United States

β-gallium oxide (β-Ga2O3) is a promising semiconductor for power device applications owing to its ultra-wide bandgap (4.8 eV) and high breakdown field (8 MV/cm). The high-power performances of the vertical-type Schottky barrier diodes (SBDs) have been demonstrated. However, killer defects are the potential leakage current paths which lowered the off-state voltages and degrade the SBD. [1,2] Therefore, identifying the killer defects and their generation mechanism is essential to enhance the SBD performance. In this talk, we reveal the identification of the killer defects in halide vapor phase epitaxial (HVPE) (001) β-Ga2O3 SBDs using ultrahigh sensitive emission microscopy and synchrotron X-ray topography. The N-type β-Ga2O3 epitaxial layer was grown by HVPE on an n-type 4&quot; EFG-grown (001) single-crystal wafer substrate. The net donor doping density, ND NA, was ca. 1016 cm−3. The epitaxial thickness was approximately 10 μm. Vertical-type SBDs were fabricated. For the ohmic contact, Ti/Au was evaporated on the entire back face. For the Schottky barrier (SB) contacts, Ni/Au was evaporated on the surface. The SBDs were observed by using Synchrotron X-ray topography, emission microscopy, SEM, and STEM. The current-voltage (I-V) characteristics of the Ga2O3 SBD showed that SBD exhibited a leakage current of 0.46 μA at a reverse bias of 100 V. However, emission patterns were observing during the I-V measurements and five emission spots were appeared in the squircle-shaped SBD with the width of 1600 μm (area of 2.37 × 10−2 cm2). Emission spots are corresponding to the reverse leakage current paths. Further investigation on these emission spots of the crystal using SEM showed a surface defect accompanies (100) cracks with a straight line along [100]. Next, cross-sectional STEM observation was performed. Dislocations were found to be propagated along [100] direction at a depth of 0.5 μm below the surface. In conclusion, we have found killer defects that act as leakage current paths in HVPE (001) β-Ga2O3 SBD. The defects contain (100) cracks and dislocations along [100] direction. [1] S. Sdoeung, M. Kasu, et al., Appl. Phys. Lett. 120, 092101 (2022). [2] S. Sdoeung, M. Kasu, et al., Appl. Phys. Lett. 120, 122107 (2022).

EMC 2023

Killer Defects Responsible for Leakage Current in HVPE (001) β-Ga2O3 SBD Observed by Emission Microscopy and Synchrotron X-Ray Topography

schottky barrier diode

gallium oxide

defect

β-gallium oxide (β-Ga2O3) is a promising semiconductor for power device applications owing to its ultra-wide bandgap (4.8 eV) and high breakdown field (8 MV/cm). The high-power performances of the vertical-type Schottky barrier diodes (SBDs) have been demonstrated. However, killer defects are the potential leakage current paths which lowered the off-state voltages and degrade the SBD. [1,2] Therefore, identifying the killer defects and their generation mechanism is essential to enhance the SBD performance. In this talk, we reveal the identification of the killer defects in halide vapor phase epitaxial (HVPE) (001) β-Ga2O3 SBDs using ultrahigh sensitive emission microscopy and synchrotron X-ray topography. The N-type β-Ga2O3 epitaxial layer was grown by HVPE on an n-type 4" EFG-grown (001) single-crystal wafer substrate. The net donor doping density, ND NA, was ca. 1016 cm−3. The epitaxial thickness was approximately 10 μm. Vertical-type SBDs were fabricated. For the ohmic contact, Ti/Au was evaporated on the entire back face. For the Schottky barrier (SB) contacts, Ni/Au was evaporated on the surface. The SBDs were observed by using Synchrotron X-ray topography, emission microscopy, SEM, and STEM. The current-voltage (I-V) characteristics of the Ga2O3 SBD showed that SBD exhibited a leakage current of 0.46 μA at a reverse bias of 100 V. However, emission patterns were observing during the I-V measurements and five emission spots were appeared in the squircle-shaped SBD with the width of 1600 μm (area of 2.37 × 10−2 cm2). Emission spots are corresponding to the reverse leakage current paths. Further investigation on these emission spots of the crystal using SEM showed a surface defect accompanies (100) cracks with a straight line along [100]. Next, cross-sectional STEM observation was performed. Dislocations were found to be propagated along [100] direction at a depth of 0.5 μm below the surface. In conclusion, we have found killer defects that act as leakage current paths in HVPE (001) β-Ga2O3 SBD. The defects contain (100) cracks and dislocations along [100] direction. [1] S. Sdoeung, M. Kasu, et al., Appl. Phys. Lett. 120, 092101 (2022). [2] S. Sdoeung, M. Kasu, et al., Appl. Phys. Lett. 120, 122107 (2022).

