Reversion method, system and detection device of semiconductor epitaxial growth parameters and related products

Abstract

This application provides a method, system, detection device, and related products for inverting semiconductor epitaxial growth parameters, relating to the field of semiconductor detection technology. This application calculates the absorption characteristic parameters of the gas phase components based on current process parameters and a pre-stored gas molecular spectral absorption database, and further obtains the gas phase transmittance of the probe optical path. It then quantitatively compensates for the measured reflectance spectrum during the semiconductor epitaxial growth process, making the compensated target reflectance spectrum closer to the true optical response of the semiconductor epitaxial film. This effectively reduces spectral distortion and systematic errors caused by the superposition of gas phase absorption, significantly improving the accuracy and stability of the inverted growth parameters. Furthermore, the quantitatively compensated values ​​can be adaptively updated according to fluctuations in operating conditions, reducing the inversion deviation introduced by operating condition drift, improving process consistency and comparability, and facilitating more reliable endpoint control and parameter closed-loop optimization.

Description

Technical Field

[0001] This application relates to the field of semiconductor testing technology, and in particular to a method, system, testing device, and related products for inverting semiconductor epitaxial growth parameters. Background Technology

[0002] In metal-organic chemical vapor deposition (MOCVD) epitaxial growth, in-situ optical monitoring technologies such as reflectance spectroscopy and thermal radiation spectroscopy are widely used for real-time monitoring of key process parameters such as growth rate, film thickness, composition, and temperature due to their non-contact and fast response characteristics. Their monitoring principle is typically based on the propagation and interference mechanism of light in multilayer thin film structures. An optical transmission model is used to fit the measured spectral data, thereby retrieving physical parameters related to epitaxial growth. These physical parameters are then further used for process control and consistency management.

[0003] However, traditional monitoring techniques generally overlook the absorption effect of gaseous components in the reaction chamber on the probe light. For example, when growing phosphorus- or indium-containing compounds, or when growing aluminum arsenide materials at high temperatures, the metal-organic source or its pyrolysis products have characteristic absorption lines in the near-infrared band. These gaseous absorption signals are superimposed on the reflection interference signal of the solid film, resulting in distortion of the reflection spectrum. This leads to significant model fitting errors and affects the inversion accuracy of growth parameters such as growth rate and composition. Summary of the Invention

[0004] The purpose of this application is to provide a method, system, detection device and related products for inverting semiconductor epitaxial growth parameters, so as to overcome the defects of traditional reflectance spectral monitoring technology, which leads to reflectance spectrum distortion and decreased accuracy of growth parameter inversion due to neglecting the absorption effect of gas phase components in the reaction chamber on the probe light.

[0005] In a first aspect, this application provides a method for inverting semiconductor epitaxial growth parameters, the method comprising:

[0006] Acquire reflection spectral data and real-time process parameters; wherein the reflection spectral data is generated based on the reflected light signal produced during semiconductor epitaxial growth;

[0007] Based on the real-time process parameters and the pre-stored gas molecule spectral absorption database, the absorption characteristic parameters of the gas phase components are calculated.

[0008] Calculate the gas phase transmittance of the probe optical path based on the absorption characteristic parameters.

[0009] Based on the gas phase transmittance, the reflectance spectral data is compensated to obtain the target reflectance spectrum;

[0010] Based on the target reflection spectrum, the growth parameters for semiconductor epitaxial growth are obtained by inversion.

[0011] In one embodiment, the real-time process parameters include gas flow rate, reaction chamber pressure, wafer temperature, and reaction chamber geometry parameters; the absorption characteristic parameters include effective concentration and absorption cross-section.

[0012] The calculation of absorption characteristic parameters of gas phase components based on the real-time process parameters and the pre-stored gas molecular spectral absorption database includes:

[0013] Based on the real-time process parameters, the fluid dynamics model is solved to obtain the velocity distribution of the flow field within the reaction chamber;

[0014] Based on the flow field velocity distribution, the mass transport equation is solved to obtain the concentration distribution of each gaseous component along the optical path; wherein, the mass transport equation is used to characterize the influence of convective transport and diffusion transport on the concentration distribution, as well as the influence of gas phase response on concentration generation;

[0015] The concentration distribution is integrated along the optical path to obtain the effective concentration of each gas phase component.

[0016] The spectral parameters of each gas phase component are retrieved from a pre-stored gas molecular spectral absorption database, and the absorption cross section of each gas phase component is calculated based on the spectral parameters under preset parameters, wherein the preset parameters include preset wavelength, preset pressure and preset temperature; the spectral parameters include line intensity and linewidth.

[0017] In one embodiment, the step of querying the spectral parameters of each gas phase component from a pre-stored gas molecular spectral absorption database, and calculating the absorption cross section of each gas phase component under preset parameter conditions based on the spectral parameters, includes:

[0018] For each of the gas phase components, spectral parameters corresponding to the preset wavelength range are queried from the gas molecule spectral absorption database;

[0019] The spectral parameters are corrected under the preset pressure and the preset temperature.

[0020] Based on the corrected spectral parameters, the absorption cross section of each gas phase component within the preset wavelength range is calculated using the Voigt line shape function.

[0021] In one embodiment, the absorption characteristic parameters include effective concentration and absorption cross-section;

[0022] The step of calculating the gas phase transmittance of the probe optical path based on the absorption characteristic parameters includes:

[0023] Based on the absorption cross-section and corresponding effective concentration of each gas phase component, the gas phase transmittance of the probe optical path is calculated according to the Beer-Lambert law; wherein, the expression for the gas phase transmittance is:

[0024]

[0025] in, λ represents the gas phase transmittance at wavelength λ; K represents the total number of gas phase components. This represents the absorption cross section of the k-th gas phase component at wavelength λ. denoted by , which represents the effective concentration of the k-th gaseous component along the probe optical path; L represents the effective optical path length after refractive index correction of the geometric length of the probe optical path propagating inside the reaction cavity.

[0026] In one embodiment, the step of compensating the reflectance spectral data based on the gas phase transmittance to obtain the target reflectance spectrum includes:

[0027] Based on the gas phase transmittance, the reflectance spectral data is subjected to transmission attenuation compensation to obtain the compensated target reflectance spectrum; wherein, the transmission attenuation compensation includes: normalizing the reflectance spectral data according to a preset power of the gas phase transmittance, and then superimposing an adaptive compensation term for dynamic correction.

[0028] In one embodiment, the method further includes:

[0029] An adaptive Kalman filter is used to filter the reflectance spectral data, and transmission attenuation compensation is performed on the filtered reflectance spectral data.

[0030] In one embodiment, the method further includes:

[0031] The adaptive compensation term is updated using a machine learning model trained based on historical process data, and the updated adaptive compensation term is used to dynamically correct the transmission attenuation compensation process in the next sampling period.

[0032] In one embodiment, the growth parameters include growth rate and composition;

[0033] The process of inverting the growth parameters for semiconductor epitaxial growth based on the target reflection spectrum includes:

[0034] The growth parameters to be inverted are input into a preset transmission matrix model to obtain the predicted reflection spectrum;

[0035] Based on the difference between the target reflectance spectrum and the predicted reflectance spectrum, an objective function is constructed to characterize the difference;

[0036] The objective function is iteratively solved using a nonlinear least squares optimization algorithm until a preset convergence condition is met, so that the growth parameters are updated to the optimal growth parameters.

