MoO3 films were grown on stepped c-plane sapphire substrates by molecular beam epitaxy using MoO3 vapor from a conventional Knudsen cell. Stepped sapphire (0001) substrates were prepared by ex situ annealing at 1100–1300 °C in dry air. Step bunching typically resulted in multistepped surfaces with wide atomically smooth terraces. Ex situ annealing at 1100 °C followed by in vacuo annealing at 700 °C provided clean substrates for growth. Ultrathin films were grown at 450 °C via a self-limiting process that represents a balance between the incident MoO3 flux and the desorption flux. Elongated bilayer islands (0.7-nm thick) were formed on sapphire (0001) terraces. Monocrystalline α-MoO3 (010) thin films [(010)α-MoO3∥(0001)sapphire] were grown at 450 °C using a higher incident MoO3 flux and characterized by atomic force microscopy, x-ray photoelectron spectroscopy, x-ray diffraction, and cross-sectional transmission electron microscopy. The step-terrace surface morphology of the monocrystalline films strongly suggests multilayer growth.

MoO3 has a variety of technological applications, including heterogeneous catalysis,1–4 gas sensing,5 advanced Li-based batteries,6 and electronics.7–9 Such a wide range of applications is possible thanks to its unique structural and electronic properties. The morphology of MoO3 films has been investigated extensively and numerous preparation techniques have been reported—including, but not limited to, molecular beam epitaxy (MBE),10 precipitation with hydrotreatment,11,12 electrodeposition,13 and reactive sputtering.14 Based on the preparation conditions, the resulting deposits form a film made of particles of various shapes ranging from microballs15 and nanoribbons5,11,16 to nanoflowers.12 Film stoichiometry is also very important, because oxygen deficiency has a profound effect on the properties of the layer, such as catalytic activity and selectivity.17 Through oxygen deficiency, one may also adjust the bandgap, which is crucial for many electronics applications.7,9,18 Ultrathin MoO3 films (2D materials) with a variety of novel properties and potential new applications have been prepared by ambient-pressure growth techniques19,20 and by MBE.21 In this work, we were primarily interested in ultrathin MoO3 films on sapphire as model catalysts for oxidative dehydrogenation of light alkanes.1–4 

Orthorhombic α-MoO3 (space group Pbmn; a = 0.396, b = 1.396, and c = 0.370 nm), the most thermodynamically stable polymorph of molybdenum trioxide, has a unique layered structure (Fig. 1).22 Bilayer (BL) sheets comprised of distorted MoO6 octahedra linked via corner and edge sharing in the a-axis and c-axis directions, respectively, are stacked in the b-axis direction. The 0.7-nm thick sheets are held together weakly, resulting in so-called van der Waals gaps. These BL sheets are sometimes called MoO3 monolayers in the materials science literature.20,21

FIG. 1.

Crystal structure of α-MoO3. The orthorhombic unit cell, van der Waals (vdW) gaps, and bilayer (BL) thickness are shown.

FIG. 1.

Crystal structure of α-MoO3. The orthorhombic unit cell, van der Waals (vdW) gaps, and bilayer (BL) thickness are shown.

Close modal

The c plane is the most thermodynamically stable crystal face of sapphire;23 however, it forms steps and terraces upon heating at >1000 °C in air.24–27 Moreover, the surface smoothens via atom migration, removing mechanical polishing defects, e.g., scratches. The resulting step-terrace morphology depends on the miscut angle and its orientation relative to the c plane. A small miscut (<0.1°) is typically unavoidable during wafer manufacturing. Single atomic steps (monosteps) form at lower annealing temperatures (1000–1100 °C) on crystals with low miscut angles (<1°). A single atomic step on sapphire (0001) is 0.22-nm high, which equates to the distance (c/6) between two oxygen atom planes in the [0001]-direction. With increasing the annealing temperature and time, steps grow via bunching and terraces broaden, leading to multistepped surfaces. A coalescence of nonparallel (meandering) steps results in cross-stepped surfaces.

Sapphire substrates with steps and atomically smooth terraces can be advantageous for heteroepitaxial growth.25,28 Atomically smooth terraces facilitate layer-by-layer growth and an abrupt interface between a film and a substrate. Moreover, “funneling” of monomers from terraces toward ascending steps can promote two-dimensional (2D) growth modes.28,29 For example, in LiNbO3 deposition, vicinal sapphire (0001) substrates with 10-nm steps provided smooth films with a preserved step-terrace structure. In contrast, LiNbO3 films deposited on nominal sapphire (0001) surfaces with 0.2-nm monosteps had rough surfaces without steps.28 

