In-situ monitoring method, system, in-situ monitoring device for solution thin film manufacturing process

Abstract

The application provides an in-situ monitoring method, system and device for a solution thin film preparation process, relates to the technical field of semiconductor detection, and obtains multi-sequence data including an image sequence, a reflection spectrum sequence and an interference signal sequence of a wafer surface to be detected; processes the reflection spectrum sequence to obtain a spectrum unmixing result; performs constraint solving on the spectrum unmixing result based on the interference signal sequence to obtain a volume fraction change curve; and calculates a process parameter based on the volume fraction change curve to generate a spatial distribution map. The application synchronously acquires and jointly analyzes image information, spectrum information and interference information, realizes the collaborative monitoring of chemical evolution, physical thickness change and spatial distribution characteristics in the solution thin film preparation process, effectively overcomes the problems of single information dimension, difficulty in decoupling coupled sub-processes and difficulty in accurately reflecting the dynamic law of the whole preparation process in the traditional technology, and thus improves the monitoring precision of the solution thin film preparation process.

Description

Technical Field

[0001] This application relates to the field of semiconductor detection technology, and in particular to an in-situ monitoring method, system, and device for solution-based thin film preparation processes. Background Technology

[0002] In-situ monitoring technology refers to the real-time, continuous, and non-invasive measurement of key physical and chemical parameters in the actual environment and dynamic process of material preparation or reaction. Compared with endpoint analysis methods that rely on detection after preparation, in-situ monitoring technology can directly obtain dynamic evolution information during the material formation process, thereby providing a basis for process mechanism research, parameter optimization, and process control.

[0003] In the field of advanced electronic and optoelectronic materials fabrication, especially in the preparation of novel soft material thin films using solution processing methods for perovskite and organic semiconductor materials, the transformation of the thin film from a liquid precursor to a solid functional layer is typically accompanied by the coupled evolution of multiple physicochemical processes, including solvent evaporation, nucleation, grain growth, phase transition, and film thickness shrinkage. These processes directly affect not only the crystallinity, phase purity, defect density, and uniformity of the thin film, but also further determine the optoelectronic performance and stability of the device. Therefore, how to obtain key process parameters related to film formation kinetics in real time during solution-based thin film preparation has become a crucial issue for achieving controllable fabrication and industrial applications in this field.

[0004] Traditional techniques typically employ a single optical sensor for online monitoring of the film formation process, an in-situ monitoring system developed specifically for vapor deposition (VCD) for real-time monitoring, and offline characterization to analyze the film structure and properties after formation. While these methods can acquire information on film thickness, morphology, or structure to some extent, they still have significant limitations: single optical monitoring provides limited information dimensions, making it difficult to distinguish coupled sub-processes such as solvent evaporation, crystallization, and phase transition; VCD monitoring systems are primarily designed for high-temperature, vacuum, or inert environments, making them unsuitable for application scenarios with high risks of atmospheric pressure, organic vapors, and contamination in solution-based processes; and offline characterization only provides the endpoint results after formation, failing to reflect the dynamic evolution during film formation. Therefore, existing technologies cannot meet the demands for simultaneous acquisition of multi-dimensional information and real-time analysis of kinetic processes in solution-based thin film preparation.

[0005] In view of this, there is an urgent need in the field for an in-situ monitoring method and system that is suitable for solution-based thin film preparation environments, can simultaneously acquire multi-dimensional raw data in the same target area, and can jointly analyze the chemical evolution and physical changes during the film formation process, so as to achieve real-time and accurate monitoring of key kinetic parameters in the solution-based thin film preparation process. Summary of the Invention

[0006] The purpose of this application is to provide an in-situ monitoring method, system, and device for solution-based thin film preparation processes, so as to overcome the shortcomings of traditional technologies, such as difficulty in simultaneously acquiring multi-dimensional process information, difficulty in accurately analyzing film formation kinetics, and inability to achieve real-time in-situ monitoring of the entire process.

[0007] Firstly, this application proposes an in-situ monitoring method for solution-based thin film preparation processes, comprising:

[0008] The method acquires multi-sequence data of at least one target region on the surface of the wafer under test, generated at the same time reference during the fabrication process based on a measurement light source, wherein the measurement light source includes an imaging illumination source, a white broadband light source, and a laser interference source, and the multi-sequence data includes an image sequence, a reflection spectrum sequence, and an interference signal sequence, wherein the multi-sequence data is generated based on the measurement light source sharing at least a portion of the optical path and focusing on the corresponding target region;

[0009] The reflectance spectrum sequence is processed using a signal decoupling algorithm to obtain spectral unmixing results, which include the spectra of individual components in multiple components and their corresponding concentration change curves.

[0010] The volume fraction variation curve of the target region preparation process is obtained by constraining the spectral unmixing result based on the interference signal sequence.

[0011] Based on the volume fraction change curve, process parameters are calculated, spatial correlation between the process parameters and the image sequence is established, and a spatial distribution map is generated.

[0012] In one embodiment, the method further includes:

[0013] The acquired multi-sequence data is spatiotemporally registered and preprocessed before being processed by a signal decoupling algorithm.

[0014] The spatiotemporal registration method includes: aligning the multi-sequence data along the time axis; and establishing a spatial correspondence between the micro-area of ​​reflectance spectroscopy measurement, the interferometric measurement point, and the image pixel position based on a common optical path design.

[0015] The preprocessing methods include at least one of image cropping, denoising, spectral baseline correction, and thickness filtering.

[0016] In one embodiment, the step of processing the reflectance spectral sequence using a signal decoupling algorithm to obtain the spectral demixing result includes:

[0017] The reflectance spectrum sequence is constructed into a spectral data matrix according to the time dimension and the wavelength dimension;

[0018] The spectral data matrix is ​​initialized based on the preset number of components and the initial single-component spectra of each target component;

[0019] A signal decoupling algorithm is used to iteratively demix the initialized spectral data matrix until a preset convergence condition is met. The iterative demixing method includes: calculating the concentration value of each target component at each sampling time when the single component spectrum is fixed, and updating the single component spectrum of each target component when the concentration value of each target component is fixed.

[0020] The single-component spectra of each component when the preset convergence condition is met, as well as the concentration change curves of each component over time, are used as the spectral unmixing results.

[0021] In one embodiment, the step of constraining the spectral unmixing result based on the interference signal sequence to obtain the volume fraction change curve of the target region preparation process includes:

[0022] Based on the interference signal sequence, the thickness change information of the target region is extracted, and the spectral unmixing result is constrained and solved based on the thickness change information to obtain the volume fraction change curve of the target region preparation process; the thickness change information includes at least one of the thickness value, thickness change amount, and thickness change rate of the target region at each sampling time.

[0023] In one embodiment, the step of constraining the spectral unmixing result based on the thickness change information to obtain the volume fraction change curve of the target region preparation process includes:

[0024] Establish the functional relationship between the concentration change curves of each component and the volume fraction of each component in the spectral unmixing results;

[0025] Based on the thickness variation information and the functional relationship, a joint objective function for the target region is constructed;

[0026] Under the constraints of the joint objective function, the concentration change curves of each component are calibrated to volume fraction change curves.