technical paper

The Electronic Materials Conference (EMC) is the premier annual forum on the preparation, characterization and use of electronic materials. This year’s event will feature a plenary session, topical sessions, a poster session and more.

EMC is coordinated with the Device Research Conference (DRC). This recognizes the strong interaction between device and electronic materials research and provides fruitful exchange of information between attendees of both Conferences.

The Electronic Materials Conference (EMC) is the premier annual forum on the preparation, characterization and use of electronic materials. 

Under high electric field, wavefunctions of electrons and holes leak into the bandgap, then, a semiconductor exhibits the tailing and oscillation in an optical absorption coefficient for below-bandgap (hv < Eg) and above-bandgap (hv > Eg) light, respectively. This is so called the Franz-Keldysh (FK) effect. The FK effect is often used as a powerful tool to investigate physical properties of materials. The understanding of the optical absorption characteristics under high electric field is very important toward various characterizations. In this study, we investigated the voltage dependence of the photocurrent in a β-Ga2O3 Schottky barrier diode under below-bandgap light illumination. Significant increase in photocurrent was observed, which is explained by the electro-absorption induced by the FK effect quantitatively. Figure 1 shows a schematic cross-section of a β-Ga2O3 Schottky barrier diode (SBD) fabricated in this study. The Si and Cl doped 10 μm thick n-type β-Ga2O3 layer was grown by hydride vapor phase epitaxy (HVPE) on a single-crystal (001)-oriented Sn-doped β-Ga2O3 bulk substrate. The threading dislocation density of the β-Ga2O3 substrate was less than104 cm-2. The carrier concentration of the substrate is approximately the 1×1018 cm-3. After the epitaxial growth, a Ti (50 nm)/Au (200 nm) metal stack was deposited on the surface of the epitaxial layer as Ohmic electrodes, and rapid thermal annealing (RTA) was performed in 450°C N2 ambient for 1 min. A Ni (20 nm)/Au (500 nm) metal stack was deposited on the surface of the epitaxial layer as Schottky electrodes. The measured device has a circular electrode with a 400 um diameter. Before the reverse-biased photocurrent measurements, current-voltage (I-V), capacitance-voltage (C-V) and internal photoemission (IPE) were carried out to characterize the Schottky barrier height of Ni/β-Ga2O3 interface. Figures 2, 3 and 4 shows the forward I-V, C-V and IPE characteristics of the SBD, respectively. The ideality factor of 1.03 was obtained from the I-V measurement, suggesting that the device showed nearly ideal thermionic emission transport. From the C-V measurement, the net donor concentration of 6.5×1015 cm-3 and its uniform distribution along the depth were confirmed. From these three measurements, the consistent and reliable barrier height of 1.16-1.17 eV was obtained, indicating that the Schottky interface has very high homogeneity. Figure 5 shows the reverse voltage dependence of the photocurrent in the β-Ga2O3 SBD under the monochromatic illumination with the wavelength of 275-300 nm (the band width is 5 nm). The irradiated light penetrated into β-Ga2O3 around top electrodes. The light reached the Schottky interface from the backside through multiple reflection. The light with photon energy higher than the Schottky barrier height causes IPE, resulting the photocurrent. The photocurrent originating from IPE has very small voltage dependence, since the barrier height is lowered by image force with increasing electric field near the Schottky interface. However, the photocurrent showed significant increase with reverse voltage. The increase was more striking as the wavelength approached to the fundamental absorption edge of β-Ga2O3. These voltage and wavelength dependences of the photocurrent indicate that the photocurrent is originating from the optical absorption induced by the FK effect. We calculated the photocurrent by considering the optical absorption induced by the FK effect in the depletion layer based on Aspnes’s theory. In the calculation, the material parameters of β-Ga2O3 were used, which is summarized in Table I. Figure 6 shows the comparison between the experimental data (open circles) and the calculated photocurrent (red broken lines). The calculated curves show excellent agreement with the experimental data. These results are very useful for optical absorption in β-Ga2O3 devices under high electric field such as a photomultiplication measurement [1, 2].