MBE has the capability to produce MoO3 films in vacuo with simple and precise control over deposition parameters. High-purity MoO3 powder sublimes readily and congruently above ∼500 °C; the primary vapor species is the trimer (MoO3)3.30 Koike et al.10 employed a conventional Knudsen cell at temperatures of 570–620 °C to grow MoO3 films on c-plane sapphire at rates of 80–100 nm/h. Films deposited at substrate temperatures <350 °C were either amorphous or β-MoO3, and films grown at 350 °C were α-MoO3; film sublimation began to limit growth at 400 °C. Du et al.21 reported self-limiting growth of ultrathin epitaxial α-MoO3 (010) films on SrTiO3 (001) by MBE at 450 °C. Self-limiting growth occurs when the incident MoO3 flux and the desorption flux are equal, and interfacial bonding stabilizes the first bilayer of MoO3 on the substrate. We previously reported a deposition of α-MoO3 (010) films on sapphire (0001) by MBE at 400 °C; however, sublimation began to limit film growth at higher temperatures.31 MoO3 deposited on sapphire (0001) at 580 °C forms 3D islands surrounded by bare substrates. X-ray photoelectron spectroscopy (XPS) demonstrated that these islands are thermally stable due to their oxygen deficiency (average composition MoO2.7).

Because the α-MoO3 (010) surface has no dangling bonds, heteroepitaxial growth on completely lattice-mismatched substrates can be achieved by van der Waals epitaxy, provided the substrate surface is atomically smooth and chemically inert. For example, a growth of α-MoO3 (010) thin films via van der Waals epitaxy on 2D substrates, including mica32 and graphene,22 has been reported. Surface termination of stepped c-plane sapphire by OH groups may facilitate van der Waals epitaxy of α-MoO3 films. Previously, only a very weak interaction between an α-MoO3 film deposited by MBE at 400 °C and the sapphire (0001) substrate was demonstrated based on in vacuo annealing experiments.31 

In this paper, we report α-MoO3 films grown by MBE on stepped c-plane sapphire substrates. Growth modes and film morphology were investigated with respect to substrate temperature, deposition time, MoO3 flux, sapphire step morphology, and surface cleanliness. The films were characterized by atomic force microscopy (AFM), x-ray photoelectron spectroscopy (XPS), x-ray diffraction (XRD), and cross-sectional transmission electron microscopy (TEM) on a focused ion beam (FIB)-prepared specimen.

Polished sapphire (0001) substrates (1 × 1 cm2) were cut from a 3-in. sapphire (0001) wafer (Roditi International). The substrates were rinsed and sonicated in isopropanol for 15 min prior to use. For preparing stepped surfaces, the substrates were placed in an alumina boat inside a clean alumina tube (MTI Corporation) in flowing dry air and annealed in a tube furnace at 1000–1300 °C for 4–24 h.

A detailed description of the multichamber MBE apparatus was provided in a previous publication.31 Briefly, the apparatus consists of a cluster of UHV chambers that are connected via a load lock allowing sample introduction and in vacuo transfer between the chambers. Each sapphire substrate was glued to a Mo sample holder with silicate-based carbon paste (Ted Pella) and cured in air as recommended by the manufacturer (room temperature overnight, 110 °C for a day, and 250 °C overnight before loading the sample). After loading into the apparatus, the substrate was outgassed and cleaned by annealing to 700 °C in 5 × 10−6 Torr O2 with a 30 min hold; sample heating was done through radiation from the back portion. The substrate temperature was measured with a K-type thermocouple in contact with the Mo sample holder and/or by pyrometer for temperatures >550 °C. A conventional Knudsen effusion cell (SVT Associates) was loaded with MoO3 (99.95%, Alfa Aesar). The effusion cell temperature was maintained at 615 or 640 °C during the MBE growth experiments.

After deposition, the sample was cooled to ambient temperature and the surface structure was checked by in situ using reflection high-energy electron diffraction (RHEED). After in vacuo transfer (<10−8 Torr), the samples were analyzed by XPS using a PHI 3057 instrument with a dual anode x-ray source using the same conditions as published elsewere.31 Binding energies were referred to as the Al 2p peak at 74.1, or alternatively, as the O 1s peak at 531.1 eV. The latter was used only when necessary (i.e., for thicker films). The XP spectra were analyzed using casaxps software. Ex situ AFM, XRD, and cross-sectional TEM were performed on the selected samples. The sapphire substrates and deposited films were imaged via AFM using a Digital Instruments Dimension 3000 scanning probe microscope using a Si tip in tapping mode. XRD patterns were acquired with a Rigaku SmartLab x-ray diffractometer with a Cu Kα source (λ = 0.1542 nm) operating at 40 kV and 44 mA. The range of 10−80° (2θ) was measured in step mode with a 0.05° increment and 3 s dwell time. A continuous epitaxial film was cross-sectioned via focused ion beam (FIB) using an FEI Quanta 3D FEG instrument and imaged with FEI Talos F200× at the Analytical Instrumentation Facility of North Carolina State University.

