Device and method for acquiring image information for semiconductor defect detection
By using short-wave infrared microscopy and polarization imaging technology in semiconductor defect detection, combining specific light sources, microscopes and CCD imaging detectors, the problem of obtaining enough image information is solved, and the rapid identification of defect characteristics of semiconductor materials is achieved and position.
Patent Information
- Application Number
- CN202211030590.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-08-26
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2042-08-26
AI Technical Summary
It is difficult to obtain enough image information for semiconductor defect detection, especially when semiconductor materials are opaque.
Using a device including a specific light source, a microscope and a CCD imaging detector, combined with a polarization detection module, up to 8 image information parameters are obtained through short-wave infrared microscopy and polarization imaging technology.
It can effectively analyze defect characteristics in semiconductor materials, such as cracks, warping, scratches, solid particles, metal pollution and organic pollution, and achieve rapid identification, judgment and positioning.
Smart Images

Figure CN115598059B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a method for detecting semiconductor defects, and in particular to a device and method for acquiring image information for semiconductor defect detection. Background Art
[0002] Semiconductor devices are widely used in various fields and various electronic products. They have become the lifeline of economic development and national information security, and have a profound impact on modern human life. At present, with the development of microelectronics technology and the integrated circuit industry, the number of devices integrated on a chip is still growing, and highly integrated circuits are constantly developing. While the integration of circuits is increasing, it also faces great challenges, not only in semiconductor manufacturing, but also in packaging and testing. The manufacturing process of integrated circuits involves a series of processing technologies in the physical and chemical fields, such as raw material production, cleaning, photolithography, etching film formation, doping, and mechanical planarization. After the wafer is processed, a series of packaging tests must be carried out. In the manufacture of semiconductor integrated circuits, particles or defects will be introduced in each process for various reasons. As the requirements for high integration and high performance of ultra-large-scale integrated circuits gradually increase, semiconductor technology is developing towards smaller feature sizes, and the impact of these particles or defects on the quality of integrated circuits is becoming more and more significant. Defects or cracks in silicon ingots and finished wafers, and surface defects including solid particles, metal pollution, organic pollution, etc., will have a serious impact on the quality of semiconductor products. In addition, during the semiconductor chip packaging and manufacturing process, various defects and even damage will inevitably occur on the chip surface, which will directly affect the chip's operating performance.
[0003] The traditional defect detection method for semiconductor raw materials and chips is manual visual inspection, which uses visible light microscopes and other equipment to identify the quality of wafers based on the experience of technicians. However, this method has the following shortcomings: (1) Human eyes are easily fatigued and cannot be tested continuously. In addition, the test results are easily affected by factors such as human experience and psychological state, resulting in differences in test results; (2) The human eye can only make general good or bad evaluations, but it is difficult to classify and summarize defects, and it is impossible to make substantive suggestions for the improvement of processing technology; (3) Wafers are high-precision materials, and manual operation can easily cause product damage or introduce new defects; (4) Chip manufacturing presents the characteristics of large-scale products, and the current labor cost is high, which will undoubtedly greatly increase the labor cost of enterprises. With the continuous improvement of the integration of semiconductor integrated circuits and the continuous development of microelectronics process technology, the processing technology is constantly improving, and its complexity is also increasing, resulting in more and more diversified sources of defects; the existence of defects has become a bottleneck inhibiting the development of the semiconductor industry. Traditional manual visual inspection methods have been difficult to adapt to the high-speed and high-precision inspection requirements of semiconductor chip manufacturing and packaging. With the development of computer technology and image processing technology, and the deepening of research and application in the field of machine vision, more and more high-end equipment manufacturers have begun to use machine vision technology to perform automated detection of chip defects. For example, Cognex in the United States, VMT in Germany, KEYENCE in Japan, STI in Singapore, etc. have all provided chip surface defect detection devices.
[0004] However, the low contrast, blurred edges, and complex background of chip surface features all pose challenges to machine vision inspection technology. At present, in the analysis of optical parts, the technology of obtaining image information through CCD and then analyzing the defects of optical parts through image fusion has been widely used. However, in order to improve the accuracy of analysis, it is necessary to obtain enough image information. Since most semiconductor materials such as silicon, germanium, gallium arsenide, and indium phosphide are opaque in the visible light / near-infrared light band, it is difficult to obtain enough necessary image information. Summary of the invention
[0005] The technical problem to be solved by the present invention is to provide a device and method for acquiring image information for semiconductor defect detection by acquiring sufficient image information.