Photocurrent Induced by Franz-Keldysh Effect in β-Ga2O3 Schottky Barrier Diode Under High Reverse Bias Voltage

Gallium oxide has rapidly advanced as a material for high-power electronics and optically blind photodetectors, due not only to its wide band gap and availability of melt-grown substrates, but also its dopability with n-type doping being demonstrated from 1014 cm-3-1022 cm-3. However, numerous reports have shown dopant compensation when annealing in temperatures greater or equal to 1000 oC. To further device development, it is critical to understand the compensation mechanism, and to study defect formation and diffusion. It is of course difficult to study the diffusion of native defects such as gallium vacancies (VGa); typically isotope diffusion experiments are conducted. It is also of high interest to investigate the mechanisms by which native defects are introduced; does the material generate them homogeneously in its bulk or are they generated at surfaces? Do vacancies diffuse alone or as complexes with other defects? We sought an alternate method to determine the introduction and diffusion of Ga vacancies. We began with the hypothesis that Al could be used as a tracer for cation vacancy diffusion because of the assumption that Al would diffuse primarily substitutionally via cation vacancies. We designed superlattices of alternating epitaxial layers of Al0.1Ga1.9O3 / Ga2O3 deposited on both Sn- and Fe-doped (010) β-Ga2O3 substrates. The presence of many repeated layers diffusing during the same experiment can be used to reveal gradients in the diffusion constant vs depth and time, which in turn can indicate the direction of vacancy flux from the source. Dynamic Secondary Ion Mass Spectrometry (D-SIMS) profiles were collected before and after annealing in O2 at 1000-1100 oC. A novel transient simulation of vacancy- mediated diffusion was developed in Matlab to model the evolution of Al profiles under different scenarios of transient vacancy diffusion. A surprising finding is that there was no evidence of vacancy injection from the free surface, but that a large flux of vacancies flows into the superlattices from Sn-doped substrates. In the case of superlattices on Fe-doped substrates, no such flux is detected and the diffusion constant of Al is uniform and much smaller in depth, consistent with smaller uniform vacancy concentrations or other non-vacancy-mediated diffusion mechanisms. We present our results for the diffusion constant of VGa or its dominant complex mediating Al diffusion as well as the diffusivity for Al by the vacancy mechanism. The role of macroscopic crystallographic defects on the diffusion results was explored via ToF-SIMS 3D reconstructions and shown to be negligible under well-chosen conditions. Meanwhile, we also investigated the diffusion of Sn and Fe from the wafers into the superlattices. We provide a lower bound for the Sn diffusion coefficient, and discuss the strong possibility that it retards the diffusion of VGa by complexation. Fe profiles are very complex and evolve via multiple diffusion channels. We provide the diffusion coefficient for the slower channel, and model the faster surface pile-up by interaction with surface potential.

Measuring Diffusion of Al, Sn, and Fe in Ga2O3 Using β-(Al,Ga)2O3/Ga2O3 Superlattices

Alloying β-Ga2O3 with Al2O3 to create (AlxGa1-x)2O3 enables ultra-wide bandgap material suitable for applications deep into the ultraviolet. In this work, photoluminescence (PL) spectra of Cr3+ were investigated in monoclinic single crystal β-Ga2O3:Cr, and 10 mol.% Al2O3 alloyed with β-Ga2O3, denoted β-(Al0.1Ga0.9)2O3 or AGO. Temperature-dependent PL properties were studied for Cr3+ in AGO and β-Ga2O3 from 295 K to 16 K. For both materials at room temperature, the red-line emission doublet R1 and R2 occurs at 696 nm (1.78 eV) and 690 nm (1.80 eV), respectively, along with a broad emission band at 709 nm (1.75 eV). The linewidths for AGO are larger for all temperatures due to alloy broadening. For both materials, the R-lines blue-shift with decreasing temperature. The (lowest energy) R1 line is dominant at low temperatures due to the thermal population of the levels. For temperatures above ~50 K, however, the ratio of R2 to R1 peak areas is dominated by nonradiative combination. Additionally, Hall effect measurements were taken at low and elevated temperatures which demonstrated n-type behavior. Two different activation energies were identified, 57 meV associated with the Cr level, and 26 meV associated with an impurity band at lower temperatures.

