A Method for Detecting Facets of SiC Ingot Based on Raman Combined with Resistivity Detection

Abstract

The present invention relates to a method for detecting facets of a SiC ingot based on Raman combined with resistivity detection, which comprises the following steps: S1 provides a silicon carbide workpiece, and based on a preset resistivity detection scanning path, obtains the resistivity ρ of different detection regions on the surface layer of the workpiece, and further preliminarily screens out regions with abnormal resistivity; S2 based on a preset Raman detection scanning path, obtains the Raman characteristic peak frequencies ω of different detection points in the edge region on the surface layer of the workpiece and the regions with abnormal resistivity obtained in S1, detection images, and XY coordinate positions on the XY plane perpendicular to the height direction of the workpiece; S3 based on a preset n-ρ inverse proportional function model of doping concentration n and resistivity ρ and a Δω-n linear model of Raman frequency shift Δω and doping concentration n, determines and correlates the XY coordinate positions in S2 and the corresponding resistivity ρ, and constructs a resistivity distribution map of the surface layer of the workpiece. The present invention has the advantages of small detection error and short detection time. k , and further preliminarily screens out regions with abnormal resistivity; S2 based on a preset Raman detection scanning path, obtains the Raman characteristic peak frequencies ω of different detection points in the edge region on the surface layer of the workpiece and the regions with abnormal resistivity obtained in S1, detection images, and XY coordinate positions on the XY plane perpendicular to the height direction of the workpiece; S3 based on a preset n-ρ inverse proportional function model of doping concentration n and resistivity ρ and a Δω-n linear model of Raman frequency shift Δω and doping concentration n, determines and correlates the XY coordinate positions in S2 and the corresponding resistivity ρ, and constructs a resistivity distribution map of the surface layer of the workpiece. The present invention has the advantages of small detection error and short detection time. ij , detection images, and XY coordinate positions on the XY plane perpendicular to the height direction of the workpiece; S3 based on a preset n-ρ inverse proportional function model of doping concentration n and resistivity ρ and a Δω-n linear model of Raman frequency shift Δω and doping concentration n, determines and correlates the XY coordinate positions in S2 and the corresponding resistivity ρ, and constructs a resistivity distribution map of the surface layer of the workpiece. The present invention has the advantages of small detection error and short detection time. ij , and constructs a resistivity distribution map of the surface layer of the workpiece. The present invention has the advantages of small detection error and short detection time.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor device processing, and in particular to a method for detecting SiC ingot facets based on Raman combined with resistivity detection. Background Art

[0002] Semiconductor materials have the characteristics of high temperature resistance and radiation resistance, and have excellent properties such as a large bandgap width and a high breakdown electric field. Compared with traditional silicon (Si) and gallium arsenide (GaAs) semiconductor materials, silicon carbide (SiC) is more excellent in terms of thermal conductivity (3 to 13 times higher), critical electric field (4 to 20 times higher), and saturated carrier velocity (2 to 2.5 times higher). At the same time, it also has better chemical inertness and radiation resistance. SiC is considered to be a semiconductor material with great potential and is widely used in high-end fields such as aerospace, turbomachinery, and nuclear power. In order to endow the SiC ingot with conductive or semi-insulating properties, impurities such as nitrogen, boron, or aluminum are usually introduced during the crystal growth process. During this doping process, due to the difference in doping concentration (i.e., carrier concentration), facet regions and non-facet regions with different resistivities and refractive indices are likely to be generated. The existence of the facet region not only causes a large loss when the SiC ingot is peeled off the wafer, but also causes a mismatch between the substrate and the epitaxial layer. Therefore, accurately characterizing the doping concentration of the ingot is a key step in current wafer production capacity improvement and subsequent process guidance.