[0027] In one embodiment, the process parameters include precursor solvent evaporation rate, crystal growth rate, nucleation time, and phase transition completion time.

[0028] The process of calculating process parameters based on the volume fraction change curve, establishing a spatial correlation between the process parameters and the image sequence, and generating a spatial distribution map includes:

[0029] The volume fraction change curves of each component are differentiated to obtain the change rate parameters of each target region during the preparation process; and the characteristic time parameters of each target region during the preparation process are determined based on the differential processing results and / or the volume fraction change curves themselves; the change rate parameters include the precursor solvent evaporation rate and crystal growth rate, and the characteristic time parameters include the nucleation time and the phase transition completion time.

[0030] Based on the spatial location of each target region in the image sequence, the process parameters corresponding to each target region are mapped to the corresponding positions in the image sequence;

[0031] A spatial distribution map characterizing the spatial differences in the fabrication process of the wafer surface under test is generated based on the mapped process parameters.

[0032] Secondly, this application proposes an in-situ monitoring system for a solution-based thin film preparation process, the system comprising:

[0033] The acquisition module is used to generate multi-sequence data based on a measurement light source at the same time reference during the fabrication process of at least one target area on the surface of the wafer under test. The measurement light source includes an imaging illumination source, a white broadband light source, and a laser interference source. The multi-sequence data includes an image sequence, a reflection spectrum sequence, and an interference signal sequence. The multi-sequence data is generated based on the measurement light source focusing on the corresponding target area after sharing at least a part of the optical path.

[0034] The processing module is used to process the reflectance spectral sequence using a signal decoupling algorithm to obtain spectral unmixing results, which include the spectra of a single component among multiple components and their corresponding concentration change curves; constrain the spectral unmixing results based on the interference signal sequence to obtain the volume fraction change curve of the target region preparation process; calculate process parameters based on the volume fraction change curve, establish the spatial correlation between the process parameters and the image sequence, and generate a spatial distribution map.

[0035] Thirdly, this application also provides an in-situ monitoring device. The device includes:

[0036] An integrated optical probe module is disposed above the thin film preparation equipment. It is used to provide a measurement light source and focus the measurement light source onto a target area on the surface of the wafer to be measured after sharing at least a portion of the optical path, and to acquire the raw data of the target area based on the measurement light source. The measurement light source includes an imaging illumination source, a white broadband light source, and a laser interference source.

[0037] The data acquisition module is used to acquire the raw data at the same time reference to obtain multi-sequence data; the multi-sequence data includes image sequences, reflectance spectrum sequences, and interference signal sequences;

[0038] The in-situ monitoring system for the solution-based thin film preparation process described in the second aspect is used to process the multi-sequence data and generate a spatial distribution map.

[0039] In one embodiment, the integrated optical probe module includes:

[0040] An imaging unit, including an imaging illumination source and a camera, is used to perform illumination imaging on the target area to obtain image data characterizing the macroscopic spatiotemporal evolution of the target area;

[0041] The microscopic spectroscopy unit includes a white light broadband light source and a fiber optic spectrometer, used to provide broadband measurement light to the target region and collect the corresponding reflectance spectral data of the target region;

[0042] The laser interferometer unit includes a laser interferometer source and an interferometer, used to provide interferometric measurement light to the target region and acquire interferometric signal data characterizing the thickness change of the target region;

[0043] A coaxial objective lens is used to coaxially integrate at least a portion of the shared optical paths of the imaging unit, the microspectral unit, and the laser interferometer unit, and to focus the corresponding measurement light onto the same target area on the surface of the wafer under test.

[0044] In one embodiment, the device further includes:

[0045] A sealed housing, within which the integrated optical probe module is disposed;

[0046] The sealed housing is provided with an annular gas curtain outlet, which is used to spray inert gas onto the outer surface of the probe's optical window to form a protective gas curtain, thereby reducing the contamination of the optical window by solvent vapor.

[0047] The above-mentioned in-situ monitoring method, system, and device for solution-based thin film preparation have at least the following advantages:

[0048] This application acquires image sequences, reflection spectrum sequences, and interference signal sequences under the same time reference, and focuses the imaging illumination source, white broadband light source, and laser interference source on the corresponding target area after sharing at least part of the optical path, ensuring the spatiotemporal synchronous acquisition of multi-source data. By decoupling the reflection spectrum sequence, the spectra of a single component and its corresponding concentration change curve are obtained from multiple components. This decomposes the originally mixed spectral responses into component information with clear physicochemical significance, thereby characterizing the component evolution during the thin film preparation process. By constraining the spectral demixing results based on the interference signal sequence, the independent physical quantity of thickness change can be introduced into the component analysis process, improving the physical consistency and solution accuracy of the volume fraction change curve. Finally, the process parameters are calculated based on the volume fraction change curve, and the spatial correlation between the process parameters and the image sequence is established to generate a spatial distribution map, which can simultaneously obtain the key kinetic parameters and their spatial distribution characteristics during the thin film preparation process. By adopting the above scheme, this application simultaneously acquires and jointly analyzes image information, spectral information and interference information, and realizes the coordinated monitoring of chemical evolution, physical thickness change and spatial distribution characteristics in the solution method thin film preparation process. It effectively overcomes the problems of single information dimension, difficulty in decoupling and coupling sub-processes and difficulty in accurately reflecting the dynamic law of the entire preparation process in traditional technology, thereby improving the monitoring accuracy of the solution method thin film preparation process. Attached Figure Description

[0049] Figure 1 This is a structural block diagram of an in-situ monitoring device in one embodiment;

[0050] Figure 2 This is a flowchart illustrating an in-situ monitoring method for solution-based thin film preparation in one embodiment.

[0051] Figure 3 This is a flowchart illustrating the steps for obtaining spectral unmixing results in one embodiment;

[0052] Figure 4 This is a flowchart illustrating the steps for obtaining the volume fraction change curve in one embodiment;

[0053] Figure 5 This is a flowchart illustrating the steps for generating a spatial distribution map in one embodiment;

[0054] Figure 6 This is a structural block diagram of an in-situ monitoring system for a solution-based thin film preparation process in one embodiment. Detailed Implementation

[0055] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0056] Some exemplary embodiments of this application have been described for illustrative purposes. It should be understood that this application may be implemented in other ways not specifically shown in the accompanying drawings.

[0057] Please see Figure 1 In one exemplary embodiment, this application provides an in-situ monitoring device, the device comprising: an integrated optical probe module, a data acquisition module, and an in-situ monitoring system for solution-based thin film preparation processes.

[0058] An integrated optical probe module is positioned above the thin film preparation equipment to provide a measurement light source and focus the measurement light source onto the target area on the surface of the wafer to be measured after sharing at least part of the optical path, and to acquire the raw data of the target area based on the measurement light source; the measurement light source includes an imaging illumination source, a white broadband light source, and a laser interference source.

[0059] Specifically, the thin film preparation equipment in this embodiment is a solution-based film deposition equipment, used to deposit a precursor solution onto the surface of the wafer under test to form a target thin film. The integrated optical probe module is located above the thin film preparation equipment and directly over the wafer under test area.

