Method for monitoring warping of silicon carbide wafer
By setting sampling points on the surface of the silicon carbide wafer and measuring roughness, monitoring the grinding quality, the compatibility of warping measurement of ultra-thin silicon carbide wafers is solved, timely detection and prevention of warping is achieved, and processing quality is ensured. stability.
Patent Information
- Application Number
- CN202410971832.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-19
- Publication Date
- 2024-11-22
AI Technical Summary
The existing wafer warpage measurement equipment is not compatible with silicon carbide ultra-thin sheets, making it difficult to monitor warpage defects during the grinding process, affecting subsequent back process operations.
N sampling points are set on the surface of the silicon carbide wafer, the roughness of the sampling points is measured by a microscope, and the grinding quality is monitored according to the roughness. If abnormal, the grinding equipment will be replaced or the process parameters will be adjusted to ensure that the warping is within a controllable range.
By monitoring the roughness of the sampling point surface of the silicon carbide wafer, abnormal grinding quality can be discovered in a timely manner, avoid warping and exceeding specifications, ensure processing quality, and ensure the normal progress of subsequent process operations.
Smart Images

Figure CN119008440A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of wafer processing, and in particular to a method for monitoring the warping of a silicon carbide wafer. Background Art
[0002] For silicon carbide power devices such as silicon carbide diodes, the thinner the device thickness, the lower the threshold voltage VF and the higher the surge parameters. The larger the size of the silicon carbide wafer, the more single dies can be made, reducing the cost of a single die. Therefore, silicon carbide wafers are developing towards ultra-thin and large sizes. However, during the thinning process of silicon carbide wafers, if there is a problem with the quality of the grinding wheel, such as changing the grinding wheel combination, changing the grinding process, abnormal silicon carbide wafer incoming materials and / or abnormal front-end processes, etc., the grinding quality will be affected, resulting in warping defects in silicon carbide ultra-thin wafers.
[0003] During the grinding process of silicon carbide wafers, a large amount of stress will accumulate in the silicon carbide wafers, causing excessive warping of the silicon carbide wafers and limiting the back-side process operations. The silicon carbide wafer mass production line needs to monitor the grinding quality and promptly discover the warping caused by the grinding quality not meeting the requirements, and take corresponding measures to prevent abnormal wafer transmission during the subsequent back-side process operations, avoid fragments and damage to the robot arm.
[0004] Currently, wafer warpage measurement equipment is mainly used for conventional thickness and is not compatible with ultra-thin silicon carbide wafers. Therefore, a method is needed to measure the warpage of ultra-thin silicon carbide wafers. Summary of the invention
[0005] In order to overcome the problems existing in the related art, the purpose of the present invention is to provide a method for monitoring the warpage of silicon carbide wafers. The method can measure the warpage of silicon carbide ultra-thin wafers, so as to timely discover abnormal grinding quality, avoid abnormal wafer transmission during the back-side process operation, and ensure the processing quality of silicon carbide wafers.
[0006] A method for monitoring warpage of a silicon carbide wafer, comprising:
[0007] N sampling points are set on the surface of the silicon carbide wafer; wherein N ≥ 2;
[0008] Measuring the roughness of N sampling points, wherein the roughness is measured under a microscope;
[0009] monitoring the grinding quality according to the roughness, wherein the roughness is positively correlated with the warpage of the silicon carbide wafer;
[0010] If the grinding quality is abnormal, replace the grinding equipment or adjust the process parameters.
[0011] In a preferred technical solution of the present invention, the replacement of the grinding equipment comprises:
[0012] Replace the current grinding wheel with an alternative grinding wheel;
[0013] Replace current SiC wafers with replacement SiC wafers.
[0014] In a preferred technical solution of the present invention, the adjusting process parameters includes:
[0015] Increase or decrease the current grinding speed;
[0016] Increase or decrease the current feed rate.
[0017] In a preferred technical solution of the present invention, the step of monitoring the grinding quality according to the roughness includes:
[0018] Calculating N reference intervals of the roughness;
[0019] Determining the roughness outside the reference interval as abnormal roughness;
[0020] Measuring the roughness of the sampling points corresponding to the abnormal roughness for multiple times;
[0021] If the roughness of the sampling point is still abnormal after multiple measurements, it is determined that the grinding is abnormal, the processing is stopped, and the cause of the grinding abnormality is checked.
[0022] In a preferred technical solution of the present invention, the calculation of N reference intervals of roughness includes:
[0023] Calculating an average value of N roughness values to obtain an average roughness;
[0024] Calculating the standard deviations of the N roughnesses to obtain the roughness standard deviation;
[0025] Subtract three times the roughness standard deviation from the average roughness to obtain the lower limit of the value distribution;
[0026] Add three times the standard deviation of the roughness to the average roughness to obtain the upper limit of the value distribution;
[0027] The lower limit of the numerical distribution and the upper limit of the numerical distribution constitute a reference interval.
