Wafer film thickness uniformity correction method based on double chemical mechanical polishing devices

Abstract

The invention relates to a wafer film thickness uniformity correction method based on double chemical mechanical polishing equipment, and belongs to the technical field of wafer processing. The method comprises the following steps that (1) a first polishing machine and a second polishing machine are provided, the first polishing machine shows the characteristic that the flat edge area is systematically thicker, and the second polishing machine shows the characteristic that the flat edge area is systematically thinner; 2) polishing the wafer by a first polishing machine; (3) the second polishing machine polishes the wafer at a lower speed relative to the first polishing machine; the target removal amount of the polishing operation in the step 2) accounts for 70-90% of the total target removal amount of the polishing operation in the step 2) and the step 3). According to the method, starting from production line system integration, collaborative correction is carried out through inherent differences between devices, the intractable area problem that the flat edge is too thick or too thin can be solved, and the important practical production application value is achieved.

Description

Technical Field

[0001] This application relates to a method for correcting wafer film thickness uniformity based on dual chemical mechanical polishing equipment, belonging to the field of wafer processing technology. Background Technology

[0002] Chemical mechanical polishing (CMP) is a key process for achieving global planarization in semiconductor manufacturing. Its basic principle is to press the wafer onto a rotating polishing pad and provide a polishing slurry containing tiny abrasive particles and chemical reagents. By utilizing the synergistic effect of mechanical abrasion and chemical reaction, the material on the wafer surface is selectively or non-selectively removed, thereby obtaining a flat surface with nanometer-level precision.

[0003] The ideal goal of CMP (Chip Motion Processing) is to achieve highly uniform material removal across the entire wafer surface. However, achieving perfect global uniformity is a significant challenge due to the complexity of the physical, mechanical, and chemical processes involved. Wafer edge regions, especially those with flat edges or notches (positioning edges), become major sources of CMP uniformity problems because their unique geometry disrupts the wafer's axisymmetry. This results in flow fields, pressure distributions, and dynamic conditions that differ drastically from those in the center region.

[0004] Due to individual differences in polishing equipment (such as polishing pad pressure distribution, polishing head design, and polishing fluid flow characteristics), wafers may repeatedly exhibit regional film thickness unevenness after polishing, such as excessively thick or thin film at the wafer's flat edges. Each CMP machine, due to minute tolerances in its mechanical assembly and different historical wear conditions of the polishing pad and carrier ring, will form its own unique and inherent non-uniformity distribution pattern, with "excessively thick flat edges" and "excessively thin flat edges" being the two most common and difficult-to-handle defect patterns.

[0005] To avoid the above problems, existing technologies often employ the following operational methods: 1) Optimize the process parameters of a single piece of equipment, such as the selection of polishing fluid, the pressure setting of the polishing head and retaining ring, and the speed setting of the polishing head and polishing pad; 2) Modify the equipment hardware, modifying or updating key components; 3) Purchase higher-performance equipment, directly purchasing the latest and most advanced CMP equipment.

[0006] For optimizing process parameters on a single piece of equipment, modifying the entire recipe is necessary to correct membrane non-uniformity in a specific area. This requires repeated verification experiments and is prone to creating a "robbing Peter to pay Paul" situation—increasing the cost of one area while introducing a new non-uniform region. Furthermore, the physical limitations and inherent conditions of the equipment restrict the effectiveness of process parameter optimization. Hardware modifications are costly, time-consuming, and still require repeated testing after modification, leading to significant uncertainty. Purchasing higher-performance equipment is not a viable non-technical solution and is not economical; the new equipment may also have inherent defects.

[0007] Patent application CN202510621808.X describes a method that modifies the secondary polishing process on the same or the same type of equipment, such as using a softer polishing pad, finer abrasive, or polishing slurry with different chemical properties. This method, through secondary polishing, focuses more on improving the surface roughness of GaN rather than its macroscopic uniformity, and cannot systematically correct the inherent macroscopic thickness non-uniformity distribution formed during the first polishing step.

