essential, because it determines the quantum efficiency (QE) of the photocathodes, where QE is a rate of free electrons generated per incident photons. Electrons overcome the work function via thermal energy to generate free electrons in thermionic cathodes. For field emission cathodes, free electrons are generated by electrons tunneling through the potential barrier of cathode surfaces via an electric field, where the barrier height is the work function. The challenge in improving the cathode performance lies in the preparation of robust low work function materials that allow cathodes to generate more free electrons at a given illumination wavelength, temperature, or electric field, at which the cathodes operate.

Here, we investigated the effects of a two-dimensional (2D) material coating on lanthanum hexaboride (LaB₆) for photo- and thermionic-cathodes. 2D materials, such as hexagonal boron nitride (hBN) and graphene, are thermally stable up to at least 2200 °C, are chemically inert, and possess an excellent gas barrier property.^2–7 In addition, there are experimental and theoretical reports on lowering the work function of metals and alkali antimonide semiconductors by using 2D material coating.^8–12 Thus, if coating LaB₆ with 2D materials lowers its work function, this could provide a breakthrough in achieving both robustness and a higher yield of free electron generation. In general, lowering the work function of a thermionic cathode by coating is challenging, because the coating materials need to withstand a high temperature of ∼1500 °C and should not increase the work function. This combination is extremely rare. For photocathodes, lowering of the work function can readily be achieved by well-established cesium (Cs) termination as demonstrated on gallium arsenide (GaAs) and other materials.¹³ The problem is that those surfaces are highly sensitive to environmental conditions; high chemical reactivity of Cs requires an ultrahigh vacuum (UHV) of 10⁻¹¹Torr for its operation, which severely limits operation environments. Even for the cathodes with Cs incorporation, in contrast to termination, such as the case of cesium antimonide (Cs₃Sb) and cesium potassium antimonide (CsK₂Sb), their operations still require 10⁻¹⁰Torr. An exception is Cs₂Te, which could operate in one or two orders of magnitude higher pressure; however, it requires 266 nm excitation due to its higher work function in contrast to 532 nm for other Cs terminated and incorporated photocathodes.

We chose the LaB₆ cathode for our purpose, because it is a well-established thermionic cathode material that is commercially available for electron microscopes and spectrometers. Its clean surface has a low work function of ∼2.4 eV (Ref. 14) and a long lifetime of ∼5000 h due to a low evaporation rate at its operation temperature of ∼1500 °C. The low work function of LaB₆ is also attractive as a photocathode—although the level of interest in this aspect is low compared to thermionic cathodes,^15–20 interest in using LaB₆ as a photocathode exists none-the-less due to its high temporal flexibility of generated free electrons compared to other types of cathodes.

We first coated the single crystal LaB₆ (100) (AP-Tech) with hBN and graphene via an established wet-transfer method. Poly(methyl methacrylate) (PMMA) polymer support was coated onto hBN on copper (Cu) and graphene on Cu that were synthesized via chemical vapor deposition (CVD).^21–26 Cu foils were then etched away using ammonium persulfate (APS) followed by transfers of PMMA supported hBN and graphene onto LaB₆ (100) by exposure to water.²⁷ As shown in Fig. 1(a), half of LaB₆ (100) was first covered with graphene and dried in air. PMMA was then removed by immersing the entire assembly in an acetone bath for three cycles of 15 min each. Subsequently, a quarter of bare and graphene covered LaB₆ (100) was covered by hBN, dried in air, and then PMMA was removed by the same protocol as that of graphene. The LaB₆ (100) surface was flat with a roughness of below 1 nm by atomic force microscopy (AFM) [Fig. 1(b)]. We also performed AFM on transferred hBN and graphene and confirmed that they are continuous sub-nm films, thus, atomically thin [Figs. 1(c) and 1(d)]. Raman spectroscopy on hBN and graphene indicated that their crystal quality is high [Figs. 1(e) and 1(f)]. Specifically, the Raman spectrum of hBN had a signature E_2g peak around 1370 cm⁻¹,²⁸ and that of graphene had signature G and 2D peaks around 1550 and 2700 cm⁻¹ with no observable defect D peak around 1350 cm⁻¹.²⁹ 