[0028] In a preferred technical solution of the present invention, the step of setting N sampling points on the surface of the silicon carbide wafer includes:
[0029] Randomly select N sampling points on the surface of the silicon carbide wafer;
[0030] Or select a first straight line and a second straight line on the surface of the silicon carbide wafer, select K1 sampling points on the first straight line, and select K2 sampling points on the second straight line; wherein the first straight line intersects or is parallel to the second straight line, and N=K1+K2.
[0031] In a preferred technical solution of the present invention, the measuring the roughness of the N sampling points includes:
[0032] Select a sampling pattern, wherein the sampling pattern is a rectangle or a circle;
[0033] Determine the area of the i-th sampling point according to the sampling graph;
[0034] The roughness in the area of the i-th sampling point is measured.
[0035] In a preferred technical solution of the present invention, the step of determining the area of the i-th sampling point according to the sampling graph includes:
[0036] If the sampling pattern is a rectangle, the area of the i-th sampling point is determined based on the length and width, taking the i-th sampling point as the center;
[0037] If the sampling pattern is a circle, the area of the i-th sampling point is determined according to the radius with the i-th sampling point as the center.
[0038] The beneficial effects of the present invention are:
[0039] The method for monitoring the warpage of silicon carbide wafers provided by the present invention includes setting N sampling points on the surface of silicon carbide wafers and measuring the roughness of the N sampling points. The grinding quality is monitored according to the roughness. If the grinding quality is abnormal, the grinding equipment is replaced or the process parameters are adjusted. The roughness of the silicon carbide wafer surface is positively correlated with the warpage of the silicon carbide wafer. N sampling points are set on the surface of the silicon carbide wafer, N≥2, and the roughness of the N sampling points can more comprehensively reflect the processing conditions of various positions on the surface of the silicon carbide wafer. If the roughness of a sampling point is large, it means that the local warpage of the silicon carbide wafer at this point is large; if the roughness of the N sampling points is large, it means that the overall warpage of the silicon carbide wafer is large. In the process of processing silicon carbide wafers, the roughness of the N sampling points on the surface of the silicon carbide wafer is monitored. According to the roughness, it can be timely judged whether the grinding quality is abnormal. When the grinding quality is abnormal, the grinding equipment is replaced in hardware, or the process parameters are adjusted in software, so as to avoid processing abnormalities in subsequent processing. The microscope has a high resolution and can reflect the roughness of the surface of the silicon carbide ultra-thin wafer. The roughness of the sampling point can be measured under the microscope to monitor the warpage of the silicon carbide ultra-thin wafer. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] Figure 1is a schematic flow diagram of a method for monitoring warpage of a silicon carbide wafer;
[0041] Figure 2 It is a flowchart for determining whether there is grinding abnormality and troubleshooting the factors causing grinding abnormality;
[0042] Figure 3 It is a schematic diagram of the process of calculating N reference intervals of roughness;
[0043] Figure 4 It is the correspondence between roughness and warpage under different grinding conditions;
[0044] Figure 5 It is a schematic diagram of multiple sampling points on the surface of SiC wafer. DETAILED DESCRIPTION
[0045] The preferred embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the preferred embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.
[0046] Example 1
[0047] like Figure 1 As shown, this embodiment provides a method for monitoring the warpage of a silicon carbide wafer, comprising:
[0048] S1: N sampling points are set on the surface of the silicon carbide wafer; wherein N ≥ 2.
[0049] The more sampling points N on the surface of the silicon carbide wafer, the more comprehensive the overall roughness of the silicon carbide wafer can be. However, if the number of sampling points N is too large, it will take too long to detect the surface roughness of each silicon carbide wafer. Therefore, N cannot be set too small or too large. In this embodiment, N is set to 5 as an example.
[0050] As an example, sampling point 2 is at the center, and sampling point 1, sampling point 3, sampling point 4 and sampling point 5 are distributed around sampling point 2, thereby more comprehensively reflecting the roughness of the silicon carbide wafer surface.
[0051] S2: Measuring the roughness of the N sampling points, wherein the roughness is measured under a microscope.
[0052] The surface of the silicon carbide wafer is observed under an AFM atomic mechanics microscope. Since the thickness of the silicon carbide wafer detected in this embodiment is relatively small and it is an ultra-thin silicon carbide wafer, an instrument is used to tap the surface of the silicon carbide wafer. Under the combined action of the force of the instrument tapping the surface of the silicon carbide wafer and the reaction force of the surface of the silicon carbide wafer, the surface scanned by the AFM atomic mechanics microscope is up and down. The roughness of the sampling point can be judged based on the surface scanned by the AFM atomic mechanics microscope.
[0053] S3: Monitoring the grinding quality according to the roughness, wherein the roughness is positively correlated with the warpage of the silicon carbide wafer.