[0037] In one embodiment, the objective function is the squared L2 of the difference between the target reflectance spectrum and the predicted reflectance spectrum;

[0038] The step of iteratively solving the objective function using a nonlinear least squares optimization algorithm until a preset convergence condition is met, and updating the growth parameters to the optimal growth parameters, includes:

[0039] Using the objective function, the target reflectance spectrum and the predicted reflectance spectrum are compared to construct a residual vector; and the sensitivity matrix of the residual vector with respect to the current growth parameters is calculated.

[0040] Based on the residual vector and the sensitivity matrix, a nonlinear least squares iterative algorithm is used to solve for the parameter increment. The growth parameters are updated based on the parameter increment until the objective function converges to a preset threshold or the number of iterations reaches a preset number. The growth parameters at this point are taken as the optimal growth parameters.

[0041] In one embodiment, the step of incrementally updating the growth parameters based on the parameters until the objective function converges to a preset threshold or the number of iterations reaches a preset number includes:

[0042] Candidate growth parameters are obtained based on the parameter increments, and candidate objective function values ​​are calculated based on the candidate growth parameters.

[0043] If the decrease in the candidate function value compared to the previous iteration is less than a preset value, the candidate growth parameter is used as the growth parameter for the next iteration, and the damping factor is reduced according to a preset step size until the objective function converges to a preset threshold or the number of iterations reaches a preset number.

[0044] Secondly, this application provides a semiconductor epitaxial growth parameter inversion system, disposed within a detection device, wherein the detection device is partially disposed within the reaction chamber of a semiconductor device; the system includes:

[0045] The acquisition module is used to acquire reflectance spectral data and real-time process parameters; wherein the reflectance spectral data is generated based on the reflected light signal generated during the semiconductor epitaxial growth process;

[0046] The calculation module is used to calculate the absorption characteristic parameters of the gas phase components based on the real-time process parameters and the pre-stored gas molecule spectral absorption database; and to calculate the gas phase transmittance of the probe optical path based on the absorption characteristic parameters.

[0047] The compensation module is used to compensate the reflectance spectral data based on the gas phase transmittance to obtain the target reflectance spectrum;

[0048] The inversion module is used to invert the growth parameters of semiconductor epitaxial growth based on the target reflection spectrum.

[0049] Thirdly, this application also provides a detection device, partially disposed within the reaction chamber of a semiconductor device, the detection device comprising:

[0050] An optical detection unit is used to output a target light signal to the wafer surface inside the reaction cavity, and to generate reflection spectrum data based on the target light signal reflected from the wafer surface.

[0051] The acquisition unit is used to acquire real-time process parameters in the reaction chamber at a preset frequency.

[0052] The semiconductor epitaxial growth parameter inversion system described in the second aspect is used to calculate the absorption characteristic parameters of the gas phase components based on the real-time process parameters and a pre-stored gas molecule spectral absorption database; calculate the gas phase transmittance of the probe optical path based on the absorption characteristic parameters; compensate the reflection spectral data based on the gas phase transmittance to obtain the target reflection spectrum; and invert the growth parameters of semiconductor epitaxial growth based on the target reflection spectrum.

[0053] In one embodiment, the optical detection unit includes:

[0054] The target light source is used to output a target light signal that covers a preset wavelength range.

[0055] An optical fiber probe is used to transmit the target optical signal to the wafer surface and to receive the target optical signal reflected from the wafer surface.

[0056] A spectrometer is used to convert the target light signal reflected from the wafer surface into reflectance spectral data;

[0057] An optical integrating sphere is used to spatially homogenize the target optical signal.

[0058] Fourthly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method steps of the first aspect.

[0059] Fifthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the method steps of the first aspect.

[0060] The aforementioned methods, systems, detection devices, and related products for inverting semiconductor epitaxial growth parameters have at least the following advantages:

[0061] This application acquires reflectance spectral data and process parameters. The reflectance spectral data is generated based on the reflected light signal produced during semiconductor epitaxial growth, and the semiconductor epitaxial growth process is executed based on real-time process parameters. Based on the real-time process parameters and a pre-stored gas molecular spectral absorption database, the absorption characteristic parameters of the gas phase components are calculated. According to the absorption characteristic parameters, the gas phase transmittance of the probe optical path is calculated. Based on the gas phase transmittance, the reflectance spectral data is compensated to obtain the target reflectance spectrum. Based on the target reflectance spectrum, the growth parameters for semiconductor epitaxial growth are inverted. Using the above scheme, this application calculates the absorption characteristic parameters of the gas phase components based on current process parameters and a pre-stored gas molecular spectral absorption database, and further obtains the gas phase transmittance of the probe optical path. This quantitatively compensates for the measured reflectance spectrum during the semiconductor epitaxial growth process, making the compensated target reflectance spectrum closer to the true optical response of the semiconductor epitaxial film. This effectively reduces spectral distortion and systematic errors caused by gas phase absorption superposition, significantly improving the accuracy and stability of the inverted growth parameters. Furthermore, the quantitative compensation values ​​can be adaptively updated according to the fluctuations in operating conditions, reducing the inversion bias introduced by operating condition drift, improving process consistency and comparability, and facilitating more reliable endpoint control and parameter closed-loop optimization. Attached Figure Description

[0062] Figure 1 This is a structural block diagram of the detection device in one embodiment;

[0063] Figure 2 This is a flowchart illustrating the method for inverting semiconductor epitaxial growth parameters in one embodiment;

[0064] Figure 3 This is a flowchart illustrating the steps for calculating the absorption characteristic parameters of gaseous components in one embodiment.

[0065] Figure 4 This is a structural block diagram of a semiconductor epitaxial growth parameter inversion system in one embodiment;

[0066] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0067] 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.

[0068] 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.

[0069] Please see Figure 1 In one exemplary embodiment, this application provides a detection device, which is partially disposed within the reaction chamber of a semiconductor device. The reaction chamber includes a base for placing a wafer, the base being made of high-purity graphite and supported by a rotating shaft driven by a motor, allowing continuous rotation during wafer processing. Under reaction conditions where a vapor precursor is introduced into the reaction chamber and suitable temperature and pressure are provided, semiconductor material is directionally deposited layer by layer on the wafer surface using epitaxial growth methods such as atomic layer epitaxy (ALE) or vapor phase epitaxy (VPE), forming an epitaxial thin film with controllable thickness.

[0070] Specifically, the detection device includes an optical detection unit, a data acquisition unit, and a semiconductor epitaxial growth parameter inversion system.

[0071] The optical detection unit is used to output the target light signal to the wafer surface in the reaction cavity and generate reflection spectrum data based on the target light signal reflected from the wafer surface.

[0072] Optionally, the optical detection unit includes a target light source, a fiber optic probe, a spectrometer, and an optical integrating sphere.

[0073] The target light source is used to output a target optical signal covering a preset wavelength range during the semiconductor epitaxial growth process. For example, in this embodiment, a broadband halogen tungsten lamp with a wavelength range of 350-2500nm is selected and equipped with constant current drive and temperature control to output a stable target optical signal.