The sample was then introduced into a system that can perform photoemission electron microcopy (PEEM), photoelectron spectroscopy (PES), and thermionic emission electron microscopy (TEEM) (Fig. 2). The sample temperature was gradually increased from room temperature up to 905 °C. TEEM was performed at elevated temperatures, and PEEM and PES were performed at room temperature after each annealing temperature to exclude thermal effects. In PEEM, ∼4.9 eV photons generated by a mercury (Hg) lamp were illuminated on the sample surface and photoelectrons with a spatial resolution of ∼20 nm to gain information on the spatial distribution of the work function. We followed our PEEM measurements with PES to quantify the difference in the work function for the regions of interest. For PES, the energy resolution was <0.2 eV, and the sample was biased to −3 V to efficiently collect photoelectrons. The Fermi level of the samples was determined using the Fermi edge of an evaporated gold (Au) thin film. We first used the helium (He) I line with an excitation energy of 21.22 eV to determine the Fermi edge of an evaporated Au film. We then used the Hg lamp on the same evaporated Au film and used the energy difference between He I and Hg lamp excitation to determine the work function of the samples. The purpose of performing TEEM was to correlate the work function observed in PEEM and PES with thermionic emission. We also performed synchrotron-radiation x-ray photoelectron spectroscopy (SR-XPS) separately to investigate the chemical state of our LaB₆ (100) surfaces. Aside from experiments, we performed a density functional theory (DFT) calculation to theoretically investigate the effect of 2D material coating on the work function of LaB₆ (100) surfaces.

Figure 3(a) is a schematic of 2D material coverage on the LaB₆ (100) surface, which indicates graphene and hBN in red and green dashed lines, respectively. Figure 3(b) is a PEEM image of a sample at room temperature after being heated at 905 °C, which was the maximum temperature of our system. Interestingly, we observed a higher intensity of photoelectrons at the region corresponding to the hBN coating, which reveals that the work function is lower for the hBN coated LaB₆ (100) surface compared to the non-coated and graphene-coated counterparts. Note that the hBN region is brighter than the graphene region, even where there is a graphene coating underneath [i.e., in between hBN and LaB₆ (100)] confirming that hBN is playing a critical role in lowering the work function. To demonstrate that our concept is also applicable to thermionic cathodes, we performed TEEM. The regions of 2D material coverage on the LaB₆ (100) surface are indicated in the same way as that of PEEM. 905 °C is the temperature that we started to observe thermionic emission as shown in Fig. 3(c). The results were consistent with PEEM and PES in that the brightest region aligned well with the hBN coated region, and the graphene coated region had brightness in between the hBN-coated region and the non-coated bare regions. The bright star shaped spot at the bottom left of the TEEM image is an artificial effect of the measurements. No observable signal was obtained for TEEM at 806 °C, indicating that thermionic emission at the temperature was below our detection limit [Fig. 3(d)].

We followed up PEEM measurements with PES to quantify the brightness difference. As you can see from Fig. 4, the trend we observed in PEEM measurements was consistent with PES. The work function for non-coated and hBN coated regions was 4.2 and 3.8 eV, respectively. This demonstrates that as large as ∼0.4 eV of the work function can be lowered by simply coating LaB₆ (100) with 2D materials such as hBN. hBN coated and hBN/graphene coated regions had similar work function; thus, it was difficult to distinguish them by the brightness in the PEEM image [Fig. 3(b)]. PES showed that the hBN/graphene coated region had a work function that is comparable to that of the hBN coated region (3.9 eV). The work function for the graphene coated region was 4.1 eV, slightly lower than the non-coated, bare surface. Our PEEM and PES results demonstrated that the work function lowering of LaB₆ (100) that we achieved is directly applicable to photocathodes because both techniques generate free electrons via photoexcitation.