[0054] like Figure 4 As shown, after multiple tests on the production line, it is determined that the roughness of the silicon carbide wafer surface can characterize the warpage of silicon carbide. The greater the roughness of the silicon carbide wafer surface, the greater the warpage of the silicon carbide wafer; the smaller the roughness of the silicon carbide wafer surface, the smaller the warpage of the silicon carbide wafer.
[0055] There is a linear relationship between the roughness of the silicon carbide wafer surface and the warpage of the silicon carbide wafer, which means that roughness is the main factor affecting the warpage. It is convenient and quick to detect the roughness alone, which can quickly reflect the warpage of the silicon carbide wafer, thereby monitoring the warpage of the silicon carbide wafers on the production line in real time and detecting warpage abnormalities in time.
[0056] S4: If the grinding quality is abnormal, replace the grinding equipment or adjust the process parameters.
[0057] When the roughness of one or several sampling points among the N sampling points is too large, it means that the local roughness of the silicon carbide wafer surface is too large, and the warping of the silicon carbide wafer may be too large; when the roughness of the N sampling points is too large, it means that the overall roughness of the silicon carbide wafer surface is too large, and the warping of the silicon carbide wafer may be too large.
[0058] If the grinding quality is abnormal, check whether the model of the grinding equipment is compatible with the size of the silicon carbide wafer, and check whether the process parameters that are compatible with the size of the silicon carbide wafer are used for processing. If the model of the grinding equipment is wrong, replace the grinding equipment so that the processing performance of the grinding equipment can meet the processing requirements of the silicon carbide wafer. If the process parameters are wrong, read the process parameter template that matches the silicon carbide wafer from the database, and detect whether the process parameters have been changed by comparison. If so, adjust the changed process parameters to the parameters in the process parameter template.
[0059] The method for monitoring the warpage of silicon carbide wafers provided in this embodiment includes setting N sampling points on the surface of silicon carbide wafers and measuring the roughness of the N sampling points. The grinding quality is monitored according to the roughness. If the grinding quality is abnormal, the grinding equipment is replaced or the process parameters are adjusted. The roughness of the silicon carbide wafer surface is positively correlated with the warpage of the silicon carbide wafer. N sampling points are set on the surface of the silicon carbide wafer, N ≥ 2, and the roughness of the N sampling points can more comprehensively reflect the processing conditions of various positions on the surface of the silicon carbide wafer. If the roughness of a sampling point is large, it means that the local warpage of the silicon carbide wafer at this point is large; if the roughness of the N sampling points is large, it means that the overall warpage of the silicon carbide wafer is large. In the process of processing silicon carbide wafers, the roughness of N sampling points on the surface of the silicon carbide wafer is monitored. According to the roughness, it can be timely judged whether the grinding quality is abnormal. When the grinding quality is abnormal, the grinding equipment is replaced in hardware, or the process parameters are adjusted in software, which can avoid processing abnormalities in subsequent processing. The microscope has a high resolution and can reflect the roughness of the surface of the silicon carbide ultra-thin wafer. The roughness of the sampling point can be measured under the microscope to monitor the warpage of the silicon carbide ultra-thin wafer.
[0060] Example 2
[0061] like Figure 1 As shown, this embodiment provides a method for monitoring the warpage of a silicon carbide wafer, comprising:
[0062] S1: N sampling points are set on the surface of the silicon carbide wafer; wherein N ≥ 2.
[0063] S2: Measuring the roughness of the N sampling points, wherein the roughness is measured under a microscope.
[0064] S3: Monitoring the grinding quality according to the roughness, wherein the roughness is positively correlated with the warpage of the silicon carbide wafer.
[0065] S4: If the grinding quality is abnormal, replace the grinding equipment or adjust the process parameters.
[0066] In the production process, the fluctuation of the processing size of the product is inevitable. It is caused by the fluctuation of basic factors such as people, machines, materials, methods and environment. There are two types of fluctuations: normal fluctuations and abnormal fluctuations. Normal fluctuations are caused by accidental reasons, that is, inevitable factors. They have little impact on product quality and are difficult to eliminate technically. Abnormal fluctuations are caused by system reasons, that is, abnormal factors. They have a great impact on product quality, but measures can be taken to avoid and eliminate them. The purpose of process control is to eliminate abnormal fluctuations and keep the production process in a normal fluctuation state.
[0067] The roughness of N sampling points is collected, and the average value of the N roughness is calculated to obtain the average roughness. The standard deviation of the N roughness is calculated to obtain the roughness standard deviation. The average roughness is subtracted by 3 times the roughness standard deviation to obtain the lower limit of the numerical distribution. The average roughness is added by 3 times the roughness standard deviation to obtain the upper limit of the numerical distribution. The lower limit of the numerical distribution and the upper limit of the numerical distribution form a reference interval, and the reference interval can reflect the fluctuation state of the production process of the silicon carbide wafer.