[0003] Chinese Patent with Publication No. CN117316791A discloses a facet detection method, including: irradiating a first surface of an ingot with a laser beam to generate a peeling layer on the ingot; applying ultrasonic waves to the formed ingot based on an ultrasonic vibration component to obtain a wafer corresponding to the peeling layer; emitting a detection beam to the wafer based on a light source with a specific wavelength and recording the light source emission intensity at the wafer receiving point; comparing the transmission intensity of the wafer and the light source emission intensity to obtain the transmittance of the wafer; positioning the facet region in the wafer based on a preset transmittance threshold and the transmittance of the wafer, and determining the position information corresponding to the facet region; using the position information corresponding to the facet region as guiding information to update the laser processing conditions of the next wafer adjacent to the wafer, and peeling the next wafer based on the laser processing conditions. The above method completes the detection of the first peeled wafer by comparing the transmission intensity of the wafer and the light source emission intensity, and then optimizes the laser processing conditions of the next wafer according to the position information corresponding to the facet region of the previous wafer. Laser energy fluctuations or other types of damage on the wafer surface will affect the measurement results, and there is a problem of insufficient detection accuracy.

[0004] Chinese Patent with Publication No. CN111162017A discloses a method for detecting the facet region of a SiC ingot, which is a method for detecting the facet region of a SiC ingot. Among them, this method has the following steps: a fluorescence brightness detection step of irradiating the SiC ingot with excitation light of a specified wavelength from the upper surface of the SiC ingot and detecting the fluorescence brightness inherent to SiC; and a coordinate setting step of setting the region with a fluorescence brightness of a specified value or more in the fluorescence brightness detection step as a non-facet region, setting the region with a fluorescence brightness lower than the specified value as a facet region, and setting the coordinates of the boundary between the facet region and the non-facet region. The above method uses a laser with a specific wavelength to irradiate the surface of the ingot, and distinguishes the facet region and the non-facet region by detecting the intensity of the fluorescence signal on the upper surface of the ingot. However, this characterization method of scanning the entire surface of the ingot with a focused small laser spot has the problem of low efficiency.

[0005] Chinese Patent with Publication No. CN113042915A discloses a method for detecting facets, which is characterized by including: irradiating a first surface of an ingot with a laser beam to cause the ingot to generate a peeling layer; applying ultrasonic waves to the formed ingot based on an ultrasonic vibration assembly to obtain a wafer corresponding to the peeling layer; emitting a detection beam to the wafer based on a light source with a specific wavelength, and recording the light source emission intensity at the receiving point of the wafer; comparing the transmission intensity of the wafer and the light source emission intensity to obtain the transmittance of the wafer; positioning the facet region in the wafer based on a preset transmittance threshold and the transmittance of the wafer, and determining the position information corresponding to the facet region; using the position information corresponding to the facet region as guiding information to update the laser processing conditions of the next wafer adjacent to the wafer, and peeling the next wafer based on the laser processing conditions, where the laser processing conditions at least include the laser focusing position and the laser power. The above method uses a resistivity measuring instrument to measure the resistivity of the surface of the ingot, discriminates the facet region through the resistivity distribution, and is used to guide the parameter setting of the peeling process. However, the facets of the ingot are usually distributed near the edge of the surface, and there are easily organic residues at the edge, interfering with the resistance measurement. In addition, the resistance probe is relatively large, and it is easy to measure inaccurately when measuring the edge. Summary of the Invention

[0006] The problem to be solved by the present invention is to provide a method for detecting the facets of a SiC ingot based on Raman combined with resistivity detection, which has the advantages of small detection error and short detection time in view of the above deficiencies in the prior art.

[0007] The above object of the present invention is achieved by the following technical solutions:

[0008] A method for detecting the facets of a SiC ingot based on Raman combined with resistivity detection includes the following steps

[0009] S1 provides a silicon carbide workpiece, and based on a preset resistivity detection scanning path, obtains the resistivity ρ of different detection regions on the surface layer of the workpiece k , and then preliminarily screens out the resistivity abnormal regions;

[0010] S2, based on a preset Raman detection scanning path, obtains the Raman characteristic peak frequencies ω of different detection points in the edge region on the surface layer of the workpiece and the resistivity abnormal regions obtained in S1 ij , detection images, and the XY coordinate positions (X i , Y j ) on the XY plane perpendicular to the height direction of the workpiece;

[0011] S3, based on a preset n-ρ inverse proportional function model of doping concentration n and resistivity ρ, and a Δω-n linear model of Raman frequency shift Δω and doping concentration n, determines and correlates the XY coordinate positions in S2 and the corresponding resistivity ρ ij , and constructs a resistivity distribution map of the surface layer of the workpiece.

