Method and apparatus for adjusting wafer polishing parameters

Abstract

The present disclosure relates to the technical field of wafer polishing, and in particular, to a method and device for adjusting wafer polishing parameters. The method for adjusting wafer polishing parameters comprises: determining thickness measurement data of a wafer at a target time point collected by a sensor; wherein the thickness measurement data comprises detection point positions on the wafer and corresponding thickness values; the target time point is a collection time point of the wafer at a target thickness; determining a detection trajectory based on the detection point positions, and establishing a detection coordinate system with a center point of the detection trajectory as an origin and a tangent direction as a horizontal axis; determining statistical characteristics of the thickness measurement data in the detection coordinate system, and determining a surface flatness of the wafer; in a case where it is determined based on an adjustment rule that the polishing parameters are to be adjusted, adjusting the polishing parameters according to the statistical characteristics; wherein the adjustment rule is determined according to the surface flatness.

Description

Technical Field

[0001] This disclosure relates to the field of wafer polishing technology, and in particular to a method and apparatus for adjusting wafer polishing parameters. Background Technology

[0002] Semiconductor wafers require polishing processes to achieve high flatness. Various parameters of the polishing equipment affect the quality of wafer polishing.

[0003] When wafer unevenness is identified, technicians adjust various parameters of the polishing equipment based on experience, and then fine-tune the parameters based on the flatness of the next batch of wafers to achieve a higher degree of flatness. However, this method relies on manual judgment, requires a high level of professional expertise from technicians, is difficult to standardize and implement, and has high labor costs. Summary of the Invention

[0004] To overcome the problems existing in the related technologies, this disclosure provides a method and apparatus for adjusting wafer polishing parameters, which can solve the above problems.

[0005] According to a first aspect of the present disclosure, a method for adjusting wafer polishing parameters is provided. The method includes: determining thickness measurement data of the wafer collected by a sensor; determining thickness measurement data of the wafer collected by the sensor at a target time; wherein the thickness measurement data includes detection points of the sensor on the wafer and corresponding thickness values; the target time is the acquisition time of the wafer at a target thickness; determining a detection trajectory based on the detection points, and establishing a detection coordinate system with the center point of the detection trajectory as the origin and the tangent direction as the horizontal axis; determining statistical characteristics of the thickness measurement data in the detection coordinate system, and determining the surface flatness of the wafer; and adjusting the polishing parameters according to the statistical characteristics when an adjustment rule is determined based on the adjustment rules; wherein the adjustment rule is determined based on the surface flatness.

[0006] According to a second aspect of the present disclosure, a device for adjusting wafer polishing parameters is provided. The device includes: a thickness measuring unit configured to determine thickness data of a wafer collected by a sensor at a target time; wherein the thickness measuring data includes detection points of the sensor on the wafer and corresponding thickness values; the target time is the acquisition time of the wafer at a target thickness; a coordinate establishment unit configured to determine a detection trajectory based on the detection points and establish a detection coordinate system with the center point of the detection trajectory as the origin and the tangent direction as the horizontal axis; a determination unit configured to determine statistical characteristics of the thickness measuring data in the detection coordinate system and determine the surface flatness of the wafer; and an adjustment unit configured to adjust the polishing parameters according to the statistical characteristics when the polishing parameters are determined to be adjusted based on an adjustment rule; wherein the adjustment rule is determined based on the surface flatness.

[0007] According to a third aspect of the present disclosure, an electronic device is provided, comprising: a processor and a memory; the memory being used to store a computer program; and the processor being used to execute, by invoking the computer program, a method for adjusting wafer polishing parameters as described in the first aspect.

[0008] According to a fourth aspect of the present disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the method as described in the first aspect.

[0009] The technical solutions provided by the embodiments of this disclosure may include the following beneficial effects: After acquiring the thickness measurement data of the wafer, this disclosure establishes a detection coordinate system based on the detection trajectory at the target time, and determines the statistical characteristics of the thickness measurement data in the detection coordinate system and the surface flatness of the wafer. Based on preset adjustment rules, it determines whether polishing parameters need to be adjusted using the statistical characteristics and surface flatness. If polishing parameters need to be adjusted, they are then adjusted according to the statistical characteristics. After establishing the detection coordinate system based on the detection trajectory, the surface flatness determined in the detection coordinate system can be used to characterize the flatness of the wafer. Furthermore, the statistical characteristics of the thickness measurement data are related to the parameters of the polishing equipment. Different parameters can result in different manifestations of the statistical characteristics of the thickness measurement data. Therefore, based on the statistical characteristics, the relevant polishing parameters can be adjusted more accurately, reducing reliance on manual experience, improving the accuracy and objectivity of parameter adjustment, increasing efficiency, and achieving higher quality and stability.

[0010] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0011] The accompanying drawings, which are incorporated in and form part of this disclosure, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0012] Figure 1 This is a schematic flowchart illustrating a method for adjusting wafer polishing parameters according to an exemplary embodiment of the present disclosure.

[0013] Figure 2 This is a schematic diagram illustrating a detection trajectory and a detection coordinate system according to an exemplary embodiment of the present disclosure.

[0014] Figure 3 This is a schematic flowchart illustrating a face type classification according to an exemplary embodiment of the present disclosure.

[0015] Figure 4 This disclosure is a schematic diagram illustrating the training of a large model for determining polishing parameters according to an exemplary embodiment.

[0016] Figure 5 This is a block diagram illustrating an apparatus for adjusting wafer polishing parameters according to an exemplary embodiment of the present disclosure.

[0017] Figure 6 This is a schematic diagram of the structure of a computer device according to an exemplary embodiment of the present disclosure. Detailed Implementation

[0018] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0019] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “a,” “the,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0020] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0021] To address the aforementioned technical problems, this disclosure proposes a method for adjusting wafer polishing parameters.

[0022] Figure 1 This is a schematic flowchart illustrating a method for adjusting wafer polishing parameters according to an embodiment of the present disclosure. This method can be used to adjust the parameters of a wafer polishing equipment, thereby polishing wafers that meet standards.

[0023] like Figure 1 As shown, the methods for adjusting wafer polishing parameters include: In step S101, the thickness measurement data of the wafer collected by the sensor at the target time is determined; wherein, the thickness measurement data includes the detection points of the sensor on the wafer and the corresponding thickness values; the target time is the acquisition time of the wafer at the target thickness. In step S102, a detection trajectory is determined based on the detection point, and a detection coordinate system is established with the center point of the detection trajectory as the origin and the tangent direction as the horizontal axis. In step S103, the statistical characteristics of the thickness measurement data in the detection coordinate system are determined, and the surface flatness of the wafer is determined. In step S104, if the polishing parameters are to be adjusted based on the adjustment rules, the polishing parameters are adjusted according to the statistical characteristics; wherein the adjustment rules are determined based on the surface flatness.