[0068] like Figure 4 As shown, for example, under grinding condition 2, the average roughness is 9nm, the roughness standard deviation is 1nm, the lower limit of the numerical distribution is 6nm, the upper limit of the numerical distribution is 12nm, and the reference interval is 6nm-12nm. If there is a sampling point with a roughness greater than 12nm or a roughness less than 6nm under grinding condition 2, the roughness of the sampling point is judged as abnormal roughness, and the grinding quality of the silicon carbide wafer is abnormal; if the roughness of N sampling points under grinding condition 2 is within the reference interval, the roughness of the N sampling points is normal, and the grinding quality of the silicon carbide wafer is normal.
[0069] This embodiment monitors the grinding quality according to the roughness, including calculating the average roughness and roughness standard deviation of N roughnesses, and determining the reference interval by the average roughness and roughness standard deviation. Under the same grinding conditions, the roughness of the N detection points fluctuates only within a very small range. If the roughness of the N detection points fluctuates greatly, and there are detection points with a roughness greater than the upper limit of the numerical distribution of the reference interval, or a roughness less than the lower limit of the numerical distribution of the reference interval, it is determined that the grinding is abnormal, the warpage of the silicon carbide wafer exceeds the range, and the processing needs to be stopped to check the factors of the grinding abnormality.
[0070] Example 3
[0071] like Figure 1 As shown, this embodiment provides a method for monitoring the warpage of a silicon carbide wafer, characterized by comprising:
[0072] S1: N sampling points are set on the surface of the silicon carbide wafer; wherein N ≥ 2.
[0073] S2: Measuring the roughness of the N sampling points, wherein the roughness is measured under a microscope.
[0074] S3: Monitoring the grinding quality according to the roughness, wherein the roughness is positively correlated with the warpage of the silicon carbide wafer.
[0075] S4: If the grinding quality is abnormal, replace the grinding equipment or adjust the process parameters.
[0076] The replacement of the grinding equipment comprises the following steps S41-S42:
[0077] S41: Replace the current grinding wheel with a replacement grinding wheel.
[0078] S42: Replace the current silicon carbide wafer with a replacement silicon carbide wafer.
[0079] The process parameters are adjusted, including the following steps S43-S44:
[0080] S43: Increase or decrease the current grinding speed.
[0081] S44: Increase or decrease the current feed speed.
[0082] Step S41 and step S43 are in parallel. Step S42 is executed after executing step S41, and step S44 is executed after executing step S43.
[0083] When other process parameters are determined, the grinding quality is basically stable and the roughness fluctuates within a very small range. Figure 4 As shown, Grinding Condition 1 adopts a better grinding process and has a higher grinding quality, and Grinding Condition 1 is the condition used daily on the production line. When a sampling point with excessive roughness appears, for example, the roughness of the sampling point changes from 4-5nm to 12nm, it means that there is a problem with the grinding quality and the warping has reached 2nm. When the roughness is abnormal as described above, it is necessary to check the grinding equipment and process parameters. First, test whether the grinding wheel is normal, compare the roughness of the original grinding wheel with the roughness after replacing the new grinding wheel. If the roughness fluctuates over a large range, it means that the quality of the grinding wheel is abnormal and a new grinding wheel needs to be replaced. If the roughness fluctuates only within a very small range, that is, the roughness of the N sampling points are all within the reference range, it means that the quality of the grinding wheel is normal and there is no need to replace a new grinding wheel.
[0084] After confirming that the quality of the grinding wheel in the grinding equipment is normal, start to detect the silicon carbide wafer. Under the same conditions, only replace the new silicon carbide wafer. If the roughness fluctuates greatly, for example, the roughness of some of the N sampling points is greater than the upper limit of the numerical distribution of the reference interval, and the roughness of the other sampling points is within the reference interval, it means that the quality of the silicon carbide wafer is abnormal and a new silicon carbide wafer needs to be replaced. If the roughness of some of the N sampling points is less than the lower limit of the numerical distribution of the reference interval, it is necessary to measure the roughness of these sampling points again to check whether there is a calculation error. If the roughness fluctuates only within a very small range, that is, the roughness of the N sampling points is within the reference interval, it means that the quality of the silicon carbide wafer is normal, and there is no need to replace a new silicon carbide wafer. When it is determined that the quality of the grinding wheel and silicon carbide wafer in the grinding equipment is normal, start to adjust the process parameters, which include the grinding speed and the feed speed, and adjust the size of the grinding speed and the feed speed. If the adjusted process parameters can restore the roughness of the silicon carbide wafer surface to normal, it proves that the abnormality of the roughness is caused by the process parameters, and the process parameters need to be re-determined at this time. Read the process parameter template number in the database, match the process parameter template corresponding to the currently processed silicon carbide wafer, compare the current process parameters with the process parameter template, find the modification items, and adjust the modification items to the corresponding parameters in the process parameter template.
