A silicon carbide wafer and a method for detecting dislocations in a silicon carbide wafer with high accuracy

Abstract

The present invention relates to the field of semiconductor technology, and discloses a silicon carbide wafer and a method for detecting dislocations of a silicon carbide wafer with high accuracy. After the silicon carbide wafer is etched with potassium hydroxide, residues invisible to the naked eye will adhere to its surface, and these residues cannot be removed by water washing or alcohol washing. Even solvents such as DMF and DMSO that can be miscible with most organic substances cannot effectively remove the above-mentioned residues. In the prior art, when using potassium hydroxide etching combined with image recognition method to detect dislocations, it is easy to cause incomplete cleaning, resulting in residues still remaining on the wafer surface, and further resulting in misjudging the number of dislocation points and dislocation density as the true values. The present invention makes the dislocation detection result of the silicon carbide wafer close to its true dislocation result and has a high accuracy by cleaning the surface of the silicon carbide wafer with DMAC after the silicon carbide wafer to be tested is etched with molten potassium hydroxide and before the dislocation detection.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and particularly to a silicon carbide wafer and a method for detecting dislocations in a silicon carbide wafer with high accuracy. Background Art

[0002] Silicon carbide (SiC) is a typical representative of the third-generation semiconductor materials. Due to its excellent comprehensive performance, SiC-based devices can operate in extreme environments such as high temperature, high pressure, and strong radiation, and have broad application prospects in the fields of power electronics and microwave communication. Therefore, SiC materials have important value.

[0003] The quality of SiC crystals has a great impact on the performance of power electronic devices. Therefore, it is necessary to control and characterize the quality of SiC crystal products. The production process of SiC crystal products generally includes: crystal growth, forming, cutting, thinning, polishing, packaging, and leaving the factory. Before leaving the factory, samples need to be taken for dislocation detection to characterize and evaluate the quality of this batch of SiC crystal products.

[0004] For the dislocation detection of SiC crystals, it is mainly carried out by combining potassium hydroxide (KOH) etching with image recognition method. The specific principle is as follows: Since the lattice around dislocations in a silicon carbide single crystal is distorted, when the surface of a silicon carbide single crystal is etched with a potassium hydroxide molten solution, the etching rate is faster at the dislocation emergence points on the surface of the silicon carbide single crystal, and thus it is easy to form etching pits with specific shapes composed of certain low-index planes. Then, the shape and size of the etching pits are observed by an optical microscope, and the etching pits are classified and the dislocation density is calculated through image recognition technology. In the prior art, for example, the patent with the application publication number CN115896954 A discloses a method for detecting basal plane dislocations in a silicon carbide single crystal. It etches a silicon carbide single crystal wafer in a molten potassium hydroxide solution and then observes the shape of the etching pits through a microscope to identify and determine the type of dislocations. However, the dislocation detection method in patent CN115896954 A has the following problems: The etching residues on the sample surface after etching are not thoroughly cleaned, and the etching residues will affect the recognition of dislocations. Only by the size and shape of the etching pits in the image to determine dislocations, the current calculation method is prone to misjudgment problems. Summary of the Invention

[0005] In order to reduce the misjudgment problem of microscopic dislocation recognition and improve the accuracy of dislocation detection. More specifically, to solve the problem of low accuracy of dislocation detection caused by potassium hydroxide molten solution etching residues, the present invention provides a silicon carbide wafer and a method for detecting dislocations in a silicon carbide wafer with high accuracy.

[0006] The specific technical solution of the present invention is as follows:

[0007] On the one hand, the present invention provides a method for detecting dislocations in a silicon carbide wafer with high accuracy. Compared with the prior art, its outstanding feature is that after etching with molten potassium hydroxide and before the dislocation detection, the silicon carbide wafer is cleaned with N,N-dimethylacetamide.

[0008] Generally, the dislocation detection of a silicon carbide wafer is achieved by etching the silicon surface of the silicon carbide wafer with a molten potassium hydroxide solution to present etch pits at the dislocation sites, and then identifying the number and density of dislocations by the image method through the etch pits. However, according to the above detection method, it is easy to leave etch residues attached to the wafer surface when etching the silicon carbide wafer with a molten potassium hydroxide solution. Although there is also prior art suggesting that the etch residues on the sample surface after the etching with the molten potassium hydroxide solution will affect the dislocation identification, most of the prior art only targets the residues visible to the naked eye or directly conducts routine water washing and alcohol washing without pertinence.

[0009] When the present inventors detected dislocations through the above detection method, they found that in addition to the residues visible to the naked eye attached to the wafer surface after etching the silicon carbide wafer surface with a molten potassium hydroxide solution, there were also some residues invisible to the naked eye attached. These residues could not be removed by cleaning with water, organic alcohol solvents such as absolute ethanol, etc. These residues were similar in size and shape to threading screw dislocations (TSDs), threading edge dislocations (TEDs), or basal plane dislocations (BPDs) under a microscope, so they were easily misidentified as one of the above three types of dislocations, resulting in a decrease in the accuracy of dislocation detection. Further research showed that the diameter range of these residues invisible to the naked eye was about 1 μm to 120 μm, and the diameter range of more than 50% of the residues was 10 μm to 80 μm. Based on the above findings, in order to overcome the influence of the residues attached to the silicon carbide wafer surface caused by etching with a molten potassium hydroxide solution on the accuracy of dislocation detection, the present invention has improved the method for detecting dislocations in a silicon carbide wafer.

