Wafer cleaning nozzle efficiency quantitative characterization method and system based on fluorescence tracing imaging technology

Abstract

The invention discloses a wafer cleaning nozzle efficiency quantitative characterization method and system based on a fluorescent tracing imaging technology, and belongs to the technical field of semiconductor cleaning. The method comprises the following steps: uniformly mixing a pollutant standard solution with controllable viscosity and a fluorescent tracer agent to form a pollutant simulation solution capable of being quantitatively traced; spin-coating quantitative pollutant simulation liquid on the surface of the test substrate based on spin-coating equipment to form a uniform pollution layer; a to-be-evaluated cleaning nozzle is adopted to carry out directional cleaning on a preset polluted substrate, a fluorescence tracer is excited through an excitation light source, and a fluorescence signal distribution image in the cleaning process is collected through a charge-coupled device camera imaging system; the pollutant residual rate is accurately calculated through an image processing algorithm, and a quantitative evaluation index of the nozzle cleaning efficiency is established. According to the invention, visual monitoring of the whole cleaning process and accurate quantitative evaluation of the cleaning efficiency are realized, and a scientific quantitative basis is provided for parameter optimization and process verification of high-precision cleaning equipment.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor cleaning technology, and in particular to a method and system for quantitatively characterizing the efficiency of a wafer cleaning nozzle based on fluorescent tracer imaging technology, which is particularly suitable for parameter optimization and process verification of nozzles in a dual-fluid nozzle wet cleaning process. Background Art

[0002] During the semiconductor manufacturing process, the cleanliness of the wafer surface will directly affect the performance, reliability and yield of the device, so the cleaning of its surface contaminants is a key process link in the entire process.

[0003] Traditional wafer cleaning technologies, such as RCA (Radio Corporation of America Cleaning) cleaning and ultrasonic cleaning, generally remove contaminants through chemical solution reactions, physical and mechanical effects, or a combination of the two. However, with the continuous advancement of process technology nodes, the complexity of the cleaning process has increased significantly, posing unprecedented challenges to contaminant residue rates, wafer surface structural damage, and surface roughness. Traditional cleaning technologies have gradually revealed their limitations in addressing these challenges, driving the industry to transition to more efficient gas-liquid two-phase nozzle cleaning technology. As a core component of cleaning equipment, the spray performance of the cleaning nozzle directly affects cleaning efficiency and chip yield. In this context, how to accurately characterize the cleaning efficiency of the gas-liquid two-phase nozzle and establish a correlation model between process parameters and cleaning efficiency has become a major challenge in optimizing the cleaning process and improving yield.

[0004] Methods for characterizing the cleaning efficiency of wafer cleaning nozzles primarily rely on physical methods to assess nozzle cleaning efficiency through optical inspection techniques such as laser scattering particle detection and surface analysis. However, these methods often rely on complex and costly instrumentation and are unable to capture the uniformity of cleaning fluid coverage and contaminant removal efficiency under the nozzle's dynamic operating state. This makes them unsuitable for applications requiring full-process, high-frequency evaluation of nozzle cleaning efficiency to optimize nozzle performance.

[0005] In summary, developing a low-cost, high-efficiency quantitative characterization method and system for cleaning nozzle cleaning efficiency that can monitor the entire cleaning process, and establishing a correlation model between process parameters and cleaning efficiency, has become an important way to promote wafer cleaning nozzle parameter optimization and process verification. Summary of the Invention

[0006] The purpose of the present invention is to overcome the limitations of the above-mentioned existing technologies and propose a method and system for quantitative characterization of wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology. On the basis of lightweight system, the standardized quantitative characterization of the cleaning efficiency of wafer cleaning nozzles is realized with high accuracy and good repeatability, providing a scientific quantitative basis for parameter optimization and process verification of nozzles in the dual-fluid nozzle wet cleaning process.

[0007] The technical solution of the present invention is: The pollutant standard solution and the fluorescent tracer are uniformly mixed to prepare a pollutant simulation solution that can be quantitatively traced; A spin coating device is used to spin-coat a predetermined amount of pollutant simulation liquid on the surface of the test substrate to form a uniform pollutant layer with controllable thickness; The cleaning nozzle to be evaluated is used to perform targeted cleaning on the pre-contaminated substrate. The fluorescent tracer is excited by an excitation light source of a specific wavelength, and a high-resolution charge-coupled device camera imaging system is used to capture the distribution image of the fluorescent signal during the cleaning process. Based on the image processing algorithm, the area where the fluorescence signal response is lower than the threshold is defined as the clean area. The pollutant residual rate is further calculated based on the area of ​​the clean area and the preset contaminated area, which is used as a quantitative evaluation indicator of the nozzle cleaning efficiency.