Representative AFM images of the sapphire (0001) substrates used in this work are shown in Fig. 2. The as-received sapphire wafer [Fig. 2(a)] has scratches due to chemical-mechanical polishing; however, the mirror-polished surface is smooth with a height variation of 1nm (Fig. 3) and root-mean-square (RMS) surface roughness of 0.19 nm. After annealing at 900 °C for 4 h, the surface remains relatively smooth [Fig. 2(b)], and the polishing scratches are not as prominent. In agreement with the literature,24–27 annealing at 1100 °C for 2 h [Fig. 2(c)] led to the development of a step-terrace morphology with relatively small meandering steps with heights corresponding to 1–3 monosteps (MS) [or i × 0.22 nm (i = 1–3)]. Many step crossings (intersections) are observed creating a surface morphology resembling fish scales with 100–150 nm wide terraces (measured at the widest point) that are 150–350 nm long. The corresponding AFM step profile (Fig. 3) is jagged and irregular. After annealing at 1100 °C for 24 h, the steps had grown and the terraces had widened. The resultant surface [Fig. 2(d)] comprises large meandering multilayer steps (multisteps) [i × 0.22 nm (i = 4–7)] delineating atomically flat terraces: 300–500 nm wide × 700–3000 nm long. The corresponding step profile (Fig. 3) suggests the formation of facets via step bunching. The substrates whose AFM images are shown in Figs. 2(d) and 2(e) were annealed in a single batch in the same alumina boat under equivalent conditions, but the resulting surface morphologies are quite different. The substrate in Fig. 2(e) shows a regular pattern of parallel steps without crossings or step coalescence points. The uniform steps are 2 MS high [i × 0.22 nm (i = 2)] with 80–220 nm wide terraces (mostly in the 100–120 nm range). The miscut angle relative to the (0001) plane was estimated from the step heights and terrace widths as ∼0.16°—which is within the tolerance quoted by the manufacturer. Other samples annealed in the same batch had surface morphologies between these two extremes. We suggest that the differences stem from a furnace temperature gradient, because the substrate surface morphology changed gradually in the same order as the wafers were arranged in the alumina boat. Finally, sapphire substrates annealed at 1300 °C for 24 h [Fig. 2(f)] have large well-defined steps and terraces with fewer crossing points. The step heights range from 3 to 9 MS [i × 0.22 nm (i = 3–9)] (with most being 7–8 MS); the atomically flat terraces are 250–500 nm wide (Fig. 3) and >1 μm long.

FIG. 2.

AFM images of sapphire (0001) substrates: (a) as-received; (b) annealed in air at 900 °C, 4 h; (c) annealed in air at 1100 °C, 2 h; (d) annealed in air at 1100 °C, 24 h; (e) annealed in air at 1100 °C, 24 h; (f) annealed in air at 1300 °C, 24 h.

FIG. 2.

AFM images of sapphire (0001) substrates: (a) as-received; (b) annealed in air at 900 °C, 4 h; (c) annealed in air at 1100 °C, 2 h; (d) annealed in air at 1100 °C, 24 h; (e) annealed in air at 1100 °C, 24 h; (f) annealed in air at 1300 °C, 24 h.

Close modal
FIG. 3.

AFM step profiles along the lines depicted in Fig. 2. For clarity, lines D and E have been offset by adding 1, and 2.5 nm, respectively.

FIG. 3.

AFM step profiles along the lines depicted in Fig. 2. For clarity, lines D and E have been offset by adding 1, and 2.5 nm, respectively.

Close modal

Carbon contamination was detected by XPS on the as-received and ex situ annealed sapphire substrates (Fig. 4). The C 1s binding energy (BE) at 284.8–285 eV is typical of adventitious carbon (C–C type).33 The C 1s peaks of the as-received and ex situ annealed sapphire substrates (1100 °C, 24 h) also have higher BE shoulders (>286 eV), suggesting C–O–C type carbon contamination.33 After in situ annealing to 700 °C in a 5 × 10−6 Torr O2 ambient, the surface carbon concentration dropped to 2.5 at. % on the as-received sapphire and to barely detectable levels on ex situ annealed sapphire. The quantitative results (Table I) demonstrate that ex situ annealing at 1100 °C in air, followed by in situ annealing in vacuo to 700 °C, results in an essentially carbon-free surface for MoO3 film growth.

FIG. 4.

C 1s XP spectra of sapphire substrates: (a) as-is and after annealing for 30 min in vacuo at 700 °C in 5 × 10−6 Torr O2; (b) post ex situ annealing at 1100 °C, 24 h in dry air and after subsequent annealing for 30 min in vacuo at 700 °C in 5 × 10−6 Torr O2. Spectra were baseline subtracted and normalized by Al 2p peak area.

FIG. 4.

C 1s XP spectra of sapphire substrates: (a) as-is and after annealing for 30 min in vacuo at 700 °C in 5 × 10−6 Torr O2; (b) post ex situ annealing at 1100 °C, 24 h in dry air and after subsequent annealing for 30 min in vacuo at 700 °C in 5 × 10−6 Torr O2. Spectra were baseline subtracted and normalized by Al 2p peak area.

Close modal
TABLE I.

Surface compositions of sapphire substrates as determined by XPS.