[0006] The technical solution adopted by the present invention to solve the above technical problems is: a device for obtaining image information for semiconductor defect detection, comprising:
[0007] A light source for providing lighting;
[0008] A beam expansion and collimation module, used for expanding and collimating the light source;
[0009] A sample stage, used to place samples to be analyzed;
[0010] Microscope lens, used for focusing light and imaging;
[0011] CCD imaging detector, used to receive optical image signals and perform photoelectric conversion;
[0012] Control the computer to calculate and obtain the image information of the sample to be analyzed based on the signal received from the CCD imaging detector;
[0013] The light emitted by the light source is expanded and collimated by the beam expansion and collimation module, and then incident on the sample to be analyzed on the sample stage. The light passing through the sample is converged and imaged by a microscope lens, received by a CCD imaging detector and transmitted to a control computer after photoelectric conversion, and image information of the sample to be analyzed is obtained in the control computer. The sample to be analyzed is a semiconductor raw material or a semiconductor chip. The light source has a short-wave infrared spectrum with a wavelength exceeding 1200nm. The microscope lens is a short-wave infrared microscope lens. A polarization detection module is arranged between the microscope lens and the CCD imaging detector. The polarization detection module consists of an ultra-wideband achromatic 1 / 4 wave plate corresponding to a wavelength of 900nm to 1700nm with an adjustable azimuth angle and a polarizer. The CCD imaging detector is an InGaAs focal plane array detector. The typical values of the main parameters of the detector are: a spectral response range of 900nm to 1700nm, a pixel scale of 640×512, a pixel center distance of 15μm, and a quantum efficiency greater than 70% in the spectral range of 1000nm to 1600nm.
[0014] Preferably, the light source is a tungsten halogen lamp.
[0015] The method for obtaining image information for semiconductor defect detection using the above device comprises the following specific steps:
[0016] S1. Place the sample to be analyzed on the sample stage, the sample to be analyzed is a semiconductor raw material or a semiconductor chip, rotate the ultra-wideband achromatic quarter wave plate in the polarization detection module, adjust the fast axis azimuth to four positions of 0°, 22.5°, 45° and 135°, measure the polarization images of the four positions respectively through the CCD imaging detector and convert them into polarization image electrical signals, recorded as I0, I 22.5 ,I 45 and I 135 , input it into the control computer;
[0017] S2. The control computer polarizes the image electrical signals I0, I 22.5 ,I 45 and I 135 Substitute into the formula
[0018]
[0019] Obtain four Stokes vector signals S = (S0, S1, S2, S3) of the polarization image of the sample to be analyzed;
[0020] S3. Using the four Stokes vector signals of the polarization image, the degree of polarization (DOP), degree of linear polarization (DOLP), polarization phase angle ψ and ellipsoid angle χ of the polarization image are solved according to the following formula;
[0021]
[0022] S4. The result of the above formula combined with four Stokes vectors gives a total of 8 parameters, which are the image information of the sample to be analyzed.
[0023] Compared with the prior art, the advantages of the present invention are: by using microscopic imaging, by setting a specific light source, a microscopic lens and a CCD imaging detector, and by setting a polarization detection module between the microscopic lens and the CCD imaging detector, polarization imaging is introduced on the basis of short-wave infrared microscopic imaging, which can not only obtain the intensity information of the defect, but also obtain the polarization information of the defect, and can provide up to 8 image information. For most opaque semiconductor materials such as silicon, germanium, gallium arsenide, indium phosphide, etc., it is beneficial to analyze the defect characteristics such as cracks, warping, scratches, solid particles, metal pollution, organic pollution, etc., and complete the rapid identification, judgment and positioning of different types of defects. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 It is a schematic diagram of the structure of a defect detection device for semiconductor raw materials and chips according to an embodiment of the present invention.
[0025] In the figure: 1 is a tungsten halogen light source; 2 is a beam expansion and collimation module; 3 is a microscope stand; 4 is a sample stage; 5 is a short-wave infrared microscope lens; 6 is a polarization detection module; 7 is a CCD imaging detector; 8 is a stepper motor controller; and 9 is a control computer. DETAILED DESCRIPTION
[0026] The present invention is further described in detail below with reference to the accompanying drawings.