[0008] Patent application CN202510405863.5 describes a method that uses an online measurement system to acquire the wafer film thickness distribution in real time and employs a complex algorithm model to dynamically adjust the pressure in different areas of the polishing head for polishing. This method requires integrating a high-precision real-time measurement system and a fast-response, zoned-pressure polishing head, resulting in high equipment and maintenance costs. Furthermore, the control effect heavily relies on the accuracy of the algorithm model. In actual polishing, the polishing pad state, slurry concentration, and other factors are constantly changing, making it difficult for the model to be completely accurate.

[0009] Patent application CN202111607221.1 describes a method that measures the location of the non-uniform area (thicker area) on the front side of the wafer and then applies a correction pad of a specific shape and thickness to the corresponding location on the back side of the wafer. This causes the thicker area on the front side to experience greater actual pressure during secondary polishing due to the pad, resulting in more material removal. However, this method is cumbersome, requiring individual measurement and customized application of correction pads to each wafer, significantly reducing production efficiency and throughput. Furthermore, attaching foreign matter to the back side of the wafer poses a risk of contamination from detachment. Additionally, excessive localized stress may lead to wafer breakage.

[0010] Therefore, how to improve the persistent problem of "overly thick or thin flat edges" is a technical problem that urgently needs to be solved. Summary of the Invention

[0011] To address the aforementioned issues, this application provides a wafer film thickness uniformity correction method based on dual polishing machine collaboration. Starting from production line system integration, it utilizes the inherent differences between equipment for collaborative correction, which can significantly improve the persistent problem of "overly thick or thin flat edges" and has important production application value and significance.

[0012] This application provides a method for correcting wafer film thickness uniformity based on a dual chemical mechanical polishing device, the method comprising the following steps: 1) Provide a first polishing machine and a second polishing machine, wherein the first polishing machine exhibits a systematic thickening characteristic in the flat edge area, and the second polishing machine exhibits a systematic thinning characteristic in the flat edge area; 2) The first polishing machine polishes the wafer with the following parameters: polishing head pressure 350±50g / cm. 2 Maintain ring pressure of 410±40g / cm 2 The polishing head speed is 71±5 rpm, the polishing pad speed is 68±5 rpm, and the polishing fluid flow rate is 90±10 ml / min. 3) The second polishing machine polishes the wafer with the following parameters: polishing head pressure 50±5g / cm. 2 Maintain ring pressure at 140±10g / cm 2 Polishing head speed 49±5 rpm, polishing pad speed 51±5 rpm, polishing fluid flow rate 90±10 ml / min; The target removal amount in step 2) of the polishing operation accounts for 70-90% of the total target removal amount in steps 2) and 3).

[0013] The first step, "high-pressure, high-speed" polishing, undertakes the task of removing the main material and ensures a high material removal rate to meet mass production efficiency requirements. At the same time, under high pressure, the inherent fingerprint of the equipment's "systematic thickness on the flat edge" is fully revealed and solidified, becoming a clear target for the second step of correction. The second step, "low-pressure, low-speed" polishing, under lower pressure, allows the contact between the wafer and the polishing pad to occur preferentially in the systematically thicker areas, thereby achieving selective correction of "protrusion preferential removal". If the pressure in the second step is too high, this selectivity will be lost; if the pressure is too low, the removal rate will be too low, reducing the correction efficiency.

[0014] Meanwhile, the low rotation speed prolongs the residence time of the polishing slurry on the wafer surface, enhancing the dominant role of the chemical reaction and facilitating selective correction. The first-step parameters being greater than the second-step parameters essentially represents the physical manifestation of the two process modes: "rough polishing" and "fine finishing." If the first-step parameters are insufficient, the fingerprint magnification will be inadequate, leaving no target for the second-step repair; if the second-step parameters are too high, the correction process will introduce new non-uniformities, destroying the compensation effect.