[0012] Among them, in the resistivity ρ k , k is an integer ≥0, and its maximum value corresponds to the number of detection regions on the resistivity detection scanning path; i and j are integers ≥0, and their maximum values correspond to the number of detection points in the X and Y directions of the Raman detection scanning path respectively.

[0013] Further, in S1, the Z-axis coordinate positions (Z k ) of different detection points on the surface layer of the workpiece are pre-obtained, and these Z-axis coordinate positions are correlated with the corresponding two-dimensional resistivity detection scanning path to construct a three-dimensional resistivity detection scanning path.

[0014] Further, in S1, the resistivity ρ k is obtained by resistivity detection using a contact resistivity probe or a non-contact resistivity probe.

[0015] Still further, in S1, the detection range of the resistivity detection probe is 0.1~100 mΩ·cm, and the measurement accuracy ≤0.5 mΩ·cm.

[0016] Further, in S1, the resistivity ρ of different detection regions k is compared with a resistivity threshold. If the resistivity ρ k exceeds the resistivity threshold, the corresponding detection region is determined as a resistivity abnormal region.

[0017] Further, in S2, the Z-axis coordinate positions (Z ij), associate these Z-axis coordinate positions with the corresponding XY coordinate positions to construct the topography distribution map of the workpiece surface layer, and further determine the Raman detection scanning path.

[0018] Further, in the step S2, the Raman characteristic peak frequency ω ij is obtained by performing Raman detection in the longitudinal optical phonon-plasmon coupling (LOPC) mode.

[0019] Still further, in the step S2, the detection laser of the Raman detection laser is a continuous laser, the laser wavelength is 450~1000nm, the laser detection power is adjustable within 0~500mW, and the beam quality factor M 2 <1.2, and the wavelength drift <10pm.

[0020] Further, in the step S3, the fitting function of the n-ρ inverse proportional function model is ρ = k1 / n, where k1 = 1×10 19 ~1×10 20 .

[0021] Still further, in the step S3, the fitting function of the n-ρ inverse proportional function model is ρ = 437.7×10 17 / n (R = 99%).

[0022] Further, in the step S3, the fitting function of the Δω-n linear model is n = k2Δω + b, where k2 = 1.00×10 17 ~1.50×10 17 , and b = 0~1.

[0023] Still further, in the step S3, the fitting function of the Δω-n linear model is n = 1.27×10 17 Δω (R = 99%).

[0024] In summary, the beneficial technical effects of the present invention are as follows:

[0025] 1. The present invention adopts the method of combining Raman detection with resistivity detection, which overcomes the defects of the original Raman detection with high spatial resolution but dense sampling points and long time consumption (4~5h), and overcomes the defects of the original resistivity detection with short time consumption (about 1min) but low spatial resolution and large edge detection error. Also, since the small facet regions are mostly concentrated on the edge of the workpiece, the resistivity probe is used to quickly detect the surface layer of the workpiece and screen the regions with abnormal resistivity, and Raman detection is supplemented to obtain the information of the edge regions and the edge abnormal regions on the surface layer of the workpiece, so as to achieve efficient and accurate identification of the resistivity distribution and control the detection time within 10 minutes;

[0026] 2. By constructing the topographic information of the ingot surface in the Z-axis direction, the present invention can guide the resistivity probe to contact the surface layer of the workpiece, avoid false measurement or damage to the workpiece due to the too deep contact of the resistivity probe with the surface layer, and can know the laser focusing position in Raman detection, so that the position where the Raman signal is generated is always on the surface of the workpiece, thereby improving the signal-to-noise ratio of the Raman characteristic peak and being beneficial to further improving the detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic connection diagram between the silicon carbide workpiece, the resistivity detection scanning path and the Raman detection scanning path in Embodiment 1 of the present invention.

[0028] Figure 2 It is a fitting function diagram of the n-ρ inverse proportional function model in Embodiment 4 of the present invention.

