A substrate double-sided processing method, a processing device and a substrate

Abstract

The application discloses a substrate double-sided processing method and device and a substrate, relates to the technical field of semiconductor substrate processing, and specifically comprises the following steps: obtaining an initial disc type spacing difference of a processing disc; obtaining original data of a substrate, wherein the original data at least includes a surface type of the substrate and a maximum total thickness deviation in the substrate; the number of the substrate is greater than 1; based on the original data of the substrate and the initial disc type spacing difference of the processing disc, a target trimmer is matched from a pre-defined rule library; the processing disc is trimmed according to the target trimmer; the substrate is placed in a carrier of the processing disc, and the substrate is processed, so that the grinding amount of each position of the substrate in the processing process is uniform, and the TTV value of the substrate after double-sided processing is effectively reduced.

Description

Technical Field

[0001] This application relates to the field of semiconductor substrate processing technology, and in particular to a method, apparatus and substrate for double-sided substrate processing. Background Technology

[0002] The semiconductor industry is a strategic, fundamental, and pioneering industry that supports the development of modern information society. Semiconductor substrates are the core basic carriers for the fabrication of various semiconductor devices, equivalent to the foundation of semiconductor devices. The surface flatness and thickness uniformity of the substrate directly determine the crystal quality and defect density of the semiconductor functional layers subsequently grown on the substrate surface, thus affecting the performance, production yield, and long-term reliability of the final semiconductor device.

[0003] Total Thickness Variation (TTV) is a core indicator for measuring the thickness uniformity of semiconductor substrates. It refers to the thickness difference between the thickest and thinnest points on the same substrate. The smaller the TTV value, the more uniform the substrate thickness, the easier the subsequent processing, and the better the device consistency. With the continuous improvement in the performance, yield, and reliability requirements of semiconductor devices, effectively reducing the TTV of semiconductor substrates has become a core technical problem that urgently needs to be solved in the current semiconductor substrate processing field. Currently, the mainstream double-sided grinding / polishing processes in the industry generally adopt fixed dressing parameters and uniform processing modes, which are difficult to adapt to the different surface shapes and thickness deviations of different substrates to be processed. This results in unstable TTV control of the finished substrate, failing to meet the mass production requirements of semiconductor devices. Summary of the Invention

[0004] In view of the above problems, this application provides a method, apparatus, and substrate for double-sided substrate processing to reduce the TTV value of the substrate. The specific solution is as follows: In a first aspect, embodiments of this application provide a method for double-sided substrate processing, including: Obtain the initial disc spacing difference of the processing discs; Obtain raw data of the substrate, the raw data including at least the surface profile of the substrate and the maximum total thickness deviation in the substrate; the number of substrates is greater than 1; Based on the original data of the substrate and the initial disk spacing difference of the processing disk, a target trimmer is obtained by matching from a predefined rule base; the predefined rule base contains multiple trimmers, and each trimmer has a corresponding relationship with the surface shape of the substrate, the maximum value of the total thickness deviation of the substrate, and the initial disk spacing difference of the processing disk. The processing disc is dressed according to the target dressing device; The substrate is placed in the carrier of the processing tray and processed.

[0005] Optionally, the processing disk includes an annular effective processing area, and the disk spacing difference is a first height minus a second height, where the first height is the distance between the upper and lower disks at the edge of the effective processing area, and the second height is the spacing between the upper and lower disks at the outer edge of the effective processing area.

[0006] Optionally, placing the substrate within the carrier of the processing tray includes: Obtain the center point thickness of the entire batch of substrates, and sort all substrates in descending order of the center point thickness value; Select the top N substrates, where N is an integer greater than 1, and the number of carriers of the processing disk is an integer multiple of N; The selected N substrates are placed in a uniform rotationally symmetrical position that completely overlaps after rotating 360° / N with the geometric center of the carrier of the processing disk as the rotation center, so that the N points with the largest center point thickness values ​​are evenly distributed in the processing disk. The remaining substrate is placed in the remaining carrier of the processing tray.

[0007] Optionally, placing the remaining substrate within the remaining carrier of the processing tray includes: The remaining substrates are sorted according to a preset rule, and the top M substrates are selected from the remaining substrates, where M is an integer equal to N. The selected M-shaped substrates are placed in the remaining uniform rotationally symmetrical empty spaces that completely overlap after rotating 360° / M with the geometric center of the carrier of the processing disk as the rotation center. Repeat the above steps until all the carriers of the processing trays are filled.

[0008] Optionally, the processing of the substrate includes: The substrate is first processed using a first pressure value; When the preset switching conditions are met, the substrate is then switched to the second pressure value for processing. The second pressure value is greater than the first pressure value.

[0009] Optionally, the first pressure value is 10 to 15 times the first value; Wherein, the first value is the difference between the maximum and minimum values ​​of the center point thickness in the first type of substrate; the first type of substrate consists of N substrates placed in a uniformly rotationally symmetrical position with the center point thickness values ​​ranked first.

[0010] Optionally, the preset switching conditions include: the spindle torque fluctuation value of the processing equipment is stable within 1% for a continuous preset time, and the substrate thickness reaches a preset value.

[0011] Optionally, the surface profile of the substrate includes: an overall inclined surface profile and a center-high, edge-low surface profile.

[0012] Secondly, embodiments of this application provide a substrate double-sided processing apparatus, comprising: The processing tray status acquisition module is used to obtain the initial tray spacing difference of the processing tray; A substrate parameter acquisition module is used to acquire raw data of the substrate, the raw data including at least the surface profile of the substrate and the maximum total thickness deviation in the substrate; the number of substrates is greater than 1; The parameter matching module is used to match the target trimmer from a predefined rule base based on the original data of the substrate and the initial disk spacing difference of the processing disk; the predefined rule base contains multiple trimmers, and each trimmer has a corresponding relationship with the surface shape of the substrate, the maximum value of the total thickness deviation of the substrate, and the initial disk spacing difference of the processing disk. The dressing execution module is used to dress the processing disc according to the target dressing device; A fabric sheet and processing module are used to place the substrate in the carrier of the processing tray and process the substrate.

[0013] Thirdly, embodiments of this application provide a substrate prepared by the substrate double-sided processing method described in the first aspect of this application.