[0015] Besides the previous one, the selection of the specific parameter ranges in the first and second steps is of great significance. The inventors have verified that only when the parameter ranges are within the above-mentioned ranges can a better effect be achieved. If the ranges are exceeded, the effect will be significantly worse or fail.

[0016] In the first polishing step, the polishing head pressure is 350±50g / cm.2 If the pressure is below the lower limit, the removal efficiency is insufficient, and the non-uniformity and thicker edge characteristics are not obvious enough; if it is above the upper limit, edge stress concentration increases, the risk of fragmentation increases, and it is easy to cause excessive edge polishing, introducing new non-uniformity. Maintain a ring pressure of 410±40 g / cm. 2 If the pressure is below the lower limit, insufficient edge pressure and low edge removal will easily lead to a thicker edge ring; conversely, if the pressure is above the upper limit, it will exacerbate over-polishing of the edges. The polishing head speed should be 71±5 rpm. If it is below the lower limit, the relative speed decreases, the removal rate decreases, and the abrasive grain trajectory coverage may be uneven; if it is above the upper limit, polishing slurry will splash, causing a significant temperature rise, and uneven temperature distribution will lead to new non-uniformity. The polishing pad speed should be 68±5 rpm, which needs to be matched with the polishing head speed. If the speed is mismatched, the relative motion will be insufficient, uniformity will decrease, and periodic abrasive grain trajectories may appear.

[0017] In the second polishing step, the polishing head pressure is 50±5g / cm. 2 If the pressure is below the lower limit, the removal rate is extremely low, resulting in low production efficiency; if it is above the upper limit, the excessive pressure leads to a loss of selective correction capability. Maintain the ring pressure at 140±10 g / cm³. 2 If the speed is below the lower limit, the edge constraint is insufficient, and the wafer edge lifting removal rate becomes uncontrolled; if it is above the upper limit, the edge pressure is too high, which may introduce new edge non-uniformity. The polishing head speed is 49±5 rpm. If it is below the lower limit, the polishing slurry renewal rate is too slow, the chemical reaction is limited, and the efficiency is low; if it is above the upper limit, the mechanical action is enhanced, and the selectivity decreases. The polishing pad speed is 51±5 rpm, and this parameter is matched with the polishing head speed.

[0018] The polishing operation aims to remove 70-90% of the total material, covering the theoretically feasible range while retaining a process margin to accommodate production variables such as equipment condition fluctuations and incoming material differences. From a production efficiency perspective, the first-step high-pressure, high-speed polishing has a high removal rate, handling the majority of the removal and ensuring overall production efficiency. If the first step accounts for less than 70%, the second step must handle over 30% of the removal, while the removal rate of the second step's low-pressure, low-speed polishing is only 1 / 10 to 1 / 15 of the first step, significantly extending the total polishing time and impacting production line throughput. From a polishing accuracy perspective, if the first step accounts for more than 90%, the second step's removal is less than 10%. Since the correction effect of the second step is gradual, insufficient removal cannot completely offset the non-uniformity deviation left by the first step. Repeated experimental results show that when the first step accounts for >90%, the final deviation Δfinal often exceeds 10nm, at which point the correction effect significantly decreases, making it unsuitable for actual production.

[0019] Optionally, the target removal amount in the polishing operation in step 2) accounts for 75-85% of the total target removal amount in the polishing operations in steps 2) and 3).

[0020] When the first step accounts for 75% to 85%, the theoretically calculated Δfinal is closest to 0nm. And after repeated experimental data verification, it is shown that in this ratio range, the final CV is stable at 3 to 4%, which is significantly better than other ratio ranges. Moreover, this range provides sufficient process margin to accommodate equipment status fluctuations and material differences, and the process robustness is the highest.

[0021] Optionally, the target removal amount of the polishing operation in step 2) accounts for 80% of the total target removal amount of the polishing operations in steps 2) and 3).

[0022] Optionally, the polishing time of the first polishing machine and the second polishing machine is calculated and determined based on the target removal amount of the polishing operation.