[0085] In this embodiment, when the grinding quality is abnormal, the quality of grinding equipment such as grinding wheels and silicon carbide wafers is first detected. If the quality of the grinding wheels and / or silicon carbide wafers is abnormal, the grinding quality abnormality is solved by replacing new grinding wheels and / or silicon carbide wafers; if the quality of the grinding wheels and silicon carbide wafers is normal, there is no need to replace new grinding wheels and silicon carbide wafers. At this time, it is detected whether the process parameters are abnormal. If so, the process parameters are further adjusted, such as adjusting the grinding speed and feed speed to solve the grinding quality abnormality.
[0086] Example 4
[0087] like Figure 1 As shown, this embodiment provides a method for monitoring the warpage of a silicon carbide wafer, comprising:
[0088] S1: N sampling points are set on the surface of the silicon carbide wafer; wherein N ≥ 2.
[0089] S2: Measuring the roughness of the N sampling points, wherein the roughness is measured under a microscope.
[0090] S3: Monitoring the grinding quality according to the roughness, wherein the roughness is positively correlated with the warpage of the silicon carbide wafer.
[0091] S4: If the grinding quality is abnormal, replace the grinding equipment or adjust the process parameters.
[0092] like Figure 2 As shown, the monitoring of grinding quality according to the roughness includes:
[0093] S31: Calculate N reference intervals of the roughness.
[0094] S32: Determine the roughness outside the reference interval as abnormal roughness.
[0095] S33: measuring the roughness of the sampling points corresponding to the abnormal roughness for multiple times.
[0096] S34: If the roughness of the sampling point is still abnormal after multiple measurements, it is determined that the grinding is abnormal, the processing is stopped, and the cause of the grinding abnormality is checked.
[0097] like Figure 3 As shown, step S31 includes the following steps S311-S315:
[0098] S311: Calculate the average value of the N roughnesses to obtain the average roughness.
[0099] S312: Calculate the standard deviations of the N roughnesses to obtain the roughness standard deviation.
[0100] S313: Subtract three times of the roughness standard deviation from the average roughness to obtain a lower limit of the value distribution.
[0101] S314: Add three times of the roughness standard deviation to the average roughness to obtain an upper limit of the value distribution.
[0102] S315: The lower limit of the numerical distribution and the upper limit of the numerical distribution form a reference interval.
[0103] like Figure 4 As shown in the figure, under grinding condition 2, if there is no sampling point with abnormal roughness, the roughness range of N sampling points is 7-10nm. Taking 5 sampling points as an example, the roughness of the 5 sampling points is 7.5nm, 7.8nm, 8nm, 9nm and 9.5nm respectively, then the average roughness is 8.36nm, and the standard deviation of roughness is 0.7605nm. Subtract 3 times the standard deviation of roughness from the average roughness to get the lower limit of the numerical distribution of 6.0785nm. Add 3 times the standard deviation of roughness to the average roughness to get the upper limit of the numerical distribution of 10.6415nm, and the reference interval is [6.0785,10.6415nm].
[0104] From the 1st roughness detection to the Nth roughness, if the qth roughness is greater than or equal to the lower limit of the numerical distribution, and less than or equal to the upper limit of the numerical distribution, the qth roughness is judged to be normal roughness; if the qth roughness is greater than the upper limit of the numerical distribution, the qth roughness is judged to be abnormal roughness, 1≤q≤N; if the qth roughness is less than the lower limit of the numerical distribution, the qth roughness may be calculated incorrectly, and the qth roughness is judged to be abnormal roughness, 1≤q≤N.
[0105] If the qth roughness is abnormal roughness, the roughness of the sampling point corresponding to the qth roughness is measured multiple times to verify whether the roughness of the sampling point is abnormal. If the roughness of the sampling point is still abnormal after multiple measurements, it is determined that the grinding is abnormal, the processing is stopped, and the factors of the grinding abnormality are checked. The factors of the grinding abnormality may be abnormal grinding equipment, abnormal process parameters and / or errors in the process of measuring roughness.
[0106] This embodiment monitors the grinding quality according to the roughness, including calculating N reference intervals of the roughness, determining the roughness outside the reference interval as abnormal roughness, and measuring the roughness of the sampling points corresponding to the abnormal roughness multiple times to verify whether the roughness of the sampling points is indeed abnormal, thereby improving the accuracy of detecting the sampling points with abnormal roughness. If the roughness of the sampling points is still abnormal after multiple measurements, the grinding is determined to be abnormal, the processing is stopped, and the factors of the grinding abnormality are checked.
[0107] Example 5
[0108] like Figure 1 As shown, this embodiment provides a method for monitoring the warpage of a silicon carbide wafer, comprising:
[0109] S1: N sampling points are set on the surface of the silicon carbide wafer; wherein N ≥ 2.
[0110] S2: Measuring the roughness of the N sampling points, wherein the roughness is measured under a microscope.