[0014] This application simultaneously obtains the initial disk spacing difference of the processing disk and the original data of the substrate. Based on the above data, a target trimmer is obtained by matching from a predefined rule base. The target trimmer is used to trim the disk so that the disk shape of the trimmed processing disk complements the surface features of the entire batch of substrates. This ensures that the grinding amount at each position of the substrate is uniform during the processing, alleviating the problem of insufficient or excessive grinding caused by the mismatch between the disk shape and the substrate surface in traditional fixed disk processing. This effectively reduces the TTV value of the substrate after double-sided processing. Attached Figure Description

[0015] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0016] Figure 1 This is a top view schematic diagram of a processing disc transmission system provided in an embodiment of this application; Figure 2 This is an exploded view of the side structure of a processing disc provided in an embodiment of this application; Figure 3A schematic diagram of the central cross-sectional structure of a processing disk provided in an embodiment of this application; Figure 4 This is a schematic flowchart of a substrate double-sided processing method provided in an embodiment of this application; Figure 5 A schematic cross-sectional view of the principle for measuring the difference in the spacing between processing discs provided in this application embodiment; Figure 6 A schematic diagram illustrating the dressing principle of a convex dressing block provided in an embodiment of this application; Figure 7 A schematic diagram illustrating the principle of a tilting trimming block trimming plate provided in an embodiment of this application; Figure 8 A three-dimensional topographic cloud image of the original surface of a substrate provided in this application embodiment; Figure 9 The three-dimensional topography cloud image of the substrate surface after conventional processing provided in Comparative Example 1 for the embodiments of this application; Figure 10 The three-dimensional topography cloud image of the substrate surface after processing by the substrate double-sided processing method of this application is provided for the embodiments of this application; Figure 11 This is a structural block diagram of a substrate double-sided processing apparatus provided in an embodiment of this application. Detailed Implementation

[0017] The embodiments of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0018] Various modifications and variations can be made to this application without departing from its spirit or scope, which will be apparent to those skilled in the art. Therefore, this application is intended to cover modifications and variations falling within the scope of the claimed technical solutions and their equivalents. It should be noted that the implementation methods provided in the embodiments of this application can be combined with each other without contradiction.

[0019] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] As described in the background section, how to effectively reduce the TTV value of semiconductor substrates has become an urgent problem to be solved in the current semiconductor substrate processing field.

[0021] In actual industrial production, the mainstream processing technologies for semiconductor substrates are mainly divided into two categories: one is wire cutting-double-sided processing, and the other is wire cutting-thinning-single-sided processing. Among them, the double-sided processing technology can simultaneously grind both the front and back surfaces of the substrate, which can not only effectively ensure the surface symmetry of the two sides of the substrate and obtain better overall surface accuracy, but also has the advantages of high processing efficiency and suitability for large-scale mass production. Therefore, it has become the preferred mainstream process for controlling the TTV of semiconductor substrates in the industry.

[0022] However, traditional double-sided processing technology still faces many core technical bottlenecks that are difficult to resolve effectively, severely restricting further improvements in semiconductor substrate processing quality and reductions in production costs. Because semiconductor crystals generally possess high hardness and brittleness, the substrates obtained through wire cutting inevitably exhibit surface inhomogeneity. This is mainly manifested in an overall tilted substrate thickness or a significantly thicker central region than the edge regions. Furthermore, the degree and type of surface differences between different batches of substrates also vary considerably.

[0023] To address the aforementioned issues, traditional double-sided machining processes typically employ the same dressing tool (dresser) and uniform dressing parameters to dress the core component (machining disc) used for grinding the substrate surface in double-sided machining equipment. This uniform dressing method limits the range of disc shape adjustments that can be made, making it impossible to customize a disc shape that complements the specific surface characteristics of each batch of substrates. This results in the high-thickness points on the substrate not being preferentially ground, leaving residual original surface defects on the finished substrate. Furthermore, traditional processes often place multiple substrates haphazardly on the grinding disc carrier. This haphazard placement tends to concentrate the high-thickness points of the entire batch of substrates in a specific area of ​​the grinding disc, causing that area to wear much faster than other areas. This leads to uneven disc shape drift, which in turn further exacerbates the surface deterioration of subsequent substrates, creating a vicious cycle that results in the finished substrate retaining its original defects.

[0024] The combined effect of these problems not only makes it difficult to effectively reduce the TTV value of the substrate after processing, but also increases the processing time and material loss of subsequent polishing processes, while affecting the growth quality of the epitaxial layer, ultimately resulting in low production yield and high production cost of semiconductor devices.

[0025] In view of this, this application simultaneously collects the initial disc spacing difference of the processing disc and the original data of the substrate, matches the most suitable dresser and performs targeted disc dressing, so that the disc shape of the dressed processing disc complements the surface shape of the substrate, ensuring that the grinding amount at each position of the substrate is uniform from the source of the process, thereby reducing the TTV value of the substrate.

[0026] It should be noted that the double-sided processing method of this application is compatible with various semiconductor substrate materials. It can be applied not only to silicon carbide substrates, but also to various commonly used semiconductor substrates such as silicon substrates, gallium nitride substrates, sapphire substrates, and gallium arsenide substrates. At the same time, this application does not limit the size of the substrate and can be adapted to batch double-sided processing scenarios of different mainstream substrate sizes such as 4 inches, 6 inches, 8 inches, and 12 inches. The overall process has strong versatility and adaptability. The process parameters can be flexibly adjusted according to different substrate materials, sizes, and processing requirements, and the processing effect of effectively reducing TTV can be achieved.

[0027] To facilitate understanding of the technical solutions in the embodiments of this application, the structure, working principle, and conventional dressing techniques of the processing disk, the core equipment used in double-sided substrate processing, will be described in detail first. The processing disk is the core execution component of double-sided grinding or polishing equipment; its structural precision, disk shape, and movement directly determine the flatness, TTV, and surface quality of the substrate after processing. The processing disk structure involved in this application can be achieved through… Figures 1 to 3 Full display, Figure 1 This is a top view schematic diagram of a processing disc transmission system provided in an embodiment of this application; Figure 2 This is an exploded view of the side structure of a processing disc provided in an embodiment of this application; Figure 3 This is a schematic diagram of the central cross-sectional structure of a processing disk provided in an embodiment of this application. It should be noted that the same numerical labels in different views represent the same component, and should be understood in conjunction with the corresponding views.

[0028] like Figure 1 As shown, the transmission system of the processing disk adopts a planetary gear structure, mainly composed of three parts: an outer gear ring 11, an inner gear ring 12, and planetary gears 13. The outer gear ring 11 is the outermost ring gear, usually fixed to the equipment frame or driven by an independent motor, providing the power basis for the revolution of the planetary gears 13. The inner gear ring 12 is a ring gear located at the center of the processing disk, driven by another independent motor, and can achieve different speeds and directions of rotation than the outer gear ring 11. The planetary gears 13 are gear-shaped disks evenly distributed between the outer gear ring 11 and the inner gear ring 12. In this embodiment, there are six planetary gears 13, each with multiple evenly distributed circular substrate placement holes 131 for fixing the semiconductor substrate to be processed. It should be noted that the number of planetary gears and substrate placement holes here is only an example; the number of planetary gears and substrate placement holes will vary for different types of processing disks, and this application does not limit this number.

[0029] During the operation of the processing disk, the external teeth of planetary gear 13 simultaneously mesh with the internal teeth of external gear ring 11 and the external teeth of internal gear ring 12. When external gear ring 11 and internal gear ring 12 rotate at different speeds, planetary gear 13 generates two motions simultaneously: one is a revolution around the center of the processing disk, and the other is a rotation around its own center. This composite motion is called planetary motion. The core advantage of this planetary motion is that it can drive the substrate to perform complex and uniform planar motion on the surface of the processing disk, so that every point on the substrate surface can traverse all areas of the processing disk, thereby ensuring that the grinding amount at each position of the substrate is uniform and consistent, avoiding the problem of local over-grinding or under-grinding. It should be noted that the annular area covered by all planetary gear motion trajectories is called the effective processing area. Only within this area can the substrate simultaneously contact the upper and lower disks and be effectively ground.

