Doped element distribution detection method and system based on hyperspectral imaging technology

Abstract

The invention relates to a doped element distribution detection method and system based on a hyperspectral imaging technology. The doped element distribution detection method based on the hyperspectral imaging technology comprises the steps that light of different wavelengths is adopted to irradiate an ion implantation area on the back face of a wafer; the method comprises the following steps: capturing spectral characteristic data of doping elements in an ion implantation region on the back surface of a wafer through hyperspectral imaging equipment, wherein the spectral characteristic data comprises space-spectral cube data; and identifying the three-dimensional distribution of the doped elements through a spectrum matching method according to the spectral characteristic data, and matching the spectral characteristic data with a quantitative relation model of the spectral intensity and the doping concentration to obtain the doping concentration of the doped element.

Description

Technical Field

[0001] The present application relates to the field of wafer manufacturing technology, and in particular to a method and system for detecting doping element distribution based on hyperspectral imaging technology. Background Art

[0002] In semiconductor manufacturing, ion implantation is a key step in achieving device doping, and the three-dimensional distribution of the doping element directly impacts device performance. Traditional detection methods rely primarily on sheet resistance measurements, which only provide an average value of the surface resistance and fail to capture the depth distribution, concentration gradient, or lateral distribution of the doping element. Precise control of doping depth and distribution is crucial for back-side ion implantation processes, such as those used in RF devices or 3D packaging. The limitations of traditional methods can easily lead to device performance parameters deviating from design targets.

[0003] While existing technologies like secondary ion mass spectrometry and transmission electron microscopy can provide depth-based distribution information, these methods require destructive sample processing, have long inspection cycles, and are costly, making them inadequate for real-time monitoring on production lines. Furthermore, these methods lack the ability to provide two-dimensional information on the dopant element distribution across the wafer surface, making it difficult to fully assess the quality of the ion implantation process. Summary of the Invention

[0004] The embodiments of the present application provide a method and system for detecting the distribution of doping elements based on hyperspectral imaging technology, which has the advantage of obtaining the three-dimensional distribution and concentration gradient of doping elements non-destructively and with high resolution.

[0005] To achieve the above objectives, in a first aspect, the present application provides a method for detecting doping element distribution based on hyperspectral imaging technology, comprising: Light of different wavelengths is used to illuminate the ion implantation area on the back side of the wafer; Capturing spectral characteristic data of doping elements in the ion implantation area on the back side of the wafer by a hyperspectral imaging device, wherein the spectral characteristic data includes spatial-spectral cube data; and The three-dimensional distribution of the doping element is identified according to the spectral characteristic data through a spectral matching method; and the spectral characteristic data is matched with a quantitative relationship model between spectral intensity and doping concentration to obtain the doping concentration of the doping element.

[0006] In some embodiments of the present application, after the step of "identifying the three-dimensional distribution of the doping element according to the spectral characteristic data and through the spectral matching method; and matching the spectral characteristic data with the quantitative relationship model between spectral intensity and doping concentration to obtain the doping concentration of the doping element", the step is also included: generating a two-dimensional doping concentration distribution map and a three-dimensional depth-concentration profile map and comparing them with the design value to achieve process monitoring.

[0007] In some embodiments of the present application, the steps of "identifying the three-dimensional distribution of the doping element according to the spectral characteristic data and through a spectral matching method; and matching the spectral characteristic data with a quantitative relationship model between spectral intensity and doping concentration to obtain the doping concentration of the doping element" include: comparing the spectral characteristic data with a spectral characteristic library of doping elements to identify the type and position of the doping element; and inputting the spectral characteristic data into the quantitative relationship model between spectral intensity and doping concentration to calculate the doping concentration of the doping element.

[0008] In some embodiments of the present application, in the step of "comparing the spectral feature data with the spectral feature library of doping elements to identify the type and location of the doping elements", multi-wavelength penetration depth difference is used to analyze the depth distribution of the doping elements, with short wavelengths analyzing shallow layers and long wavelengths analyzing deep layers.

[0009] In some embodiments of the present application, in the step of analyzing the doping depth distribution using multi-wavelength penetration depth difference, it is also necessary to analyze the depth distribution of the doping element in combination with the injection energy and annealing process parameters during ion doping.

[0010] In some embodiments of the present application, before the step of "identifying the three-dimensional distribution of the doping element according to the spectral characteristic data and through a spectral matching method; and matching the spectral characteristic data with a quantitative relationship model between spectral intensity and doping concentration to obtain the doping concentration of the doping element", the step is also included: preprocessing the collected hyperspectral data; wherein the preprocessing includes at least one of denoising, background subtraction, and normalization.

[0011] In some embodiments of the present application, the step of "capturing spectral characteristic data of specific doping elements in the ion implantation area on the back side of the wafer through a hyperspectral imaging device" includes: the incident angle of the light of different wavelengths is 30°-70°.