PretreatmentElemental composition (at. %)a
AlOCO/Al
As-received 33.8 60.6 5.6 1.8 
In situ annealing, 700 °C, 30 min 39.5 58.0 2.5 1.5 
Ex situ annealing, 1100 °C, 24 h 27.0 65.7 7.3 2.4 
Ex situ + in situ annealing 39.6 60.4 0.0 1.5 
PretreatmentElemental composition (at. %)a
AlOCO/Al
As-received 33.8 60.6 5.6 1.8 
In situ annealing, 700 °C, 30 min 39.5 58.0 2.5 1.5 
Ex situ annealing, 1100 °C, 24 h 27.0 65.7 7.3 2.4 
Ex situ + in situ annealing 39.6 60.4 0.0 1.5 

aBased on Al 2p, O 1s, and C 1s spectra.

The O 1s spectra of the as-received and ex situ annealed sapphire substrates [Figs. 5(a) and 5(c)] contain two components separated by 1.6 eV. The 531.1-eV BE component is assigned to oxide ions (O2−) and the 532.7-eV component primarily to surface OH groups.34 The ex situ annealed sample has a higher percentage of the 532.7-eV component and a higher O/Al ratio (Table I) than the as-received sapphire. After subsequent in situ annealing at 700 °C [Figs. 5(b) and 5(d)], the relative intensity of the 532.7-eV component and the O/Al ratio decrease significantly for both samples consistent with the removal of surface OH and carbonate groups.35 The final surface OH percentages are similar, and the O/Al ratios are equivalent to the bulk stoichiometry.

FIG. 5.

O 1s XP spectra of sapphire substrates: (a) as-received, (b) sample after heating to 700 °C in 5 × 10−6 Torr O2 for 30 min, (c) sample annealed ex situ at 1100 °C, 24 h in dry air, (d) ex situ annealed sample after heating to 700 °C in 5 × 10−6 Torr O2 for 30 min. Spectra normalized to maximum counts.

FIG. 5.

O 1s XP spectra of sapphire substrates: (a) as-received, (b) sample after heating to 700 °C in 5 × 10−6 Torr O2 for 30 min, (c) sample annealed ex situ at 1100 °C, 24 h in dry air, (d) ex situ annealed sample after heating to 700 °C in 5 × 10−6 Torr O2 for 30 min. Spectra normalized to maximum counts.

Close modal

MBE growth of α-MoO3 thin films from the vapor is very sensitive to substrate temperature. AFM images of MoO3 films deposited on stepped sapphire (0001) substrates at 350, 450, and 475 °C are shown in Fig. 6. For each film, the K-cell temperature and deposition time were 640 °C and 15 min, respectively. Films deposited at 350 [Fig. 6(a)] and 450 °C [Figs. 6(b)6(e)] are very rough with micrometer-scale surface features. The latter is characterized by larger facets. Sublimation becomes so significant at 475 °C that only a single incomplete BL of MoO3 is deposited with small 1–2 BL high islands on top [Fig. 6(f)]. The surface morphology of the sapphire is visible beneath the film. Thus, at this incident MoO3 flux, the substrate temperature should be kept below 475 °C in order to deposit continuous multilayer films.

FIG. 6.

AFM images of MoO3 films deposited on stepped sapphire (0001) using a K-cell temperature of 640 °C and 5 × 10−6 Torr O2 ambient. Deposition time: 15 min. Substrate temperature (corresponding substrate AFM image): (a) 350 °C [Fig. 2(d)]; (b) 450 °C [Fig. 2(c)]; (c) 450 °C [Fig. 2(d)]; (d) 450 °C [Fig. 2(e)]; (e) 450 °C [Fig. 2(f)]; and (f) 475 °C [Fig. 2(d)].

FIG. 6.

AFM images of MoO3 films deposited on stepped sapphire (0001) using a K-cell temperature of 640 °C and 5 × 10−6 Torr O2 ambient. Deposition time: 15 min. Substrate temperature (corresponding substrate AFM image): (a) 350 °C [Fig. 2(d)]; (b) 450 °C [Fig. 2(c)]; (c) 450 °C [Fig. 2(d)]; (d) 450 °C [Fig. 2(e)]; (e) 450 °C [Fig. 2(f)]; and (f) 475 °C [Fig. 2(d)].

Close modal

MoO3 films were deposited at 450 °C on stepped sapphire (0001) substrates having different surface morphologies. The film in Fig. 6(b) was deposited on a substrate with smaller and less regular steps [see Fig. 2(c)], and it shows relatively small pyramidal facets with no preferred orientation. A film deposited on a substrate with uniform 0.44-nm steps and relatively narrow terraces [see Fig. 2(e)] has a similar surface morphology [AFM image, Fig. 6(d)], except for some larger hexagonal slabs oriented roughly parallel to the substrate. As evidenced by AFM [Fig. 6(c)], a film comprising smooth rectangular slabs (∼500 nm × ∼ 1 μm) that are oriented parallel to the substrate surface was grown on the cross-stepped sapphire substrate shown in Fig. 2(d). Finally, a film [AFM image, Fig. 6(e)] grown on the sapphire substrate with the largest steps and widest terraces [Fig. 2(f)] also contains large (≥1 μm) smooth hexagonal and rectangular MoO3 slabs. Sapphire steps are inferred to create diffusion barriers for depositing MoO3 molecules, and multilayer MoO3 islands (domains) initially grow independently on each terrace. Once their thickness exceeds the step height, domains from adjacent terraces can connect, forming a continuous film; however, this process generates strain in the film because of misalignment between domains on adjacent terraces. The resulting stress is relaxed by forming facets through surface roughening. Increasing substrate step density increases the rate of roughening, leading to greater surface segmentation via faceting. Larger (taller) steps and wider terraces are expected to have the opposite effect.