[0074] The fiber optic probe, partially housed within the reaction chamber of the semiconductor device, is used to transmit target optical signals to the wafer surface and receive target optical signals reflected from the wafer surface. Exemplarily, the fiber optic probe employs a Y-shaped six-channel fiber bundle, made of quartz material and designed for high-temperature resistance. Further, the fiber optic probe is partially sealed within a quartz window at the top of the reaction chamber via a flange, vertically aligned with the wafer surface, with an incident angle typically less than 5°.

[0075] A spectrometer, located outside the reaction chamber, is used to convert the target light signal reflected from the wafer surface into reflectance spectral data. For example, this embodiment employs a high-resolution spectrometer from the Ocean Insight FX series, equipped with dual detectors: a CCD (200-1100nm) and an InGaAs (900-2500nm). This allows for continuous and stable coverage and stitching of a wide wavelength range of 200–2500nm within their respective high-sensitivity ranges, thereby improving the signal-to-noise ratio, dynamic range, and inversion reliability of the reflectance spectrum measurement.

[0076] An optical integrating sphere is used to spatially homogenize the target optical signal, thereby improving signal collection efficiency. For example, the optical integrating sphere can be positioned between the fiber optic probe and the quartz window.

[0077] The acquisition unit is used to acquire real-time process parameters within the reaction chamber at a preset frequency. For example, in this embodiment, the preset frequency is 1 second.

[0078] Optionally, the real-time process parameters in this embodiment include gas flow rate, reaction chamber pressure, wafer temperature, and reaction chamber geometry parameters. The gas flow rate is read by a mass flow controller with an accuracy of ±0.1 sccm. This mass flow controller can read various reaction source gases, such as TMGa, TMAl, TMin, NH3, and TBP. The reaction chamber pressure is read by a pressure sensor with a reading range of [missing information]. The accuracy is ±0.1%. The wafer temperature is read by a thermocouple placed on the wafer tray, with a reading range of 300-1200℃ and an accuracy of ±0.5℃. It should be understood that the actual process parameters in the reaction chamber will vary with operating conditions and execution errors, and therefore there will usually be a certain deviation from the process parameters set by the machine.

[0079] Furthermore, the reaction cavity geometric parameters refer to a set of parameters used to characterize the internal structural dimensions, relative positional relationships, and geometric boundaries of the optical and gas flow channels of the reaction cavity. For example, the reaction cavity geometric parameters include the overall cavity dimensions, gas flow structure geometric parameters, and optical port and optical path related geometric parameters.

[0080] The semiconductor epitaxial growth parameter inversion system is used to calculate the absorption characteristic parameters of gas phase components based on real-time process parameters and a pre-stored gas molecule spectral absorption database; calculate the gas phase transmittance of the probe optical path based on the absorption characteristic parameters; compensate the reflection spectral data based on the gas phase transmittance to obtain the target reflection spectrum; and invert the growth parameters of semiconductor epitaxial growth based on the target reflection spectrum.

[0081] Furthermore, the aforementioned detection optical path refers to the entire optical transmission path from the target light source, through the quartz window of the reaction chamber to the wafer surface, then reflected and / or scattered from the wafer surface, and finally collected and transmitted to the spectrometer; it includes the incident optical path and the recovery optical path, and the two can be coaxial or separate.

[0082] The aforementioned detection device uses an optical detection unit to generate reflectance spectral data, an acquisition unit to obtain real-time process parameters, and a semiconductor epitaxial growth parameter inversion system. Based on the real-time process parameters and a pre-stored gas molecule spectral absorption database, it calculates the absorption characteristic parameters of the gas phase components; based on the absorption characteristic parameters, it calculates the gas phase transmittance of the probe optical path; based on the gas phase transmittance, it compensates the reflectance spectral data to obtain the target reflectance spectrum; and based on the target reflectance spectrum, it inverts the growth parameters for semiconductor epitaxial growth. Using this scheme, this application calculates the absorption characteristic parameters of the gas phase components based on current process parameters and a pre-stored gas molecule spectral absorption database, and further obtains the gas phase transmittance of the probe optical path. It then quantitatively compensates for the measured reflectance spectrum during the semiconductor epitaxial growth process, making the compensated target reflectance spectrum closer to the true optical response of the semiconductor epitaxial film. This effectively reduces spectral distortion and systematic errors caused by gas phase absorption superposition, significantly improving the accuracy and stability of the inverted growth parameters. Furthermore, the quantitative compensation values ​​can be adaptively updated according to the fluctuations in operating conditions, reducing the inversion bias introduced by operating condition drift, improving process consistency and comparability, and facilitating more reliable endpoint control and parameter closed-loop optimization.

[0083] Please see Figure 2 In one exemplary embodiment, this application provides a method for inverting semiconductor epitaxial growth parameters, including:

[0084] Step 202: Obtain reflection spectrum data and real-time process parameters; wherein, the reflection spectrum data is generated based on the reflected light signal generated during the semiconductor epitaxial growth process.

[0085] Specifically, reflectance spectral data refers to the time-series reflectance spectral data composed of multiple frames of reflectance spectra collected by a spectrometer during the semiconductor epitaxial growth process. This data is obtained after the broadband target light signal output by the target light source is incident on the wafer surface in the reaction cavity, and then reflected back by the wafer surface and its epitaxial layer structure. It is used to characterize the dynamic changes of the reflectance spectrum over time during the epitaxial growth process.

[0086] Real-time process parameters refer to the parameter values ​​that are continuously collected and output by the sensors and control system of the main equipment during the semiconductor epitaxial growth process at a preset sampling period. These parameters can characterize the current operating status and reflect the actual operating status measurement value in the reaction chamber. Generally speaking, they deviate to a certain extent from the nominal values ​​set in the process recipe.

[0087] Step 204: Based on real-time process parameters and a pre-stored gas molecule spectral absorption database, calculate the absorption characteristic parameters of the gas phase components; based on the absorption characteristic parameters, calculate the gas phase transmittance of the probe optical path.

[0088] Specifically, a gas molecule spectral absorption database refers to a library of absorption lines that provides fundamental data on the spectral absorption of gaseous molecules at different wavelengths. This database is pre-stored locally or on a server in the form of data tables or set of model parameters, allowing inversion algorithms to query and calculate on demand during runtime. This gas molecule spectral absorption database typically includes the spectral line positions, intensities, broadening parameters, applicable temperature, and pressure dependencies for each target gas phase component. For example, this gas molecule spectral absorption database can be HITRAN, MODTRAN, or a dedicated absorption gas molecule spectral absorption database derived from them and constructed for common gas phase components in MOCVD.

[0089] Absorption characteristic parameters refer to a set of spectral parameters used to characterize the absorption effect of a certain gas phase component on a target optical signal at various wavelengths under given temperature, pressure, and component concentration conditions. The absorption characteristic parameters are then used to determine the variation of the absorption intensity of the gas with wavelength.

[0090] Step 206: Based on the gas phase transmittance, the reflectance spectral data is compensated to obtain the target reflectance spectrum.

[0091] Specifically, gas phase transmittance refers to the proportion of the light intensity that can pass through and reach the fiber optic probe at a certain wavelength after the target light signal passes through the gas phase medium in the reaction chamber, relative to the incident light intensity. Its value is usually between 0 and 1 and varies with wavelength. It characterizes the degree of attenuation of light by the gas phase, and the attenuation source is the absorption of gas phase molecules at that wavelength.

