Method for estimating twin defect density

Abstract

Disclosed is a method for estimating twin defect density in a single-crystal sample, including: (A) etching the observed surface of a single crystal to form etch pits; (B) selecting bar-shaped etch pits caused by twin defect; (C) from the long-axis direction lengths of the etch pits caused by twin defect, estimating the twin defect density by using the following equation: twin defect density=Σkx′.sub.i/area of sample, wherein 2≤k≤3, and x′.sub.i is the long-axis direction length of an etch pit caused by the i-th twin.

Claims

1. A method of evaluating a twin defect density of a single crystal sample, comprising steps of: (A) forming etch pits by etching an observation surface of a single crystal; (B) selecting an etch pit according to a twin defect; and (C) evaluating a twin defect density using an equation below from a long-axis direction length of the etch pit according to the twin defect, wherein the twin defect density=Σkx′.sub.i/the area of the sample, and wherein 2≤k≤≤3, x′.sub.i is the long-axis direction length of the etch pit based on an i-th twin defect.

2. The method of claim 1, wherein: the etch pit according to the twin defect in the step (B) has a bar shape, in the etch pit, a width that is a short-axis direction length is increased according to a lapse of an etching time, but a long-axis direction length is not changed.

3. The method of claim 1, wherein: k=2 if the single crystal sample is a substrate, and k=3 if the single crystal sample is a thin film.

4. The method of claim 1, wherein the single crystal sample is a gallium oxide single crystal or a gallium oxide single crystal thin film.

5. The method of claim 4, wherein the gallium oxide single crystal thin film is an epi layer formed on a gallium oxide (−201) single crystal substrate.

6. The method of claim 5, wherein the long-axis direction of the etch pit according to the twin defect is a [010] direction.

7. The method of claim 2, wherein: k=2 if the single crystal sample is a substrate, and k=3 if the single crystal sample is a thin film.

8. The method of claim 2, wherein the single crystal sample is a gallium oxide single crystal or a gallium oxide single crystal thin film.

9. The method of claim 8, wherein the gallium oxide single crystal thin film is an epi layer formed on a gallium oxide (−201) single crystal substrate.

10. The method of claim 9, wherein the long-axis direction of the etch pit according to the twin defect is a [010] direction.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 show images illustrating RHEED patterns of a gallium oxide thin film according to an embodiment.

(2) FIG. 2 is XRC of the gallium oxide thin film according to an embodiment.

(3) FIG. 3 is a Cross-sectional TEM image of the gallium oxide thin film according to an embodiment.

(4) FIG. 4 shows AFM images of the gallium oxide thin film according to an embodiment.

(5) FIG. 5 shows optical microscope images of the gallium oxide thin film according to an etching time.

(6) FIG. 6 shows SEM and TEM images of a gallium oxide thin film fragment.

(7) FIG. 7 shows TEM images for investigating the origin of a bullet-shaped etch pit.

(8) FIG. 8 shows SEM and TEM images for investigating the origin of a bar-shaped etch pit.

(9) FIG. 9 shows a plane-view TEM image and a schematic diagram illustrating a twin defect.

(10) FIG. 10 shows TEM images for investigating the origin of a bar-shaped etch pit.

(11) FIG. 11 is a schematic diagram illustrating a twin defect-caused etch pit.

DETAILED DESCRIPTION

(12) Hereinafter, the present invention is described in detail with reference to attached embodiments. However, such embodiments are only illustrative in order to easily describe the contents and range of the technical spirit of the present invention, and the technical range of the present invention is not limited by the embodiments. Furthermore, various modifications and changes based on the examples within the range of the technical spirit of the present invention will be evident to those skilled in the art.

EMBODIMENT

Embodiment 1: Fabrication of Gallium Oxide Thin Film

(13) Commercial gallium oxide (−201) substrates was washed by ultrasonic waves stirring for 10 minutes sequentially using acetone, methanol and deionized water (DI water). The washed substrate was chemically washed using a solution with volume ratio of 1:1:4 mixture of deionized water, 30% oxygenated water and 96% sulfuric acid for 5 minutes. The substrate was additionally washed using deionized water, and dried using nitrogen gas. The dried substrate was introduced into a plasma-assisted molecular beam epitaxy growth chamber, and was prepared by thermal cleaning the substrate at 850° C. for 30 minutes using oxygen radicals generated by excited plasma using RF power of 300 W and an oxygen gas flow rate of 2 sccm.