[0012] In some embodiments of the present application, the establishment of the doping element spectral characteristic library includes: for different doping elements, establishing their characteristic absorption / reflection spectral curves in the activated state through experiments or simulations, wherein the characteristic absorption / reflection spectral curves include peak wavelength parameters and half-height width parameters; and / or The establishment of the quantitative relationship model between spectral intensity and doping concentration includes: using training data of samples with known concentrations of doping elements to calibrate samples, and using machine learning or physical models to achieve concentration inversion.

[0013] On the second aspect, the present application also provides a doping element distribution detection system based on hyperspectral imaging technology, including: a light source for emitting light of different wavelengths and irradiating the ion implantation area on the back of the wafer; a hyperspectral camera for capturing spectral characteristic data of specific doping elements in the ion implantation area on the back of the wafer; an inclined incidence optical module for adjusting the incident angle of light emitted by the light source; and a data processing unit for storing a quantitative relationship model between spectral intensity and doping concentration and a doping element spectral characteristic library; and comparing the spectral characteristic data with the doping element spectral characteristic library to identify the type and position of the doping element, and inputting the spectral characteristic data into the quantitative relationship model between spectral intensity and doping concentration to calculate the doping concentration of the doping element.

[0014] In some embodiments of the present application, the data processing unit is further configured to generate a two-dimensional doping concentration distribution map and a three-dimensional depth-concentration profile map and compare them with the design values ​​to achieve process monitoring.

[0015] The present application provides a method and system for detecting the distribution of doping elements based on hyperspectral imaging technology, the detection method comprising: irradiating the ion implantation area on the back of the wafer with light of different wavelengths; capturing the spectral characteristic data of the specific doping element in the ion implantation area on the back of the wafer by a hyperspectral imaging device, the spectral characteristic data comprising spatial-spectral cube data; and identifying the three-dimensional distribution of the doping element based on the spectral characteristic data and by a spectral matching method; and matching the spectral characteristic data with a quantitative relationship model between spectral intensity and doping concentration to obtain the doping concentration of the doping element. The present application introduces a hyperspectral imaging device and combines the spectral matching method and the quantitative relationship model between spectral intensity and doping concentration to obtain the concentration gradient, injection depth and lateral distribution information (three-dimensional distribution and concentration gradient of the doping element) of the specific doping element in the ion implantation area on the back of the wafer without destroying the wafer. Therefore, the present application provides a method and system for detecting the distribution of doping elements based on hyperspectral imaging technology with the advantages of non-destructiveness and high resolution.

[0016] Other features and advantages of the present application will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS To more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present application. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0017] In order to more completely understand the present application and its beneficial effects, the following description will be given in conjunction with the accompanying drawings, wherein the same drawing numbers represent the same parts in the following description.

[0018] Figure 1 A schematic flow chart of a method for detecting doping element distribution based on hyperspectral imaging technology is provided for some exemplary embodiments of the present application.

[0019] Figure 2 for Figure 1 The flowchart of step S2 in the flowchart of the method for detecting doping element distribution based on hyperspectral imaging technology is shown.

[0020] Figure 3 Schematic diagram of modules of a doping element distribution detection system based on hyperspectral imaging technology provided for some exemplary embodiments of the present application.

[0021] Description of reference numerals: 100. Detection system; 10. Light source; 20. Hyperspectral device; 30. Oblique-incidence optical module; 40. Data processing unit. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0023] The semiconductor manufacturing industry has long relied on the sheet resistance method to assess the doping effect after ion implantation. However, this method only reflects the average surface resistance and cannot reveal the depth distribution and concentration gradient of the doping element. For backside implantation processes, precise control of doping depth and lateral distribution directly affects device performance. The two-dimensional detection characteristics of traditional methods result in a lack of effective data support for process parameter adjustments. While destructive testing methods can obtain depth information, they require wafer slicing and cannot meet the needs of online testing, resulting in extended production cycles and increased costs.

[0024] In order to solve the above problems, the inventors noticed that hyperspectral imaging technology can simultaneously obtain spatial information of the target and continuous or quasi-continuous narrow-band spectral data. Its high-resolution spectrum usually contains characteristic absorption or reflection peaks related to the chemical composition of the material (such as the fingerprint characteristics of a specific element), but it is necessary to combine calibration data or algorithms to analyze overlapping signals. By studying the optical response characteristics of the activated state of the doping elements, it was found that different elements have distinguishable absorption or reflection characteristics at specific wavelengths. It was further found that the difference in penetration depth of light of different wavelengths can reflect the depth distribution information. Combined with the physical correlation between spectral intensity and concentration, a technical route for inverting three-dimensional distribution and concentration through optical signals was formed. The doping element distribution detection method based on hyperspectral imaging technology of this application will be described in detail below with reference to the accompanying drawings.

