Patent Application
20250087484
Various embodiments of the present technology generally relate to silicon (Si)-based semiconductor manufacturing. More specifically, some embodiments of the present technology relate to systems and methods for integrating gallium oxide thin films on silicon (001) via a buffer layer.
Ga2O3 is a promising wide band-gap semiconductor with potential applications in high power electronics, solar-blind ultraviolet (UV) photodetectors, UV-transparent conductive films, UV detectors, electro-optical devices, and possibly microwave switching and amplification [1-3, 106, 107]. Benefits of the material are an ultra-wide band gap of ˜5 eV, transparency into the deep-UV spectral range, and the availability of relatively low-cost bulk single crystals, compared to other wide band gap materials. Ga2O3 has a high theoretical breakdown field of about 8 MV/cm [108], higher than GaN (3.3 MV/cm) and SiC (2.5 MV/cm), which are, besides the ubiquitous Si (0.3 MV/cm), the most commonly used materials in high-power electronics [4, 5]. Larger breakdown fields can lead to the miniaturization of power electronic devices with associated reductions in cost and weight. The enhanced radiation hardness of Ga2O3 [6] additionally makes it suitable for space applications. Today an estimated 30% of all electricity flows through power electronics and this is projected to reach 80% in the future [7]. Efficient n-type doping of Ga2O3 can be achieved by Si, Ge, Sn and Nb incorporation [3]. Recently p-type doping with H has been demonstrated [8].
Six different polymorphs of Ga2O3 have been described: monoclinic (β), rhombohedral/trigonal (α), cubic defective spinel (γ), cubic bixbyite (δ), orthorhombic (κ), and hexagonal (ε) [9]. All Ga2O3 polymorphs are based on a close-packed arrangement of oxide ions (hexagonal close-packed (hcp) or face-centered cubic (fcc); distorted to some degree in most cases), and Ga ions partially fill octahedral and tetrahedral voids, which show local disorder for some structures [9]. δ-Ga2O3 has been speculated to be a mixture of β- and ε—Ga2O3 [9], and hexagonal ε—Ga2O3 has been shown by electron microscopy to be a multi-domain structure of orthorhombic κ-Ga2O3 (isostructural to κ-Al2O3) [10]. Recently, the orthorhombic polymorph has been referred to as either as ε—Ga2O3 [11, 12] or as κ-Ga2O3 [10]; due to the analogy with κ-Al2O3 and to avoid ambiguity, it will be referred to herein as κ-Ga2O3. The crystallographic relationships between κ-, β-, and γ-polymorphs have been described [13]. In summary, there appear to be only four truly distinct polymorphs: α, β, γ and κ.
β-Ga2O3 is the stable polymorph under normal conditions and crystallizes in space group C2/m with lattice parameters a=12.22 Å, b=3.04 Å, c=5.80 Å and β=103.75° [9]. A major advantage of β-Ga2O3 is that low-cost and high-quality single crystals are available: 4 inch β-Ga2O3 wafers are commercially available [14], and 6 inch wafers are in development [15]. The production of 6 inch β-Ga2O3 bulk wafers is projected to be about a ⅓ cheaper than the production of 6″ SiC wafers [7]. A major disadvantage of Ga2O3, however, is the low thermal conductivity of 0.1-0.3 Wcm−1K−1 [5, 108], which is detrimental for high-power applications that dissipate energy in the form of Joule heating [16, 109]. Therefore, the integration of epitaxial Ga2O3 onto Si would be advantageous due to the 750% higher thermal conductivity of Si compared to Ga2O3 [17, 110]. Currently, large-scale epitaxial films of Ga2O3 are unavailable commercially and bulk crystal wafers are only available up to 4 inch diameter [111]. Another benefit would be cost reduction [17], based on even larger diameter Ga2O3 substrates and direct process integration with Si technology. Epitaxial growth onto Si (001) wafers would open up numerous avenues for the large scale integration but, unfortunately, Ga2O3 cannot be directly grown on Si.
The challenge of epitaxial integration is the structural difference between silicon and gallium oxide. Si has a crystal structure with cubic symmetry, whereas β-Ga2O3 has monoclinic symmetry. This symmetry mismatch will unavoidably lead to a multi-orientation growth of epitaxial domains [18, 19]. Crystal grain boundaries between epitaxial domains and defects arising from the lattice mismatch to the substrate can be pernicious to the electronic properties of the epilayer.
In prior work on the integration of Ga2O3 with Si, a textured 1 μm β-Ga2O3 film was grown by HVPE on Si (111) with a ˜100 nm 3C—SiC (zinc blende structure) buffer layer [24]; randomly oriented β-Ga2O3 rods as well as nanosheets were grown by chemical vapor deposition (CVD) on 3C—SiC/Si (001) [25, 26]; and polycrystalline β-Ga2O3 was grown directly on Si (001) and Si (111) without buffer layers by pulsed laser deposition (PLD) [27]. β-Ga2O3 of unclear structural orientation was also grown by HVPE on SiC-buffered Si (001), Si (011) and Si (111) [28, 29]. Additional prior work includes the deposition of amorphous Ga2O3 on Si (001) by CVD [30], and crystallization of amorphous Ga2O3 on Si (111) by annealing leading to grains of β-Ga2O3 [31]. Recently, β-Ga2O3 wafers have been integrated directly, and with an alumina interlayer, onto Si and SiC wafers by wafer bonding and ion cutting [32, 33].
In many instances, Ga2O3 has been grown heteroepitaxially on widely available, high-structural-quality and low-cost c-plane sapphire substrates by a variety of physical and chemical vapor deposition methods: MBE [34-43], MOCVD [42, 44, 45], PLD [42, 46], rf-sputtering [47], CVD [11, 48, 49], HVPE [50-53], PVD [54, 55]. Depending on the growth method, growth parameters such as temperature and post-growth treatments, various phases and mixtures of phases of epitaxial Ga2O3 are created on c-plane sapphire: β-Ga2O3 (
Previous attempts at integrating Ga2O3 on Si highlight the difficulty of this process. A textured 1-μm-thick β-Ga2O3 film was grown by hydride vapor phase epitaxy (HVPE) on Si (111) with a ˜100-nm 3C—SiC (zincblende structure) buffer layer [
Epitaxial Ga2O3 thin films have previously been grown on some oxide substrates. The (100)-oriented epitaxial relationship was found for Ga2O3 grown by MOCVD on SrTiO3 (STO) (001) after annealing the sample above 1000° C. [122]. Other examples of Ga2O3 growth on oxide surfaces include the deposition of κ-Ga2O3 (001) on STO (111) via tin-assisted pulsed-laser deposition and by mist CVD [124]. We have also recently reported the integration of β-Ga2O3 by plasma-assisted molecular beam epitaxy onto a γ-Al2O3 (111) buffer layer grown at 840° C. by e-beam evaporation on a clean Si (001) surface [125].
Integration of β-Ga2O3 on Si (001) is highly challenging due to multiple fundamental differences between the sesquioxide and the semiconductor substrate.
In this disclosure, references to the above sources, which are numbered in single bracket “[ ]”, are made using the corresponding number in the single brackets, in the same manner as those sources are listed above. Each of the above listed references, along with other patents, patent applications, and non-patent literature not specifically listed above, are incorporated herein by reference in their entireties.
Systems and techniques for growing Ga2O3 on a Si substrate are provided herein. In an aspect, a buffer layer is provided herein that allows for integration of a thin film of Ga2O3 on a Si surface without impacting the stability of Ga2O3 or impacting the crystal symmetry of Ga2O3. In particular, the buffer layer provided herein provides an interface between Ga2O3 and Si such to thermodynamically stabilize Ga2O3 and reduce lattice defects or disorder. By providing the buffer layer, Ga2O3 can be formed in an epitaxial stack structure formed by growing the buffer layer on a Si substrate, and then forming a thin film of Ga2O3 on the buffer layer.
As will be described in greater detail below, a buffer layer containing a metal oxide may be formed on the Si substrate. The metal oxide may include STO, MgO, alumina, a rare earth oxide, or a combination thereof. In some cases, the metal oxide may be STO or MgO onto which a seed layer may be formed. The seed layer may be an aluminum oxide material that provides for further lattice alignment at the interface between the buffer layer and the thin film of Ga2O3. In some embodiments, in addition to the thin film of Ga2O3, one or more additional layers of Ga2O3 may be formed on the thin film to increase the thickness of the thin film. Various techniques may be employed to form each of the buffer layer, seed layer, and thin film of Ga2O3, such as MBE, PAMBE, CVD and MOCVD.
In some embodiments, the epitaxial stack structure, which may also be referred to herein as a wafer, may be integrated into an electronic device. For example, the epitaxial stack structure may be in electrical communication with one or more of a positive electrode and a negative electrode to provide electrical support for an electronic device. By integrating the epitaxial stack structure into an electronic device, the advantageous characteristics and properties of Ga2O3 can be harnessed to enhance the performance of the device.
The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fec.
Embodiments of the present technology will be described and explained through the use of the accompanying drawings.
100
and
110
azimuths for a 20-nm film.