[0010] According to the residues adhering to the surface of the silicon carbide wafer caused by the corrosion of molten potassium hydroxide, after the silicon carbide wafer is corroded with molten potassium hydroxide in the present invention and before dislocation detection, various solvents are tried for cleaning. It is found that only N,N-dimethylacetamide (DMAC) has an effective effect on removing the residues. Specifically, on the surface of the silicon carbide wafer cleaned with DMAC, after being cleaned with DMAC again, the number of dislocations in the unit field of view measured remains unchanged; while on the surface of the silicon carbide wafer cleaned with acetone, or N,N-dimethylformamide (DMF), or dimethyl sulfoxide (DMSO), after being cleaned with DMAC again, the number of dislocations in the unit field of view measured all decreases. Among them, the decrease in the number of dislocations measured after cleaning with acetone or DMF and then with DMAC again is more obvious, indicating that the removal effect of acetone and DMF on the residues is poor. The number of dislocations measured after cleaning with DMSO and then with DMAC again decreases to a certain extent, indicating that DMSO has a slightly effective effect on removing the residues. The difficulty in improving the accuracy rate of dislocation detection using potassium hydroxide corrosion combined with the image recognition method is caused by the following reasons: First, the residues on the wafer surface caused by potassium hydroxide corrosion are invisible to the naked eye and are often overlooked, resulting in the dislocation number and dislocation density when there are misjudged dislocation points being the true values; second, the composition of the above-mentioned residues invisible to the naked eye is unknown, and its adhesion performance cannot be determined. The difficulty in removing the above-mentioned residues is relatively large. Even DMF, which is similar in nature to DMAC, cannot effectively remove the above-mentioned residues. Based on the above attempts, the present invention provides a high-accuracy dislocation detection method for silicon carbide wafers, that is, after the silicon carbide wafer to be tested is corroded with molten potassium hydroxide and before the dislocation detection, the silicon carbide wafer is cleaned with N,N-dimethylacetamide.

[0011] Specifically, the present invention provides a high-accuracy dislocation detection method for silicon carbide wafers, which includes the following steps:

[0012] Step S1: After the silicon carbide wafer is put into molten potassium hydroxide for corrosion, take it out and cool it;

[0013] Step S2: Rinse the cooled silicon carbide wafer;

[0014] Step S3: Use N,N-dimethylacetamide to clean the surface to be detected of the silicon carbide wafer;

[0015] Step S4: Rinse the surface to be detected after being cleaned with N,N-dimethylacetamide;

[0016] Step S5: Use a dislocation automatic tester to detect the dislocations on the surface to be detected.

[0017] As an optimization of the above detection method, in step S1, the temperature of the corrosion is 480-550°C.

[0018] Preferably in the above detection method, in step S1, the corrosion time is 10 to 25 minutes.

[0019] Preferably in the above detection method, in step S2, water is used for the rinsing.

[0020] Preferably in the above detection method, in step S3, the cleaning method is: pouring N,N-dimethylacetamide liquid onto the surface to be detected of the silicon carbide wafer to form a N,N-dimethylacetamide liquid pile with a thickness on the surface to be detected, then placing a cleaning tool on the surface to be detected, applying pressure to the cleaning tool to form a frictional force on the surface to be detected, and moving the tool in a circular motion or back-and-forth friction manner.

[0021] Preferably in the above detection method, the tool is selected from lint-free paper, brush, sponge, and lint-free cloth.

[0022] Preferably in the above detection method, in step S4, absolute ethanol is used for the rinsing.

[0023] Preferably in the above detection method, in step S5, the field of view area for the detection is not less than 0.01 cm 2 .

[0024] On the other hand, the present invention provides a silicon carbide wafer for dislocation detection. Compared with the prior art, its outstanding feature is that: after the silicon carbide wafer is corroded with molten potassium hydroxide, the diameter of the residue on the surface of the silicon carbide wafer caused by the corrosion is 1 μm to 120 μm;

[0025] And, after forming the residue on the surface of the silicon carbide wafer caused by the corrosion, it is cleaned with N,N-dimethylacetamide.

[0026] Compared with the prior art, the present invention has the following technical effects:

[0027] After potassium hydroxide corrodes the silicon carbide wafer, residues invisible to the naked eye will adhere to its surface. These residues cannot be removed by water washing or alcohol washing. Even solvents such as DMF and DMSO that can be miscible with most organic substances cannot effectively remove the above residues. In the prior art, when using potassium hydroxide corrosion combined with image recognition method to detect dislocations, the residues invisible to the naked eye on the surface of the wafer caused by potassium hydroxide corrosion are easily overlooked in the cleaning step, resulting in misjudgment of the number of dislocation points and dislocation density as the true values. The present invention makes the dislocation detection result of the silicon carbide wafer close to its true dislocation result and has a high accuracy by cleaning the surface of the silicon carbide wafer with DMAC after the silicon carbide wafer to be detected is corroded with molten potassium hydroxide and before the dislocation detection. Description of the Drawings

[0028] Figure 1 A flowchart of a method for detecting dislocations on a silicon carbide wafer with high accuracy provided for an exemplary embodiment;

[0029] Figure 2 A result diagram of dislocation detection for a unit field of view on the surface of a silicon carbide wafer cleaned with DMAC for the first time;

[0030] Figure 3 For the surface of a silicon carbide wafer cleaned with DMAC for the first time and DMAC for the second time, and Figure 2 A result diagram of dislocation detection for the same unit field of view;

[0031] Figure 4 A result diagram of dislocation detection for a unit field of view on the surface of a silicon carbide wafer cleaned with DMF for the first time;

[0032] Figure 5 For the surface of a silicon carbide wafer cleaned with DMF for the first time and DMAC for the second time, and Figure 4 A result diagram of dislocation detection for the same unit field of view. Detailed Description of the Invention

[0033] Here, the exemplary embodiments will be described in detail. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with one or more embodiments of the present invention. On the contrary, they are merely examples of devices and methods consistent with some aspects of one or more embodiments of the present invention as detailed in the appended claims.