[0025] See also Figure 1 The present application provides a method for detecting the distribution of doping elements based on hyperspectral imaging technology, comprising the steps of: S1, using light of different wavelengths to illuminate the ion implantation area on the back side of the wafer; S3, capturing spectral characteristic data of doping elements in the ion implantation area on the back side of the wafer by a hyperspectral imaging device, wherein the spectral characteristic data includes spatial-spectral cube data; and S4, identifying the three-dimensional distribution of the doping element according to the spectral characteristic data through a spectral matching method; and matching the spectral characteristic data with a quantitative relationship model between spectral intensity and doping concentration to obtain the doping concentration of the doping element.

[0026] Hyperspectral imaging equipment is an imaging device that simultaneously records the spatial location of a target area and continuous or quasi-continuous spectral information. This is achieved using a combination of a beam splitter prism and an array detector. The beam splitter decomposes the incident light into hundreds of continuous bands, with each pixel corresponding to a complete spectral curve. This hyperspectral imaging device provides the data foundation for analyzing the three-dimensional distribution of doping elements.

[0027] Among them, spectral feature data refers to the data cube containing spatial position information and continuous or quasi-continuous band spectral response obtained by hyperspectral imaging equipment. Specifically, it can be achieved by scanning the back area of ​​the wafer line by line in the ultraviolet to near-infrared band through a hyperspectral camera and recording the full-band reflectivity or transmittance of each pixel point. It is used to characterize the differences in spectral characteristics at different positions. Its spatial dimension information is used to locate the doping area, and the spectral dimension information is used to identify the type and concentration of elements.

[0028] Among them, the spectral matching method refers to a method of identifying element distribution by comparing the similarity between the measured spectrum and the characteristic spectrum library. Specifically, it can be implemented using the spectral angle mapping algorithm or the least squares fitting algorithm, and the element type and spatial position are determined by calculating the angle or residual between the spectral curves.

[0029] Among them, the quantitative relationship model refers to an algorithmic model that establishes a mathematical association between spectral data and doping concentration values. Specifically, it can be achieved by using machine learning methods to train the mapping relationship between the spectral characteristics of samples with known concentrations and concentration labels, which is used to convert spectral intensity into quantitative concentration values.

[0030] Specifically, the hyperspectral imaging device illuminates the back of the wafer at a specific incident angle to reduce surface reflection interference and enhance the optical response signal of the injection area. Each pixel in the collected spatial-spectral cube data contains a spectral curve of a continuous or quasi-continuous band. Through spectral angle mapping, least squares fitting and other spectral matching algorithms, it is compared with the pre-established doping element feature library to identify the type and lateral distribution of the doping element. The difference in penetration depth of light of different wavelengths is used to analyze the depth distribution. Short-wavelength light reflects shallow doping information, and long-wavelength light reflects deep distribution. After the spectral intensity data is input into the quantitative relationship model, the concentration value of each pixel is output according to the calibration curve or training model, and finally the concentration distribution data under three-dimensional spatial coordinates is formed.

[0031] Compared to existing technologies, traditional sheet resistance methods only provide average surface resistance values ​​and are unable to distinguish concentration differences at different depths or lateral positions. This application, however, leverages the three-dimensional nature of spectral data to achieve depth distribution detection with micron-level spatial resolution. Compared to destructive detection methods, this application's dopant element distribution detection method, based on hyperspectral imaging technology, can complete full-area scanning without destroying the wafer, making the detection process compatible with production lines.

[0032] Through the above-mentioned technical solution, this application solves the problem that traditional methods cannot obtain the depth distribution and concentration gradient of doping elements, and realizes non-destructive three-dimensional distribution detection of doping elements. By combining spectral matching with quantitative modeling, concentration inversion is completed while maintaining high spatial resolution, providing accurate data support for process parameter optimization. Utilizing optical penetration depth differences to analyze longitudinal distribution avoids complex physical stripping steps and significantly improves detection efficiency and repeatability.

[0033] Please continue reading Figure 1 In some embodiments of the present application, after step S4, the doping element distribution detection method based on hyperspectral imaging technology also includes step S5: generating a two-dimensional doping concentration distribution map and a three-dimensional depth-concentration profile map and comparing them with the design value to achieve process monitoring.

[0034] Among them, the two-dimensional doping concentration distribution map refers to an image generated by mapping the lateral spatial coordinates in the hyperspectral data with the doping concentration values ​​at the corresponding positions. Specifically, pseudo-color coding technology can be used to convert different concentration values ​​into color gradients for visualization, which is used to intuitively display the concentration gradient distribution of the doping element on the wafer surface. The three-dimensional depth-concentration profile map refers to a stereoscopic image generated by obtaining longitudinal depth information through multi-wavelength penetration depth difference analysis technology and reconstructing the concentration data corresponding to each depth layer in three dimensions. Specifically, tomography algorithms can be used to deconvolute the spectral data of different wavelengths to reveal the concentration variation pattern of the doping element in the vertical direction. Design value comparison refers to the difference analysis between the measured two-dimensional concentration distribution data and three-dimensional depth-concentration data and the distribution parameters preset in the process design file. Specifically, image registration technology can be used to spatially align the measured data with the design template and calculate the deviation amount, which is used for real-time feedback of the process execution status.

