CN-121655621-B - MOCVD process multi-parameter synchronous detection method, system, equipment and storage medium

CN121655621BCN 121655621 BCN121655621 BCN 121655621BCN-121655621-B

Abstract

The application provides a MOCVD process multiparameter synchronous detection method, a system, equipment and a storage medium, which relate to the technical field of semiconductor detection, and the application divides single-mode laser into object light and reference light, and utilizes the reference light to form spherical reference waves through point diffraction, so that the object light and the spherical reference waves are spatially overlapped near the surface of a semiconductor to form stable interference fringes; and synchronously extracting a phase distribution map and a contrast distribution map from each frame of interference image, and synchronously decoupling and inverting multiple parameters under a unified data source and a spatial reference system, so that in-situ, real-time and full-field synchronous acquisition and association analysis of semiconductor morphology, temperature and optical reflection characteristics can be realized under MOCVD high-temperature strong radiation and semiconductor rotation conditions, the problems of time-space asynchronism and information splitting caused by a plurality of sets of independent monitoring systems are avoided, the complexity of system integration and calibration is reduced, and a unified, accurate and comprehensive data base is provided for process optimization.

Inventors

LIU XINYANG
ZHANG LIMING
Xuan Guangpu

Assignees

上海车仪田科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260206

Claims (11)

