Film thickness analysis method

By obtaining the measurement reflectivity matches the simulation spectral library in a multi-layer film system, and using the LM nonlinear regression algorithm to optimize the film thickness value, combined with automatic de-trend processing, the problems of high computational complexity and insufficient accuracy in traditional methods are solved. It realizes efficient and accurate optical characteristic simulation and data matching.

CN118960583AActive Publication Date: 2024-11-15WUHAN YISIPU TECH CO LTD

Patent Information

Application Number
CN202410983884.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-22
Publication Date
2024-11-15
Estimated Expiration
2044-07-22

AI Technical Summary

Technical Problem

Traditional methods have high computational complexity and low efficiency in the calculation of optical characteristics of multilayer film systems, making it difficult to accurately simulate the optical characteristics of the actual system, and it is difficult to flexibly cope with multilayer film systems with different materials and thickness combinations.

Method used

By obtaining the measured light intensity of the sample to be measured, calculating the measured reflectance, and matching it with the theoretical reflectance in the simulation spectrum library, the film thickness value is iteratively optimized using the LM nonlinear regression algorithm, combined with automatic de-trend processing, to realize experimental data and simulation Exact match of data.

Benefits of technology

It improves the efficiency and accuracy of the optical reflection characteristic simulation of multi-layer thin film systems, provides accurate material and structural information, and has stronger adaptability and scalability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118960583A_ABST
    Figure CN118960583A_ABST
Patent Text Reader

Abstract

The invention provides a film thickness analysis method which comprises the following steps: calculating the measured reflectivity of a to-be-measured sample according to the measured light intensity of the to-be-measured sample; matching the measured reflectivity with the theoretical reflectivity in a simulation spectrum library; obtaining the theoretical reflectivity corresponding to the minimum variance, and taking the corresponding thickness value of each layer of film as the film thickness initial value of the sample to be detected; performing iterative optimization on the film thickness initial value of the to-be-measured sample based on an LM nonlinear regression algorithm, and solving an optimal film thickness value of the to-be-measured sample; calculating the correlation between the measured reflectivity and the theoretical reflectivity; if the correlation is smaller than a set threshold value, the measured reflectivity is subjected to de-trending processing, matching is conducted again in the simulation spectrum library, and the film thickness value is obtained. According to the method disclosed by the invention, the optical reflection characteristics of the multilayer film system can be efficiently and accurately simulated by combining spectrograph measurement data and simulation calculation, the optical reflection characteristics are matched with actually measured data, accurate material and structure information is provided, and an important technical support is provided for the field of modern optics and material science.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of optical scattering measurement, and more specifically, to a film thickness analysis method suitable for spectral analysis of complex multilayer thin film systems. Background Art

[0002] In the field of modern optics and materials science, the design and analysis of multilayer thin film systems are of great significance. Multilayer thin films are widely used in optical coatings, semiconductor manufacturing, biosensors and other high-tech fields. The system achieves specific optical functions by controlling the reflection, transmission and absorption characteristics of light. Optical scattering measurement is an analytical technique that uses the scattering phenomenon that occurs when light interacts with matter to obtain physical properties and structural information of the sample. When light encounters the surface of a material or passes through a medium, scattering occurs. The distribution, intensity and polarization state of this scattered light carry important information about the surface properties and internal structure of the material.

[0003] With the development of science and technology, accurate measurement and simulation of the optical properties of multilayer thin film systems have become a research focus. These systems are usually composed of multiple layers of different materials, and the thickness and refractive index of each layer of material have a significant impact on the reflection and transmission characteristics of light. Traditional calculation methods involve a large number of matrix operations and complex optical theories in calculating the optical properties of multilayer thin film systems. The calculation complexity is high and the efficiency is low when dealing with multilayer structures. Due to the complexity of multilayer thin film systems, traditional calculation methods are insufficient in accuracy and it is difficult to accurately simulate the optical properties of actual systems. In addition, existing technologies are difficult to flexibly deal with multilayer thin film systems with different material and thickness combinations, and lack adaptability and scalability. Summary of the invention

[0004] In order to overcome the deficiencies of the film thickness analysis technology in the prior art, the present invention provides a film thickness analysis method in optical scattering measurement, which effectively solves the problem of large error in solving parameters due to large difference in data scales.