[0035] Specifically, after preprocessing the hundreds of bands of spectral data acquired by hyperspectral imaging equipment, the characteristic spectral curve of each pixel is extracted. A spectral matching algorithm is used to identify the type of doping element and determine its spatial distribution. Combined with a quantitative concentration model, the doping concentration value of each pixel is calculated to form a two-dimensional concentration distribution map. At the same time, based on the difference in penetration depth of light of different wavelengths in silicon materials, short-wavelength data reflects shallow concentration information, and long-wavelength data reflects deep concentration information. A three-dimensional depth-concentration profile is constructed through multi-wavelength tomography analysis. The generated distribution map is compared point by point with the target distribution range specified in the process design document. When the measured concentration distribution exceeds the design tolerance band, an alarm is triggered to guide real-time adjustment of process parameters.

[0036] Compared with existing technologies, traditional sheet resistance measurement can only obtain the average surface resistance value and cannot distinguish lateral distribution differences or depth-directed concentration gradients. Destructive detection methods such as secondary ion mass spectrometry require offline sampling and take several hours. This application uses hyperspectral imaging technology to achieve non-destructive testing. While maintaining the integrity of the wafer, it can complete the two-dimensional concentration distribution and three-dimensional depth profile detection of the entire wafer in five minutes. The detection data is directly linked to the production line control system to achieve closed-loop process control.

[0037] Through the above technical solution, this application solves the technical bottleneck that traditional detection methods cannot obtain three-dimensional distribution information of doping elements online. It realizes the instant identification of process deviations through visual concentration distribution maps, and feeds back the detection results to the ion implanter for dose compensation or scan path optimization, effectively preventing device performance failure caused by abnormal doping distribution.

[0038] See also Figure 2In some embodiments of the present application, step S4 includes: step S41, comparing the spectral feature data with the spectral feature library of the doping element to identify the type and position of the doping element; and step S42, inputting the spectral feature data into the quantitative relationship model between the spectral intensity and the doping concentration to calculate the doping concentration of the doping element.

[0039] Among them, the doping element spectral feature library refers to a pre-established database containing the characteristic absorption or reflection spectral curves of different doping elements in the activated state. It can be achieved by simulating the spectral responses of different elements at specific wavelengths through experimental measurement or simulation, and is used to identify the element type through spectral matching.

[0040] Specifically, after completing hyperspectral data acquisition, the spectral curve of each pixel is first calculated for similarity with the standard spectrum in the feature library. The type of doping element and its spatial distribution are determined by matching the peak wavelength and half-width at half maximum parameters. Subsequently, the matched spectral data is input into a trained quantitative relationship model between spectral intensity and doping concentration. Based on the nonlinear relationship between spectral intensity and concentration established within the model, the concentration value of each pixel is output. Through layer-by-layer analysis of spatial-spectral data, the three-dimensional distribution of the doping element in the lateral and depth directions and its concentration gradient are ultimately obtained.

[0041] In this embodiment, in step S1, a halogen lamp (400 nm-2500 nm) is incident on a silicon wafer that has undergone backside ion implantation and annealing. The light source is a halogen lamp (400 nm-2500 nm) with an incident angle of 60°. In step S3, the hyperspectral device has a resolution of 5 μm / pixel. In step S41, characteristic wavelengths (e.g., the 1050 nm reflection valley corresponding to phosphorus doping) are extracted, and an element distribution map is generated using spectral angle mapping. In step S42, the concentration distribution differences at different wavelengths (450 nm and 1000 nm) are compared to estimate the deviation of the doping depth from the designed value.

[0042] Compared to existing technologies, traditional sheet resistance methods can only measure average surface resistance and are unable to distinguish element types or depth distribution. Destructive detection methods such as secondary ion mass spectrometry require wafer cleavage and cannot be applied online. This solution, through a dual matching mechanism of spectral signature libraries and quantitative models, simultaneously achieves element type identification, three-dimensional localization, and concentration inversion under non-destructive conditions, overcoming the limitations of traditional single-dimensional detection techniques.

[0043] Through the above technical solution, the present application can obtain the three-dimensional distribution and concentration gradient data of the doping elements in the ion implantation area in real time during the semiconductor manufacturing process, avoid damaging the wafer structure, provide a quantitative basis with high spatial resolution for process parameter adjustment, and effectively improve the doping depth and concentration control accuracy.