[0060] The target area refers to the same local detection micro-region on the surface of the wafer under test used to receive the measurement light source during a single measurement process. Multiple target areas can be preset on the wafer surface. Throughout the monitoring process, the integrated optical probe module measures all target areas sequentially according to a preset order, obtaining raw data for each target area.

[0061] Optionally, the integrated optical probe module includes: an imaging unit, a microspectral unit, a laser interferometer unit, and a coaxial objective lens.

[0062] An imaging unit, including an imaging illumination source and a camera, is used to illuminate and image a target area to acquire image data characterizing the macroscopic spatiotemporal evolution of the target area; wherein the imaging illumination source includes a bright-field or polarized illumination source. Exemplarily, the imaging unit includes a high-speed CMOS camera and a switchable bright-field or polarized illumination source. The camera is coupled to the main optical path via a dichroic mirror to record the macroscopic spatiotemporal evolution of the sample area at a rate of 100-1000 frames per second, including flow fronts, color changes, and birefringent textures.

[0063] The microscopic spectroscopy unit includes a white broadband light source and a fiber optic spectrometer, used to provide broadband measurement light to the target area and acquire the corresponding reflectance spectral data of the target area. Exemplarily, the microscopic spectroscopy unit includes a halogen lamp and a fiber optic spectrometer. The light emitted by the halogen lamp is transmitted through an optical fiber, passes through a beam splitter, and is then incorporated into the main optical path. The spectrometer is used to acquire the reflectance spectral sequence of the measurement point in the wavelength range of 400-1000 nm, with a frequency up to 10 Hz, for analyzing chemical composition and phase evolution. In this embodiment, the diameter of the acquisition point is approximately 50 μm.

[0064] A laser interferometer unit, comprising a laser interferometer source and an interferometer, is used to provide interferometric measurement light to a target region and acquire interferometric signal data characterizing the thickness variation of the target region. Exemplarily, the interferometer can be a low-coherence interferometer, whose laser beam is precisely coaxial with the main optical path via another beam splitter. This laser interferometer unit measures the absolute vertical displacement of a sample at the same point in real time with nanometer-level resolution and a sampling rate up to 1 kHz. In other embodiments, the laser interferometer unit may also use a confocal displacement sensor as needed.

[0065] A coaxial objective lens is used to coaxially integrate at least part of the shared optical paths of the imaging unit, the microspectral unit, and the laser interferometry unit, and to focus the corresponding measurement beams onto the same target area on the surface of the wafer under test. All optical paths are ultimately integrated by a set of shared coaxial objectives, ensuring that the three measurement beams are precisely focused on the same micro-area on the surface of the wafer under test.

[0066] Optionally, the above-mentioned in-situ monitoring device further includes: a sealed housing.

[0067] The integrated optical probe module is housed within a sealed housing. An annular gas curtain outlet is provided on the sealed housing, which sprays inert gas onto the outer surface of the probe's optical window to form a protective gas curtain, reducing solvent vapor contamination of the optical window. In this embodiment, high-purity nitrogen is used as the inert gas.

[0068] The data acquisition module is used to acquire raw data at the same time reference to obtain multi-sequence data; the multi-sequence data includes image sequences, reflectance spectrum sequences and interference signal sequences.

[0069] Optionally, the data acquisition module includes a synchronization controller and a multi-channel data acquisition card. The synchronization controller is communicatively connected to the thin film preparation equipment and is used to receive trigger signals from the equipment during the initial processes of dripping, spin coating, blade coating, slot coating, or annealing. For example, upon receiving a trigger signal, the synchronization controller generates synchronization pulses at a unified time reference and sends them to the camera, fiber optic spectrometer, and interferometer, respectively, to control the high-speed camera, spectrometer, and interferometer to begin data acquisition at a unified timestamp and continuously output the corresponding raw signals at a preset frequency during monitoring.

[0070] A multi-channel data acquisition card is connected to a camera, a fiber optic spectrometer, and an interferometer to synchronously receive raw signals from each sensor. The camera outputs image frame data, the spectrometer outputs a reflectance spectrum array, and the interferometer outputs an interference displacement signal or a corresponding electrical signal characterizing the thickness change of the target region. The multi-channel data acquisition card performs high-speed caching, time stamping, and data packaging on the received raw signals to form multi-sequence data under the same time reference. This multi-sequence data is then transmitted to the in-situ monitoring system of the solution-based thin film preparation process for subsequent processing.

[0071] An in-situ monitoring system for the solution-based thin film fabrication process is connected to a multi-channel data acquisition card to process multi-sequence data and generate spatial distribution maps. Specifically, after acquiring multi-sequence data, the in-situ monitoring system uses a signal decoupling algorithm to process the reflectance spectral sequence, obtaining spectral unmixing results. The unmixing results include the spectra of individual components in multiple components and their corresponding concentration change curves. Constraints are applied to the unmixing results based on the interference signal sequence to obtain the volume fraction change curve of the target region during the fabrication process. Process parameters are calculated based on the volume fraction change curve, and a spatial correlation between the process parameters and the image sequence is established to generate a spatial distribution map.

[0072] The aforementioned in-situ monitoring device acquires image sequences, reflection spectrum sequences, and interference signal sequences under the same time reference. It ensures that the imaging illumination source, white broadband light source, and laser interference source share at least part of their optical paths and focus on the corresponding target area, guaranteeing spatiotemporal synchronous acquisition of multi-source data. By decoupling the reflection spectrum sequence, it obtains the spectra of individual components and their corresponding concentration change curves from multiple components. This decomposes the originally mixed spectral responses into component information with clear physicochemical significance, thus characterizing the component evolution during thin film preparation. Constraint-solving of the spectral demixing results based on the interference signal sequence allows the introduction of thickness change as an independent physical quantity into the component analysis process, improving the physical consistency and solution accuracy of the volume fraction change curve. Finally, process parameters are calculated based on the volume fraction change curve, and a spatial correlation between the process parameters and the image sequence is established to generate a spatial distribution map. This allows for the simultaneous acquisition of key kinetic parameters and their spatial distribution characteristics during thin film preparation. By adopting the above scheme, this application simultaneously acquires and jointly analyzes image information, spectral information and interference information, and realizes the coordinated monitoring of chemical evolution, physical thickness change and spatial distribution characteristics in the solution method thin film preparation process. It effectively overcomes the problems of single information dimension, difficulty in decoupling and coupling sub-processes and difficulty in accurately reflecting the dynamic law of the entire preparation process in traditional technology, thereby improving the monitoring accuracy of the solution method thin film preparation process.

[0073] Please see Figure 2 In one exemplary embodiment, this application provides an in-situ monitoring method for solution-based thin film preparation processes, specifically including the following steps:

[0074] Step 202: Obtain multi-sequence data of at least one target region on the surface of the wafer under test, generated based on the measurement light source at the same time reference during the fabrication process.

[0075] Specifically, the measurement light sources include imaging illumination sources, white broadband light sources, and laser interferometric sources. Multi-sequence data is generated based on the measurement light sources sharing at least a portion of their optical paths and focusing onto the corresponding target region; it includes image sequences, reflectance spectrum sequences, and interferometric signal sequences.