[0027] Example: Figure 1As shown, an image information acquisition device for semiconductor defect detection includes a halogen tungsten lamp 1, a beam expansion and collimation module 2, a microscope frame 3, a sample stage 4, a short-wave infrared microscope lens 5, a polarization detection module 6, a CCD imaging detector 7, a stepper motor controller 8 and a control computer 9, which are arranged in sequence. The short-wave infrared light emitted by the halogen tungsten lamp 1 is incident on the semiconductor raw material or semiconductor chip sample to be analyzed on the sample stage 4 through the beam expansion and collimation of the beam expansion and collimation module 2. The transmitted light is converged by the short-wave infrared microscope lens 5 to form an image, and a polarization parameter image is obtained after passing through the polarization detection module 6, wherein the polarization detection module 6 is composed of a corresponding An ultra-wideband achromatic 1 / 4 wave plate with a wavelength of 900nm to 1700nm and an analyzer are formed. The ultra-wideband achromatic 1 / 4 wave plate is fixed on a stepper motor (not shown in the figure). The CCD imaging detector 7 uses an InGaAs focal plane array detector with a spectral response range of 900nm to 1700nm, a pixel size of 640×512, a pixel center distance of 15μm, and a quantum efficiency greater than 70% in the spectral range of 1000nm to 1600nm. The CCD imaging detector 7 performs photoelectric conversion on the polarization parameter image and transmits it to the control computer 9, and obtains the image information of the sample to be analyzed by calculation in the control computer 9.
[0028] The acquisition method and principle of the device of the present invention are as follows:
[0029] In order to conveniently describe the polarization characteristics of the sample, the Stokes vector is introduced to describe the polarization state of light. Since the Stokes vector can describe both fully polarized light and partially polarized light and natural light, the following four parameters (S0, S1, S2, S3) are used to describe the light intensity and polarization state, where S0 represents the total light intensity I, S1 represents the intensity difference between 0° and 90°, S2 represents the intensity difference between 45° and 135°, and S3 represents the intensity difference between right-handed circularly polarized light and left-handed circularly polarized light. The Stokes vector of the sample polarization image is defined as
[0030]
[0031] The polarized light transmitted through the sample is collected by the short-wave infrared microscope lens 5 and incident on the wave plate and analyzer of the polarization detection module 6. After polarization analysis, the sample polarization image is obtained. The Stokes vector and Muller matrix can well describe the light polarization transmission characteristics and complete the polarization analysis. The polarization transmission characteristics of the wave plate and the analyzer are described by the Muller matrix. The wave plate uses an ultra-wideband achromatic 1 / 4 wave plate, and the phase delay generated is π / 2. The fast axis of the wave plate is set at the position of ρ, then the Muller matrix of the ultra-wideband achromatic wave plate is
[0032]
[0033] The transmittance axis of the analyzer is set at 0°, and its polarization transmission characteristics are described by the Muller matrix:
[0034]
[0035] According to the polarization transmission characteristics, the Stokes vector of the output light of the sample polarization information after passing through the ultra-wideband achromatic quarter-wave plate and the 0° analyzer to reach the CCD can be expressed as
[0036] S out =M 检偏 M 波片 S (4)
[0037] In order to solve the polarization information of the sample to be analyzed, the fast axis azimuth angle ρ of the ultra-wideband achromatic 1 / 4 wave plate is set at 0°, 22.5°, 45° and 135° respectively. In order to achieve high-precision polarization imaging, a stepper motor is used to rotate the wave plate to reduce the measurement error and realize the automation of the measurement system. After the stepper motor in the polarization detection module 6 receives the unit pulse signal from the stepper motor controller 8, the driver drives the stepper motor to rotate a fixed step angle. The rotation angle of the stepper motor is subdivided by changing the phase current of the stepper motor through the subdivision drive. In this embodiment, a two-phase four-wire stepper motor is selected with a step angle of 1.8° and a step angle of 32 subdivisions. The step angle resolution can reach 1.8° / 32=0.05625°. The fast axis azimuth of the ultra-wideband achromatic quarter-wave plate in the polarization detection module 6 is rotated to four angle positions respectively, and considering that the light intensity that the detector can detect is the first component of the Stokes vector, (1), (2) and (3) are substituted into (4) and the four cases of the fast axis azimuth of the ultra-wideband achromatic quarter-wave plate are considered, and the detection light intensity obtained by the CCD detector 7 can be solved as follows:
[0038]
[0039] The sample polarization image measurement can be completed by equation (5), and the obtained Stokes vector is:
[0040]
[0041] The Stokes vector microscopic image of the sample to be analyzed is obtained by measurement. For most semiconductor materials such as silicon, germanium, gallium arsenide, and indium phosphide, although they are opaque in the visible light / near infrared light band, their light transmittance is relatively good in the short-wave infrared region, especially short-wave infrared light with a wavelength exceeding 1200nm has good transmittance. In order to realize the transmission microscopic imaging measurement of semiconductor chip samples and obtain the short-wave infrared image of semiconductor chip samples, the CCD imaging detector 7 of the present invention uses an InGaAs focal plane array detector with low noise and high sensitivity. The typical values of the detector parameters are: spectral response range of 900nm to 1700nm, 640×512 pixel size, 15μm pixel center distance, quantum efficiency greater than 70% in the spectral range of 1000nm to 1600nm, and high resolution and detection accuracy.