The drawings have not necessarily been drawn to scale. Similarly, some components and/or operations may be separated into different blocks or combined into a single block for the purposes of discussion of some of the embodiments of the present technology. Moreover, while the technology is amenable to various modifications and alternative forms, specific embodiments have been shown by way of example in the drawings and are described in detail below. The intention, however, is not to limit the technology to the particular embodiments described. On the contrary, the technology is intended to cover all modifications, equivalents, and alternatives falling within the scope of the technology as defined by the appended claims.
The present disclosure relates to the epitaxial integration of gallium oxide (Ga2O3), a wide band-gap semiconductor material, onto Si substrates for advanced electronic applications. Ga2O3 has garnered significant attention in high-power electronics, ultraviolet (UV) photodetectors, and other optoelectronic devices due to its ultra-wide band-gap (˜5 eV), high theoretical breakdown field (˜8 MV/cm), and relatively low cost for bulk crystal production compared to other wide band-gap semiconductors. However, despite these advantages, the integration of Ga2O3 onto Si poses several technical challenges, including structural and thermal incompatibilities that must be addressed to enable scalable, high-quality Ga2O3-based devices.
This detailed description explores the methods and materials required for overcoming these integration challenges, focusing on improving the epitaxial growth of Ga2O3 on Si substrates. The embodiments described herein provide solutions to the crystal symmetry mismatch and the need for buffer layers, enabling successful epitaxial growth while reducing defect densities and enhancing the performance of the resulting electronic devices. In the following description, for the purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of embodiments of the present technology. It will be apparent, however, to one skilled in the art that embodiments of the present technology may be practiced without some of these specific details.
One of the key challenges in integrating Ga2O3 with Si substrates arises from the structural differences between the two materials. Silicon possesses a cubic crystal structure, whereas β-Ga2O3, the stable phase of gallium oxide, has a monoclinic crystal structure. This symmetry mismatch results in multi-orientation growth of Ga2O3 epitaxial domains on the Si substrate. The resulting crystal grain boundaries between these domains, along with defects arising from lattice mismatch, can severely degrade the electronic properties of the Ga2O3 epilayer. Therefore, improving the crystal quality of hetero-epitaxially grown Ga2O3 by reducing defect densities and increasing grain sizes is critical for high-performance device applications.
Another challenge is the low thermal conductivity of Ga2O3 (0.1-0.3 Wcm−1K−1), which makes heat dissipation difficult in high-power applications. Silicon, on the other hand, has a thermal conductivity over 750% higher than Ga2O3, making it a desirable substrate for heat management. However, direct growth of Ga2O3 on Si using molecular beam epitaxy (MBE), the preferred growth technique for integration due to its ability to control growth at the atomic level, is not feasible without a buffer layer. MBE requires an activated oxygen source, such as plasma or ozone, for Ga2O3 growth, which would oxidize the clean Si surface, forming amorphous silicon dioxide (SiO2) and preventing epitaxial alignment. Furthermore, Ga2O3 is not thermodynamically stable on Si, leading to the spontaneous formation of SiO2 at the interface, which further inhibits epitaxial growth.
To address at least the above-identified shortcomings, a buffer layer is necessary to achieve high-quality epitaxial growth of Ga2O3 on Si. The buffer layer prevents the oxidation of the Si surface and enables the epitaxial alignment of Ga2O3. Suitable buffer layers include materials that can grow as single crystals on Si without oxidizing it, and that also possess in-plane lattice spacings compatible with at least two dimensions of the β-Ga2O3 unit cell. Examples of such metal-oxide buffer layers include alumina (Al2O3), strontium titanate (SrTiO3), and magnesium oxide (MgO). These buffer layers facilitate the successful integration of Ga2O3 onto Si, enabling the growth of high-quality epitaxial films suitable for advanced electronic applications. The following sections will describe these buffer layers and the epitaxial growth process in greater detail, providing a comprehensive approach to overcoming the challenges associated with integrating Ga2O3 onto Si substrates.
β-Gallium Oxide (β-Ga2O3) is an emerging wide band-gap semiconductor with promising applications in high-power electronics and optoelectronic devices. Due to its ultra-wide band-gap of approximately 5 eV, β-Ga2O3 offers significant advantages over traditional semiconductors such as Si, gallium nitride (GaN), and silicon carbide (SiC), particularly in high-voltage and high-temperature applications. Its high theoretical breakdown field (˜8 MV/cm) allows for the miniaturization of power electronics, enabling smaller, lighter, and more efficient devices. However, despite its benefits, integrating β-Ga2O3 into conventional semiconductor technologies, particularly onto Si substrates, presents challenges due to lattice mismatch and thermal incompatibilities.
Structural Properties of β-Ga2O3
β-Ga2O3 crystallizes in a monoclinic structure, which poses integration challenges when attempting epitaxial growth on substrates like silicon, which has a cubic crystal structure. The mismatch in symmetry can lead to multi-orientation growth of epitaxial domains, resulting in grain boundaries and other defects that degrade the electronic properties of the material. For high-performance device applications, it is essential to minimize these defects by optimizing epitaxial growth methods that reduce defect densities and increase grain size.
The crystal structure of β-Ga2O3 is classified under space group C2/m, with lattice parameters a=12.2 Å, b=3.0 Å, c=5.8 Å, and B=104°. The gallium ions in β-Ga2O3 occupy octahedral and tetrahedral sites within a distorted face-centered cubic oxygen sublattice. This unique structural arrangement is responsible for the material's exceptional properties, including its high breakdown field and transparency into the deep ultraviolet (UV) range. Additionally, β-Ga2O3's high radiation hardness makes it an attractive candidate for space-based applications.
Systems and Methods for Epitaxial Growth of β-Ga2O3
In an embodiment of the present technology, MBE is utilized for the growth of high-quality β-Ga2O3 thin films on Si substrates. MBE offers unparalleled atomic-level control, making it the preferred method for heteroepitaxial growth of semiconductors. However, due to the chemical and thermal instability of direct β-Ga2O3 growth on Si, a buffer layer is necessary to prevent the formation of amorphous silicon dioxide (SiO2) and to facilitate epitaxial alignment. It should be appreciated that while the buffer layer is discussed in the following paragraphs, the below “Metal-oxide Buffer Layer” section provides more details on the specifics of the buffer layer.
In one embodiment, a buffer layer is formed having a thickness between 2-10 nm on a 2-inch diameter Si (001) substrate prior to the deposition of β-Ga2O3. The buffer layer is deposited using e-beam evaporation at a temperature greater than 800° C., resulting in a smooth, single-crystal layer that matches the underlying Si substrate without causing oxidation. The use of the buffer layer not only prevents the oxidation of Si but also provides a surface that closely matches the lattice spacing of β-Ga2O3, thereby enabling high-quality epitaxial growth.
Following the buffer layer deposition, β-Ga2O3 is deposited at substrate temperatures ranging from 625° C. to 675° C. using MBE. Gallium (Ga) is evaporated from an effusion cell at 880° C., while an activated oxygen plasma source is employed to provide the oxygen flux required for the formation of Ga2O3. The total growth time and conditions are controlled to achieve a stoichiometric β-Ga2O3 layer, with post-growth characterization confirming the formation of a high-quality epitaxial film.
The β-Ga2O3 thin film may be grown to have a nominal thickness of between 10-200 nm, 20-150, 30-125, 40-100, 40-80, 40-70, or 50-60 nm using MBE. Reflection high-energy electron diffraction (RHEED) patterns can be recorded at various stages of growth to confirm the epitaxial nature of the β-Ga2O3 layer. X-ray photoelectron spectroscopy (XPS) and x-ray reflectivity (XRR) can be used to characterize the stoichiometry and thickness of the deposited layers. XPS reveals a stoichiometric composition of Ga2O3, while XRR confirms the layer thickness. In some embodiments, the β-Ga2O3 thin film may have an average sticking coefficient of 59% under the given growth conditions.
To further increase the film thickness for x-ray diffraction (XRD) characterization, in some embodiments, one or more additional layers of Ga2O3 film may be deposited at a lower substrate temperature, such as a temperature of less than 640° C. (e.g., 630° C.). With the additional layers of Ga2O3 film, a total thin film thickness may be approximately 66 nm, which is sufficient for detailed structural analysis. Symmetric θ/2θ XRD scans can be performed to reveal clear diffraction peaks corresponding to the β-Ga2O3 polymorph. Analysis of the peak intensities and positions can also confirm the presence of multiple out-of-plane orientations of β-Ga2O3, with estimated volume fractions of 71% for the (310) orientation, 13% for the (101) orientation, and 16% for the (
The epitaxial growth of β-Ga2O3 on a buffer layer may result in a textured thin film with distinct in-plane and out-of-plane orientations. Due to the monoclinic symmetry of β-Ga2O3, there are several possible growth planes, including (
The epitaxial growth of β-Ga2O3 is further complicated by the presence of twin domains, which arise due to 180° rotations about the c-axis. These twin domains can be observed in XRD and RHEED patterns and contribute to the overall complexity of the film's microstructure. By carefully controlling the growth temperature and oxygen flux during MBE deposition, it is possible to minimize the formation of twin domains and other defects.