[0076] The aforementioned measurement light sources share at least a portion of their optical paths on the side closest to the surface of the wafer under test, and are focused onto the target region via a coaxial objective lens. This ensures that image acquisition, spectral acquisition, and interferometric acquisition correspond to the same micro-region on the surface of the wafer under test. Therefore, at any time t during the thin film fabrication process, the selected target region can be measured, yielding corresponding multi-sequence data.

[0077] For example, to facilitate the description of the correspondence between multimodal data in the time and spatial dimensions, the multi-sequence data obtained at any time t and at a spatial point (x, y) on the surface of the wafer under test are defined as including an image sequence V(t), a reflection spectrum sequence I(t, λ), and an interference signal sequence H(t), where λ is the wavelength; the image sequence V(t) represents the change of the pixel value corresponding to the spatial point (x, y) in the camera image over time, and the pixel value can be a gray value, a color vector, or a polarization intensity; the reflection spectrum sequence I(t, λ) represents the reflection spectrum measured by the spectrometer at time t at the spatial point (x, y); and H(t) represents the absolute physical thickness or thickness change information measured by the interferometer at time t at the spatial point (x, y).

[0078] Step 204: The reflectance spectrum sequence is processed using a signal decoupling algorithm to obtain the spectral unmixing results. The spectral unmixing results include the spectra of individual components in multiple components and their corresponding concentration change curves.

[0079] Specifically, since multiple components such as precursor solution, solvent residue, intermediate phase, crystalline phase, and / or final solid phase usually coexist in the target region during thin film preparation, a single spectrum in the reflectance spectrum sequence is a mixed spectrum formed by the superposition of the spectral responses of these multiple components, rather than an independent spectrum of a single component. Based on this, this application employs a signal decoupling algorithm to decompose the mixed spectrum into multiple single-component spectra with physical or chemical significance and their corresponding concentration change curves. These concentration change curves characterize the contribution of a single component to the changes over time during thin film preparation.

[0080] Step 206: Based on the interference signal sequence, constrain the spectral unmixing results to obtain the volume fraction change curve of the target region preparation process.

[0081] Specifically, since the concentration change curve represents the relative contribution of each component to the mixed spectrum, the response intensity of the reflectance spectrum is affected not only by the components but also by factors such as the refractive index, extinction coefficient, scattering characteristics, optical path length, and film thickness changes of each component. Therefore, it does not directly correspond to the actual volume fraction of each component in the target region. In contrast, the interference signal sequence represents the true thickness change information. By introducing the interference signal sequence for constraint solving, an independent physical constraint is added to the spectral inversion, ensuring that the changes in each component not only explain the changes in the reflectance spectrum but also align with the actual film thickness evolution, thus obtaining evolution results that are closer to the actual evolution process.

[0082] Step 208: Calculate process parameters based on the volume fraction change curve, establish spatial correlation between process parameters and image sequence, and generate spatial distribution map.

[0083] Specifically, the volume fraction change curve refers to the functional relationship between the volume fraction of each component in the target region and time during the thin film preparation process, reflecting the actual evolution state of different components in the target region. Process parameters refer to parameters extracted based on the volume fraction change curve, used to characterize the kinetics and process state of thin film preparation. For example, process parameters include precursor solvent evaporation rate, crystal growth rate, nucleation time, and phase transition completion time. Among these, the solvent evaporation rate characterizes the rate of solvent component reduction; the crystal growth rate characterizes the rate of crystalline phase formation and growth; the nucleation time characterizes the moment when stable nucleation begins in the target region; and the phase transition completion time characterizes the moment when the transformation from the precursor state to the target crystalline phase is essentially complete.

[0084] Furthermore, since each target region corresponds to a specific pixel location or pixel region in the image sequence, establishing a spatial correlation between the process parameters at different locations and the image sequence allows the obtained process parameters to have not only evolutionary significance in the temporal dimension but also distributional significance in the spatial dimension. The resulting spatial distribution map represents the graphical result of the distribution of process parameters at different locations on the surface of the wafer under test. It can intuitively reflect the differences in nucleation sequence, crystallization rate, volatilization uniformity, and phase transition process in different regions during thin film preparation, thereby characterizing the spatial uniformity of thin film preparation, identifying abnormal regions, and assisting in subsequent process optimization.

[0085] The aforementioned in-situ monitoring method for solution-based thin film preparation acquires image sequences, reflection spectrum sequences, and interference signal sequences at the same time reference. By ensuring that the imaging illumination source, white broadband light source, and laser interference source share at least part of their optical paths and focus on the corresponding target region, the spatiotemporal synchronous acquisition of multi-source data is guaranteed. By decoupling the reflection spectrum sequence, the spectra of individual components and their corresponding concentration change curves are obtained, decomposing the originally mixed spectral responses into component information with clear physicochemical significance, thus characterizing the component evolution during thin film preparation. Constraints are applied to the spectral demixing results based on the interference signal sequence, introducing the independent physical quantity of thickness change into the component analysis process, improving the physical consistency and accuracy of the volume fraction change curve. Finally, process parameters are calculated based on the volume fraction change curve, and a spatial correlation between the process parameters and the image sequence is established to generate a spatial distribution map, enabling the simultaneous acquisition of key kinetic parameters and their spatial distribution characteristics during thin film preparation. By adopting the above scheme, this application simultaneously acquires and jointly analyzes image information, spectral information and interference information, and realizes the coordinated monitoring of chemical evolution, physical thickness change and spatial distribution characteristics in the solution method thin film preparation process. It effectively overcomes the problems of single information dimension, difficulty in decoupling and coupling sub-processes and difficulty in accurately reflecting the dynamic law of the entire preparation process in traditional technology, thereby improving the monitoring accuracy of the solution method thin film preparation process.

[0086] Optionally, the in-situ monitoring method for the above-mentioned solution-based thin film preparation process further includes:

[0087] After spatiotemporal registration and preprocessing of the acquired multi-sequence data, signal decoupling algorithm processing is performed. The spatiotemporal registration methods include: aligning the time axis of the multi-sequence data; and establishing the spatial correspondence between the micro-area of ​​reflectance spectral measurement, the interferometric measurement point and the image pixel position based on the common optical path design. The preprocessing methods include at least one of image cropping, denoising, spectral baseline correction and thickness filtering.

[0088] Specifically, spatiotemporal registration refers to establishing a correspondence between multiple data sequences in both time and space dimensions to ensure that data acquired from different channels consistently point to the same fabrication state and target region on the surface of the wafer under test. In this embodiment, time axis alignment can be achieved through hardware timestamps recorded by a synchronization controller. Spatial correspondence can be established based on a common optical path design, ensuring that the measurement micro-area of ​​the reflection spectrum, the measurement point of the interference signal, and the pixel position or pixel region in the camera image correspond to each other, thereby maintaining spatial consistency between spectral information, thickness information, and image information.

[0089] The purpose of preprocessing is to adaptively correct multi-sequence data before signal decoupling, thereby improving the stability and accuracy of subsequent analysis. In this embodiment, the preprocessing methods include at least one of image cropping, denoising, spectral baseline correction, and thickness filtering. Image cropping is used to extract the effective image region corresponding to the target region; denoising is used to reduce the impact of acquisition noise on the image, spectrum, or interference signal; spectral baseline correction is used to reduce the interference of background drift or uneven system response on the reflection spectrum; and thickness filtering is used to smooth random fluctuations in the interference signal and preserve the true thickness variation trend.