[0042] When the fast axis azimuth of the ultra-wideband achromatic quarter-wave plate in the polarization detection module 6 is at 0°, 22.5°, 45° and 135°, the polarization image signals I0, I1 are obtained by photoelectric conversion. 22.5 ,I 45 and I 135 , substituting into formula (6) to obtain the four Stokes vector polarization image information S = (S0, S1, S2, S3) of the sample to be analyzed, the four Stokes vectors can be used to further solve the image information such as polarization degree (DOP), linear polarization degree (DOLP), polarization phase angle ψ and ellipsoid angle χ.
[0043]
[0044] The above result is combined with four Stokes vector images S = (S0, S1, S2, S3), and a total of 8 parameters are the image information of the sample to be analyzed.
[0045] The drawings and embodiments of the present invention are only used to illustrate the technical solution rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the specific implementation methods of the present invention can still be modified or some technical features can be replaced by equivalents without departing from the spirit of the present invention, which should be included in the scope of the technical solution for protection of the present invention.
Claims
1. A method for acquiring image information for semiconductor defect detection, using an acquisition device, the acquisition device comprising: A light source for providing lighting; A beam expansion and collimation module, used for expanding and collimating the light source; A sample stage, used to place samples to be analyzed; Microscope lens, used for focusing light and imaging; CCD imaging detector, used to receive optical image signals and perform photoelectric conversion; The control computer calculates and obtains the image information of the sample to be analyzed according to the signal received from the CCD imaging detector; The light emitted by the light source is expanded and collimated by the beam expansion and collimation module, and then incident on the sample to be analyzed on the sample stage. The light passing through the sample is converged and imaged by the microscope lens, received by the CCD imaging detector and transmitted to the control computer after photoelectric conversion, and the image information of the sample to be analyzed is calculated and obtained in the control computer. The sample to be analyzed is a semiconductor raw material or a semiconductor chip. The light source has a short-wave infrared spectrum with a wavelength exceeding 1200nm. The microscope lens is a short-wave infrared microscope lens. A polarization detection module is arranged between the microscope lens and the CCD imaging detector. The polarization detection module consists of an ultra-wideband achromatic 1 / 4 wave plate corresponding to a wavelength of 900nm~1700nm with an adjustable azimuth angle and a polarizer. The CCD imaging detector is an InGaAs focal plane array detector. The typical values of the main parameters of the detector are: a spectral response range of 900nm~1700nm, a pixel scale of 640´512, a pixel center distance of 15mm, and a quantum efficiency greater than 70% in the spectral range of 1000nm~1600nm. The specific steps are as follows: S1. Place the sample to be analyzed on the sample stage, the sample to be analyzed is a semiconductor raw material or a semiconductor chip, rotate the ultra-wideband achromatic quarter wave plate in the polarization detection module, adjust the fast axis azimuth to four positions of 0°, 22.5°, 45° and 135°, and measure the polarization images of the four positions respectively through the CCD imaging detector and convert them into polarization image electrical signals, which are recorded as I 0. I 22.5 , I 45 and I 135 , input it into the control computer; S2. The control computer polarizes the image electrical signal I 0. I 22.5 , I 45 and I 135 Substitute into the formula , Get the four Stokes vector signals S=( S 0, S 1, S 2, S 3); S3. Use the four Stokes vector signals of the polarization image to solve the degree of polarization (DOP), degree of linear polarization (DOLP), and polarization phase angle of the polarization image according to the following formulas. ψ and ellipsoid angle Χ ; , S4. The result of the above formula combined with four Stokes vectors gives a total of 8 parameters, which are the image information of the sample to be analyzed.
Citation Information
Patent Citations
Hyperspectral microscopy imaging and polarization modulation apparatus based on LCTF and method thereof
CN107101949A
Non-intuitive microscopic fluorescence imaging device and method of gene molecules
CN107290279A
Hyperspectral microscopic polarization imaging device and method
CN108414088A
Super-resolution microscopic imaging method and device based on active time modulation frequency mixing excitation irradiation
CN112485230A
Microscopic scattering polarization imaging surface defect measuring device and measuring method
WO2021003802A1