[0023] Optionally, the parameters for polishing the wafer by the first polishing machine are as follows: polishing head pressure 350g / cm. 2 Maintain ring pressure of 410 g / cm 2 The polishing head speed was 71 rpm, the polishing pad speed was 68 rpm, and the polishing fluid flow rate was 90 ml / min. The parameters for polishing the wafer by the second polishing machine are as follows: polishing head pressure 50±5g / cm 2 Maintain ring pressure at 140±10g / cm 2 Polishing head speed 49±5 rpm, polishing pad speed 51±5 rpm, polishing fluid flow rate 90±10 ml / min.

[0024] The first-step polishing parameters (polishing head pressure 350g / cm², retaining ring pressure 410g / cm², polishing head speed 71rpm, polishing pad speed 68rpm, polishing fluid flow rate 90ml / min) are the benchmark operating points determined through system optimization, and have the following technical significance: 1) The above-mentioned point parameters represent the most stable and repeatable manifestation conditions for the "systematic thickening of the flat edge" characteristic of the first polishing machine. Numerous experiments have shown that when the first-step polishing parameters are fixed at these point values, the inherent deviation of equipment A remains stable within a certain range with a small coefficient of variation. If the parameters fluctuate around these point values, the manifestation of the equipment fingerprint may fluctuate by ±20%, affecting the accuracy of subsequent compensation. 2) Fixing the first-step parameters at these point values ​​makes ΔA a stable constant, thereby simplifying the compensation model and improving the accuracy of proportional calculations. 3) These point parameters balance the removal rate and process stability.

[0025] The second-step polishing parameter range (polishing head pressure 50±5 g / cm², holding ring pressure 140±10 g / cm², polishing head speed 49±5 rpm, polishing pad speed 51±5 rpm, polishing fluid flow rate 90±10 ml / min) collectively defines the working range for selective correction. The polishing head pressure prioritizes the removal of protrusions, the holding ring pressure provides edge constraint without compromising selectivity, and the speed ratio ensures uniform coverage of the abrasive grain trajectory and sufficient chemical reaction. Exceeding this range will fundamentally alter the correction mechanism. For example, pressure > 60 g / cm² enters the linear region, resulting in loss of selectivity; speed > 55 rpm leads to excessive mechanical force and decreased selectivity; holding ring pressure < 130 g / cm² results in edge instability, and > 150 g / cm² introduces new non-uniformity.

[0026] Optionally, the second polishing machine can be fine-tuned within a fixed range during operation, wherein only the polishing head pressure is adjusted while other parameters remain unchanged.

[0027] Pressure is the most sensitive parameter affecting the removal rate. Fine-tuning the pressure within the range of 50±5 g / cm² can compensate for changes in the removal rate caused by factors such as polishing pad aging and batch differences in polishing solution, allowing the target removal amount to be accurately achieved without altering the physical mechanism of selective correction.

[0028] If the first step is fixed to the point value parameter, and the second step is limited to the range parameter and only some parameters are fine-tuned, a collaborative mode of "stable benchmark + flexible correction" is formed: 1) The point value of the first step provides a stable and repeatable "thick fingerprint" as a benchmark target for compensation; 2) The range of the second step provides an adjustable "correction capability" to adapt to production fluctuations; 3) Only some parameters are fine-tuned, which can achieve precise control while keeping the correction mechanism unchanged.

[0029] In this application, by fixing the first step parameter to a point value and limiting the second step parameter to a specific range with only minor adjustments to key parameters, the best balance between correction accuracy, process stability, and operational flexibility is achieved. This is the preferred implementation method for the actual industrial application of this technical solution. After repeated experimental verification, this operation method has obvious advantages.

[0030] Optional, when T flat -T global If the thickness is greater than 8nm, the flat-edge region is considered to be systematically thicker. When T flat -T global If the thickness is less than -8nm, the flat-edge region is considered to be systematically thin. Among them, T global This represents the average global film thickness of the wafer after polishing. This represents the average thickness of the film at the flat edge after polishing.