[0005] The present invention provides a film thickness analysis method, comprising:

[0006] Step S1, obtaining the measured light intensity of the sample to be tested, and calculating the measured reflectivity of the sample to be tested according to the measured light intensity;

[0007] Step S2, calculating the variance between the measured reflectivity and each theoretical reflectivity in a simulation spectrum library, wherein the simulation spectrum library constructs a corresponding relationship between the thickness value of each layer of a multilayer thin film system and the theoretical reflectivity;

[0008] Step S3, obtaining the theoretical reflectivity corresponding to the minimum variance, and taking the corresponding thickness value of each layer of the thin film as the initial value of the film thickness of the sample to be tested;

[0009] Step S4, iteratively optimizing the initial value of the film thickness of the sample to be tested based on the LM nonlinear regression algorithm to solve the optimal film thickness value of the sample to be tested;

[0010] Step S5, calculating the correlation between the measured reflectivity and the theoretical reflectivity corresponding to the optimal film thickness value;

[0011] Step S6, if the correlation is less than the set threshold, detrending the measured reflectance, obtaining the measured spectrum after detrending, and returning to step S2; if the correlation is greater than or equal to the set threshold, the process ends.

[0012] The present invention provides a film thickness analysis method, which calculates the measured reflectivity of the sample to be tested according to the measured light intensity of the sample to be tested; matches the measured reflectivity with the theoretical reflectivity in the simulation spectrum library; obtains the theoretical reflectivity corresponding to the minimum variance, and uses the corresponding thickness value of each layer of the film as the initial value of the film thickness of the sample to be tested; iteratively optimizes the initial value of the film thickness of the sample to be tested based on the LM nonlinear regression algorithm to solve the optimal film thickness value of the sample to be tested; calculates the correlation between the measured reflectivity and the theoretical reflectivity; if the correlation is less than the set threshold, the measured reflectivity is detrended and re-matched in the simulation spectrum library to obtain the film thickness value. The method of the present invention can efficiently and accurately simulate the optical reflection characteristics of a multilayer film system by combining spectrometer measurement data and simulation calculations, and matches with measured data, providing accurate material and structural information, and providing important technical support for the fields of modern optics and materials science. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 A flow chart of a film thickness analysis method provided by the present invention;

[0014] Figure 2 It is a schematic diagram of the overall process of a film thickness analysis method;

[0015] Figure 3 Schematic diagram of the effect of using LM iteration to solve the fitted reflectivity;

[0016] Figure 4-1 This is a schematic diagram of the fitting effect when the downward trend algorithm is not used;

[0017] Figure 4-2 This is a schematic diagram of the fitting effect when using the descending trend algorithm. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. In addition, the technical features in the various embodiments or single embodiments provided by the present invention can be arbitrarily combined with each other to form a feasible technical solution. This combination is not subject to the constraints of the sequence of steps and / or the structural composition mode, but must be based on the ability of ordinary technicians in this field to achieve. When the combination of technical solutions is contradictory or cannot be achieved, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0019] In order to solve the problems raised in the background technology, the present invention proposes a film thickness analysis method applied to optical scattering measurement. This method can efficiently and accurately simulate the optical reflection characteristics of multilayer thin film systems by combining spectrometer measurement data and simulation calculations, and match them with measured data to provide accurate material and structural information, providing important technical support for the fields of modern optics and materials science.

[0020] See also Figure 1 , is a flow chart of a film thickness analysis method provided by the present invention, the method mainly comprising:

[0021] Step S1, obtaining the measured light intensity of the sample to be tested, and calculating the measured reflectivity of the sample to be tested according to the measured light intensity.