[0111] S3: Monitoring the grinding quality according to the roughness, wherein the roughness is positively correlated with the warpage of the silicon carbide wafer.
[0112] S4: If the grinding quality is abnormal, replace the grinding equipment or adjust the process parameters.
[0113] After determining that the grinding quality is abnormal, the method further includes:
[0114] The grinding quality is calculated according to the following formula:
[0115] Bow=b+a*Ra;
[0116] Among them, a is the first coefficient, b is the second coefficient, Bow is the grinding quality, and Ra is the roughness.
[0117] There is a linear relationship between grinding quality and roughness. The first coefficient and the second coefficient corresponding to different production lines are different. It is necessary to fit the first coefficient and the second coefficient according to the actual data on the production line. In this embodiment, the first coefficient is 0.1303 and the second coefficient is 0.1815 as an example.
[0118] The step of setting N sampling points on the surface of the silicon carbide wafer includes:
[0119] S11: Randomly select N sampling points on the surface of the silicon carbide wafer.
[0120] S12: or selecting a first straight line and a second straight line on the surface of the silicon carbide wafer, selecting K1 sampling points on the first straight line, and selecting K2 sampling points on the second straight line; wherein the first straight line intersects or is parallel to the second straight line, and N=K1+K2.
[0121] Step S11 and step S12 are parallel, and a sampling point distance threshold can be set. N sampling points are randomly selected on the surface of the silicon carbide wafer according to the sampling point interval to ensure that the distance between any two sampling points is greater than the sampling point distance threshold. A first straight line and a second straight line can also be selected on the surface of the silicon carbide wafer, and the first straight line and the second straight line intersect or are parallel. K1 sampling points are randomly selected on the first straight line, and K2 sampling points are randomly selected on the second straight line to ensure that the distance between any two sampling points is greater than the sampling point distance threshold.
[0122] The measuring the roughness of the N sampling points comprises:
[0123] S21: Select a sampling pattern, where the sampling pattern is a rectangle or a circle.
[0124] S22: Determine the area of the i-th sampling point according to the sampling pattern.
[0125] S23: Measure the roughness in the area of the i-th sampling point.
[0126] Step S22 includes the following steps S221-S222:
[0127] S221: If the sampling pattern is a rectangle, taking the ith sampling point as the center, determine the area of the ith sampling point according to the length and width.
[0128] S222: If the sampling pattern is a circle, taking the i-th sampling point as the center, determine the area of the i-th sampling point according to the radius.
[0129] like Figure 5As shown, when the sampling pattern is a rectangle, the length of the rectangle is 10um, the width is 10um, and the coordinates of the i-th sampling point are (100,50), then the coordinates of the four vertices of the corresponding sampling area are (90,60)(110,60)(90,40)(110,40). Figure 5 Here, 1 represents the first sampling point, 2 represents the second sampling point, 3 represents the third sampling point, 4 represents the fourth sampling point, and 5 represents the fifth sampling point.
[0130] When the sampling pattern is a circle, the radius of the circle is 6um, and the coordinates of the i-th sampling point are (100,50), then the corresponding sampling area is a circle center area with the sampling point coordinates (100,50) as the center and a radius of 6um.
[0131] This embodiment provides two sampling patterns and two ways of setting N sampling points. The sampling pattern can be rectangular or circular. When the sampling point coordinates are the same, different sampling patterns correspond to different measurement areas, and the measured roughness is different. N sampling points can be randomly selected on the surface of the silicon carbide wafer, or the first straight line and the second straight line can be selected on the surface of the silicon carbide wafer, and multiple sampling points can be selected on the first straight line and the second straight line respectively, and the total number of sampling points on the first straight line and the second straight line is N. The operator can select the appropriate sampling pattern and the way of setting N sampling points according to the processing needs to adapt to a variety of processing scenarios.
[0132] Example 6
[0133] like Figure 1 As shown, this embodiment provides a method for monitoring the warpage of a silicon carbide wafer, comprising:
[0134] S1: N sampling points are set on the surface of the silicon carbide wafer; wherein N ≥ 2.
[0135] S2: Measuring the roughness of the N sampling points, wherein the roughness is measured under a microscope.
[0136] S3: Monitoring the grinding quality according to the roughness, wherein the roughness is positively correlated with the warpage of the silicon carbide wafer.
[0137] S4: If the grinding quality is abnormal, replace the grinding equipment or adjust the process parameters.
[0138] The step of monitoring the grinding quality according to the roughness comprises:
[0139] S31: Detecting the SPC fluctuation value of the roughness.
[0140] S32: If the SPC fluctuation value is less than or equal to the fluctuation value upper limit, it is determined that the grinding quality is normal.
[0141] S33: If the SPC fluctuation value is greater than the fluctuation value upper limit, it is determined that the grinding quality is abnormal.