Photoluminescence Spectroscopy of Cr3+ in β-Ga2O3 and (Al0.1Ga0.9)2O3

Ferroelectrics functioning at reduced dimensionality have long been pursued for the rich physics behind, including domain wall motion and energy transfer, charge compensation, phase stability, strain and interface coupling, as well as the devices enabled for highly scaled, energy efficient, multifunctional memory/sensing/computing devices and architectures. The emergent nitride ferroelectrics, including ScAlN and ScGaN, have been proposed as promising ferroelectrics with wide and tunable bandgap, large polarization and high tunable coercive field, with potentially high compatibility with both GaN and Si technology. For this new class of ferroelectrics, the ferroelectric wurtzite phase is energetically stable at room temperature, making it highly possible to maintain ferroelectric order in nitride ferroelectrics even at extremely reduced dimensionality. In this work, the thickness scaling down behavior of ScAlN thin films in the thickness range of 100 nm to 5 nm, has been studied to reveal the ferroelectric order in nitride ferroelectrics at reduced dimensionality. The ScAlN samples were grown by molecular beam epitaxy (MBE) on CMOS-compatible Mo/Si substrates. Reflection high energy electron diffraction (RHEED), X-ray diffraction (XRD), and scanning transmission electron microscopy (STEM) confirmed the wurtzite structure of the as-grown films. The ferroelectric properties of the as-grown films have been confirmed and investigated by capacitance-voltage (C-V) loops and positive-up-negative-down (PUND) measurements. An increase in the coercive field and a significant reduction of remnant polarization are found when the ferroelectric layer is scaled down to below 20 nm. Significantly, a switching voltage of 2-3.8 V, and a saturated remnant polarization of ~23 µC/cm2 are measured in ~ 5 nm thick ScAlN. X-ray diffractions and transmission electron microscopy studies suggest that the increase in the coercive field and diminishment in switchable polarization can be attributed to the naturally formed surface oxide and the compressive strain in ultrathin ScAlN films due to a larger in-plane lattice constant of ScAlN compared to Mo (011). Our work sheds light on the fundamental thickness-scaling properties of epitaxial ferroelectric ScAlN thin films and paves the way for next-generation compact and power-efficient devices and applications based on nitride ferroelectrics.

Epitaxial Ferroelectric ScAlN—Thickness Scaling to the Nanometer Scale

Extreme bandgap (EBG) semiconductors such as AlxGa1-xN (x>0.6, bandgap energy EG > 4.88 eV) are ideal for high power, high voltage, and high temperature power electronics. Their critical breakdown field (Ec) in lateral devices scales as EG2.5, leads to an estimated value of the Baliga’s Figure of Merit (BFOM) significantly higher than that for GaN[1],[2],[3],[4] For achieving high BFOM, in addition to a large breakdown voltage (VBR), a simultaneous low device specific ON-resistance(RON,SP) is also required. For the EBG AlxGa1-xN channel HEMTs to have a low RON,SP , a 2-dimensional electron gas (2DEG) concentration >1x1013 cm-2 is needed, requiring a barrier AlxGa1-xN layer with x > 0.80. At such high value of “x”, it is challenging to make linear ohmic contacts due to large Schottky barrier height between metal-AlGaN interface. To solve these issues, we demonstrate EBG n-Al0.87Ga0.13N-Al0.64Ga0.36N heterostructures over bulk AlN substrate grown by Low Pressure-MOCVD. To facilitate ohmic contact formation, highly doped reverse-graded AlGaN layer(x=0.87 to 0.40) was grown in-situ. Reciprocal Space Map(105) and AFM analysis revealed pseudomorphic epilayers with step-flow morphology and RMS roughness ~0.18 nm. X-Ray Topography estimation of control sample gave us a TDD ~ 7 x 103 cm-2, indicating the high quality of our pseudomorphic epilayers. A planar Zr-based metal stack was deposited and subsequent removal of the reverse graded n-AlxGa1-xN layer in the access region gave linear ohmic contacts, RC~4.3 Ω-mm. Additionally, planar metal-oxide-semiconductor (MOS) diodes were fabricated and 2-terminal gate-source breakdown measurement was performed. A VBR of 739 V was measured at metal flashover point for 0.65 µm gate-source spacing, giving an electric field value ~ 11.37 MV/cm. From the breakdown data and the RON,SP (~2.4 × 10-4 Ω-cm2) value, we estimate the BFOM for our EBG n-Al0.87Ga0.13N-A0.64Ga0.36N heterostructures to be as high as 2.27 GW/cm2. This, to the best of our knowledge is the highest reported value to date for an extreme bandgap AlxGa1-xN (x>0.6) channel heterostructure.[5] References [1] B. J. Baliga, IEEE Electron Dev. Lett., 10, 455 (1989). [2] R. J. Kaplar, A. A. Allerman, A. M. Armstrong, M. H. Crawford, J. Ray Dickerson, A. J. Fischer, A. G. Baca, and E. A. Douglas, ECS J. Solid State Sci., 6, Q3061 (2016). [3] T. Nanjo, A. Imai, Y. Suzuki, Y. Abe, T. Oishi, M. Suita, E. Yagyu, and Y. Tokuda, IEEE Trans. Electron Devices, 60, 1046 (2013). [4] N. Yafune, S. Hashimoto, K. Akita, Y. Yamamoto, H. Tokuda, and M. Kuzuhara, Electronics Lett., 50, 211 (2014) [5] K. Hussain et al., Appl. Phys. Express (2023)