[0022] Wherein, the step S1, obtaining the measured light intensity of the sample to be tested, and calculating the measured reflectivity of the sample to be tested according to the measured light intensity, includes:

[0023] The spectrometer collects the measured light intensity of the sample to be tested at N measuring wavelength points, and calculates the corresponding measured reflectivity according to the measured light intensity at each measuring wavelength point:

[0024]

[0025] The measured reflectance is normalized:

[0026]

[0027] Among them, I sample,i It represents the measured light intensity of the sample to be tested at the i-th measurement wavelength point, I dark,i It represents the measured light intensity at the i-th measurement wavelength point when there is no light irradiating the sample to be measured. It is used to eliminate the influence of the instrument's own noise. std,irepresents the reflectance of a known standard sample at the i-th measurement wavelength point, Indicates the minimum measured reflectance among all measured wavelength points, Indicates the maximum measured reflectance among all measured wavelength points, The reflectance is measured at the normalized i-th measurement wavelength point.

[0028] It is understood that light is incident from the air at a certain angle to the surface of the film. This incident light is divided into two parts when it reaches the surface of the film: one part is reflected back into the air, and the other part enters the film. When the incident light reaches the surface of the film, part of the light is reflected. The direction of this part of the reflected light is symmetrical to the incident light, forming the first reflected light. The light entering the film propagates inside the film and undergoes multiple reflections between the bottom surface and the top of the film. At each reflection, part of the light is transmitted to the substrate to form transmitted light. Transmitted light is light that finally penetrates the film and enters the substrate after multiple reflections from the film. These light rays undergo multiple reflections and interferences in the process of penetrating the film, resulting in changes in their phase and intensity. The spectral band, incident angle, morphology and other data measured by the spectrometer can be used to accurately describe and simulate the propagation process of this light in a multilayer film system.

[0029] Therefore, the measured spectral band and light intensity can be given by the spectrometer, that is, the light intensity information (called measured light intensity) of multiple measurement wavelength points (for example, N measurement wavelength points of the spectral band) is collected by the spectrometer, and the measured reflectivity of each measurement wavelength point is calculated according to the above-mentioned formula for calculating the measured reflectivity, and the measured reflectivity of each measurement wavelength point is normalized using the above-mentioned formula for calculating the theoretical reflectivity.

[0030] Step S2, calculating the variance between the measured reflectivity and each theoretical reflectivity in the simulation spectrum library, wherein the simulation spectrum library constructs a corresponding relationship between the thickness value of each layer of the multilayer thin film system and the theoretical reflectivity.

[0031] Among them, according to the input number of thin film layers and the thickness range of each thin film layer, multiple possible thickness values ​​of each thin film layer are uniformly generated; the thickness value of each thin film layer is taken according to the possible multiple thickness values ​​of each thin film layer to form different thin film structure combinations; for each thin film structure combination, the optical theoretical reflectivity is calculated, and the corresponding relationship between the thickness value of each layer of the multilayer thin film system and the theoretical reflectivity is generated to construct a simulation spectrum library.

[0032] It is understandable that, according to the parameters input by the user (such as the number of film layers, the thickness range of each layer, etc.), the possible thickness values ​​of each film layer are uniformly generated, that is, there are multiple values ​​of the thickness value of each film layer, so different film structures can be formed by taking different thickness values ​​of each film layer. The optical theoretical reflectivity of each film structure is calculated, and a simulation spectrum library is constructed. The corresponding relationship between the thickness value of each film layer (hereinafter referred to as thickness value) and the theoretical reflectivity is stored in the simulation spectrum library.

[0033] Among them, the method for calculating the theoretical reflectivity according to the thickness value of the thin film structure is as follows:

[0034] Define the possibility of different film structure combinations and calculate the optical theoretical reflectivity for each thickness combination. Define the characteristic matrix D of the kth layer of the multilayer film system at each measurement wavelength point k And the phase change matrix P k , the characteristic matrix describes the behavior of light at the interface, while the phase change matrix describes the phase change of light as it propagates within the layer.

[0035] Among them, for p-polarized light, the characteristic matrix D of the k-th film is k,p It is expressed as:

[0036]

[0037] For s-polarized light, the characteristic matrix D of the kth film k,s It is expressed as:

[0038]

[0039] Phase change matrix P k It is expressed as:

[0040]

[0041] Phase change angle β k for:

[0042]

[0043] Among them, θ k is the incident angle of the kth film, n k is the refractive index of the kth film layer, d k is the thickness of the kth film layer, and λ is the measurement wavelength.