[0029] Figure 3 It is a fitting function diagram of the Δω-n linear model in Embodiment 4 of the present invention.

[0030] Figure 4 It is a schematic connection diagram between the silicon carbide workpiece, the facet area and the detection points in Embodiment 4 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0031] In order to make the technical means, creative features, achieved purposes and functions of the present invention clearer and easier to understand, the present invention will be further described below with reference to the drawings and specific embodiments.

[0032] Embodiment 1: Refer to Figure 1 , a SiC ingot facet detection method based on Raman combined with resistivity detection disclosed in the present invention, includes the following steps.

[0033] S1 Provide a silicon carbide workpiece, and based on a preset resistivity detection scanning path, obtain the resistivity ρ of different detection regions on the surface layer of the workpiece k , and then preliminarily screen out the resistivity abnormal regions.

[0034] S2 Based on a preset Raman detection scanning path, obtain the Raman characteristic peak frequency ω of different detection points in the edge region on the surface layer of the workpiece and the resistivity abnormal regions obtained in S1 ij , detection images, and the XY coordinate positions (X i , Y j ) on the XY plane perpendicular to the height direction of the workpiece.

[0035] S3 Based on a preset n-ρ inverse proportional function model of doping concentration n and resistivity ρ, and a Δω-n linear model of Raman frequency shift Δω and doping concentration n, determine and associate the XY coordinate positions in S2 and the corresponding resistivity ρij , construct a resistivity distribution map of the workpiece surface layer.

[0036] Embodiment 2: A method for detecting SiC ingot facets based on Raman combined with resistivity detection disclosed in the present invention. The difference from Embodiment 1 is that the specific implementation of S1 is as follows: Provide a silicon carbide workpiece. For a contact resistivity probe, pre-acquire the Z-axis coordinate positions (Z k ) of different detection points on the workpiece surface layer, associate these Z-axis coordinate positions with the corresponding two-dimensional resistivity detection scanning paths to construct a three-dimensional resistivity detection scanning path. For a non-contact resistivity probe, directly perform detection according to the two-dimensional resistivity detection scanning path; then place the ingot to be measured on a three-dimensional displacement stage, and use a contact resistivity detection probe to perform resistivity detection to obtain the resistivity ρ k , and the detection range of the resistivity detection probe is 0.1~100 mΩ·cm, and the measurement accuracy ≤ 0.5 mΩ·cm; then compare the resistivity ρ k of different detection regions with the resistivity threshold. If the resistivity ρ k exceeds the resistivity threshold, determine the corresponding detection region as a resistivity abnormal region.

[0037] Embodiment 3: A method for detecting SiC ingot facets based on Raman combined with resistivity detection disclosed in the present invention. The difference from Embodiment 2 is that the specific implementation of S2 is as follows: Pre-acquire the Z-axis coordinate positions (Z ij ) of different detection points on the workpiece surface layer, associate these Z-axis coordinate positions with the corresponding XY coordinate positions to construct a topography distribution map of the workpiece surface layer, and then determine the Raman detection scanning path. The scanning path is in a vortex-like linear shape on the XY plane; then perform Raman detection in the longitudinal optical phonon-plasma coupling (LOPC) mode to obtain the Raman characteristic peak frequencies ω ij , detection images, and the XY coordinate positions (X i , Y j ) on the XY plane perpendicular to the height direction of the workpiece. And the detection laser of the Raman detection laser is a continuous laser, the laser wavelength is 450~1000 nm, the laser detection power is adjustable from 0 to 500 mW, the beam quality factor M 2 < 1.2, and the wavelength drift < 10 pm.

[0038] Embodiment 4: A method for detecting SiC ingot facets based on Raman combined with resistivity detection disclosed in the present invention. The difference from Embodiment 3 is that the specific implementation of S3 is as follows: Refer to Figure 2 and Figure 3, by previously detecting silicon carbide standard samples with different doping concentrations n, the fitting function of the n-ρ inverse proportional function model between the doping concentration n and the resistivity ρ is determined to be ρ = 437.7×10 17 / n (R = 99%), and the fitting function of the Δω-n linear model between the Raman shift Δω and the doping concentration n is determined to be n = 1.27×10 17 Δω (R = 99%). Based on the detection values of different detection points in Example 3, the resistivity ρ is calculated ij , and the resistivity ρ is correlated ij , and the corresponding XY coordinate positions (X i , Y j ) are correlated into two-dimensional data (X i , Y j , ρ ij ) to construct a resistivity distribution map of the workpiece surface layer.

