Wafer polishing laser interference closed loop pressure control method

Abstract

The application discloses a wafer polishing laser interference closed-loop pressure control method, and relates to the technical field of semiconductor manufacturing, and comprises the following steps: S1: a ceramic plate is triggered to automatically measure a program after being turned over with a wafer; S2: a laser interference measurement head is driven by a six-axis mechanical arm positioning system to non-contact scan T2 / T4 points of the wafer, and thickness distribution data is acquired; S3: a Zernike polynomial is adopted to fit the thickness distribution data to generate a three-dimensional distribution graph, and a pressure correction amount is calculated based on a pressure prediction model constructed by a process kinetics model and a support vector machine regression algorithm; S4: an abnormal processing module is used to detect data outliers in real time, and a device protection strategy is triggered when a deviation exceeds a set threshold; and S5: a pressure control instruction is transmitted to a polishing head by a PLC controller through a PROFINET bus. By means of non-contact laser interference measurement, the laser interference technology is applied to wafer thickness detection after polishing for the first time, and single measurement time is less than 20 seconds, so that the detection efficiency is significantly improved.

Description

Technical Field

[0001] This invention relates to the field of semiconductor manufacturing technology, and more specifically to a closed-loop pressure control method for laser interference in wafer polishing. Background Technology

[0002] Wafer polishing is a process that uses mechanical grinding and chemical action to remove unevenness, defects, and impurities from the surface of a silicon wafer, making the wafer surface flat, smooth, and defect-free. Wafer polishing is a crucial surface treatment process in semiconductor manufacturing, commonly used in the production of chips, LEDs, and solar cells. The main purpose of wafer polishing is to ensure the planarization of the wafer surface to meet the demands of manufacturing high-precision chips. During chip manufacturing, especially in the photolithography stage, the wafer surface must be extremely flat because as chip manufacturing processes shrink, the imaging resolution requirements of photolithography machines become increasingly higher, and the depth of focus decreases. The surface undulations of the wafer must be within the depth of focus to ensure the clarity of the lithographic image. However, existing technologies have the following problems:

[0003] In traditional wafer polishing processes, thickness measurement relies on manual measurement of T2 / T4 points using a dial indicator, and manual calculation and adjustment of equipment pressure parameters. This method has the following problems:

[0004] 1. Response delay: Each adjustment takes 5-10 minutes, resulting in low process efficiency;

[0005] 2. Risk of human error: A deviation rate of ±0.5μm exceeding 15% will affect product consistency;

[0006] 3. Low equipment utilization rate: Downtime for adjustment accounts for 8% to 12%, resulting in serious waste of resources.

[0007] The aforementioned problems result in high costs and low precision in wafer polishing processes, making it difficult to meet the demands of high-precision semiconductor manufacturing. Summary of the Invention

[0008] This invention provides a closed-loop pressure control method for laser interference in wafer polishing to solve the problems mentioned in the background art.

[0009] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0010] A closed-loop pressure control method for laser interferometry in wafer polishing includes the following steps:

[0011] S1: Automatic measurement program is triggered after the ceramic plate and wafer are flipped;

[0012] S2: The six-axis robotic arm positioning system drives the laser interferometer head to perform non-contact scanning of the T2 / T4 points on the wafer to obtain thickness distribution data;

[0013] S3: The thickness distribution data is fitted with Zernike polynomials to generate a three-dimensional distribution map, and the pressure correction is calculated based on the pressure prediction model constructed by the process dynamics model and the support vector machine regression algorithm.

[0014] S4: Real-time detection of outliers in data via the anomaly handling module; triggering device protection strategies when the deviation exceeds a set threshold.

[0015] S5: Transmits pressure control commands to the polishing head via the PLC controller and PROFINET bus;

[0016] S6: The polishing head performs pressure adjustment and provides feedback confirmation signal. The process parameters and adjustment status are displayed in real time through the industrial-grade HMI interactive terminal, completing closed-loop control.

[0017] A further improvement of the technical solution of the present invention is that the laser interferometric measuring head adopts the Michelson interferometry principle, the measurement accuracy is ±0.1μm, the single non-contact scanning time is less than 20 seconds, and the scanning range covers 20% to 80% of the wafer diameter.