[0044] According to the characteristic matrix D of each film k and the phase change matrix P k , calculate the transmission matrix M of each film structure combination. The transmission matrix M of the entire multilayer film structure can be expressed as the product of the characteristic matrix of each layer and the phase change matrix:

[0045]

[0046] in, is the inverse matrix of the characteristic matrix of the first film, is the inverse matrix of the characteristic matrix of the kth film, and j is the number of film layers.

[0047] According to the transmission matrix M, the reflection coefficient r is calculated:

[0048]

[0049] Among them, M 21 and M 11 are the elements of the transmission matrix M.

[0050] According to the reflection coefficient r of p-polarized light and s-polarized light p and r s , the total reflectivity R (theoretical reflectivity) of each film structure combination is:

[0051]

[0052] The theoretical reflectivity R is normalized to:

[0053]

[0054] in, Represents the minimum theoretical reflectivity among all measured wavelength points, Represents the maximum theoretical reflectivity among all measured wavelength points, is the normalized theoretical reflectivity.

[0055] The above method can be used to calculate the theoretical reflectivity of the thin film structure of each thickness value at different measurement wavelength points, and the corresponding relationship between the thickness value and the theoretical reflectivity can be stored in the simulation spectrum library.

[0056] The measured reflectivity calculated in step S1 is matched with the theoretical reflectivity in the simulation spectrum library. During the matching process, the variance between the measured reflectivity and each theoretical reflectivity in the simulation spectrum library can be calculated. The variance can be expressed as:

[0057]

[0058] in, represents the measured reflectance at the i-th measurement wavelength point, represents the theoretical reflectivity of the i-th measurement wavelength point, and N is the number of measurement wavelength points.

[0059] Step S3, obtaining the theoretical reflectivity corresponding to the minimum variance, and taking the corresponding thickness value of each thin film layer as the initial value of the film thickness of the sample to be tested.

[0060] It can be understood that, according to the minimum variance, the theoretical reflectivity that is most similar to the measured reflectivity is determined, and the thickness corresponding to the theoretical reflectivity with the minimum variance is considered to be the initial thickness value d of the sample to be measured. 0

[0061] Step S4, iteratively optimizing the initial value of the film thickness of the sample to be tested based on the LM nonlinear regression algorithm to solve the optimal film thickness value of the sample to be tested.

[0062] It can be understood that according to the initial thickness value d obtained in the previous step 0 , use nonlinear regression algorithm to iteratively find the optimal solution.

[0063] In the process of measuring film thickness, the Levenberg-Marquardt (LM) algorithm can be used to effectively optimize the calculation of film thickness by combining the single-layer film structure model. The following are detailed steps and formula derivation:

[0064] First, the initial film thickness value d of the sample to be tested is obtained in step S3. 0 , and then calculate the measured reflectivity and initial film thickness d 0 The mean square error MSE between the corresponding theoretical reflectivities is:

[0065]

[0066] Where N is the number of measurement wavelength points, is the measured reflectance at the ith measurement wavelength point, is the initial value of film thickness d 0 The corresponding theoretical reflectivity at the i-th measurement wavelength point.

[0067] Measure the initial reflectivity and film thickness d 0 The optimization goal is to minimize the mean square error MSE between the corresponding theoretical reflectivities and solve the film thickness update step δd 1 ;

[0068] Update the step size δd based on the film thickness 1 For the initial value of the film thickness d 0 Perform the first update and get d 1 =d 0 +δd 1 , based on the updated film thickness value d 1 , recalculate the measured reflectivity and updated film thickness value d 1 Mean square error MSE between the corresponding theoretical reflectivities;

[0069] The thickness update step length δd of the second update is calculated based on the mean square error MSE after the first update. 2 , for the initial value of film thickness d0 Perform the second update and get d 2 =d 1 +δd 2 , by continuously updating the initial value of the film thickness until the calculated mean square error MSE is less than or equal to the set deviation threshold, the optimal film thickness value is obtained.