In Secs. III C and III D, we describe thinner films deposited at 450 °C on stepped sapphire substrates using a lower MoO3 flux and/or shorter growth times.

In order to investigate the initial stages of growth, the incident MoO3 flux was reduced by lowering the K-cell temperature to 615 °C, and films were deposited at 450 °C on a sapphire substrate with parallel steps and terraces [AFM image, Fig. 2(e)]. Ex situ AFM images measured after MoO3 deposition for 5 min under these conditions are shown in Figs. 7(a) and 7(b). The sapphire is almost completely covered with small circular islands (nuclei) of 1 BL thickness (1 BL = 0.7 nm), some islands of 2 BL thickness (brighter spots), and a few 2–4 nm thick larger particles [the brightest spots in Fig. 7(a)]. The surface steps on the sapphire substrate are clearly visible beneath the film. Online XPS indicates that this film contains primarily MoO3 species and that the sapphire surface is stoichiometric with approximately the same OH contribution (15%) to the O 1s peak as before deposition. The Mo 3d spectral envelope [Fig. 8(a)] can be fit adequately using only one spin–orbit doublet; its components are separated by 3.1 eV with the 3d5/2/3d3/2 area ratio fixed at 1.5. The Mo 3d5/2 BE 233.0 eV is in excellent agreement with expected values for Mo+VI species.2,36–38 Adding a small (∼3%) contribution with a Mo 3d5/2 BE of 231.0 eV improves the fit. This contribution is consistent with Mo+V species, as observed by others for ultrathin MoO3 films on other substrates;21,22 however, a definitive assignment cannot be made. The Al 2p XP spectra of the sapphire substrate and MoO3 film are shown in Fig. 8(b). A film thickness of 0.48 nm was estimated from the Mo/Al intensity ratio using the equation reported by Cumpson39 and the measured Al 2p and Mo 3d intensities for clean sapphire and a thick MoO3 film, respectively. This estimate compares reasonably well with the AFM-determined thickness, especially given the incomplete coverage of the substrate. Carbon was not detected in the film by XPS.

FIG. 7.

AFM images of MoO3 deposited on sapphire (0001) with parallel steps and terraces at 450 °C using a K-cell temperature of 615 °C: (a) 5 min deposition (3 × 3) μm2 scan, (b) 5 min deposition (1 × 1) μm2 scan, (c) 15 min deposition (3 × 3) μm2 scan, (d) 15 min deposition (1 × 1) μm2 scan.

FIG. 7.

AFM images of MoO3 deposited on sapphire (0001) with parallel steps and terraces at 450 °C using a K-cell temperature of 615 °C: (a) 5 min deposition (3 × 3) μm2 scan, (b) 5 min deposition (1 × 1) μm2 scan, (c) 15 min deposition (3 × 3) μm2 scan, (d) 15 min deposition (1 × 1) μm2 scan.

Close modal
FIG. 8.

XP spectra of ultrathin MoO3 film deposited on stepped sapphire (0001) at 450 °C using a K-cell temperature of 615 °C; (a) Mo 3d region and (b) Al 2p region, including clean substrate and ultrathin film.

FIG. 8.

XP spectra of ultrathin MoO3 film deposited on stepped sapphire (0001) at 450 °C using a K-cell temperature of 615 °C; (a) Mo 3d region and (b) Al 2p region, including clean substrate and ultrathin film.

Close modal

The film morphology slowly evolves with time under these growth conditions; however, net MoO3 deposition is negligible after 5 min. After 15 min, long (500–1000 nm) 1-BL islands have formed on the terraces, and some smaller 2-BL islands (bright spots) are seen on top of these islands [Figs. 7(c) and 7(d)]. We infer that this morphology change occurs via island coalescence driven by a favorable (i.e., wetting) interaction with the sapphire surface. The ultrathin film, however, is not continuous over a large area, and some smaller islands directly on the substrate remain. The 3 × 3 μm2 AFM scan [Fig. 7(c)] shows a few 4–6 nm high particles; however, these could arise from contamination. The 1-BL islands on the terraces [Fig. 7(d)] are closely similar to those observed after 15-min deposition on stepped sapphire at 475 °C using a higher MoO3 flux [Fig. 6(f)], suggesting self-limiting growth under both conditions. Moreover, online XPS indicates that the 15-min ultrathin film has an equivalent Mo surface concentration to the 5-min ultrathin film and contains only Mo+VI species. The surface, however, contains less oxygen that the 5-min film, and the OH contribution to the O 1s peak (8.5%) is lower than for the bare sapphire substrate. This suggests that MoO3 reacts with surface OH groups as it spreads over the sapphire substrate. The interfacial reaction consumes surface OH groups but does not change the Mo oxidation state. The Mo 3d spectral envelope can be fit very well using only one spin–orbit doublet; its components are separated by 3.1 eV with a 3d5/2/3d3/2 area ratio of 1.5. The Mo 3d5/2 BE is 233.0 eV, as expected for Mo+VI species.2,36–38 The carbon concentration was below our XPS detection limit.