As noted above, there are significant challenges to growing β-Ga2O3 on Si. For successful epitaxial integration of β-Ga2O3 onto Si, a metal-oxide buffer layer, also referred to herein as a “buffer layer” or “epi-oxide buffer layer,” is required since the interface between Ga2O3 and Si is not thermodynamically stable [126]. Integration via a metal-oxide buffer layer is therefore an important step that provides a high-quality, well-defined template layer for subsequent growth by a faster method such as CVD, metal-organic CVD (MOCVD), or the recently reported suboxide MBE [127-129]. By using the metal-oxide buffer layers described herein, β-Ga2O3 can successfully be grown on a Si substrate. As such, large scale β-Ga2O3 thin films may be formed for use in electrical products, thereby allowing the electrical transport properties of β-Ga2O3 epitaxial films to be leveraged.
The metal-oxide buffer layer described herein may be formed from a variety of metal oxides that grow as a single crystal on the Si surface without oxidizing the underlying Si within the growing conditions required for Ga2O3. Moreover, the metal oxide used to form the buffer layer should provide in-plane lattice spacing that closely matches two of the dimensions of the unit cell of β-Ga2O3. As described in greater detail below, metal oxides that are able to meet the above requirements include alumina, STO (SrTiO3) and MgO. Accordingly, as provided herein the buffer layer used to integrate Ga2O3 onto a Si substrate includes STO and MgO, and in some embodiments, alumina.
As noted above, one factor in selecting a respective metal oxide as a buffer layer for growing β-Ga2O3 on Si is that the metal oxide be unreactive with Si within the operating conditions required for growing the Ga2O3 thin film. As exemplified in the below Examples 2 and 5 for STO and MgO, respectively, both STO and MgO are stable and non-reactive with Si within the oxidizing conditions required to grow Ga2O3 thin film. While STO is stable on Si only at temperatures below 850° C., the STO is stable at the growing temperature of Ga2O3 which is below 800° C. MgO on the other hand is completely unreactive with Si at all practical temperatures up to the melting point of Si (˜1400° C.) due to the large formation energy of MgO.
Another factor for selecting a metal oxide as a buffer layer for Ga2O3 growth on Si is matching of at least two crystal dimensions of β-Ga2O3. That is, the selected metal oxide should match at least two of the lattice parameters of β-Ga2O3, which are a=12.2 Å, b=3.0 Å, c=5.8 Å, β=104, as noted above. STO and MgO approximately match the b and c lattice parameters of β-Ga2O3 when their interplanar spacing is doubled. For example, for STO (110), the interplanar spacing has a length of 5.52 Å, which represents an 4.8% mismatch to the β-Ga2O3 c lattice parameter of 5.8 Å, and for MgO (110), the interplanar spacing has a length of 5.96 Å which represents a 0.7% mismatch to twice the β-Ga2O3 b lattice parameter of 3.0 Å. Based on the relatively low mismatch between the interplanar spacing of the STO and MgO and the b and c lattice parameters of β-Ga2O3, both STO and MgO are suitable metal oxides for forming a buffer layer to stabilize the stacking of β-Ga2O3 along its 100 planes.
In some embodiments, matching of the lattice structure with the Si surface may also be desirable. Both STO and MgO are able to comprehensively match the lattice structure of Si. For example, the primitive surface unit cell of Si when it is unreconstructed is a square with length 3.84 Å rotated by 45 deg from the conventional unit cell. This is well matched to the lattice constant of SrTiO3 of 3.9 Å (e.g., a 1.6% mismatch) but not as closely matched with MgO (e.g., a 8.8% mismatch). However, because the MgO crystal has a large formation energy, MgO adopts a 4:3 coincident site epitaxy with Si such that there are 4 MgO conventional unit cells (4.21 Å) for every 3 Si conventional unit cells (5.43 Å), which results in a more manageable 3.3% mismatch. Accordingly, when viewed comprehensively, both MgO and STO are considered to contain a matching lattice structure to Si.
As demonstrated by Examples 2-5 provided below, STO and MgO, respectively, can each be epitaxially integrated onto Si, and hence can act as a buffer layer for growth of Ga2O3 on Si, which would then lead to the availability of large area Ga2O3 epitaxial films. Per Example 2, strontium titanate (SrTiO3 or STO) is an oxide material that can be integrated with Si substrates, and thus can be used as a buffer layer for the epitaxial growth of other materials such as Ga2O3. These examples illustrate that STO and MgO provide for structural compatibility with Si. As illustrated by Example 2, due to the lattice match between STO and Si, using STO as a buffer layer allows for high-quality epitaxial integration, which is crucial for subsequent epitaxial growth processes. As illustrated by Example 5, while MgO provides for a slightly lower match with Si, due in part to a partial plane tilt of MgO planes relative to Si, MgO still allows for the (100) orientation of Ga2O3 to remain intact when grown on top of the MgO.
Another consideration for selecting a metal oxide for the buffer layer, in some embodiments, includes the ability of that respective metal oxide to be doped. For example, STO is able to be doped with niobium (Nb), allowing it to function as a conductive template layer and as a bottom electrode, which is particularly useful in electronic devices where the Ga2O3 layer acts as the active material. This dual functionality is particularly advantageous in Ga2O3-based devices, where the buffer layer not only acts as a structural buffer but also contributes to the overall electrical performance of the device.
To integrate Ga2O3 onto a Si substrate (e.g., wafer), an epitaxial stack structure may be formed. As will be described in the following discussion, the epitaxial stack structure, which is also referred to herein as a layered stack, may include a Si substrate or wafer, a metal-oxide buffer layer, and a thin film of Ga2O3. In some embodiments, an optional seed layer of alumina (Al2O3) may be included as part of the buffer layer or on top of the buffer layer to improve lattice matching.
Referring now to
The epitaxial stack structure 100 includes a buffer layer 120 to address the challenge of directly growing the thin film 130 of Ga2O3 on the Si substrate 110. Without the buffer layer 120, the Ga2O3/Si interface formed between the thin film 130 and the Si substrate 110 becomes thermodynamically unstable, favoring the formation of silicon dioxide (SiO2). To prevent the formation of SiO2 and allow for integration of the thin film 130 onto the Si substrate 110, the buffer layer 120 may be formed. As described above, the buffer layer 120 may include a metal-oxide, such as STO or MgO.
In some embodiments, the Si substrate 110, typically a silicon, single-side-polished wafer, undergoes one or more cleaning processes. This can involve immersion of the Si substrate 110 in one or more of acetone, 2-propanol, and deionized water, with each step performed in an ultrasonic bath to effectively remove organic contaminants and particles. After the wet cleaning, the Si substrate 110 may be subjected to ultra-high vacuum (UHV) degassing to remove any residual moisture or organic residues, ensuring the surface is pristine for subsequent growth.
Once cleaned and degassed, the Si substrate 110 is loaded into an MBE chamber. In the chamber, the Si substrate 110 is exposed to a high-vacuum environment and preheated to a temperature between 800° C. and 900° C. to desorb the native SiO2 layer.
With Si surface prepared, the buffer layer 120 is epitaxially grown on the Si substrate 110. The growth process is carefully controlled, with the substrate temperature maintained between 400° C. and 800° C., which optimizes the quality of the buffer layer 120. The buffer layer 120 is grown to a thickness ranging from 5 nm to 20 nm, which provides a smooth and thermally stable interface that prevents reactions between the Si substrate 110 and any subsequent materials grown on top.
The epitaxial growth of the buffer layer 120 on the Si substrate 110 is monitored using reflection high-energy electron diffraction (RHEED) to confirm the formation of a high-quality, lattice-matched interface. This interface serves as an ideal template for subsequent growth steps, facilitating the deposition of additional layers, such as the β-Ga2O3 thin film 130, without inducing defects or compromising the structural integrity of the films.
2. Thin Film 130 of Gallium Oxide (Ga2O3) Epitaxial Growth
Following the formation of the buffer layer 120, one or more Ga2O3 thin films 130 are deposited, typically ranging from 20 to 100 nm in thickness. The Ga2O3 growth of the thin film 130 may be achieved using either PAMBE or MOCVD techniques. In the case of PAMBE, the Si substrate 110 temperature is maintained between 670° C. and 775° C., with the gallium cell heated to approximately 880° C. Oxygen plasma may be employed throughout the process to enhance the growth kinetics, leading to deposition rates of 1-4 Å/min. The oxygen plasma also increases the Ga2O3 growth rate by a factor of ˜2.8 due to improved surface kinetics compared to metal gallium. In some cases, RHEED may be employed during the growth process to monitor the crystallinity of the thin film 130 in real-time, and post-deposition analysis using XRD, XPS, and TEM ensures that the Ga2O3 thin film 130 adheres to the desired crystalline quality. Post-growth treatments such as annealing can further improve the thin film's 130 structural properties.
In some embodiments, a compositionally graded alumina (Al2O3) seed layer 125 may be introduced between the buffer layer 120 and the Ga2O3 thin film 130 to improve lattice matching. This seed layer 125 is typically grown using MBE or MOCVD, with the Al2O3 seed layer 125 thickness ranging from 5 to 20 nm. By employing an Al2O3 seed layer 125, any lattice mismatch between the buffer layer 120 and the Ga2O3 thin film 130 can be minimized, resulting in improved initial film quality.