[0030] like Figure 2 and Figure 3 As shown, the main structure of the processing disk consists of an upper disk 21 and a lower disk 22 arranged opposite each other. The upper disk 21 is located above the processing system and is usually made of high-strength cast iron or stainless steel, possessing extremely high rigidity and initial flatness. The upper disk can move up and down vertically to apply stable pressure to the substrate during processing, and can also be driven by a motor to rotate in the opposite direction to the lower disk. The lower disk 22 is located below the processing system, made of the same material as the upper disk 21, and is driven to rotate by the main motor. It is the basic component supporting the planetary gear 13 and the substrate. A polishing pad is attached to the lower surface of the upper disk 21 and the upper surface of the lower disk 22, respectively. The polishing pad is a key component that directly contacts the substrate surface and is usually made of materials such as polyurethane, synthetic fibers, or non-woven fabric, with a specific texture and pore structure designed on its surface. The main functions of a polishing pad are threefold: first, to store and transport the polishing slurry, which contains micron-sized abrasive particles and is the core medium for substrate grinding; second, to evenly transfer the pressure applied by the upper platen to the substrate surface, preventing excessive local pressure that could cause substrate breakage; and third, to compensate for minor surface deviations between the substrate and the processing platen through its own elastic deformation, thereby improving processing uniformity. Polishing pads with different hardness, thickness, and texture are suitable for different processing stages. For example, harder polishing pads are typically used in the rough grinding stage to improve grinding efficiency, while softer polishing pads are used in the fine polishing stage to achieve better surface quality.

[0031] The basic working principle of double-sided substrate processing is as follows: First, the substrates to be processed are placed one by one into the substrate placement holes 131 of the planetary gears 13, and then all the planetary gears are evenly placed in the effective processing area of ​​the lower platen 22. Next, the upper platen 21 is slowly lowered until it contacts the upper surface of the substrate, and a set processing pressure is applied. Then, the drive system is activated, causing the upper platen 21 and the lower platen 22 to rotate in opposite directions at a set speed, while simultaneously driving the internal gear ring 12 to rotate at a specific speed. At this time, the planetary gears 13 begin to perform planetary motion under the combined action of the internal and external gear rings, causing the substrate to perform complex planar motion between the upper and lower grinding pads. Simultaneously, the grinding fluid is continuously and evenly delivered between the upper and lower grinding pads, and the abrasive particles roll and slide between the grinding pads and the substrate surface, performing micro-mechanical grinding on the substrate surface, gradually removing the wire-cut damage layer and the original thickness deviation from the substrate surface. Due to the planetary motion characteristics of the substrate, the grinding path at each point on its surface is random and uniformly distributed. Therefore, after a certain processing time, extremely high flatness and uniform thickness can be obtained.

[0032] Dresser 23 is a special tool used to correct machining discs and grinding pads. It is usually made of diamond particles sintered with a metal binder and has extremely high hardness and wear resistance.

[0033] This application provides a method for double-sided substrate processing, such as... Figure 4 As shown, Figure 4 This application provides a schematic flowchart of a substrate double-sided processing method, which includes: S401: Obtain the initial disc spacing difference of the processing discs; The purpose of this step is to characterize the overall unevenness of the processing pad with quantifiable values ​​before processing each batch of substrates. This provides a quantitative basis for matching the corresponding dresser and formulating targeted dressing parameters based on pad shape deviations. Traditional dressing processes skip this quantitative inspection step, using the same dresser and dressing parameters regardless of the wear pattern of the processing pad. This results in the dressed pad shape failing to compensate for the original surface deviation of the substrate, making it difficult to consistently control the total substrate thickness deviation within the acceptable range. This step, by obtaining quantitative data on the pad shape in advance, achieves closed-loop control of inspection before dressing and matching as needed, fundamentally improving the problems of traditional processes.

[0034] Optionally, in one embodiment of this application, the processing disk includes an annular effective processing area, and the disk spacing difference is a first height minus a second height, where the first height is the distance between the upper and lower disks at the edge of the effective processing area, and the second height is the spacing between the upper and lower disks at the outer edge of the effective processing area.

[0035] like Figure 5 As shown, Figure 5This is a cross-sectional schematic diagram of the principle of measuring the spacing difference between processing discs provided in an embodiment of this application. In this diagram, label a corresponds to the first height, which is the vertical distance between the lower surface of the upper disc and the upper surface of the lower disc at the edge of the effective processing area; label b corresponds to the second height, which is the vertical distance between the lower surface of the upper disc and the upper surface of the lower disc at the outer edge of the effective processing area.

[0036] Among them, the disc spacing difference is a surface shape characterization parameter specifically designed for the annular processing disc of double-sided processing equipment. Through the height difference between the two characteristic positions mentioned above, a single value can uniquely and unambiguously characterize the overall surface shape of the processing disc. The effective processing area boundary refers to the largest annular area covered by the movement trajectory of all planetary gears. Its inner boundary is the inscribed circle when the planetary gears are closest to the center of the processing disc, and its outer boundary is the circumscribed circle when the planetary gears are furthest from the center of the processing disc. This boundary is the area where the processing disc actually participates in grinding. Disc surface wear beyond this area does not affect the substrate processing quality and is therefore not included in the measurement range. The vertical distance refers to the absolute distance along the central axis of the processing disc, not the inclined plane distance or distance in other directions. At the same time, this parameter characterizes the overall concavity and convexity trend of the processing disc, rather than local minor defects. Local scratches, pits, and other defects cannot be reflected by this parameter and need to be identified by other specialized detection methods.

[0037] The physical significance of the disc spacing difference has a direct and clear impact on the processing results: When the disc spacing difference is greater than 0, that is, the first height is greater than the second height, the processed disc has a convex structure with a higher inner area and a lower outer area. During processing, the pressure on the central area of ​​the substrate is greater than that on the edge area, resulting in a greater grinding amount in the central area than in the edge area. If no correction is made, the processed substrate will have a surface shape that is thinner in the center and thicker at the edges, and the total thickness deviation will increase significantly. When the disc spacing difference is equal to 0, that is, the first height is equal to the second height, the processed disc has a theoretically ideal planar structure. During processing, the pressure on each position of the substrate is theoretically uniform, and the grinding amount is evenly distributed. This is the ideal disc shape state, but it is difficult to maintain for a long time in actual continuous production. When the disc spacing difference is less than 0, that is, the first height is less than the second height, the processed disc has a concave structure with a lower inner area and a higher outer area. During processing, the pressure on the edge area of ​​the substrate is greater than that on the central area, resulting in a greater grinding amount on the edge area than on the central area. If no correction is made, the processed substrate will have a surface shape that is thinner at the edges and thicker at the center, which will also lead to an unqualified total thickness deviation.