Sapphire (0001) substrates with multilayer steps and wide atomically smooth terraces [as shown in Fig. 2(d)] were selected for the deposition of continuous thin films at 450 °C using a K-cell temperature of 640 °C. Relatively short growth times (1 and 5 min) resulted in thin films with smooth surfaces as indicated by their streaky RHEED patterns. See the supplementary material. The well-defined step-terrace structure of these films is apparent in the ex situ AFM images [Figs. 9(a) and 9(b)].

FIG. 9.

MoO3 films grown on stepped c-plane sapphire at 450 °C using 640 °C K-cell temperature. AFM images: (a) 1 min. and (b) 5 min growth times and corresponding step profiles: (c) and (d), respectively.

FIG. 9.

MoO3 films grown on stepped c-plane sapphire at 450 °C using 640 °C K-cell temperature. AFM images: (a) 1 min. and (b) 5 min growth times and corresponding step profiles: (c) and (d), respectively.

Close modal

The AFM step profiles [Figs. 9(c) and 9(d)] reveal that the steps are ∼0.7 nm high—equivalent to the thickness of one BL of MoO6 octahedra in the b-direction of α-MoO3 [0.693 nm (Ref. 22)]. The overall appearance of the film is typical of multilayer growth.40 In this growth mode, after a terrace reaches a certain width, adsorbed monomers cannot descend the steps rapidly enough to prevent the nucleation of another layer. The magnitude of the Ehrlich–Schwoebel barrier controls the rate of step decent.41,42 The critical terrace width can be estimated from the AFM images to be 50–80 nm (the maximum width of terraces). An additional feature in Fig. 9(a), highlighted by a black circle, is a small island that appears to have nucleated on an open terrace. No grain boundaries or cracks were detected by AFM (multiple spots analyzed); therefore, we conclude that the films are monocrystalline, as supported by XRD and cross-sectional TEM (vide infra).

Figure 10 shows the Mo 3d XP spectrum of the 5-min MoO3 film grown at 450 °C [Fig. 9(b)]. An excellent fit of the spectral envelope was obtained using only one spin–orbit doublet; its components are separated by 3.15 eV with a 3d5/2/3d3/2 area ratio of 1.50. The Mo 3d5/2 binding energy (233.0 eV) is in excellent agreement with values expected for Mo+VI species.2,36–38 No carbon was detected by XPS, indicating good film purity. The Al 2p peak from the substrate was not observed, demonstrating that the film is continuous with a thickness exceeding the XPS sampling depth (∼three times the photoelectron attenuation length).39 

FIG. 10.

Mo 3d XP spectrum of MoO3 film grown on stepped c-plane sapphire at 450 °C using 640 °C K-cell temperature and 5 min growth time.

FIG. 10.

Mo 3d XP spectrum of MoO3 film grown on stepped c-plane sapphire at 450 °C using 640 °C K-cell temperature and 5 min growth time.

Close modal

The crystalline phase and orientation of selected films were determined by XRD. For comparison, the topmost trace in Fig. 11 is a diffractogram of α-MoO3 powder showing the diffraction peaks of randomly oriented crystallites. In contrast, the films grown on sapphire at 450 °C show exclusively diffraction peaks corresponding to (0k0) (k = 2, 4, 6, and 10) crystalline planes [12.8, 25.7, 39.0, 67.6°, literature43 12.8 (020), 25.7 (040), 39.0 (060), 67.5° (0(10)0)]; the (080) peak has negligible intensity. Thus, conventional XRD indicates that the films are α-MoO3 with (010) parallel to the sapphire surface, which is in agreement with the AFM results. The interlayer spacing calculated from the (020) Bragg reflection equals 0.692 nm, which is in excellent agreement with the expected value (0.693 nm) for α-MoO3.22 The peak at 41.7° corresponds to the (006) reflection of α-Al2O3 (sapphire).

FIG. 11.

XRD patterns of α-MoO3 powder and selected MoO3 films on stepped sapphire (0001).

FIG. 11.

XRD patterns of α-MoO3 powder and selected MoO3 films on stepped sapphire (0001).

Close modal

The α-MoO3 film shown in Fig. 9(b) was cross-sectioned via FIB, and the resulting specimen was imaged by high-resolution TEM. The 4.9-μm-long specimen [Fig. 12(a)] comprises several layers, some of which are remnants of FIB preparation. The two outer layers (bottom of image) are ∼300-nm ion beam- and ∼220-nm electron-beam-deposited Pt layers that protect the sample surface during ion milling. The third layer is a ∼25-nm thick sputter-deposited Pd–Au film used to make the surface conductive. Below this layer is the MoO3 film of interest, and below that is a 1.2-μm-thick sapphire substrate layer. The MoO3 film is ∼60 nm (54–64 nm range) thick over the length of the specimen. The α-MoO3 film is continuous and monocrystalline, i.e., no cracks or grain boundaries were observed. The α-MoO3 bilayers are parallel to sapphire (0001) [i.e., (010)α-MoO3∥(0001)sapphire], which is in agreement with the AFM and XRD results. The measured interlayer spacing of the α-MoO3 (010) sheets is 0.695 nm [Fig. 12(b)], which is very close to the established value of 0.693 nm.22 The in-plane orientation of the film cannot be assessed from the available data.