[0142] In the production process, the fluctuation of the processing size of the product is inevitable. It is caused by the fluctuation of basic factors such as people, machines, materials, methods and environment. There are two types of fluctuations: normal fluctuations and abnormal fluctuations. Normal fluctuations are caused by accidental reasons, that is, unavoidable factors. They have little impact on product quality and are difficult to eliminate technically. Abnormal fluctuations are caused by system reasons, that is, abnormal factors. They have a great impact on product quality, but measures can be taken to avoid and eliminate them. The purpose of process control is to eliminate abnormal fluctuations and keep the production process in a normal fluctuation state.
[0143] Statistical process control, referred to as SPC, is a process control tool that uses mathematical statistics. It analyzes and evaluates the production process, promptly discovers systematic factors based on feedback information, and takes measures to eliminate their influence, so that the production process is maintained in a controlled state affected only by random factors, in order to achieve the purpose of quality control. When the process is only affected by random factors, the process is in a state of statistical control; when there are systematic factors in the process, the process is in a state of statistical out-of-control. Since process fluctuations have statistical regularity, when the production process is controlled, the process characteristics generally obey a stable random distribution; when the production process is out of control, the process distribution will change. SPC uses the statistical regularity of process fluctuations to analyze and control the process. Therefore, it emphasizes that the production process is operated in a controlled and capable state so that the product meets customer requirements.
[0144] Collect the roughness of N sampling points and detect the SPC fluctuation values of the N roughness. The normal range of the SPC of the roughness of the silicon carbide wafer surface is 0.2-0.4, that is, the upper limit of the fluctuation value is 0.4. When the SPC fluctuation values of the N roughness are less than the upper limit of the fluctuation value, the grinding quality is judged to be normal, for example, the SPC fluctuation values of the N roughness are in the range of 0.32-0.38. When the SPC fluctuation value of the roughness is greater than the upper limit of the fluctuation value, the grinding quality is judged to be abnormal, for example, the SPC fluctuation value of the roughness of one or more sampling points is 0.5. By detecting the SPC fluctuation value of the roughness, the grinding process of the silicon carbide wafer can be quality monitored.
[0145] The SPC fluctuation value of the roughness is calculated by first calculating the lower limit and the upper limit of the value distribution according to steps S311-S314:
[0146] S311: Calculate the average value of the N roughnesses to obtain the average roughness.
[0147] S312: Calculate the standard deviations of the N roughnesses to obtain the roughness standard deviation.
[0148] S313: Subtract three times of the roughness standard deviation from the average roughness to obtain a lower limit of the value distribution.
[0149] S314: Add three times of the roughness standard deviation to the average roughness to obtain an upper limit of the value distribution.
[0150] After calculating the lower limit and the upper limit of the numerical distribution, the method further includes: subtracting the lower limit of the numerical distribution from the upper limit of the numerical distribution to obtain a numerical distribution difference, and then dividing the numerical distribution difference by 6 times the standard deviation to obtain an SPC fluctuation value of the roughness.
[0151] The replacement of the grinding equipment comprises the following steps S41-S42:
[0152] S41: Replace the current grinding wheel with a replacement grinding wheel.
[0153] S42: Replace the current silicon carbide wafer with a replacement silicon carbide wafer.
[0154] The process parameters are adjusted, including the following steps S43-S44:
[0155] S43: Increase or decrease the current grinding speed.
[0156] S44: Increase or decrease the current feed speed.
[0157] Step S41 and step S43 are in parallel. Step S42 is executed after executing step S41, and step S44 is executed after executing step S43.
[0158] Under the condition that other process conditions and parameters are determined, the grinding quality is basically stable, and the SPC fluctuation value of the roughness is within a normal range. This embodiment takes the normal range of 0.2-0.4 as an example. Figure 4 As shown, grinding condition 1 adopts a better grinding process, has a higher grinding quality, and grinding condition 1 is the condition used daily on the production line. When a jump point appears, for example, the SPC fluctuation value of the roughness of one or more sampling points is greater than 0.4, it means that there is a problem with the grinding quality and the warping of the silicon carbide wafer is large. When the roughness is abnormal as described above, it is necessary to check the grinding equipment and process parameters. First, test whether the grinding wheel is normal. By comparing the SPC fluctuation value of the roughness obtained using the original grinding wheel and the SPC fluctuation value of the roughness obtained after replacing the new grinding wheel, it can be determined whether the grinding wheel has quality abnormalities. For example, when the SPC fluctuation value of the roughness of one or more sampling points is greater than 0.4, such as 0.5, it is considered that the grinding wheel quality is abnormal and a new grinding wheel needs to be replaced. If the SPC fluctuation value of the roughness of all sampling points is within 0.2-0.4, it means that the quality of the grinding wheel is normal and there is no need to replace a new grinding wheel.