[0090] By adopting the above scheme, spatiotemporal registration of multi-sequence data before signal decoupling can ensure that the image sequence, reflectance spectrum sequence and interference signal sequence are synchronized in time and correspond in space, so that different modal data can jointly reflect the real fabrication state of the same target area on the surface of the wafer under test at the same time. At the same time, preprocessing can reduce the impact of noise interference and background drift on subsequent unmixing analysis, thereby improving the accuracy and stability of signal decoupling results.

[0091] Please see Figure 3 Optionally, a signal decoupling algorithm is used to process the reflectance spectrum sequence to obtain the spectral unmixing result, including:

[0092] Step 302: Construct a spectral data matrix from the reflectance spectral sequence according to the time dimension and the wavelength dimension.

[0093] Step 304: Initialize the spectral data matrix based on the preset number of components and the initial single-component spectra of each target component.

[0094] Step 306: The signal decoupling algorithm is used to iteratively demix the initialized spectral data matrix until the preset convergence condition is met. The iterative demixing method includes: calculating the concentration value of each target component at each sampling time when the single component spectrum is fixed, and updating the single component spectrum of each target component when the concentration value of each target component is fixed.

[0095] Step 308: The single-component spectra of each component when the preset convergence condition is met, as well as the concentration change curves of each component over time, are taken as the spectral unmixing results.

[0096] Specifically, for the reflectance spectrum sequence of the target area acquired at multiple sampling times, the collected spectral data at T times and L wavelengths are arranged into a two-dimensional matrix D (dimension T×L).

[0097] The preset number of components is used to characterize the number of major components involved in the spectral response of the target region preparation process; the initial single-component spectra of each target component are used as initial data for iterative unmixing. For example, the preset number of components can be set based on prior information; the initial single-component spectra can be initial characteristic spectral estimates of the solvent component, intermediate phase component, crystallized phase component, and / or the final solid phase component. Initializing the number of components and the initial single-component spectra provides initial boundaries and convergence directions for subsequent iterative unmixing.

[0098] Furthermore, in this embodiment, the multivariate curve-resolved alternating least squares (MCR-ALS) method is used to process the reflectance spectral sequence I(t,λ). It is assumed that the measured spectrum I(t,λ) is a linear combination of the spectra of several pure components (endmembers) in the system, and its expression is:

[0099]

[0100] in: The characteristic spectrum of the kth single component is represented, for example, S_solvent(λ), S_intermediate phase(λ), S_perovskite crystal(λ); Let represent the relative concentration of the k-th pure component at time t; e represents the residual term; and K represents the preset number of components.

[0101] Within one iteration cycle, the individual component spectra of each target component are kept unchanged, and the concentration values ​​of each target component at each sampling time are obtained based on least squares solution. Then, keeping the obtained concentration values ​​of each target component unchanged, the individual component spectra of each target component are updated based on least squares solution. Subsequently, the above alternating solution process is repeated so that the reconstruction result of the spectral data matrix gradually approximates the actual measurement result.

[0102] Optionally, the preset convergence conditions include at least one of the following: the change in residual between two adjacent iterations is less than a preset threshold, the decrease in the objective function is less than a preset threshold, or the preset number of iterations is reached.

[0103] After reaching the preset convergence condition, the system outputs the individual component spectra of each target component and the concentration change curves of each target component at each sampling time. The individual component spectra characterize the characteristic response of the corresponding component in the wavelength dimension, while the concentration change curves characterize the relative evolution trend of the corresponding component in the time dimension. For example, the individual component spectra and concentration change curves of the solvent component and the crystallized phase component can be obtained, thereby separating the contributions of different components in the mixed reflectance spectrum.

[0104] By employing the above scheme, constructing the reflectance spectral sequence into a spectral data matrix and initializing it based on a preset number of components and initial single-component spectra, a clear solution boundary can be provided for signal decoupling. Through iterative demixing between fixed single-component spectra and fixed concentration values, the independent spectral characteristics of each target component in the reflectance spectral sequence and its time-varying concentration contribution can be gradually separated, thereby achieving the distinction between different sub-processes. Thus, this application can decompose the originally coupled mixed spectral signals into interpretable component information, improving the ability to identify the component evolution laws during solution-based thin film preparation and the accuracy of subsequent process parameter solutions.

[0105] Please see Figure 4 Optionally, the spectral unmixing results are constrained and solved based on the interference signal sequence to obtain the volume fraction change curve of the target region preparation process, including:

[0106] Step 402: Extract thickness variation information of the target region based on the interference signal sequence.

[0107] Step 404: Based on the thickness change information, constrain the spectral unmixing results to obtain the volume fraction change curve of the target region preparation process; the thickness change information includes at least one of the thickness value, thickness change amount, and thickness change rate of the target region at each sampling time.

[0108] Specifically, the interference signal sequence is a time-series signal continuously acquired by the interferometer during the preparation of the target region, used to characterize the thickness state of the target region at each sampling time. Based on the interference signal sequence, at least one of the following can be extracted as thickness change information: the thickness value of the target region at each sampling time, the amount of thickness change relative to the initial time, and the rate of thickness change.

[0109] Optionally, the spectral unmixing results are constrained and solved based on the thickness variation information to obtain the volume fraction variation curve of the target region preparation process, including:

[0110] Establish the functional relationship between the concentration change curves of each component and the volume fraction of each component in the spectral unmixing results; construct a joint objective function for the target region based on the thickness change information and the functional relationship; and calibrate the concentration change curves of each component to volume fraction change curves under the constraint of the joint objective function.

[0111] Specifically, this application first establishes a functional relationship between the concentration variation curves of each component and the volume fraction of each component in the spectral unmixing results. For example, when the thin film in the target region is considered as a mixed layer composed of solvent and solid components, without considering scattering, the relationship between the effective refractive index of the mixed layer and the volume fraction of each component can be established based on the effective medium approximation theory, and its expression is:

[0112]

[0113] in, Indicates the target region at time... The effective refractive index, Indicates the solvent component at time [time]. volume fraction, Indicates the solid component at time [time]. volume fraction, Represents the refractive index of the solvent component. It represents the refractive index of the solid component.

[0114] Secondly, based on the thickness variation information and the aforementioned functional relationship, a joint objective function for the target region is constructed, ensuring that the solution simultaneously satisfies the requirements for reflecting spectrum fitting and thickness variation fitting. For example, the joint objective function can be expressed as:

[0115]

[0116] in, This represents the measured reflectance spectrum. This represents the model reflectance spectrum calculated based on volume fraction and effective refractive index. This indicates the measured thickness variation information. This indicates the model thickness variation information calculated based on changes in component volume fraction. This represents the weighting factor.