[0034] To further illustrate the present invention, the following embodiments are provided:

[0035] The production process of silicon carbide crystal products generally includes: crystal growth, forming, cutting, thinning, polishing, packaging, and shipping. Before shipping, samples need to be taken for dislocation detection to characterize and evaluate the quality of this batch of silicon carbide crystal products. The dislocation detection in the present invention is sampling detection. Therefore, in the present invention, the cleaning step before dislocation detection is different from the cleaning steps in production processes such as cutting or polishing, and the cleaning step before dislocation detection is only for the impurities formed by the corrosion of the silicon carbide wafer with molten potassium hydroxide.

[0036] First, the present invention provides a method for detecting dislocations on a silicon carbide wafer with high accuracy, which includes the steps of: after being corroded with molten potassium hydroxide and before the dislocation detection, the silicon carbide wafer is cleaned with N,N-dimethylacetamide.

[0037] Generally, the dislocation detection of silicon carbide wafers is achieved by etching the silicon surface of the silicon carbide wafer with a potassium hydroxide melt to present etch pits at the dislocation sites, and then identifying the etch pits by image method to detect the number and density of dislocations, thereby realizing the detection of dislocations. However, according to the above detection method, when etching the surface of the silicon carbide wafer with a potassium hydroxide melt, it is easy to leave etch residues attached to the wafer surface. Although there are also prior arts suggesting that the etch residues on the sample surface after the potassium hydroxide melt etching will affect the identification of dislocations, most of the prior arts only target the residues visible to the naked eye, or directly perform routine water washing and alcohol washing without pertinence.

[0038] When the present inventors achieved the detection of dislocations through the above detection method, it was found that after etching the surface of the silicon carbide wafer with a potassium hydroxide melt, in addition to the residues visible to the naked eye attached to the wafer surface, there were also some residues invisible to the naked eye attached. These residues could not be removed by cleaning with organic alcohol solvents such as clean water and absolute ethanol. Under the microscope, due to their size and shape being similar to those of threading screw dislocation (TSD), threading edge dislocation (TED), or basal plane dislocation (BPD), these residues were easily misidentified as one of the above three types of dislocations, resulting in a decrease in the accuracy of dislocation detection. Further research found that the diameter range of these residues invisible to the naked eye was about 1μm - 120μm, and the diameter range of more than 50% of the residues was 10μm - 80μm. Based on the above findings, in order to overcome the influence of the residues attached to the surface of the silicon carbide wafer caused by the potassium hydroxide melt etching on the accuracy of dislocation detection, the present invention improved the dislocation detection method of the silicon carbide wafer.

[0039] According to the residues attached to the surface of the silicon carbide wafer caused by the potassium hydroxide melt etching, after the silicon carbide wafer is etched with molten potassium hydroxide and before dislocation detection, the present invention attempts to use a variety of solvents for cleaning and finds that only N,N-dimethylacetamide (DMAC) has an effective effect on removing the residues. Specifically, on the surface of the silicon carbide wafer cleaned with DMAC, after being cleaned with DMAC again, the number of dislocations in the unit field of view measured remains unchanged, as Figure 2 and Figure 3 As can be seen from the comparison, the number of dislocations is the same before and after being cleaned with DMAC again. Thus, it can be seen that the residues have been removed when DMAC is used for the first cleaning; while on the surface of the silicon carbide wafer cleaned with acetone, or acetone alcohol, or N,N-dimethylformamide (DMF), or dimethyl sulfoxide (DMSO), after being cleaned with DMAC again, the number of dislocations in the unit field of view measured all decreases, as Figure 4 andFigure 5 As can be seen from the comparison, DMF is used for the first cleaning and DMAC is used for the second cleaning. At this time, there is a significant difference in the number of detected dislocations before and after, and the number of dislocations decreases after using DMAC for the second time. It can be seen that a large part of this residue is not removed when using DMF for the first cleaning. Among them, the number of dislocations measured after using acetone or DMF for cleaning and then using DMAC for cleaning again decreases significantly, indicating that acetone and DMF have poor removal effects on this residue. The number of dislocations measured after using DMSO for cleaning and then using DMAC for cleaning again decreases to a certain extent, indicating that DMSO has a slightly better removal effect on this residue. The difficulty in improving the accuracy of detecting dislocations by using potassium hydroxide corrosion combined with image recognition method is caused by the following reasons: First, the residues on the surface of the wafer caused by potassium hydroxide corrosion are invisible to the naked eye and are often ignored, which may lead to misjudgment of the number of dislocation points and dislocation density as the true values; Second, the composition of the above-mentioned residues invisible to the naked eye is unknown, and its adhesion performance cannot be determined. It is difficult to remove the above-mentioned residues. Even DMF, which is similar in nature to DMAC, cannot effectively remove the above-mentioned residues. Based on the above attempts, the present invention provides a method for detecting dislocations on a silicon carbide wafer with high accuracy, that is, after the silicon carbide wafer to be tested is corroded with molten potassium hydroxide and before the dislocation detection, the silicon carbide wafer is cleaned with N,N-dimethylacetamide.

