Advances in actinide thin films: synthesis, properties, and future directions

In early attempts to fabricate bulk single crystals, actinide compounds such as oxides or alloys were placed in crucibles. When subjected to extreme temperatures and pressures, the required materials are evaporated from the solid source onto a substrate (figure 1(a)). The high melting points and chemical activity of actinide and rare-earth materials used in this early attempts presented a challenge for substrate and crucible material selection. Among the first reports attempting to grow actinide thin films was mineralization. In this approach, the crucible is placed on a rotating, water-cooled support, connected to a direct current (DC) power supply. A tungsten wire acts as a cathode, emitting electrons that weld a lid onto the crucible. Then, by heating the crucible very close to its melting temperature, single crystals can grow inside from the source material over a period of weeks [3, 11, 15]. In this way, Ac₂O₃ was chemically converted to a face-centered cubic (fcc) Ac metal powder, along with the presence of actinium hydride discovered through x-ray diffraction (XRD) [16]. Similar approaches have yielded samples of Th₃N₄ [17], UN [18], U₂N₃ [18], USb [15], UTe [15], and U metal [19]. Different transuranic elements can be synthesized into thin films by reducing their oxide compounds. These materials are loaded into a Ta crucible, together with a reducing agent. The crucible is then heated with DC filaments in order to evaporate the pure metal and condense it onto a cooler substrate outside of the crucible [19]. Pyrolitic decomposition has been used to thermally break down precursor molecules containing U and Am, and form thin films of these active species [20]. In general, these early methods introduced large numbers of impurities, and resulted in poor control over the crystal quality of each sample. Nonetheless they helped researchers improve their understanding of growth parameters and physical properties, paving the way for more innovative approaches.

The levels of radioactivity in some actinide materials make it impractical to handle of more than trace amounts in facilities not designed for this purpose. Dissolving these trace amounts in solution helps minimize the amount of the radioactive elements involved in single-crystal growth (figure 2(b)). One of the most successful approaches to solution deposition to date is polymer assisted deposition (PAD), where precursor molecules are made by binding the metal ions of interest to a polymer. As this polymer solution evaporates, the metal film grows. This technique allows researchers to bypass the need for vacuum equipment by controlling the growth process through the viscosity and homogeneity of the metal-bearing precursor polymers [21–24]. Electrodeposition is another solution-based approach. Atoms of the required metal are suspended in solution and mixed with a metallic stirrer that acts as the anode, while a suitable metal plate onto which the film will grow is kept near the cathode [25–27]. This approach has provided an opportunity to study the oxidation states in uranium oxides [22], and to grow thin films of uranium oxides on foils [28].

Zoom In Zoom Out Reset image size

Other solution-based approaches include hydrothermal synthesis, in use since 1839 when the first barium and strontium carbonate crystals were formed [29]. Pressure and temperature gradients are used to precipitate the desired material from a high purity feedstock (figure 2(e)). A mineralized solution (e.g., CsF in the case of UO₂) is often used to help transport the nutrient to the substrate where the thin film can crystallize and grow [30, 31]. The main difference between conventional solution growth and hydrothermal synthesis is the viscosity of the liquid. Several actinide compounds have been successfully synthesized using solution approaches, including ThO₂ [32, 33], UN₂ [23], UO₂ [8, 22, 23, 28, 30, 31, 34], U₃O₈ [23], UO₃ [21], UC₂ [23], NpO₂ [23, 24], and PuO₂ [23].

A key challenge associated with solution-based growth approaches is presence of trace amounts of solvent in the final samples. Although the quality of thin films prepared in the last few years is remarkable, further reductions in defect and impurity densities is still a goal for higher precision measurements.

Several types of physical vapor deposition (PVD) have been used to synthesize thin films of actinides and surrogate materials (figure 2(c)). Among these, sputtering and electron beam evaporation (EBE) have been most commonly used due to their comparatively low cost and ease of use. Different types of sputtering are used for the deposition of thin films depending on the physical properties of the targets required: radio frequency (RF), DC, high frequency, ion beam, and magnetron sputtering are among the most common. Multiple types can be combined to produce enhanced results in terms of faster deposition rates or more uniform films (figure 2(d)). EBE is commonly used for its ability to heat relatively small areas of a source to very high temperatures, allowing for high precision in deposition rates.