[0008] In the above-mentioned method and system for quantitative characterization of wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology, the pollutant standard liquid is a dimethyl silicone oil solution, and the fluorescent tracer is an oily color paste substance.

[0009] In the above-mentioned method and system for quantitative characterization of wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology, the spin coating equipment is a brushless motor with a maximum speed of 12,000 r / min and controllable parameters such as the spin coating speed and time, and the test substrate is a polyvinyl chloride plate, a metal plate or a substrate of other different materials.

[0010] In the above-mentioned method and system for quantitative characterization of wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology, the cleaning nozzle to be evaluated is an internally mixed gas-liquid two-phase nozzle with nitrogen as the gas path and deionized water as the liquid path, the excitation light source of the specific wavelength is a 365 nm ultraviolet light source, and the high-resolution charge-coupled device camera imaging system includes an ultraviolet light source, a high-resolution charge-coupled device camera, an optical imaging lens and a data processing terminal.

[0011] In the above-mentioned method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology, the image processing algorithm includes the following steps: (1) Filter the collected cleaning process images to suppress image noise while preserving the boundary features of the cleaning area as much as possible; (2) Thresholding the filtered image is performed, and the area where the fluorescence signal response is lower than the threshold is defined as the clean area; (3) Perform morphological processing on the thresholded image to remove internal holes and edge burrs generated after thresholding; (4) Performing contour detection on the image after morphological processing to accurately identify the contour of the clean area and calculate the area of ​​the clean area, and further calculating the pollutant residual rate based on the area of ​​the clean area and the preset polluted area.

[0012] In the above-mentioned method and system for quantitative characterization of wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology, the calculation formula for the pollutant residual rate R, a quantitative evaluation indicator of nozzle cleaning efficiency, is as follows: Where: A T The area of ​​the preset contaminated area; A C is the area of ​​the cleaning area.

[0013] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention constructs a lightweight detection system architecture. By integrating a UV excitation light source, a charge-coupled device camera, an optical imaging lens, and a data processing terminal, an optical imaging system with high spatial resolution is constructed. This system can achieve quantitative evaluation of the cleaning nozzle efficiency without relying on complex high-precision detection instruments, providing a low-cost, highly compatible process evaluation technology path for the semiconductor manufacturing field. (2) The present invention demonstrates a high degree of process adaptability and compatibility with application scenarios, and constructs a cleaning efficiency characterization system with multi-parameter coordinated regulation. It can compare and evaluate the cleaning efficiency of nozzles under different fluid mechanics parameters, surface interface characteristics and nozzle geometric characteristic parameters, providing common technical support for the cleaning efficiency evaluation of strategic industries such as semiconductors, biomedicine, and new energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a flow chart of a method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology of the present invention. DETAILED DESCRIPTION

[0015] The present invention will be further described below with reference to the accompanying drawings.

[0016] Figure 1 This is a flow chart of a method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology of the present invention, which mainly includes the following processes: Process 1: Prepare pollutant simulation solution based on mixing pollutant standard solution with fluorescent tracer; Process 2: A quantitative and controllable contamination layer construction method based on spin coating technology; Process 3: Capturing dynamic images of the cleaning process using a high-resolution imaging system; Process 4: Constructing a quantitative evaluation index for cleaning efficiency based on image processing algorithm.

[0017] The specific implementation of Process 1 is as follows: a dimethyl silicone oil solution of a specific viscosity that meets the required adhesion strength and an oil-based colorant fluorescent tracer with good solubility and dispersibility in the dimethyl silicone oil solution are selected. Depending on the viscosity of the dimethyl silicone oil, the dimethyl silicone oil solution and the fluorescent tracer are uniformly mixed in a specific volume ratio using either a stirring rod or a magnetic stirrer to form a pollutant simulant solution that can be quantitatively tracked.

[0018] The specific implementation of process 2 is as follows: a clean test substrate and motor specifications that meet the requirements are selected, and the test substrate is fixed to a rotating motor. A syringe or pipette is used to control the amount of contaminant simulating liquid applied each time during spin coating, and a motor controller is used to ensure consistent spin coating speed and duration. Ultimately, a contamination layer of uniform thickness and distribution is formed on the surface of the test substrate, resulting in the predetermined contaminated substrate.

[0019] The specific implementation of process 3 is as follows: The presumed contaminated substrate is placed at a specific distance from the nozzle outlet and directionally cleaned using the internally mixed gas-liquid two-phase nozzle to be evaluated. An imaging system consists of a 365 nm ultraviolet excitation light source, a high-resolution charge-coupled device (CCD) camera, an optical imaging lens, and a data processing terminal. The UV light source is turned on to excite fluorescent markers on the surface of the presumed contaminated substrate to emit a fluorescent signal. A synchronizer controls the opening of a solenoid valve, triggering the high-resolution CCD camera to capture images of the fluorescent signal distribution during the cleaning process.