[0092] The target reflectance spectrum refers to the reflectance spectrum obtained after gas-phase transmittance compensation of the measured reflectance spectrum data. Its purpose is to minimize the wavelength-selective absorption attenuation effect caused by the target optical signal during its propagation in the gas phase of the reaction cavity, allowing the spectrum to primarily characterize the true optical response of the epitaxial thin film system on the wafer surface. Compared to the measured reflectance spectrum, the target reflectance spectrum more closely approximates the true reflectance response of the epitaxial film, and its spectral shape is more interpretable.

[0093] Step 208: Based on the target reflection spectrum, the growth parameters for semiconductor epitaxial growth are obtained by inversion.

[0094] Specifically, growth parameters refer to a set of key parameters used to quantitatively describe the growth state and material properties of epitaxial films on wafer surfaces. These parameters directly characterize information such as growth thickness, growth rate, material composition and optical properties, and interface quality during the epitaxial growth process. Reflectance spectroscopy is an indirect observation of the aforementioned epitaxial growth state.

[0095] The aforementioned method for inverting semiconductor epitaxial growth parameters acquires reflection spectral data and process parameters. The reflection spectral data is generated based on the reflected light signals produced during semiconductor epitaxial growth, and the semiconductor epitaxial growth process is executed based on real-time process parameters. Based on the real-time process parameters and a pre-stored gas molecular spectral absorption database, the absorption characteristic parameters of the gas phase components are calculated. Based on these absorption characteristic parameters, the gas phase transmittance of the probe optical path is calculated. Based on the gas phase transmittance, the reflection spectral data is compensated to obtain the target reflection spectrum. Based on the target reflection spectrum, the growth parameters for semiconductor epitaxial growth are inverted. Using this scheme, this application calculates the absorption characteristic parameters of the gas phase components based on current process parameters and a pre-stored gas molecular spectral absorption database, and further obtains the gas phase transmittance of the probe optical path. This quantitatively compensates for the measured reflection spectrum during the semiconductor epitaxial growth process, making the compensated target reflection spectrum closer to the true optical response of the semiconductor epitaxial film. This effectively reduces spectral distortion and systematic errors caused by gas phase absorption superposition, significantly improving the accuracy and stability of the inverted growth parameters. Furthermore, the quantitative compensation values ​​can be adaptively updated according to the fluctuations in operating conditions, reducing the inversion bias introduced by operating condition drift, improving process consistency and comparability, and facilitating more reliable endpoint control and parameter closed-loop optimization.

[0096] Please see Figure 3 Optionally, when the real-time process parameters include gas flow rate, reaction chamber pressure, wafer temperature, and reaction chamber geometry; and the absorption characteristic parameters include effective concentration and absorption cross-section, the absorption characteristic parameters of the gas phase components are calculated based on the real-time process parameters and a pre-stored gas molecular spectral absorption database, including:

[0097] Step 302: Based on real-time process parameters, solve the fluid dynamics model to obtain the velocity distribution of the flow field in the reaction chamber.

[0098] Step 304: Based on the flow field velocity distribution, solve the mass transport equation to obtain the concentration distribution of each gas phase component along the optical path; wherein, the mass transport equation is used to characterize the influence of convective transport and diffusion transport on the concentration distribution, as well as the influence of gas phase response on concentration generation.

[0099] Step 306: Integrate the concentration distribution along the optical path to obtain the effective concentration of each gas phase component.

[0100] Step 308: Query the spectral parameters of each gas phase component from the pre-stored gas molecular spectral absorption database, and calculate the absorption cross section of each gas phase component based on the spectral parameters under preset parameters. The preset parameters include preset wavelength, preset pressure and preset temperature, and the spectral parameters include line intensity and linewidth.

[0101] Specifically, the fluid dynamics model in this embodiment adopts the Navier-Stokes equations, which are a set of partial differential equations used to describe the fluid motion of, for example, gases or liquids.

[0102] Optionally, based on real-time process parameters, a fluid dynamics model is solved to obtain the velocity distribution of the flow field within the reaction chamber, including:

[0103] The computational domain and mesh of the reaction chamber are established based on the geometric parameters of the reaction chamber; the fluid properties of the mixed gas, such as density and viscosity, are determined based on the reaction chamber pressure and wafer temperature; the flow rates of each gas are obtained and converted into inlet velocity boundary conditions at the inlet end; the reaction chamber pressure is used as the pressure boundary condition at the exhaust end, and the outlet boundary is set in combination with the exhaust port geometric parameters; with the base rotating continuously, the base rotation speed is converted into rotating wall boundary conditions; under the constraints of the above boundary conditions and physical property parameters, the simplified Navier-Stokes equations are solved to obtain the velocity vector distribution at each position in the reaction chamber, which is then output as the flow field velocity distribution.

[0104] The mass transport equation is a conservation equation used to describe how a certain gaseous component changes its distribution over time and space in a flowing gas.

[0105] Optionally, based on the flow field velocity distribution, the mass transport equation is solved to obtain the concentration distribution of each gas phase component along the optical path, including:

[0106] The flow velocity distribution is used as the input for the convection term, and the molecular diffusion coefficients of each gaseous component under the current operating conditions are determined based on the reaction chamber pressure and wafer temperature. The inlet concentration of each gaseous component is calculated using the real-time flow rate and proportion at the inlet as the inlet boundary condition, and the convective outflow at the exhaust port as the outlet boundary condition. Deposition consumption boundaries are set on the cavity wall and the base surface. Simultaneously, a reaction source term is constructed based on a preset gaseous reaction mechanism to characterize the generation and consumption of the gaseous component in the gaseous reaction. Finally, the mass transport equation is solved under the above boundary conditions to obtain the concentration field of the gaseous component in the reaction chamber space. The concentration field is further sampled along the probe optical path to obtain the concentration distribution of the gaseous component along the optical path direction.

[0107] Optionally, the concentration distribution is integrated along the optical path to obtain the effective concentration of each gas phase component, including:

[0108] The path segment of the optical path in the reaction cavity is determined based on the geometric parameters of the detection optical path. This path is then discretized into multiple sampling points arranged sequentially along the optical path. For each sampling point, the concentration value corresponding to that sampling point is obtained by interpolation from the concentration distribution, forming a discrete concentration sequence of the gaseous component along the optical path direction. The discrete concentration sequence is then weighted and summed according to the optical path segment length between each adjacent sampling point to obtain the path integral value of the gaseous component on the detection optical path. The path integral value is then divided by the total path length of the optical path in the reaction cavity to obtain the effective concentration of the gaseous component.

[0109] Optionally, the spectral parameters of each gas phase component are queried from a pre-stored gas molecular spectral absorption database, and the absorption cross section of each gas phase component is calculated under preset parameter conditions based on the spectral parameters, including:

[0110] For each gas phase component, spectral parameters corresponding to a preset wavelength range are queried from a gas molecular spectral absorption database. These spectral parameters include line intensity and linewidth. The spectral parameters are then corrected at a preset pressure and temperature. Based on the corrected spectral parameters, the absorption cross section of each gas phase component within the preset wavelength range is calculated using the Voigt line shape function.

[0111] Specifically, the purpose of the correction process is to convert the dashed line intensity and linewidth parameters given in the gas molecule spectral absorption database under the reference operating conditions to the target operating conditions, thereby obtaining the line intensity and linewidth under the target operating conditions.