[0038] It should be noted that this application does not limit the specific detection methods for the first height and the second height; at the same time, the initial disc spacing difference can also be obtained by other direct or indirect calculation methods, and this application does not limit this as well.

[0039] S402: Obtain raw data of the substrate, the raw data including at least the surface profile of the substrate and the maximum total thickness deviation in the substrate; the number of substrates is greater than 1.

[0040] The core objective of this step is to understand the original quality characteristics of the substrate batch to be processed before repairing, so that subsequent repair parameters can not only match the wear state of the repaired pad, but also specifically compensate for the surface deviation of the substrate itself, ultimately achieving bidirectional complementarity between the pad shape and the substrate surface shape. Traditional repair processes completely ignore the differences in the original substrate data, using uniform repair and processing parameters regardless of the surface shape or thickness deviation of the substrate to be processed, resulting in large fluctuations in substrate processing quality within a batch and a high rate of surface shape defects. It should be noted that the execution order of steps S401 and S402 can be interchanged.

[0041] The requirement of having more than one substrate is because semiconductor substrates are processed in batches. Multiple substrates within the same batch may exhibit variations in surface shape and thickness due to equipment precision and process fluctuations during pre-processing such as cutting and rough grinding. Measuring data from only a single substrate is unreliable and cannot represent the overall characteristics of the entire batch. By measuring data from multiple substrates, the mainstream surface shape characteristics and extreme deviations of the entire batch can be statistically analyzed, preventing errors in rework parameters due to a few abnormal substrates.

[0042] The maximum total thickness deviation of the substrate refers to the maximum difference between the thickest and thinnest points on a single substrate among all sampled substrates. It reflects the most extreme thickness deviation of the entire batch of substrates and is the core basis for determining the dressing amount and grinding amount. If the maximum total thickness deviation is large, it indicates that there are substrates with significant thickness differences within the batch, requiring a corresponding increase in the dressing amount and an extension of the substrate processing time to ensure that the original thickness deviation of all substrates can be completely removed. If parameters are determined solely based on the average total thickness deviation, individual substrates with large total thickness deviations may still fail to meet the standards after processing, resulting in batch-wide losses.

[0043] The raw data must include the substrate surface shape and the maximum total thickness deviation. The substrate surface shape determines the direction of the dressing, i.e., the complementary shape to which the processed disk needs to be dressed. The maximum total thickness deviation determines the dressing extent, i.e., the dressing amount and the minimum grinding amount for subsequent substrate processing; both are indispensable. It should be noted that the raw data may also include other parameters, such as substrate warpage and curvature, etc., which are not limited in this application.

[0044] Optionally, the raw substrate data can be obtained using the common 5-point thickness measurement method in the art. During measurement, thickness values ​​need to be collected at five points: the center of the substrate and the top, bottom, left, and right points. The four edge points can be located 2mm to 10mm from the substrate edge. Alternatively, non-contact laser interferometry or other conventional measurement methods in the art can be used; this application does not impose any restrictions on this. Surface shape values ​​can be obtained using conventional surface shape analysis methods in the art; this application does not impose any restrictions on this.

[0045] Optionally, the surface profile of the substrate includes: an overall inclined surface profile and a center-high, edge-low surface profile.

[0046] Among them, the center-high and edge-low surface type is the most common surface type after the front-end processing of semiconductor substrates. Its characteristic is that the thickness of the center region of the substrate is greater than the thickness of the edge region, and the whole presents a dome-shaped shape with a convex center. If no targeted compensation is made, after processing with an ideal planar processing disk, the grinding amount of the center region of the substrate will be less than that of the edge region, and the center-high and edge-low surface type will still be retained, resulting in the total thickness deviation exceeding the standard.

[0047] The overall tilted surface is a type of surface that is easily overlooked by traditional processes but has a significant impact on processing quality. Its characteristic is that the thickness of the substrate gradually decreases linearly from one side to the other, and the whole presents a tilted planar shape. If no targeted compensation is made, the substrate will retain the original tilted surface after processing, resulting in a large TTV value.

[0048] It should be noted that the surface shape of the substrate may also include a low center and high edge surface shape and other common surface shapes in the art, which are not limited in this application.

[0049] S403: Based on the original data of the substrate and the initial disk spacing difference of the processing disk, a target trimmer is obtained by matching from a predefined rule base; the predefined rule base contains multiple trimmers, and each trimmer has a corresponding relationship with the surface shape of the substrate, the maximum value of the total thickness deviation of the substrate, and the initial disk spacing difference of the processing disk.

[0050] This step performs three-dimensional correlation matching between the initial disk spacing difference of the processing disk obtained in S401 and the original data of the substrate obtained in S402, and selects the target trimmer most suitable for the current process conditions. This allows the trimmed processing disk surface to simultaneously offset the wear of the processing disk itself and the original surface defects of the substrate, ultimately achieving the best processing effect and reducing the TTV value of the substrate.

[0051] In this embodiment, optionally, the trimmer is a trimming block commonly used in the industry. The predefined rule base is a standardized matching system established in advance through a large number of process experiments. Table 1 is a specific embodiment of this rule base. The row dimension of the rule base corresponds to the substrate surface type and the initial disk spacing difference range of the processing disk, and the column dimension corresponds to the maximum range of the total thickness deviation of the substrate. The content in each cell is the target trimming block specification to be selected under the corresponding working condition. Each set of correspondences in the rule base has been repeatedly verified by the process to ensure that the surface type of the trimmed processing disk can complement the original surface type of the substrate.

[0052] Table 1:

[0053] The specific matching rules in Table 1 are illustrated below: When the substrate surface is generally tilted, the initial disk spacing difference is -60μm to -30μm (i.e., the disk is noticeably concave), and the maximum total substrate thickness deviation is 0 to 2μm, a 2 to 4μm tilt trimming block is matched; when the substrate surface is center-high and edge-low, the initial disk spacing difference is 0μm to 30μm (i.e., the disk is slightly convex), and the maximum total substrate thickness deviation is 4 to 6μm, a 0 to 1μm trimming block is matched. Through this three-dimensional matching method, the optimal trimmer can be selected for each specific substrate condition, maximizing the trimming effect and thus minimizing the substrate's TTV value. It should be noted that there is no fixed relationship between the substrate surface shape and the shape of the trimming block. For example, when the substrate surface is center-high and edge-low, a tilt trimming block or a convex trimming block can be used; this application does not impose any limitations.

[0054] Optionally, in this embodiment, the dressing block is made of a composite structure of a metal substrate support and a diamond electroplated grinding wheel. This structure combines the advantages of both metal substrate and diamond abrasive, resulting in strong structural stability, high dressing accuracy, and excellent wear resistance.

[0055] S404: The processing disc is dressed according to the target dressing device.