[0024] In some embodiments, the current polishing parameters of the polishing equipment can be obtained.

[0025] The polishing parameter adjustment method proposed in this disclosure is based on adjusting the polishing parameters of the current processing batch, thereby determining the accurate polishing parameters for the next batch (run to run), thereby improving the surface quality and yield of the wafer processing results.

[0026] Therefore, the current polishing parameters of the polishing equipment can be obtained, and subsequent parameter adjustments can be made based on the current polishing parameters by adjusting the amount.

[0027] Polishing parameters can cover key process parameters in the polishing process, such as, but not limited to, the rotation speed of the upper and lower polishing discs, the applied pressure, the polishing disc temperature, the polishing slurry temperature and flow rate, the pH value and conductivity of the polishing slurry, the cooling water flow rate, and the wafer thickness information.

[0028] In some embodiments, the thickness data of the wafer acquired by the sensor at a target time is determined.

[0029] The sensor can be a laser thickness sensor, which is installed on the upper plate of the polishing machine. It can scan and detect the wafer in the cavity in real time during the polishing process to determine the thickness data.

[0030] The sensor can acquire and record the spatial coordinates (X, Y) of each detection point and the corresponding thickness value (T), thereby forming thickness measurement data, which is high-density wafer three-dimensional morphology data.

[0031] Based on real-time wafer thickness measurement data, thickness features correlated with the polishing parameters of the polishing equipment can be further extracted. These features may include the thickness before polishing, the thickness during the main / rough polishing stage jump, and the wafer thickness after polishing. These thickness features can be used to analyze the impact of different polishing processes on wafer surface quality, thereby adjusting the relationship between wafer surface quality and polishing parameters.

[0032] Throughout the polishing process, the sensor can traverse every detection point on the wafer, thereby acquiring thickness measurement data for the entire wafer. However, this thickness measurement data is substantial, and data processing is complex. Therefore, it is necessary to determine the thickness measurement data of the sensor at a specific target time.

[0033] The target time can be either the intermediate time or the end time.

[0034] The intermediate moment can be the moment when the wafer thickness is reduced by a preset thickness (e.g., 5 micrometers) relative to the thickness before polishing. The thickness measurement data at this moment can initially reflect the wafer's surface profile and is related to some polishing parameters. Based on the analysis of the thickness measurement data at the intermediate moment, the adjustment strategy for some polishing parameters related to the surface profile at the intermediate moment can be determined.

[0035] The end time can be the moment before the polishing process ends. At the end time, the sensor trajectory can scan the thickness measurement data within a first length range (e.g., within a radius of 10 mm) of the wafer center. This part of the thickness measurement data can be used to evaluate the processing quality of the entire polishing process and is related to another part of the polishing parameters.

[0036] In some embodiments, a detection trajectory is determined based on the detection points, and a detection coordinate system is established with the center point of the detection trajectory as the origin and the tangent direction as the horizontal axis.

[0037] In the device coordinate system, since the sensor's detection trajectory may be located in different areas on the wafer, the detection points of the thickness measurement data belong to different intervals. In order to extract the common data features of wafer thickness measurement data at multiple target times, this embodiment uses the detection trajectory itself as a reference frame to transform the coordinate system of the thickness measurement data. By taking the center point of the detection trajectory as the origin, the tangent direction of the detection trajectory at the center point as the horizontal axis, and the normal direction as the vertical axis, a detection coordinate system is constructed. This achieves data standardization, reduces the interference of scanning position differences on data statistics, and improves the accuracy of statistical feature extraction.

[0038] Figure 2 This is a schematic diagram illustrating a detection trajectory and a detection coordinate system according to an embodiment of the present disclosure.

[0039] like Figure 2 As shown, the left figure is a schematic diagram of the sensor's detection trajectory (thickness measurement scan trajectory) at the target time in the original coordinate system. The sensor can usually detect thickness measurement data at points on and near the detection trajectory. It can be seen that in the coordinate system of the left figure, it is difficult to analyze and determine the characteristics of the thickness measurement data in terms of coordinate distribution.

[0040] Based on embodiments of this disclosure, the center point of the detection trajectory can be taken as the origin, the tangent direction as the horizontal axis, and the normal direction as the vertical axis, constructing a system as follows: Figure 2 The right-hand figure shows the detection coordinate system (X', Y'), which transforms the coordinates of the thickness measurement data determined based on the detection trajectory from the original spatial coordinates to the coordinates in the detection coordinate system. Under this detection coordinate system, the thickness measurement data, after simpler processing, can exhibit its symmetry and other statistical characteristics based on the coordinates in the detection coordinate system.

[0041] In some embodiments, the coordinates of the thickness measurement data are transformed based on the detection coordinate system, and the horizontal axis coordinates of the thickness measurement data in the detection coordinate system are mapped to the wafer diameter range.

[0042] Since there can be multiple detection trajectories at the target time, the corresponding thickness measurement data may be distributed in multiple ranges on the horizontal axis. To facilitate subsequent data processing, the data can be standardized and normalized.

[0043] The horizontal axis coordinates of thickness measurement data in the detection coordinate system can be mapped to the diameter range of the wafer through a linear mapping of relative spacing. For example, for a wafer with a diameter of 300 mm, the horizontal axis coordinate range of the thickness measurement data in the detection coordinate system is [-143, 143], which can be mapped to [-150, 150]. Combined with... Figure 2 In this way, the thickness measurement data of the detection trajectory on the wafer can all be distributed within the same horizontal coordinate range.

[0044] In some embodiments, the statistical characteristics of the thickness measurement data in the detection coordinate system are determined, and the surface flatness of the wafer is determined.

[0045] The statistical characteristics of thickness measurement data in the detection coordinate system can be determined. These characteristics may include, but are not limited to, the average thickness, maximum and minimum thickness, and the distribution characteristics of the thickness measurement data. Based on these statistical characteristics, the distribution characteristics of the wafer thickness in the detection coordinate system can be determined, which in turn allows for the determination of wafer profile, surface flatness, and other indicators. These indicators can be used to measure the polishing quality of the wafer, determining whether it meets standards and whether parameters need adjustment. Furthermore, the polishing parameters of the polishing equipment also affect these indicators, creating a correlation between them. Therefore, identifying these indicators can help determine adjustment strategies for the polishing parameters.

[0046] The surface flatness (Peak to Valley, PV value) of a wafer can also be determined based on thickness measurement data. Surface flatness measures whether the wafer surface is flat; the smaller the surface flatness, the flatter the wafer. For example, a preset threshold can be used to compare with the surface flatness to measure whether the wafer's flatness meets the standard.