[0018] A further improvement of the technical solution of the present invention is that the six-axis robotic arm positioning system is driven by a servo motor and equipped with a grating ruler feedback device, with a positioning accuracy of ±0.05mm, and can realize the attitude adjustment of the measuring head within a ±1° tilt range on the wafer surface.

[0019] A further improvement of the technical solution of the present invention is that: the pressure prediction model is trained on 1000 sets of historical process parameters and corresponding thickness data by means of support vector machine regression algorithm. The process parameters include polishing pressure, rotation speed and slurry concentration. Radial basis function is used as kernel function and the parameters are optimized so that the prediction error is less than ±0.15μm.

[0020] A further improvement of the technical solution of the present invention is that: the anomaly handling module uses the 3σ principle to detect outliers in the data. When the deviation between the measured thickness and the predicted thickness exceeds ±0.1μm, the polishing head pressure locking mechanism is automatically triggered, and an audible and visual alarm is issued through the HMI terminal.

[0021] A further improvement of the technical solution of the present invention is that the PLC controller and the polishing head communicate in real time through PROFINETIO, adopting IRT mode, with a command transmission delay of less than 50ms, a communication cycle of 1ms, and support for data verification and retransmission mechanisms.

[0022] A further improvement of the technical solution of the present invention is that the industrial-grade HMI interactive terminal is equipped with a 10.1-inch touch screen, which displays the laser measurement three-dimensional cloud map, pressure correction curve and equipment OEE in real time, and supports historical data query and process parameter formula storage functions.

[0023] A further improvement of the technical solution of the present invention is that the Zernike polynomial fitting algorithm adopts a 36th-order polynomial, and the fitting error of the wafer thickness data is less than 0.08μm, which can identify local thickness anomaly regions with a diameter greater than 50μm.

[0024] Due to the adoption of the above technical solution, the technical progress achieved by this invention compared to the prior art is as follows:

[0025] 1. This invention provides a closed-loop pressure control method for wafer polishing laser interferometry. Through non-contact laser interferometry measurement, laser interferometry technology is applied to the detection of wafer thickness after polishing for the first time. The single measurement time is less than 20 seconds, which significantly improves the detection efficiency.

[0026] 2. This invention provides a closed-loop pressure control method for laser interferometry in wafer polishing. Through an adaptive pressure control algorithm, based on a process dynamics model, the pressure correction amount is calculated in real time, improving thickness uniformity by 35%, achieving laser interferometry measurement accuracy of ±0.1μm, and reducing the deviation rate to <5%, thus realizing high-precision detection.

[0027] 3. This invention provides a closed-loop pressure control method for laser interference in wafer polishing. Through digital twin pre-optimization, a pressure prediction model is constructed using a support vector machine regression algorithm to optimize process parameters, reduce manual intervention by more than 90%, and achieve full-process closed-loop control from detection to pressure adjustment, thereby improving equipment uptime. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0029] The present invention will be further described in detail below with reference to embodiments:

[0030] Example 1

[0031] like Figure 1 As shown, this invention provides a closed-loop pressure control method for laser interferometry in wafer polishing, comprising the following steps:

[0032] S1: Automatic measurement program is triggered after the ceramic plate and wafer are flipped;

[0033] S2: The six-axis robotic arm positioning system drives the laser interferometer head to perform non-contact scanning of the T2 / T4 points on the wafer to obtain thickness distribution data;

[0034] S3: The thickness distribution data is fitted with Zernike polynomials to generate a three-dimensional distribution map, and the pressure correction is calculated based on the pressure prediction model constructed by the process dynamics model and the support vector machine regression algorithm.

[0035] S4: Real-time detection of outliers in data via the anomaly handling module; triggering device protection strategies when the deviation exceeds a set threshold.

[0036] S5: Transmits pressure control commands to the polishing head via the PLC controller and PROFINET bus;

[0037] S6: The polishing head performs pressure adjustment and provides feedback confirmation signal. The process parameters and adjustment status are displayed in real time through the industrial-grade HMI interactive terminal, completing closed-loop control.