[0040] The above Figure 4 is a result diagram of dislocation detection for a unit field of view on the surface of a silicon carbide wafer where DMF is used for the first cleaning. The above Figure 5 is a result diagram of dislocation detection for the same unit field of view on the surface of a silicon carbide wafer where DMF is used for the first cleaning and DMAC is used for the second cleaning. However, due to normal instrument errors, the points in the unit field of view fluctuate, Figure 4 and the position on the wafer surface corresponding to the unit field of view has a slight displacement from Figure 5 the position on the wafer surface corresponding to the unit field of view, which is within the normal error range, Figure 4 and most of the sites of the true dislocations in Figure 4 and Figure 5 still overlap.

[0041] Based on the results of attempting to use various solvents to clean the above-mentioned residues, it shows the unexpected effect of using DMAC to clean and remove the residues attached to the surface of the silicon carbide wafer caused by the corrosion of potassium hydroxide molten liquid. DMF and DMAC have similar structures and properties, and in many practical examples, it shows that DMF and DMAC have similar solubility. For example, linear polyester resins in the form of viscous liquids or solids are usually insoluble in water and ethanol, but this linear polyester resin has good solubility in both DMF and DMAC. However, the above-mentioned residues have good solubility in DMAC and poor solubility in DMF. Therefore, it can be shown that the present invention has a better effect of using DMAC to dissolve and remove the above-mentioned residues; DMSO is well-known for its excellent dissolving ability and can dissolve a relatively large variety of substances. However, the above-mentioned residues have poor solubility in DMSO, while they have good solubility in DMAC. Therefore, it can be shown that the present invention has a better effect of using DMAC to dissolve and remove the above-mentioned residues; Acetone and acetone alcohol have a certain structural similarity, but there is a large difference in the effect of dissolving and removing the above-mentioned residues between acetone and acetone alcohol. Acetone has a poor dissolving effect on the above-mentioned residues, while acetone alcohol has a certain effect on the dissolution of the above-mentioned residues, and the dissolving effect of acetone alcohol is greater than that of DMSO and DMF, and the dissolving effect of DMAC is greater than that of acetone alcohol. Therefore, it can be shown that the present invention has a better effect of using DMAC to dissolve and remove the above-mentioned residues. In summary, the results of cleaning the above-mentioned residues with various solvents jointly prove the unexpected effect of using DMAC to clean and remove the residues attached to the surface of the silicon carbide wafer caused by the corrosion of potassium hydroxide molten liquid before the present invention was proposed.

[0042] Figure 1 FIG. 4 is a flowchart of a method for detecting dislocations on a silicon carbide wafer with high accuracy provided by an exemplary embodiment of the present invention. The dislocation detection method includes the following steps:

[0043] Step S1: After the silicon carbide wafer is corroded in molten potassium hydroxide, take it out and cool it;

[0044] Step S2: Rinse the cooled silicon carbide wafer;

[0045] Step S3: Use N,N-dimethylacetamide to clean the surface to be detected of the silicon carbide wafer;

[0046] Step S4: Rinse the surface to be detected after being cleaned with N,N-dimethylacetamide;

[0047] Step S5: Use a dislocation automatic tester to detect the dislocations on the surface to be detected.

[0048] In this embodiment, after using molten potassium hydroxide to etch the silicon carbide wafer, general impurities on the surface of the silicon carbide wafer are removed by rinsing respectively, then residues difficult to remove adhering to the surface of the silicon carbide wafer caused by the etching of the potassium hydroxide melt are dissolved and removed by cleaning with N,N-dimethylacetamide, and then N,N-dimethylacetamide and the residues dissolved therein are removed by rinsing. Finally, the dislocation situation of the surface to be detected, including the number of dislocations and the dislocation density, is measured by a dislocation automatic tester. Among them, the dislocation density is: the total number of dislocations in all unit fields of view / the total area of the field of view.

[0049] The general impurities can be understood as impurities on the surface of the silicon carbide wafer that can be removed by clean water or with a certain flushing force. For example, dust adhering to the surface of the silicon carbide wafer, potassium hydroxide, residues adhering to the surface of the silicon carbide wafer caused by the etching of the potassium hydroxide melt with relatively small adhesion force, etc., including, but not limited to, these.

[0050] The dislocation automatic tester can be understood as an automated instrument that counts the types and numbers of dislocations according to the shape of the etch pits. It can be a common dislocation automatic tester on the market. For example, the SiC etch wafer defect detector of the FabXLab-LFM series of QianShi Technology.

[0051] In one embodiment, in the above step S1, the temperature of the etching is 480 - 550 °C, and the time of the etching is 10 - 25 minutes.

[0052] Etching for 10 - 25 minutes at an etching temperature of 480 - 550 °C can expose the dislocations on the silicon surface of the silicon carbide wafer according to the theory of preferential chemical etching, presenting etch pits of a certain shape and size. Specifically: screw dislocations present hexagonal etch pits with a diameter of about 100 μm - 120 μm, a pointed bottom and slightly biased to one side; edge dislocations present approximately circular etch pits with a diameter of about 50 μm - 80 μm, a pointed bottom and slightly biased to one side; basal plane dislocations present elliptical etch pits with a diameter of about 30 μm - 80 μm, a bottom and severely biased to one side of the ellipse. After the etch pits characterizing the dislocations are exposed, they are identified and counted using a dislocation automatic tester.

[0053] In one embodiment, in the above step S2, water is used for the rinsing.