To further optimize lattice matching, a compositionally graded AlxGa1-xO3 alloy may be used for the seed layer 125 in which the aluminum content (x) gradually transitioning from 1 to 0, enabling a smooth transition from the Al2O3 seed layer 125 to the Ga2O3 thin film 130. This approach helps reduce defects during early stages of Ga2O3 growth, ultimately improving the crystalline quality of thicker films. As with the Ga2O3 films 130, the seed layer 125 benefits from post-growth treatments like annealing and etching to improve smoothness and crystalline quality.
While initial epitaxial Ga2O3 growth on the Si substrate 110 may result in some defects due to lattice mismatch, the crystalline quality improves significantly with increasing film thickness. In some embodiments, the buffer layer 120 may include the seed layer 125 that supports further Ga2O3 growth through MOCVD or similar methods. Compositionally graded seed layers 125 such as AlxGa1-xO3 further enhance lattice matching, and post-growth treatments like annealing, polishing, and etching contribute to producing high-quality, free-standing Ga2O3 films 130 or epitaxial stack structure the 100.
Referring now to
In one embodiment, forming the buffer layer (210) may involve a multi-step process. In such embodiments, forming the buffer layer may include forming a first layer of a metal oxide on the Si substrate, such as STO or MgO, followed by forming a second layer including a second metal oxide, such as aluminum oxide. The second layer may serve as an additional seed layer, such as the seed layer 125, which can further improve the lattice matching between the gallium oxide film and the underlying substrate.
In one embodiment, the buffer layer may be formed of aluminum oxide as the metal oxide material for the buffer layer. For example, the buffer layer may be or include gamma-aluminum oxide (γ-Al2O3), which provides a suitable template for the subsequent growth of gallium oxide. In another embodiment, the buffer layer could include an aluminum-gallium oxide alloy, either in addition to or as a replacement for γ-Al2O3, to improve the lattice matching between the buffer layer and the gallium oxide thin film. In other words, in some cases the buffer layer 120 may be formed from aluminum oxide (or an alloy thereof), while in other cases, the buffer layer 120 may be formed from another metal oxide (e.g., STO or MgO), with a subsequent seed layer 125 containing aluminum oxide formed thereon.
In another embodiment, the buffer layer may be formed using a rare earth oxide such as gadolinium oxide (Gd2O3) or erbium oxide (Er2O3). These materials provide an alternative to the metal oxides described above and can be used to support the heteroepitaxial growth of gallium oxide in cases where better lattice matching is required. In such cases, the buffer layer may be formed via a first layer of rare earth oxide on the Si substrate, followed by a second layer of gallium-rare earth alloy oxide. The thin film of gallium oxide may then be grown on this layered structure to create a stable, high-quality epitaxial stack.
In some embodiments, forming the buffer layer (210) may involve creating a compositionally graded buffer layer to improve the transition between the Si substrate and the gallium oxide film.
Once the buffer layer is formed on the Si substrate (210), the flow 200 may include forming a thin film of gallium oxide, such as the thin film 130, on the buffer layer (220). The Ga2O3 film may be grown epitaxially as described above to ensure high-quality crystal formation. In some embodiments, the thin film of Ga2O3 may be formed to be or include a gamma-gallium oxide (γ-Ga2O3), while in other embodiments, the thin film of Ga2O3 may be or include beta-gallium oxide (β-Ga2O3), which can be grown either in place of or in addition to gamma-gallium oxide to enhance the functionality of the film for specific applications. In some cases, both variants of Ga2O3 may be grown, depending on the desired material properties for the final application.
In an example, forming the thin film of gallium oxide (220) may involve growing the thin film Ga2O3 layer on the seed layer. As described above, the seed layer may include aluminum oxide layer formed as γ-Al2O3, which enhances the stability of the gallium oxide growth. In another example, forming the thin film of gallium oxide (220) may involve hetero-epitaxially integrating the Ga2O3 thin film onto the Si substrate via the buffer layer.
In some embodiments, the flow 200 may also include forming one or more additional layers of gallium oxide on the initial thin film (230). In an example, forming the additional layers Ga2O3 (230) may include growing the additional layers using CVD, which enables the production of thicker, bulk-like Ga2O3 layers. For example, forming the additional layers of Ga2O3 films (230) may form bulk-like crystals that are essential for certain high-performance applications. In some embodiments an aluminum-gallium oxide alloy may formed as part of the additional thin film layers (230) to improve the quality and stability of the epitaxial stack structure. In cases where the buffer layer includes aluminum oxide, formation of the additional layers of Ga2O3 may include first forming a layer of aluminum-gallium alloy oxide, followed by further layers of Ga2O3. In cases where the buffer layer includes a rare earth oxide, the process may involve forming a layer of gallium-rare earth alloy oxide before the additional Ga2O3 layers are deposited.
As noted above, once the thin film of Ga2O3 is grown on the buffer layer, the flow 200 may include one or more additional steps such as etching, polishing, or annealing to smoothen the surface and enhance the crystalline quality of the Ga2O3 thin film.
In any of the embodiments described herein with reference to
The epitaxial stack structure 100 formed by the Si substrate 110, the buffer layer 120, and the Ga2O3 thin film 130 offers a scalable solution for growing Ga2O3 films on large-area silicon wafers (200-300 mm). This is critical for applications in power electronics and optoelectronics, where large-area, bulk-like Ga2O3 films are required. The large-scale production capability enabled by this stack structure 100 supports a wide range of applications, including the fabrication of field-effect transistors, power devices, and UV-sensitive optoelectronic devices. Moreover, the combination of β-Ga2O3 with Si provides improved heat dissipation, addressing another limitation of β-Ga2O3 in high-power applications.
The successful integration of β-Ga2O3 on Si substrates has significant implications for the development of next-generation power electronics. The ability to grow high-quality-Ga2O3 thin films on large-scale Si wafers enables the fabrication of devices that leverage the high breakdown field and thermal stability of β-Ga2O3 while benefiting from the established infrastructure of Si technology. Examples of such electronic devices are illustrated in
Accordingly, the present technology provides a scalable method for the epitaxial growth of β-Ga2O3 on Si substrates using buffer layers. By addressing the challenges of lattice mismatch, thermal instability, and defect formation, the approach provided herein paves the way for the widespread adoption of β-Ga2O3 in power electronics, UV photodetectors, and other advanced semiconductor devices.
In an embodiment of the present technology, a 5 nm γ-Al2O3 (111) layer was deposited on a 2 inch diameter Si (001) substrate with 2×1 surface reconstruction, akin to references [60-62].
Following the characterization, the alumina (“Al2O3”)-buffered 2 inch Si wafer was re-loaded into the ultra-high vacuum (UHV) system after ex situ cleaning with acetone, 2-propanol, de-ionized water, drying with N2 and UV/ozone for 30 min. Ga2O3 was then deposited at 670° C. substrate temperature. Ga (99.99% purity) was evaporated from a dual filament (hot lip) effusion cell held at 880° C., and an O plasma was generated by an rf-plasma source operated at 2.0×10−5 torr O2 partial pressure and 200 W plasma power. The total pressure was monitored by an ion gauge located in the upper part of the molecular-beam epitaxy (MBE) chamber (above the sample), whereas a residual gas analyzer/quadrupole mass spectrometer (RGA/QMS) located at the side of the chamber (below the sample) measured an O2 to O ratio of approximately 10, from which an absolute O flux of ˜10 nm−2s−1 is estimated. The sample exposed to oxygen plasma was annealed for 30 min at 670° C. prior to growth. The Ga2O3 deposition rate on the quartz crystal microbalance (QCM) sensor before the anneal was 3.15 Å/min.
The film was grown for 174.5 min for a nominal Ga2O3 thickness of 55 nm. Post-growth RHEED showed a streaky surface with qualitatively similar pattern as the underlying γ-Al2O3, although with more modulation along the streaks as shown in
The resulting Ga2O3 film was too thin for comprehensive characterization with a laboratory x-ray diffractometer (Al and Ga are weak Rayleigh x-ray scatterers [69]). Hence, additional Ga2O3 was deposited. The 2 inch wafer was first diced into 10×10 mm2 pieces. After dicing, a 10×10 mm2 piece obtained from the center of the wafer was re-cleaned following the above cleaning protocol and loaded into UHV on a molybdenum sample holder. After a 40 min anneal at 630° C. in oxygen plasma, an additional Ga2O3 layer was deposited at the lower substrate temperature of 630° C. to enhance the sticking probability [70], with GaOx rate ˜0.3 nm/min, at otherwise identical growth conditions for 256 min (76.8 nm additional Ga2O3 nominally). The XRR revealed an additional Ga2O3 thickness of 45 nm (see “Supplemental Material For Example 1,” below), indicating an increased average sticking probability of 59%, leading to a total film thickness of ˜66 nm, sufficiently thick for XRD characterization. RHEED post-growth shown in
In order to identify the polymorph and orientation of the stoichiometric and crystalline Ga2O3 layer, an ex situ θ/2θ XRD scan with sample alignment to Si (004) was conducted on a Rigaku Ultima IV diffractometer with Cu-Kα anode, parallel beam geometry (multilayer mirror), and a 2-bounce Ge monochromator. The scan was recorded with slit settings of 1.0 mm for the divergence slit and 0.5 mm for both scattering and receiving slits, without the use of Soller slits. The settings were chosen to enhance the weak film signal and allow the observation of lattice planes that have normals slightly tilted away from the Si surface. A 5.0 mm height-limiting slit was used to limit the spread of the beam perpendicular to the scattering plane and keep the x-ray beam from spilling over the sample edges.