[0070] Among them, the method for solving the film thickness update step δd includes:

[0071] Theoretical reflectivity R theory (d 1 ) is approximated by the first-order Taylor expansion:

[0072]

[0073] Among them, δd 1 Indicates a small increment in film thickness value, is the initial value of the film thickness d 0 The partial derivative at ;

[0074] Taking the minimization of the mean square error (MSE) between the measured reflectivity and the theoretical reflectivity as the optimization goal, we solve δd 1 Minimize the following relationship:

[0075]

[0076] In order to minimize the above relationship, δd needs to satisfy the following relationship equation:

[0077]

[0078] Among them, μ>0 is the damping factor, ε is the variance between the measured reflectivity and the theoretical reflectivity; by solving the above relationship equation, the film thickness update step δd is obtained 1 :

[0079]

[0080] Initial value of film thickness d 0 Each time the update is performed, the corresponding film thickness update step δd is calculated. h ,δd h Indicates the initial value of film thickness d 0 The film thickness update step size when performing the nth update.

[0081] According to each updated film thickness value, the corresponding theoretical reflectivity is calculated based on the method of calculating the theoretical reflectivity in step S2, and then the mean square error MSE between the measured reflectivity and the theoretical reflectivity is calculated to determine whether the mean square error MSE is less than the set error threshold. If so, the iterative update process is terminated to obtain the updated film thickness value. If the mean square error MSE is still greater than the set error threshold, then at the film thickness value d 1 The film thickness value is updated again based on the value obtained by continuously updating the film thickness value until the calculated mean square error MSE is less than the set error threshold. The film thickness value at this time is obtained, which is the optimal film thickness value of the sample to be tested. Figure 3 This is a schematic diagram of the effect of using LM iteration to solve the fitted reflectance, where the red curve represents the graph of measured reflectance and wavelength, the blue curve represents the graph of theoretical reflectance and wavelength, the horizontal axis is wavelength, and the vertical axis is reflectance.

[0082] Step S5, calculating the correlation between the measured reflectivity and the theoretical reflectivity corresponding to the optimal film thickness value.

[0083] It is understandable that when the thickness of a single film layer in a thin film structure is large, simply normalizing the measured reflectivity can sometimes produce poor fitting results. To address this situation, an automatic downtrend algorithm for the measured reflectivity is added based on the previous algorithm steps to improve the fitting effect. After the theoretical reflectivity corresponding to the measured reflectivity and the optimal film thickness value is calculated in the above steps, the measured reflectivity R is calculated. meas And the theoretical reflectivity R theory The correlation gof is used to determine whether detrending is needed. Among them, the measured reflectivity R meas And the theoretical reflectivity R theory The correlation gof can be calculated as follows:

[0084] a. Calculate the mean:

[0085] First, calculate the average of the measured reflectivity and the theoretical reflectivity using the following formula:

[0086]

[0087] Where N is the number of measurement wavelength points, R meas,i represents the measured reflectance at the i-th measurement wavelength point, R theory,i represents the theoretical reflectivity at the i-th measurement wavelength point, S meas Represents the mean value of the measured reflectivity, S theory Represents the mean value of theoretical reflectivity.

[0088] b. Based on the average of the measured reflectivity and the theoretical reflectivity, the measured reflectivity and the theoretical reflectivity are respectively centralized:

[0089] R′ meas,i =R meas,i -S meas , R′ theory,i =R theory,i -S theory ;

[0090] Among them, R′ meas,i represents the decentralized measured reflectivity at the i-th measurement wavelength point, R′ theory,i represents the decentralized theoretical reflectivity at the i-th measurement wavelength point.

[0091] c. Calculate the mean of the measured reflectivity after centering d meas and the mean value of theoretical reflectivity d theory :

[0092]

[0093] d. Calculate R' meas The variance of r and R' theory The variance f is:

[0094]

[0095] e. Calculate the covariance rf between the measured reflectivity and the theoretical reflectivity:

[0096]

[0097] f. Calculate the correlation gof between the measured reflectivity and the theoretical reflectivity and standardize it to [0,1]:

[0098]

[0099] Step S6, if the correlation is less than the set threshold, detrending the measured reflectance, obtaining the measured spectrum after detrending, and returning to step S2; if the correlation is greater than or equal to the set threshold, the process ends.