FIG. 12.

Cross-sectional TEM images of a MoO3 film grown on stepped c-plane sapphire at 450 °C using 640 °C K-cell temperature and 5 min growth time: (a) all layers in the FIB-prepared specimen and (b) high-magnification view of the MoO3 film.

FIG. 12.

Cross-sectional TEM images of a MoO3 film grown on stepped c-plane sapphire at 450 °C using 640 °C K-cell temperature and 5 min growth time: (a) all layers in the FIB-prepared specimen and (b) high-magnification view of the MoO3 film.

Close modal

Sapphire (0001) substrates with various step-terrace surface morphologies were prepared via annealing at 1100–1300 °C in air. The combination of ex situ annealing at 1100 °C and in vacuo annealing to 700 °C was found to be effective at surface cleaning. MoO3 films were deposited on c-plane sapphire via MBE using MoO3 vapor from a conventional Knudsen cell. Ultrathin films (1 BL thick) grow at 450 °C via a self-limiting process that represents a balance between the incident MoO3 flux and the desorption flux. The films are stoichiometric MoO3 as evidenced by XPS. For thick (∼1 μm) films, substrates with large steps (0.9–1.5 nm) and wide terraces provide less segmented films than substrates with small steps (0.22–0.66 nm) of high surface density. Monocrystalline thin films grown at 450 °C exhibit multilayer growth, resulting in the stepped surfaces comprised of 0.7-nm thick α-MoO3 (010) BL sheets. The monocrystalline [(010)α-MoO3∥(0001)sapphire] films were characterized by AFM, XPS, XRD, and FIB cross-sectional TEM.

This research was sponsored by the National Science Foundation (NSF) (No. CBET-1604605). It was performed in part at the Analytical Instrumentation Facility (AIF) at North Carolina State University, which is supported by the State of North Carolina and the NSF (No. ECCS-1542015). This work made use of instrumentation acquired with support from the NSF (No. DMR-1726294). The AIF is a member of the North Carolina Research Triangle Nanotechnology Network (RTNN), a site in the National Nanotechnology Coordinated Infrastructure (NNCI).

The data that support the findings of this study are available within the article and its supplementary material.