Reactive PVD occurs when a chemical reaction occurs between the deposited material and the atmosphere surrounding the substrate, often during an annealing step. In contrast, pulsed laser deposition (PLD) is a form of PVD where a high power, low bandwidth laser is used to melt, evaporate, and ionize material from the surface of a target [6].

Other techniques used mostly for rare-earth compounds include chemical vapor deposition (CVD) [35, 36] and molecular beam epitaxy (MBE) [37–42]. There have been numerous successful research efforts to use rare-earths as surrogates for actinide materials [43–47]. Similarities in their densities, electronic structure, and high melting temperatures give researchers an opportunity to develop equipment and growth regimes that can be translated from the rare earths to synthesize actinide thin films.

Thin films of actinium metal have been produced since 1953, while thin films of metallic protactinium were synthesized through electrodeposition as early as 1964 [27]. These metallic thin films were used for α-spectrometry and nuclear reaction experiments. Following this work, deposition of thin films of protactinium pentoxide (Pa₂O₅) by electro-spraying was achieved [26].

Reports of thorium-based thin film properties remain relatively scarce due to the absence of a 5f electron orbital. Instead, most attention has focused on thorium's bulk properties as a candidate fuel for nuclear reactors. With this objective in mind, films of ThO₂ have been synthesized using sputtering [48, 49], photochemical deposition [32], and EBE [50]. Uno et al used a nitridation approach to synthesize Th₃N₄ [51]. Gouder et al prepared thin films of ThN and Th₃N₄ using sputter deposition, where photoemission spectroscopy demonstrated a high density of 6d states in ThN and a non-metallic character in Th₃N₄ [52].

UN thin films have been grown via reactive and DC magnetron sputtering using a Nb buffer layer on the $(1\bar{1}02)$ facet of sapphire [53], as well as on Si(111), polycrystalline Ta foil and glass substrates [54]. Black et al used the same set of substrates [Si(111), polycrystalline Ta foil, and glass] in attempts to grow U₂N₃ [54]. Scott et al synthesized UN₂ on LAO(001) substrates using PAD [23]. Long et al also achieved the growth of both of these compounds by nitriding U metal flakes [55]. Gouder grew this compound on CaF₂ using DC magnetron sputtering [56]. Adamska et al grew UMo films on a Nb(110) buffer layer on a sapphire substrate as well as U–Zr, U–Mo, and U–Nb [57]. The study by Black et al also included thin films of UC₂ grown on YSZ(110) and U₃O₈ in different crystal structures that depended on the orientation of the sapphire substrates [54]. Thin films of uranium oxicarbides and oxinitrides were prepared on Si(111) substrates by Eckle and Gouder using reactive DC sputtering [58]. Two compounds of uranium oxide were grown using PAD on LSAT(100) substrates: α-UO₃ and α-U₃O₈ [21].

Several attempts at synthesizing UO₂ thin films have been reported with various results. The first reported attempt was by Bierlein and Mastel in 1960, where they studied the effects of irradiation by growing UO₂ on carbon substrates [59]. Elbakhshwan and Heuser summarized the evolution of UO₂ thin film growth up to 2017 [60], and discussed the use of CVD and sputtering techniques (see table 1). Their contribution also includes growth on several different substrates using magnetron sputtering. Further advances in this field since then include reactive sputtering on LSAT(110) [61, 62], magnetron sputtering on YSZ(100) [63], DC sputtering on SrTiO₃(110) [64], PAD of UO_2+x by Zhang et al [65], and spin coat combustion on aluminum sheets by Roach et al [66]. U_x Th_1−x O₂ thin film synthesis was achieved by Cakir et al to study surface reduction using ice [67].

Work on superconducting materials UNi₂Al₃ and UPd₂Al₃ realized the first MBE synthesis of actinide containing materials [37–40]. Using electron beam evaporating sources for this materials, depleted U (with no specified purity) was deposited on a heated substrates (listed in table 1) at temperatures of about 700 °C.