[0031] Optional, when T flat -T global If the thickness is greater than 10nm, the flat-edge region is considered to be systematically thicker. When T flat -T global If the thickness is less than -10nm, the flat edge region is considered to be systematically thin.

[0032] Optionally, when 40nm > T flat -T global If the thickness is greater than 10nm, the flat-edge region is considered to be systematically thicker. When 40nm < T flat -T global If the thickness is less than -10nm, the flat edge region is considered to be systematically thin.

[0033] For the lower limit of 10nm in this application, if the absolute value of the deviation is too small ( This indicates that the flat edge uniformity of the wafer is at an acceptable level. At this point, conventional polishing with a single machine can meet the process requirements, without the need to initiate a dual-machine collaborative correction process. Introducing a second polishing step not only increases process complexity but also reduces production efficiency. Therefore, setting a lower limit can avoid unnecessary operations and ensure that this method is only applicable to equipment with significant regional deviations.

[0034] For the upper limit of 40nm in this application: This method is based on the principle of linear superposition of the inherent deviations of the two devices, and achieves complementary compensation by adjusting the time ratio of the two polishing steps. The effectiveness of this mechanism requires that the deviation presented in the first polishing step ( The deviation between the second polishing step and the first polishing step ( The magnitudes are comparable, and the final deviation can be controlled within a certain range. If the absolute value of the deviation is too large ( This will lead to the following problems: 1) Excessive deviation is often accompanied by more complex non-uniform distribution, which may exceed the applicable range of the linear superposition model. After correction, the residual deviation is difficult to stably control within the acceptable range. 2) The second polishing step uses low pressure and low speed, with a removal rate only about one-third that of the first step. If the deviation in the first step is too large, the second step needs to handle a larger proportion of the removal, resulting in a significant extension of polishing time and affecting the production line throughput. (Based on the total removal amount...) Taking the removal rate in the second step as an example, when the deviation in the first step exceeds a certain threshold, the time required for the second step often exceeds the threshold as well. This significantly reduces production efficiency. 3) When the absolute value of the deviation exceeds a certain threshold, the corrected inter-wafer variation (CV) increases significantly, and process stability decreases substantially. The judgment range for systematic over-thickness / under-thinness is limited to... This approach eliminates the possibility of producing uniform products that require no correction, while also preventing method failure due to excessive deviation. Within this range, equipment can achieve stable and reliable correction results through two-step synergistic polishing.

[0035] Optional, after correction -5nm < T flat -T global If the value is less than 5nm, the correction is considered complete.

[0036] T flat -T global The closer to 0nm, the better the correction effect. This application adopts polishing machines with complementary characteristics and limits the use of polishing machines that meet specific conditions. By controlling the specific parameters and polishing ratio of the first and second polishing steps and setting the fine-tuning method of the second polishing step, a significantly improved synergistic and complementary polishing effect can be achieved.

[0037] The beneficial effects of this application include, but are not limited to: 1. The wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment of this application utilizes the inherent differences between the equipment for collaborative correction. Its core is not to eliminate the differences, but to actively utilize these differences to transform the "non-uniform fingerprints" of the two equipment into mutually canceling "compensators". It identifies the non-uniform distribution of different equipment and forms complementarity, transforming the inherent differences of equipment in production into usable process compensation resources, thereby fundamentally solving the inherent systematic deviations that cannot be overcome by a single equipment.

[0038] 2. The wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment according to this application can avoid the huge hardware investment of "real-time partitioned pressure" technology, as well as the cumbersome operation and capacity loss of "back-mounted correction pad" technology. Without expensive equipment modification, without introducing additional consumables, and without significantly reducing production throughput, it maximizes the potential of existing production line equipment and provides a highly cost-effective and efficient mass production solution.