1. The MOCVD process multiparameter synchronous detection method is characterized by comprising the following steps of: the method comprises the steps of obtaining a reference phase distribution diagram of a process initial stage and an interference image sequence in a process processing stage, wherein the interference image sequence is obtained by dividing single-mode laser into object light and reference light, forming spherical reference waves by point diffraction of the reference light, enabling the object light and the spherical reference waves to overlap in space near a semiconductor surface to form interference fringes, and then carrying out imaging acquisition on the interference fringes based on trigger signals, wherein the trigger signals are generated based on synchronous pulse signals of a rotary encoder; For each frame of interference image, respectively carrying out phase extraction and contrast extraction on each frame of interference image to obtain a phase distribution map and a contrast distribution map; based on the difference between the phase distribution map and the reference phase distribution map, and the contrast distribution map, performing multi-parameter synchronous decoupling and inversion processing to synchronously obtain distribution results of semiconductor morphology-related parameters, temperature-related parameters, and optical reflection-related parameters, including: The method comprises the steps of calculating a phase difference distribution diagram of the phase distribution diagram and a reference phase distribution diagram, performing spatial domain component separation processing on the phase difference distribution diagram to obtain a geometric phase component representing the appearance of a semiconductor surface, performing height conversion on the geometric phase component to obtain a height variation distribution of the semiconductor surface, and performing two-dimensional numerical differentiation on the height variation distribution to obtain a curvature distribution; The geometrical phase component is subtracted from the difference value to obtain thermotropic phase distribution, a temperature change field is obtained by inversion based on the thermotropic phase distribution and preset material parameters, and the temperature field distribution is determined based on the temperature change field, wherein the preset material parameters comprise a thermo-optical coefficient, a thermal expansion coefficient, an initial refractive index and a thickness corresponding to a semiconductor material to be detected; and carrying out normalized calibration on the contrast distribution map and standard parameters to obtain relative reflectivity distribution, wherein the standard parameters are obtained based on standard reflector plate measurement.
2. The method of claim 1, wherein for each frame of interference image, performing phase extraction to obtain a phase profile comprises: performing two-dimensional Fourier transform on each frame of interference image to obtain a frequency domain spectrum; After the frequency domain spectrum is subjected to frequency domain filtering, primary frequency spectrum components comprising phase information are separated; Performing inverse Fourier transform after the primary frequency spectrum component is shifted to zero frequency to obtain complex amplitude distribution; And calculating a wrapped phase distribution diagram based on the argument of the complex amplitude distribution, and performing phase unwrapping treatment on the wrapped phase distribution diagram to obtain the phase distribution diagram.
3. The method of claim 1, wherein for each frame of interference image, contrast extraction is performed to obtain a contrast profile, comprising: performing sliding window traversal according to a preset size on each frame of interference image to obtain a plurality of local window areas; Calculating the ratio of the mean value and the standard deviation of the pixel intensity in each local window area and executing normalization processing to obtain the local contrast corresponding to the local window area; And performing space splicing and interpolation reconstruction on each local contrast to generate the contrast distribution map, wherein the value range of the contrast distribution map is 0 to 1.
4. The method according to claim 1, wherein the method further comprises: performing visual coding processing on the distribution result, displaying the distribution result in parallel on the same display interface and updating the distribution result in real time along with the interference image sequence; And calculating and displaying a key statistical index based on the distribution result, wherein the key statistical index comprises at least one of average temperature, maximum temperature difference, warping degree and reflectivity non-uniformity.
5. The MOCVD process multiparameter synchronous detection system is characterized by comprising the following components: The system comprises an acquisition module, an imaging module, a processing module and a control module, wherein the acquisition module is used for acquiring a reference phase distribution diagram of a process initial stage and an interference image sequence in a process processing stage, the interference image sequence is formed by dividing single-mode laser into object light and reference light, the reference light forms reference waves through point diffraction, and after the object light and the reference waves are spatially overlapped near the surface of a semiconductor to generate interference fringes, the interference fringes are imaged and acquired based on a trigger signal, the trigger signal is generated based on a synchronous pulse signal of a rotary encoder, and a narrow-band filter matched with the center wavelength of the single-mode laser is arranged in an imaging path so as to inhibit heat radiation background light in an epitaxial growth cavity; the information extraction module is used for respectively carrying out phase extraction and contrast extraction on each frame of interference image to obtain a phase distribution map and a contrast distribution map; the synchronous decoupling module is configured to perform multi-parameter synchronous decoupling and inversion processing based on a difference value between the phase distribution diagram and the reference phase distribution diagram, and the contrast distribution diagram, so as to synchronously obtain distribution results of semiconductor morphology-related parameters, temperature-related parameters, and optical reflection-related parameters, and includes: The method comprises the steps of calculating a phase difference distribution diagram of the phase distribution diagram and a reference phase distribution diagram, performing spatial domain component separation processing on the phase difference distribution diagram to obtain a geometric phase component representing the appearance of a semiconductor surface, performing height conversion on the geometric phase component to obtain a height variation distribution of the semiconductor surface, and performing two-dimensional numerical differentiation on the height variation distribution to obtain a curvature distribution; The geometrical phase component is subtracted from the difference value to obtain thermotropic phase distribution, a temperature change field is obtained by inversion based on the thermotropic phase distribution and preset material parameters, and the temperature field distribution is determined based on the temperature change field, wherein the preset material parameters comprise a thermo-optical coefficient, a thermal expansion coefficient, an initial refractive index and a thickness corresponding to a semiconductor material to be detected; and carrying out normalized calibration on the contrast distribution map and standard parameters to obtain relative reflectivity distribution, wherein the standard parameters are obtained based on standard reflector plate measurement.
6. A detection apparatus, characterized by comprising: a light source module for outputting a single-mode laser having a preset wavelength; The point diffraction interference module comprises a point diffraction element and a light source, wherein the point diffraction element is used for dividing the single-mode laser into object light and reference light, diffracting the reference light point into spherical reference waves through the point diffraction element, and enabling the object light and the spherical reference waves to be spatially overlapped near the surface of a semiconductor to form interference fringes; the imaging detection module is used for imaging the interference fringes based on the trigger signal and outputting an interference image sequence, and comprises a narrow-band filter which is arranged in an imaging path and matched with the central wavelength of the single-mode laser, wherein the narrow-band filter is used for inhibiting heat radiation background light in the epitaxial growth cavity; the trigger module is used for generating the trigger signal based on the synchronous pulse signal of the rotary encoder; the MOCVD process multi-parameter synchronous detecting system according to claim 5, wherein the MOCVD process multi-parameter synchronous detecting system is used for performing data processing on the interference image sequence to obtain distribution results of semiconductor morphology-related parameters, temperature-related parameters and optical reflection-related parameters.
7. The apparatus of claim 6, wherein the light source module comprises: The single-mode frequency stabilization laser is used for outputting single-mode laser with preset wavelength; and the optical fiber isolator is used for inhibiting the back-end reflected light from returning to the single-mode frequency stabilization laser.
8. The apparatus of claim 6, wherein the point diffraction interference module further comprises: a transmission fiber for guiding the single-mode laser; the optical fiber coupler is used for dividing the single-mode laser into an object light path and a reference light path according to a preset light splitting ratio; The collimating mirror is used for collimating the emergent light of the object light path and irradiating the emergent light to the surface of the semiconductor; and the three-dimensional precise adjusting frame is used for adjusting the three-dimensional position of the point diffraction element so as to optimize the interference fringe quality.
9. The apparatus of claim 6, wherein the imaging detection module comprises: The imaging lens group is used for imaging the interference area to the target surface of the high-speed area-array camera; the narrow-band filter is arranged at the front end of the imaging lens group.
10. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1-5 when the computer program is executed.
11. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1-5.