Synthesis of transuranic compound thin films has been historically challenging due to the high toxicity and scarcity of material precursors [68]. An early attempt to produce thin films of PuO₂ is reported by Shaw et al [69]. More recently, Wilkerson et al summarized the state-of-the-art and reported the growth of thin films using PAD on YSZ substrates [70]. Roussel discussed the sputtering of PuO₂ and α-Pu₂O₃ films used to study inverse photoemission [71]. Subsequent attempts to grow bulk single crystals of these oxides used solution based approaches [72–74]. Other Pu-based single crystal compounds synthesized include PuAs, PuSb, and PuBi for studies of anisotropic magnetization [3]. Mannix et al used the same fabrication method to grow NpS and PuSb single crystals to investigate their magnetic structure [75]. Durakiewicz et al investigated the spectral features of USb, NpSb, PuSb, NpTe and PuTe using angle-resolved photoemission spectroscopy (ARPES) on samples grown using a mineralization technique [15]. Gouder et al grew thin films of PuSe and PuSb for study using ultraviolet photoelectron spectroscopy [76], and thin films of PuSi_x (with x varying from 4 to 0.5) [56]. A particularly successful research program involved PAD of PuO₂ and NpO₂ with high enough crystalline quality that they enabled studies of band dispersion and the Fermi surface, and helped validate several characteristics used in 5f electron modeling approaches [23]. NpO₂ was grown on LAO(001) substrates using a UO₂ buffer layer. Elements beyond Np have scarcely been fabricated in sufficient amounts to attempt high quality crystal growth. Through a method of oxidation, Radchenko et al managed to fabricate high quality bulk samples of Am, Cm, Bk, and Cf on a La crucible [19].

We can divide materials characterization techniques into categories that deal with different properties of the thin films in question. Figure 2 illustrates how one can use various techniques to explore structural, surface and electronic properties together in concert.

We can obtain plan view or cross-sectional structural images of a thin film using transmission electron microscopy, scanning electron microscopy (SEM) (see figures 2(c) and (h)), atomic force microscopy, and scanning tunneling microscopy (STM). These can be used to determine the size of crystallites, the density of defects such as dislocations twins and stacking faults, and crystalline quality, in addition to quantifying the roughness of the surfaces and the shape, size and areal density of any structures present on them. 'Indirect' structural characterization may include Rutherford back scattering (RBS), XRD, and Raman spectroscopy (see figures 2(a), (d)–(f), (i) and (k)). These techniques can provide information that complements what can be seen in microscopy, and help identify properties such as alloy composition, lattice parameter, and in the case of ion-damaged samples, penetration depth.

Several techniques have been used to measure the thermophysical properties of bulk actinide materials [7]. However, thin films have been targeted in specific cases as a means to increase measurement accuracy while using non-destructive techniques. It is here where photothermal approaches offer a viable option for studying samples with small physical dimensions without damage to the sample or the need for destructive sample preparation approaches [105]. Several comprehensive reviews have been recently published on the thermophysical properties of actinide oxides [7], nitrides [106], and carbides [107, 108] due to the importance of these compounds to nuclear science and industry.

Electronic structure characterization includes angle-resolved photemission spectroscopy (ARPES), x-ray photoemission spectroscopy (XPS), STM-based scanning tunneling spectroscopy (STS), and photoluminescence spectroscopy. These approaches can shed light on charge carrier behavior, and the electronic states present in a given material. These characteristics are of particular interest for actinide thin films due to the interesting physics governed by their 5f electron states. XPS allows one to determine the elements present within a material, along with their oxidation and bonding state. XPS is especially useful for samples with reduced dimensionality within which the localization of 5f electrons occurs [109]. Gouder and Havela identified the complications of applying XPS to the study of actinides: an asymmetrical tail in metallic systems can often prove difficult to differentiate from scattering of electrons onto the Fermi level. In addition, the anisotropic distribution of atoms at different temperatures contributes to higher noise levels in the XPS signal [109]. The team lead by Black et al identified the 5f states of UN around 0.8 eV [54]. Havela et al indicate that 5f localization behavior can be studied in PuSb (localized), PuSe (intermediately localized), and PuN (potentially delocalized) [96]. In addition, Gouder et al studied the localization of 5f electrons in PuSi_x [89]. Durakiewicz et al investigated the systematic changes in the 5f binding energy peak between antimonide and tellurides of U, Pu, and Np [15]. Joyce et al studied the dual nature of 5f electrons using photoemission of δ-Pu, PuIn₃, and PuCoGa₅ [110, 111]. On the same line, Eloirdi et al reported results from XPS studies of PuCoGa₅ deposited using sputtering [112]. Long et al studied the electronic structure of α-U₂N₃ and found hints of metallic bonding [81]. Pentavalent uranium was found in U₂O₅ by Gouder et al using photoemission studies [113].