1.
P.
Novotný
,
S.
Yusuf
,
F.
Li
, and
H. H.
Lamb
,
J. Chem. Phys.
152
,
044713
(
2020
).
2.
E.
Heracleous
,
F.
Lee
,
I.
Vasalos
, and
A. A.
Lemonidou
,
Catal. Lett.
88
,
47
(
2003
).
3.
A.
Christodoulakis
,
E.
Heracleous
,
A.
Lemonidou
, and
S.
Boghosian
,
J. Catal.
242
,
16
(
2006
).
4.
K.
Chen
,
S.
Xie
,
A. T.
Bell
, and
E.
Iglesia
,
J. Catal.
198
,
232
(
2001
).
5.
E.
Comini
,
L.
Yubao
,
Y.
Brando
, and
G.
Sberveglieri
,
Chem. Phys. Lett.
407
,
368
(
2005
).
6.
L.
Zhou
,
L.
Yang
,
P.
Yuan
,
J.
Zou
,
Y.
Wu
, and
C.
Yu
,
J. Phys. Chem. C
114
,
21868
(
2010
).
7.
C.
Battaglia
 et al,
Nano Lett.
14
,
967
(
2014
).
8.
W.
Zeng
,
K. S.
Yong
,
Z. M.
Kam
,
Z.
Chen
, and
Y.
Li
,
Synth. Met.
161
,
2748
(
2012
).
9.
J. J.
Jasieniak
,
J.
Seifter
,
J.
Jo
,
T.
Mates
, and
A. J.
Heeger
,
Adv. Funct. Mater.
22
,
2594
(
2012
).
10.
K.
Koike
,
R.
Wada
,
S.
Yagi
,
Y.
Harada
,
S.
Sasa
, and
M.
Yano
,
Jpn. J. Appl. Phys.
53
,
058007
(
2014
).
11.
X. W.
Lou
and
H. C.
Zeng
,
Chem. Mater.
14
,
4781
(
2002
).
12.
G.
Wei
,
W.
Qin
,
D.
Zhang
,
G.
Wang
,
R.
Kim
,
K.
Zheng
, and
L.
Wang
,
J. Alloys Compd.
481
,
417
(
2009
).
13.
D.
Di Yao
,
J. Z.
Ou
,
K.
Latham
,
S.
Zhuiykov
,
A. P.
O’Mullane
, and
K.
Kalantarzadeh
,
Cryst. Growth Des.
12
,
1865
(
2012
).
14.
P. F.
Carcia
and
E. M.
McCarron
,
Thin Solid Films
155
,
53
(
1987
).
15.
Q. P.
Ding
,
H. B.
Huang
,
J. H.
Duan
,
J. F.
Gong
,
S. G.
Yang
,
X. N.
Zhao
, and
Y. W.
Du
,
J. Cryst. Growth
294
,
304
(
2006
).
16.
X. W.
Lou
and
H. C.
Zeng
,
J. Am. Chem. Soc.
125
,
2697
(
2003
).
17.
A.
Hanif
,
T.
Xiao
,
A. P. E.
York
,
J.
Sloan
, and
M. L. H.
Green
,
Chem. Mater.
14
,
1009
(
2002
).
18.
H.-S.
Kim
,
J. B.
Cook
,
H.
Lin
,
J. S.
Ko
,
S. H.
Tolbert
,
V.
Ozolins
, and
B.
Dunn
,
Nat. Mater.
16
,
454
(
2017
).
19.
A. J.
Molina-Mendoza
 et al,
Chem. Mater.
28
,
4042
(
2016
).
20.
J. H.
Kim
,
J. K.
Dash
,
J.
Kwon
,
C.
Hyun
,
H.
Kim
,
E.
Ji
, and
G.-H.
Lee
,
2D Mater.
6
,
015016
(
2019
).
21.
22.
L.
Kihlborg
,
Ark. Kemi
21
,
357
(
1963
).
23.
S. I.
Bakholdin
and
V. N.
Maslov
,
Phys. Solid State
57
,
1236
(
2015
).
24.
J. R.
Heffelfinger
and
C. B.
Carter
,
Surf. Sci.
389
,
188
(
1997
).
25.
M.
Yoshimoto
,
T.
Maeda
,
T.
Ohnishi
,
H.
Koinuma
,
O.
Ishiyama
,
M.
Shinohara
,
M.
Kubo
,
R.
Miura
, and
A.
Miyamoto
,
Appl. Phys. Lett.
67
,
2615
(
1995
).
27.
T.
Isono
,
T.
Ikeda
,
R.
Aoki
,
K.
Yamazaki
, and
T.
Ogino
,
Surf. Sci.
604
,
2055
(
2010
).
28.
G.-H.
Lee
,
Mater. Sci. Eng. B
138
,
41
(
2007
).
29.
S. C.
Wang
and
G.
Ehrlich
,
Phys. Rev. Lett.
70
,
41
(
1993
).
30.
K. M.
Adkison
,
S.-L.
Shang
,
B. J.
Bocklund
,
D.
Klimm
,
D. C.
Schlom
, and
Z.-K.
Liu
,
APL Mater.
8
,
081110
(
2020
).
31.
P.
Novotný
and
H. H.
Lamb
,
J. Vac. Sci. Technol. A
37
,
051504
(
2019
).
32.
D.
Wang
,
J.-N.
Li
,
Y.
Zhou
,
D.-H.
Xu
,
X.
Xiong
,
R.-W.
Peng
, and
M.
Wang
,
Appl. Phys. Lett.
108
,
053107
(
2016
).
33.
C. D.
Wagner
,
NIST X-Ray Photoelectron Spectroscopy (XPS) Database
(
National Instutute of Standards and Technology
,
Oakland
,
1991
).
34.
D.
Yang
,
M.
Krasowska
,
R.
Sedev
, and
J.
Ralston
,
Phys. Chem. Chem. Phys.
12
,
13724
(
2010
).
35.
J.
Stoch
and
J.
Gablankowska-Kukucz
,
Surf. Interface Anal.
17
,
165
(
1991
).
36.
P. A.
Spevack
and
N. S.
McIntyre
,
J. Phys. Chem.
97
,
11020
(
1993
).
37.
T. A.
Patterson
,
J. C.
Carver
,
D. E.
Leyden
, and
D. M.
Hercules
,
J. Phys. Chem.
80
,
1700
(
1976
).
38.
J.-G.
Choi
and
L. T.
Thompson
,
Appl. Surf. Sci.
93
,
143
(
1996
).
39.
P. J.
Cumpson
,
Surf. Interface Anal.
29
,
403
(
2000
).
40.
H.
Brune
, in
Surface and Interface Science: Solid-Solid Interfaces and Thin Films
, edited by
K.
Wandelt
(
Wiley VCH
,
Weinheim
,
2014
), p.
421
.
41.
G.
Ehrlich
and
F. G.
Hudda
,
J. Chem. Phys.
44
,
1039
(
1966
).
42.
R. L.
Schwoebel
and
E. J.
Shipsey
,
J. Appl. Phys.
37
,
3682
(
1966
).
43.
JCPDS Card No. PDF# 00-05-0508 (
2019
).
44.
See the supplementary material at http://www.scitation.org/doi/suppl/10.1116/6.0000962 for RHEED patterns of a clean stepped sapphire substrate and a film after 5 min growth.

Supplementary Material