[0047] In some embodiments, when the polishing parameters are determined to be adjusted based on adjustment rules, the polishing parameters are adjusted according to the statistical characteristics; wherein the adjustment rules are determined based on the surface flatness.

[0048] The need for adjusting polishing parameters is determined based on adjustment rules. Surface flatness reflects the flatness of the wafer, therefore, adjustment rules can be determined at least based on surface flatness.

[0049] If it is determined that the polishing parameters need to be adjusted, they can be further adjusted based on statistical characteristics. The statistical characteristics in this disclosure are determined based on thickness measurement data in a detection coordinate system, making these statistical characteristics related to the polishing parameters. Therefore, after determining the correlation between the polishing parameters and the statistical characteristics, the polishing parameters that need to be adjusted and the amount of adjustment can be determined based on the determined statistical characteristics, thereby realizing the adjustment of the polishing parameters.

[0050] This disclosure establishes a detection coordinate system for the detection trajectory at the target time and determines the statistical characteristics of the thickness measurement data within this coordinate system. Since polishing parameters affect the wafer's polishing effect, and the wafer's polishing effect can be reflected by the thickness measurement data, the statistical characteristics of the thickness measurement data within the detection coordinate system can more accurately reflect the correlation between statistical characteristics and polishing parameters. Therefore, based on these statistical characteristics, the polishing parameters can be adjusted more accurately. This disclosure also determines the wafer's surface flatness. Based on the wafer's surface flatness and adjustment rules, it is possible to more accurately determine whether polishing parameters need adjustment, ensuring polishing quality and reducing unnecessary parameter adjustments.

[0051] In some embodiments, the method further includes: transforming the thickness measurement data from the original coordinate system to the detection coordinate system.

[0052] In some embodiments, transforming the thickness measurement data from the original coordinate system to the detection coordinate system includes: determining the rotation angle between the coordinate system of the detection point and the detection coordinate system; transforming the coordinates of the detection point based on the rotation angle to determine the detection coordinates of the detection point in the detection coordinate system; and determining the statistical characteristics of the thickness measurement data in the detection coordinate system includes: determining the statistical characteristics based on the detection coordinates and the corresponding thickness value.

[0053] The rotation angle between the horizontal axis of the detection coordinate system and the horizontal axis of the original coordinate system can be determined. θ .

[0054] For example, for any spatial point (X, Y) in the original coordinate system, its coordinates (X', Y') in the detection coordinate system are calculated through translation and rotation transformations. The transformation formula is as follows:

[0055]

[0056] in,( , ) represents the spatial coordinates of the center point of the detected trajectory in the original coordinate system.

[0057] Based on this transformation formula, coordinate transformation is performed on the thickness measurement data.

[0058] After coordinate transformation, the horizontal axis of the detection coordinate represents the distance from the detection point to the center of the wafer. The larger the absolute value of the horizontal axis coordinate, the farther the detection point is from the center of the wafer, and the closer it is to the edge of the wafer. Therefore, based on the detection coordinate, the thickness values ​​of the wafer in the edge and center regions can be more intuitively represented, thereby determining statistical characteristics.

[0059] In some embodiments, the abscissa of the detection coordinate can be mapped to the abscissa range corresponding to the diameter of the wafer.

[0060] Since the sensor's detection points are not necessarily distributed along the diameter of the wafer, the range of values ​​for the horizontal coordinate of the detection coordinate may be smaller than the wafer's diameter. This can cause problems for subsequent data processing and hinder the unified organization of data. Therefore, the horizontal coordinate of the detection coordinate can be mapped to normalize and standardize it.

[0061] The diameter range of the horizontal axis [-R, R] can be determined based on the wafer diameter D, where R is half of the diameter D.

[0062] The minimum value of the abscissa of the thickness measurement data in the detection coordinate system after statistical transformation can be obtained. and maximum value And calculate the data span .

[0063] Then, a linear transformation is used to map the abscissa of the thickness measurement data in the detection coordinate system to the standardized interval:

[0064] in, The x-axis is mapped to the thickness measurement data within the diameter range.

[0065] In some embodiments, edge region data with an absolute value of the horizontal coordinate greater than a certain threshold can be removed.

[0066] Due to edge effect interference, data near the edge of the wafer is distorted. Therefore, thickness measurement data from edge regions where the absolute value of the mapped x-coordinate exceeds a certain threshold can be discarded.

[0067] For example, if the wafer diameter is 300mm and the x-axis ranges from [-150, 150], then thickness measurement data outside the range of 148mm from the origin can be discarded, meaning that thickness measurement data within the range of [-148, 148] can be retained. The following examples can be further processed based on the thickness measurement data after edge trimming. It should be noted that further processing can also be performed on thickness measurement data without edge trimming; this disclosure does not impose any limitations on this.

[0068] In some embodiments, the statistical features include at least one of the following: the surface type classification of the wafer, the edge classification of the wafer; determining the statistical features of the thickness measurement data in the detection coordinate system includes: determining the fitting curve corresponding to the thickness measurement data; and determining the surface type classification and / or edge classification of the wafer based on the fitting curve.

[0069] Statistical characteristics and wafer surface flatness can be determined based on the transformed coordinates. Statistical characteristics may include, but are not limited to: wafer surface type classification and wafer edge classification.

[0070] Wafer profile classification is used to characterize the overall morphology of the wafer, and different profile classifications require different polishing parameters to be adjusted. For example, a wafer profile classified as convex (n-type) may be related to the pressure between the upper and lower polishing pads and the temperature inside the polishing pads. For instance, when the pressure on the upper polishing pad is high, the convex shape of the wafer will be reduced. Furthermore, if there is a significant temperature difference between the inside and outside of the polishing pad, due to the influence of thermal expansion and contraction temperature gradients, the wafer will bulge in the higher-temperature areas and sink in the lower-temperature areas.

[0071] Therefore, after determining the wafer's surface type classification, the polishing parameters related to the cause of the surface type classification can be determined, and these polishing parameters can be adjusted to correct the wafer's surface type.

[0072] Edge classification of a wafer is used to characterize the morphological features of its edges. Polishing parameters also affect the wafer edges, resulting in various edge classifications. Therefore, determining the edge classification of a wafer is helpful in accurately identifying the polishing parameters that need adjustment, thus allowing for more precise adjustment of these parameters.

[0073] It should be noted that surface classification can be determined based on the corresponding fitted curve, and edge classification can also be determined based on the corresponding fitted curve.

[0074] In some embodiments, a fitting curve corresponding to the thickness measurement data can be determined.

[0075] Fitted curves can separate signals from noise and quantify trend characteristics.