Fitting of the peak profiles to pseudo-Voigt functions is presented in
The volume fractions obtained indicate that there are differences in nucleation and growth rate for the different basal growth planes. A non-uniform vertical distribution of the three out-of-plane orientations is also possible. By using the full-width-at-half-maxima of the peak profiles in the Debye-Scherrer equation, the coherence length of the scattering crystallites along the out-of-plane direction is estimated to be ˜21 nm for the (
The crystal structure of β-Ga2O3 has a low symmetry, base-centered monoclinic C2/m space group (12) with only one mirror plane (perpendicular to b-axis in standard setting [71]), one two-fold rotation axis (parallel to b-axis), and an inversion center (origin) as its non-trivial symmetry operations. However, β-Ga2O3 is characterized by oxygen ions forming a sublattice with a distorted fcc arrangement [72], where the smaller gallium ions fill interstitial spaces of tetrahedral (Ga1) or octahedral (Ga2) symmetry.
By looking at the distorted tetrahedra (blue, labeled 3050 in
The basal growth planes under discussion can also be thought as of being the close-packing planes of the distorted fcc O sublattice of β-Ga2O3, i.e. the planes that are perpendicular to the body-diagonal directions of the fcc cage in
Due to the missing global three-fold symmetry (indeed absence of any non-trivial rotational symmetry) of β-Ga2O3 when rotated about the basal growth plane normals, each distinct growth plane should give rise to a minimum of six in-plane variants of β-Ga2O3 when grown on the six-fold symmetric surface of γ-Al2O3 (111) shown in
Representative atomistic models of feasible β-Ga2O3/γ-Al2O3 interfaces with alignment of the oxygen ions in the interface plane are presented in
The crystallographic relationships between γ-Al2O3 and Si are (
To confirm the presence of 48 in-plane variants, a full pole figure scan with 2θBragg angle of 31.7° was conducted on the Rigaku Ultima IV diffractometer with use of parallel beam in a non-coplanar measurement geometry with an in-plane detector axis [73]. The scan was done with 2.5° vertical Soller slits on both source and detector sides, 0.5 mm divergence slit, 5.0 mm width-limiting slit, 1.0 mm receiving slit, and 1.0 mm scattering slit. A thin Ni foil was used to block out Cu—KB radiation. The 2θ angle of 31.7° corresponds to the (002) and (202) reflections of β-Ga2O3, which are strong intensity reflections of β-Ga2O3 with only two-fold multiplicity, which makes them suitable for a pole figure scan.
The resulting pole figure is shown in
The polar angle (measured from the north pole downwards) is represented linearly on the radial axis of the plot of
The morphology of the β-Ga2O3 thin films was observed using an image-corrected FEI Titan 80-300 high-resolution electron microscope (HREM) operated at 300 kV and a probe-corrected JEOL ARM200F scanning transmission electron microscope (STEM) operated at 200 kV. Suitable cross-section samples were prepared in the form of thin (50-100 nm) lamellae using focused-ion-beam milling and lift-out techniques, and all images were recorded with the incident electron beam aligned parallel with a Si zone axis.
Depending on the growth method and temperature, different ratios of (
In conclusion, strongly textured growth of β-Ga2O3 with (
This Supplemental Material contains additional data and analysis for the above described Example 1, including growth rate analysis by QCM and XRR, stoichiometry analysis by in situ XPS, additional XRD experiments and supporting analysis and supplemental tables for the XRD data of this Example 1, a further discussion on the growth model and the oxygen to gallium ratio, and finally density functional theory surface energy calculations of growth surfaces discussed above.
For a homogenous thin film f containing the element f1 with smooth coverage on the substrate s containing the element s1 we can write for the ratio Af,f1nl/As,s1n′l′ of measured photoelectron intensity of subshell n1 of element f1 to the measured photoelectron intensity of subshell n′l′ of element $1 [80]:
Here T is the analyzer response function also denoted as transmission function, EKf1,nl is the kinetic energy of the ejected photoelectron from element f1 and subshell nl, λf (λs) is the effective attenuation length (EAL) for electrons in the film (substrate), d is the film thickness, σf1,nl is the photoelectric cross section of element f1 and subshell nl, Mf (Ms) is the molar mass of the molecules comprising the film (substrate) material, nf1 (ns1) is the stoichiometric coefficient of element f1 (s1) in the molecules comprising the film (substrate). This equation can be solved numerically as an estimate of the thin film thickness d, as long as d≤5λ. The transmission function of our VG Scienta spectrometer was characterized, for the used experimental conditions, by a method devised by the National Physical Laboratory (NPL) of the United Kingdom [81], where survey spectra from in situ Ar sputter-cleaned copper, silver and gold samples were acquired and then referenced back to a spectrum determined on a well characterised spectrometer at NPL. The binding energy (BE) scale was calibrated using the Au 4f7/2, Ag 3d5/2 and Cu 2p3/2 core-levels [82]. The EAL is estimated by using the so-called CS2-formula [80, 83]. For the theoretical photoelectric cross-sections we use the Scofield values [84].
Once the thickness d of the film layer is known (by the above estimate or another more accurate method like XRR) we can evaluate the stoichiometry of the film. The ratio of the stoichiometric coefficients nf2/nf1 of elements f1 and f2 contained in the film layer is given by [80]:
Using this approach of determining the stoichiometry from fundamental parameters we found that it is necessary to adjust the O 1s Scofield cross-section [84], empirically to 0.75 times its tabulated value for a variety of investigated oxides (including Ga2O3, Al2O3, SrTiO3, La2O3, Nb2O5) in order to arrive at the correct oxygen to metal ratio for these materials. γ-Al2O3 growth on Si (001)
Silicon (001) wafers with 2 inch diameter (50.8 mm), n-type phosphorous doped, single-side polished of 200 μm thickness were used as substrates (Virginia Semiconductor). The substrates were ex situ cleaned by immersion in solvent and sonication for 10 min each with acetone, 2-propanol and de-ionized water. Thereafter the substrate was blow dried with N2 and exposed to an ozone/UV treatment for 15 min to break-up residual hydrocarbons on the surface, after which the substrate was loaded into the UHV system on a molybdenum sample holder without backplate. The UHV system consisted of a DCA MBE chamber connected via a buffer line to a VG Scienta XPS system. The SiO2 layer was desorbed in situ in the MBE chamber using the Sr desorption method [85], resulting in a 2×1 surface reconstruction of Si (001) observed by reflection high-energy electron diffraction (RHEED) shown in
The alumina source consisted of alumina pellets (1-3 mm pieces of 99.99% purity). Alumina was deposited on the substrate at 840° C. without supplying additional oxygen for two hours. This resulted in a stoichiometric Al2O3 layer of 5.0 nm without the formation of SiO2 at the Al2O3/Si interface as determined ex situ by x-ray reflectivity (XRR) and in situ by x-ray photoelectron spectroscopy (XPS) (see
The diffraction patterns discussed above with reference to 11
γ-Al2O3[100]∥Si and
110
γ-Al2O3∥[100] Si. This geometry results in a surface consisting of equilateral triangles of oxygen ions, as displayed in
The γ-Al2O3 lattice constant was calculated as @Al2O3=7.89 Å by measuring (4
Ga2O3 Growth and Characterization
However, based on intensities this polymorph and orientation can likely be excluded. The γ-, κ-, and β-polymorphs in the given orientations are all likely candidates based on their Q/Q0 values. However, we also see that within the considered possibilities, the harmonic Peaks 2 and 5 are only explained well by β-Ga2O3 in (310)-orientation. Based on intensities, a mixture of β-Ga2O3 epitaxial orientation variants seems to explain the data best. The absence of Peak 6 in the experimental data seems plausible for B Ga2O3 due to the low intensity of the (804)/(404) peaks; this is not so for γ- and κ-polymorphs. Therefore, we deem it highly likely that all peaks originate from β-Ga2O3. This is corroborated by the elevated growth temperatures and the pole figure scan shown in the main text of the article. By means of the pole figure scan we can exclude significant volume fractions of the α-, γ- and κ-polymorphs and confirm β-Ga2O3 with 48 epitaxial orientation variants as described in the main article.