Description

MOCVD process multi-parameter synchronous detection method, system, equipment and storage medium Technical Field The application relates to the technical field of semiconductor detection, in particular to a method, a system, equipment and a storage medium for synchronously detecting multiple parameters of an MOCVD process. Background Metal Organic Chemical Vapor Deposition (MOCVD) is a mainstream process equipment and a technical route for preparing second-generation and third-generation compound semiconductor epitaxial materials such as gallium nitride (GaN), gallium arsenide (GaAs) and the like. In the MOCVD epitaxial growth process, spatial uniformity of substrate temperature, surface warpage state caused by thermal stress coupling, and epitaxial layer growth rate and composition uniformity are generally considered key process parameters affecting epitaxial material quality, device performance and production yield. Therefore, the parameters are monitored in situ, in real time and in a full field, and the method has important significance for process development optimization, mechanism model correction, online process control and yield improvement. In the traditional technology, the key process parameters are mainly monitored by mutually independent technologies, such as temperature monitoring by an infrared radiation temperature measurement method, curvature monitoring by a laser beam reflection method or a multi-beam stress meter, reflectivity monitoring by an in-situ laser reflection spectrum or an ellipsometer, and further growth rate and film thickness information are obtained according to the reflectivity. The monitoring mode has inherent limitations under the environment that MOCVD high temperature strong radiation and the wafer surface state changes dynamically. For example, infrared temperature measurement is easily affected by dynamic change of emissivity to cause limited absolute temperature measurement precision, laser curvature is mostly spot measurement or linear measurement, full-field two-dimensional morphology and curvature distribution are difficult to obtain, and reflection spectrum or ellipsometry methods often rely on spectrum scanning or polarization analysis to complicate a system and limit response speed, so that high precision, high timeliness and full-field real-time measurement requirements are difficult to be simultaneously considered. In addition, the key process parameters are respectively measured by a plurality of independent systems, so that the sampling time, the space reference system and the calibration reference of each system are difficult to be strictly aligned, the data time-space dyssynchrony is easy to be caused, the temperature, the curvature and the growth state at the same position and at the same time are difficult to be accurately associated and analyzed, meanwhile, the plurality of systems are integrated to occupy a plurality of observation ports, the complexity of structure, maintenance and calibration is increased, the equipment cost is obviously raised, the process image is incomplete due to the data dimension splitting, the full-field multi-parameter characterization with consistent time-space is difficult to be formed, and the fine process analysis and the closed-loop control capability are further restricted. Disclosure of Invention The application aims to provide a MOCVD process multi-parameter synchronous detection method, a system, equipment and a storage medium, which are used for overcoming the defects that the traditional technology adopts a mutually independent mode to monitor key process parameters, so that each monitoring method is difficult to consider the requirements of precision, response speed and full field due to inherent limitations of a measurement principle, and multi-system data are difficult to be aligned in a consistent way on a sampling time and a space reference, and multi-parameter synchronous correlation analysis cannot be realized. In a first aspect, the present application provides a method for synchronously detecting multiple parameters in an MOCVD process, including: The method comprises the steps of obtaining a reference phase distribution diagram of a process initial stage and an interference image sequence in a process processing stage, wherein the interference image sequence is obtained by dividing single-mode laser into object light and reference light, forming spherical reference waves by point diffraction of the reference light, and collecting the object light and the spherical reference waves after overlapping the space near the surface of a semiconductor to form interference fringes; For each frame of interference image, respectively carrying out phase extraction and contrast extraction on each frame of interference image to obtain a phase distribution map and a contrast distribution map; Based on the difference value between the phase distribution diagram and the reference phase distribution diagram and the contr