[0076] For example, an nth-order (e.g., n can be 6) polynomial model can be used. Least-squares fitting is performed on the thickness measurement data to obtain an approximate fitting curve characterizing the wafer surface shape. Based on this fitting curve, the wafer surface shape classification can be determined.

[0077] Of course, the fitted curve can also include the approximate fitted curve of the wafer edge, which will not be elaborated on here, but will be introduced in detail later.

[0078] In some embodiments, determining the surface flatness of the wafer includes: determining the surface flatness of the wafer based on a fitted curve characterizing the wafer surface profile.

[0079] Based on the fitted curve y(x), the effective range of the thickness measurement data after edge removal can be determined. Calculate the peak-to-valley value, also known as the surface smoothness (PV value):

[0080] In some embodiments, the statistical features may also include the average thickness of the fitted curve.

[0081] After determining the fitted curve, the average thickness of the fitted curve can be further determined. Statistical characteristics may also include the average thickness, which can be used to determine the adjustment amount for polishing parameters. The integral formula is:

[0082] in, This represents the average thickness.

[0083] In some embodiments, determining the wafer's surface type classification based on the fitted curve includes: determining the maxima and minima of the fitted curve; performing noise reduction processing on adjacent maxima and minima; and determining the wafer's surface type classification based on the number of denoised maxima and minima. Specifically, if the fitted curve has one maxima, the surface type is classified as convex; if the fitted curve has one minima, the surface type is classified as concave; if the number of maxima in the fitted curve is greater than the number of minima, the surface type is classified as multi-point convex; if the number of maxima in the fitted curve is less than the number of minima, the surface type is classified as multi-point concave; and if the number of maxima in the fitted curve is equal to the number of minima, the surface type is classified as asymmetrical.

[0084] Based on the fitted curve, the maximum and minimum values ​​of the fitted curve can be determined, and noise reduction processing can be performed on the maximum and minimum values. Then, the number of maximum and minimum values ​​in the fitted curve can be counted, and the surface type classification can be determined based on the number of maximum and minimum values.

[0085] When the first derivative of the fitted curve is 0 and the second derivative is less than 0, the fitted curve has a maximum value, which satisfies the following:

[0086] The number of maxima in the fitted curve is denoted as . .

[0087] When the first derivative of the fitted curve is 0 and the second derivative is greater than 0, the fitted curve has a local minimum, which satisfies the following:

[0088] The number of local minima in the fitted curve is denoted as . .

[0089] However, some adjacent maxima and minima have small variation ranges and small spans on the horizontal axis. These adjacent maxima and minima do not actually affect the judgment of the entire wafer surface. Therefore, these maxima and minima can be removed during data processing by noise reduction.

[0090] In some embodiments, the noise reduction processing of adjacent maxima and minima includes: determining the thickness difference and radius difference between any extreme point and another adjacent extreme point; if the thickness difference is less than a longitudinal threshold or the radius difference is less than a transverse threshold, discarding the extreme point and merging adjacent extreme points of the same type; wherein the longitudinal threshold is determined based on the surface flatness of the wafer, and the transverse threshold is determined based on the diameter of the wafer.

[0091] The vertical threshold can be determined based on the PV value, and the horizontal threshold can be determined based on the thickness measurement data within the effective range of the horizontal axis.

[0092] Vertical threshold can be The horizontal threshold can be .in This is the longitudinal morphology discrimination factor, which can be set to, for example, 0.5; This is the lateral morphology discrimination factor, which can be set to, for example, 0.15-0.2.

[0093] If the thickness difference between adjacent extreme points is less than the vertical threshold, it indicates that the extreme point has relatively small fluctuations in the thickness dimension; if the difference in the horizontal coordinates (radius difference) between adjacent extreme points is less than the horizontal threshold, it indicates that the extreme points are relatively close in the horizontal coordinate dimension. These two types of extreme points have little impact on the morphology and can be ignored.

[0094] After ignoring some extreme points, there may be two adjacent maximum points or two adjacent minimum points. In this case, adjacent extreme points of the same type (e.g., both being maximum points) can be merged, and the wafer surface type can be classified according to the number of maximum and minimum points.

[0095] In some embodiments, the wafer's surface type is determined based on the number of the maxima and minima after noise reduction.

[0096] After removing adjacent extreme points with minimal variation, the wafer surface type can be classified into convex (n-type), concave (u-type), multi-point convex (m-type), multi-point concave (w-type), and asymmetric type based on the distribution and overall trend of extreme points in the fitted curve.

[0097] If the fitted curve has a maximum value but no minimum value, the surface type is classified as type n. If the fitted curve has a minimum value but no maximum value, the surface shape is classified as U-shaped.

[0098] In some embodiments, the criteria for determining a multi-point convex shape are: the number of maxima is greater than the number of minima, and the absolute thickness difference between adjacent maxima and minima is greater than a preset longitudinal threshold, and the absolute radius difference between adjacent maxima and minima is greater than a transverse threshold.

[0099] The criteria for judgment can be: Furthermore, the absolute thickness difference between adjacent maxima and minima is greater than the longitudinal threshold. The absolute radius difference between adjacent maxima and minima is greater than the horizontal threshold. .

[0100] Alternatively, one can first identify the extreme points where the absolute thickness difference is greater than the longitudinal threshold and the absolute radius difference is greater than the transverse threshold, and then determine the number of maxima and minima among these extreme points. This method can first screen out some maxima and minima that do not meet the specifications, and then determine the surface shape based on the screened extreme points.

[0101] In some embodiments, the criteria for determining multi-point concaveness are: the number of maxima is less than the number of minima, and the absolute thickness difference between adjacent maxima and minima is greater than a preset longitudinal threshold, and the absolute radius difference between adjacent maxima and minima is greater than a transverse threshold.

[0102] In some embodiments, the number of maxima is equal to the number of minima ( In the case of ), the face shape is classified as asymmetrical.

[0103] Figure 3 This is a schematic flowchart illustrating a face shape classification process according to an embodiment of the present disclosure.

[0104] like Figure 3 As shown, the specific methods of face shape classification can be demonstrated by combining the above embodiments.

[0105] Determine the number of extreme points. If the number of extreme points is 1, determine whether the extreme point is a maximum or a minimum. If it is a maximum, the surface type is classified as n-type; if it is a minimum, the surface type is classified as u-type.

[0106] If the number of extreme points is not 1, determine whether the number of extreme points is greater than or equal to 3. If not, that is, the number of extreme points is 2, it is impossible for there to be only two maximum points or only two minimum points. There must be one maximum point and one minimum point, which means it is an asymmetric type.

[0107] If the number of extreme points is greater than or equal to 3, further compare the number of maximum points and the number of minimum points: if there are many maximum points, and the difference in thickness and radius between adjacent extreme points is greater than the corresponding threshold, then it is an m-type; if there are few maximum points, and the difference in thickness and radius between adjacent extreme points is greater than the corresponding threshold, then it is a w-type; if the number of maximum points is equal to the number of minimum points, then it is an asymmetric type.