[0117] Finally, under the constraint of the joint objective function, the concentration change curves of each component are calibrated to convert them into volume fraction change curves. For example, the concentration change curve of the solvent component can be calibrated into a solvent volume fraction change curve. The concentration variation curve of the crystalline phase component was calibrated to a solid volume fraction variation curve. This yields the volume fraction change curve during the preparation of the target region. In one example, the above calibration process can also be combined with the boundary conditions of the preparation process. For example, at the beginning of the process... The target region is mainly in a liquid state, and the volume fraction of the solvent component can be set to meet the requirements. At the end of the process, the final thickness is determined based on the interferometer measurement. This allows for the determination or calibration of the final solid component volume fraction. Furthermore, during calibration, physical consistency can be determined based on the synergistic relationship between the component change rate and the thickness change rate. For example, if the crystalline phase concentration change curve obtained from spectral demixing... The rate of change and the thickness shrinkage rate measured by the interferometer If the preset association relationship is not met, the corresponding unmixing result is determined to not meet the physical consistency requirements, and the unmixing result is corrected or removed.

[0118] By adopting the above scheme and introducing the effective refractive index model formula, a clear relationship can be established between the volume fraction of the components and the changes in the optical parameters of the target region. By introducing the joint objective function formula, the fitting error of the reflection spectrum and the fitting error of the thickness change can be incorporated into the unified solution framework at the same time. This makes the calibration results of the concentration change curve not only conform to the spectral response law, but also conform to the true thickness evolution law of the target region, thereby improving the physical consistency and solution accuracy of the volume fraction change curve.

[0119] Please see Figure 5 Optionally, when the process parameters include precursor solvent evaporation rate, crystal growth rate, nucleation time, and phase transition completion time, the process parameters are calculated based on the volume fraction change curve, a spatial correlation between the process parameters and the image sequence is established, and a spatial distribution map is generated, including:

[0120] Step 502: Differentiate the volume fraction change curves of each component to obtain the change rate parameters of each target region during the preparation process; and determine the characteristic time parameters of each target region during the preparation process based on the differential processing results and / or the volume fraction change curves themselves; the change rate parameters include the precursor solvent evaporation rate and crystal growth rate, and the characteristic time parameters include the nucleation time and the phase transition completion time.

[0121] Step 504: Based on the spatial location of each target region in the image sequence, map the process parameters corresponding to each target region to the corresponding location in the image sequence.

[0122] Step 506: Generate a spatial distribution map characterizing the spatial differences in the fabrication process of the wafer surface under test based on the mapped process parameters.

[0123] Specifically, the volume fraction variation curve characterizes the actual evolution of each component in the target region during the preparation process. For example, the volume fraction variation curve of the precursor solvent is obtained... Volume fraction variation curve of crystalline phase Then, it is subjected to differential processing to determine the precursor solvent evaporation rate. and crystal growth rate Their expressions are as follows:

[0124]

[0125]

[0126] Among them, the precursor solvent evaporation rate is used to characterize how quickly the volume fraction of the precursor solvent decreases over time, and the crystal growth rate is used to characterize how quickly the volume fraction of the crystallized phase increases over time.

[0127] Furthermore, characteristic time parameters can be determined based on the differential processing results and / or the volume fraction change curve itself. For example, nucleation time can be defined as the crystal growth rate. The moment when the volume fraction of the crystalline phase is significantly greater than the noise starting from zero, or the curve showing the change in the volume fraction of the crystalline phase. The inflection point at which the acceleration begins; the time to complete the phase transformation can be defined as the crystal growth rate. The time at which the temperature drops to a preset ratio after reaching its peak value. Thus, at least one process parameter can be obtained for each target region, including the precursor solvent evaporation rate, crystal growth rate, nucleation time, and phase transition completion time.

[0128] Furthermore, each target region corresponds to a different pixel position or pixel region in the image sequence. Therefore, based on the spatial position of each target region in the image coordinate system, the process parameters corresponding to each target region can be mapped to the corresponding positions in the image sequence. For example, the nucleation time, peak crystal growth rate, precursor solvent evaporation rate, or phase transition completion time of each target region can be assigned to the corresponding image coordinate points to establish a spatial correlation between process parameters and the image sequence.

[0129] After completing the spatial mapping of process parameters and image sequences, the mapping results can be processed to generate a spatial distribution map. The processing methods include interpolation, meshing, color encoding, and visualization. For example, when mapping and visualizing the nucleation time of each target region, a nucleation time distribution map can be generated to characterize the differences in the order of nucleation at different locations on the surface of the wafer under test; when mapping and visualizing the crystal growth rate of each target region, a crystal growth rate distribution map can be generated to characterize the differences in crystallization kinetics at different locations on the surface of the wafer under test; when mapping and visualizing the precursor solvent evaporation rate of each target region, an evaporation rate distribution map can be generated to characterize the evaporation uniformity in different regions of the surface of the wafer under test; and when mapping and visualizing the phase transition completion time of each target region, a phase transition completion time distribution map can be generated to characterize the differences in the phase transition process in different regions of the surface of the wafer under test.

[0130] Ultimately, the output analysis data includes: spectra of multiple single components, volume fraction variation curves of each component, process parameters extracted based on the volume fraction variation curves, and spatial distribution maps generated based on the process parameters corresponding to multiple target regions.

[0131] The above-mentioned in-situ monitoring method for solution-based thin film preparation acquires image sequences, reflection spectrum sequences, and interference signal sequences at the same time reference, and focuses each measurement light source on the corresponding target area after sharing at least part of the optical path. This enables synchronous acquisition of multi-source data in the time dimension and corresponding acquisition in the spatial dimension, thereby ensuring that data from different modes jointly characterize the true preparation state of the same target area on the surface of the wafer under test at the same time, and improving the accuracy and consistency of subsequent joint analysis.

[0132] Furthermore, this application obtains the spectra of a single component and its corresponding concentration change curve from multiple components by performing signal decoupling processing on the reflection spectrum sequence. This enables the separation of different component responses that were originally mixed in the reflection spectrum, thereby achieving the identification of different material states such as solvent, precursor, intermediate phase, crystalline phase and / or final solid phase, and improving the ability to analyze the chemical evolution law during thin film preparation.

[0133] Furthermore, this application uses the interference signal sequence to constrain the solution of the spectral unmixing results, which enables joint calibration of the relative concentration change results obtained from spectral unmixing with the actual thickness change process of the target region. This further transforms the concentration change curve into a volume fraction change curve with physical meaning, thereby improving the physical consistency, reliability and solution accuracy of the component evolution results.

[0134] Furthermore, by extracting process parameters such as precursor solvent evaporation rate, crystal growth rate, nucleation time, and phase transition completion time from the volume fraction change curve, this application can further transform the component evolution information into characteristic parameters that can directly characterize the kinetic behavior of thin film preparation, thereby providing a quantitative basis for analyzing the thin film preparation mechanism and evaluating the process status.

[0135] Furthermore, by establishing a spatial correlation between process parameters and image sequences and generating a spatial distribution map, this application can intuitively reflect the spatial differences at different locations on the surface of the wafer under test in terms of nucleation sequence, crystallization rate, volatilization uniformity, and phase transition process, thereby achieving a visual characterization of the uniformity of thin film preparation, abnormal regions, and the trend of local defect formation.

[0136] Furthermore, this application combines image information, spectral information, and thickness information for joint analysis, thereby effectively overcoming the shortcomings of existing technologies, such as insufficient information dimensions of single sensing methods, difficulty in distinguishing coupled sub-processes, and difficulty in accurately characterizing the dynamic preparation process of solution-based thin films, and realizing qualitative identification and quantitative analysis of complex non-equilibrium film formation processes.