[0056] This step involves using a target trimmer to trim the processing disk, replicating the surface features of the target trimmer onto the surface of the processing disk to achieve a preset target gap difference (GAP). Unlike traditional fixed-time trimming, this step employs closed-loop control of the gap difference. Trimming stops immediately when the measured GAP value of the processing disk reaches the target value, rather than ending at a fixed time, ensuring the consistency and repeatability of trimming accuracy. Optionally, in this embodiment, the adjustable gap difference (GAP) range of the processing disk is -60μm to 60μm, which can cover the surface compensation requirements of most semiconductor substrate processing. After trimming, the target gap difference of the processing disk is calculated using the formula: GAP(target) = GAP(original) + GAP(adjustment), where GAP(original) is the initial gap difference of the processing disk measured in step S401, and GAP(adjustment) is the change in gap difference that needs to be achieved through trimming.

[0057] like Figure 6 As shown, Figure 6This application provides a schematic diagram of the dressing principle of a convex dressing block. In the diagram, label 'a' represents the vertical distance between the lower surface of the upper disk and the upper surface of the lower disk at the edge of the effective processing area; label 'b' represents the vertical distance between the lower surface of the upper disk and the upper surface of the lower disk at the outer edge of the effective processing area; label 'c' represents the maximum central convex height of the convex dressing block; and label 'd' represents the total radial length of the convex dressing block. The convex dressing block has an arc-shaped convex structure that is thicker in the middle and thinner at the edges. Its selection is calculated based on the following formula: c / d = GAP (adjustment) / r, where r is the difference between the outer radius and the inner radius of the effective processing area of ​​the processing disk. This formula establishes a quantitative relationship between the surface features of the dressing block and the target disk shape adjustment amount. By matching the ratio of the convex height of the dressing block to its own length, the required change in disk spacing difference is achieved. During dressing, the central convex portion of the convex dressing block generates greater grinding pressure on the central area of ​​the processing disk, thereby removing more material; while the thinner edge section of the dressing block exerts less grinding pressure on the edge area of ​​the processing disk, removing less material. Ultimately, the processing disk will be reshaped into a concave disk shape with a low center and high edges. This disk shape will result in a greater grinding amount in the central region of the substrate than in the edge region during subsequent processing, thus offsetting the original surface defect of a high center and low edges, and effectively reducing the TTV value of the substrate after double-sided processing.

[0058] like Figure 7 As shown, Figure 7 This application provides a schematic diagram of the tilt trimming block's trimming principle. In the diagram, label 'a' represents the vertical distance between the lower surface of the upper plate and the upper surface of the lower plate at the edge of the effective processing area; label 'b' represents the vertical distance between the lower surface of the upper plate and the upper surface of the lower plate at the outer edge of the effective processing area; label 'c' represents the thickness difference between the two ends of the tilt trimming block; and label 'e' represents the total radial length of the tilt trimming block. The tilt trimming block has a wedge-shaped structure, thicker at one end and thinner at the other, with its thickness gradually changing linearly along the radial direction. Its selection is calculated using the following formula: e / R = GAP(adjustment) / r, where R is the diameter of the semiconductor substrate to be processed, and r is the difference between the outer radius and inner radius of the effective processing area of ​​the processing plate. This formula matches the overall tilt amount of the substrate that needs to be compensated through the proportional relationship between the thickness difference of the trimming block and the substrate diameter. During the dressing process, the thicker end of the tilting dressing block will produce a larger grinding amount at the corresponding position of the processing disk, while the thinner end will produce a smaller grinding amount. Ultimately, the processing disk will be dressed into a tilted disk shape opposite to the tilting direction of the substrate. This disk shape will cause the grinding amount on the thicker side of the substrate to be greater than that on the thinner side during subsequent processing, thus offsetting the overall tilting surface defect of the substrate and effectively reducing the TTV value of the substrate after double-sided processing.

[0059] Optionally, this application embodiment provides a dressing operation process: First, the matched target dresser is placed in the effective processing area of ​​the lower plate, ensuring that the dresser is completely in contact with the surface of the lower plate; then, the upper plate is slowly lowered until it contacts the upper surface of the target dresser, and a set dressing pressure is applied; then, the drive system is started, and the dressing is performed using a speed ratio of n1 (upper plate speed):n2 (lower plate speed):n3 (internal gear ring speed):n4 (external gear ring speed) = (10~15r / min):(-10~-15r / min):(5~10r / min):(0.1~3r / min), where the negative sign before the speed value indicates that the rotation direction of the component is opposite to that of the upper plate; during the dressing process, the disc gap difference GAP of the processing disc is monitored in real time. When the measured GAP value reaches the preset target GAP value, the equipment operation is stopped immediately, the upper plate is lifted and the dresser is removed, and the dressing operation is completed.

[0060] It should be noted that this application does not limit the specific dressing pressure or the precise rotation speed of each component during the dressing process. Those skilled in the art can make adjustments according to actual processing needs.

[0061] S405: Place the substrate in the carrier of the processing tray and process the substrate.

[0062] Based on the target disk shape obtained by the aforementioned disk repair, this step processes the substrate. During the processing, the grinding amount at each position of the substrate is uniform and consistent, which alleviates the problem of insufficient or excessive grinding in local areas caused by the mismatch between the disk shape and the substrate surface shape in traditional fixed disk shape processing, thereby effectively reducing the TTV value of the substrate after double-sided processing.

[0063] Optionally, in one embodiment of this application, placing the substrate within the carrier of the processing tray includes: Obtain the center point thickness of the entire batch of substrates, and sort all substrates in descending order of the center point thickness value; Select the top N substrates, where N is an integer greater than 1, and the number of carriers of the processing disk is an integer multiple of N; The selected N substrates are placed in a uniform rotationally symmetrical position that completely overlaps after rotating 360° / N with the geometric center of the carrier of the processing disk as the rotation center, so that the N points with the largest center point thickness values ​​are evenly distributed in the processing disk. The remaining substrate is placed in the remaining carrier of the processing tray.

[0064] In this embodiment, the purpose of the substrate arrangement is to solve the problem of localized load concentration on the disk surface caused by traditional random wafer loading by symmetrically distributing thick substrates. After the previous cutting and rough grinding processes, there are generally differences in the thickness of individual substrates in a batch. If the thicker substrates are concentrated in a local area of ​​the processing disk, the loading disk will tilt slightly due to excessive local pressure in the early stage of processing, directly destroying the parallelism of the disk surface, causing uneven grinding of substrates at different positions, and ultimately increasing the batch TTV deviation. This arrangement method ensures balanced force on the processing disk surface from the source of wafer loading by allocating the positions of thick substrates, maintaining the horizontal state of the loading disk throughout the process, and ensuring the consistency of grinding of all substrates.

[0065] In the specific implementation process, the industry-standard 5-point thickness measurement method can be used to collect the center point thickness data of each substrate. Based on the measured thickness data, the center point thickness of the entire batch of substrates can be sorted from largest to smallest to provide data support for subsequent differentiated arrangement.