[0112] The preset pressure is set to the nominal pressure of the reaction chamber, and the preset temperature is set to the nominal temperature of the reaction chamber. Based on the preset pressure and preset temperature, the spectral parameters are corrected. For example, the line intensity is corrected according to the preset temperature, and the collision broadening linewidth is corrected according to the preset pressure and preset temperature, so as to obtain the corrected spectral line intensity and corrected linewidth parameters.

[0113] The Voigt line shape function is a function used to describe the shape of a single spectral line. It is the convolution of the Gaussian line shape and the Lorentz line shape. The Gaussian part mainly corresponds to the Doppler broadening caused by molecular thermal motion, while the Lorentz part mainly corresponds to the collision broadening caused by molecular collisions. Therefore, the Voigt line shape can simultaneously characterize the actual spectral line shape under the combined effects of temperature and pressure.

[0114] Furthermore, using the modified spectral parameters as input, the Voith line function is calculated point by point within a preset wavelength range at preset wavelength steps, and the Voith absorption contribution of each spectral line at each wavelength point is superimposed to obtain the absorption cross-section curve of the gas phase component within the preset wavelength range; the above process is performed on all gas phase components to obtain the absorption cross-section corresponding to each gas phase component.

[0115] Using the above scheme, real-time process parameters can explicitly map operating condition information such as gas flow rate, reaction chamber pressure, wafer temperature, and reaction chamber geometric parameters into the flow field and mass transport solution. This yields the true concentration distribution along the probe optical path, which is then integrated through the optical path to form the effective concentration. Simultaneously, by utilizing the spectral line parameters of the gas molecule absorption database, combined with pressure and temperature corrections and Voigt line shape calculations, an absorption cross section consistent with the current operating conditions is obtained. This avoids the systematic errors introduced by simply treating gas phase concentration and absorption capacity as constants, thereby improving the accuracy and stability of subsequent gas phase transmittance calculations and reflectance spectrum compensation. It also reduces inversion drift caused by operating condition fluctuations, improves the consistency and repeatability of growth parameter inversions such as epitaxial growth thickness, growth rate, and composition, and enhances online deployability and closed-loop control effects under mass production conditions.

[0116] Optionally, the gas phase transmittance of the probe optical path is calculated based on absorption characteristic parameters, including:

[0117] Based on the absorption cross-section and corresponding effective concentration of each gas phase component, the gas phase transmittance of the probe optical path is calculated according to the Beer-Lambert law; where the expression for the gas phase transmittance is:

[0118]

[0119] in, λ represents the gas phase transmittance at wavelength λ; K represents the total number of gas phase components. This represents the absorption cross section of the k-th gas phase component at wavelength λ. denoted by , which represents the effective concentration of the k-th gaseous component along the probe optical path; L represents the effective optical path length after refractive index correction of the geometric length of the probe optical path propagating inside the reaction cavity.

[0120] Specifically, the basic form of Beer-Lambert's law is T = exp(-σ×C×L), which describes the relationship between transmittance T and absorption cross-section σ, concentration C, and optical path L for a single component and a single wavelength. In this specific embodiment, it is necessary to calculate the gas phase transmittance of the entire detection optical path, and this gas phase transmittance is a transmittance function that varies with wavelength. This involves the comprehensive absorption effect of multiple gas phase components over a continuous spectral range. In the above expression, for each gas phase component k, at a wavelength... The contribution of this component to light attenuation is expressed as... The total optical thickness is obtained by summing the results from the summation of the multiple components. Further, the total optical thickness is converted into a multiplicative superposition of transmittance, that is, the transmittance caused by each gas phase component is multiplied to obtain the total transmittance: .

[0121] Using the above scheme, the expression for gas phase transmittance can comprehensively characterize the absorption contributions of multiple gas phase components over a continuous spectral range within a unified physical framework; wherein, the absorption cross-section of each component... This demonstrates the selective absorption characteristics of spectral lines as a function of wavelength. The summation of multiple components reflects the additivity of the optical thickness of the multi-component absorption, and the exponential form further ensures that the transmittance calculation satisfies the light intensity attenuation law, thus yielding a transmittance function that varies with wavelength. Since the above formula can directly superimpose the contributions of each component at each wavelength point, even if the absorption spectra of different components overlap, they can be naturally included in a unified calculation, avoiding the accumulation of errors and instability in fitting caused by treating the absorption effect separately.

[0122] Optionally, the reflectance spectral data is compensated based on the gas phase transmittance to obtain the target reflectance spectrum, including:

[0123] Based on gas phase transmittance, transmission attenuation compensation is performed on the reflectance spectral data to obtain the compensated target reflectance spectrum. The transmission attenuation compensation includes: normalizing the reflectance spectral data to a preset power of the gas phase transmittance, and then dynamically correcting it by superimposing an adaptive compensation term.

[0124] Specifically, the expression for transmission attenuation compensation is:

[0125]

[0126] in, The target's reflection spectrum; These are measured reflectance spectrum data; To detect the gas phase transmittance of the optical path; This is an adaptive compensation term.

[0127] In this embodiment, the preset power is set to 2, which means that the attenuation of the gas phase component on the incident target light signal twice is quantitatively separated from the measured reflection spectrum data, thereby more closely approximating the true reflection response of the epitaxial film.

[0128] The adaptive compensation term characterizes unmodeled systematic errors, including slowly varying terms introduced by instrument response drift, scattering, and window contamination. Compensating reflectance spectral data based on the adaptive compensation term improves the robustness of the compensation and avoids misattributing these errors to epitaxial growth parameters.

[0129] Optionally, the above-mentioned method for inverting semiconductor epitaxial growth parameters further includes:

[0130] An adaptive Kalman filter is used to filter the reflectance spectral data, and transmission attenuation compensation is applied to the filtered reflectance spectral data. Using this scheme, the filtering process can suppress transient disturbances and random noise while preserving effective spectral characteristics. Furthermore, the adaptive Kalman filter can dynamically adjust the filter gain based on the real-time estimated noise level, avoiding the problems of over-smoothing or under-filtering that occur with fixed-parameter filtering under different operating conditions.

[0131] Optionally, the above-mentioned method for inverting semiconductor epitaxial growth parameters further includes:

[0132] The adaptive compensation term is updated using a machine learning model trained on historical process data, and the updated adaptive compensation term is used to dynamically correct the transmission attenuation compensation process in the next sampling period.

[0133] Specifically, historical process data includes raw reflectance spectral data collected from past processes, process parameter sets, and historical growth parameters obtained and verified through inversion using the method of this invention.

[0134] A machine learning model is employed, using raw reflectance spectrum data and partial features of the process parameter set as input, with the goal of training the model to make the compensation terms output by the model more accurately help to deduce the growth parameters.

[0135] Optionally, when the growth parameters include growth rate and composition, the growth parameters for semiconductor epitaxial growth are obtained by inversion based on the target reflectance spectrum, including:

[0136] The growth parameters to be inverted are input into a preset transfer matrix model to obtain the predicted reflection spectrum. Based on the difference between the target reflection spectrum and the predicted reflection spectrum, an objective function is constructed to characterize the difference. The objective function is iteratively solved using a nonlinear least squares optimization algorithm until the preset convergence condition is met, so that the growth parameters are updated to the optimal growth parameters.

