This invention concerns a multilayer structure. In particular, it relates to a multilayer structure comprising at least two monocrystalline layers A and B, wherein layer A comprises K-Ga2Os and layer B comprises 0-Ga2O3 and wherein layers A and B are adjacent. The invention also relates to a method of producing K- Ga2Os, said method comprising the step of irradiating 0-Ga2O3 with an ion beam. The invention further relates to a semiconductor device comprising the multilayer structure.

Background

Polymorphs are common in nature and the basic principles of polymorphism in crystals are known: the lattices adapt to the minimum energy in respect of temperature and pressure. Controllable stabilization of the metastable phases, e.g. at room temperature, is of great interest because it offers the possibility to realise new properties. Certain polymorphs can be stabilized by applying external pressure in materials, however the application of colossal pressures in these scenarios is undesirable. For that reason, reaching metastable conditions via ion beam assisted processes with displaced atoms provoking strain accumulation is an interesting alternative. However, the ion-induced disorder may also result in amorphization, as typically occurs in semiconductors such as Si or SiC. Even for so-called “radiation hard” semiconductors, e.g. GaN or ZnO, the ion irradiation still ends up with high disorder level or partial amorphization.

As a promising ultra-wide bandgap oxide semiconductor material, gallium oxide has potential applications in the wide range of enabling technologies, from power electronics to optoelectronic devices. Ga2Ch is well known for its polymorphism, however, at present, Ga2Ch fabrication technology is immature. In contrast to the stable -phase, the metastable Ga20s polymorphs are not available in the form of bulk crystals of technologically significant size. In order to prepare such polymorphs, film deposition on other Ga2C>3 substrates is typically employed, utilising strain to facilitate these polymorph formations. For example, CN103924298 describes a method of spontaneous formation of a K-Ga2O3/ -Ga2O3 heterogeneous structure. The obtained K-Ga2C>3 is a new crystal structure in a gallium oxide system and has orthorhombic symmetry. However, the prepared K-Ga2O3/0-Ga2O3 heterogeneous structure comprises merely random inclusions of K-Ga2O3. In the majority of cases, film deposition methods result in inclusions of the second phase, or lead to multiphase growth.

In order to maximise the potential applications of gallium oxide, fabricating regular multi -polymorph structures is of particular interest, paving the way for the realization of polymorph heterostructures. Of specific interest is the attainment of consistent polymorph interfaces, which may be very difficult to fabricate by more- close-to-equilibrium techniques (as all other conventional deposition techniques in comparison to ion implantation) because the growing material will try to inherit the crystalline structure of the substrate.

Summary of Invention

The present inventors have surprisingly found that ion implantation in gallium oxide can be exploited to generate sufficient strain, triggering the polymorph transitions exactly in the irradiated regions. Unexpectedly, this allows for controllable ion-beam induced phase engineering, resulting in the formation of highly oriented K-Ga2O3 on top of 0-Ga2O3, not previously achievable by conventional thin film deposition methods. Spectacularly, such fabricated structures, which comprise a layer comprising K-Ga2O3 adjacent to a layer comprising 0-Ga2O3 exhibit unprecedentedly high radiation tolerance.

Thus, viewed from a first aspect the invention provides a multilayer structure comprising at least two monocrystalline layers A and B, wherein layer A comprises K-Ga2O3 and layer B comprises 0-Ga2O3 and wherein layers A and B are adjacent.

Viewed from another aspect, the invention provides a method for producing K-Ga2O3, said method comprising the step of irradiating 0-Ga2O3 with an ion beam.

Viewed from a further aspect, the invention provides a semiconductor device comprising a multilayer structure as hereinbefore defined. Definitions

The term “monocrystalline” as used herein means a single crystal phase. Thus, each of the layers in the structures of the present invention may consist of a single crystal phase of gallium oxide, namely layer A comprises gallium oxide in the form K-Ga2Os and layer B comprises gallium oxide in the form 0-Ga2O3. This term therefore excludes structures which comprise more than one crystal phase of gallium oxide in a layer.

It will be appreciated that, in the context of the present invention, the terms “film” and “layer” are equivalent and have the same meaning.

Detailed Description of Invention

The present invention relates to a multilayer structure comprising at least two monocrystalline layers A and B, wherein layer A comprises K-Ga2O3 and layer B comprises 0-Ga2O3 and wherein layers A and B are adjacent. The invention also relates to a method for producing K-Ga2O3, said method comprising the step of irradiating 0-Ga2O3 with an ion beam.

Multilayer Structure

The multilayer structure of the invention comprises at least two layers A and B as hereinbefore defined. In one embodiment, the structure comprises or consists of two layers, i.e. layer A and layer B. In another embodiment, the structure comprises more than two layers, such as three, four or five layers. In such structures, it is preferable if each of the additional layers is a layer A or B as hereinbefore defined. In one particularly preferred embodiment, the multilayer structure comprises three layers in the order BAB, wherein layers A and B are as hereinbefore defined.

Each layer A and B is monocrystalline. Ideally, each layer is homogenous, i.e. comprising no other phase inclusions. Layer A comprises, preferably consists of, K-Ga20s. K-Ga20s is a polymorph of gallium oxide with an orthorhombic unit cell.