[0054] According to the purpose of the above step S2, it is preferably to use water for rinsing. Under the condition of meeting the requirement of removing general impurities on the surface of the silicon carbide wafer, the cost can be reduced, and it can play a certain cooling role for the silicon carbide wafer.

[0055] In one embodiment, in the above step S3, the cleaning method is as follows: Pour N,N-dimethylacetamide liquid onto the surface to be detected of the silicon carbide wafer to form an N,N-dimethylacetamide liquid pile with a thickness on the surface to be detected. Then, place the cleaning tool on the surface to be detected, apply pressure to the cleaning tool to form a frictional force on the surface to be detected, and move the tool in a back-and-forth friction manner.

[0056] Specifically, the back-and-forth friction can be performed in a circular motion or in other ways, as long as the surface formed by the movement trajectory covers the wafer surface.

[0057] Although N,N-dimethylacetamide can dissolve the residues attached to the surface of the silicon carbide wafer caused by the corrosion of the above-mentioned potassium hydroxide molten liquid and has a good dissolution ability, due to the relatively stubborn residual adhesion force of the above-mentioned residues, certain means are needed to completely remove them. For example, by increasing the use of frictional force to remove the above-mentioned residues on the premise of the dissolution of N,N-dimethylacetamide, the removal effect is more excellent.

[0058] In one embodiment, the tool is selected from lint-free paper, brush, sponge, and lint-free cloth. The above tools have equivalent effects as the tools used in step S3.

[0059] In one embodiment, in the above step S4, anhydrous ethanol is used for the rinsing.

[0060] According to the purpose of the above step S4, it is preferably to use anhydrous ethanol for rinsing, which can quickly wash away N,N-dimethylacetamide and the residues dissolved therein, and has a short drying time, so that the next step can be carried out quickly. Among them, the drying can be drying in an oven, natural air drying, or drying with lint-free paper or lint-free cloth.

[0061] In one embodiment, in the above step S5, the field of view area of the detection is not less than 0.01 cm 2 .

[0062] For dislocation detection, a larger field of view area is selected, and the number of sampling detection samples is large, which can more accurately represent the number of dislocations on the surface of the tested silicon carbide wafer and can improve the detection accuracy. The field of view area should be understood as the total field of view area.

[0063] Secondly, the present invention provides a silicon carbide wafer for dislocation detection. Compared with the prior art, its outstanding feature is that after the silicon carbide wafer is corroded with molten potassium hydroxide, the diameter of the residues on the surface of the silicon carbide wafer caused by the corrosion is 1 μm to 120 μm;

[0064] And after forming the residues on the surface of the silicon carbide wafer caused by the corrosion, it is cleaned with N,N-dimethylacetamide.

[0065] In the research of the present invention, it is found that after a silicon carbide wafer is etched with molten potassium hydroxide, invisible residues will be formed on its surface. Moreover, these residues are difficult to remove by ordinary washing with water or ethanol. Further research reveals that the diameter range of these invisible residues is 1 μm to 120 μm. Since the size and surface shape of these residues are close to those of the corrosion pits formed by dislocations, they are easily mistaken for dislocations, which results in a low accuracy rate of dislocation testing. Suppressing the influence of these residues can improve the accuracy rate of dislocation detection. More specifically, since the shape and size of residues larger than 120 μm are quite different from those of the corrosion pits formed by dislocations and are more obvious under a microscope, the probability of misjudgment is relatively low. And residues smaller than 1 μm are difficult to be recognized by the automatic dislocation tester, and the probability of misjudgment is also relatively low. Therefore, suppressing the influence of residues with a diameter range of 1 μm to 120 μm can more effectively improve the accuracy rate of dislocation detection.

[0066] Based on the present invention, the above-mentioned "residues" can be defined as substances other than silicon carbide formed when a silicon carbide wafer is placed in molten potassium hydroxide, which remain on the surface of the silicon carbide wafer and can be formed by the etching of the silicon carbide wafer with molten potassium hydroxide.

[0067] The present invention will be further described below in conjunction with embodiments. In the following embodiments, the automatic dislocation tester used is the SiC etching wafer defect detector of the FabXLab-LFM series of QianShi Technology, and the total field of view area for dislocation detection is 26.88 cm 2 (a total of 1574 test points), and the 8-inch silicon carbide wafers used for dislocation detection are the same batch of silicon carbide single crystal wafers. Those of ordinary skill in the art will be able to implement the present invention based on these descriptions. In addition, the embodiments of the present invention involved in the following descriptions are usually only a part of the embodiments of the present invention, rather than all the embodiments. Therefore, all other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0068] Embodiment 1

[0069] A method for detecting dislocations in a silicon carbide wafer is provided, including the following steps:

[0070] Step S1: Place an 8-inch silicon carbide wafer into molten potassium hydroxide, etch it at 500 °C for 20 minutes, take it out and cool it down;

[0071] Step S2: Rinse the cooled silicon carbide wafer with water;

[0072] Step S3: Pour 5 mL of N,N-dimethylacetamide liquid onto the surface to be detected of the silicon carbide wafer to form an N,N-dimethylacetamide liquid pile with a thickness on the surface to be detected. Then place a lint-free paper on the surface to be detected, apply pressure to the lint-free paper to form a frictional force on the surface to be detected, and move the lint-free paper in a circular motion so that the surface formed by the movement track of the lint-free paper covers the entire surface to be detected;

[0073] Step S4: Rinse the surface to be detected after being cleaned in Step S3 with absolute ethanol, and then dry it with a lint-free paper;

[0074] Step S5: Use a dislocation automatic tester to detect the number and dislocation density of screw dislocations, edge dislocations, and basal plane dislocations on the surface to be detected. The results of screw dislocations, edge dislocations, and basal plane dislocations are shown in Table 1 respectively. This test record is the first quantity and the first density.