[0020] The specific implementation of process 4 is as follows: (1) Filter the collected cleaning process images to suppress image noise while preserving the boundary features of the cleaning area as much as possible; (2) Thresholding the filtered image is performed, and the area where the fluorescence signal response is lower than the threshold is defined as the clean area; (3) Perform morphological processing on the thresholded image to remove internal holes and edge burrs generated after thresholding; (4) Performing contour detection on the image after morphological processing to accurately identify the contour of the clean area and calculate the area of ​​the clean area, and further calculating the pollutant residual rate based on the area of ​​the clean area and the preset polluted area.

Claims

1. A method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology, characterized in that: The method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology include the following steps: uniformly mixing a contaminant standard solution with controllable viscosity with a fluorescent tracer to form a quantitatively traceable contaminant simulation solution; spin-coating a quantitative amount of the contaminant simulation solution on the surface of a test substrate using a spin coating device to form a uniform contamination layer; using a cleaning nozzle to be evaluated to perform directionally cleaning on a preset contaminated substrate, exciting the fluorescent tracer with an excitation light source, and collecting a fluorescent signal distribution image during the cleaning process using a charge-coupled device camera imaging system; and accurately calculating the contaminant residual rate using an image processing algorithm to establish a quantitative evaluation index for nozzle cleaning efficiency.

2. The method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology according to claim 1, characterized in that: The pollutant standard solution is a dimethyl silicone oil solution that can be fully mixed with the fluorescent tracer and maintain a stable state during the spin coating process. The viscosity range can be selected within the range of 100-10000 cSt according to the adhesion strength requirement.

3. The method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology according to claim 1, characterized in that: The fluorescent tracer is an oily colorant substance that can emit a fluorescent signal with an intensity sufficient to be captured by a camera under the irradiation of an excitation light source of the specific wavelength and has good solubility and dispersibility in the pollutant standard solution.

4. The method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology according to claim 1, characterized in that: The spin coating equipment is a brushless motor with a maximum speed of 12000 r / min and controllable parameters such as the spin coating speed and time. Adjusting the parameters can control the thickness and distribution of the contamination layer formed by the contaminant simulation liquid on the surface of the test substrate.

5. The method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology according to claim 1, characterized in that: The test substrate is a material that does not react with the pollutant standard solution and the fluorescent tracer. The test substrate can be made of polyvinyl chloride, metal or other materials according to actual needs to simulate the pollution conditions on different substrates.

6. The method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology according to claim 1, characterized in that: The cleaning nozzle to be evaluated is an internally mixed gas-liquid two-phase nozzle, the gas path medium is nitrogen, and the liquid path medium is deionized water. During spray cleaning, its cleaning parameters such as the spray pressure (0.3-0.6 MPa), spray distance (10-30 mm), and spray angle (30-90°) are set according to the actual cleaning process requirements.

7. The method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology according to claim 1, characterized in that: The excitation light source of the specific wavelength is a 365 nm ultraviolet light source, the wavelength and intensity of which can effectively excite the fluorescent tracer to emit a fluorescent signal without causing other adverse effects on the test substrate and the pollutant simulation liquid.

8. The method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology according to claim 1, characterized in that: The high-resolution charge-coupled device camera imaging system includes the ultraviolet excitation light source with a wavelength of 365nm, a charge-coupled device camera with a resolution of 2200×2752 pixels, an optical imaging lens and a data processing terminal, which can fully collect high-resolution images during the cleaning process.

9. The method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology according to claim 1, characterized in that: The image processing algorithm comprises the following steps: (1) Filter the collected cleaning process images to suppress image noise while preserving the boundary features of the cleaning area as much as possible; (2) Thresholding the filtered image is performed, and the area where the fluorescence signal response is lower than the threshold is defined as the clean area; (3) Perform morphological processing on the thresholded image to remove internal holes and edge burrs generated after thresholding; (4) Performing contour detection on the image after morphological processing to accurately identify the contour of the clean area and calculate the area of ​​the clean area, and further calculating the pollutant residual rate based on the area of ​​the clean area and the preset polluted area.

10. The method and system for quantitatively characterizing wafer cleaning nozzle efficiency based on fluorescent tracer imaging technology according to claim 1, characterized in that: The calculation formula of the pollutant residual rate R, a quantitative evaluation index of the nozzle cleaning efficiency, is as follows: Where: A T The area of ​​the preset contaminated area; A C is the area of ​​the cleaning area.

Images