[0066] This application uses a total of 12 carriers and N=3 as an example: the substrate with the thickest center point is placed at carrier 1, the substrate with the second thickest center point is placed at carrier 5, and the substrate with the third thickest center point is placed at carrier 9. These three carrier positions are evenly distributed along the geometric center of the processing disk, forming a stable equilateral triangular support structure. In the initial processing stage, this symmetrical triangular support structure provides uniform support to the upper and lower processing disks, effectively constraining the offset and tilt of the upper disk. Simultaneously, it prioritizes grinding the substrates with higher thicknesses, eliminating local thickness discrepancies in advance and completely avoiding disk tilting caused by excessively thick substrates in certain areas. This also prevents problems such as poor processing and excessive thickness deviations of other substrates due to disk imbalance. After completing the symmetrical arrangement of the thicker substrates, all remaining substrates are then placed sequentially in the empty carrier positions of the processing disk.

[0067] Optionally, this application does not limit the specific value of N. N only needs to be set to an integer greater than 1. At the same time, it can be adapted to the substrate arrangement process of this application and to double-sided processing equipment of different specifications and different numbers of carriers, provided that the total number of carriers of the processing disk is an integer multiple of N and the force balance of the processing disk surface is maintained after the N substrates are symmetrically arranged.

[0068] Optionally, based on the above embodiments, in one embodiment of this application, placing the remaining substrate within the remaining carrier of the processing tray includes: The remaining substrates are sorted according to a preset rule, and the top M substrates are selected from the remaining substrates, where M is an integer equal to N. The selected M-shaped substrates are placed in the remaining uniform rotationally symmetrical empty spaces that completely overlap after rotating 360° / M with the geometric center of the carrier of the processing disk as the rotation center. Repeat the above steps until all the carriers of the processing trays are filled.

[0069] This embodiment is a secondary optimization arrangement process after the symmetrical arrangement of the central thick sheets. On the basis of completing the balanced arrangement of the substrate with the central point thickness, the remaining substrate is arranged symmetrically to address the edge thickness differences. This solves the grinding deviation problem caused by uneven thickness at the edges of the batch substrates and the concentration of local thin points, further improving the grinding uniformity of the entire substrate and ensuring the stability of the TTV value after grinding.

[0070] The preset rule can adopt an edge thickness sorting rule. In specific implementation, based on the data results of the substrate 5-point thickness measurement method, the thickness values ​​of the four edge points other than the center point of each remaining substrate are extracted, and all edge thickness point values ​​are uniformly sorted from smallest to largest. At the same time, a sorting avoidance rule is set: if two consecutive minimum thickness points come from the same substrate during the sorting process, the duplicate wafer point is skipped, and the minimum edge thickness point of the next different substrate is selected in sequence, so as to avoid multiple thin points on a single substrate being concentrated in one place and to ensure the dispersion of point sources.

[0071] Taking a 12-carrier processing tray with M=3 as an example: Place the substrate corresponding to the minimum edge thickness point (ranked first) in carrier 2; place the substrate corresponding to the minimum edge thickness point of the second-ranked substrate in carrier 6; and place the substrate corresponding to the minimum edge thickness point of the third-ranked substrate in carrier 10. Then place the substrate corresponding to the minimum edge thickness point of the fourth-ranked substrate in carrier 3; the substrate corresponding to the minimum edge thickness point of the fifth-ranked substrate in carrier 7; and the substrate corresponding to the minimum edge thickness point of the sixth-ranked substrate in carrier 11. Finally, place the remaining unplaced substrates into the remaining empty spaces in carriers 4, 8, and 12. Repeat the above steps until all carriers in the processing trays are filled.

[0072] By using the above-mentioned layered and symmetrical staggered arrangement, the thick and thin points of the entire substrate can be evenly distributed in various areas of the processing disk, completely avoiding the problem of concentrated accumulation of thick and thin points, making the grinding amount in each area of ​​the processing disk more consistent, and greatly improving the processing uniformity of the entire batch of substrates.

[0073] Optionally, this application does not limit the specific type of the preset rule. In addition to the above-mentioned edge thickness sorting rule, the substrate center thickness sorting rule or other conventional substrate sorting and arrangement rules in the art can also be used, as long as the remaining substrates can be arranged uniformly and symmetrically and the force on the disk surface can be balanced. This application does not limit this.

[0074] Optionally, in one embodiment of this application, the processing of the substrate includes: The substrate is first processed using a first pressure value; When the preset switching conditions are met, the substrate is then switched to the second pressure value for processing. The second pressure value is greater than the first pressure value.

[0075] This embodiment employs a segmented pressure processing technology, dividing the substrate grinding process into a low-pressure calibration stage and a high-pressure fine grinding stage. This effectively addresses the shortcomings of traditional single-pressure processing, which cannot simultaneously ensure thickness calibration accuracy and processing efficiency. The low-pressure processing stage is primarily used to smoothly correct the initial thickness differences of the entire batch of substrates, gradually unifying the thickness level of all substrates and eliminating minor thickness deviations remaining from the initial substrate mounting arrangement. The high-pressure processing stage, after the substrate thickness has become uniform, efficiently removes the residual wire-cut damage layer and remaining processing allowance from the substrate surface, completing precision grinding and polishing to ensure the final thickness accuracy and surface quality of the substrate.

[0076] Optionally, the process parameters for segmented processing can be as follows: First, use low-pressure processing (100N-500N) for 5-15 minutes to calibrate the thickness difference of the entire substrate, making the thickness of each substrate more consistent. After meeting the switching conditions, switch to high-pressure processing (600N-1000N) for 20-40 minutes to perform deep grinding and polishing of the substrate until it reaches the target thickness specification. This segmented pressure processing mode can gradually correct substrate thickness deviations, avoid uneven grinding caused by direct high-pressure processing, and ensure the stability of the TTV value after grinding and reduce the TTV value.

[0077] Optionally, based on the above embodiments, in one embodiment of this application, the preset switching conditions include: the spindle torque fluctuation value of the processing equipment is stable within 1% for a continuous preset time, and the substrate thickness reaches a preset value.

[0078] Spindle torque fluctuation is a key indirect indicator reflecting substrate thickness uniformity. When the thickness difference of all substrates is basically eliminated, the upper plate is subjected to uniform force, and the spindle torque fluctuation will significantly decrease and tend to stabilize. When the spindle torque fluctuation remains stable within 1% for a preset time and the substrate thickness reaches the preset switching thickness, it indicates that the low-pressure calibration stage has been completed. At this point, switching to high pressure for fine grinding can significantly improve processing efficiency while reducing the TTV value.

[0079] Optionally, in one embodiment of this application, the first pressure value is 10 to 15 times the first value; Wherein, the first value is the difference between the maximum and minimum values ​​of the center point thickness in the first type of substrate; the first type of substrate consists of N substrates placed in a uniformly rotationally symmetrical position with the center point thickness values ​​ranked first.

[0080] This embodiment provides a quantitative selection basis for the first pressure value. The low pressure value is determined based on the thickness difference of the thick substrates arranged at symmetrical points, avoiding blind selection of low pressure parameters. The low pressure value is adaptively matched according to the actual maximum thickness difference of the substrates. The larger the thickness difference, the higher the corresponding initial low pressure. The low pressure set according to the above rules can slowly grind the ultra-thick substrate and gradually reduce the thickness difference of the entire batch of substrates, without causing excessive wear damage to the thin substrate due to excessive pressure, which is suitable for the purpose of early low pressure calibration.