[0159] This embodiment monitors the grinding quality according to the roughness, including detecting the SPC fluctuation value of the roughness. If the SPC fluctuation value is less than or equal to the upper limit of the fluctuation value, the grinding quality is determined to be normal. If the SPC fluctuation value is greater than the upper limit of the fluctuation value, the grinding quality is determined to be abnormal. By detecting the SPC fluctuation value of the roughness, the grinding process of the silicon carbide wafer can be quality monitored. Under normal circumstances, the SPC fluctuation values of the roughness of N detection points are within the normal range. If the SPC fluctuation value of the roughness of one or more detection points is greater than the upper limit of the fluctuation value, the grinding is determined to be abnormal, the warping of the silicon carbide wafer is out of range, and the processing needs to be stopped to check for abnormalities.
[0160] Unless otherwise specifically stated, the relative arrangement, numerical expressions and numerical values of the parts and steps set forth in these embodiments do not limit the scope of the application. In all examples shown and discussed here, any specific value should be interpreted as being merely exemplary, rather than as a restriction. Therefore, other examples of exemplary embodiments can have different values. It should be noted that similar reference numerals and letters represent similar items in the following drawings, and therefore, once a certain item is defined in an accompanying drawing, it does not need to be further discussed in the subsequent accompanying drawings.
[0161] It should be understood that spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. For example, if the device in the figures is inverted, a device described as "above" or "above" other devices or structures would then be positioned "below" or "below" the other devices or structures. Thus, the exemplary term "above" can include both the orientations of "above" and "below." The device may also be positioned in other different ways (rotated 90 degrees or in other orientations), and the spatially relative descriptions used herein should be interpreted accordingly.
[0162] In addition, it should be noted that the use of terms such as "first" and "second" to limit components is only for the convenience of distinguishing the corresponding components. Unless otherwise stated, the above terms have no special meaning and therefore cannot be understood as limiting the scope of protection of this application.
[0163] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A method for monitoring warpage of a silicon carbide wafer, characterized in that: include: N sampling points are set on the surface of the silicon carbide wafer, where N ≥ 2; Measuring the roughness of N sampling points, wherein the roughness is measured under a microscope; monitoring the grinding quality according to the roughness, wherein the roughness is positively correlated with the warpage of the silicon carbide wafer; If the grinding quality is abnormal, replace the grinding equipment or adjust the process parameters.
2. The method for monitoring warpage of a silicon carbide wafer according to claim 1, characterized in that: The replacement of the grinding equipment comprises: Replace the current grinding wheel with an alternative grinding wheel; Replace current SiC wafers with replacement SiC wafers.
3. The method for monitoring warpage of a silicon carbide wafer according to claim 1, characterized in that: The adjusting process parameters comprises: Increase or decrease the current grinding speed; Increase or decrease the current feed rate.
4. The method for monitoring warpage of a silicon carbide wafer according to claim 1, characterized in that: The step of monitoring the grinding quality according to the roughness comprises: Calculating N reference intervals of the roughness; Determining the roughness outside the reference interval as abnormal roughness; Measuring the roughness of the sampling points corresponding to the abnormal roughness for multiple times; If the roughness of the sampling point is still abnormal after multiple measurements, it is determined that the grinding is abnormal, the processing is stopped, and the cause of the grinding abnormality is checked.
5. The method for monitoring warpage of a silicon carbide wafer according to claim 4, characterized in that: The calculating of N reference intervals of the roughness comprises: Calculating an average value of N roughness values to obtain an average roughness; Calculating the standard deviations of N roughnesses to obtain the roughness standard deviation; Subtract three times the standard deviation of the roughness from the average roughness to obtain the lower limit of the numerical distribution; Add three times the standard deviation of the roughness to the average roughness to obtain the upper limit of the value distribution; The lower limit of the numerical distribution and the upper limit of the numerical distribution constitute a reference interval.
6. The method for monitoring warpage of a silicon carbide wafer according to claim 1, characterized in that: The step of setting N sampling points on the surface of the silicon carbide wafer includes: Randomly select N sampling points on the surface of the silicon carbide wafer; Or select a first straight line and a second straight line on the surface of the silicon carbide wafer, select K1 sampling points on the first straight line, and select K2 sampling points on the second straight line; wherein the first straight line intersects or is parallel to the second straight line, and N=K1+K2.
7. The method for monitoring warpage of a silicon carbide wafer according to claim 1, characterized in that: The measuring the roughness of the N sampling points comprises: Select a sampling pattern, wherein the sampling pattern is a rectangle or a circle; Determine the area of the i-th sampling point according to the sampling graph; The roughness in the area of the i-th sampling point is measured.
8. The method for monitoring warpage of a silicon carbide wafer according to claim 7, characterized in that: The step of determining the area of the i-th sampling point according to the sampling graph comprises: If the sampling pattern is a rectangle, the area of the i-th sampling point is determined based on the length and width, taking the i-th sampling point as the center; If the sampling pattern is a circle, the area of the i-th sampling point is determined according to the radius with the i-th sampling point as the center.