Layer B comprises, preferably consists of, 0-Ga2O3. 0-Ga2O3 is a stable form of gallium oxide and is a polymorph with monoclinic structure.

In all embodiments, it is possible for the gallium oxide to be doped with other metal atoms.

The interface between layers A and B may be considered to be continuous, linear and sharp.

Each layer in the multilayer structure of the invention will typically have a thickness of the order of nanometers or micrometers. The skilled person will appreciate that the thickness of the layers may be tailored towards a particular structure or application. The thickness of the layers may be controlled, for example, by altering the ion energy and/or by the angle of incidence of the ion beam. Thus, each layer may have a thickness in the range of 10 nm to 10 pm, such as 100 nm to 300 nm.

The multilayer structure may form the bulk of the structure or, alternatively, may form part of a more complex structure.

Method

The methods of the invention comprise irradiating 0-Ga2O3 with an ion beam. This irradiation produces K-Ga2O3 in a controllable manner.

The ion beam may be any suitable beam of ions which are capable of inducing a sufficiently high level of internal strain within the 0-Ga2O3, so as to induce the phase transition. Typically, therefore, the ion beam is a medium to heavy ion beam. By “medium to heavy ion beam” we generally mean a beam composed of particles which are heavier than helium. Specific examples of ions which may be employed include nickel, gallium or gold ions.

Typically, the ion beam has a dosage in the range 1 x 10 ¹³ to 1 x 10 ¹⁷ ions cm' ² depending on the ion type used. More preferably dosage ranges include 1 x 10 ¹⁵ to 1 x 10 ¹⁶ ions cm' ² . The K-Ga20s produced by the methods of the invention may form part of any larger overall structure. However, in a preferable embodiment, the K-Ga2Os forms a layer in a multilayer structure as hereinbefore defined. Hence, the methods of the invention may also be considered as methods of preparing the multilayer structures of the invention.

The K-Ga2Os produced by the methods of the invention is ideally monocry stalline.

The P-Ga2C>3 which is irradiated with an ion beam may take any suitable form, such as a single crystal wafer.

The method of the invention preferably does not involve applying external pressure. Typically, the methods may be performed at room temperature (e.g. 18 to 35 °C).

The angle of incidence of the ion beam is an additional option to tune the layer thickness, it can be in the range 0 to 89°.

The multilayer structure of the present invention is particularly applicable for use in a semiconductor device. Thus, the invention also relates to a semiconductor device comprising a multilayer structure as hereinbefore defined. The semiconductor device may find application in a wide range of end uses, such as electronic applications, in particular in high voltage systems in the automotive industry.

The invention will now be described with reference to the following nonlimiting examples and figures.

Figure 1. (a) ABF-STEM image of P-Ga2C>3 sample irradiated with l * 10 ¹⁶ Ni/cm ² Panels (b) and (c) illustrate SAED patterns from the unimplanted and implanted regions respectively; (d-f) FFTs from high-resolution images taken from different regions of the sample as indicated in the panel (a).

Figure 2. (a) ABF-STEM and (b) HAADF-STEM high-resolution images acquired simultaneously from the maximum ion-concentration region of (010) P- Ga2O3 irradiated with I / I O ¹⁶ Ni/cm ² . FFTs extracted from two adjacent grains, indicated as 1 and 2 in panel (a), are shown in the corresponding bottom panels. Figure 3. EELS spectra of the oxygen-K edge, acquired from different areas of the high-dose and low-dose implanted 0-Ga2O3, illustrating characteristic intensity changes correlated with 0-to-K transition.

Figure 4. (a) RBS/C spectra and (b) XRD 2theta scans of the (010) oriented -Ga2O3 crystals implanted with 400 keV ⁵⁸ Ni ⁺ (1 * 10 ¹⁵ cm’ ² ), 500 keV ⁶⁹ Ga ⁺ (l *10 ¹⁵ cm’ ² ) and 1.2 MeV ¹⁹⁷ Au ⁺ ions (3*10 ¹⁴ cm’ ² ).

Examples

(010) and (-201) oriented 0-Ga2O3 single crystal wafers (Tamura Corp.) were implanted with different ions, specifically ⁵⁸ Ni ⁺ , ⁶⁹ Ga ⁺ , and ¹⁹⁷ Au ⁺ . The ballistic defect production rates (without accounting for non-linear cascade density effects) for ⁶⁹ Ga ⁺ and ¹⁹⁷ Au ⁺ implants were normalized to that of ⁵⁸ Ni ⁺ ion implanted with 400 keV in a wide dose range of 6*10 ¹³ to l *10 ¹⁶ cm’ ² . The implantations were performed at room temperature (if not indicated otherwise) and 7° off the normal direction.

The samples were analyzed by a combination of Rutherford backscattering spectrometry in channeling mode (RBS/C), x-ray diffraction (XRD) and scanning transmission electron microscopy (STEM) combined with electron energy loss spectroscopy (EELS). The RBS/C was performed using 1.6 MeV He ⁺ ions incident along [010] direction and 165° backscattering geometry. The XRD 2theta measurements were performed using Bruker AXS D8 Discover diffractometer using Cu n radiation in locked-coupled mode.