The work function of 4.2 eV we observed for non-coated LaB₆ (100) was much higher than ∼2.4 eV reported for its clean surfaces. This is probably due to the presence of an oxide layer.^14,30 It is known that LaB₆ surfaces, including the (100) form native oxide layer in air, and annealing at ∼1500 °C under 1 × 10⁻⁷Torr is necessary for its complete removal. Our annealing temperature of 905 °C was well below ∼1500 °C. Moreover, our observed work function of 4.2 eV is close to the reported value of ∼4.0 eV for the LaB₆ (100) surface with a saturated oxide layer.¹⁴ We performed SR-XPS at the surface chemistry end-station (BL23SU) of SPring-8 in Japan using soft x-rays (711 eV) to gain information on the oxidation states of our LaB₆ (100) surfaces. As expected, we observed lanthanum oxide contribution in the O 1s peak even after annealing at 905 °C (Fig. 5). Entire peak positions and shapes as well as peak positions of deconvoluted peaks were consistent with those reported for La₂O₃ prior to and after annealing at 800 °C.^31,32 Specifically, while there was a dominant broad peak around 531.47 eV prior to annealing, the peak shape changed drastically after the annealing and peaks formed at 531.03 and 531.85 eV. The peak at lower binding energy has been assigned to lattice oxygen. There is ongoing discussion on the exact assignment of the peak at the higher binding energy due to the complex structure, which contains a mixture of both carbonates and hydroxyls. The results suggest that the annealing of our air exposed LaB₆ (100) surfaces at 905 °C forms La₂O₃. We observed the B 1s peak around 187.5 eV for wide scans prior to and after annealing, indicating that the thickness of our oxide layer is below a few nm as expected for a native oxide layer.

We performed density functional theory (DFT) calculations on the work function of LaB₆ (100) with 2D material coatings to gain insight into the mechanism of observed work function lowering. Specifically, we calculated the work function of oxidized LaB₆ (100) surfaces and these surfaces after coating with hBN and graphene monolayers. We also calculated the case for pristine LaB₆ (100) surfaces as a reference. Our calculations were performed with the use of DFT and the projector augmented-wave (PAW) method³³ as implemented in the Vienna ab initio simulation package (VASP).³⁴ The generalized gradient approximation (GGA) of the Perdew–Burke–Ernzerhof (PBE)³⁵ functional was used to represent the exchange–correlation interaction. The substrates were represented with slab models with a thickness of ∼12 Å and a vacuum gap in the direction normal to the surface to eliminate the interactions between the replicas due to the periodic boundary conditions. The surfaces of the slabs are terminated with La atoms based on previous experiments. The size of the primitive cell of the LaB₆ slab was 4.155 × 4.155 Ų, and the size of a rectangular cell of graphene/hBN was 2.467 × 4.274 Ų. Therefore, a 3 × 1 supercell of the LaB₆ slab (12.465 × 4.155 Ų) and a 5 × 1 supercell of graphene/hBN (12.335 × 4.274 A2) had the lattice mismatch of ∼3%.

We started our calculation on a pristine LaB₆ (100) surface as a reference to ensure that our calculated value was consistent with experiment values. The calculated work function was 2.20 eV, which was indeed in good agreement with an experimental value of 2.3 eV (Ref. 36) [Fig. 6(a), Table I]. After coating with the hBN monolayer, its work function decreased to 1.87 eV. However, the calculated work function increased to 3.44 eV with graphene coating [Fig. 6(b), Table I]. This modification originates from the interfacial dipoles induced by the coating as discussed later. These calculations qualitatively support the work function modifications of LaB₆ that we observed experimentally by the 2D material coatings.