[0100] It is understandable that if the correlation between the measured reflectivity and the theoretical reflectivity is less than the set threshold, the measured reflectivity needs to be detrended. The method for detrending the measured reflectivity is:

[0101] According to the measured reflectance at multiple measurement wavelength points, the reflectance is measured based on a third-order polynomial fitting, and the third-order polynomial fitting formula is:

[0102] R meas (λ) = a 3 λ 3 +a 2 λ 2 +a1 λ+a 0 ;

[0103] Among them, a=[a 3 ,a 2 ,a 1 ,a 0 ] T are the polynomial coefficients;

[0104] Establishing a fitting objective function based on the least squares method, wherein the objective function is to minimize the sum of squares of errors between the measured reflectivity and the measured reflectivity after fitting;

[0105]

[0106] Solve the objective function and get a=[a 3 ,a 2 ,a 1 ,a 0 ] T .

[0107] Calculate the fitted reflectance at each measured wavelength point:

[0108] R fit,i =(a 3 λ i 3 +a 2 λ i 2 +a 1 λ i +a 0 )) 2 ;

[0109] Detrend the measured reflectance:

[0110] R detrend,i =R meas,i -R fit,i ;

[0111] Among them, R detrend,i is the measured reflectance after detrending at the ith measurement wavelength point, R meas,i is the measured reflectivity at the i-th measurement wavelength point.

[0112] Based on the measured reflectivity after detrending, step S2 is executed, that is, matching is performed in the simulation spectrum library, and the optimal film thickness value of the sample to be measured is solved by executing steps S2 to S4 again.

[0113] in, Figure 4-1 This is a schematic diagram of the effect of LM iteration solution fitting reflectivity without using the descending trend algorithm. Figure 4-2The figure is a schematic diagram of the effect of LM iteration using the descending trend algorithm to solve the fitted reflectance. The red curve represents the graph of measured reflectance and wavelength, and the blue curve represents the graph of theoretical reflectance and wavelength. The horizontal axis is wavelength and the vertical axis is reflectance. Figure 4-1 and Figure 4-2 It can be seen that the LM iterative solution using the descending trend algorithm has a significantly better effect on fitting reflectivity.

[0114] The film thickness analysis method provided by the present invention has the following beneficial effects compared with the prior art:

[0115] (1) In optical scattering measurement, large differences in data scales often lead to large errors in solution parameters. The present invention can effectively reduce the impact of data scale differences on solution results through normalization processing, thereby improving matching accuracy and reliability.

[0116] (2) Traditional methods make it difficult to efficiently and accurately match experimental measurement data with simulation calculation data. The present invention can quickly find the simulated reflectance most similar to the data to be measured by constructing a simulated spectrum library, combining the LM algorithm and the automatic detrending algorithm, and achieve accurate matching of experimental data with simulation data.

[0117] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0118] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0119] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A film thickness analysis method, characterized in that: include: Step S1, obtaining the measured light intensity of the sample to be tested, and calculating the measured reflectivity of the sample to be tested according to the measured light intensity; Step S2, calculating the variance between the measured reflectivity and each theoretical reflectivity in a simulation spectrum library, wherein the simulation spectrum library constructs a corresponding relationship between the thickness value of each layer of a multilayer thin film system and the theoretical reflectivity; Step S3, obtaining the theoretical reflectivity corresponding to the minimum variance, and taking the corresponding thickness value of each layer of the thin film as the initial value of the film thickness of the sample to be tested; Step S4, iteratively optimizing the initial value of the film thickness of the sample to be tested based on the LM nonlinear regression algorithm to solve the optimal film thickness value of the sample to be tested; Step S5, calculating the correlation between the measured reflectivity and the theoretical reflectivity corresponding to the optimal film thickness value; Step S6, if the correlation is less than the set threshold, detrending the measured reflectance, obtaining the measured spectrum after detrending, and returning to step S2; if the correlation is greater than or equal to the set threshold, the process ends.