[0137] Furthermore, since the monitoring results obtained in this application simultaneously include single-component spectra, volume fraction change curves, process parameters, and spatial distribution parameters, they can not only be used for process monitoring, but also provide a high-quality data foundation for process optimization, parameter adjustment, anomaly diagnosis, and subsequent establishment of intelligent control models, thereby improving the monitoring depth and process control capability of the solution-based thin film preparation process.

[0138] In summary, by adopting the above scheme, this application can achieve simultaneous perception and joint analysis of chemical evolution, physical thickness changes and spatial distribution characteristics during the preparation of the wafer surface under test, accurately reflect the dynamic law of the entire preparation process, improve the accuracy, reliability and information integrity of in-situ monitoring, and provide effective technical support for the precision preparation and intelligent optimization of solution-based thin film materials.

[0139] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0140] Based on the same inventive concept, this application also provides an in-situ monitoring system for solution-based thin film preparation process. This system is applicable to the above-mentioned in-situ monitoring method for solution-based thin film preparation process. The solution provided by this system is similar to the solution described in the above method. Therefore, the specific limitations in one or more system embodiments provided below can be found in the limitations of the method above, and will not be repeated here.

[0141] Please see Figure 6 In one embodiment, the in-situ monitoring system for the solution-based thin film preparation process includes an acquisition module and a processing module.

[0142] The acquisition module is used to acquire multiple sequence data generated by a measurement light source based on the same time reference during the fabrication process of at least one target area on the surface of the wafer under test. The measurement light source includes an imaging illumination source, a white broadband light source, and a laser interference source. The multiple sequence data includes an image sequence, a reflection spectrum sequence, and an interference signal sequence. The multiple sequence data is generated based on the measurement light source sharing at least part of the optical path and focusing on the corresponding target area.

[0143] The processing module is used to process the reflection spectrum sequence using a signal decoupling algorithm to obtain the spectral unmixing result, which includes the spectrum of a single component in multiple components and its corresponding concentration change curve; constrain the spectral unmixing result based on the interference signal sequence to obtain the volume fraction change curve of the target region preparation process; calculate the process parameters based on the volume fraction change curve, establish the spatial correlation between the process parameters and the image sequence, and generate a spatial distribution map.

[0144] Optionally, the in-situ monitoring system for the above-mentioned solution-based thin film preparation process further includes: a registration and pretreatment module.

[0145] The registration and preprocessing module is used to perform spatiotemporal registration and preprocessing on the acquired multi-sequence data before signal decoupling algorithm processing. The spatiotemporal registration methods include: aligning the time axis of the multi-sequence data; and establishing the spatial correspondence between the micro-area of ​​reflectance spectroscopy measurement, the interferometric measurement point and the image pixel position based on the common optical path design. The preprocessing methods include at least one of image cropping, denoising, spectral baseline correction and thickness filtering.

[0146] Optionally, the processing module uses a signal decoupling algorithm to process the reflectance spectral sequence to obtain spectral unmixing results, including: constructing a spectral data matrix from the reflectance spectral sequence according to the time and wavelength dimensions; initializing the spectral data matrix based on the preset number of components and the initial single-component spectra of each target component; using a signal decoupling algorithm to iteratively unmix the initialized spectral data matrix until a preset convergence condition is met; the iterative unmixing method includes: calculating the concentration values ​​of each target component at each sampling time while fixing the single-component spectra, and updating the single-component spectra of each target component while fixing the concentration values ​​of each target component; and using the single-component spectra of each component when the preset convergence condition is met and the concentration change curves of each component over time as the spectral unmixing results.

[0147] Optionally, the processing module performs constraint solving on the spectral unmixing results based on the interference signal sequence to obtain the volume fraction change curve of the target region preparation process, including: extracting thickness change information of the target region based on the interference signal sequence; performing constraint solving on the spectral unmixing results based on the thickness change information to obtain the volume fraction change curve of the target region preparation process; the thickness change information includes at least one of the thickness value, thickness change amount, and thickness change rate of the target region at each sampling time. Specifically, performing constraint solving on the spectral unmixing results based on the thickness change information to obtain the volume fraction change curve of the target region preparation process includes: establishing a functional relationship between the concentration change curves of each component and the volume fraction of each component in the spectral unmixing results; constructing a joint objective function for the target region based on the thickness change information and the functional relationship; and calibrating the concentration change curves of each component to volume fraction change curves under the constraint of the joint objective function.

[0148] Optionally, when the process parameters include precursor solvent evaporation rate, crystal growth rate, nucleation time, and phase transition completion time, the processing module calculates the process parameters based on the volume fraction change curve, establishes a spatial correlation between the process parameters and the image sequence, and generates a spatial distribution map. This includes: differentiating the volume fraction change curves of each component to obtain the change rate parameters of each target region during the preparation process; and determining the characteristic time parameters of each target region during the preparation process based on the differential processing results and / or the volume fraction change curves themselves; the change rate parameters include the precursor solvent evaporation rate and crystal growth rate, and the characteristic time parameters include the nucleation time and phase transition completion time; mapping the process parameters corresponding to each target region to the corresponding positions in the image sequence according to the spatial position of each target region; and generating a spatial distribution map characterizing the spatial differences in the preparation process of the wafer surface under test based on the mapped process parameters.

[0149] The in-situ monitoring system for the solution-based thin film preparation process described above acquires image sequences, reflection spectrum sequences, and interference signal sequences at the same time reference. It ensures that the imaging illumination source, white broadband light source, and laser interference source share at least part of their optical paths and focus on the corresponding target region, guaranteeing simultaneous spatiotemporal acquisition of multi-source data. By decoupling the reflection spectrum sequence, it obtains the spectra of individual components and their corresponding concentration change curves, decomposing the originally mixed spectral responses into component information with clear physicochemical significance, thereby characterizing the component evolution during thin film preparation. Constraint-based solutions to the spectral demixing results based on the interference signal sequence introduce the independent physical quantity of thickness change into the component analysis process, improving the physical consistency and solution accuracy of the volume fraction change curve. Finally, process parameters are calculated based on the volume fraction change curve, and a spatial correlation between the process parameters and the image sequence is established to generate a spatial distribution map, enabling the simultaneous acquisition of key kinetic parameters and their spatial distribution characteristics during thin film preparation. By adopting the above scheme, this application simultaneously acquires and jointly analyzes image information, spectral information and interference information, and realizes the coordinated monitoring of chemical evolution, physical thickness change and spatial distribution characteristics in the solution method thin film preparation process. It effectively overcomes the problems of single information dimension, difficulty in decoupling and coupling sub-processes and difficulty in accurately reflecting the dynamic law of the entire preparation process in traditional technology, thereby improving the monitoring accuracy of the solution method thin film preparation process.