The ARPES technique has been used to determine precise band structures and Fermi surfaces of actinide-based single crystals [23, 98, 114–137]. Largely focused on bulk actinide compounds, these ARPES studies have created new opportunities for determining the origin of the unusual physical properties of 5f-electron materials. Going forward, ARPES-based studies of epitaxial actinide thin films could lead to the discovery of a wide range of emerging electronic, magnetic, and structural phenomena (figure 2(j)). Interactions between the epitaxial thin film and the underlying substrate can produce properties that are substantially different from those of the bulk materials. For instance, it has been reported that one can obtain an unusual hexagonal close-packed phase of uranium (hcp-U) by depositing U thin films onto a W (110) substrate [138, 139]. Theoretical calculations predict that the hcp-U phase has an electronic instability, that could lead to a possible charge density wave or magnetic ordering [140]. Different orientations of the α-U orthorhombic phase can also be obtained by depositing U onto a variety of buffer/seed layers on sapphire substrates [92, 141]. ARPES was used to obtain the band structure of α-U single crystals at 173 K [142]. The valence band structure of α-U was studied in details using ultraviolet photoemission spectroscopy and XPS [143],

For ordered overlayers of U metal grown on a W (110) substrate, band-like properties of the U 5f states were observed, which were proposed to arise from direct f–f electron interactions [138]. STM/STS results showed that the U density of states close to the Fermi level is dominated by the 5f states [139]. STS and ARPES showed both the Fermi surface topology and electronic structure of ordered α-U films [144]. ARPES studies of epitaxial thin films of the oxides PuO₂ and NpO₂ [24, 145], and uranium carbide (UC₂), reveal the importance of the underlying substrate structure in stabilizing the epitaxial film [146]. But in general, compared with bulk actinide crystals, only a few ARPES studies have been performed on actinide thin films. Therefore, the use of ARPES to measure novel actinide thin films synthesized with higher quality will produce results that are critical to understanding their reactivity, aging, nuclear fuel behavior, and environmental fate, as well as the potentially rich physics of these exotic 5f materials.

Thermophysical properties play a central role in determining the characteristics of actinide materials. Atomic structure dictates macroscopic behavior, and it is hence important to understand the temperature dependence of these properties. A key research driver in actinide compounds is the qualification process for nuclear fuel materials. This process is heavily reliant on characterizing thermophysical properties: heat capacity, thermal conductivity, elastic constants, and thermal expansion [7, 147]. Heat capacity has been measured in several important actinide dioxides, such as ThO₂ [105], UO₂ [148], NpO₂ [149], PuO₂ [150]. UO₂ and NpO₂ show an interesting magnetic alignment transition that creates anomalies in their heat capacity. Thermal conductivities have been measured for ThO₂ [151], UO₂ [152], NpO₂ [153], and PuO₂ [154]. The heat capacities of select actinide nitride systems have been measured, including ThN [155], and more recently for the UN–ThN mixtures used in nuclear fuels [106].

Among these properties, thermal diffusivity and conductivity have been recently studied as important factors in determining the properties of a nuclear fuel candidate material [7]. The qualification process requires that intrinsic thermophysical properties be resolved independently of any particular specimen geometry, without the competing influence of lattice defects or other chemical or structural imperfections. Building a fundamental understanding of intrinsic properties allows for the effects of these additional complexities, for example the types of structural defects generated under exposure to irradiation, to be systematically included in computational frameworks for engineering-scale performance. Much success to date has been found in the use of non-destructive photothermal techniques to measure these intrinsic thermophysical properties [105]. Such methods broadly consider microscale thermal and acoustic wave propagation, which can be measured via surface deformation, refractive index gradient, acoustic expansion/contraction, or changes in optical reflectivity (thermoreflectance). This last approach can be further subdivided into frequency-domain thermoreflectance, time-domain thermoreflectance, spatial-domain thermoreflectance, and hybrid methods [147, 156]. With small sampling volumes, these methods are broadly applicable to the study of thin films [157]. Although they have not yet been applied to actinide thin films, we anticipate that these techniques will help to shed light on their thermophysical properties in the future.