[0039] 3. Based on the wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment of this application, a process design sequence of rough polishing and fine polishing is designed. The first polishing step is responsible for the main removal task, and the second polishing step is responsible for the selective correction task. The process parameters of the second polishing step are set specifically to offset the fixed non-uniform distribution of the first polishing step. The optimal proportion of the two polishing steps in the total removal amount is also clearly defined. This allocation ratio is the key to ensuring the correction effect and efficiency.

[0040] 4. The wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment according to this application can systematically, cost-effectively and efficiently overcome the inherent regional non-uniformity defects of a single polishing equipment, and ultimately achieve ultra-high precision global planarization effect in large-scale production, which has important industrial application value. Attached Figure Description

[0041] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a schematic diagram showing the film thickness of an unpolished wafer. Figure 2 This is an image of an unpolished wafer. Figure 3 A schematic diagram showing the wafer film thickness after polishing by the first polishing machine; Figure 4 This is a photograph of the wafer after polishing by the first polishing machine. Figure 5 A schematic diagram showing the wafer film thickness after polishing by the second polishing machine; Figure 6 This is an image of the wafer after polishing by the second polishing machine. Figure 7 A schematic diagram showing the wafer film thickness after co-polishing by the first and second polishing machines; Figure 8 This is a photograph of the wafer after it has been polished by the first and second polishing machines working together. Detailed Implementation

[0042] The present application is described in detail below with reference to the embodiments, but the present application is not limited to these embodiments. Unless otherwise specified, the raw materials and reagents in the embodiments of the present application are all purchased through commercial channels.

[0043] The present application solution will be described below through specific embodiments.

[0044] Example 1 Provide a first polishing machine (equipment A) and a second polishing machine (equipment B). These two machines can be the same model but perform differently due to aging or maintenance history, or they can be different models with inherent characteristics.

[0045] Then, the inherent non-uniformity distribution of the two devices is calibrated: using a large amount of historical production data or specialized test wafers, typical film thickness distribution maps of devices A and B after polishing are accurately measured and established. It is confirmed that device A exhibits a "first non-uniformity distribution" (e.g., systematically thicker in flat-edge regions), while device B exhibits a "second non-uniformity distribution" (e.g., systematically thinner in flat-edge regions).

[0046] For example, one criterion for judging whether a flat-edge region is systematically thinner or thicker is as follows: when T flat -T global If the thickness is greater than 8nm, the flat-edge region is considered to be systematically thicker; when T flat -T global If the thickness is less than -8nm, the flat-edge region is considered to be systematically thin; where T globalThis represents the average global film thickness of the wafer after polishing. This represents the average thickness of the film at the flat edge after polishing. Another standard is as follows: when T... flat -T global If the thickness is greater than 10nm, the flat-edge region is considered to be systematically thicker; when T flat -T global If the thickness is less than -10nm, the flat edge region is considered to be systematically thin.

[0047] These two spatially distributed features (especially in the target correction area, such as the flat edge) are complementary, meaning that one flat edge region in device A and device B is systematically thicker and the other is systematically thinner, which is the basis for the implementation of the proposed solution.

[0048] The coefficient of variation (CV) for thickness removal during polishing is defined as the percentage of the standard deviation of the thickness removal (thickness difference before and after polishing) at each measurement point on the wafer to the average thickness removal. , To remove the standard deviation of the thickness sample.

[0049] The polishing parameters for both equipment A and equipment B were set and tested under the following conditions: polishing head pressure 350 g / cm. 2 Maintain ring pressure of 410 g / cm 2 The polishing head speed was 71 rpm, the polishing pad speed was 68 rpm, and the polishing fluid flow rate was 90 ml / min.

[0050] The parameters of the wafer after polishing to 150nm using only equipment A are shown in Table 1 below: Table 1. Parameters of Equipment A after polishing to 150nm

[0051] The parameters of the wafer after polishing to 150nm using only equipment B are shown in Table 2 below: Table 2 Parameters after polishing at 150nm using Equipment B

[0052] It is determined that devices A and B have complementary characteristics and can be used subsequently.