[0075] Example 2

[0076] Provide a method for detecting dislocations of a silicon carbide wafer, which is only different from Example 1 in that: Step S1 is different. Other steps are the same as those in Example 1.

[0077] The Step S1 of this example is as follows:

[0078] Put an 8-inch silicon carbide wafer into molten potassium hydroxide, corrode it at 480 °C for 25 minutes, take it out and cool it.

[0079] Example 3

[0080] Provide a method for detecting dislocations of a silicon carbide wafer, which is only different from Example 1 in that: Step S1 is different. Other steps are the same as those in Example 1.

[0081] The Step S1 of this example is as follows:

[0082] Put an 8-inch silicon carbide wafer into molten potassium hydroxide, corrode it at 550 °C for 10 minutes, take it out and cool it.

[0083] Example 4

[0084] Provide a method for detecting dislocations of a silicon carbide wafer, which is only different from Example 1 in that: the cleaning method in Step S3 is different. Other steps are the same as those in Example 1.

[0085] The Step S3 of this example is as follows:

[0086] Completely immerse the silicon carbide wafer in N,N-dimethylacetamide and soak it for 10 minutes.

[0087] Example 5

[0088] A method for detecting dislocations in a silicon carbide wafer is provided. The difference from Example 1 is only that: the cleaning method in step S3 is different. Other steps are the same as those in Example 1.

[0089] The step S3 of this embodiment is as follows:

[0090] Rinse the surface to be detected of the silicon carbide wafer with N,N-dimethylacetamide for 12 minutes.

[0091] Example 6

[0092] A method for detecting dislocations in a silicon carbide wafer is provided. The difference from Example 1 is only that: 10 mL of N,N-dimethylacetamide is used in step S3. Other steps are the same as those in Example 1.

[0093] The step S3 of this embodiment is as follows:

[0094] Pour 10 mL of N,N-dimethylacetamide liquid onto the surface to be detected of the silicon carbide wafer to form an N,N-dimethylacetamide liquid stack with a thickness on the surface to be detected. Then place a lint-free paper on the surface to be detected, apply pressure to the lint-free paper to form a frictional force on the surface to be detected, and move the lint-free paper in a circular motion so that the surface formed by the movement track of the lint-free paper covers the entire surface to be detected.

[0095] Example 7

[0096] A method for detecting dislocations in a silicon carbide wafer is provided. The difference from Example 1 is only that: 1 mL of N,N-dimethylacetamide is used in step S3. Other steps are the same as those in Example 1.

[0097] The step S3 of this embodiment is as follows:

[0098] Pour 1 mL of N,N-dimethylacetamide liquid onto the surface to be detected of the silicon carbide wafer, then place a lint-free paper on the surface to be detected, apply pressure to the lint-free paper to form a frictional force on the surface to be detected, and move the lint-free paper in a circular motion so that the surface formed by the movement track of the lint-free paper covers the entire surface to be detected.

[0099] Example 8

[0100] A method for detecting dislocations in a silicon carbide wafer is provided. The difference from Example 1 is only that: 5 mL of acetone alcohol is used in step S3. Other steps are the same as those in Example 1.

[0101] The step S3 of this embodiment is as follows:

[0102] Pour 5 mL of acetone alcohol liquid onto the surface to be detected of the silicon carbide wafer to form an N,N-dimethylacetamide liquid pile with a thickness on the surface to be detected. Then, place the lint-free paper on the surface to be detected, apply pressure to the lint-free paper to form a frictional force on the surface to be detected, and move the lint-free paper in a circular motion so that the surface formed by the movement trajectory of the lint-free paper covers the entire surface to be detected.

[0103] Example 9

[0104] Provide a method for detecting dislocations in a silicon carbide wafer. The difference from Example 1 is only that: 10 mL of acetone alcohol is used in step S3. Other steps are the same as those in Example 1.

[0105] The step S3 of this example is as follows:

[0106] Pour 10 mL of acetone alcohol liquid onto the surface to be detected of the silicon carbide wafer to form an N,N-dimethylacetamide liquid pile with a thickness on the surface to be detected. Then, place the lint-free paper on the surface to be detected, apply pressure to the lint-free paper to form a frictional force on the surface to be detected, and move the lint-free paper in a circular motion so that the surface formed by the movement trajectory of the lint-free paper covers the entire surface to be detected.

[0107] Example 10

[0108] Provide a method for detecting dislocations in a silicon carbide wafer. The difference from Example 1 is only that: step S4 is different. Other steps are the same as those in Example 1.

[0109] The step S4 of this example is as follows:

[0110] Rinse the surface to be detected after being cleaned in step S3 with water, and then dry it with lint-free paper.

[0111] Comparative Example 1

[0112] Provide a method for detecting dislocations in a silicon carbide wafer. The difference from Example 1 is only that: 5 mL of DMF is used in step S3. Other steps are the same as those in Example 1.

[0113] The step S3 of this example is as follows:

[0114] Pour 5 mL of DMF liquid onto the surface to be detected of the silicon carbide wafer to form an N,N-dimethylacetamide liquid pile with a thickness on the surface to be detected. Then, place the lint-free paper on the surface to be detected, apply pressure to the lint-free paper to form a frictional force on the surface to be detected, and move the lint-free paper in a circular motion so that the surface formed by the movement trajectory of the lint-free paper covers the entire surface to be detected.