[0108] In some embodiments, polishing parameters are adjusted according to the determined wafer surface type classification.

[0109] Different surface type classifications correspond to different types of polishing parameters. Based on the determined surface type classification, the polishing parameters that need to be adjusted can be determined. Then, based on statistical characteristics (which may include surface type classification), the adjustment amount of these polishing parameters that need to be adjusted can be determined to achieve the adjustment of polishing parameters.

[0110] In some embodiments, the wafer profile can be determined as flat or tilted based on the wafer's surface flatness.

[0111] Flat or tilted types are used to determine whether the polishing parameters of the wafer need to be adjusted, and the adjustment rules can be determined based on the flat or tilted type standards.

[0112] In some embodiments, the adjustment rule includes at least one of the following: determining to adjust the polishing parameters when the surface flatness is not less than a preset threshold; determining to adjust the polishing parameters when the thickness difference between the two edges of the detection trajectory is greater than a tilt threshold based on the statistical characteristics; wherein the tilt threshold is determined based on the surface flatness.

[0113] The surface flatness of the wafer determines whether the wafer's shape is flat or tilted.

[0114] The thickness difference between the two edges of the thickness measurement data can be the longitudinal difference of the fitted curve at the left and right edge points, and the tilt threshold can be the product of a preset tilt morphology factor and the PV value. Under the condition that... In this case, the surface shape can be defined as inclined, and the classification of inclined surface shapes requires adjustment of polishing parameters. Among these, This represents the longitudinal difference between the left and right edge points of the fitted curve. This is a preset tilt morphology judgment factor, for example, it can be 0.5.

[0115] In satisfying In this case, the surface shape can be defined as flat, where For the flatness specification line, when the target time is the middle time, the flatness specification line can be 300-500 micrometers; when the target time is the end time, the flatness specification line can be determined according to the actual production needs, for example, it can be 200 micrometers.

[0116] If the surface is not flat, the polishing parameters need to be adjusted. The preset threshold can be... This means that the flatness is half of the flatness specification line. When the surface flatness is not less than the preset threshold, it indicates that the surface undulation of the wafer is large and the polishing parameters need to be adjusted.

[0117] Therefore, based on this embodiment, when the surface shape is determined to be tilted, and / or when the wafer PV value is large and does not belong to the flat type, the polishing parameters need to be adjusted. The adjustment of the polishing parameters needs to be determined based on statistical characteristics (such as surface shape classification, edge classification, PV value, etc.).

[0118] In some embodiments, determining to adjust the polishing parameters includes: based on the changing trend of surface flatness of wafer polishing in multiple consecutive batches, predicting that the next batch will meet the adjustment rules, and then determining to adjust the polishing parameters.

[0119] When performing a series of wafer polishing processes, the trend of surface flatness variation can be determined, and the surface flatness of the next batch of wafers can be predicted based on this trend.

[0120] If the surface flatness of the next batch is predicted to meet the adjustment rules (tilt or excessive PV value), the polishing parameters can be adjusted and corrected in advance before the actual polishing of the next batch to ensure that the surface flatness of the wafer is within a certain range and to ensure the quality of polishing multiple batches of wafers.

[0121] In addition to classifying the wafer's surface shape as described above, statistical features can also include edge classification. Edge classification and surface shape classification have different relationships with polishing parameters; therefore, determining the wafer's edge classification can also help to more accurately adjust polishing parameters.

[0122] In some embodiments, determining the edge classification of the wafer based on the fitted curve includes: determining the mean range of thickness measurement data within sliding windows on both sides of the detection trajectory, and determining an edge threshold based on the mean range; determining the measured mean of the thickness measurement data located at the edges on both sides of the detection trajectory, and the fitted mean of the fitted curve located at the edges on both sides; determining the edge difference between the measured mean and the fitted mean on both sides respectively; comparing the edge difference with the edge threshold, and determining the edge of the wafer based on the comparison result on both sides.

[0123] It should be noted that the fitting curve required to determine edge classification may not be the same fitting curve used to determine face shape classification.

[0124] The thickness sequence T of the thickness measurement data can be normalized by taking the maximum and minimum values ​​for normalization. After normalization, the data is then fitted with an nth-order (e.g., 6th-order) polynomial to obtain the fitted curve. The expression for the fitted curve here can be found in the expression for the fitted curve corresponding to face shape classification in the above embodiments.

[0125] Thickness data located at both edges (left and right edges) can be determined. Measured mean and and the fitted curves located on both sides. Fitted mean and And determine the margin difference between the measured mean and the fitted mean on both sides respectively.

[0126] The difference between the left and right edges is The difference on the right edge is .

[0127] Then, compare the left edge difference with the edge threshold and the right edge difference with the edge threshold to determine the edge type on both sides, and then determine the edge of the wafer based on the comparison results on both sides.

[0128] The edge threshold can be a preset threshold, for example, based on practical experience, the edge threshold can be set to 0.2.

[0129] Because the collected thickness measurement data exhibits certain fluctuations, there may be fluctuation errors in the edge threshold. Furthermore, in some embodiments, the edge threshold can be determined based on the mean range of the thickness measurement data at the edges on both sides of the detection trajectory.

[0130] Some thickness measurement data can be extracted using a sliding window method. The wafer thickness data can be extracted separately using the sliding window feature extraction method. The sliding range is calculated for a certain range (e.g., one-sixth of the total data) on both the left and right sides. The sliding range is the difference between the maximum and minimum values ​​within the sliding window. The sliding window is moved across this range to obtain multiple sets of sliding ranges. After determining these multiple sets of sliding ranges, the mean of these sliding extreme values ​​is determined. For example, the length of the sliding window can be 10% of the data length, and the sliding step size can be 1. Based on this method, the mean of the left-side range can be determined. and the mean of the right range .

[0131] The edge threshold is corrected based on the determined mean of the left and right ranges. The left threshold is corrected as follows: The threshold on the right is corrected to .

[0132] In some embodiments, the edge difference is compared with the edge threshold, and the edge of the wafer is determined based on the comparison results on both sides; wherein, if the absolute value of the edge difference on both sides is less than the edge threshold, the edge is classified as a flat edge; if the edge difference on both sides is greater than the edge threshold, the edge is classified as a warped edge; if the edge difference on both sides is less than the opposite of the edge threshold, the edge type is a collapsed edge; if the edge types determined by the comparison results on both sides are different, the edge type is an asymmetrical edge.