[0053] 3) Calculate the polishing time based on the above wafer polishing characteristic data.

[0054] For device A, the average T flat -T global =ΔA=12.334(nm); For device B, the average T flat -T global =ΔB=-39.806(nm).

[0055] Assume that the polishing time ratio of device A is α (corresponding to the removal rate ratio), and the ratio of device B is β, satisfying: .

[0056] The thickness difference of the flat edge after two polishing steps is: .

[0057] At this time, it is required that Solving for: 0.6676 < α < 0.8593 0.1407 < β < 0.3324.

[0058] To balance efficiency and stability, a middle value is chosen: .

[0059] at this time It fully meets the requirements.

[0060] 4) Polish the wafer according to the polishing parameters and polishing time above to complete the correction process.

[0061] Based on the inherent properties of equipment A and equipment B as described above, when processing other wafers, equipment A and equipment B can be directly used according to the required thickness. For example, if equipment A alone requires time to polish to the target thickness... Then the first polishing time is set to If equipment B is used alone to polish to the target thickness, it will take time. Then the polishing time for the second step is set to , and , respectively, represent the time (or equivalent time) required for equipment A and equipment B to polish the wafer to the target thickness individually. This approach directly utilizes the inherent non-uniformity fingerprints of the two devices, achieving complementary cancellation through precise control of the time ratio of the two steps. No other parameters need to be adjusted, making it easy to implement and reproducible.

[0062] The specific plan is as follows: Step 1: Polishing, removal of the main material. Polishing is performed according to the first set of polishing parameters set in the process. Polishing head pressure: 350g / cm. 2 Maintain ring pressure of 410 g / cm 2 The polishing head rotates at 71 rpm, the polishing pad rotates at 68 rpm, and the polishing slurry flow rate is 90 ml / min. Under these conditions, removing a 150 nm thick layer of SiO2 takes 20 seconds, i.e., the removal rate is 7.5 nm / s. Therefore, according to the above formula, the polishing time for this step is t = 0.763TA ≈ 15 seconds. The wafer is then transferred to polishing machine B for the second polishing step. Maintaining the above process parameters, the time is t = 0.237TB ≈ 5 seconds.

[0063] like Figure 1 and Figure 2 The images show a schematic diagram of the film thickness on an unpolished wafer and an actual photograph of an unpolished wafer. Figure 3 and Figure 4 The images show a schematic diagram of the wafer film thickness after polishing using equipment A and an actual image of the polished wafer. Figure 5 and Figure 6 The images show a schematic diagram of the wafer film thickness after polishing using device B and an actual image of the polished wafer.

[0064] And such Figure 7 and Figure 8 This diagram shows the wafer film thickness after co-polishing by equipment A+B, and an actual image of the polished wafer.

[0065] The experimental results of this scheme are shown in Table 3 below.

[0066] Table 3. Test results with consistent parameters for the first and second polishing steps.

[0067] Example 2 Based on the scheme of Example 1, optimization was carried out, specifically by reducing the polishing pressure and rotation speed of the second polishing device B, controlling them within a reasonable range. Specifically, the second polishing device B operated under the following parameters: polishing head pressure 50g / cm. 2 Maintain ring pressure of 140 g / cm 2 The polishing head speed was 49 rpm, the polishing pad speed was 51 rpm, and the polishing fluid flow rate was 90 ml / min. The removal rate was measured to be 0.75 nm / s. The parameters of the first polishing step remained unchanged: polishing head pressure 350 g / cm³. 2 Maintain ring pressure of 410 g / cm 2 The polishing head speed was 71 rpm, the polishing pad speed was 68 rpm, and the polishing fluid flow rate was 90 ml / min.

[0068] Then calculate the polishing time required for equipment A and equipment B corresponding to a removal amount of 150nm: According to the formula and Calculations yielded The value ranges from 137.76nm to 144.81nm. The value ranges from 5.19nm to 12.24nm.

[0069] Select , The polishing time can be calculated as follows: , .