[0115] Comparative Example 2

[0116] A method for detecting dislocations in a silicon carbide wafer, which is only different from Example 1 in that: 5 mL of DMSO is used in step S3. Other steps are the same as those in Example 1.

[0117] Step S3 of this embodiment is as follows:

[0118] Pour 5 mL of DMSO liquid onto the surface to be detected of the silicon carbide wafer to form an N,N-dimethylacetamide liquid pile with a thickness on the surface to be detected. Then place a lint-free paper on the surface to be detected, apply pressure to the lint-free paper to form a frictional force on the surface to be detected, and move the lint-free paper in a circular motion so that the surface formed by the movement trajectory of the lint-free paper covers the entire surface to be detected.

[0119] Comparative Example 3

[0120] A method for detecting dislocations in a silicon carbide wafer, which is only different from Example 1 in that: 5 mL of acetone is used in step S3. Other steps are the same as those in Example 1.

[0121] Step S3 of this embodiment is as follows:

[0122] Pour 5 mL of acetone liquid onto the surface to be detected of the silicon carbide wafer to form an N,N-dimethylacetamide liquid pile with a thickness on the surface to be detected. Then place a lint-free paper on the surface to be detected, apply pressure to the lint-free paper to form a frictional force on the surface to be detected, and move the lint-free paper in a circular motion so that the surface formed by the movement trajectory of the lint-free paper covers the entire surface to be detected.

[0123] Characterization of method accuracy

[0124] For the silicon carbide wafers after dislocation detection in Examples 1 to 10 and Comparative Examples 1 to 3, after all the silicon carbide wafers are subjected to surface cleaning of the surface to be detected through step S3 and step S4 of Example 1, the number and dislocation density of screw dislocations, edge dislocations, and basal plane dislocations on the surface to be detected are detected again using a dislocation automatic tester. The results of screw dislocations, edge dislocations, and basal plane dislocations are shown in Table 1, Table 2, and Table 3 respectively, and are recorded as the number of repeated tests and the density of repeated tests. Among them, the smaller the error rate, the higher the accuracy. The calculation formula for the error rate is: Error rate = (first density - repeated test density) / repeated test density.

[0125] Table 1 Detection results of screw dislocations

[0126] Initial Quantity (pcs) <![CDATA[Initial density (pcs / cm 2 )]]> Repeated Test Quantity (pcs) <![CDATA[Repeat test density (pcs / cm 2 )]]> Error Rate (%) Example 1 902 34 902 34 0.00 Example 2 917 34 917 34 0.00 Example 3 901 34 901 34 0.00 Example 4 909 34 900 33 1.00 Example 5 906 34 901 34 0.55 Example 6 894 33 894 33 0.00 Example 7 998 37 990 37 0.81 Example 8 1010 38 910 34 10.99 Example 9 1019 38 941 35 8.29 Example 10 934 35 934 35 0.00 Comparative Example 1 1069 40 901 34 18.65 Comparative Example 2 1121 42 972 36 15.33 Comparative Example 3 1106 41 894 33 23.71

[0127] Table 2 Detection results of edge dislocations

[0128] Initial Quantity (pcs) <![CDATA[Initial density (pcs / cm 2 ).]]> Repeated Test Quantity (pcs) <![CDATA[Repeated test density (pcs / cm 2 )]]> Error Rate (%) Example 1 50124 1865 50124 1865 0.00 Example 2 50235 1869 50235 1869 0.00 Example 3 50121 1865 50121 1865 0.00 Example 4 50044 1862 49870 1860 0.35 Example 5 50178 1867 50113 1864 0.13 Example 6 50220 1868 50220 1868 0.00 Example 7 50127 1865 50120 1865 0.01 Example 8 50989 1897 50091 1864 1.79 Example 9 50734 1887 50064 1863 1.34 Example 10 50746 1888 50744 1888 0.00 Comparative Example 1 51126 1902 50026 1861 2.20 Comparative Example 2 51124 1902 50017 1861 2.21 Comparative Example 3 51237 1906 50039 1862 2.39

[0129] Table 3 Detection results of basal plane dislocations

[0130] Initial Quantity (pcs) <![CDATA[Initial density (number / cm 2 ).]]> Repeated Test Quantity (pcs) <![CDATA[Repeated test density (pcs / cm 2 )]]> Error Rate (%) Example 1 12356 460 12356 460 0.00% Example 2 12098 450 12098 450 0.00% Example 3 12458 463 12458 463 0.00% Example 4 12249 456 12217 455 0.26% Example 5 12147 452 12131 451 0.13% Example 6 12088 450 12088 450 0.00% Example 7 12476 464 12469 464 0.06% Example 8 12457 463 12267 456 1.55% Example 9 12346 459 12222 455 1.01% Example 10 12350 459 12350 459 0.00% Comparative Example 1 13251 493 12794 476 3.57% Comparative Example 2 13156 489 12655 471 3.96% Comparative Example 3 13448 500 12427 462 8.22%

[0131] It can be seen from the data analysis in Tables 1 to 3 that:

[0132] (1) The corrosion of silicon carbide wafers by molten potassium hydroxide solution will cause residues difficult to remove to adhere to the wafer surface. The size and shape of these residues are similar to the corrosion pits caused by screw dislocations, edge dislocations or basal plane dislocations, resulting in these residues being easily misjudged as dislocations. Furthermore, when there are misjudged dislocation points, the number and density of dislocations are the true values. For example, the initial number and initial density in Comparative Examples 1 to 3 would be misjudged as the true number and density of dislocations before the present invention.