We would like to mention here that the presence of the (101) phase can be easily overlooked in a θ/2θ scan since its peaks overlap with the peaks of the (
Density functional theory (DFT), as implemented in the Vienna ab initio Simulation Package (VASP) code, is used to perform all calculations [99]. The Perdew-Burke-Ernzerhof (PBE) generalized-gradient approximation (GGA) is used for the exchange correlation energy functional. We use projector-augmented-wave potentials [100], to describe Ga and O with a cut-off energy of 500 eV for bulk optimizations. The valence electron configurations used are 4s23d104p1 for Ga and 2s22p4 for O. Each self-consistent electronic calculation is converged to within 10−6 eV per cell, and the ionic relaxation is iterated until the forces are less than 0.01 eV/Å. For the Brillouin zone integration of bulk Ga2O3 (space group C2/m, 12), a 5×9×11 Monkhorst-Pack grid is used. For bulk Ga (space group Abma, 64), a 12×12×12 Monkhorst-Pack k-point mesh is used.
The lattice parameters calculated for bulk Ga2O3 are a=12.445, b=3.084 Å, c=5.876 Å, and β=103.73°, which compare favorably with the experimental values of a=12.22 Å, b=3.04 Å, c=5.80 Å and β=103.75° [101], and theoretical values of a=12.446, b=3.083 Å, c=5.876 Å, and β=103.70° [102].
To create the slabs, we used surface-oriented basis transformations on the conventional unit cell using methods described in reference [103]. The transformation matrices used are shown in Table 7, below. Each unit cell is repeated along the c-axis to create a 15-20 Å thick supercell and a 15 Å vacuum is added to create the slab. For slab optimizations, a 600 eV energy cutoff is used, and the convergence criteria are the same as those used for the bulk optimizations. For Brillouin zone integration of the (
We calculate the surface energies using the following equation,
In this equation, Ysurface is the surface energy, A is the surface area with a factor of 2 accounting for both surfaces, Eslab is the energy of the Ga2O3 slab, NX is the number of X atoms in the Ga2O3 slab, EX is the energy of one unit cell of bulk X, and μX is the chemical potential of bulk or molecular X.
To simplify the number of variables in this equation, we use the relationship:
We assume that the surface is in equilibrium with its own bulk, hence μGa
Using this definition and rearranging the equation above, we get the relationship
We can then allow the chemical potential to vary over the range
where uO
represents Ga-rich conditions. We determined that the formation energy of Ga2O3 is −9.31 eV, which is less stable than the experimental value of −11.3 eV [104], but comparable to the formation energy value of −9.3 eV calculated by Zacherle et. al [102].
To determine the lowest energy surface reconstruction of the (310) plane, we studied one Ga-terminated and one O-terminated surface (
For epitaxial integration, a buffer layer is essential since the interface between Ga2O3 and Si is not thermodynamically stable [126]. Integration via an oxide buffer layer is therefore an important step that provides a high-quality, well-defined template layer for subsequent growth by a faster method such as CVD, metal-organic CVD (MOCVD), or the recently reported suboxide MBE [127-129]. SrTiO3 (001) can be routinely epitaxially integrated onto Si (001) [130], and hence it could act as a buffer layer for growth of Ga2O3 on Si, which would then lead to the availability of large scale Ga2O3 epitaxial films. Electrical transport properties of β-Ga2O3 films grown on STO (100) were recently reported by Wang et al. and we expect films grown on STO-buffered Si to behave similarly [131].
Here we report the deposition of β-Ga2O3 thin films grown on both STO (001) and STO-buffered Si (001) substrates using PAMBE and investigate their structural properties in detail. STO has the attraction of being one of the few oxide materials that can routinely be directly integrated with Si (001) [130]. Since it can be n-type doped with Nb [132], STO can also function in the dual role of epitaxial template layer for Ga2O3 and as conductive bottom electrode. The β-Ga2O3 thin films have been characterized by RHEED, XPS, XRD, XRR, reflection electron-energy-loss spectroscopy (REELS), and TEM.
In an embodiment of the present technology, single-side-polished STO (001) substrates with dimensions of 10×10×0.5 mm3 were cleaned ex situ in acetone, 2-propanol, and de-ionized water, each for 10 min in an ultrasonic bath. The samples were loaded into the MBE growth chamber (base pressure 5×10−10 torr) on Mo sample holders held in place by a Mo spring clip exerting pressure onto the backside of the sample, and a ˜0.5 mm protruding lip on the growth side of the sample. The sample was heated radiatively from the backside via a SiC heating coil placed above the sample with a set point of 775° C. (corresponding to a calibrated substrate temperature of ˜670° C. measured for a Si substrate with a pyrometer) in an oxygen plasma for 30 min. The oxygen plasma was generated by an Oxford Applied Research HD25 radio-frequency plasma generator at 200 W forward power and 2.0×10−5 torr nominal O2 pressure measured by an ion gauge located at the top of the growth chamber. A residual gas analyzer (RGA) located on the sidewall of the chamber confirmed that O2 was the predominant gas species, with an atomic oxygen amount of about 10% at the RGA position. Before introducing oxygen into the chamber, the dual filament Ga effusion cell was heated to 900° C. and outgassed for 20 minutes. The cell temperature was then lowered to 880° C. and the Ga rate was measured in the high-vacuum atmosphere of typically low 10−8 torr (caused by the hot effusion cell, and predominantly H2 as measured by the RGA) by a QCM. Typical metal growth rates at this cell temperature were about 1 Å/min. After ignition of the oxygen plasma, the Ga rate was measured again, usually achieving a deposition rate that was a factor of ˜2.8 higher than for the bare Ga rate (using the same density setting for the QCM of 5.95 g/cm3 since Ga and Ga2O3 densities are nearly identical). This difference is most likely due to a lower sticking coefficient of bare Ga due to larger surface tension. The Ga2O3 growth was performed under the same Ga cell and O-plasma conditions at a manipulator set-point temperature of 775° C. This growth temperature was selected as it is the highest temperature that gives unity sticking coefficient for our oxygen flux, based on the model developed by Vogt and Bierwagen [133]. Higher substrate temperatures provide better epitaxy but should not be too high as to result in significant Ga suboxide desorption. The STO buffer layer on Si was grown on a 2-inch Si (001) wafer [132], and the Ga2O3 film was grown on this pseudo-substrate under nominally the same conditions as used for the bare STO substrates. The QCM-measured GaOx growth rate matched the observed rate of Ga2O3 on bare STO, as confirmed by ex situ XRR thickness measurements. The film surfaces grown on bare STO were monitored in situ during growth by RHEED with a 21-keV electron beam and in situ after growth in a secondary analysis chamber with a 15-keV electron beam for the STO-Si substrate. The thin film stoichiometry was measured in situ and post-growth by XPS with monochromatic Al-Kα and a VG Scienta R3000 hemispherical analyzer set at 200-eV pass energy and 0.8-mm slit size. The sample grown on STO-Si was also studied by REELS with 1.9-keV electrons and a Staib Auger probe. To determine the sample thickness, morphology, and crystal structure, XRD and XRR measurements were performed ex situ on a Rigaku Ultima IV diffractometer (Cu-Kα). TEM observations were made using an image-corrected Titan 80-300 operated at 300 kV.
100
and
110
azimuths for a 20-nm film, while
XPS measurements conducted in situ after growth indicated stoichiometric Ga2O3 films, as shown in
The epitaxial relationships observed can be written as: β-Ga2O3 (100) ∥ SrTiO3 (110), and β-Ga2O3 (
The epitaxial texture of the films is confirmed by cross-sectional TEM images displayed in
Based on the observed epitaxial relationships, a structural model is presented in
Motivated by the epitaxial matching, despite rotational symmetry mismatch, of monoclinic β-Ga2O3 and cubic SrTiO3 (001), we have performed epitaxial integration with a 2-inch-diameter Si wafer using a SrTiO3 buffer layer under nominally identical growth conditions as used for the bare STO substrate. The structural, morphology, and thickness characterization results are shown in
Additional electron-energy-loss spectroscopy (EELS) experiments were performed on the β-Ga2O3 film grown on Si. A band gap of 4.5-5.2 eV is extracted from the fits, consistent with literature data for Ga2O3 [106]. For details see
Unlike the MOCVD growth of Ga2O3 on STO [122], in the present study we observe crystallization of Ga2O3 as-grown with the use of MBE and growth temperature of 670° C. Beside the (100)-orientation, we additionally observe the (
The (
To calculate the surface energy, we used the following relation:
In this equation, Ysurface is the surface energy, A is the surface area with a factor of 2 accounting for the top and bottom surfaces, Eslab is the energy of the Ga2O3 slab, NGa (NO) is the number of Ga (O) atoms in the Ga2O3 slab, EGa (EO
The details of the methods used for these calculations are outlined, below, in the section entitled “Supplemental Material for Example 2.” For the (100) surface, we used two stoichiometric slabs and found that the surface energy of the mixed-terminated surface is 0.5 J/m2 and it is 0.8 J/m2 for the O-terminated surface. The surface energy values we have calculated are lower than the values reported by Bermudez who used different computational methods. However, we do find that the 0.3 J/m2 difference in energy between the (100)-A and (100)-B is comparable to the energy difference of 0.5 J/m2 reported by Bermudez. To our knowledge, calculations of the (
The horizontal RHEED spot spacings in
The epitaxial growth of (100) β-Ga2O3, even in the case of homo-epitaxy on a non-vicinal substrate, gives rise to stacking faults and twin boundaries (mirror about (100) with a c/2-glide) [137]. The mirror twin and stacking faults could possibly be avoided by using a vicinal (100) substrate with off-cut along at an optimal angle of ˜6°. Such substrates give rise to terrace steps formed by the energetically low (
In summary, we have demonstrated the successful epitaxial integration of β-Ga2O3 on SrTiO3 (001) and SrTiO3-buffered Si (001). The films are crystalline as grown. Two basal growth planes (100) and (
This supplementary material for Example 2 provides additional supporting figures and tables, electron energy loss spectroscopy measurements, and for the details of density functional calculations.