[0044] In some embodiments of the present application, in the step of "comparing the spectral feature data with the spectral feature library of doping elements to identify the type and location of the doping elements", multi-wavelength penetration depth difference is used to analyze the depth distribution of the doping elements, with short wavelengths analyzing shallow layers and long wavelengths analyzing deep layers.

[0045] Multi-wavelength penetration depth variation refers to the difference in the ability of light of different wavelengths to penetrate semiconductor materials. This can be achieved using light in the ultraviolet and near-infrared bands. Light in the ultraviolet band has weaker penetration, while light in the near-infrared band has stronger penetration. Short wavelength refers to light with a wavelength range of 300-450 nanometers, specifically achieved using ultraviolet light sources or blue-violet lasers. The penetration depth of such light in silicon materials is less than 100 nanometers. Long wavelength refers to light with a wavelength range of 800-2500 nanometers, specifically achieved using halogen lamps or near-infrared lasers. The penetration depth of such light in silicon materials can reach micrometers.

[0046] Specifically, during the data acquisition phase, a broadband light source sequentially projects light of different wavelengths onto the back of the wafer, and the hyperspectral camera synchronously collects the reflection or absorption spectral data corresponding to each wavelength. During the data processing phase, for the same spatial location, spectral data corresponding to short wavelengths are extracted to analyze the distribution of shallow doping elements, and spectral data corresponding to long wavelengths are extracted to analyze the distribution of deep doping elements. By superimposing the spectral matching results of different wavelength layers, a three-dimensional depth distribution model is constructed. For example, for the phosphorus-doped area, the spectral data with a wavelength of 450 nanometers reflects the doping concentration within 50 nanometers of the surface, and the spectral data with a wavelength of 1050 nanometers reflects the doping concentration gradient at a depth of 1 micron.

[0047] Compared to existing technologies, traditional spectral matching methods utilize only a single wavelength or fixed wavelength combination for element identification, making it impossible to distinguish doping distributions at different depths. This approach leverages physical penetration depth differences to enable tomographic analysis, obtaining depth-dimensional information without destroying the sample. While existing spectral analysis only generates two-dimensional distribution maps, this approach directly constructs three-dimensional depth-concentration profiles through multi-wavelength data fusion.

[0048] Through the above-mentioned technical solution, this application solves the problem that traditional detection methods cannot non-destructively analyze the depth distribution of ion implanted doping elements. Through the synergistic effect of short and long wavelengths, shallow and deep doping element information can be simultaneously obtained during the same detection process, realizing three-dimensional concentration gradient analysis from the surface to the interior. This method avoids destructive sampling and eliminates the information loss in the depth dimension of traditional spectral matching, providing complete doping distribution data for process monitoring.

[0049] In some embodiments of the present application, in the step of analyzing the doping depth distribution using multi-wavelength penetration depth difference, it is also necessary to analyze the depth distribution of the doping element in combination with the injection energy and annealing process parameters during ion doping.

[0050] Among them, the injection energy refers to the initial ion implantation depth determined by the acceleration voltage during the ion injection process. This value can be obtained through process parameter records or equipment logs and is used to constrain the initial distribution position of the doping element.

[0051] The annealing process parameters refer to the temperature, time and atmosphere conditions during the annealing process. Specifically, temperature gradient data and time series data can be used to correct the influence of the diffusion behavior of doping elements caused by annealing on the depth distribution.

[0052] Specifically, when analyzing the doping depth distribution, preliminary depth stratification information is first obtained through multi-wavelength spectral data, and then the injection energy parameters are used as physical constraints for the initial depth distribution. For example, when the injection energy is high, the initial position of the doping element is limited to a deeper area. At the same time, the annealing process parameters are used to calculate the diffusion range of the doping element during the annealing process. For example, high-temperature annealing will cause the doping element to diffuse to the shallow layer, thereby adjusting the analysis results based on the spectral penetration depth. By fusing the process parameters with the optical data, a depth analysis model that contains physical laws is established, so that the depth distribution results are consistent with both the spectral detection data and the physical mechanisms of ion implantation and annealing.

[0053] Compared to existing technologies, which rely solely on multi-wavelength penetration differences for depth analysis, this approach fails to consider the impact of actual process parameters on the distribution of dopant elements, potentially leading to analytical results that deviate from the true physical process. For example, when the annealing process causes dopant element diffusion, simple spectral penetration depth analysis cannot accurately reflect the actual distribution after diffusion. This approach effectively eliminates analytical errors introduced by process variables by incorporating process parameters as constraints.

[0054] Through the above technical solution, the present application solves the problem of physical mechanism deviation that may exist when relying solely on optical penetration difference to analyze depth distribution, improves the analysis accuracy of the depth distribution of doped elements, and makes the detection results more consistent with the element distribution state under actual process conditions.