[0038] Example 2

[0039] like Figure 1 As shown, based on Embodiment 1, the present invention provides a technical solution: Preferably, the laser interferometric measuring head adopts the Michelson interferometry principle, with a measurement accuracy of ±0.1μm, a single non-contact scanning time of less than 20 seconds, and a scanning range covering 20% ​​to 80% of the wafer diameter. The six-axis robotic arm positioning system is driven by a servo motor and equipped with a grating ruler feedback device, achieving a positioning accuracy of ±0.05mm, and can realize attitude adjustment of the measuring head within a ±1° tilt range on the wafer surface.

[0040] In this embodiment, the laser interferometer head incorporates a semiconductor laser with a wavelength of 635nm±5nm, a beam splitter, and a charge-coupled device detector. The measurement optical path forms a 0.1mm diameter spot through a collimating lens, ensuring that the wafer surface is not damaged during non-contact measurement. Through a temperature compensation algorithm with a compensation range of 20℃±5℃ and an optical path calibration mechanism, a thickness measurement accuracy of ±0.1μm is achieved, with a single scan time of ≤18 seconds and a scanning area covering 25% to 75% of the wafer diameter in a ring-shaped region.

[0041] Example 3

[0042] like Figure 1 As shown, based on Embodiment 1, the present invention provides a technical solution: Preferably, the pressure prediction model is trained on 1000 sets of historical process parameters and corresponding thickness data using a support vector machine regression algorithm. The process parameters include polishing pressure, rotation speed, and slurry concentration. A radial basis function is used as the kernel function, and the parameters are optimized to make the prediction error less than ±0.15μm. The anomaly handling module uses the 3σ principle to detect outliers in the data. When the deviation between the measured thickness and the predicted thickness exceeds ±0.1μm, the polishing head pressure locking mechanism is automatically triggered, and an audible and visual alarm is issued through the HMI terminal. The PLC controller and the polishing head communicate in real time through PROFINETIO, using IRT mode, with a command transmission delay of less than 50ms and a communication cycle of 1ms, supporting data verification and retransmission mechanisms.

[0043] In this embodiment, historical process data includes parameters such as polishing pressure (1-5 bar), rotation speed (50-200 rpm), and slurry concentration (5%-15%). The kernel function is a radial basis function with parameter γ = 0.5 and penalty factor C = 10. Based on the process dynamics model, the pressure correction is calculated in real time with a correction step of 0.05 bar to ensure thickness uniformity improvement of ≥35%. Simultaneously, after the anomaly handling module is triggered, the system automatically performs the following operations: the polishing head pressure is locked at the current value ±5%; the HMI terminal issues an audible and visual alarm, including an 85dB buzzer and a red warning light; an anomaly log is generated, including timestamps, deviation data, and equipment status. The PLC controller is a Siemens S7-1500 series PLC equipped with a PROFINETIO controller module, supporting IRT communication mode with a command transmission delay ≤40ms. PROFINET communication uses a 1ms periodic synchronization mechanism with a data transmission rate of 100Mbps. Redundant power supplies and shielded cables ensure communication anti-interference capabilities.

[0044] Example 4

[0045] like Figure 1 As shown, based on Embodiment 1, the present invention provides a technical solution: Preferably, the industrial-grade HMI interactive terminal is equipped with a 10.1-inch touch screen, which displays the laser measurement three-dimensional cloud map, pressure correction curve and equipment OEE in real time, supports historical data query and process parameter recipe storage functions, and the Zernike polynomial fitting algorithm adopts a 36th-order polynomial, with a fitting error of less than 0.08μm for wafer thickness data, and can identify local thickness anomaly areas with a diameter greater than 50μm.

[0046] In this embodiment, the Zernike polynomial fitting algorithm fits the 200×200 sampling point data on the wafer surface with a fitting error ≤0.07μm. It can identify local thickness anomaly regions with a diameter ≥50μm and generate an STL format three-dimensional model based on the fitted data. The thickness deviation is displayed intuitively through color gradation mapping with a color gradation accuracy of 0.1μm.

[0047] The working principle of this wafer polishing laser interference closed-loop pressure control method will be explained in detail below.