The fitted peak positions and plane spacings in
These cells shown in
Additional electron-energy-loss spectroscopy (EELS) experiments were performed on the β-Ga2O3 film grown on Si.
To perform our calculations, we used density functional theory (DFT) as implemented in the Vienna ab Initio Simulation Package (VASP). [146] For the exchange energy correlation functional, we used the generalized-gradient approximation (GGA) parametrized by Perdew-Burke-Ernerhof (PBE). We use a cutoff energy of 500 eV for all bulk optimizations and 600 eV for all slab calculations. We use projector-augmented-wave potentials to describe Ga and O and the valence electron configurations used are 4s23d104p1 for Ga and 2s22p4 for O. Each self-consistent electronic calculation is converged to within 10−6 eV per cell, and the ionic relaxation is iterated until the forces are less than 0.01 eV/Å. For the Brillouin zone integration of bulk Ga2O3 (space group C2/m) and bulk Ga (space group Abma), 5×9×11 and 12×12×12 Monkhorst-Pack k-point grids are used, respectively. The lattice parameters calculated for bulk Ga2O3 are a=12.445, b=3.084 Å, c=5.876 Å, and B=103.73°, which compare favorably with the experimental values of a=12.22 Å, b=3.04 Å, c=5.80 Å, and β=103.75° and theoretical values of a=12.446, b=3.083 Å, c=5.876 Å, and β=103.70°. [149]
The transformation matrices used to create unit cells for each surface were derived using the methods described in reference and are provided in Table 10. For Brillouin zone integration of the (100) and (
To calculate the surface energy values listed in Table 11, we used the following relation:
In this equation, Y surface is the surface energy, A is the surface area with a factor of 2 accounting for the top and bottom surfaces, Eslab is the energy of the Ga2O3 slab, NGa (NO) is the number of Ga (O) atoms in the Ga2O3 slab, EGa (EO
To simplify the number of variables in this equation, we use the relationship:
Assuming the surface is in equilibrium with its own bulk, μGa
Using this definition and rearranging the previous equation, we define the chemical potential of bulk Ga as
We can then allow the chemical potential to vary over the range
where μO
represent O-rich and Ga-rich conditions, respectively. We determined that the formation energy of Ga2O3 is −9.31 eV which is comparable to the formation energy value of −9.3 eV calculated by Zacherle et al. [149], but lower in magnitude compared to the experimental value of −11.3 eV [151].
We present a route to integrate crystalline gallium oxide grown epitaxially on a silicon (001) substrate via an epitaxial strontium titanate (STO) buffer layer. Specifically, we produce samples with the following layer stack: silicon (001) substrate/thin film SrTiO3 (buffer layer)/thin film Ga2O3. Such a materials system could be used as a seed layer for thick bulk-like MOCVD-grown Ga2O3 films or layers directly integrated on silicon, or can yield, via an etching process, free-standing gallium oxide films that can be transferred to other substrates. Gallium oxide is likely to find use in a wide variety of microelectronic (e.g., integrated circuits), micro-electromechanical system (MEMS), optoelectronic and power-electronics applications and devices. The STO buffer can be as thin as 5 nm or can be thicker (up to 20 nm) while the Ga2O3 epitaxial seed layer can range from 20 nm to 100 nm.
The layer stack according to the present technology has not previously been prepared, to the best of our knowledge. Present technologies for power-electronics applications tend to focus on millimeter-sized bulk-crystals of Ga2O3 grown by conventional crystal growth methods or thin films grown on isostructural substrates. The hetero-epitaxial integration of Ga2O3 onto silicon allows processing of gallium oxide technology in state-of-the-art semiconductor fabs specialized in the processing of silicon devices on large scale wafer diameters allowing for cost-effective production.
Ga2O3 cannot be directly grown on silicon due to a thermodynamically unstable interface favoring the formation of SiO2 hence the SrTiO3 seed, or buffer, layer is a crucial part of the present technology of epitaxial integration of Ga2O3 directly onto Si. The layer stacks according to the present technology enable the growth of large scale Ga2O3 bulk-like films on the Si wafer platform by serving as a template for subsequent fast deposition techniques like MOCVD or sputtering. An additional etching and polishing process could produce large scale free-standing Ga2O3 films that can be layer transferred to other substrates for use in power-electronics devices. Power-electronics devices may require large area, bulk-like, Ga2O3 crystals and therefore a substrate like silicon that is available in large wafer diameters (200-300 mm) is more suitable than the available isostructural substrates commonly used for hetero-epitaxy of Ga2O3. “Bulk-like” in this context refers to the Ga2O3 being epitaxial rather than polycrystalline, and with thicknesses on the order of microns.
The present technology has the advantage of integrating the wide-gap semiconductor Ga2O3 directly on the Si semiconductor platform for which many etching and processing technologies are widely available. The current hetero-epitaxial growth of Ga2O3 are limited to the small wafer sizes of the currently available substrate materials.
Since the silicon/STO platform is not very well lattice matched to Ga2O3, initially grown films are defective. However, if they act as a seed layer for further deposition of Ga2O3 by MOCVD or a similar method, this problem can be overcome as it is expected that the crystalline quality will drastically improve with increasing film thickness as long as there is a highly ordered nucleation layer. Furthermore, post-processing methods like annealing can further improve crystalline quality and polishing or etching can smoothen out a rough surface of such grown films to obtain high-quality free-standing Ga2O3 or Si/STO/Ga2O3 stacks. Another method to improve the lattice matching is to use a compositionally graded AlxGa1-xO3 alloy as the initial seed layer, as Al2O3 has a smaller lattice constant more closely-matched with STO.
Ga2O3 is a material with prospective use in high-power-electronics. In order to be useful for power-electronics applications, bulk-size crystals are needed. The size of currently grown epitaxial thin films is limited by the substrate wafer size.
The present technology comprises the epitaxial integration of a Ga2O3 layer on a silicon (001) substrate by means of a thin SrTiO3 buffer layer. The present technology also enables the integration and processing of Ga2O3 directly into the silicon processing line and allows for growth on large scale wafer substrates.
Layer stacks according to the present technology would allow the growth of large scale Ga2O3 bulk-like films on the Si wafer platform, which are required for power-electronics applications of Ga2O3. The present technology opens the door for a multitude of new use cases that have not yet been considered. Ga2O3 is also a potential semiconductor for optoelectronic applications due to its large bandgap which makes it transparent into the UV range. The epitaxial integration onto silicon layer enables the construction of a wide range of novel devices relying on the electronic and optical properties of Ga2O3. Potential examples include field-effect transistors, semiconductor devices, and sensor technology.
As shown in
In one embodiment, gallia was grown on STO buffered Si under the following process conditions. The STO-buffered Si wafer was ex situ cleaned (acetone, 2-propanol, distilled water, each for minutes in sonicator). UHV was loaded on a molybdenum sample holder. Degassing was performed under O-plasma exposure (200 W, 3×10−5 torr O2 partial pressure). Ga2O3 growth was done at 775° C. with O-plasma and Ga cell temperature of 880° C. Then, a cool down in O-plasma to 200° C. was done. In situ analysis included RHEED and XPS (using floodgun), and ex situ measurements were done using AFM, XRR and XRD.
(100)-oriented β-Ga2O3 thin film is formed on the (100) MgO substrate with a fourfold domain structure rotated every 90°. In addition, the c-axis direction of β-Ga2O3 was parallel to the <011> direction of MgO. These results were the same as that observed for the crystal orientation of β-Ga2O3 formed on the (100) MgO substrate, as shown in reference [66], although different methods for film formation were employed. This should show the in-plane spacings (020) and (11-2) in RHEED & XRD at ˜45° (actual separation 46.3°) azimuthal direction.
Table 11, below, provides the expected 20 peak positions in an out-of-plane XRD scan for the example.
Measured XRD reflections in the example indicated a mix of β and γ polymorphs. The following relation provides lattice constants d obtained from the Miller indices (hkl) for a monolithic crystal:
where
and where
ξ4=√{square root over (ξ1ξ3)} cos β
For the γ polymorph, a=8.01 Å (in-plane 20 nm film), and a=8.3 Å (oop RSM 50 nm film), compared to the bulk value of 8.23 Å. For the β polymorph, reflections were not sufficient to determine all lattice parameters (need on more lattice parameter off-symmetry). The b lattice parameter can be extracted as 2.99 Å (20 nm film), compared to the bulk value of 3.04 Å.