[0055] In some embodiments of the present application, before the step of "identifying the three-dimensional distribution of the doping element according to the spectral characteristic data and through a spectral matching method; and matching the spectral characteristic data with a quantitative relationship model between spectral intensity and doping concentration to obtain the doping concentration of the doping element", the step is also included: preprocessing the collected hyperspectral data; wherein the preprocessing includes at least one of denoising, background subtraction, and normalization.

[0056] Hyperspectral data preprocessing is a key step in eliminating interference and enhancing effective signals, directly impacting the reliability of subsequent analysis. It reduces random noise (such as sensor noise and environmental interference) and improves the signal-to-noise ratio (SNR) to prevent noise interference with subsequent quantitative analysis (such as concentration inversion). In this embodiment, Savitzky-Golay filtering can be used for denoising. This denoising method smoothes the spectral curve using a local polynomial fit, preserving peak characteristics. Background subtraction can eliminate the influence of non-target signals (such as light source fluctuations, dark current, and stray light) and highlight the characteristic spectrum of the material itself. For example, dark current correction, whiteboard correction, baseline correction, and spectral differentiation can be used for background subtraction. Normalization can eliminate dimensional differences, making data from different samples or bands comparable and reducing the influence of factors such as light source intensity and distance on the spectrum. For example, maximum normalization, standard normal transformation, multivariate scatter correction, and band ratio methods can be used for normalization.

[0057] This application preprocesses the collected hyperspectral data to eliminate interference and enhance effective signals, thereby enhancing the reliability of subsequent three-dimensional distribution and concentration analysis of doping elements.

[0058] In some embodiments of the present application, the establishment of the doping element spectral characteristic library includes: for different doping elements, establishing their characteristic absorption / reflection spectral curves in the activated state through experiments or simulations, wherein the characteristic absorption / reflection spectral curves include peak wavelength parameters and half-height width parameters; and / or Establishing the quantitative relationship model between spectral intensity and doping concentration includes: using training data of samples with known concentrations of doping elements to calibrate samples, and using machine learning (such as support vector regression, neural networks) or physical models (such as the modified formula of the Beer-Lambert law) to achieve concentration inversion.

[0059] For example, phosphorus doping has a characteristic reflectivity change in the near-infrared band (such as 900-1100 nm), and boron doping exhibits specific absorption in the visible light band (such as 500-600 nm).

[0060] Among them, the characteristic absorption or reflection spectral curve refers to the spectral response characteristics of the doping element under specific energy level transitions. Specifically, it can be achieved by using experimental measurements or electromagnetic simulation software to simulate the optical properties of the activated doping element, which is used to accurately identify the type and activation state of the element.

[0061] The peak wavelength parameter refers to the wavelength position of the absorption or reflection characteristic peak in the spectral curve. Specifically, it can be achieved by measuring the spectral response of different doping elements with a spectrometer and extracting the peak position to distinguish the characteristic spectra of different elements.

[0062] The half-width parameter refers to the wavelength width of the characteristic peak at half of the maximum intensity, which can be achieved by calculating the half-width value of the spectral curve. It is used to quantify the broadening degree of the spectral characteristic peak to improve the recognition specificity.

[0063] Among them, the quantitative relationship model between spectral intensity and doping concentration refers to an algorithm that establishes a mathematical association between spectral data and concentration values. Specifically, it can be achieved by using a neural network to train the mapping relationship between the spectral intensity of the calibration sample and the known concentration, or by establishing a physical equation based on the modified Beer-Lambert law to convert the spectral signal into a concentration value.

[0064] Specifically, when establishing a characteristic absorption or reflection spectrum curve, the spectral data of the doping element in the activated state is obtained through experimental measurement or simulation, and the peak wavelength and half-width are extracted as characteristic parameters and stored in the database. For example, after annealing the boron-doped sample, a hyperspectral imaging system is used to collect its absorption spectrum in the visible light band, and it is determined that the characteristic peak is located at 550 nanometers and the half-width is 30 nanometers. When establishing a quantitative relationship model, a calibration sample set with known concentration is used, and the collected hyperspectral data is input into the neural network for training to establish a nonlinear mapping relationship between spectral intensity and concentration. For example, the gradient descent algorithm is used to optimize the neural network weights so that the error between the concentration prediction value output by the model and the actual measured value is minimized.

[0065] Compared to existing technologies, traditional methods rely on destructive detection methods and lack a reliable spectral signature library, making them incapable of non-destructive element identification and concentration inversion. Existing technologies lack established spectral signature parameters for activated dopant elements, resulting in low element identification accuracy. Furthermore, they lack concentration inversion methods that combine machine learning and physical models, making them difficult to adapt to detection requirements under complex process conditions.

[0066] Through the above technical solution, this application achieves the technical effect of non-destructively detecting the type and concentration of doping elements. By establishing a spectral feature library containing peak wavelength and half-width at half maximum parameters, the spectral fingerprint characteristics of different doping elements can be accurately identified. By integrating data-driven and physical mechanism-based concentration inversion models, hyperspectral data can be converted into quantitative concentration values, solving the technical problem that traditional methods cannot achieve simultaneous element identification and concentration measurement.