[0133] The specific judgment rules are as follows: , is a flat edge; The edge is curled up; This is considered a collapsed edge. Based on this determination rule, the edge difference on the left side is compared with the left threshold, and the edge difference on the right side is compared with the right threshold to determine the edge type on both sides. If the edge types on both sides are the same, the edge type of the wafer is the determined edge type; if the edge types on both sides are different, the edge type of the wafer is an asymmetrical edge.

[0134] In some embodiments, if the edge difference on either side is greater than the upper limit threshold, the edge type on that side is determined to be either a large collapsed edge or a large warped edge.

[0135] When the edge difference If the value is too large, it can be classified as a large collapsed edge or a large warped edge according to the judgment rules. The upper limit threshold is one such threshold. It can be preset, for example, it can be 0.4. The upper limit threshold can also be modified based on the mean of the left and right ranges, which will not be elaborated here.

[0136] In some embodiments, the polishing parameters can be adjusted based on the determined edge classification.

[0137] In some embodiments, edge classification can be combined with surface classification, and then the polishing parameters can be adjusted.

[0138] If there are anomalies in the edge classification results, such as warped or collapsed edges, they can be combined with the surface classification, and the corresponding polishing parameters can be determined based on the combination.

[0139] For example, even though both are warped edges, the polishing parameters associated with U-shaped and N-shaped warped edges differ.

[0140] In some embodiments, the number of wafers can be multiple.

[0141] A single polishing process can polish multiple wafers in a batch (e.g., 15 or 25). However, even when polishing with the same polishing parameters in the same batch, some wafers may still have different polishing results.

[0142] In this case, the number and proportion of surface types of multiple wafers can be statistically analyzed to determine the number and proportion of each type of abnormal surface type and abnormal edge. Then, based on the combination of the majority of abnormal surface types and abnormal edges, the polishing parameters can be adjusted.

[0143] In some embodiments, adjusting the polishing parameters based on the statistical characteristics includes: determining the correlation between the wafer's surface type classification and / or the wafer's edge classification and the polishing parameters; determining the corresponding polishing parameter type based on the determined wafer's surface type classification and / or edge classification; and determining the adjustment amount for the polishing parameter type based on the surface flatness and the thickness measurement data.

[0144] After determining the anomaly type of the wafer under the current polishing parameters (which may include surface type classification and edge classification), the correlation between the anomaly type and the polishing parameters can be determined. Then, the polishing parameter type can be determined based on the anomaly type, and the specific adjustment amount for the polishing parameter type can be determined by combining the PV value and thickness measurement data.

[0145] For example, based on the predetermined correlation between the average wafer thickness and the average pressure of the upper polishing pad, and combined with the average thickness in the thickness measurement data, the amount of adjustment required for the average pressure of the upper polishing pad to achieve the target average thickness can be determined.

[0146] Figure 4 This is a schematic diagram illustrating the training of a large model for determining polishing parameters according to an embodiment of the present disclosure. Figure 4 The following example illustrates the training process of the large model.

[0147] In some embodiments, the adjusted polishing parameters can be determined using a large model.

[0148] The large-scale model requires a built-in knowledge base for process adjustment based on expert experience. This knowledge base clearly defines the correlation between various surface anomalies (such as U-shaped warping) and polishing parameters (such as pressure and temperature). This knowledge base can provide a theoretical basis for preliminary diagnosis and parameter localization. Within this knowledge base, the mapping relationship between wafer morphology characteristics and polishing parameter adjustment directions is stored. Based on this mapping relationship, when a specific morphology anomaly occurs on the wafer, the large-scale model can indicate one or more polishing parameters that need to be adjusted first, along with the direction of adjustment (increase or decrease, and the corresponding adjustment amount).

[0149] For example, when it is diagnosed that the wafer processing has an n-type morphology and the PV value is greater than the specification line, it can indicate that it is necessary to focus on inspection and adjust the disk temperature and polishing pressure; when it is diagnosed that the wafer has edge warping defects, it can indicate that it is necessary to focus on inspection and adjust the step thickness setting value.

[0150] In some embodiments, large models can be built by mining and analyzing historical processing data.

[0151] By mining and analyzing massive amounts of historical processing data, nonlinear multivariate regression prediction models can be constructed to achieve intelligent mapping from wafer morphology diagnostic results to optimal process parameters.

[0152] Input the feature vector X into the large model. The feature vector X can include full-dimensional data of the wafers in the current batch. X can be represented as:

[0153] Where P_current represents the complete set of polishing parameters for the current batch (e.g., pressure, disk temperature, polishing fluid flow rate, rotation speed, etc.); T_metrics represents the set of thickness measurement metrics, including surface flatness PV and average thickness. C_shape and C_edge are encoded classification variables, corresponding to surface shape classification (e.g., n-shaped, u-shaped) and edge classification (e.g., flat edge, warped edge, collapsed edge), respectively; e_edge is a continuous quantized value of the degree of collapse or warping of the edge.

[0154] The target polishing parameter Y is predicted based on the input feature vector X. This target polishing parameter is the set value of the polishing parameter for the next batch that has been verified as successfully optimized by the model output.

[0155]

[0156] Where P_next is the target polishing parameter vector corresponding to the P_current dimension, representing the polishing parameters that should be used in the next batch after adjustment.

[0157] In some embodiments, a learning model based on a nonlinear decision tree framework can be used as the base model.

[0158] The XGBoost (eXtreme Gradient Boosting) ensemble learning model, based on a nonlinear decision tree framework, can be used as the base model for regression prediction. This model uses an additive training strategy to sequentially construct a series of sub-decision trees, each of which acts as a weak learner, to progressively correct the prediction residuals from the previous round.

[0159] Predicted output The decision is determined by the outputs of K decision trees, and the expression is:

[0160] in, Represents the k-th decision tree, It is the function space consisting of all possible decision trees.

[0161] The goal of model training is to minimize the regularization objective function. :

[0162] in, It is a measure of recommendation value Compared with actual value The loss function of the error between them (e.g., mean squared error, MSE); This is a regularization term used to control model complexity and prevent overfitting. T is the number of leaf nodes in the tree. The weight score of the leaf node. and This is a hyperparameter.

[0163] To enhance the model's learning of key morphological diagnostic features, a feature weighting mechanism based on domain knowledge can be introduced. This mechanism can assign importance weights to different features in the loss function. To achieve this.

[0164] In some embodiments, higher weights can be assigned to the features of face shape classification C_shape and edge classification C_edge. and .

[0165] For example, when constructing a decision tree for feature splitting, a weighted information gain can be calculated. For a splitting scheme containing feature j, the weighted information gain is expressed as:

[0166] Where G_L and G_R are the sum of the first gradients (first derivatives of the loss function) of the left and right child node samples, respectively; H_L and H_R are the sum of the second gradients (second derivatives of the loss function) of the left and right child node samples, respectively; λ_j is the domain knowledge weight assigned to feature j. When j is an input face shape classification or edge classification feature, λ_j is set to a value greater than 1, while for other features λ_j is set to 1.