[0048] like Figure 1As shown, the laser beam is split into measurement light and reference light by a beam splitter. The measurement light is reflected after illuminating the wafer surface and forms interference fringes with the reference light on the CCD surface. The wafer thickness change Δd is calculated by the relationship between the fringe displacement Δx and the laser wavelength λ: Δd = Δx·λ / 2, achieving sub-micron level precision detection. The support vector machine maps process parameters to a high-dimensional space through the RBF kernel function to construct a nonlinear relationship model between pressure and thickness. When the current measured thickness data is input, the model outputs the optimal polishing pressure value, reducing the manual intervention rate by ≥90%. The system establishes a polishing pressure-material removal rate model based on Newton's second law: dH / dt = K·P·v, where K is the process constant, P is the pressure, and v is the polishing head rotation speed. The thickness H is dynamically controlled by real-time correction of the pressure P, ensuring that the equipment utilization rate is ≥95%.

[0049] The present invention has been described in detail above. However, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, any modifications or improvements that do not depart from the spirit of the present invention are within the scope of protection of the present invention.

Claims

1. A closed-loop pressure control method for laser interferometry in wafer polishing, characterized in that: Includes the following steps: S1: Automatic measurement program is triggered after the ceramic plate and wafer are flipped; S2: The six-axis robotic arm positioning system drives the laser interferometer head to perform non-contact scanning of the T2 / T4 points on the wafer to obtain thickness distribution data; S3: The thickness distribution data is fitted with Zernike polynomials to generate a three-dimensional distribution map, and the pressure correction is calculated based on the pressure prediction model constructed by the process dynamics model and the support vector machine regression algorithm. S4: Real-time detection of outliers in data via the anomaly handling module; triggering device protection strategies when the deviation exceeds a set threshold. S5: Transmits pressure control commands to the polishing head via the PLC controller and PROFINET bus; S6: The polishing head performs pressure adjustment and provides feedback confirmation signal. The process parameters and adjustment status are displayed in real time through the industrial-grade HMI interactive terminal, completing closed-loop control.

2. The wafer polishing laser interferometry closed-loop pressure control method according to claim 1, characterized in that: The laser interferometer head adopts the Michelson interferometry principle, with a measurement accuracy of ±0.1μm, a single non-contact scan time of less than 20 seconds, and a scanning range covering 20% ​​to 80% of the wafer diameter.

3. The wafer polishing laser interferometry closed-loop pressure control method according to claim 1, characterized in that: The six-axis robotic arm positioning system is driven by a servo motor and equipped with a grating ruler feedback device, achieving a positioning accuracy of ±0.05mm. It can adjust the posture of the measuring head within a ±1° tilt range on the wafer surface.

4. The wafer polishing laser interferometry closed-loop pressure control method according to claim 1, characterized in that: The pressure prediction model is trained on 1000 sets of historical process parameters and corresponding thickness data using a support vector machine regression algorithm. The process parameters include polishing pressure, rotation speed, and slurry concentration. A radial basis function is used as the kernel function, and the parameters are optimized to make the prediction error less than ±0.15μm.

5. The wafer polishing laser interferometry closed-loop pressure control method according to claim 1, characterized in that: The anomaly handling module uses the 3σ principle to detect outliers in the data. When the measured thickness deviates from the predicted thickness by more than ±0.1μm, the polishing head pressure locking mechanism is automatically triggered, and an audible and visual alarm is issued through the HMI terminal.

6. The wafer polishing laser interferometry closed-loop pressure control method according to claim 1, characterized in that: The PLC controller and the polishing head communicate in real time via PROFINETIO, using IRT mode. The instruction transmission delay is less than 50ms, the communication cycle is 1ms, and it supports data verification and retransmission mechanisms.

7. The wafer polishing laser interferometry closed-loop pressure control method according to claim 1, characterized in that: The industrial-grade HMI interactive terminal is equipped with a 10.1-inch touch screen, which displays the laser measurement 3D cloud map, pressure correction curve and equipment OEE in real time, and supports historical data query and process parameter formula storage functions.

8. The wafer polishing laser interferometry closed-loop pressure control method according to claim 1, characterized in that: The Zernike polynomial fitting algorithm uses a 36th-order polynomial and has a fitting error of less than 0.08 μm for wafer thickness data. It can identify local thickness anomaly regions with a diameter greater than 50 μm.

Images