Interfacing between the monoclinic β-gallia and silicon requires a buffer that grows epitaxially on silicon while at the same time having structural commonality with the β-gallia crystal structure. SrTiO3 (STO) is well-known to be able to grow epitaxially on Si (100) also with 100 orientation. The b lattice vector and twice the c lattice vector match reasonably well with the STO (110) spacing. Thus, the β-Ga2O3 may grow with its bc-plane epitaxially aligned on the STO (001) surface with a 45° rotation. Additionally, the oxygen sublattice in this orientation of β-Ga2O3 is only a slightly distorted continuation from that in STO, which can allow for a low energy interface.
Another approach is to grow a buffer layer that has a crystal structure which Ga2O3 can also take. While the β-gallia structure is unique to Ga2O3, Ga2O3 can also undergo the corundum (α), spinel (γ), and bixbyite (d) crystal structures. Both spinel (γ-Al2O3) and bixbyite (rare earth oxides) materials have been demonstrated as epitaxial films on Si(100) previously. Such buffers may force the Ga2O3 to initially take these alternative crystal structures and transition to its normal β-gallia structure as it grows thicker. Because the atomic distortions among the different polymorphs is not large, such buffers may enable the continuation of the atomic structure from the buffer to the β-gallia structure. In the case of γ-Al2O3, it has been reported that when grown on Si (100), it initially grows pseudomorphically as γ-Al2O3 (100), but surface energy considerations cause it to transition to (111) orientation. The oxygen framework of γ-Al2O3 (111) (which is the same as corundum 0001) induces the (
In some embodiments, use of compositional grading in the case of spinel or bixbyite buffers may alleviate lattice mismatch. For example, after the γ-Al2O3 layer, one may then switch to an intermediate AlGaO3 composition before putting pure Ga2O3. In another embodiment, one can also use continuously graded layers AlxGa2-xO3 with x going from 1 to 0 smoothly, The same concept can be used for bixbyite oxides, e.g., Gd2O3 initially then GdxGa2-xO3 then Ga2O3. Additionally, use of vicinal surfaces may eliminate other domains, as discussed above in Example 2.
In an experiment, to grow a Ga2O3 film on MgO-buffered Si, the Ga2O3 film was first grown on a single-side polished 10 mm×10 mm×0.5 mm MgO single crystal using rf magnetron sputtering to confirm that the deposition process and lattice matching would work. Prior to Ga2O3 deposition, the as-received MgO substrate was exposed, without any solvent cleaning, to 30 min of UV/ozone to volatilize any adsorbed hydrocarbon contaminants [154]. The MgO substrate was then outgassed at a substrate temperature of 700-800° C. in high vacuum (˜10−7 Torr) for 15 minutes to outgas the surface. After this substrate preparation, the Ga2O3 growth was done using off-axis rf magnetron sputtering in a gas mixture with 5 to 20% oxygen and with argon for the balance. The forward power used for the Ga2O3 deposition was 75 W (37 W/cm2 power density). A substrate temperature of 700-800° C. and a total sputtering pressure of 15-25 mtorr were used for this process. These growth conditions resulted in a growth rate of 5.4 nm/min. It is theorized that other methods of depositing Ga2O3 should work as well so long as the growth rate is not too high to prevent epitaxy.
After confirmation that the Ga2O3 film can be grown in a MgO without significant formation of undesirable orientations, we then grew the Ga2O3 film on MgO-buffered Si. The MgO buffer was formed by first outgassing the Si wafer at 700° C. then thermally desorbing the native SiO2 layer at 900° C. Once a clean Si (001) surface is achieved, the substrate temperature is reduced to 350° C. and MgO is evaporated from single crystal pieces of MgO using electron beam evaporation. The MgO buffer layer is grown to a thickness of 8-10 nm prior to Ga2O3 deposition by off-axis sputtering, using the same growth conditions as for the trial growth on MgO single crystal.
Additionally,
Larger breakdown fields can lead to the miniaturization of power electronic devices with associated reductions in cost and weight. The enhanced radiation hardness of Ga2O3 [6] additionally makes it suitable for space applications. Today an estimated 30% of all electricity flows through power electronics and this is projected to reach 80% in the future [7]. Efficient n-type doping of Ga2O3 can be achieved by Si, Ge, Sn and Nb incorporation [3]. Recently p-type doping with H has been demonstrated [8]. This, in combination with a way to create shallow p-type doping, could lead to realization of new kinds of optoelectronics operating in the deep-UV due to the large band gap of Ga2O3.
One or more additional layers of gallia may be formed on the thin film 6030 including a gallium oxide. In the embodiment illustrated in
In some embodiments, as shown for example in
One or more additional layers of gallia may be formed on the thin film 6130 including a gallium oxide. In the embodiment illustrated in
Epitaxial growth onto Si wafers would open up numerous avenues for the large scale integration but, unfortunately, Ga2O3 cannot be directly grown on Si. In additional to the examples of power electronics devices described above with reference to
The phrases “in some embodiments,” “according to some embodiments,” “in the embodiments shown,” “in other embodiments,” and the like generally mean the particular feature, structure, or characteristic following the phrase is included in at least one implementation of the present technology and may be included in more than one implementation. In addition, such phrases do not necessarily refer to the same embodiments or different embodiments.
Unless the context clearly requires otherwise, throughout the description and the claims, the words “comprise,” “comprising,” and the like are to be construed in an inclusive sense, as opposed to an exclusive or exhaustive sense; that is to say, in the sense of “including, but not limited to.” As used herein, the terms “on,” “connected,” or “coupled” means having any attachment, connection or coupling, either direct or indirect, between two or more elements; the attachment, coupling. or connection between the elements can be physical, logical, or a combination thereof. Similarly, the phrase “directly on” means a direct attachment, connection, or coupling without any intermediate elements, layers, etc. Additionally, the words “herein,” “above,” “below,” and words of similar import, when used in this application, refer to this application as a whole and not to any particular portions of this application. Where the context permits, words in the above Detailed Description using the singular or plural number may also include the plural or singular number respectively. The word “or,” in reference to a list of two or more items, covers all of the following interpretations of the word: any of the items in the list, all of the items in the list, and any combination of the items in the list.
The above Detailed Description of examples of the technology is not intended to be exhaustive or to limit the technology to the precise form disclosed above. While specific examples for the technology are described above for illustrative purposes, various equivalent modifications are possible within the scope of the technology, as those skilled in the relevant art will recognize. For example, while processes or blocks are presented in a given order, alternative implementations may perform routines having steps, or employ systems having blocks, in a different order, and some processes or blocks may be deleted, moved, added, subdivided, combined, and/or modified to provide alternative or subcombinations. Each of these processes or blocks may be implemented in a variety of different ways. Also, while processes or blocks are at times shown as being performed in series, these processes or blocks may instead be performed or implemented in parallel, or may be performed at different times. Further any specific numbers noted herein are only examples: alternative implementations may employ differing values or ranges.
The teachings of the technology provided herein can be applied to other systems, not necessarily the system described above. The elements and acts of the various examples described above can be combined to provide further implementations of the technology. Some alternative implementations of the technology may include not only additional elements to those implementations noted above, but also may include fewer elements.
These and other changes can be made to the technology in light of the above Detailed Description. While the above description describes certain examples of the technology, and describes the best mode contemplated, no matter how detailed the above appears in text, the technology can be practiced in many ways. Details of the system may vary considerably in its specific implementation, while still being encompassed by the technology disclosed herein. As noted above, particular terminology used when describing certain features or aspects of the technology should not be taken to imply that the terminology is being redefined herein to be restricted to any specific characteristics, features, or aspects of the technology with which that terminology is associated. In general, the terms used in the following claims should not be construed to limit the technology to the specific examples disclosed in the specification, unless the above Detailed Description section explicitly defines such terms. Accordingly, the actual scope of the technology encompasses not only the disclosed examples, but also all equivalent ways of practicing or implementing the technology under the claims.
To reduce the number of claims, certain aspects of the technology are presented below in certain claim forms, but the applicant contemplates the various aspects of the technology in any number of claim forms. For example, various aspects may be presented in other system claims, composition of matter claims, method claims, or in other forms, such as being embodied in a means-plus-function claim. Any claims intended to be treated under 35 U.S.C. § 112(f) will begin with the words “means for”, but use of the term “for” in any other context is not intended to invoke treatment under 35 U.S.C. § 112(f). Accordingly, the applicant reserves the right to pursue additional claims after filing this application to pursue such additional claim forms, in cither this application or in a continuing application.
This application claims the benefit of and priority to International Application No. PCT/US2023/017927, filed on Apr. 7, 2023, which claims benefit and priority to U.S. Provisional Application No. 63/329,729, filed on Apr. 11, 2022, both of which are incorporated by reference herein in their entirety.
This invention was made with government support under Grant No. FA9550-18-1-0053 and Grant No. FA9550-12-1-0494 awarded by the Air Force Office of Scientific Research. The government has certain rights in the invention.
| Number | Date | Country | |
|---|---|---|---|
| 63329729 | Apr 2022 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | PCT/US2023/017927 | Apr 2023 | WO |
| Child | 18911910 | US |