[0167] In some embodiments, adjusting the polishing parameters according to the statistical features includes: inputting the wafer's surface shape classification, the wafer's edge classification, and the surface flatness into a trained parameter adjustment model; adjusting the polishing parameters according to the output of the parameter adjustment model; wherein the parameter adjustment model is trained through weighted learning, and the weights of the surface shape classification and the edge classification in the weighted learning are greater than the weights of other features.

[0168] Based on the weighting mechanism in this embodiment, when searching for the best splitting features, the large model can tend to select those "face shape classification" features and "edge classification" features that are considered more important by the domain knowledge base, thereby learning the complex nonlinear mapping relationship between face shape anomalies and polishing parameter adjustment more accurately, thus generating a highly customized parameter recommendation large model with clear physical meaning.

[0169] Corresponding to the embodiments of the wafer polishing parameter adjustment method of this disclosure, this disclosure also provides embodiments of a corresponding wafer polishing parameter adjustment apparatus.

[0170] Please see Figure 5 , Figure 5 This is a block diagram of a wafer polishing parameter adjustment device according to one embodiment of this disclosure. Figure 5 As shown, the device for adjusting wafer polishing parameters includes: Thickness measurement unit 510 is configured to determine the thickness measurement data of the wafer collected by the sensor at a target time; wherein, the thickness measurement data includes the detection point of the sensor on the wafer and the corresponding thickness value; the target time is the acquisition time of the wafer at a target thickness; The system establishment unit 520 is configured to determine the detection trajectory based on the detection point and establish a detection coordinate system with the center point of the detection trajectory as the origin and the tangent direction as the horizontal axis. The determining unit 530 is configured to determine the statistical characteristics of the thickness measurement data in the detection coordinate system and to determine the surface flatness of the wafer; The adjustment unit 540 is configured to adjust the polishing parameters according to the statistical characteristics when the adjustment of the polishing parameters is determined based on the adjustment rules; wherein the adjustment rules are determined based on the surface flatness.

[0171] In some embodiments, the adjustment rule includes at least one of the following: determining to adjust the polishing parameters when the surface flatness is not less than a preset threshold; determining to adjust the polishing parameters when the thickness difference between the two edges of the detection trajectory is greater than a tilt threshold based on the statistical characteristics; wherein the tilt threshold is determined based on the surface flatness.

[0172] In some embodiments, determining to adjust the polishing parameters includes: based on the changing trend of surface flatness of wafer polishing in multiple consecutive batches, predicting that the next batch will meet the adjustment rules, and then determining to adjust the polishing parameters.

[0173] In some embodiments, the device is further configured to: determine the rotation angle between the coordinate system of the detection point and the detection coordinate system; transform the coordinates of the detection point based on the rotation angle to determine the detection coordinates of the detection point in the detection coordinate system; and determine the statistical characteristics of the thickness measurement data in the detection coordinate system, including: determining the statistical characteristics based on the detection coordinates and the corresponding thickness value.

[0174] In some embodiments, the apparatus is further configured to map the abscissa of the detection coordinates to the abscissa range corresponding to the diameter of the wafer.

[0175] In some embodiments, the statistical features include at least one of the following: the surface type classification of the wafer, the edge classification of the wafer; determining the statistical features of the thickness measurement data in the detection coordinate system includes: determining the fitting curve corresponding to the thickness measurement data; and determining the surface type classification and / or edge classification of the wafer based on the fitting curve.

[0176] In some embodiments, determining the wafer's surface type classification based on the fitted curve includes: determining the maxima and minima of the fitted curve; performing noise reduction processing on adjacent maxima and minima; and determining the wafer's surface type classification based on the number of denoised maxima and minima. Specifically, if the fitted curve has only one maxima, the surface type is classified as convex; if the fitted curve has only one minima, the surface type is classified as concave; if the number of maxima in the fitted curve is greater than the number of minima, the surface type is classified as multi-point convex; if the number of maxima in the fitted curve is less than the number of minima, the surface type is classified as multi-point concave; and if the number of maxima in the fitted curve is equal to the number of minima, the surface type is classified as asymmetrical.

[0177] In some embodiments, the noise reduction processing of adjacent maxima and minima includes: determining the thickness difference and radius difference between any extreme point and another adjacent extreme point; if the thickness difference is less than a longitudinal threshold or the radius difference is less than a transverse threshold, discarding the extreme point and merging adjacent extreme points of the same type; wherein the longitudinal threshold is determined based on the surface flatness of the wafer, and the transverse threshold is determined based on the diameter of the wafer.

[0178] In some embodiments, determining the edge classification of the wafer based on the fitted curve includes: determining the mean range of thickness measurement data within sliding windows on both sides of the detection trajectory, and determining an edge threshold based on the mean range; determining the measured mean of the thickness measurement data located at the edges on both sides of the detection trajectory, and the fitted mean of the fitted curve located at the edges on both sides; determining the edge difference between the measured mean and the fitted mean on both sides respectively; comparing the edge difference with the edge threshold, and determining the edge of the wafer based on the comparison results on both sides; wherein, if the absolute value of the edge difference on both sides is less than the edge threshold, the edge is classified as a flat edge; if the edge difference on both sides is greater than the edge threshold, the edge is classified as a warped edge; if the edge difference on both sides is less than the negative number of the edge threshold, the edge type is a collapsed edge; if the edge types determined by the comparison results on both sides are different, the edge type is an asymmetrical edge.

[0179] In some embodiments, adjusting the polishing parameters based on the statistical characteristics includes: determining the correlation between the wafer's surface type classification and / or the wafer's edge classification and the polishing parameters; determining the corresponding polishing parameter type based on the determined wafer's surface type classification and / or edge classification; and determining the adjustment amount for the polishing parameter type based on the surface flatness and the thickness measurement data.

[0180] In some embodiments, adjusting the polishing parameters according to the statistical features includes: inputting the wafer's surface shape classification, the wafer's edge classification, and the surface flatness into a trained parameter adjustment model; adjusting the polishing parameters according to the output of the parameter adjustment model; wherein the parameter adjustment model is trained through weighted learning, and the weights of the surface shape classification and the edge classification in the weighted learning are greater than the weights of other features.

[0181] The specific implementation process of the functions and roles of each unit in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0182] Embodiments of this disclosure also provide an electronic device, including: a processor and a memory; the memory for storing a computer program; and the processor for executing a wafer polishing parameter adjustment method as described in any of the above embodiments by invoking the computer program.