The STEM and EELS investigations were conducted on an FEI Titan G2 60- 300 kV at 300 kV with a probe convergence angle of 24 mrad. The simultaneous STEM imaging was conducted with 3 detectors: high-angle annular dark field (HAADF) (collection angles 101.7-200 mrad), annular dark field (ADF) (collection angles 22.4-101.7 mrad) and annular bright field (ABF) (collection angles 8.5-22.4 mrad). The resulting spatial resolution achieved was approximately 0.08 nm. EELS was performed using a Gatan Quantum 965 imaging filter. The energy dispersion was 0.1 eV/channel and the energy resolution measured using the full width at half maximum (FWHM) of the zero-loss peak was 1.1 eV. Electron transparent TEM samples with a cross-sectional wedge geometry were prepared by mechanical polishing with the final thinning performed by Ar ion milling and plasma cleaning.

Formation of the new phase in 0-Ga2C>3 due to high dose implantation is supported by STEM investigations and Fig. l(a-f) summarizes the STEM data for the sample implanted with l *10 ¹⁶ Ni/cm ² . Specifically, Fig.1(a) shows an ABF- STEM image and strain contrast reveals the formation of two distinct regions - the film and the substrate - of the initially homogeneous 0-Ga2O3 wafer. Selected area electron diffraction (SAED) patterns taken from the unimplanted and implanted regions, i.e. Figs.1(b) and 1(c), illustrate a prominent transformation from monoclinic - to ordered orthorhombic K-phase. This phase transformation extends to -300 nm from the surface and stops abruptly, forming a sharp interface with the 0-phase wafer/substrate, see Fig.1(a). The contrast associated with defects/strain inside the K-Ga2O3 film gradually increases towards the K/0 interface, see Fig.1(a). However, fast Fourier transforms (FFTs) from high-resolution images taken at the interfacial area (e) and the upper part (f) of the implanted region show that the ordered orthorhombic phase is retained through the depth of the film as compared with the FFT at the 0-Ga2O3 substrate, see Fig.1(f). Thus, SAED and FFTs show the formation of a single-phase ordered orthorhombic K-phase region both at meso and nano-scale. The sharp spots, in addition to the lack of extra reflections and/or striking of the main reflections, indicate a highly-oriented crystalline film with no signs of high-angle mis-orientations, high-density of mis-oriented grains or amorphization as also supported by Fig. 2 showing an analysis of the adjacent K- phase grains exhibiting different strain contrast. Indeed, even though the film in Fig. 1(a) was clearly interpreted as single K-phase film, there are still remaining questions, in particular related to potential mis-orientation between the grains as well as regarding potential chemical variations. For that reason, we investigated two adjacent grains as seen in Fig. 2. FFTs extracted from two adjacent grains reveal that both grains are stabilized in K-phase and have the same orientation. The stable contrast in HAADF (pure Z-contrast image) indicates no chemical variations and the contrast in ABF is attributed only to the strain. Thus, the grains are nicely cooriented and there are no chemical inhomogeneities observed. Moreover, the comparison between EELS spectra in Fig.3 provides additional arguments. Indeed, because of different atomic coordination in 0- and K- phases we detected characteristic changes in the fine structure of the EELS spectra acquired in STEM-mode, by comparing low and high-dose implanted samples. In particular, the oxygen K-edge is characterized by two main peaks, labelled A and B in Fig.3, related to the O 2p - Ga 4s and O 2p - Ga 4p bonding, respectively. As seen from Fig.3, prominent changes in the A/B intensity ratio (IA/IB) occur when the phase transition takes place. Specifically, IA/IB decreases in the K-phase. This can be attributed either to the increase in O 2p - Ga 4p hybridization or to the transfer of electrons from O 2p - Ga 4p band into another band.

The next figure demonstrates that the phenomenon of the 0-to-K phase transition is generically related to the accumulation of the lattice disorder and not to the chemical nature of the implanted ions. This was proved by performing control implants with Ga and Au ions. Note that, the ballistic defect production rates for Ga, and Au implants were normalized to that of Ni ion implanted with 400 keV in a wide dose range of 6*10 ¹³ - 1 *10 ¹⁶ cm’ ² . Fig. 4 provides a representative example of the corresponding RBS/C and XRD data. As seen from Fig. 4, both the RBS/C profiles and XRD 2theta scans exhibit very similar trends for the all ion species used. Specifically, Fig. 4(a) shows the formation of the box-shape disorder layer reaching -90% of the random signal for the all ions used. In its turn, Fig. 4(b) demonstrates that the virgin 0-Ga2O3 wafer is characterized by a strong reflection around 60.9° attributed to the (020) planes of 0-Ga2O3, while the diffraction peaks in the implanted samples centered at -63.4° was interpreted as signatures of the K- Ga2O3. Thus, the observed polymorph transitions are attributed to the disorder- induced effects with negligible impact of the chemical nature of the ions.