[0150] Each module in the in-situ monitoring system for the above-mentioned solution-based thin film preparation process can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0151] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0152] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0153] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An in-situ monitoring method for solution-based thin film preparation process, characterized in that, The method includes: The method acquires multi-sequence data of at least one target region on the surface of the wafer under test, generated at the same time reference during the fabrication process based on a measurement light source, wherein the measurement light source includes an imaging illumination source, a white broadband light source, and a laser interference source, and the multi-sequence data includes an image sequence, a reflection spectrum sequence, and an interference signal sequence, wherein the multi-sequence data is generated based on the measurement light source sharing at least a portion of the optical path and focusing on the corresponding target region; The reflectance spectral sequence is processed using a signal decoupling algorithm to obtain spectral unmixing results, including: constructing a spectral data matrix from the reflectance spectral sequence according to the time dimension and wavelength dimension, and iteratively unmixing the spectral data matrix using a signal decoupling algorithm to obtain the spectra of a single component among multiple components and its corresponding concentration change curves; The volume fraction variation curve of the target region preparation process is obtained by constraining the spectral unmixing result based on the interference signal sequence, including: extracting the thickness variation information of the target region based on the interference signal sequence, and establishing a functional relationship between the concentration variation curve of each component and the volume fraction of each component in the spectral unmixing result; constructing a joint objective function for the target region based on the thickness variation information and the functional relationship; calibrating the concentration variation curve of each component to a volume fraction variation curve under the constraint of the joint objective function; calculating process parameters based on the volume fraction variation curve, establishing a spatial correlation between the process parameters and the image sequence, and generating a spatial distribution map, including: adjusting the volume fraction variation curve of each component. The process curves are differentiated to obtain the rate of change parameters of each target region during the preparation process; and the characteristic time parameters of each target region during the preparation process are determined based on the differential processing results and / or the volume fraction change curves themselves; the process parameters include precursor solvent evaporation rate, crystal growth rate, nucleation time, and phase transition completion time; the rate of change parameters include precursor solvent evaporation rate and crystal growth rate, and the characteristic time parameters include nucleation time and phase transition completion time; according to the spatial position of each target region in the image sequence, the process parameters corresponding to each target region are mapped to the corresponding positions in the image sequence; a spatial distribution map characterizing the spatial differences in the preparation process of the wafer surface under test is generated based on the mapped process parameters.

2. The method according to claim 1, characterized in that, The method further includes: The acquired multi-sequence data is spatiotemporally registered and preprocessed before being processed by a signal decoupling algorithm. The spatiotemporal registration method includes: aligning the multi-sequence data along the time axis; and establishing a spatial correspondence between the micro-area of ​​reflectance spectroscopy measurement, the interferometric measurement point, and the image pixel position based on a common optical path design. The preprocessing methods include at least one of image cropping, denoising, spectral baseline correction, and thickness filtering.

3. The method according to claim 1, characterized in that, The step of processing the reflectance spectral sequence using a signal decoupling algorithm to obtain the spectral demixing result further includes: The spectral data matrix is ​​initialized based on the preset number of components and the initial single-component spectra of each target component; A signal decoupling algorithm is used to iteratively demix the initialized spectral data matrix until a preset convergence condition is met. The iterative demixing method includes: calculating the concentration value of each target component at each sampling time when the single component spectrum is fixed, and updating the single component spectrum of each target component when the concentration value of each target component is fixed. The single-component spectra of each component when the preset convergence condition is met, as well as the concentration change curves of each component over time, are used as the spectral unmixing results.

4. The method according to claim 1, characterized in that: The thickness change information includes at least one of the following: the thickness value of the target region at each sampling time, the amount of thickness change, and the rate of thickness change.

5. An in-situ monitoring system for a solution-based thin film preparation process, characterized in that, The system includes: The acquisition module is used to generate multi-sequence data based on a measurement light source at the same time reference during the fabrication process of at least one target area on the surface of the wafer under test. The measurement light source includes an imaging illumination source, a white broadband light source, and a laser interference source. The multi-sequence data includes an image sequence, a reflection spectrum sequence, and an interference signal sequence. The multi-sequence data is generated based on the measurement light source focusing on the corresponding target area after sharing at least a part of the optical path. The processing module is used to process the reflectance spectral sequence using a signal decoupling algorithm to obtain spectral unmixing results, including: constructing a spectral data matrix from the reflectance spectral sequence according to the time and wavelength dimensions; iteratively unmixing the spectral data matrix using the signal decoupling algorithm to obtain the spectra of individual components and their corresponding concentration change curves; constraining the spectral unmixing results based on the interference signal sequence to obtain the volume fraction change curves of the target region preparation process, including: extracting the thickness change information of the target region based on the interference signal sequence; establishing a functional relationship between the concentration change curves of each component and the volume fraction of each component in the spectral unmixing results; constructing a joint objective function for the target region based on the thickness change information and the functional relationship; and calibrating the concentration change curves of each component to volume fraction change curves under the constraints of the joint objective function. The calculation of process parameters based on the volume fraction change curves, the establishment of spatial correlation between the process parameters and the image sequence, and the generation of a spatial distribution map include: differentiating the volume fraction change curves of each component to obtain the rate of change parameters of each target region during the preparation process; and determining the characteristic time parameters of each target region during the preparation process based on the differential processing results and / or the volume fraction change curves themselves; the process parameters include precursor solvent evaporation rate, crystal growth rate, nucleation time, and phase transition completion time; the rate of change parameters include precursor solvent evaporation rate and crystal growth rate; and the characteristic time parameters include nucleation time and phase transition completion time; mapping the process parameters corresponding to each target region to the corresponding positions in the image sequence according to the spatial position of each target region; and generating a spatial distribution map characterizing the spatial differences in the preparation process of the wafer surface under test based on the mapped process parameters.

6. An in-situ monitoring device, characterized in that, The device includes: An integrated optical probe module is disposed above the thin film preparation equipment. It is used to provide a measurement light source and focus the measurement light source onto a target area on the surface of the wafer to be measured after sharing at least a portion of the optical path, and to acquire the raw data of the target area based on the measurement light source. The measurement light source includes an imaging illumination source, a white broadband light source, and a laser interference source. The data acquisition module is used to acquire the raw data at the same time reference to obtain multi-sequence data; the multi-sequence data includes image sequences, reflectance spectrum sequences, and interference signal sequences; The in-situ monitoring system for the solution-based thin film preparation process as described in claim 5 is used to process the multi-sequence data and generate a spatial distribution map.

7. The apparatus according to claim 6, characterized in that, The integrated optical probe module includes: An imaging unit, including an imaging illumination source and a camera, is used to perform illumination imaging on the target area to obtain image data characterizing the macroscopic spatiotemporal evolution of the target area; The microscopic spectroscopy unit includes a white light broadband light source and a fiber optic spectrometer, used to provide broadband measurement light to the target region and collect the corresponding reflectance spectral data of the target region; The laser interferometer unit includes a laser interferometer source and an interferometer, used to provide interferometric measurement light to the target region and acquire interferometric signal data characterizing the thickness change of the target region; A coaxial objective lens is used to coaxially integrate at least a portion of the shared optical paths of the imaging unit, the microspectral unit, and the laser interferometer unit, and to focus the corresponding measurement light onto the same target area on the surface of the wafer under test.

8. The apparatus according to claim 6, characterized in that, The device further includes: A sealed housing, within which the integrated optical probe module is disposed; The sealed housing is provided with an annular gas curtain outlet, which is used to spray inert gas onto the outer surface of the probe's optical window to form a protective gas curtain, thereby reducing the contamination of the optical window by solvent vapor.

Images