[0137] Optionally, the objective function is the squared L2 norm of the difference between the target reflection spectrum and the predicted reflection spectrum; a nonlinear least squares optimization algorithm is used to iteratively solve the objective function until a preset convergence condition is met, so as to update the growth parameters to the optimal growth parameters, including:

[0138] Using an objective function, the residual vector is obtained by comparing the target reflection spectrum and the predicted reflection spectrum; and the sensitivity matrix of the residual vector with respect to the current growth parameters is calculated.

[0139] Based on the residual vector and sensitivity matrix, a nonlinear least squares iterative algorithm is used to solve for the parameter increment, and the growth parameters are updated synchronously based on the parameter increment until the objective function converges to a preset threshold or the number of iterations reaches a preset number. The growth parameters at this point are taken as the optimal growth parameters.

[0140] Optionally, the growth parameters are updated incrementally until the objective function converges to a preset threshold or the number of iterations reaches a preset number, including:

[0141] Candidate growth parameters are obtained based on parameter increments, and candidate objective function values ​​are calculated based on the candidate growth parameters. If the decrease in the candidate function value compared to the previous iteration is less than a preset value, the candidate growth parameters are used as growth parameters for the next iteration, and the damping factor is reduced according to a preset step size until the objective function converges to a preset threshold or the number of iterations reaches a preset number.

[0142] Specifically, the processing logic of the transmission matrix model lies in treating the multilayer epitaxial thin film as multiple sequentially stacked optical layers. Each layer's thickness and optical constants describe the propagation of light within that layer, and the reflection relationship between each layer interface describes the distribution of light at that interface. The propagation effects of each layer and the reflection and transmission effects of each interface are represented by matrices and multiplied sequentially to obtain the overall response of the entire film system to the incident target light signal. From this, the reflection coefficients at different wavelengths can be calculated, and their intensities can be used to obtain the reflectivity. Thus, the theoretical reflection spectrum can be calculated point-by-point within a preset wavelength range. This embodiment improves the transmission matrix model by further incorporating the optical responses of anisotropic materials in different directions and the correction for reflection by rough interfaces, making the predicted reflection spectrum closer to actual measurements.

[0143] Specifically, the purpose of iteration is to find a set of growth parameters that cause the objective function to converge to a condition that satisfies a preset condition. In this embodiment, the preset condition is convergence to a value less than a preset threshold. For example, the iterative optimization algorithm in this embodiment can be a nonlinear least squares optimization algorithm (Levenberg–Marquardt, LM).

[0144] Before iteration, the iteration parameters are first initialized. For example, the initial damping factor λ is set to 0.01; the maximum number of iterations K is set to 50; and the preset convergence threshold τ is set to... .

[0145] In each iteration, the current growth parameters are input into the transfer matrix model to obtain the predicted reflection spectrum. The difference between the predicted reflection spectrum and the target reflection spectrum is calculated to obtain the residual vector r. The sensitivity matrix is ​​calculated based on the residual vector, whereby the sensitivity matrix is ​​used to characterize the partial derivative relationship of the residual with respect to each growth parameter.

[0146] Based on residual vector With sensitivity matrix The parameter increment is solved using a nonlinear least squares iterative rule, and the growth parameter is updated based on the parameter increment; for example, the expression for the parameter increment δ is:

[0147]

[0148] in, is the damping factor; T is the transpose.

[0149] During the iteration process, candidate growth parameters are calculated based on the current growth parameters and parameter increments. Using the same steps as above, the candidate growth parameters are used as input to the transfer matrix model to finally calculate the candidate objective function value. If the value of the subsequent objective function decreases to a preset value compared to the previous iteration, the candidate growth parameter is used as the growth parameter for the next iteration. At the same time, the damping factor is adaptively adjusted to control the iteration step size and improve the solution stability.

[0150] Repeat the above steps until the objective function converges to less than the preset threshold τ, or the number of iterations reaches the preset number K.

[0151] It should be noted that the film thickness is the product of the growth rate and time. Therefore, the film thickness can also be calculated based on the growth rate obtained from the inversion. In this case, the growth parameters also include the film thickness.

[0152] The above scheme, which constructs an objective function based on the difference between the target and predicted reflectance spectra and employs an inversion step that uses nonlinear least squares iteration, globally constrains reflectance spectral information across multiple wavelength ranges using a unified L2-squared form. The residual vector and sensitivity matrix quantify the influence of each growth parameter on the spectrum, enabling synergistic optimization of multiple growth parameters during iteration. Simultaneously, an adaptive adjustment mechanism for the damping factor is introduced, effectively suppressing divergence or oscillations caused by excessively large step sizes even under conditions of significant disturbance, thus improving the numerical stability and convergence reliability of the solution. This allows for obtaining optimal growth parameters that meet convergence conditions within fewer iterations, making the inversion results less sensitive to measurement noise and operating condition drift. This improves the accuracy, consistency, and repeatability of online inversion, supporting real-time monitoring, endpoint determination, and closed-loop process parameter tuning of the epitaxial growth process, thereby improving epitaxial quality and mass production yield.

[0153] Optionally, the above-mentioned method for inverting semiconductor epitaxial growth parameters further includes:

[0154] Based on the growth parameters obtained from the inversion, the process parameters are adjusted in real time so that the main control system executes the epitaxial growth process according to the adjusted process parameters in the next process cycle, thereby realizing closed-loop control of the epitaxial growth process.

[0155] 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.

[0156] Based on the same inventive concept, this application also provides a semiconductor epitaxial growth parameter inversion system. This system is applicable to the above-mentioned semiconductor epitaxial growth parameter inversion method. The solution provided by this system is similar to the solution described in the above-mentioned method. Therefore, the specific limitations of one or more system embodiments provided below can be found in the limitations of the method above, and will not be repeated here.

[0157] Please see Figure 4 In one embodiment, the semiconductor epitaxial growth parameter inversion system is located within the detection device, which is partially located within the reaction chamber of the semiconductor device.

[0158] The semiconductor epitaxial growth parameter inversion system includes: an acquisition module, a calculation module, a compensation module, and an inversion module.

[0159] The acquisition module is used to acquire reflectance spectral data and real-time process parameters; wherein, the reflectance spectral data is generated based on the reflected light signals generated during the semiconductor epitaxial growth process;

[0160] The calculation module is used to calculate the absorption characteristic parameters of gas phase components based on real-time process parameters and a pre-stored gas molecule spectral absorption database; and to calculate the gas phase transmittance of the probe optical path based on the absorption characteristic parameters.

[0161] The compensation module is used to compensate the reflectance spectral data based on the gas phase transmittance to obtain the target reflectance spectrum;

[0162] The inversion module is used to invert the growth parameters of semiconductor epitaxial growth based on the target reflection spectrum.