[0070] The goal of the adjustment scheme in Example 2 is to correct the first non-uniform distribution caused by the first polishing step through the second polishing step. Because device B has the characteristic of having excessively thin flat edges, it will preferentially act on the excessively thick flat edges caused by device A during the polishing process, ultimately correcting them into a uniform global contour.

[0071] While maintaining the time-ratio allocation as the main control variable, limited and purposeful adjustments can be made to key parameters such as pressure and flow rate in the second polishing step, resulting in a better process window and robustness.

[0072] The experimental results of this scheme are shown in Table 4 below: Table 4. Test results of inconsistent parameters between the first and second polishing steps.

[0073] Based on the above results, it can be seen that the optimization of the wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment has a significant improvement in effect compared with Example 1.

[0074] The above description is merely an embodiment of this application, and the scope of protection of this application is not limited to these specific embodiments, but is determined by the claims of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the technical concept and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for correcting wafer film thickness uniformity based on dual chemical mechanical polishing equipment, characterized in that, The method includes the following steps: 1) Provide a first polishing machine and a second polishing machine, wherein the first polishing machine exhibits a systematic thickening characteristic in the flat edge area, and the second polishing machine exhibits a systematic thinning characteristic in the flat edge area; 2) The first polishing machine polishes the wafer with the following parameters: polishing head pressure 350±50g / cm. 2 Maintain ring pressure of 410±40g / cm 2 The polishing head speed is 71±5 rpm, the polishing pad speed is 68±5 rpm, and the polishing fluid flow rate is 90±10 ml / min. 3) The second polishing machine polishes the wafer with the following parameters: polishing head pressure 50±5g / cm. 2 Maintain ring pressure at 140±10g / cm 2 Polishing head speed 49±5 rpm, polishing pad speed 51±5 rpm, polishing fluid flow rate 90±10 ml / min; The target removal amount in step 2) of the polishing operation accounts for 70-90% of the total target removal amount in steps 2) and 3).

2. The wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment according to claim 1, characterized in that, The target removal amount of the polishing operation in step 2) accounts for 75-85% of the total target removal amount of the polishing operations in steps 2) and 3).

3. The wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment according to claim 2, characterized in that, The target removal amount of the polishing operation in step 2) accounts for 80% of the total target removal amount of the polishing operations in steps 2) and 3).

4. The wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment according to claim 1, characterized in that, The polishing time of the first polishing machine and the second polishing machine is calculated and determined based on the target removal amount of the polishing operation.

5. The wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment according to claim 1, characterized in that, The parameters for polishing the wafer by the first polishing machine are as follows: polishing head pressure 350g / cm. 2 Maintain ring pressure of 410 g / cm 2 The polishing head speed was 71 rpm, the polishing pad speed was 68 rpm, and the polishing fluid flow rate was 90 ml / min. The parameters for polishing the wafer by the second polishing machine are as follows: polishing head pressure 50±5g / cm 2 Maintain ring pressure at 140±10g / cm 2 Polishing head speed 49±5 rpm, polishing pad speed 51±5 rpm, polishing fluid flow rate 90±10 ml / min.

6. The wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment according to claim 1, characterized in that, When T flat -T global If the thickness is greater than 8nm, the flat-edge region is considered to be systematically thicker. When T flat -T global If the thickness is less than -8nm, the flat-edge region is considered to be systematically thin. Among them, T global This represents the average global film thickness of the wafer after polishing. This represents the average thickness of the film at the flat edge after polishing.

7. The wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment according to claim 6, characterized in that, When T flat -T global If the thickness is greater than 10nm, the flat-edge region is considered to be systematically thicker. When T flat -T global If the thickness is less than -10nm, the flat edge region is considered to be systematically thin.

8. The wafer film thickness uniformity correction method based on dual chemical mechanical polishing equipment according to claim 7, characterized in that, After correction -5nm < T flat -T global If the value is less than 5nm, the correction is considered complete.

Images