2. The film thickness analysis method according to claim 1, characterized in that: The step S1, obtaining the measured light intensity of the sample to be tested, and calculating the measured reflectivity of the sample to be tested according to the measured light intensity, includes: The spectrometer collects the measured light intensity of the sample to be tested at N measuring wavelength points, and calculates the corresponding measured reflectivity according to the measured light intensity at each measuring wavelength point: The measured reflectance is normalized: Among them, I sample,i It represents the measured light intensity of the sample to be tested at the i-th measurement wavelength point, I dark,i It represents the measured light intensity at the i-th measurement wavelength point when there is no light irradiating the sample to be measured, R std,i represents the reflectance of a known standard sample at the i-th measurement wavelength point, Indicates the minimum measured reflectivity among all measured wavelength points, Indicates the maximum measured reflectance among all measured wavelength points, The reflectance is measured at the normalized i-th measurement wavelength point.

3. The film thickness analysis method according to claim 1, characterized in that: Build a simulated spectral library, including: According to the input number of film layers and the thickness range of each film layer, multiple possible thickness values ​​of each film layer are uniformly generated; The thickness value of each film layer is determined according to the possible multiple thickness values ​​of each film layer to form different film structure combinations; For each combination of thin film structures, the theoretical optical reflectivity is calculated, the corresponding relationship between the thickness value of each layer of the multilayer thin film system and the theoretical reflectivity is generated, and a simulation spectrum library is constructed.

4. The film thickness analysis method according to claim 3, characterized in that: For each combination of film structures, the optical theoretical reflectivity is calculated, including: Define the characteristic matrix Dk and phase change matrix Pk of the k-th layer of the multilayer film system at each measurement wavelength point, where for p-polarized light: For s-polarized light: Phase change matrix P k : Phase change angle β k for: Among them, θ k is the incident angle of the kth film, n k is the refractive index of the kth film layer, d k is the thickness of the kth film layer, and λ is the measurement wavelength; According to the characteristic matrix D of each film k And the phase change matrix P k , calculate the transmission matrix M of each film structure combination: in, is the inverse matrix of the characteristic matrix of the first film, is the inverse matrix of the characteristic matrix of the k-th film, j is the number of film layers; According to the transmission matrix M, the reflection coefficient r is calculated: Among them, M 21 and M 11 are the elements of the transmission matrix M; According to the reflection coefficient r of p-polarized light and s-polarized light p and r s , the theoretical reflectivity R of each film structure combination is: The theoretical reflectivity R is normalized to: in, Represents the minimum theoretical reflectivity among all measured wavelength points, Represents the maximum theoretical reflectivity among all measured wavelength points, is the normalized theoretical reflectivity; According to the thickness value of each film layer, the theoretical reflectivity at each measuring wavelength point is calculated.

5. The film thickness analysis method according to claim 1, characterized in that: The step S2, calculating the variance between the measured reflectivity and each theoretical reflectivity in the simulation spectrum library, comprises: in, represents the measured reflectance at the i-th measurement wavelength point, represents the theoretical reflectivity of the i-th measurement wavelength point, and N is the number of measurement wavelength points.

6. The film thickness analysis method according to claim 1, characterized in that: The step S4, iteratively optimizing the initial value of the film thickness of the sample to be tested based on the LM nonlinear regression algorithm to solve the optimal film thickness value of the sample to be tested, includes: Calculate the mean square error (MSE) between the measured reflectivity and the theoretical reflectivity corresponding to the initial value of the film thickness: Where N is the number of measurement wavelength points, is the measured reflectance at the ith measurement wavelength point, is the theoretical reflectivity at the i-th measurement wavelength point corresponding to the initial value of film thickness d0; Taking the minimization of the mean square error MSE between the measured reflectivity and the theoretical reflectivity corresponding to the initial value of the film thickness d0 as the optimization goal, the film thickness update step δd1 is solved; The initial film thickness value d0 is updated for the first time based on the film thickness update step δd1 to obtain d1=d0+δd1, and based on the updated film thickness value d1, the mean square error MSE between the measured reflectivity and the theoretical reflectivity corresponding to the updated film thickness value d1 is calculated again; The second updated film thickness update step δd2 is calculated based on the mean square error MSE after the first update, and the initial film thickness value d0 is updated for the second time to obtain d2=d1+δd2. The initial film thickness value is continuously updated until the calculated mean square error MSE is less than or equal to the set deviation threshold to obtain the optimal film thickness value.