High Figure of Merit Extreme Bandgap Al0.87Ga0.13N-Al0.64Ga0.36N Heterostructures Over Bulk AlN Substrates

In this work, epitaxial wurtzite phase BAlN films were grown by nitrogen plasma molecular beam epitaxy on different wurtzite phase substrates (e. g. nitrided sapphire, β-Nb2N film, GaN template). Effect of different growth parameters on composition of BAlN films was explored. B incorporation was found to increase with decreasing substrate temperature and with increasing B flux. In-situ electron diffraction and ex-situ x-ray diffraction measurements confirm the growth of wurtzite BAlN films at 300 in plane orientation with sapphire substrate and at 00 orientation with bottom Nb2N films. With increasing boron content in BAlN films, the c-parameter decreases while the a-parameter shows a non-monotonic trend. Ferroelectricity (with remnant polarization of ~16 μC.cm–2 and coercive field of ~1.45 MV.cm-1) and high dielectric constant (of ~18 at 500 kHz) were also observed in these BAlN films. A slight reduction in bandgap is seen with increasing boron content in BAlN films (from ~6.12eV for AlN films to ~5.95eV for B0.15Al0.85N film). This work also reports the first demonstration of polarization-induced two-dimensional electron gas (2DEG) in the epitaxial BAlN/GaN heterojunctions. Observation of ferroelectricity with low coercive field, remnant polarization (lower than the sputtered BAlN [1] and comparable to MBE grown AlScN films [2]) and high-k, makes BAlN an attractive barrier in GaN high electron mobility transistors (HEMTs). Therefore, BAlN/GaN heterostructures were grown on 2” and 3” wafers; showing polarization-induced 2DEG with good sheet resistance of ~650 ohm/sq, ~400-500 cm2/V.s mobility (400-500 cm2/V.s), and 1.4-1.8 e13 cm-2 charge. The results presented in this paper demonstrates MBE growth of BAlN films and heterostructures with interesting properties that would enable new electronic devices. These films could be useful for combining memory and logic functionalities in digital electronics and to mingle epitaxial ferroelectric and high-k barrier technology into RF and mm-wave power electronics. For instance, a new class of ferroelectric HEMTs devices can be formed by integrating epitaxial ferroelectric, high-k nitride barrier with III-V nitride semiconductors [3].