[0133] (2) It can be seen from Examples 1 to 3 that in the present invention, after the silicon carbide wafer to be tested is corroded with molten potassium hydroxide and before the dislocation detection, by cleaning the surface of the silicon carbide wafer with DMAC, the results of the number and density of dislocations obtained when detecting dislocations again can be exactly the same as the results of the first detection, making the dislocation detection result of the silicon carbide wafer almost close to or its true dislocation result, with a relatively high accuracy.

[0134] (3) It can be seen from the comparative analysis of Example 1 with Comparative Examples 1 to 3 that only DMAC has an effective effect on removing this residue. Specifically, on the surface of the silicon carbide wafer cleaned with DMAC, when cleaned with DMAC again, the number of dislocations in the unit field of view measured remains unchanged; while on the surface of the silicon carbide wafer cleaned with DMF (Comparative Example 1), or DMSO (Comparative Example 2), or acetone (Comparative Example 3), when cleaned with DMAC again, the number of dislocations in the unit field of view measured all decreases. Among them, the decrease in the number of dislocations measured after cleaning with acetone or DMF and then with DMAC is more obvious, indicating that acetone and DMF have a poor effect on removing this residue. The number of dislocations measured after cleaning with DMSO and then with DMAC decreases to a certain extent, indicating that DMSO has a slightly better effect on removing this residue. It can be seen from the comparative analysis of Example 1 with Example 8 and Comparative Example 3 that the effect of DMAC on removing the above residue is greater than that of acetone alcohol. Although acetone and acetone alcohol have a certain structural similarity, the effect of acetone and acetone alcohol on dissolving and removing the above residue is quite different. Acetone has a poor effect on dissolving the above residue, while acetone alcohol has a certain effect on dissolving the above residue. Through the above analysis, it can be seen that using DMAC to clean and remove the residue adhering to the surface of the silicon carbide wafer caused by the corrosion of molten potassium hydroxide solution has an unexpected effect.

[0135] (4)It can be seen from the comparative analysis of Example 1 and Examples 4 to 5 that the effect of removing the above residues can be improved by optimizing the cleaning means. Specifically, in Example 1: on the premise of dissolving in N,N-dimethylacetamide, the use of friction is increased to remove the above residues, and the removal effect is more excellent.

[0136] Unless otherwise specified, the raw materials and equipment used in the present invention are all common raw materials and equipment in the art; unless otherwise specified, the methods used in the present invention are all conventional methods in the art.

[0137] The above are only the preferred embodiments of the present invention, and do not impose any limitations on the present invention. Any simple modifications, changes and equivalent transformations made to the above embodiments according to the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A method for detecting dislocations in silicon carbide wafers with high accuracy, characterized in that: After being etched with molten potassium hydroxide and before dislocation detection, the silicon carbide wafer is cleaned with N,N-dimethylacetamide.

2. The method for detecting dislocations of a silicon carbide wafer with high accuracy according to claim 1, characterized in that: The method includes the following steps: Step S1: After the silicon carbide wafer is placed in molten potassium hydroxide for etching, it is taken out and cooled. Step S2: Rinse the cooled silicon carbide wafer. Step S3: Clean the surface to be detected of the silicon carbide wafer with N,N-dimethylacetamide. Step S4: Rinse the surface to be detected after being cleaned with N,N-dimethylacetamide. Step S5: Use a dislocation automatic tester to detect the dislocations on the surface to be detected.

3. The method for detecting dislocations of a silicon carbide wafer with high accuracy according to claim 2, characterized in that: In step S1, the temperature of the etching is 480 - 550 °C.

4. A method for detecting dislocations of silicon carbide wafers with high accuracy as claimed in claim 2 or 3, characterized in that: In step S1, the time of the etching is 10 - 25 minutes.

5. The method for detecting dislocations of a silicon carbide wafer with high accuracy according to claim 2, wherein: In step S2, water is used for the rinsing.

6. The method for detecting dislocations of a silicon carbide wafer with high accuracy according to claim 2, wherein: In step S3, the cleaning method is as follows: Pour the N,N-dimethylacetamide liquid onto the surface to be detected of the silicon carbide wafer to form a N,N-dimethylacetamide liquid pile with a thickness on the surface to be detected, then place the cleaning tool on the surface to be detected, apply pressure to the cleaning tool to form a frictional force on the surface to be detected, and move the tool in a back-and-forth friction manner.

7. The method for detecting dislocations of a silicon carbide wafer with high accuracy according to claim 6, characterized in that: The tool is selected from dust-free paper, brush, sponge, and dust-free cloth.

8. The method for detecting dislocations of a silicon carbide wafer with high accuracy according to claim 2, wherein: In step S4, anhydrous ethanol is used for the rinsing.

9. The method for detecting dislocations of a silicon carbide wafer with high accuracy according to claim 2, characterized in that: In step S5, the detected field of view area is not less than 0.01 cm 2 .

10. A silicon carbide wafer for dislocation detection, characterized in that: After the silicon carbide wafer is etched with molten potassium hydroxide, the diameter of the surface residue of the silicon carbide wafer caused by the etching is 1 μm - 120 μm; Moreover, after the surface residue of the silicon carbide wafer caused by the etching is formed, it is cleaned with N,N-dimethylacetamide.

Images