(14) Element gallium (Ga) having a purity of 7N was supplied through a Sumo effusion cell made by Veeco Co. of USA into a plasma-assisted molecular beam epitaxy growth chamber on which the substrate was mounted. A gallium oxide thin film was grown using a plasma-assisted molecular beam epitaxy method by generating and supplying oxygen radicals by plasma by supplying oxygen to a plasma cell made by SVAT Co. of USA. A thin film growth condition is as follows; substrate temperatures of 650-850° C., an oxygen gas flow rates of 1-3 sccm, Ga flux of 0.1 Å/s, a thin film growth time for 180 minutes, growth chamber pressure of about 2×10.sup.−9 Torr before the gas was introduced, and pressures of 2×10.sup.−5 Torr-4×10.sup.−5 Torr during the growth.

Embodiment 2: Evaluation of Characteristics of Gallium Oxide Epi Thin Film

(15) Prior to the evaluation of a crystal defect, it was first checked that the gallium oxide epi thin film was grown by the molecular beam epitaxy method of Embodiment 1.

(16) 1) Crystal Structure Analysis Through XRD, RHEED and Cross-Sectional TEM Analysis

(17) The crystal structure of the thin film grown by Embodiment 1 was analyzed using reflection high energy electron diffraction (RHEED) and X-ray scattering analysis (XRD, X-ray diffraction).

(18) The RHEED was analyzed under a condition with an electronic voltage of 18 kV and a beam current of 39 μA. The XRD was analyzed under conditions with Kα X-ray of Cu wherein λ=1.5406 nm, a voltage of 40 kV, a current of 40 mA, a scan step of 0.02, and a step time of 0.5 s, by using D8 Discover model made by Bruker AXS Co. of Germany. FIG. 1 illustrates RHEED images of the thin film with growth times. FIG. 2 illustrates X-ray rocking curves (XRCs) for a (−402) along the [010] and the [102] directions in the thin film grown for 180 minutes. The results of FIGS. 1 and 2 show that the thin film grown by Embodiment 1 has a monoclinc crystal structure and excellent crystallizability.

(19) FIG. 3 illustrates an X-TEM image of the thin film, and shows that an interface was not observed between the substrate and the epi layer and likewise has excellent crystallinity.

(20) 2) Measurement of Surface Roughness of Thin Film

(21) Surface roughness was measured based on a root mean square roughness by using an atomic force microscope (AFM, MFD-3D model of Asylum Research Col.) value under conditions with scan areas for 10 μm×10 μm, 5 μm×5 μm, and 2 μm×2 μm, a scan speed of 1 Hz, and a driving frequency of 70 kHz from the thin film. FIG. 4 illustrates AFM images showing surface roughness of the gallium oxide thin film grown for 180 minutes, and sequentially illustrates the AFM images for scan areas of 10 μm×10 μm, 5 μm×5 μm, 2 μm×2 μm from the top. It can be seen that flat root mean square (RMS) roughness of about 1 nm level from the scan area of 2 μm×2 μm.

Embodiment 3: Etch Pit Analysis

(22) The gallium oxide epi layer fabricated using the method of Embodiment 1 was etched using a phosphoric acid at 130° C., and an etched surface was observed using an optical microscope. The observation using the optical microscope was performed at 1000× magnification using Eclipse LV100ND model of Nikon Co. FIG. 5 illustrates optical microscope images with different etching time, and illustrates that the bar-shaped etch pits were mostly observed when etched for 5 minutes, but the bullet-shaped etch pits were formed along with the bar-shaped etch pits as the etching time was increased. Accordingly, the origin of each etch pit was analyzed.

(23) 1) Bullet-Shaped Etch Pit

(24) In order to confirm the origin of the bullet-shaped etch pit generated by etching and placed along the [102] direction of a single crystal gallium oxide epi layer surface, a TEM specimen was fabricated using a focused ion beam (FIB) method, and was observed using a TEM. FIG. 6(a) illustrates a SEM image during a process of fabricating a cross-sectional TEM specimen using the FIB method, and FIG. 6(b) is a low magnification cross-sectional bright field TEM image showing selected etch pits of the cross-sectional TEM specimen finally completed in FIG. 6(a) process. FIG. 7(a) is a cross-sectional bright field TEM image for the fifth and sixth etch pit portions in FIG. 6(b). It can be seen that the origin of the bullet-shaped etch pit is a threading dislocation (TD). FIGS. 7(b) to 7(d) illustrate two-beam analysis results for threading dislocation analysis. FIG. 7(b) is a bright field image under the zone axis, 7(c) is with a diffraction vector of g=20−2, and 7(d) shows the image observed under the diffraction vector of g=0−20. According to a visible/invisible condition of the dislocation, it can be seen that the Burgers vector of threading dislocation is [010]. Accordingly, the type of threading dislocation is analyzed as an edge dislocation.