[0163] Optionally, when the real-time process parameters include gas flow rate, reaction chamber pressure, wafer temperature, and reaction chamber geometry parameters; and the absorption characteristic parameters include effective concentration and absorption cross-section, the calculation module calculates the absorption characteristic parameters of the gas phase components based on the real-time process parameters and a pre-stored gas molecular spectral absorption database. This includes: solving a fluid dynamics model based on the real-time process parameters to obtain the flow field velocity distribution within the reaction chamber; solving the mass transport equation based on the flow field velocity distribution to obtain the concentration distribution of each gas phase component along the optical path; wherein the mass transport equation is used to characterize the influence of convective and diffusion transport on the concentration distribution, as well as the influence of gas phase response on concentration generation; integrating the concentration distribution along the optical path to obtain the effective concentration of each gas phase component; querying the spectral parameters of each gas phase component from the pre-stored gas molecular spectral absorption database, and calculating the absorption cross-section of each gas phase component under preset parameters based on the spectral parameters, wherein the preset parameters include preset wavelength, preset pressure, and preset temperature, and the spectral parameters include line intensity and linewidth.

[0164] Optionally, the calculation module calculates the gas phase transmittance of the probe optical path based on the absorption characteristic parameters, including: calculating the gas phase transmittance of the probe optical path according to Beer-Lambert's law based on the absorption cross-section and corresponding effective concentration of each gas phase component; wherein, the expression for the gas phase transmittance is:

[0165]

[0166] in, λ represents the gas phase transmittance at wavelength λ; K represents the total number of gas phase components. This represents the absorption cross section of the k-th gas phase component at wavelength λ. denoted by , which represents the effective concentration of the k-th gaseous component along the probe optical path; L represents the effective optical path length after refractive index correction of the geometric length of the probe optical path propagating inside the reaction cavity.

[0167] Optionally, the compensation module compensates the reflectance spectral data based on the gas phase transmittance to obtain the target reflectance spectrum, including: performing transmission attenuation compensation on the reflectance spectral data based on the gas phase transmittance to obtain the compensated target reflectance spectrum; wherein, the transmission attenuation compensation includes: normalizing the reflectance spectral data according to a preset power of the gas phase transmittance, and then superimposing an adaptive compensation term for dynamic correction.

[0168] Optionally, the compensation module is also used to filter the reflectance spectral data using an adaptive Kalman filter, and to compensate for transmission attenuation in the filtered reflectance spectral data.

[0169] Optionally, the compensation module is also used to update the adaptive compensation term using a machine learning model trained based on historical process data, and to dynamically correct the transmission attenuation compensation process in the next sampling period using the updated adaptive compensation term.

[0170] Optionally, when the growth parameters include growth rate and composition, the inversion module inverts the growth parameters of semiconductor epitaxial growth based on the target reflection spectrum, including: inputting the growth parameters to be inverted into a preset transfer matrix model to obtain a predicted reflection spectrum; constructing an objective function to characterize the difference between the target reflection spectrum and the predicted reflection spectrum; and iteratively solving the objective function using a nonlinear least squares optimization algorithm until a preset convergence condition is met, so as to update the growth parameters to the optimal growth parameters.

[0171] Optionally, the objective function is the squared L2 norm of the difference between the target reflection spectrum and the predicted reflection spectrum. The inversion module uses a nonlinear least squares optimization algorithm to iteratively solve the objective function until a preset convergence condition is met, so as to update the growth parameters to the optimal growth parameters. This includes: using the objective function, comparing the target reflection spectrum and the predicted reflection spectrum to obtain the residual vector; and calculating the sensitivity matrix of the residual vector relative to the current growth parameters; based on the residual vector and the sensitivity matrix, using a nonlinear least squares iterative algorithm to solve the parameter increment, and synchronously updating the growth parameters based on the parameter increment, until the objective function converges to a preset threshold or the number of iterations reaches a preset number, and the growth parameters at this time are taken as the optimal growth parameters.

[0172] Optionally, the inversion module updates the growth parameters based on parameter increments until the objective function converges to a preset threshold or the number of iterations reaches a preset number. This includes: obtaining candidate growth parameters based on parameter increments; calculating candidate objective function values ​​based on candidate growth parameters; and if the decrease in the candidate function value compared to the previous iteration is less than a preset value, using the candidate growth parameters as growth parameters for the next iteration and decreasing the damping factor according to a preset step size until the objective function converges to a preset threshold or the number of iterations reaches a preset number.

[0173] Optionally, the above-mentioned semiconductor epitaxial growth parameter inversion system further includes a process parameter adjustment module.

[0174] The process parameter adjustment module is used to adjust the process parameters in real time based on the growth parameters obtained by inversion, so that the main control system can execute the epitaxial growth process according to the adjusted process parameters in the next process cycle.

[0175] The aforementioned semiconductor epitaxial growth parameter inversion system acquires reflection spectral data and process parameters. The reflection spectral data is generated based on the reflected light signals produced during semiconductor epitaxial growth, and the semiconductor epitaxial growth process is executed based on real-time process parameters. Based on the real-time process parameters and a pre-stored gas molecule spectral absorption database, the absorption characteristic parameters of the gas phase components are calculated. Based on the absorption characteristic parameters, the gas phase transmittance of the probe optical path is calculated. Based on the gas phase transmittance, the reflection spectral data is compensated to obtain the target reflection spectrum. Based on the target reflection spectrum, the growth parameters for semiconductor epitaxial growth are inverted. Using the above scheme, this application calculates the absorption characteristic parameters of the gas phase components based on current process parameters and a pre-stored gas molecule spectral absorption database, and further obtains the gas phase transmittance of the probe optical path. This quantitatively compensates for the measured reflection spectrum during the semiconductor epitaxial growth process, making the compensated target reflection spectrum closer to the true optical response of the semiconductor epitaxial film. This effectively reduces spectral distortion and systematic errors caused by gas phase absorption superposition, significantly improving the accuracy and stability of the inverted growth parameters. Furthermore, the quantitative compensation values ​​can be adaptively updated according to the fluctuations in operating conditions, reducing the inversion bias introduced by operating condition drift, improving process consistency and comparability, and facilitating more reliable endpoint control and parameter closed-loop optimization.

[0176] Each module in the aforementioned semiconductor epitaxial growth parameter inversion system 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 the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0177] In one feasible embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 5As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements the aforementioned method for inverting semiconductor epitaxial growth parameters. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0178] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0179] In one feasible embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method steps in the above-described method for inverting semiconductor epitaxial growth parameters.

[0180] In one feasible embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method steps in the above-described method for inverting semiconductor epitaxial growth parameters.

[0181] In one feasible embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the method steps in the above-described method for inverting semiconductor epitaxial growth parameters.

[0182] 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, and when executed, it can include the processes of the embodiments described above. Any references to memory, gas molecule spectral absorption 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 gas molecule spectral absorption databases involved in the embodiments provided in this application may include at least one of relational and non-relational gas molecule spectral absorption databases. Non-relational gas molecule spectral absorption databases may include, but are not limited to, blockchain-based distributed gas molecule spectral absorption 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.

[0183] 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.

[0184] 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. A method of inversion of semiconductor epitaxial growth parameters, characterized in that, The method comprises: acquiring reflectance spectrum data and real-time process parameters; wherein the reflectance spectrum data is generated based on a reflected light signal generated during a semiconductor epitaxial growth process; based on the real-time process parameters and a pre-stored gas molecule spectral absorption database, calculating absorption characteristic parameters of gas phase components; the absorption characteristic parameters include effective concentration and absorption cross section; based on the absorption characteristic parameters, calculating the gas phase transmittance of a detection light path; based on the gas phase transmittance, compensating the reflectance spectrum data to obtain a target reflectance spectrum; based on the target reflectance spectrum, inversely calculating growth parameters of the semiconductor epitaxial growth.