MBE Growth, Characterization of BAlN Films and 2D Electron Gas in Epitaxial BAlN/GaN Heterojunction

Two-dimensional (2D) transition metal dichalcogenides (MX2, with M being a transition metal and X a chalcogen) have potential application in next-generation optoelectronics, electronics, chemical sensing and quantum technologies. Epitaxial growth of single crystal TMD monolayers at the wafer-scale is of significant interest for device applications. Metalorganic chemical vapor deposition (MOCVD) is a promising approach for TMD growth as it enables the use of high substrate temperatures (900-1000oC) and chalcogen/metal ratios (103-105) which are beneficial for epitaxy. In addition, the flow rate of precursors can be modulated during growth to enhance the surface diffusion of metal-containing species and control the lateral growth rate of TMD domains. Sapphire has emerged as a promising substrate for wafer-scale TMD monolayers and few layer films due to its crystallographic compatibility and commercial availability. Despite the broad interest in wafer-scale 2D TMDs, the fundamental understanding of how the growth conditions, e.g. growth temperature, impacts the epitaxial growth and properties of 2D TMDs has rarely been investigated. In this study, MOCVD growth of wafer scale MoS2 monolayer was carried out in a horizontal cold-wall reactor that includes an induction-heated rotating SiC-coated graphite susceptor and separate gas inlets for the metal and chalcogen precursors. Molybdenum hexacarbonyl (Mo(CO)6) and hydrogen sulfide (H2S) were used as precursors in a H2 carrier gas. A 2” c-plane sapphire substate was used for all MoS2 growths. The growth temperature was varied as 900oC, 950oC, and 1000oC. The flow rate of the molybdenum precursor used in this work ranged from 0.001 to 0.003sccm while the H2S flow rate was maintained at 400 sccm. Growths were carried out with a moderate reactor pressure of 50 Torr using H2 as the carrier gas. Using these conditions, growth of fully coalesced monolayer MoS2 was achieved in 10-30 minutes across the 2” substrate. AFM and FESEM were used to characterize the surface coverage and morphology of the layers and in-plane X-ray diffraction was used to assess the epitaxial quality. Spatially resolved Raman and photoluminescence spectroscopy were carried out to assess the uniformity of the monolayer and along with variable temperature photoluminescence measurements. A coalesced monolayer of MoS2 was achieved on 2” c-plane sapphire over the temperature range from 900oC to 1000oC. The surface coverage (growth rate) was found to be only weakly dependent on the growth temperature and scaled roughly linearly with Mo(CO)6 flow rate indicating that the growth is limited by mass transport rather than reaction kinetics. In-plane XRD phi-scans confirm that the films are epitaxially oriented with respect to the sapphire over the temperature range (900oC to 1000oC) with the (1120) of MoS2 aligned with the (1120) of α-Al2O3. The FWHM of the phi-scan peak of MoS2 decreases with increasing growth temperature, indicating a reduction in in-plane rotational misorientation of the MoS2 domains at higher growth temperature. The MoS2 monolayer sample grown at 1000oC exhibits strong photoluminescence at room temperature centered around 1.9 eV and defect-related emission which dominates at low temperature of 6 K. PL measurements also reveal an increase in the intensity ratio of the A exciton to trion emission with increasing growth temperature indicating a reduction in defects. Additional results correlating the structural and optical properties of the MoS2 monolayers to the transport properties of field-effect devices will be reported. The results of this work provide a deeper understanding of the epitaxial growth of MoS2 monolayers grown by MOCVD and the impact of defects on the optical and electrical properties of the layers.

Effects of Growth Temperature on the Properties of Wafer-Scale Epitaxial MoS2 Monolayers Grown by Metalorganic Chemical Vapor Deposition