(25) 2) Bar-Shaped Etch Pit

(26) In FIG. 5, the number of etch pits was not changed with increasing the etching time for the bar-shaped etch pit lengthily formed along the [010] direction. A thickness of the bar-shaped etch pit in a [102] direction was increased, but a length in the direction was not changed.

(27) FIG. 8(a) is an SEM image during the process of fabricating the cross-sectional TEM specimen using the FIB method. FIG. 8(b) is a low magnification cross-sectional bright field TEM image showing three etch pits of the cross-sectional TEM specimen finally completed in FIG. 8(a). FIG. 8(c) is a cross-sectional bright field TEM image showing the first etch pit of FIG. 8(b). FIG. 8(d) is a high magnification bright field TEM image of the etch pit of FIG. 8(c), and shows that the etch pit was formed at the location where the boundary of a twin defect met a surface of the epi layer. It can be seen that a pair of etch pits are observed as one etch pit in the optical microscope of FIG. 5 or in the SEM image of FIG. 8(a). A distance between the pair of etch pits was about 190 nm. FIG. 8(e) is a high resolution TEM photograph of a twin crystal portion, and shows that a matrix and a defect portion have a twin crystal relation (mirror plane symmetrical). It can be seen that the corresponding defect is a twin defect.

(28) FIG. 9(a) is a low magnification TEM photograph from (−201) plane-view specimen prepared by using the FIB method. In this case, a twin defect (rectangular portion) lengthily present in the [010] direction is shown. A shape of the twin defect present within the sample was schematically shown in FIG. 9(b) based on the cross-sectional TEM observation results of FIG. 9(d) for the twin defect and the plane-view observation results of FIG. 9(a).

(29) FIG. 10 illustrates additional results showing that the corresponding defect is a twin defect. FIG. 10(a) is a cross-sectional bright field TEM image of a defect region generated in the gallium oxide epi layer. FIG. 10(b) shows a diffraction pattern in a normal crystal matrix. FIG. 10(c) is a diffraction pattern of the defect region. Comparing the diffraction patterns of FIGS. 10(b) and 10(c), the matrix and the defect region have a twin crystal relation, and the defect corresponds to a twin defect.

(30) 3) Calculation of Twin Defect Density

(31) In conventional technology, a method of calculating a dislocation defect density from an etch pit has been known, but to calculate a twin defect density has not been known. In particular, in the case of a gallium oxide (−201) substrate, a large-area substrate can be easily supplied, and a growth rate of a gallium oxide epi layer is fast, but has a drawback in that the generation of a twin defect in the epi layer is high. If a twin defect can be quantitatively evaluated, it is more effective to develop a method of growing an epi layer with superior crystal quality by decreasing the twin defect. Accordingly, a method of quantitatively calculating a twin defect density from an observed etch pit was developed.

(32) FIG. 11 is a schematic diagram showing twin defect-caused etch pits. In the twin defect, a boundary between the twinned crystal and the matrix acts as a defect, and thus a defect density may be calculated from the area of a boundary of a twinned crystal. One boundary area TWB-S1 of the twin defect in a sample having a total volume of V (V=xyz) is defined as follows.

(33) x′ is approximately an order of 10 μm unit, and y′ is several hundreds of nm unit,

(34) x′>>y′,

(35) Accordingly, TWB-S1=2x′z+2y′z+x′y′=2(x′+y′)z+x′y′˜2x′z+x′y′=x′(2z+y′).

(36) Since 0<y′≤z,

(37) TWB-S1=kx′z (2≤k≤3)

(38) For example,

(39) 1) If the sample is a substrate,

(40) z>>y′ because z is a thickness of the substrate and has a very large value.

(41) TWB-S1=x′(2z+y′)˜2x′z, that is, k converges on 2.

(42) 2) If the sample is a thin film,

(43) z˜y′ because z is a level similar to y′, that is, a thickness of the thin film.

(44) TWB-S1=x′(2z+y′)˜3x′z, that is, k converges on 3.

(45) The sum of a twin defect density TWBD in the sample=a sum of twin crystal boundary area in the sample/the volume of the sample=Σkx′.sub.iz/xyz=Σkx′.sub.i/xy (2≤k≤3)

(46) That is, the twin defect density in the sample may be simply calculated from the sum of lengths of bar-shaped etch pits at twin defects and a surface area of the etched sample. If the sample has a constant thickness and only a specific condition for forming the thin film is changed, k may be considered as being a constant. Twin defect densities can be quantitatively compared by only calculating the sum of lengths of etch pits per unit area.