[0067] In some embodiments of the present application, the step of "capturing spectral characteristic data of specific doping elements in the ion implantation area on the back side of the wafer through a hyperspectral imaging device" includes: the incident angle of the light of different wavelengths is 30°-70°.

[0068] The term "light of different wavelengths" refers to light sources in at least two wavelengths of ultraviolet, visible, and near-infrared light. This can be achieved using a halogen lamp combined with a filter or a tunable laser. By switching between different wavelengths, the detection of doped elements at different depths can be achieved. The incident angle of 30°-70° refers to the angle formed between the light and the wafer surface normal. This can be achieved using the reflector angle adjustment mechanism in the tilted-incidence optical module. By controlling the incident angle, interference from surface-reflected light on the effective spectral signal can be suppressed.

[0069] Specifically, a broad-spectrum light source illuminates the wafer backside at an oblique incidence. Short-wavelength light (such as ultraviolet light) is preferentially absorbed by shallow (<100 nm) doped regions, while long-wavelength light (such as near-infrared light) penetrates deeper regions (micrometer-level), creating a multi-layered detection effect. Hyperspectral imaging equipment collects the reflected spectrum line by line using a line-scanning method, constructing a data cube containing two-dimensional spatial coordinates and continuous spectral dimensions. When the incident angle is controlled within the range of 30°-70°, the surface specular reflected light deviates from the detector's receiving direction, effectively capturing the diffuse reflected signal from the doped regions while ensuring sufficient light flux to maintain the signal-to-noise ratio. By synergistically optimizing wavelength selection and incident angle, the characteristic absorption peaks of the doping elements are highlighted in the spectral data. Specifically, when short-wavelength light illuminates the wafer backside at an incident angle of 30°-70° (for example, 50°), its shallow penetration effectively captures the spectral characteristics of the surface doping elements. At the same time, the oblique incidence method deflects the reflected light from the detection optical path. For deeply doped elements, long-wavelength light is used at an incident angle of 30°-70° (e.g., 65°). Its deep penetration ability can stimulate the spectral response of deeply doped elements, while oblique incidence still maintains effective signal acquisition. Hyperspectral imaging equipment records the spectral intensity of each spatial location at different wavelengths through a point-by-point scanning method, forming a three-dimensional dataset containing lateral position, depth information, and spectral characteristics.

[0070] Compared with existing technologies, traditional methods use a fixed wavelength vertical incidence method, which causes surface reflected light to overlap with the effective signal and cannot distinguish features at different depths. This application suppresses surface reflection noise and realizes layered extraction of spectral signals at different depths through the coordinated regulation of wavelength and incident angle.

[0071] Through the above technical solution, this application effectively eliminates the spectral distortion caused by wafer surface reflection, realizes the full-depth spectral feature capture from sub-micron shallow doping to several micron deep doping, and provides a high-fidelity original data foundation for subsequent three-dimensional distribution reconstruction.

[0072] The present invention can replace traditional destructive detection methods, improve the efficiency of ion implantation process monitoring, and is applicable to the stringent requirements on doping distribution in advanced processes (such as Fin Field-Effect Transistor (Fin-FET) or Three-Dimensional Integrated Circuit (3D IC)), thus having significant industrial value.

[0073] See also Figure 3 In the second aspect, the present application also provides a doping element distribution detection system 100 based on hyperspectral imaging technology. The detection system 100 includes a light source 10, a hyperspectral device 20, an oblique incidence optical module 30 and a data processing unit 40. The light source 10 is used to emit light of different wavelengths and illuminate the ion implantation area on the back of the wafer. The hyperspectral device 20 is used to capture the spectral characteristic data of a specific doping element in the ion implantation area on the back of the wafer. The oblique incidence optical module 30 is used to adjust the incident angle of the light emitted by the light source. The data processing unit 40 is used to store a quantitative relationship model between spectral intensity and doping concentration and a doping element spectral characteristic library; and compare the spectral characteristic data with the doping element spectral characteristic library to identify the type and position of the doping element, input the spectral characteristic data into the quantitative relationship model between spectral intensity and doping concentration, and calculate the doping concentration of the doping element.

[0074] In some embodiments of the present application, the data processing unit 40 is further configured to generate a two-dimensional doping concentration distribution map and a three-dimensional depth-concentration profile map and compare them with design values ​​to achieve process monitoring.

[0075] Among them, the doping element distribution detection system 100 based on hyperspectral imaging technology can be integrated into a semiconductor production line to support online detection and process feedback control.

[0076] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this application, "plurality" means two or more, unless otherwise specifically defined.

[0077] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0078] The embodiments, implementation methods and related technical features of the present application can be combined and replaced with each other without conflict.