[0081] To more clearly demonstrate the difference between the double-sided substrate processing method of this application and the traditional double-sided substrate polishing method, the applicant set up a set of parallel control verification experiments. To ensure the comparability of the test results, this verification experiment strictly followed the single variable principle. The embodiments of this application and the traditional comparative examples were carried out under completely identical test conditions. The unified test settings are as follows: the test raw materials were the same batch of 8-inch silicon carbide substrates, all of which were center-high and edge-low surface type, and the original TTV of the substrates was uniformly 3.96μm; the test equipment used was the same double-sided polishing equipment, equipped with a 12-station planetary carrier disk, and the polishing pad model, polishing slurry composition and concentration, ambient temperature and humidity, substrate cleaning process and testing equipment were all kept consistent.

[0082] In this verification experiment, Example 1 uses the double-sided substrate processing method of this application, and the specific steps are as follows: First, the initial GAP (Gap Amount) difference between the two grinding discs was measured to be -24 μm. Then, considering the substrate's maximum TTV (Total TV Value) of 4 μm and its center-high, edge-low surface characteristics, a target trimming block was matched according to a predefined rule library, and the required adjustment GAP value for trimming was calculated to be 11 μm. Based on the formula GAP(target) = GAP(original) + GAP(adjustment), the target GAP value for trimming was determined to be -13 μm. According to the matching rules, a tilted trimming block with a specification of 2–4 μm was selected, and the trimming block was used for trimming. Trimming was stopped once the measured GAP of the grinding disc reached the target value of -13 μm. Secondly, 12 substrates were randomly selected from this batch as test samples. The average TTV of the test samples was 3.5 μm, the maximum thickness of a single substrate was 372 μm, and the minimum thickness was 367 μm. Then, the thickness data of the center point and the four edge points (top, bottom, left, and right) of each of the 12 substrates were collected using the 5-point thickness measurement method. All thickness data were entered into Excel and sorted. Then, N=3 substrates with the highest center point thickness were selected and placed in carriers 1, 5, and 9 according to the symmetrical position of an equilateral triangle. Then, the edge thickness secondary symmetrical arrangement process of this application was performed. Based on the thickness data of the four edge points of the remaining substrates, the remaining substrates were sorted and placed in the remaining empty positions of carriers 2, 6, 10, 3, 7, 11, 4, 8, and 12.

[0083] Finally, the first pressure value is calculated, and the substrate is processed according to the first pressure value. After the spindle torque fluctuation stabilizes within 1% for a preset time and the substrate thickness reaches the preset switching value, the second pressure value is switched to complete the fine polishing. Then, the substrate that has completed double-sided processing is cleaned.

[0084] Meanwhile, the applicant added Comparative Example 1, which adopts the existing traditional double-sided grinding scheme throughout the process. The difference between Comparative Example 1 and Example 1 is that: the initial disk shape GAP value of the processing disk is not measured, the trimming block is not matched according to the incoming material surface shape, and the fixed specification trimming block is used for long-term trimming at a fixed cycle; the substrate is loaded in a completely random manner, without a center point or edge thickness graded symmetrical arrangement process; the grinding process adopts a single fixed pressure throughout the process, without a segmented pressure adjustment process of low pressure calibration and high pressure fine grinding.

[0085] After processing, an optical profilometer was used to acquire three-dimensional thickness contour maps of the original substrate surface, the surface of the test sample in Example 1, and the surface of the test sample in Comparative Example 1. The corresponding figures are as follows: Figure 8 , Figure 9 , Figure 10 All cloud map legends are in μm, and the colors from purple to red represent substrate thickness from thin to thick.

[0086] like Figure 8 As shown, Figure 8 This embodiment of the application provides a three-dimensional topographic cloud image of the original surface of a substrate. The central region of the substrate is distinctly red, while the edge region is purple, exhibiting a typical convex-central and concave-edge surface shape. The thickness distribution decreases rapidly from the center to the edge, and the measured original TTV is 3.96 μm. Figure 9 As shown, Figure 9 The three-dimensional topography cloud image of the substrate surface after conventional processing in Comparative Example 1 provided in this application shows that the thickness distribution of the finished substrate is still uneven, with obvious abrupt thickness changes in local areas. Due to the influence of random mounting and fixed disk type, there is a significant thick protrusion on one side of the substrate, and the overall surface distortion has not been effectively compensated. The measured TTV of the finished product is 2.12 μm. Figure 10 As shown, Figure 10 The three-dimensional topography cloud image of the substrate surface after processing by the double-sided substrate processing method of this application in Example 1 shows that the substrate thickness is uniformly and smoothly transitioned in the radial direction, the color gradient is continuous and regular, the overall surface shape is close to an ideal plane, the thickness fluctuation range is significantly narrowed, and there are no obvious local protrusions or depressions. The measured TTV of the finished product is 1.03μm.

[0087] In this verification test, the TTV of the original substrate was uniformly 3.96 μm; the TTV of the substrate after processing by the traditional process (Comparative Example 1) was 2.12 μm, and the TTV reduction was 46.5%; the TTV of the substrate after processing by the process of this application (Example 1) was 1.03 μm, and the TTV reduction was 74.0%, which is 51.4% lower than that of the traditional process.

[0088] Therefore, it is evident that traditional double-sided grinding processes do not adaptively adjust the processing disc according to the incoming material's surface shape, failing to create a disc shape that complements the original substrate surface shape. Furthermore, the random loading method easily leads to uneven local stress on the disc surface, and a single fixed pressure cannot simultaneously address thickness difference calibration and processing efficiency, only partially offsetting the original substrate surface shape deviation, resulting in limited reduction in the final product's total thickness volume (TTV). This application addresses this by selecting the target disc dresser based on the original data of the processing disc and substrate, employing a symmetrical loading of thick and thin substrates, and a coordinated process of low-pressure pre-calibration followed by high-pressure fine grinding. This process simultaneously offsets the original substrate thickness deviation and processing disc wear deviation from three dimensions: disc shape matching, stress uniformity, and segmented grinding, significantly improving the surface shape accuracy of the ground substrate and the consistency of batch processing.

[0089] This application provides a substrate double-sided processing apparatus for performing the substrate double-sided processing method described in the foregoing embodiments, such as... Figure 11 As shown, Figure 11 This application provides a structural block diagram of a substrate double-sided processing apparatus. The apparatus 110 includes: The processing tray status acquisition module 1101 is used to acquire the initial tray spacing difference of the processing tray; The substrate parameter acquisition module 1102 is used to acquire raw data of the substrate, the raw data including at least the surface shape of the substrate and the maximum total thickness deviation in the substrate; the number of substrates is greater than 1; The parameter matching module 1103 is used to match a target trimmer from a predefined rule base based on the original data of the substrate and the initial disk spacing difference of the processing disk; the predefined rule base contains multiple trimmers, and each trimmer has a corresponding relationship with the surface shape of the substrate, the maximum value of the total thickness deviation of the substrate, and the initial disk spacing difference of the processing disk. The dressing execution module 1104 is used to dress the processing disc according to the target dressing device; The substrate and processing module 1105 are used to place the substrate in the carrier of the processing tray and process the substrate.