[0039] Referring to Figure 4 , taking multiple detection points (from left to right) on a straight line passing through a small surface area as an example, the two-dimensional data corresponding to its resistivity distribution map is shown in Table 1 below.

[0040] Table 1

[0041]

[0042] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the purpose and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A method for detecting SiC ingot facets based on Raman combined with resistivity detection, characterized in that: The following steps are included: S1 provides a silicon carbide workpiece, and obtains the resistivity ρ of different detection areas on the surface layer of the workpiece based on a preset resistivity detection scanning path k , and then preliminarily screen out the resistivity abnormal area; S2 obtains the Raman characteristic peak frequency ω of the edge area on the surface layer of the workpiece and the resistivity abnormal area obtained in S1 at different detection points based on the preset Raman detection scanning path. ij , the detection image, and the XY coordinate position (X i , Y j ); S3 determines and associates the XY coordinate position of S2 and the corresponding resistivity ρ based on the preset n-ρ inverse proportional function model of doping concentration n and resistivity ρ and the Δω-n linear model of Raman frequency shift Δω and doping concentration n. ij , construct a resistivity distribution map of the surface layer of the workpiece.

2. A SiC ingot facet detection method based on Raman combined with resistivity detection according to claim 1, characterized in that: In S1, the Z-axis coordinate positions (Z k ), these Z-axis coordinate positions and the corresponding two-dimensional resistivity detection scanning paths are associated to construct a three-dimensional resistivity detection scanning path.

3. The method for detecting SiC ingot facets based on Raman combined with resistivity detection according to claim 1, characterized in that: In S1, the resistivity ρ k The resistivity is obtained by resistivity detection using a contact resistivity probe or a non-contact resistivity probe.

4. A method for detecting SiC ingot facets based on Raman combined with resistivity detection according to claim 3, characterized in that: In S1, the detection range of the resistivity detection probe is 0.1-100 mΩ·cm, and the measurement accuracy is ≤0.5 mΩ·cm.

5. The method for detecting SiC ingot facets based on Raman combined with resistivity detection according to claim 1, characterized in that: In S1, the resistivity ρ of different detection areas is k Compared with the resistivity threshold, if the resistivity ρ k If the resistivity threshold is exceeded, the corresponding detection area is determined to be a resistivity abnormal area.

6. The method for detecting SiC ingot facets based on Raman combined with resistivity detection according to claim 1, characterized in that: In S2, the Z-axis coordinate positions (Z ij ), these Z-axis coordinate positions and the corresponding XY coordinate positions are associated to construct a morphology distribution map of the surface layer of the workpiece, and then determine the Raman detection scanning path.

7. The method for detecting SiC ingot facets based on Raman combined with resistivity detection according to claim 1, characterized in that: In S2, the Raman characteristic peak frequency ω ij Obtained by Raman detection in longitudinal optical phonon-plasmon coupling mode.

8. The method for detecting SiC ingot facets based on Raman combined with resistivity detection according to claim 7, characterized in that: In S2, the detection laser of the Raman detection laser is a continuous laser with a laser wavelength of 450-1000 nm, a laser detection power adjustable from 0 to 500 mW, and a beam quality factor M 2 <1.2, wavelength drift <10pm.

9. The method for detecting SiC ingot facets based on Raman combined with resistivity detection according to claim 1, characterized in that: In S3, the fitting function of the n-ρ inverse proportional function model is ρ=k1 / n, k1=1×10 19 ~1×10 20 .

10. The method for detecting SiC ingot facets based on Raman combined with resistivity detection according to claim 1, characterized in that: In S3, the fitting function of the Δω-n linear model is n=k2Δω+b, k2=1.00×10 17 ~1.50×10 17 , b=0~1.

Images