2. The method of claim 1, wherein, The real-time process parameters include gas flow, reaction cavity pressure, wafer temperature and reaction cavity geometric parameters; The calculation of the absorption characteristic parameters of the gas phase components based on the real-time process parameters and the pre-stored gas molecule spectral absorption database comprises: based on the real-time process parameters, solving a fluid dynamics model to obtain a flow field velocity distribution in the reaction cavity; based on the flow field velocity distribution, solving a substance transport equation to obtain a concentration distribution of each gas phase component along the light path direction; wherein the substance transport equation is used to represent the influence of convective transport and diffusive transport on the concentration distribution, and the influence of gas phase reaction on the concentration generation; integrating the concentration distribution along the light path to obtain the effective concentration of each gas phase component; querying spectral line parameters of each gas phase component from the pre-stored gas molecule spectral absorption database, and calculating the absorption cross section of each gas phase component under pre-set parameters according to each spectral line parameter, wherein the pre-set parameters include a pre-set wavelength, a pre-set pressure and a pre-set temperature; the spectral line parameters include line strength and line width.

3. The method of claim 2, wherein: the querying of the spectral line parameters of each gas phase component from the pre-stored gas molecule spectral absorption database, and the calculation of the absorption cross section of each gas phase component under the pre-set parameters according to each spectral line parameter, comprises: for each gas phase component, querying spectral line parameters corresponding to a pre-set wavelength range from the gas molecule spectral absorption database; correcting the spectral line parameters under the pre-set pressure and the pre-set temperature; based on the corrected spectral line parameters, calculating the absorption cross section of each gas phase component in the pre-set wavelength range by using Voigt line shape function.

4. The method of claim 1, wherein, The absorption characteristic parameters include effective concentration and absorption cross section; The calculation of the gas phase transmittance of the detection light path based on the absorption characteristic parameters comprises: calculating the gas phase transmittance of the detection light path according to Beer-Lambert law based on the absorption cross section and the corresponding effective concentration of each gas phase component; wherein the expression of the gas phase transmittance is: wherein, T(λ) represents the gas phase transmittance at wavelength λ; K represents the total number of gas phase components; σk(λ) represents the absorption cross section of the kth gas phase component at wavelength λ; ck represents the effective concentration of the kth gas phase component along the probe light path; L represents the effective optical path length of the geometric length of the propagation path of the probe light path inside the reaction chamber after the refractive index correction.

5. The method of claim 1, wherein, The compensation of the reflectance spectrum data based on the gas phase transmittance to obtain the target reflectance spectrum comprises: performing transmission attenuation compensation on the reflectance spectrum data based on the gas phase transmittance to obtain the compensated target reflectance spectrum; wherein the transmission attenuation compensation comprises: after normalizing the reflectance spectrum data by a pre-set power of the gas phase transmittance, superimposing an adaptive compensation term for dynamic correction.

6. The method of claim 5, wherein, The method further comprises: The adaptive Kalman filter is used to filter the reflection spectrum data, and the reflection spectrum data after the filtering is compensated for transmission attenuation.

7. The method of claim 5, wherein, The method further comprises: The adaptive compensation term is updated by using a machine learning model trained based on historical process data, and the transmission attenuation compensation process of the next sampling period is dynamically corrected by using the updated adaptive compensation term.

8. The method of claim 1, wherein, The growth parameters include growth rate and composition; The growth parameters of the semiconductor epitaxial growth are inversely derived based on the target reflection spectrum, including: The growth parameters to be inversely derived are input into a preset transfer matrix model to obtain a predicted reflection spectrum; A target function for characterizing the difference is constructed based on the difference between the target reflection spectrum and the predicted reflection spectrum; The growth parameters are updated to optimal growth parameters by iteratively solving the target function by using a nonlinear least squares optimization algorithm until a preset convergence condition is met.

9. The method of claim 8, wherein, The target function is the square of the two-norm of the difference between the target reflection spectrum and the predicted reflection spectrum. The growth parameters are updated to optimal growth parameters by iteratively solving the target function by using a nonlinear least squares optimization algorithm until a preset convergence condition is met. The target function is used to compare the target reflection spectrum and the predicted reflection spectrum to construct a residual error vector, and a sensitivity matrix of the residual error vector with respect to the current growth parameters is calculated. Based on the residual error vector and the sensitivity matrix, a parameter increment is solved by using a nonlinear least squares iterative algorithm, and the growth parameters are updated based on the parameter increment until the target function converges to a preset threshold or the number of iterations reaches a preset number, and the growth parameters at this time are taken as the optimal growth parameters.

10. The method of claim 9, wherein, The growth parameters are updated based on the parameter increment until the target function converges to a preset threshold or the number of iterations reaches a preset number, including: A candidate growth parameter is obtained based on the parameter increment, and a candidate target function value is calculated based on the candidate growth parameter. In the case that the reduction of the candidate function value compared with the last iteration is less than a preset value, the candidate growth parameter is taken as the growth parameter of the next iteration, and the damping factor is reduced by a preset step size until the target function converges to a preset threshold or the number of iterations reaches a preset number.

11. A system for inversion of semiconductor epitaxial growth parameters, characterized in that The detection device is arranged in the semiconductor equipment reaction chamber, and the system comprises: An acquisition module is configured to acquire reflection spectrum data and real-time process parameters, wherein the reflection spectrum data is generated based on reflection light signals generated during a semiconductor epitaxial growth process; A calculation module is configured to calculate absorption characteristic parameters of a gas phase component based on the real-time process parameters and a pre-stored gas molecule spectrum absorption database, and calculate a gas phase transmittance of a detection light path according to the absorption characteristic parameters, wherein the absorption characteristic parameters include effective concentration and absorption cross section; A compensation module is configured to compensate the reflection spectrum data based on the gas phase transmittance to obtain a target reflection spectrum; An inversion module is configured to inversely derive growth parameters of the semiconductor epitaxial growth based on the target reflection spectrum.

12. A detection device, characterized in that Partially disposed in a semiconductor device reaction cavity, the detection device comprises: An optical detection unit for outputting a target light signal to a wafer surface in the reaction cavity and generating reflection spectrum data according to the target light signal reflected by the wafer surface; A collection unit for collecting real-time process parameters in the reaction cavity according to a preset frequency; The inversion system of the semiconductor epitaxial growth parameter as claimed in claim 11 is used for calculating the absorption characteristic parameters of the gas phase component based on the real-time process parameters and the pre-stored gas molecule spectrum absorption database, calculating the gas phase transmittance of the probe light path according to the absorption characteristic parameters, compensating the reflection spectrum data based on the gas phase transmittance to obtain a target reflection spectrum, and inversely obtaining the growth parameters of the semiconductor epitaxial growth based on the target reflection spectrum.

13. The detection device of claim 12, wherein, The optical detection unit comprises: A target light source for outputting a target light signal covering a preset wavelength range; An optical fiber probe for conducting the target light signal to the wafer surface and receiving the target light signal reflected by the wafer surface; A spectrometer for converting the target light signal reflected by the wafer surface into reflection spectrum data; An optical integrating sphere for spatially homogenizing the target light signal.

14. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to realize the steps of the method of any one of claims 1-10.

15. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the method of any one of claims 1-10.

Images