Single photon emitters (SPEs) in 2D transition metal dichalcogenides (TMDCs) offer a compelling solution to create bright, spectrally narrow quantum emitters with high photon-extraction efficiencies. Quantum emitters in TMDC monolayers are typically associated with point defects or strain in engineered structures. Another potential avenue for SPEs is the formation of lateral heterostructures in which a TMDC nanodot is embedded in a matrix comprised of a wider bandgap TMDC. The nanodot can capture excitons from the surrounding matrix and may also host quantum states. Despite these intriguing prospects, there have been few studies focused on nanodot formation in TMDC monolayer material. In this work, we report on the synthesis of MoSe2 nanodots embedded within a WSe2 monolayer matrix via metal-organic chemical vapor deposition (MOCVD). MOCVD growth was carried out at 800-1000 °C in a cold-wall reactor using Mo(CO)6 and W(CO)6 as metal precursors and H2Se as the chalcogen source. C-plane sapphire was used as the substrate with H2 as the carrier gas. A three-step process was employed whereby a short nucleation step was initially used to form nanoscale MoSe2 domains followed by a ripening step in which the Mo(CO)6 was switched out of the reactor and the sample was annealed under H2Se. W(CO)6 was then introduced resulting in lateral growth of a WSe2 matrix surrounding the MoSe2 domains. The nucleation time was varied to change the size of the MoSe2 nanodots and the duration of ripening was modified to control the interfacial abruptness. Raman spectroscopy of the monolayers confirmed the presence of both the A1g peaks for WSe2 and MoSe2 at 247 cm-1 and 239 cm-1, respectively. High resolution scanning transmission electron microscopy (STEM) and energy dispersive spectroscopy (EDS) revealed the presence of triangular MoSe2 nanodots with edge lengths on the order of 10-100 nm embedded within a continuous WSe2 matrix. Variable temperature photoluminescence (PL) measurements were carried out on monolayer samples removed from the sapphire growth substrate and encapsulated in hBN. At room temperature, PL peaks were observed at 1.65 eV and 1.57 eV which correlate with exciton-related emission from WSe2 and MoSe2, respectively. As the temperature was decreased from 280K to 80K, the MoSe2 PL intensity increased relative to the WSe2 emission which is suggestive of charge transport from the WSe2 matrix to the MoSe2 nanodots.

MOCVD Growth and Characterization of MoSe2 Nanodots within a WSe2 Monolayer Matrix

Defects can occur during the growth of two-dimensional (2D) transition metal dichalcogenide (TMD) crystals. Exposure to these defects in 2D TMD to air and harsh conditions initiates the reaction of exposed transition metal atoms with active species such as O2 and H2O molecules, triggering an oxidation process. Understanding the oxidation behavior of 2D TMDs is crucial for their practical applications in electronics, optoelectronics, and electro-/photocatalysis because any structural deformation under ambient conditions directly affects their unique properties. As 2D semiconductor, 2H-MoTe2 is highly susceptible to air, it has the potential to control the conversion between p- and n-type characteristics. Nevertheless, previous research has been conducted on macroscopic scale oxidation, there is limited understanding of the oxidation behavior at the atomic level. It is imperative to explore the oxidation behavior of 2H-MoTe2 in greater detail to effectively modify its properties. In this study, we provide a detailed investigation on the oxidation of 2H-MoTe2 under various environmental conditions and insights for oxidation control. Exposure of synthesized 2H-MoTe2 flakes to air at room temperature for 17 months led to two striking morphological changes: 1) Hexagonal or triangular etching was observed from defects such as screw dislocations, crystal edges, and grain boundaries, resulting in the preferential oxidation and formation of pits with an average lateral size of several micrometers. 2) After a relatively slow reaction, triangular protruding oxide layer consisting of a stable MoO2 hillocks were formed on the basal plane of MoTe2, as confirmed by surface and structure characterization using scanning electron microscope (SEM), transmission electron microscope (TEM) and Raman. Experiments monitoring at the oxidation process at 230 °C for 3 h and TEM analysis showed that the substrates and active species have a significant effect on oxidation mechanism. Interestingly, under an O2 atmosphere, MoO2 layers was formed on the entire surface of 2H-MoTe2, while in a H2O atmosphere, most of the surface remained intact but oxidation etching occurred at defects in the basal plane. Furthermore, using theoretical calculations, we investigated the reaction of the surface Te atoms of defective 2H-MoTe2 with O2 and H2O molecules, which provide volatile TeO2 and H2Te and create a Te vacancy in the 2H-MoTe2 basal plane. It was found that, although O2 dissociation in the basal plane of 2H-MoTe2 with a Te vacancy is energetically favored over H2O dissociation, the subsequent reaction involving H is a dominant mechanism in forming Te vacancies than reactions involving an O atom. More importantly, when TeO2 volatilizes, the additional adsorbed O atom suppresses the lattice reconstruction around TeO2, resulting in an increased kinetic barrier (∼2.22 eV) to volatile TeO2. This finding is consistent with the TEM analysis results of the monitoring experiment, which demonstrated the formation of a stable oxide film under an O2 atmosphere. Our study provides valuable insight into the oxidation dynamics of 2H-MoTe2 at the atomic level and can inspire a new approach to improve the reliability and durability of TMD-based 2D devices.