Heterostructure devices that integrate different functional layers can encounter performance issues as a result of restricted heat dissipation across the heterointerfaces (Kapitza resistance). This restriction in heat dissipation correlates with factors such as interfacial roughness, disorder, dislocations, and bonding. As a result, computational modeling of these interfaces often fails to capture all of their vibrational properties, for example changes in mass density, phonon mean free path, and carrier scattering across the interface [158] (figure 3).

Zoom In Zoom Out Reset image size

Different aspects of electron and phonon behavior have been carefully studied under the influence of magnetic fields. Among these, transport and its different manifestations enabled the observations of several noteworthy phenomena. These physical properties are part of what makes actinides interesting to researchers beyond their nuclear applications [159]. Most transition metal elements have itinerant d electrons, while lanthanide elements exhibit mostly localized 4f electrons. These facts help us distinguish characteristics related to the conduction band of a material, and its magnetic properties relating to the 4f electrons [159]. Due to the scarcity, high cost of ²²⁷Ac, and intense radioactivity single crystal samples of this metal have not been synthesized to date, making it difficult to study its magnetic properties experimentally [160]. A similar situation exists for the actinide elements after californium.

Considering the intervening elements, Th is a paramagnetic metal with three phases that are heavily influenced by the concentration of impurities present [1]. Pa metal was synthesized relatively early on in the development of actinide research, and it was hence possible to determine its phase transition and superconducting behavior [1], as discussed below. The early actinide metals (U, Np, Pu) show no sign of magnetic ordering mostly because the bandwidth of the 5f electrons in these elements is too wide to satisfy the Stoner criterion [161]. U metal exhibits a weak Pauli-type paramagnetism, with α-U having a charge-density-wave anomaly around 43 K that has been observed in magnetic susceptibility measurements [1]. Np exhibits spin–orbit coupling in the same order as other crystal-related electric effects, leading to its metallic form and paramagnetic behavior [1]. Elemental Pu can be found in both α- and δ-Pu phases. Recent studies point to a lack of evidence favoring any ordered or disordered magnetism in δ-Pu [162], which preferentially forms a valence fluctuation ground state [163]. Am metal shows a non-magnetic ground state coming from the 5f⁶ electronic configuration (J = 0). Curium (Cm) is the first element in the actinide series that orders magnetically. Cm is antiferromagnetic (AF) below 65 K in the double hexagonal close packed (dhcp) phase, while in the fcc phase it is ferromagnetic (FM) below 205 K [164]. Berkelium (Bk) also exists in these two crystallographic phases (dhcp and fcc) and similarly orders either AF or FM. Californium (Cf) also exhibits a large effective magnetic moment, similar to Cm and Bk, and likely has a FM ground state [165].

A variety of thin films of actinide compounds have provided insight into the magnetic properties of these systems, such as the magnetic structures of NpS and PuSb [75], itinerant antiferromagnetism [77] and magneto-optical Kerr effect [166] in UN, weak ferromagnetism [167] and antiferromagnetism [82] of UO₂, and the high pressure behavior of electrical resistance in NpAs and NpBi [168]. Single crystal compounds have shown to have different critical temperatures for magnetic reordering [37], and special epitaxial compounds such as UNi₂Al₃ have demonstrated superconductivity at 0.97 K [40]. Magnetic susceptibility measurements of transplutonium elements Am, Cm, and Bk helped determine the effective magnetic moments and Curie–Weiss constants for some of their compounds [169]. Magnetic properties have also been extensively studied in these materials using computational methods ([170] and references therein). Bright et al studied the magnetic and electronic structure of UN and U₂N₃ epitaxial films using x-ray scattering [171]. Recently, using high-field magnetization and magnetostriction measurements, the presence of a metamagnetic transition and a tri-critical point was observed in UN single crystals, leading to the development of a magnetic phase diagram for this material [172, 173]. Using micro-structured single crystals, Hamann et al found strong reconstruction of the Fermi surface at the high-field transition [174].