7. The film thickness analysis method according to claim 6, characterized in that: The method takes minimization of the mean square error MSE between the measured reflectivity and the theoretical reflectivity corresponding to the initial value d0 of the film thickness as the optimization goal, and solves the film thickness update step δd1, including: Theoretical reflectivity R theory (d1) is approximated by the first-order Taylor expansion: Where δd1 represents a small increment in film thickness. is the partial derivative at the initial value of film thickness d0; Taking the minimization of the mean square error MSE between the measured reflectivity and the theoretical reflectivity as the optimization goal, δd1 is solved to minimize the following relationship: In order to minimize the above relationship, δd needs to satisfy the following relationship equation: Among them, μ>0 is the damping factor, ε is the variance between the measured reflectivity and the theoretical reflectivity; By solving the above relationship equation, the film thickness update step δd1 is obtained: Each time the initial value of the film thickness d0 is updated, the corresponding film thickness update step δd is calculated h ,δd h Indicates the film thickness update step size when the initial film thickness value d0 is updated for the nth time.

8. The film thickness analysis method according to claim 1, characterized in that: Step S5, calculating the correlation between the measured reflectivity and the theoretical reflectivity corresponding to the optimal film thickness value, comprising: Calculate the mean of the measured reflectivity and the mean of the theoretical reflectivity corresponding to the optimal film thickness value: Where N represents the number of measurement wavelength points, R meas,i represents the measured reflectance at the i-th measurement wavelength point, R theory,i represents the theoretical reflectivity at the i-th measurement wavelength point; The measured reflectivity and theoretical reflectivity are centered based on the mean of the measured reflectivity and the mean of the theoretical reflectivity, respectively. The formula is as follows: R′ meas,i =R meas,i -S meas ,R′ theory,i =R theory,i -S theory ; Calculate the mean of the measured reflectivity after decentralization and the mean of the theoretical reflectivity corresponding to the optimal film thickness value: Calculate the variance r of R'meas and the variance f of Rt'heory: Compute the covariance: Calculate the correlation gof and standardize it to [0,1]:

9. The film thickness analysis method according to claim 1, characterized in that: The step S6, if the correlation is less than a set threshold, detrending the measured reflectivity to obtain a detrended measured spectrum, comprises: According to the measured reflectance at multiple measurement wavelength points, the measured reflectance is fitted based on a third-order polynomial. The third-order polynomial fitting formula is: R meas (λ)=a3λ 3 +a2λ 2 +a1λ+a0; Where a=[a3,a2,a1,a0] T are the polynomial coefficients; Establishing a fitting objective function based on the least squares method, wherein the objective function is to minimize the sum of squares of errors between the measured reflectivity and the measured reflectivity after fitting; Solve the objective function and get a=[a3,a2,a1,a0] T ; Calculate the fitted reflectance at each measured wavelength point: R fit,i =(a3λ i 3 +a2λ i 2 +a1λ i +a0)) 2 ; Detrend the measured reflectance: R detrend,i =R meas,i -R fit,i ; Among them, R detrend,i is the measured reflectance after detrending at the ith measurement wavelength point, R meas,i is the measured reflectivity at the i-th measurement wavelength point.

Citation Information

Patent Citations

  • Method and device for measuring thickness of semiconductor film

    CN101865641A

  • Back compensation-based transparent substrate film thickness measurement system

    CN102243065A

  • Partial reflection multi-layer optical thin film

    CN105652358A

  • Thin film measuring method

    CN111426275A

  • Optical characteristic modeling method, photoacoustic measurement method and device

    CN114543690A

Cited By

  • Filter film thickness detection device and detection method

    CN120141319A

  • A filter film thickness detection device and a detection method

    CN120141319B

  • Intelligent wide-spectrum film thickness real-time monitoring system and method

    CN120778013A

  • Intelligent wide-spectrum film thickness real-time monitoring system and method

    CN120778013B

  • Method and device for detecting thickness of thin film

    CN121163387A