[0183] Embodiments of this disclosure also provide a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method for adjusting wafer polishing parameters as described in any of the above embodiments.

[0184] Embodiments of this disclosure also provide a computer program product, including a computer program that, when executed by a processor, implements the methods described in any of the foregoing embodiments.

[0185] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 6 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system).

[0186] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0187] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0188] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device as shown by a landing page for an app. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0189] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0190] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0191] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0192] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

[0193] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0194] The methods and apparatus provided in the embodiments of this disclosure have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this disclosure. The descriptions of the embodiments above are only for the purpose of helping to understand the methods and core ideas of this disclosure. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this disclosure. Therefore, the content of this specification should not be construed as a limitation of this disclosure.

Claims

1. A method for adjusting wafer polishing parameters, characterized in that, The method includes: Determine the thickness measurement data of the wafer collected by the sensor at a target time; wherein, the thickness measurement data includes the detection points of the sensor on the wafer and the corresponding thickness values; the target time is the acquisition time of the wafer at the target thickness; The detection trajectory is determined based on the detection points, and a detection coordinate system is established with the center point of the detection trajectory as the origin and the tangent direction as the horizontal axis. Determine the statistical characteristics of the thickness measurement data in the detection coordinate system, and determine the surface flatness of the wafer; When the polishing parameters are adjusted based on the adjustment rules, the polishing parameters are adjusted according to the statistical characteristics; wherein the adjustment rules are determined based on the surface flatness.

2. The method according to claim 1, characterized in that, The adjustment rules include at least one of the following: If the surface flatness is not less than a preset threshold, it is determined that the polishing parameters should be adjusted. If, based on the statistical characteristics, the difference in thickness between the two edges of the detection trajectory is greater than a tilt threshold, it is determined that the polishing parameters should be adjusted; wherein, the tilt threshold is determined based on the surface flatness.

3. The method according to claim 2, characterized in that, The step of determining to adjust the polishing parameters includes: Based on the changing trend of surface flatness of multiple consecutive batches of wafer polishing, and predicting that the next batch will meet the adjustment rules, the polishing parameters will be adjusted accordingly.

4. The method according to claim 1, characterized in that, The method further includes: Determine the rotation angle between the coordinate system of the detection point and the detection coordinate system; The coordinates of the detection point are transformed based on the rotation angle to determine the detection coordinates of the detection point in the detection coordinate system. Determining the statistical characteristics of the thickness measurement data in the detection coordinate system includes: The statistical features are determined based on the detection coordinates and the corresponding thickness values.

5. The method according to claim 4, characterized in that, The method further includes: The horizontal coordinate of the detection coordinate is mapped to the horizontal coordinate range corresponding to the diameter of the wafer.

6. The method according to claim 1, characterized in that, The statistical features include at least one of the following: the surface type classification of the wafer, the edge classification of the wafer; Determining the statistical characteristics of the thickness measurement data in the detection coordinate system includes: Determine the fitting curve corresponding to the thickness measurement data; The surface type and / or edge type of the wafer are determined based on the fitted curve.

7. The method according to claim 6, characterized in that, The step of determining the surface type classification of the wafer based on the fitted curve includes: Determine the maximum and minimum values ​​of the fitted curve; Noise reduction is performed on adjacent maxima and minima; The surface type classification of the wafer is determined based on the number of maxima and minima after noise reduction; wherein, If the fitted curve has only one maximum value, the surface shape is classified as convex. When the fitted curve has only one minimum value, the surface shape is classified as concave. When the number of maxima of the fitted curve is greater than the number of minima, the surface shape is classified as a multi-point convex shape. When the number of maxima of the fitted curve is less than the number of minima, the surface shape is classified as a multi-point concave shape. When the number of maxima of the fitted curve is equal to the number of minima, the surface type is classified as asymmetric.

8. The method according to claim 7, characterized in that, The noise reduction process for adjacent maxima and minima includes: Determine the thickness difference and radius difference between any extreme point and its adjacent extreme point; If the thickness difference is less than the longitudinal threshold, or the radius difference is less than the transverse threshold, then either extreme point is discarded, and adjacent extreme points of the same type are merged; wherein, The longitudinal threshold is determined based on the surface flatness of the wafer, and the transverse threshold is determined based on the diameter of the wafer.

9. The method according to claim 6, characterized in that, Determining the edge classification of the wafer based on the fitted curve includes: Determine the mean range of thickness measurement data within the sliding window on both sides of the detection trajectory, and determine the edge threshold based on the mean range; Determine the measured mean of the thickness data located at both edges of the detection trajectory, and the fitted mean of the fitted curve located at both edges; Determine the margin difference between the measured mean and the fitted mean on both sides respectively; The edge difference is compared with the edge threshold, and the edge of the wafer is determined based on the comparison results on both sides; wherein, If the absolute value of the difference between the two edges is less than the edge threshold, the edge is classified as a flat edge. If the difference between the two edges is greater than the edge threshold, the edge is classified as a warped edge; When the difference between the two edges is less than the negative of the edge threshold, the edge type is a collapsed edge; If the edge types determined by the comparison results on both sides are different, the edge type is an asymmetric edge.

10. The method according to claim 6, characterized in that, The step of adjusting the polishing parameters based on the statistical characteristics includes: Determine the correlation between the wafer's surface type classification and / or the wafer's edge classification and the polishing parameters; The corresponding polishing parameter type is determined based on the identified surface type and / or edge type of the wafer; The adjustment amount for the polishing parameter type is determined based on the surface flatness and the thickness measurement data.

11. The method according to claim 6, characterized in that, The step of adjusting the polishing parameters based on the statistical characteristics includes: The surface shape classification, edge classification, and surface flatness of the wafer are input into the trained parameter adjustment model; The polishing parameters are adjusted based on the output of the parameter adjustment model; wherein the parameter adjustment model is trained through weighted learning, and the weights of the face shape classification and the edge classification in the weighted learning are greater than the weights of other features.

12. A device for adjusting wafer polishing parameters, characterized in that, The device includes: The thickness measurement unit is configured to determine the thickness measurement data of the wafer collected by the sensor at a target time; wherein, the thickness measurement data includes the detection points of the sensor on the wafer and the corresponding thickness values; the target time is the acquisition time of the wafer at a target thickness; The system establishment unit is configured to determine the detection trajectory based on the detection points, and to establish a detection coordinate system with the center point of the detection trajectory as the origin and the tangent direction as the horizontal axis. The determining unit is configured to determine the statistical characteristics of the thickness measurement data in the detection coordinate system and to determine the surface flatness of the wafer; An adjustment unit is configured to adjust the polishing parameters according to the statistical characteristics when the adjustment is determined based on an adjustment rule; wherein the adjustment rule is determined based on the surface flatness.

Images