Certain actinide metals are known to exhibit superconductivity at low temperature. These include Th at 1.4 K [175], Pa at 1.4 K [176, 177], α-U at 0.7 K and γ-U at 1.8 K [2, 178–180]. Am was predicted to be a superconductor, although confirming this experimentally was a challenge due to its large self-heating, scarcity, and radioactivity. However, this prediction was confirmed by Smith and Haire; when its fcc phase is stabilized by impurities, Am becomes superconducting at 1.05 K, and pure dhcp Am is superconducting at 0.79 K [181]. α-Np and α-Pu have no reported superconducting signature down to 0.5 K [182] and Cm, Bk, and Cf are not expected to show superconductivity because of their magnetic ground-state [182]. There are several actinide compounds that show heavy-fermion superconductivity, including UBe₁₃, UPt₃, URu₂Si₂, UNi₂Al₃, UPd₂Al₃, NpPd₅Al₂, PuCoIn₅, PuRhGa₅, PuRhIn₅ and PuCoGa₅ [98, 183–190]. Ott et al showed that UB₁₃ is superconducting below 0.85 K [191]. Joyce et al studied photoemission data from PuCoGa₅, PuIn₃, and δ-Pu metal. A comparison between their results and model studies indicate characteristics of both itinerant and localized Pu 5f electrons [110]. Furthermore, Joyce compared T_c in PuCoGa₅ against UCoGa₅ and found no evident characteristics of strongly correlated electrons [111]. Thus, valence fluctuations have been proposed as an alternative mechanism for the high-temperature superconductivity observed in PuCoGa₅ [192, 193]. Recently, unconventional [194], spin-triplet [195] superconductivity with multiple superconducting phases [196] has been reported in UTe₂. In addition, there are indications for spontaneously broken time-reversal symmetry [197], and chiral Majorana edge and surface states [198], the nature of which are not yet understood. Brun, Cren, and Roditchev summarized the superconductivity of 2D materials as a strong case for epitaxial monolayer materials growth [199]. Huth et al demonstrated deviations from typical heavy-fermion material characteristics in superconducting UPt₃ [91], attributed to the hybridization of 5f electrons and itinerant states, although the nature of the pairing mechanism is still under debate. Superconductivity was also observed in UPd₂Al₃ [38] and UNi₂Al₃ [40], while the optical conductivity of this last compound was studied by Ostertag et al using terahertz frequencies [200].

The study of conducting states in topologically nontrivial materials is often hindered by the presence of bulk conducting states [201]. High quality thin films offer one way of minimizing bulk contributions by increasing a sample's surface area-to-volume ratio, simplifying measurements of these topological effects. Due the nature of their electronic structure, actinide materials represent a fascinating, albeit challenging area for topological discoveries. Promising technological applications and insights into fundamental physics continue to motivate investigations on the complex manifold of 5f, 6p, 6d, and 7s orbital shells [202]. These interactions, present in an analogous manner in the rare earths, are expected to give rise to a variety of topological behaviors in different compounds and geometries.

Heavy fermion materials owe their particular properties to the competition between Coulomb repulsion of localized f electrons, and hybridization with itinerant d electrons. These electron correlations produce complex states including Mott physics, superconductivity, and correlated magnetism [203]. Understanding these electronic systems could potentially help elucidate other strongly correlated electron materials, for example high-temperature superconductors, multiferroics, and magnetoresistive materials. Topological insulators possess conducting surface states protected by the non-trivial topological property of spin-momentum locking. Several quantum materials have been proposed based on topological properties that involve quantum phenomena such as Majorana fermions, the anomalous quantum Hall effect, with applications in spintronics and fault tolerant quantum computation. Magnetic topological semimetals have been synthesized using MBE [204, 205]. These materials have a FM ground state that breaks time-reversal symmetry with a gap in the topological surface states [206]. A powerful tool to study these phenomena is density functional theory (DFT), e.g. its use in identifying phases in UNiSn corresponding to a topological insulator and a Weyl semimetal depending on magnetic ordering [207]. Rocksalt compounds of Pu and Am were predicted to form a topologically insulating ground state due to their 5f electron being right on the boundary between localized and itinerant, and coupling between electronic and spin–orbit interactions gives rise to band gap formation [208]. The predicted compounds that Zhang et al used for this study include americium monopnictides and plutonium monochalcogenide [208]. Using nuclear magnetic resonance, Dioguadi et al demonstrated that PuB₄ is a strong candidate for a topological insulator, which has also been suggested by Choi et al [209, 210]. Recent theoretical work suggests that the intermediate valence PuB₆ is a strong topological insulator with nontrivial Z₂ topological invariants [211], similar to SmB₆ [212].

Overall, much work is still to be done in understanding fully the properties of these actinide compounds, and high quality thin films promise to be the missing piece to provide a window into their discovery.