After confirming that our calculation values for the clean surface were consistent with the experimental values, we moved on to a more representative case for our experiments with an oxide layer. The calculated work function of the oxidized LaB₆ (100) surface was dependent on the coverage of oxygen (O) as reported experimentally. For the surface with O atoms at the equivalent metal bridge sites (with O coverage of 1/2 on these available sites³⁰), the calculated work function was 3.60 eV. This is close to ∼4.0 eV reported experimentally for the oxygen saturated surface.¹⁴ However, a large supercell was required to match this surface structure with graphene and hBN coatings, which means that the calculations involving 1/2 O coverage are computationally expensive. Therefore, we used an oxidized LaB₆ (100) surface with 1/3 O coverage instead in our study to represent an oxidized LaB₆ surface [Fig. 6(c), Table I]. Its calculated work function was 3.12 eV, slightly lower than the case of 1/2 O coverage as expected. When the surface was covered with an hBN monolayer, the work function decreased to 2.99 eV. This is qualitatively consistent with the case of clean LaB₆ (100), indicating that the calculations predict lowering of the work function by hBN coating on both the clean and oxide surfaces. The work function increased to 3.88 eV after graphene coating, and this was also qualitatively consistent with the case of clean surfaces. Band diagrams illustrating the work function change of oxidized LaB₆ (100) by graphene and hBN coating are shown in Fig. 7. For the semi-metallic graphene, the charge transfer from LaB₆ to the conduction bands of the coating materials induces inward pointing dipoles, which increase the work function [Figs. 7(a) and 7(b)]. The wave-like curves in Fig. 7(c) are the planar-averaged electron transfer along the direction perpendicular to the surface. The inset of Fig. 7(c) shows the corresponding electron transfers at the interface of oxidized LaB₆ (100) surface and graphene. Blue and yellow regions indicate electron depletion and accumulation, respectively. On the other hand, for monolayer hBN, because of its large bandgap, an outward dipole is formed at the interface originating from exchange repulsion, which decreases the work function [Fig. 7(d) and 7(e)]. Figure 7(f) in comparison to Fig. 7(c) clearly shows that the surface dipole has different directions for graphene and BN coated surfaces. For both the hBN and graphene coating, the degree of work function modification on oxidized surfaces was smaller compared to the clean surfaces, indicating that the oxide layer reduces the change transfers between LaB₆ and 2D materials.

In summary, lowering of the work function for LaB₆ by 2D material coating is demonstrated. Specifically, PEEM and TEEM both revealed that an hBN coated region of the LaB₆ (100) single crystal has a lower work function compared to bare (i.e., non-coated) and graphene coated regions. A broad and uniform brighter image of an hBN coated region in PEEM was quantitatively supported by a 0.4 eV decrease in the work function in photoelectron spectra compared to the bare region. TEEM results were consistent in that the hBN coated region exhibited thermionic emission at 905 °C, whereas the bare and graphene coated regions did not. A larger decrease in the work function for hBN coated LaB₆ (100) compared to graphene coated LaB₆ (100) was qualitatively supported by our density functional theory (DFT) calculations. Adding a presence of an oxide layer on LaB₆ in the calculations improved consistency between the calculations and experimental results. We followed up our calculations with SR-XPS and confirmed the presence of an oxide layer on our LaB₆ (100) single crystal.

This work was supported by the U.S. Department of Energy (DOE) Office of Science U.S.–Japan Science and Technology Cooperation Program in High Energy Physics (HEP) and the JSPS KAKENHI (Grant No. JP17KK0125). The XPS measurements were performed using synchrotron radiation at Beamline BL23SU of SPring-8 (Project Nos. 2021A-E14, 2021B-E22, 2022A-E20, and 2022B-E13) with the approval of the Japan Synchrotron Radiation Research Institute (JASRI) (Proposal Nos. 2021A3831, 2021A3801, 2021B3831, 2021B3801, 2022A3831, 2022A3801, 2022B3831, and 2022B3801). A Part of this work was supported by the JAEA Advanced Characterization Nanotechnology Platform under the remit of “Nanotechnology Platform” of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan (Grant Nos. JPMXP09A21AE0013 and JPMXP09A21AE0042). A part of this work was supported by “Advanced Research Infrastructure for Materials and Nanotechnology in Japan (ARIM)” of the MEXT, Japan (Proposal Nos. JPMXP1222AE0019 and JPMXP1222AE0029). The work was also supported, in part, by the Los Alamos National Laboratory (LANL) Laboratory Directed Research and Development (LDRD) Program through Directed Research (DR) “Cathodes And Rf Interactions In Extremes (CARIE)” (Project No. 20230011DR). Studies were performed, in part, at the Center for Integrated Nanotechnologies, an Office of Science User Facility, operated for the U.S. Department of Energy (DOE), Office of Science. Los Alamos National Laboratory, an affirmative action equal opportunity employer, is managed by Triad National Security, LLC for the U.S. Department of Energy’s NNSA, under contract 89233218CNA000001. This research used resources provided by the Los Alamos National Laboratory Institutional Computing Program, which is supported by the U.S. Department of Energy National Nuclear Security Administration under Contract No. 89233218CNA000001.