[0079] The above are merely preferred embodiments of the present application and do not constitute any form of limitation to the present application. Although the description of each embodiment in the present application has different focuses, for parts not described in detail in a certain embodiment, reference can be made to the relevant embodiments of other embodiments. However, any modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present application without departing from the content of the technical solution of the present application are still within the scope of the technical solution of the present application.

Claims

1. A method for detecting doping element distribution based on hyperspectral imaging technology, characterized in that: Including steps: Light of different wavelengths is used to illuminate the ion implantation area on the back side of the wafer; Capturing spectral characteristic data of doping elements in the ion implantation area on the back side of the wafer by a hyperspectral imaging device, wherein the spectral characteristic data includes spatial-spectral cube data; and The three-dimensional distribution of the doping element is identified according to the spectral characteristic data through a spectral matching method; and the spectral characteristic data is matched with a quantitative relationship model between spectral intensity and doping concentration to obtain the doping concentration of the doping element.

2. The method for detecting doping element distribution based on hyperspectral imaging technology according to claim 1, wherein: After the step of "identifying the three-dimensional distribution of the doping element by a spectral matching method based on the spectral characteristic data; and matching the spectral characteristic data with a quantitative relationship model between spectral intensity and doping concentration to obtain the doping concentration of the doping element", the method further includes the following steps: Generate 2D doping concentration profiles and 3D depth-concentration profiles and compare them with design values ​​for process monitoring.

3. The method for detecting doping element distribution based on hyperspectral imaging technology according to claim 1, wherein: "Identifying the three-dimensional distribution of doping elements based on the spectral characteristic data by a spectral matching method; The step of matching the spectral characteristic data with a quantitative relationship model between spectral intensity and doping concentration to obtain the doping concentration of the doping element includes: Comparing the spectral characteristic data with a doping element spectral characteristic library to identify the type and location of the doping element; and The spectral characteristic data is input into the quantitative relationship model between the spectral intensity and the doping concentration to calculate the doping concentration of the doping element.

4. The method for detecting doping element distribution based on hyperspectral imaging technology according to claim 3, wherein: In the step of "comparing the spectral feature data with the doping element spectral feature library to identify the type and location of the doping element", multi-wavelength penetration depth difference is used to analyze the depth distribution of the doping element, with short wavelengths analyzing shallow layers and long wavelengths analyzing deep layers.

5. The method for detecting doping element distribution based on hyperspectral imaging technology according to claim 4, characterized in that: In the step of analyzing the doping depth distribution by using multi-wavelength penetration depth difference, it is also necessary to analyze the depth distribution of the doping elements by combining the injection energy and annealing process parameters during ion doping.

6. The method for detecting doping element distribution based on hyperspectral imaging technology according to claim 1, wherein: Before the step of "identifying the three-dimensional distribution of the doping element by a spectral matching method based on the spectral characteristic data; and matching the spectral characteristic data with a quantitative relationship model between spectral intensity and doping concentration to obtain the doping concentration of the doping element", the method further includes the following steps: Preprocess the collected hyperspectral data; The preprocessing includes at least one of denoising, background subtraction, and normalization.

7. The method for detecting doping element distribution based on hyperspectral imaging technology according to claim 1, wherein: The incident angles of the lights of different wavelengths are 30°-70°.

8. The method for detecting doping element distribution based on hyperspectral imaging technology according to claim 3, wherein: The establishment of the doping element spectral feature library includes: For different doping elements, establishing characteristic absorption / reflection spectrum curves in their activated states through experiments or simulations, wherein the characteristic absorption / reflection spectrum curves include peak wavelength parameters and half-height width parameters; and / or The establishment of the quantitative relationship model between spectral intensity and doping concentration includes: Using training data of samples with known concentrations of doping elements, concentration inversion is achieved using machine learning or physical models.

9. A doping element distribution detection system based on hyperspectral imaging technology, characterized in that: include: A light source, configured to emit light of different wavelengths and illuminate the ion implantation area on the back side of the wafer; Hyperspectral equipment, used to capture spectral signature data of specific doping elements in the ion-implanted area on the backside of the wafer; an inclined-incidence optical module, configured to adjust the incident angle of light emitted by the light source; and A data processing unit is used to store a quantitative relationship model between spectral intensity and doping concentration and a spectral feature library of doping elements; and compare the spectral feature data with the spectral feature library of doping elements to identify the type and location of the doping element, input the spectral feature data into the quantitative relationship model between spectral intensity and doping concentration, and calculate the doping concentration of the doping element.

10. The doping element distribution detection system based on hyperspectral imaging technology according to claim 9, characterized in that: The data processing unit is further used to generate a two-dimensional doping concentration distribution map and a three-dimensional depth-concentration profile map and compare them with the design values ​​to achieve process monitoring.

Images