Oxidation Mechanism of 2H-MoTe2 at the Atomic Scale

Mixed valence manganese oxides (MnOx) have attracted significant interest in recent years due to the easily reversible redox reactions between manganese oxidation states (Mn+2, Mn+3, and Mn+4)1, which has enabled applications in catalysis2, energy storage3, and gas sensing4. Manganese dioxide (MnO2), in particular, has demonstrated tunable structural properties, due to its wide variety of synthesizable structural polymorphs ((α, β) tunnel5, (γ) spinel5, and (δ) layered5) which allow for control over its available surface area and the reactive properties. Among these structural polymorphs, the α and δ phases exhibit strongly mixed-valence character, with Mn+3 defects found throughout the crystal structure5, which is desirable due to it being a highly active site6. Water-stabilized hexagonal δ-MnO2 (HMnO2) contains many Mn+3 ions as the interlayer contains Mn+2/+3 ions to neutralize the layer charge from Mn+3 lattice defects7. However, the direct synthesis of HMnO2 has not been reported extensively in the literature, as most synthesis routes rely on permanganate salts1, which leads to the synthesis of ion intercalated δ-MnO2 and a reduction in Mn+3 defects in the lattice9. In our previous work, we presented the synthesis of high Mn+3 content HMnO2 via electrodeposition on epitaxial graphene/silicon carbide (EG/SiC) substrates4. In this work, we extend upon this by transforming also these as-grown films via intercalating with ions to form alternate birnessites and controlling the growth via patterning. The electrodeposition was carried out by a three-pulse deposition in a three-electrode electrochemical cell, as described previously4. The resulting δ-MnO2 thin films were then characterized through Raman spectroscopy, scanning electron microscopy (SEM), and atomic force microscopy(AFM). Raman spectroscopy confirmed the formation of δ-MnO2 as the peaks were consistent with the literature4,8. The AFM indicated that the surface was made up of nanofibrous nanoparticles, which SEM confirmed and indicated that the deposition on EG/SiC was made up of microplates larger than 40 μm x 40 μm in size. After this characterization, the resulting films were exfoliated mechanically and analyzed with AFM to determine if the films were epitaxial. This exfoliation generated 6-layer MnO2 nanoplates, indicating we can reach the 2D limit and confirm epitaxial growth. The resulting films were intercalated with alkali metals (Na+) via cyclic voltammetry in a similar three-electrode cell to the deposition, using a 2 M metal nitrate solution. Optical Analysis of the film indicated two distinct regions formed in the film, large multicolored regions, and small blue and black islands. Raman analysis of these regions found that the islands were unintercalated MnO2 but the colored regions were Na-type birnessite due to their Raman shifts which are consistent with the literature8. Believing ion diffusion resistance to be impeding the intercalation we patterned δ-MnO2 into 100 μm ribbons and attempted additional intercalation. Optical analysis of the film post patterned intercalation revealed two regions, a brown and blue region. Raman spectroscopy of these two regions indicated that both were Na-type birnessite. However, differing peak intensities indicate the crystallinity of the two regions is different. This demonstrates that we have successfully grown and transformed the H-type birnessite to Na-type birnessite with 100% efficiency on EG/SiC, allowing for additional fine-tuning of its properties and expanding its application in energy storage to Na+ and K+ chemistries. References: [1]L. Spinelle et al. Sensors 17(7), 2017 [2] F. Cheng et al. Chem. Mater 22(3), 2010 [3] Y.J. Huang et al. Electrochim Acta 99, 2013 [4] M. Pedowitz et al. J. Microelectromech Syst 29(5), 2020 [5] Z. Chang et al. Proc Natl Acad Sci USA 115(23), 2018 [6] Julien et al. Spectrochim Acta A 60, 2004 [7] Drits et al. Am Mineralogist 82, 1997 [8] Julien et al. Solid State Ionics 159, 2003

Premium content

Killer Defects Responsible for Leakage Current in HVPE (001) β-Ga2O3 SBD Observed by Emission Microscopy and Synchrotron X-Ray Topography

Downloads

Next from EMC 2023

Photocurrent Induced by Franz-Keldysh Effect in β-Ga2O3 Schottky Barrier Diode Under High Reverse Bias Voltage

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES