Method for optimizing process environment in epitaxial growth reaction chamber and system thereof

Abstract

The application provides an optimization method and system for a process environment in an epitaxial growth reaction cavity, relates to the technical field of semiconductor detection, and acquires a side scattering light signal of particulate matter in the reaction cavity and rotation phase information of a planetary disc. An optimal estimation algorithm is compared with a forward model database constructed based on a process scenario, the concentration information of the particulate matter and the source of the particulate matter can be inversely calculated in real time during epitaxial growth, and the process parameters related to the generation of the particulate matter are quantified, so that the dynamic optimization of the process environment of epitaxial growth is realized. By using the scheme, the particulate matter risks caused by different mechanisms can be identified in time, the process parameters are adjusted accordingly, the accumulation and pollution of the particulate matter are effectively inhibited, the stability and yield of epitaxial growth are improved, and the film quality and process consistency are ensured.

Description

Technical Field

[0001] This application relates to the field of semiconductor detection technology, and in particular to a method and system for optimizing the process environment within an epitaxial growth reaction chamber. Background Technology

[0002] Epitaxial growth is a process of growing a new single-crystal thin film along the original crystal orientation on a single-crystal substrate (e.g., silicon, sapphire, silicon carbide, gallium arsenide). This epitaxial layer determines the core electrical and optical characteristics of semiconductor devices and is a crucial step in manufacturing core chips such as high-performance transistors, light-emitting diodes (LEDs), lasers (LDs), and power devices. Mainstream epitaxial technologies include metal-organic chemical vapor deposition (MOCVD), hydride vapor phase epitaxy (HYPE), molecular beam epitaxy (MBE), and various chemical vapor deposition (CVD) techniques.

[0003] The quality of epitaxial films, especially their crystal integrity, surface morphology, thickness uniformity, and doping precision, directly determines the performance, yield, and reliability of the final device. The quality of epitaxial films is highly dependent on the cleanliness and environmental stability within the reaction chamber during growth. The reaction chamber is an extremely complex environment combining high temperature, vacuum or a specific atmosphere, and complex flow fields. Under these conditions, any micron- or submicron-sized particulate contamination—its generation, suspension, transport, and eventual settling on the wafer growth surface—will cause fatal damage to the wafer. These particles, acting as unplanned nucleation centers or physical masks, severely disrupt the continuity of the epitaxial single-crystal structure, leading to a series of defects in the epitaxial film.

[0004] In traditional technologies, particulate matter monitoring mainly relies on external optical inspection, offline surface defect detection, or periodic sample observation. This cannot achieve real-time, in-situ monitoring of particulate matter inside the cavity, nor can it identify the source and triggering mechanism of particulate matter generation in a timely manner. This can cause disturbances to the epitaxial growth process environment in subsequent processes, amplify process fluctuations, and reduce yield. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for optimizing the process environment inside the epitaxial growth reaction chamber, so as to overcome the shortcomings of traditional technology, which cannot achieve real-time, in-situ monitoring of particles inside the chamber, and it is difficult to identify the source and triggering mechanism of particle generation in a timely manner, resulting in disturbance of the epitaxial growth process environment in subsequent processes, amplifying process fluctuations and reducing yield.

[0006] Firstly, this application proposes a method for optimizing the process environment within an epitaxial growth reaction chamber, including:

[0007] Acquire the scattered light signal and the rotation phase signal of the planetary disk in the reaction cavity; wherein, the scattered light signal is generated by the laser signal irradiating the particulate matter in the reaction cavity through the lateral detection channel of the in-situ optical measurement device;

[0008] Based on the optimal estimation algorithm, the scattered light signal is compared with the preset forward model database to obtain the concentration information of particulate matter. The forward model database is used to characterize the mapping relationship between particulate matter properties and the scattering response of the detection point under given wavelength and scattering angle conditions. It is based on the process parameters of the reaction chamber and adopts the Monte Carlo method combined with the computational fluid dynamics model to simulate the optical acquisition path formation of photons under dynamic rotating process parameters.

[0009] The source of particulate matter is determined based on the concentration information and the rotational phase signal;

[0010] Based on the source of the particulate matter, process parameters are quantified to optimize the process environment for epitaxial growth in the reaction chamber.

[0011] In one embodiment, the laser signal is emitted by a pulse-modulated laser source;

[0012] The methods for acquiring the scattered light signal include:

[0013] Based on time-gating technology, the scattered light signal is acquired within a specified time window after the laser signal is emitted; wherein, the specified time window refers to a sampling time interval set based on the time difference between the laser signal emission time and the expected arrival time of the particulate scattering signal in the target monitoring area, which is used to receive the particulate scattering signal and suppress background reflection from non-target areas; the non-target areas include the reaction cavity wall and the wafer interface.

[0014] In one embodiment, determining the particulate matter source based on the concentration information and the rotational phase signal includes:

[0015] By performing correlation analysis between the concentration information and the rotation phase signal, the variation law of particulate matter concentration with phase can be obtained;

[0016] If the variation pattern exhibits a periodic characteristic synchronized with the rotation period of the planetary disk, the particulate matter is determined to originate from a rotating component.

[0017] If the change pattern is a non-periodic abrupt change, the source of the particulate matter is determined to be the reaction of process gas and / or instantaneous shedding of the cavity wall.

[0018] In one embodiment, the step of performing correlation analysis between the concentration information and the rotation phase signal to obtain the variation law of particulate matter concentration with phase includes:

[0019] The concentration information is time-aligned with the rotating phase signal;

[0020] The rotating phase signal is normalized to a preset phase range, and the concentration information is registered according to the phase.

[0021] The preset phase interval is divided into multiple phase bins, and the concentration samples in each bin are periodically superimposed and statistically analyzed to obtain the phase and concentration distribution curves.

[0022] Extract the characteristic parameters of the distribution curve, and determine the variation law of particulate matter concentration with phase based on the characteristic parameters; the characteristic parameters include at least one of peak phase angle, peak-to-valley difference, main period and peak width stability.

[0023] In one embodiment, quantifying process parameters based on the particulate matter source to optimize the process environment for epitaxial growth in the reaction chamber includes:

[0024] When the particulate matter source is a process gas reaction, reduce the precursor flow rate and / or adjust the chamber pressure to suppress gas-phase nucleation;

[0025] When the particulate matter source is a rotating component, the rotational speed of the planetary disk and / or satellite disk is adjusted to change the particulate matter transport path or reduce mechanical wear.

[0026] When the particulate matter originates from the spalling of the chamber wall, the heating rate and / or temperature uniformity of the reaction chamber are adjusted to reduce thermal stress impact.

[0027] In one embodiment, the method further includes:

[0028] Acquire the reflected light signal; wherein the reflected light signal is generated by the laser signal irradiating the particulate matter in the reaction chamber through the main detection channel of the in-situ optical measurement device;

[0029] The wafer surface state information is obtained based on the reflected light signal;

[0030] A time-series correlation analysis is performed between the abrupt changes in the concentration information and the abrupt changes in the wafer surface state information. If the two are strongly correlated, the particulate matter is determined to originate from the wafer growth surface.

[0031] In one embodiment, quantifying process parameters based on the particulate matter source to optimize the process environment for epitaxial growth in the reaction chamber includes:

[0032] When the particulate matter source is the wafer growth surface, the molar flow ratio of the group V / III precursor, and / or the growth temperature, and / or the reaction chamber pressure are adjusted to improve the crystal quality and thickness uniformity of the epitaxial film.

[0033] In one embodiment, the process parameters include at least one of reaction chamber geometric modeling parameters, material optical property parameters, and reaction gas environment parameters; the property parameters include at least one of spatial position, wavelength, beam shape, and polarization state.

[0034] Based on the process parameters of the reaction chamber, the forward model database is formed by simulating the photon acquisition path under the stated process parameter conditions using the Monte Carlo method, including:

[0035] Configure the attribute parameters of the laser signal according to the process parameters;

[0036] Based on the aforementioned attribute parameters, the Monte Carlo method is used to simulate the random walk of photons within the reaction cavity. The arrival energy of each photon is statistically analyzed within a preset detection angle to calculate the expected scattered light intensity corresponding to different combinations of equivalent size and equivalent refractive index. Based on this, a forward model database indexed by particulate matter attributes and measurement configuration is generated.

[0037] In one embodiment, the step of comparing the scattered light signal with a preset forward model database based on an optimal estimation algorithm to obtain particulate matter concentration information includes:

[0038] Based on the scattered light signal, equivalent parameters characterizing the inherent properties of particulate matter are obtained from the forward model database; wherein, the equivalent parameters include equivalent size and equivalent refractive index;

[0039] Within a preset statistical time period, the number and amplitude of the scattered light signal pulses are counted to obtain statistical results;

[0040] Based on the statistical results and the equivalent parameters, the number of particles per unit volume is calculated using the principle of light scattering particle counting, and the concentration information is obtained.

[0041] In one embodiment, the step of calculating the number of particles per unit volume and obtaining concentration information based on the statistical results and the equivalent parameters using the principle of light scattering particle counting includes:

[0042] Based on the statistical results and the equivalent parameters, query the corresponding single-particle response amplitude threshold and detection probability in the positive model database;

[0043] Based on the response amplitude threshold, valid pulses are identified and classified into the corresponding size categories according to their amplitudes to obtain the valid counts for each size category.

[0044] Based on the detection probability of each size class, the effective count is corrected for visibility to obtain the arrival rate of the particle arrival event;

[0045] Based on the monitored volume, the arrival rate is converted into particulate matter concentration information per unit volume according to the Poisson arrival model.

[0046] Secondly, this application proposes an optimization system for the process environment within an epitaxial growth reaction chamber, the system comprising:

[0047] The acquisition module is used to acquire the scattered light signal and the rotation phase signal of the planetary disk in the reaction cavity; wherein, the scattered light signal is generated by the laser signal irradiating the particulate matter in the reaction cavity through the lateral detection channel of the in-situ optical measurement device;

[0048] The comparison module is used to compare the scattered light signal with a preset forward model database based on the optimal estimation algorithm, and to obtain the concentration information of particulate matter. The forward model database is used to characterize the mapping relationship between particulate matter properties and scattering response at the detection point under given wavelength and scattering angle conditions. It is based on the process parameters of the reaction chamber and uses the Monte Carlo method combined with a computational fluid dynamics model to simulate the optical acquisition path formation of photons under dynamic rotating process parameter conditions.

[0049] The identification module is used to determine the source of particulate matter based on the concentration information and the rotation phase signal;

[0050] The optimization module is used to quantify process parameters based on the source of the particulate matter in order to optimize the process environment for epitaxial growth in the reaction chamber.

[0051] Thirdly, 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 in the first aspect.

[0052] Fourthly, 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.

[0053] The above-mentioned method and system for optimizing the process environment within the epitaxial growth reaction chamber have at least the following advantages:

[0054] This application acquires the lateral scattered light signal of particulate matter within the reaction chamber and the rotational phase information of the planetary disk. It then compares this information with an optimal estimation algorithm and a forward model database built based on the process scenario. This allows for real-time inversion of particulate matter concentration and sources during epitaxial growth, and quantification of process parameters related to particulate matter generation, enabling dynamic optimization of the epitaxial growth process environment. This approach can promptly identify particulate matter risks caused by different mechanisms, allowing for targeted adjustments to process parameters. This effectively suppresses particulate matter accumulation and contamination events, improves epitaxial growth stability and yield, and ensures film quality and process consistency. Attached Figure Description

[0055] Figure 1 This is an application environment diagram of an optimization method for the process environment within an epitaxial growth reaction chamber in one embodiment;

[0056] Figure 2 This is a flowchart illustrating a method for optimizing the process environment within an epitaxial growth reaction chamber in one embodiment.

[0057] Figure 3 This is a flowchart illustrating the steps for determining the source of particulate matter in one embodiment;

[0058] Figure 4 This is a flowchart illustrating a method for optimizing the process environment within the epitaxial growth reaction chamber in another embodiment.

[0059] Figure 5 This is a structural block diagram of an optimization system for the process environment within an epitaxial growth reaction chamber in one embodiment;

[0060] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0061] 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.

[0062] 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.

[0063] The method for optimizing the process environment within the epitaxial growth reaction chamber provided in this application embodiment can be applied to, for example... Figure 1The application environment shown is as follows. The reaction chamber 1 contains a rotatable planetary disk, on which multiple rotatable satellite disks are mounted. Each satellite disk has multiple wafer carrier disks for placing wafers. The planetary disk rotates under motor control, and each satellite disk rotates due to airflow.

[0064] An in-situ measurement device 2 is disposed above the reaction cavity 1. The in-situ measurement device 2 includes a laser source and a detector. In this embodiment, the in-situ measurement device 2 adopts a side-scattering detection layout. The laser beam generated by the laser source enters the reaction cavity 1 from one side, and a shielded optical trap is disposed on the other side of the reaction cavity 1 to completely absorb the transmitted light and reduce background noise. The detector is disposed at a specific angle to the laser beam to collect the side-scattered light signal caused by particles during epitaxial growth. For example, the specific angle in this embodiment is 90°; the laser source is a near-infrared (e.g., 850nm) VCSEL laser, which has the advantages of small size, low power consumption, good beam quality, and easy integration, and is very suitable for installation on space-constrained MOCVD equipment; the detector is a high-sensitivity photodetector. The optical acquisition path arranged at a 90° angle to the incident laser beam forms a side detection channel.

[0065] Furthermore, the in-situ measurement device 2 also includes a main detection channel. The main detection channel is positioned coaxially or quasi-coaxially with the incident laser beam. A beam splitter guides the returned light to the detector to collect specular reflection and backscattered signals, thereby obtaining the reflected light signal. Typically, the main detection channel and the lateral detection channel share a single laser source to ensure comparability and synchronization of responses from both channels to the same monitored volume at the same time. Each of the main detection channel and the lateral detection channel is equipped with an independent detector to avoid crosstalk and facilitate separate optimization of gain, bandwidth, and filtering.

[0066] An optimization system 3 for the process environment within the epitaxial growth reaction chamber is connected to an in-situ measurement device 2. This system acquires scattered light signals from the in-situ measurement device 2 and obtains the rotational phase signal of the planetary disk within the reaction chamber. Based on an optimal estimation algorithm, the scattered light signals are compared with a pre-defined forward model database to invert the particulate matter concentration information. The forward model database characterizes the mapping relationship between particulate matter properties and the scattering response at the detection point under given wavelength and scattering angle conditions. It is based on the process parameters of the reaction chamber and employs a Monte Carlo method incorporating computational fluid dynamics to simulate the optical acquisition path formation of photons under dynamically rotating process parameters. The source of particulate matter is determined based on the concentration information and the rotational phase signal. Based on the particulate matter source, the process parameters are quantified to optimize the process environment for epitaxial growth within the reaction chamber.

[0067] The aforementioned method for optimizing the process environment within the epitaxial growth reaction chamber acquires the lateral scattered light signal of particles and the rotational phase information of the planetary disk within the reaction chamber. It then compares this information with a forward model database built based on the process scenario using an optimal estimation algorithm. This allows for real-time inversion of particle concentration and source information during epitaxial growth, and quantification of process parameters related to particle generation, thus achieving dynamic optimization of the epitaxial growth process environment. This approach enables timely identification of particulate risks caused by different mechanisms, targeted adjustment of process parameters, and effective suppression of particulate accumulation and contamination events. This improves epitaxial growth stability and yield, ensuring film quality and process consistency.

[0068] To further explain, particulate matter severely disrupts the continuity of the single-crystal structure of epitaxial layers, leading to a series of defects in the epitaxial film, including: surface morphology defects such as pits, protrusions, mounds, and disordered growth steps; crystal defects, inducing dislocations and stacking faults, which become non-radiative recombination centers for charge carriers, reducing device efficiency; forming optical scattering points, causing light scattering and absorption in optoelectronic devices, significantly reducing the luminous efficiency of LEDs or the light emission threshold of lasers; and degrading electrical performance, leading to premature local breakdown, increased leakage current, and threshold voltage drift in electronic devices, severely impairing the uniformity, reliability, and lifespan of devices. Therefore, particulate matter contamination within the reaction chamber has become one of the key bottlenecks restricting the improvement of yield and technological development in advanced semiconductor epitaxial processes.

[0069] Currently, in the field of semiconductor epitaxial growth, the monitoring and control of particulate matter pollution still mainly relies on a series of traditional, passive, and lagging technologies, which have fundamental technical shortcomings.

[0070] First, the offline and non-in-situ nature of monitoring methods leads to a significant delay in problem detection. The industry generally relies on off-site offline inspection, where the wafer is removed from the reaction chamber after growth is complete, and then observed using equipment such as optical microscopes, surface defect inspection machines, and atomic force microscopes. This method cannot provide any early warning information during the process. By the time batch-level particulate contamination is detected, multiple defective wafers have often already been produced, resulting in a waste of materials, time, and energy costs.

[0071] Secondly, there is a lack of real-time, in-situ sensing capabilities for particle dynamics within the cavity. Because the epitaxial reaction cavity is typically located at temperatures ranging from hundreds to thousands of degrees Celsius, and may be exposed to corrosive, flammable reactive gases, plasma, or strong radiation, and because the optical access window is limited, traditional commercial particle monitoring instruments are difficult to integrate directly and operate stably for extended periods. This results in a significant blind spot throughout the core growth process. Users cannot know in real time when, where, or why particles are generated, nor can they track their trajectory within the cavity or whether they ultimately land on the wafer surface.

[0072] Secondly, the lack of particulate matter source traceability makes it difficult to pinpoint the root cause of problems. Due to the lack of real-time, in-situ monitoring data, process engineers struggle to trace the source when particulate contamination is detected through offline detection. The potential sources of particulate matter are extremely broad and complex, potentially originating from precursor reactions, internal components, transient changes in process parameters, or introduction during maintenance operations.

[0073] Finally, passive control amplifies process fluctuations and reduces production efficiency. The lack of monitoring and traceability capabilities directly leads to passive control. Current control measures are mostly based on fixed time periods or conservative experience, such as periodic cavity cleaning, excessively tight process windows, and the inability to intervene in real time.

[0074] Based on this, this application provides a method for optimizing the process environment inside the epitaxial growth reaction chamber, which can promptly identify particulate matter risks caused by different mechanisms, adjust process parameters accordingly, thereby effectively suppressing particulate matter accumulation and contamination events, improving epitaxial growth stability and yield, and ensuring film quality and process consistency.

[0075] Please see Figure 2 In one exemplary embodiment, this application provides a method for optimizing the process environment within an epitaxial growth reaction chamber, specifically including the following steps:

[0076] Step 202: Acquire the scattered light signal and the rotation phase signal of the planetary disk in the reaction cavity; wherein, the scattered light signal is generated by the laser signal irradiating the particles in the reaction cavity through the lateral detection channel of the in-situ optical measurement device.

[0077] Specifically, the rotating phase signal refers to the representation of the angular position or phase angle of the planetary disk at any given moment, which can be denoted as φ(t)∈[0,2π) or the corresponding coded number. It is used to align optical measurements with the spatial orientation of the wafer within the cavity. Typically, the rotating phase signal can be obtained by an angle encoder installed on the planetary disk drive shaft.

[0078] Step 204: Based on the optimal estimation algorithm, the scattered light signal is compared with a preset forward model database to obtain the concentration information of particulate matter. The forward model database characterizes the mapping relationship between particulate matter properties and the scattering response at the detection point under given wavelength and scattering angle conditions. It is based on the process parameters of the reaction chamber and employs a Monte Carlo method incorporating computational fluid dynamics to simulate the optical acquisition path formation of photons under dynamically rotating process parameter conditions.

[0079] Specifically, the optimal estimation algorithm refers to selecting the parameter estimation method that optimizes the objective function among given observation data and the model. Optimal means achieving one of the following: maximum likelihood, minimum residual, or maximum posterior probability. In this embodiment, it is used to compare the measured scattered light signal with the forward model database to retrieve concentration information such as equivalent size and equivalent refractive index.

[0080] The given wavelength refers to the center wavelength of the laser used for monitoring and simulation, such as 850 nm in the example above. The scattering angle refers to the angle between the optical axis of the probe channel and the axis of the incident laser beam. Particle properties refer to the optical and geometric eigenvalues ​​used to characterize the particles under test, including at least equivalent size and equivalent refractive index. The detector point scattering response refers to the expected signal characteristics produced by a single particle at the detector at a given wavelength and scattering angle, typically including quantitative indicators such as lateral arrival energy, expected single-pulse amplitude, phase function-weighted angular distribution, and corresponding detection probability.

[0081] With the wavelength and scattering angle known, and using particulate matter properties as input, the forward model database can provide the expected scattering response at the detection location. During measurement, the measured scattered light signal is compared with the forward model database using an optimal estimation algorithm to deduce the best-matching particulate matter properties, thereby calculating the particulate matter concentration.

[0082] Computational fluid dynamics (CFD) models are numerical models used to solve for gas flow and heat and mass transfer within a reaction chamber, outputting information such as velocity, temperature, pressure, and composition fields. By constraining the medium properties of photon motion and the probability of particle presence, they ensure that the forward model database is consistent with real-world operating conditions.

[0083] Dynamic rotation process parameters refer to the operating state of the planetary disk as it rotates continuously during the process and changes the overlapping area of ​​the effective field of view and the distribution of the gas field in the cavity over time. These dynamic rotation process parameters require the consideration of the phase and time dependence in forward modeling and comparison.

[0084] Using the current process parameters as input, the velocity field, temperature field, and composition field within the reaction chamber are solved using a computational fluid dynamics model. These are then used as boundary conditions and weights for Monte Carlo photon transport and lateral collection geometry under dynamic rotating phase. Thus, given a wavelength and scattering angle configuration, the expected scattering response at the detection point is calculated for different combinations of equivalent particle size and spatial position, and stored as a searchable forward model database.

[0085] Step 206: Determine the source of particulate matter based on concentration information and rotation phase signal.

[0086] Specifically, particulate matter sources refer to the location and mechanism of particulate matter generation and detachment within the reaction chamber, typically including rotating follower sources, fixed component sources, gas-phase nucleation and aggregation sources, flow path-carried sources, and operational disturbance sources. Since the planetary disk rotates periodically, analyzing the changes in the rotational phase signal within the concentration sequence allows for the screening of phase-related and phase-independent particulate matter sources.

[0087] Step 208: Quantify process parameters based on the source of particulate matter to optimize the process environment for epitaxial growth in the reaction chamber.

[0088] Specifically, the process environment for epitaxial growth in a reaction chamber refers to the comprehensive process conditions within the reaction chamber that support the stable deposition of the epitaxial film, including key environmental factors that affect the quality of epitaxial growth and particle formation, such as gas flow and ratio, temperature field, pressure, cleanliness of the reaction atmosphere, and surface condition of the chamber.

[0089] Process parameters refer to a set of settings that determine the growth environment and reaction kinetics. They are usually stored in the MES management system and sent to the machine for execution before the process begins.

[0090] After determining the source of particulate matter, the source and cause of particulate matter are mapped to the deviation of specific process parameters. This allows us to determine which parameters to quantify and in which direction to adjust them, thereby reducing the probability of particulate matter generation and shedding, and thus optimizing the epitaxial growth environment of the reaction chamber.

[0091] The aforementioned method for optimizing the process environment within the epitaxial growth reaction chamber acquires the lateral scattered light signal of particles and the rotational phase information of the planetary disk within the reaction chamber. It then compares this information with a forward model database built based on the process scenario using an optimal estimation algorithm. This allows for real-time inversion of particle concentration and source information during epitaxial growth, and quantification of process parameters related to particle generation, thus achieving dynamic optimization of the epitaxial growth process environment. This approach enables timely identification of particulate risks caused by different mechanisms, targeted adjustment of process parameters, and effective suppression of particulate accumulation and contamination events. This improves epitaxial growth stability and yield, ensuring film quality and process consistency.

[0092] Optionally, the laser signal is emitted by a pulse-modulated laser source; the methods for acquiring the scattered light signal include:

[0093] Based on time-gating technology, the scattered light signal is acquired within a specified time window after the laser signal is emitted. The specified time window is a sampling time interval set based on the time difference between the laser signal emission time and the expected arrival time of the particulate scattering signal in the target monitoring area. It is used to receive the particulate scattering signal and suppress background reflection from non-target areas. Non-target areas include the reaction cavity wall and the wafer interface.

[0094] Specifically, the laser signal is emitted by a pulse-modulated laser source with a pulse width on the order of tens of nanoseconds. Signal acquisition employs a time-gated method. A high-precision timer is activated after the laser pulse is triggered, and after a preset delay, effective sampling by the detector only begins within a specified time window. This specified time window is calculated and set based on the beam path length, the location of the monitored volume, and the propagation time of the scattered light. For example, a 50ns delay is used for window opening, with a window width of 30ns, ensuring that only scattered light signals from particles in the monitored area are received. During the window's open period, the detector operates synchronously with data sampling, acquiring the amplitude and occurrence time of the scattered pulse. When the window closes, the detector bias is maintained, but the signal link is in a suppressed state, thereby shielding against strong reflections or parasitic fluorescence from non-target areas such as the reaction chamber walls and wafer interfaces.

[0095] The above scheme employs a pulsed laser combined with time-gated detection. By collecting the scattered light signal only within a specified short time window after laser emission, the background interference caused by strong reflections from non-target areas such as cavity walls and wafer interfaces can be significantly suppressed, improving the signal-to-noise ratio and detection sensitivity of the particle scattered light signal. At the same time, it avoids the accumulation of thermal background and optical drift caused by continuous irradiation, which is conducive to achieving in-situ accurate particle monitoring under high temperature and strong reaction environments, providing a more reliable data foundation for subsequent concentration inversion and source identification.

[0096] Please see Figure 3 Optionally, the source of particulate matter can be determined based on concentration information and rotational phase signal, including:

[0097] Step 302: Correlation analysis is performed on the concentration information and the rotation phase signal to obtain the variation law of particulate matter concentration with phase.

[0098] Step 304: If the change pattern has a periodic characteristic that is synchronized with the rotation period of the planetary disk, the particulate matter is determined to originate from the rotating component.

[0099] Step 306: If the change pattern is a non-periodic abrupt change, determine that the source of particulate matter is the reaction of process gas and / or instantaneous shedding of the cavity wall.

[0100] Optionally, the concentration information and the rotating phase signal are correlated to obtain the variation law of particulate matter concentration with phase, including: time alignment of concentration information and rotating phase signal; normalization of rotating phase signal to a preset phase interval, and registration of concentration information according to phase; dividing the preset phase interval into multiple phase bins, and performing periodic superposition statistics on concentration samples in each bin to obtain the distribution curve of phase and concentration; extracting characteristic parameters of the distribution curve, and determining the variation law of particulate matter concentration with phase based on the characteristic parameters; the characteristic parameters include at least one of peak position phase angle, peak-to-valley difference, main period and peak width stability.

[0101] Specifically, a unified time base is used to align the concentration information obtained from the real-time acquired scattered light signals with the rotating phase signal. The continuous rotating phase signal is unfolded into a multi-period sequence, and each concentration sampling point is registered to the corresponding instantaneous phase according to the timestamp. The phase interval is divided into multiple equal-width phase bins, and the concentration samples within the same phase bin are aggregated. The data from the same phase bin within multiple rotation periods are superimposed and averaged or statistically analyzed to obtain the phase and concentration distribution curves. Feature parameters in the distribution curves are extracted to form a quantitative description of the concentration variation with phase.

[0102] When the concentration exhibits a periodic modulation consistent with the planetary disk rotation period and the peak position stably corresponds to the carrier's rotation phase, the particulate matter is considered to originate from rotating components moving with the planetary mechanism, such as micro-detachment from the graphite substrate, wafer edge, or tray surface. When the concentration shows a sudden spike or short-term jump independent of the phase over time, and then quickly returns to the baseline level, the particulate matter is considered to originate from transient agglomeration and precipitation caused by gas-phase side reactions or transient peeling off of local deposits on the cavity wall.

[0103] By adopting the above scheme and performing correlation analysis between particulate matter concentration and rotation phase, this embodiment can achieve precise location of particulate matter source, thereby supporting targeted optimization of process parameters and significantly improving the stability of epitaxial generation process and film quality.

[0104] Optionally, based on the source of particulate matter, process parameters are quantified to optimize the process environment for epitaxial growth in the reaction chamber, including:

[0105] When the particulate matter source is the process gas reaction, reduce the precursor flow rate and / or adjust the chamber pressure to suppress gas-phase nucleation;

[0106] When the particulate matter source is a rotating component, adjust the rotational speed of the planetary disk and / or satellite disk to change the particulate matter transport path or reduce mechanical wear;

[0107] When the particulate matter originates from the spalling of the chamber wall, adjust the heating rate and / or temperature uniformity of the reaction chamber to reduce thermal stress shock.

[0108] Specifically, the above adjustment process should adopt a closed-loop evaluation method. After each parameter adjustment, the concentration and phase characteristics should be continuously observed to see if they return to the safe range. If not, fine-tuning should continue according to the preset small step size until the threshold and stability criteria are met.

[0109] The above-mentioned approach enables differentiated control based on different causes. For gas-phase nucleation, supersaturation is reduced and particle formation is suppressed by fine-tuning the precursor flow rate and pressure. For detachment of rotating components, wear is reduced and particle transport paths are altered by optimizing the rotation speed ratio, thus decreasing the particle throughput within the field of view. For cavity wall spalling, thermal stress impact and spalling probability are reduced by adjusting the thermal history and temperature uniformity. This application significantly reduces peak and fluctuating particle concentration, decreases sudden contamination events, improves the surface cleanliness and inter-wafer consistency of epitaxial films, and enhances yield and process stability.

[0110] Please see Figure 4 Optionally, the above-mentioned method for optimizing the process environment within the epitaxial growth reaction chamber further includes:

[0111] Step 402: Acquire the reflected light signal; wherein, the reflected light signal is generated by the laser signal irradiating the particulate matter in the reaction chamber through the main detection channel of the in-situ optical measurement device.

[0112] Step 404: Obtain wafer surface state information based on the reflected light signal.

[0113] Step 406: Perform time-series correlation analysis on the abrupt changes in concentration information and the abrupt changes in wafer surface state information. If the two are strongly correlated, the particulate matter is determined to originate from the wafer growth surface.

[0114] Specifically, wafer surface state information refers to characteristic quantities reflecting the optical and morphological state of the wafer growth surface, including at least one of reflectivity, fringe amplitude and phase, baseline drift, and instantaneous decoherence index. Feature extraction is performed on the reflected light signal to obtain characteristic parameters such as reflectivity, fringe amplitude, fringe phase, and frequency. These characteristic parameters are then statistically analyzed within a sliding window to obtain the wafer surface state information.

[0115] A sudden change in concentration information refers to an event in which the number concentration of particulate matter deviates significantly from the baseline level within a short period of time. This is manifested as a spike, a step-like increase or decrease, the magnitude of which exceeds a preset threshold and the duration is not less than the minimum confirmation time.

[0116] A sudden change in surface state information refers to a significant change in the characteristic parameters of the reflected light signal within a short period of time, which is also determined by a threshold and a minimum duration.

[0117] If the two mutation events mentioned above are highly synchronized in time within the same time window and the consistency of their occurrence exceeds a preset threshold, then they are considered to have a strong correlation. For example, at least one of the following methods can be used to quantify this strong correlation: event-level matching rate, Jaccard coefficient, Pearson / Spearman correlation coefficient, mutual information, etc.

[0118] Optionally, based on the source of particulate matter, the process parameters are quantified to optimize the process environment for epitaxial growth in the reaction chamber, and the method further includes:

[0119] When the particulate matter source is the wafer growth surface, the molar flow ratio of the V / III group precursor, and / or the growth temperature, and / or the reaction chamber pressure can be adjusted to improve the crystal quality and thickness uniformity of the epitaxial film.

[0120] By adopting the above scheme, when the reflected light signal is obtained and the particle source is determined to be the wafer growth surface, the V / III precursor molar flow ratio, growth temperature and cavity pressure can be adjusted in a targeted manner to reduce the surface nucleation driving force and gas phase supersaturation, suppress microparticle adhesion and surface roughening, and stabilize step flow growth. This reduces crystal defects and stress fluctuations in the thin film, narrows the XRD full width at half maximum, increases PL luminescence intensity, and reduces AFM surface roughness. At the same time, it improves the arrival and reaction uniformity of the precursor, enhances the thickness uniformity of the same wafer and between wafers, and ultimately improves the structure and photoelectric quality of the epitaxial film, as well as improve yield and process stability.

[0121] Optionally, the process parameters include at least one of the following: reaction chamber geometric modeling parameters, material optical property parameters, and reaction gas environment parameters; the property parameters include at least one of the following: spatial position, wavelength, beam shape, and polarization state.

[0122] Based on the process parameters of the reaction chamber, a forward model database is formed by simulating the photon acquisition path under the process parameter conditions using the Monte Carlo method, including:

[0123] Based on the process parameters, the attribute parameters of the laser signal are configured. Based on the attribute parameters, the Monte Carlo method is used to simulate the random walk of photons in the reaction cavity. The arrival energy of each photon is counted within the preset detection angle to calculate the expected scattered light intensity corresponding to different combinations of equivalent size and equivalent refractive index. Based on this, a forward model database indexed by particulate matter attributes and measurement configuration is generated.

[0124] Specifically, before constructing the forward model database, the process parameters used for simulation are first obtained to build a cavity model. The reaction cavity geometric modeling parameters characterize the three-dimensional geometry of the reaction cavity and the positions of key components, such as the spray head, graphite base, heating screen, and viewing window. The optical property parameters of the component materials include the refractive index and transmittance of the quartz viewing window, and the absorptivity and diffuse reflectivity of the graphite base. The reaction gas environment parameters under the current process conditions include the proportion of NH3 / III group precursors, cavity pressure, approximate temperature distribution, and refractive index model. Subsequently, based on the process settings, the laser signal is configured as an 850 nm single-mode Gaussian beam, and the linear polarization direction, divergence angle, and incident position are set.

[0125] Based on this, the Monte Carlo method is used to simulate the propagation trajectory of a large number of photons in the aforementioned cavity model. During propagation, photons undergo absorption, specular reflection, or diffuse reflection according to the material's optical properties and the gas refractive index distribution. When scattered by predetermined particle properties, angular scattering occurs; these particle properties are configured with different combinations of equivalent size and equivalent refractive index. The system statistically analyzes the photon energy arriving at the detection window within a preset detection angle range of 90° laterally, obtaining the expected scattered light intensity under the corresponding particle property conditions. By scanning different particle size and refractive index combinations, a forward model database indexed by particle properties and measurement configurations is generated; the measurement configurations include wavelength, scattering angle, beam shape, and polarization state.

[0126] By employing the above-mentioned scheme, and through Monte Carlo simulation of the photon propagation and scattering process within the reaction cavity under realistic process parameters, and by constructing a forward model database indexed by particle properties and measurement configuration, this application achieves high-fidelity particle scattering prediction. This enables accurate inversion of particle size and refractive index information during epitaxial growth, significantly improving the accuracy of particle identification and concentration inversion. Simultaneously, the constructed forward model database pre-considers the effects of beam morphology, polarization state, and probe angle, making the inversion process more robust to changes in the cavity optical environment. This facilitates real-time monitoring and process closed-loop optimization, thereby improving the cleanliness and yield of the epitaxial film.

[0127] Optionally, based on the optimal estimation algorithm, the scattered light signal is compared with a pre-set forward model database to obtain the concentration information of particulate matter, including:

[0128] Based on the scattered light signal, equivalent parameters characterizing the inherent properties of particulate matter are obtained from the forward model database; among them, the equivalent parameters include equivalent size and equivalent refractive index;

[0129] Within a preset statistical time period, the number and amplitude of scattered light signals are counted to obtain statistical results;

[0130] Based on statistical results and equivalent parameters, the number of particulate matter per unit volume is calculated using the principle of light scattering particle counting, thus obtaining concentration information.

[0131] Optionally, based on statistical results and equivalent parameters, the number of particulate matter per unit volume is calculated using the principle of light scattering particle counting to obtain concentration information, including:

[0132] Based on the statistical results and equivalent parameters, the corresponding single-particle response amplitude threshold and detection probability are queried in the forward model database; based on the response amplitude threshold, valid pulses are identified and classified into the corresponding size class according to their amplitude to obtain the valid count for each size class; based on the detection probability for each size class, the valid count is corrected for visibility to obtain the arrival rate of the particle arrival event; based on the monitored volume, the arrival rate is converted into the concentration information of particulate matter per unit volume according to the Poisson arrival model.

[0133] Specifically, an objective function is constructed using optimal estimation algorithms such as maximum likelihood estimation and Bayesian maximum a posteriori estimation. The pulse amplitude distribution of the measured scattered light signal is compared with the theoretical scattering response distribution corresponding to different particulate matter properties in the forward model database. Through iterative search, the error between the theoretical and measured distributions is minimized, and the best-matching equivalent size and equivalent refractive index are obtained as the equivalent optical parameters of the particulate matter.

[0134] Furthermore, within a preset statistical time period, the number and amplitude of scattered light signals are counted. Based on the equivalent parameters derived earlier, the corresponding single-particle expected response threshold and detection probability are retrieved from the forward model database. Valid pulses are identified and retained according to the expected response threshold, and then categorized into corresponding size classes based on amplitude, yielding the effective count for each size class. Visibility correction is applied to the effective counts for each size class based on the detection probability to obtain the arrival rate of particle arrival events. Finally, combined with the monitored volume, the arrival rate is converted into particulate matter concentration information per unit volume using the Poisson arrival model, outputting the real-time concentration result. The Poisson arrival model treats the events of particulate matter entering the monitored volume as random and independent arrival processes, thus obtaining a Poisson distribution with parameter λT within a time window of length T, where λ is the arrival rate, and the interval between adjacent arrivals follows an exponential distribution.

[0135] By employing the above scheme, an optimal estimation algorithm is used to accurately match the pulse distribution of the measured scattered light signal with the theoretical response of the forward model database, enabling the accurate derivation of the equivalent size and equivalent refractive index of particulate matter. Based on the deduced equivalent properties, combined with detection probability correction and a Poisson statistical model, particulate matter concentration information can be calculated with high precision, significantly improving the sensitivity and accuracy of particulate matter detection. This application effectively overcomes the sensitivity of simple threshold methods to size heterogeneity and changes in detection efficiency, achieving robust identification of particles of different sizes and materials. It is suitable for online monitoring of epitaxially grown particles under high-temperature and high-reaction environments, providing reliable data support for epitaxial process optimization and real-time control, and improving product consistency and yield.

[0136] The aforementioned method for optimizing the process environment within the epitaxial growth reaction chamber acquires the lateral scattered light signal of particles and the rotational phase information of the planetary disk within the reaction chamber. It then compares this information with a forward model database built based on the process scenario using an optimal estimation algorithm. This allows for real-time inversion of particle concentration and source information during epitaxial growth, and quantification of process parameters related to particle generation, thus achieving dynamic optimization of the epitaxial growth process environment. This approach enables timely identification of particulate risks caused by different mechanisms, targeted adjustment of process parameters, and effective suppression of particulate accumulation and contamination events. This improves epitaxial growth stability and yield, ensuring film quality and process consistency.

[0137] 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.

[0138] Based on the same inventive concept, this application also provides an optimization system for the process environment inside an epitaxial growth reaction chamber. This system is applicable to the above-mentioned optimization method for the process environment inside an epitaxial growth reaction chamber. The solution provided by this system is similar to the solution described in the above-mentioned method. Therefore, the specific limitations in one or more device embodiments provided below can be found in the limitations of the method above, and will not be repeated here.

[0139] Please see Figure 5 In one embodiment, this application provides an optimization system for the process environment within an epitaxial growth reaction chamber, comprising: an acquisition module, a comparison module, an identification module, and an optimization module.

[0140] The acquisition module is used to acquire the scattered light signal and the rotation phase signal of the planetary disk in the reaction cavity; wherein, the scattered light signal is generated by the laser signal irradiating the particles in the reaction cavity through the lateral detection channel of the in-situ optical measurement device.

[0141] The comparison module is used to compare the scattered light signal with the preset forward model database based on the optimal estimation algorithm, and to obtain the concentration information of particulate matter. The forward model database is used to characterize the mapping relationship between particulate matter properties and the scattering response of the detection point under given wavelength and scattering angle conditions. It is based on the process parameters of the reaction chamber and uses the Monte Carlo method combined with the computational fluid dynamics model to simulate the optical acquisition path formation of photons under dynamic rotating process parameter conditions.

[0142] The identification module is used to determine the source of particulate matter based on concentration information and rotation phase signal.

[0143] The optimization module is used to quantify process parameters based on the source of particulate matter in order to optimize the process environment for epitaxial growth in the reaction chamber.

[0144] Optionally, the identification module determines the source of particulate matter based on concentration information and rotation phase signal, including: performing correlation analysis on the concentration information and rotation phase signal to obtain the variation law of particulate matter concentration with phase; if the variation law has a periodic characteristic synchronized with the rotation period of the planetary disk, the source of particulate matter is determined to be a rotating component; if the variation law is a non-periodic abrupt change, the source of particulate matter is determined to be a process gas reaction and / or instantaneous shedding of the cavity wall. Specifically, performing correlation analysis on the concentration information and rotation phase signal to obtain the variation law of particulate matter concentration with phase includes: time alignment of the concentration information and rotation phase signal; normalization of the rotation phase signal to a preset phase interval, and registration of the concentration information according to phase; dividing the preset phase interval into multiple phase bins, and periodically superimposing and statistically analyzing the concentration samples in each bin to obtain the distribution curve of phase and concentration; extracting characteristic parameters of the distribution curve, and determining the variation law of particulate matter concentration with phase based on the characteristic parameters; the characteristic parameters include at least one of peak position phase angle, peak-to-valley difference, main period, and peak width stability.

[0145] Optionally, the optimization module quantifies process parameters based on the particulate matter source to optimize the process environment for epitaxial growth in the reaction chamber, including: reducing the precursor flow rate and / or adjusting the chamber pressure when the particulate matter source is a process gas reaction to suppress gas-phase nucleation; adjusting the rotational speed of the planetary disk and / or satellite disk when the particulate matter source is a rotating component to change the particulate matter transport path or reduce mechanical wear; and adjusting the heating rate and / or temperature uniformity of the reaction chamber when the particulate matter source is chamber wall spalling to reduce thermal stress impact.

[0146] Optionally, the above-mentioned optimization system for the process environment within the epitaxial growth reaction chamber also includes a processing module.

[0147] The acquisition module is also used to acquire reflected light signals; wherein, the reflected light signals are generated by the laser signal irradiating the particles in the reaction chamber through the main detection channel of the in-situ optical measurement device.

[0148] The processing module is used to obtain wafer surface state information based on the reflected light signal; it performs time-series correlation analysis on the abrupt changes in concentration information and wafer surface state information. If the two have a strong correlation, the particulate matter is determined to originate from the wafer growth surface.

[0149] The optimization module is also used to adjust the molar flow ratio of V / III group precursors, and / or growth temperature, and / or reaction chamber pressure when the particulate matter source is the wafer growth surface, in order to improve the crystal quality and thickness uniformity of the epitaxial film.

[0150] Optionally, the above-mentioned optimization system for the process environment within the epitaxial growth reaction chamber also includes a database construction module.

[0151] When the process parameters include at least one of the following: reaction cavity geometric modeling parameters, material optical property parameters, and reaction gas environment parameters; and the attribute parameters include at least one of the following: spatial position, wavelength, beam shape, and polarization state, the database construction module is used to simulate the acquisition path of photons under the process parameter conditions based on the process parameters of the reaction cavity using the Monte Carlo method to form a forward model database. The construction steps include: configuring the attribute parameters of the laser signal according to the process parameters; simulating the random walk of photons in the reaction cavity using the Monte Carlo method based on the attribute parameters; statistically analyzing the arrival energy of each photon within a preset detection angle to calculate the expected scattered light intensity corresponding to different combinations of equivalent size and equivalent refractive index; and generating a forward model database indexed by particulate matter attributes and measurement configuration.

[0152] Optionally, the comparison module, based on the optimal estimation algorithm, compares the scattered light signal with a preset forward model database to obtain the concentration information of particulate matter. This includes: querying the forward model database to obtain equivalent parameters characterizing the inherent properties of particulate matter based on the scattered light signal; wherein, the equivalent parameters include equivalent size and equivalent refractive index; counting the number and amplitude of the scattered light signal pulses within a preset statistical time period to obtain statistical results; and calculating the number of particulate matter per unit volume based on the statistical results and equivalent parameters using the principle of light scattering particle counting to obtain the concentration information. Based on statistical results and equivalent parameters, the principle of light scattering particle counting is used to calculate the number of particles per unit volume and obtain concentration information. This includes: querying the corresponding single-particle response amplitude threshold and detection probability in the forward model database based on statistical results and equivalent parameters; identifying effective pulses based on the response amplitude threshold and classifying them into the corresponding size class according to their amplitude to obtain the effective count for each size class; performing visibility correction on the effective count based on the detection probability for each size class to obtain the arrival rate of the particle arrival event; and converting the arrival rate into the concentration information of particles per unit volume according to the Poisson arrival model based on the monitored volume.

[0153] The aforementioned optimization system for the epitaxial growth reaction chamber environment acquires the lateral scattered light signal of particles within the reaction chamber and the rotational phase information of the planetary disk. It then compares this information with a forward model database built based on the process scenario using an optimal estimation algorithm. This allows for real-time inversion of particle concentration and source information during epitaxial growth, and quantification of process parameters related to particle generation, thus achieving dynamic optimization of the epitaxial growth process environment. This approach enables timely identification of particulate risks caused by different mechanisms, targeted adjustment of process parameters, effectively suppressing particulate accumulation and contamination events, improving epitaxial growth stability and yield, and ensuring film quality and process consistency.

[0154] Each module in the aforementioned optimization system for the process environment within the epitaxial growth reaction chamber can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the corresponding operations of each module.

[0155] In one feasible embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 6As 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 optimizing the process environment within the epitaxial growth reaction chamber. 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.

[0156] Those skilled in the art will understand that Figure 6 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, but are not limited to, the following: Figure 6 The diagram shows more or fewer components, or combinations of certain components, or different component arrangements.

[0157] 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 optimizing the process environment within the epitaxial growth reaction chamber.

[0158] 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 optimizing the process environment within the epitaxial growth reaction chamber.

[0159] 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 optimizing the process environment within the epitaxial growth reaction chamber.

[0160] 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.

[0161] 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.

[0162] 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 for optimizing the process environment within an epitaxial growth reaction chamber, characterized in that, The method includes: Acquire the scattered light signal and the rotation phase signal of the planetary disk in the reaction cavity; wherein, the scattered light signal is generated by the laser signal irradiating the particulate matter in the reaction cavity through the lateral detection channel of the in-situ optical measurement device; Based on the optimal estimation algorithm, the scattered light signal is compared with a preset forward model database to obtain the concentration information of particulate matter. This includes: querying the forward model database to obtain equivalent parameters characterizing the inherent properties of particulate matter based on the scattered light signal; wherein, the equivalent parameters include equivalent size and equivalent refractive index; counting the number and amplitude of the scattered light signal pulses within a preset statistical time period to obtain statistical results; and calculating the number of particulate matter per unit volume based on the statistical results and the equivalent parameters using the principle of light scattering particle counting to obtain concentration information. The forward model database is used to characterize the mapping relationship between particulate matter properties and the scattering response of the detection point under given wavelength and scattering angle conditions. It is based on the process parameters of the reaction chamber and uses a Monte Carlo method incorporating computational fluid dynamics models to simulate the optical acquisition path formation of photons under dynamically rotating process parameter conditions. Determining the source of particulate matter based on the concentration information and the rotation phase signal includes: performing correlation analysis on the concentration information and the rotation phase signal to obtain the variation law of particulate matter concentration with phase; if the variation law has a periodic characteristic synchronized with the rotation period of the planetary disk, determining that the source of particulate matter is a rotating component; if the variation law is a non-periodic abrupt change, determining that the source of particulate matter is a process gas reaction and / or instantaneous peeling of the cavity wall. Based on the source of the particulate matter, process parameters are quantified to optimize the process environment for epitaxial growth in the reaction chamber.

2. The method according to claim 1, characterized in that, The laser signal is emitted by a pulse-modulated laser source; The methods for acquiring the scattered light signal include: Based on time-gating technology, the scattered light signal is acquired within a specified time window after the laser signal is emitted; wherein, the specified time window refers to a sampling time interval set based on the time difference between the laser signal emission time and the expected arrival time of the particulate scattering signal in the target monitoring area, which is used to receive the particulate scattering signal and suppress background reflection from non-target areas; the non-target areas include the reaction cavity wall and the wafer interface.

3. The method according to claim 1, characterized in that, The step of performing correlation analysis between the concentration information and the rotation phase signal to obtain the variation law of particulate matter concentration with phase includes: The concentration information is time-aligned with the rotating phase signal; The rotating phase signal is normalized to a preset phase range, and the concentration information is registered according to the phase. The preset phase interval is divided into multiple phase bins, and the concentration samples in each bin are periodically superimposed and statistically analyzed to obtain the phase and concentration distribution curves. Extract the characteristic parameters of the distribution curve, and determine the variation law of particulate matter concentration with phase based on the characteristic parameters; the characteristic parameters include at least one of peak phase angle, peak-to-valley difference, main period and peak width stability.

4. The method according to claim 1, characterized in that, The step of quantifying process parameters based on the source of the particulate matter to optimize the process environment for epitaxial growth in the reaction chamber includes: When the particulate matter source is a process gas reaction, reduce the precursor flow rate and / or adjust the chamber pressure to suppress gas-phase nucleation; When the particulate matter source is a rotating component, the rotational speed of the planetary disk and / or satellite disk is adjusted to change the particulate matter transport path or reduce mechanical wear. When the particulate matter originates from the spalling of the chamber wall, the heating rate and / or temperature uniformity of the reaction chamber are adjusted to reduce thermal stress impact.

5. The method according to claim 1, characterized in that, The method further includes: Acquire the reflected light signal; wherein the reflected light signal is generated by the laser signal irradiating the particulate matter in the reaction chamber through the main detection channel of the in-situ optical measurement device; The wafer surface state information is obtained based on the reflected light signal; A time-series correlation analysis is performed between the abrupt changes in the concentration information and the abrupt changes in the wafer surface state information. If the two are strongly correlated, the particulate matter is determined to originate from the wafer growth surface.

6. The method according to claim 5, characterized in that, The step of quantifying process parameters based on the source of the particulate matter to optimize the process environment for epitaxial growth in the reaction chamber includes: When the particulate matter source is the wafer growth surface, the molar flow ratio of the group V / III precursor, and / or the growth temperature, and / or the reaction chamber pressure are adjusted to improve the crystal quality and thickness uniformity of the epitaxial film.

7. The method according to claim 1, characterized in that, The process parameters include at least one of the following: reaction chamber geometric modeling parameters, material optical property parameters, and reaction gas environment parameters; the attribute parameters include at least one of the following: spatial position, wavelength, beam shape, and polarization state. Based on the process parameters of the reaction chamber, the forward model database is formed by simulating the photon acquisition path under the stated process parameter conditions using the Monte Carlo method, including: Configure the attribute parameters of the laser signal according to the process parameters; Based on the aforementioned attribute parameters, the Monte Carlo method is used to simulate the random walk of photons within the reaction cavity. The arrival energy of each photon is statistically analyzed within a preset detection angle to calculate the expected scattered light intensity corresponding to different combinations of equivalent size and equivalent refractive index. Based on this, a forward model database indexed by particulate matter attributes and measurement configuration is generated.

8. The method according to claim 1, characterized in that, The step involves calculating the number of particles per unit volume based on the statistical results and the equivalent parameters, using the principle of light scattering particle counting, to obtain concentration information, including: Based on the statistical results and the equivalent parameters, query the corresponding single-particle response amplitude threshold and detection probability in the positive model database; Based on the response amplitude threshold, valid pulses are identified and classified into the corresponding size categories according to their amplitudes to obtain the valid counts for each size category. Based on the detection probability of each size class, the effective count is corrected for visibility to obtain the arrival rate of the particle arrival event; Based on the monitored volume, the arrival rate is converted into particulate matter concentration information per unit volume according to the Poisson arrival model.

9. A system for optimizing the process environment within an epitaxial growth reaction chamber, characterized in that, The system includes: The acquisition module is used to acquire the scattered light signal and the rotation phase signal of the planetary disk in the reaction cavity; wherein, the scattered light signal is generated by the laser signal irradiating the particulate matter in the reaction cavity through the lateral detection channel of the in-situ optical measurement device; The comparison module is used to compare the scattered light signal with a preset forward model database based on an optimal estimation algorithm to obtain the concentration information of particulate matter. This includes: querying the forward model database to obtain equivalent parameters characterizing the inherent properties of particulate matter based on the scattered light signal; wherein the equivalent parameters include equivalent size and equivalent refractive index; counting the number and amplitude of the scattered light signal pulses within a preset statistical time period to obtain statistical results; and calculating the number of particulate matter per unit volume based on the statistical results and the equivalent parameters using the principle of light scattering particle counting to obtain concentration information. The forward model database is used to characterize the mapping relationship between particulate matter properties and the scattering response of the detection point under given wavelength and scattering angle conditions. It is based on the process parameters of the reaction chamber and uses a Monte Carlo method incorporating computational fluid dynamics models to simulate the optical acquisition path formation of photons under dynamically rotating process parameter conditions. The identification module is used to determine the source of particulate matter based on the concentration information and the rotation phase signal, including: performing correlation analysis on the concentration information and the rotation phase signal to obtain the variation law of particulate matter concentration with phase; if the variation law has a periodic characteristic synchronized with the rotation period of the planetary disk, determining that the source of particulate matter is a rotating component; if the variation law is a non-periodic abrupt change, determining that the source of particulate matter is a process gas reaction and / or instantaneous peeling of the cavity wall; The optimization module is used to quantify process parameters based on the source of the particulate matter in order to optimize the process environment for epitaxial growth in the reaction chamber.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1-8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-8.

Images