[0090] Optionally, the processing disk includes an annular effective processing area, and the disk spacing difference is a first height minus a second height, where the first height is the distance between the upper and lower disks at the edge of the effective processing area, and the second height is the spacing between the upper and lower disks at the outer edge of the effective processing area.

[0091] Optionally, the fabrication and processing module includes a fabrication sub-module, specifically used to obtain the center point thickness of the entire batch of substrates, sort all substrates according to the center point thickness values ​​from largest to smallest; select the top N substrates in the sorting, where N is an integer greater than 1, and the number of carriers on the processing disk is an integer multiple of N; place the selected N substrates in a uniformly rotationally symmetrical position that completely overlaps after rotating 360° / N around the geometric center of the carriers on the processing disk, so that the N points with the largest center point thickness values ​​are evenly distributed within the processing disk; and place the remaining substrates in the remaining carriers of the processing disk.

[0092] Optionally, the fabrication and processing module includes a fabrication sub-module, specifically used to sort the remaining substrates according to a preset rule, select the top M substrates from the remaining substrates, where M is an integer equal to N; place the selected M substrates in the remaining uniform rotationally symmetrical empty spaces that completely overlap after rotating 360° / M with the geometric center of the carrier of the processing disk as the rotation center; repeat the above steps until all carriers of all processing disks are filled.

[0093] Optionally, the fabric and processing module includes a processing submodule, specifically used to process the substrate with a first pressure value first; and when a preset switching condition is met, to switch to a second pressure value to process the substrate; wherein the second pressure value is greater than the first pressure value.

[0094] Optionally, the first pressure value is 10 to 15 times the first value; wherein, the first value is the difference between the maximum and minimum values ​​of the center point thickness in the first type of substrate; the first type of substrate consists of N substrates placed in a uniformly rotationally symmetrical position with the center point thickness values ​​ranked first.

[0095] Optionally, the preset switching conditions include: the spindle torque fluctuation value of the processing equipment is stable within 1% for a continuous preset time, and the substrate thickness reaches a preset value.

[0096] Optionally, the surface profile of the substrate includes: an overall inclined surface profile and a center-high, edge-low surface profile.

[0097] Furthermore, this application also provides a substrate, which is prepared by the substrate double-sided processing method provided in any of the foregoing embodiments.

[0098] The various embodiments in this specification are described in a progressive, parallel, or combined manner. Each embodiment focuses on the differences from other embodiments, and the same or similar parts between the embodiments can be referred to each other.

[0099] It should be noted that, in the description of this application, the accompanying drawings and embodiments are illustrative rather than restrictive. The same reference numerals throughout the embodiments identify the same structures. It should also be noted that, in this document, relational terms such as "first" and "second" are used merely 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. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device comprising a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in an article or device comprising the aforementioned element.

[0100] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for double-sided substrate processing, characterized in that, include: Obtain the initial disc spacing difference of the processing discs; Obtain raw data of the substrate, the raw data including at least the surface profile of the substrate and the maximum total thickness deviation in the substrate; the number of substrates is greater than 1; Based on the original data of the substrate and the initial disk spacing difference of the processing disk, a target trimmer is obtained by matching from a predefined rule base; the predefined rule base contains multiple trimmers, and each trimmer has a corresponding relationship with the surface shape of the substrate, the maximum value of the total thickness deviation of the substrate, and the initial disk spacing difference of the processing disk. The processing disc is dressed according to the target dressing device; The substrate is placed in the carrier of the processing tray and processed.

2. The method according to claim 1, characterized in that, The processing disk includes an annular effective processing area. The disk spacing difference is a first height minus a second height. The first height is the distance between the upper and lower disks at the edge of the effective processing area, and the second height is the spacing between the upper and lower disks at the outer edge of the effective processing area.

3. The method according to claim 1, characterized in that, The step of placing the substrate within the carrier of the processing tray includes: Obtain the center point thickness of the entire batch of substrates, and sort all substrates in descending order of the center point thickness value; Select the top N substrates, where N is an integer greater than 1, and the number of carriers of the processing disk is an integer multiple of N; The selected N substrates are placed in a uniform rotationally symmetrical position that completely overlaps after rotating 360° / N with the geometric center of the carrier of the processing disk as the rotation center, so that the N points with the largest center point thickness values ​​are evenly distributed in the processing disk. The remaining substrate is placed in the remaining carrier of the processing tray.

4. The method according to claim 3, characterized in that, The step of placing the remaining substrate within the remaining carrier of the processing tray includes: The remaining substrates are sorted according to a preset rule, and the top M substrates are selected from the remaining substrates, where M is an integer equal to N. The selected M-shaped substrates are placed in the remaining uniform rotationally symmetrical empty spaces that completely overlap after rotating 360° / M with the geometric center of the carrier of the processing disk as the rotation center. Repeat the above steps until all the carriers of the processing trays are filled.

5. The method according to claim 1, characterized in that, The processing of the substrate includes: The substrate is first processed using a first pressure value; When the preset switching conditions are met, the substrate is then switched to the second pressure value for processing. The second pressure value is greater than the first pressure value.

6. The method according to claim 3 or 5, characterized in that, The first pressure value is 10 to 15 times the first value; Wherein, the first value is the difference between the maximum and minimum values ​​of the center point thickness in the first type of substrate; the first type of substrate consists of N substrates placed in a uniformly rotationally symmetrical position with the center point thickness values ​​ranked first.

7. The method according to claim 5, characterized in that, The preset switching conditions include: the spindle torque fluctuation value of the processing equipment is stable within 1% for a continuous preset time, and the substrate thickness reaches a preset value.

8. The method according to claim 1, characterized in that, The surface profiles of the substrate include: an overall inclined surface profile and a surface profile with a high center and low edges.

9. A substrate double-sided processing apparatus, characterized in that, include: The processing tray status acquisition module is used to obtain the initial tray spacing difference of the processing tray; A substrate parameter acquisition module is used to acquire raw data of the substrate, the raw data including at least the surface profile of the substrate and the maximum total thickness deviation in the substrate; the number of substrates is greater than 1; The parameter matching module is used to match the target trimmer from a predefined rule base based on the original data of the substrate and the initial disk spacing difference of the processing disk; the predefined rule base contains multiple trimmers, and each trimmer has a corresponding relationship with the surface shape of the substrate, the maximum value of the total thickness deviation of the substrate, and the initial disk spacing difference of the processing disk. The dressing execution module is used to dress the processing disc according to the target dressing device; A fabric sheet and processing module are used to place the substrate in the carrier of the processing tray and process the substrate.

10. A substrate, characterized in that, The substrate is prepared by the substrate double-sided processing method according to any one of claims 1-8.

Images