A method for accurate film thickness measurement for large-scale spin coating
Through adaptive analysis and optimization of filter window length, the problem of noise interference in traditional film thickness measurement is solved, the film thickness measurement accuracy during large-size spin coating is improved, and more efficient film thickness measurement is achieved.
Patent Information
- Application Number
- CN202410993937.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-24
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-07-24
AI Technical Summary
Traditional film thickness measurement methods are not effective due to environmental and equipment noise during large-size spin coating, which reduces the film thickness measurement accuracy, especially ignores irregular noise counter-return caused by the central stacking effect during semiconductor spin coating. influence of wave signal.
Adaptive analysis method is used to divide the echo signal into different waveform intervals, and the interference degree is adaptively constructed based on the changes in the echo signal within the waveform interval, optimize the waveform interference degree, adaptively adjust the filter window length, and pass the optimal filter window to match the echo. The signal is filtered to improve the film thickness measurement accuracy.
Through adaptive analysis and optimization of the filter window length, the noise interference generated during spin coating is effectively removed, the accuracy of film thickness measurement is improved, the original signal characteristics are retained, and the accuracy of film thickness measurement is enhanced.
Smart Images

Figure CN118518042B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of coating material thickness measurement, and in particular to a method for accurately measuring the film thickness for large-size spin coating. Background Art
[0002] Photolithography technology is widely used in aerospace, optical engineering, biomedicine and other fields. Among them, the preparation of photoresist film is a key step to ensure the performance of the final device as the front-end process of photolithography technology. Spin coating is a thin film preparation process with a wide range of film thickness, simple operation and high efficiency, which is widely used in photoresist film formation. The typical spin coating process is mainly divided into three steps: glue dripping, spin coating and solvent volatilization. First, the spin coating liquid is dropped onto the surface of the substrate, and then the droplets are spread out by high-speed rotation to form a uniform film, and then the remaining solvent is dried and evaporated to obtain a film with stable performance. The thickness of the photoresist film directly affects the formation and quality of the photolithography pattern. It is usually necessary to measure the thickness of the photoresist film to promptly discover and correct problems that may affect product quality.
[0003] Traditional film thickness measurement is often performed by ultrasonic wave. However, ultrasonic wave measurement of film thickness may be affected by various noises from the environment, equipment or the material itself. Therefore, when measuring film thickness, a filtering algorithm is usually used to denoise the echo signal. The existing technology directly filters the echo signal through a preset filtering window, ignoring the influence of irregular noise caused by the center accumulation effect generated during the semiconductor spin coating process on the echo signal fluctuation, resulting in poor filtering effect, which in turn leads to low accuracy of film thickness measurement. Summary of the invention
[0004] In order to solve the above technical problems, the present application provides a method for accurately measuring the film thickness of large-size spin coating to solve the existing problems.
[0005] A method for accurately measuring the film thickness of a large-size spin coating in this application adopts the following technical solution:
[0006] An embodiment of the present application provides a method for accurately measuring the film thickness of large-size spin coating, the method comprising the following steps:
[0007] S10, collecting an excitation signal applied to the ultrasonic transducer and an echo signal at each sampling point on the surface of the semiconductor substrate;
[0008] S20, dividing the echo signal into waveform intervals according to the excitation signal and the distribution of zero points in the echo signal at each sampling point;
[0009] S30, obtaining the center point of each waveform interval, and obtaining the waveform interference degree according to the difference between the echo signal points symmetrical about the center point;
[0010] S40, optimizing the waveform interference degree according to the difference between the waveform intervals at the sampling points at different positions; the optimization process includes:
[0011] S41, obtaining a film thickness attenuation characteristic value according to the echo signals of all sampling points in each sampling point sequence obtained by arranging all sampling points;
[0012] S42, obtaining each concentric circle, and obtaining the film thickness measurement concentricity according to the film thickness attenuation characteristic value of the sampling point sequence to which the sampling points located on the same concentric circle belong;
[0013] S43, optimizing the waveform interference degree of each waveform interval by combining the film thickness attenuation characteristic value and the film thickness measurement concentricity to obtain an optimized waveform interval interference degree;
[0014] S50, obtaining the optimal filtering window length of each waveform interval according to the optimized waveform interval interference degree;
[0015] S60, filtering the echo signal at each sampling point according to the optimal filter window length to obtain a filtered echo signal; obtaining the film thickness of the semiconductor substrate coating according to the filtered echo signals at all sampling points and the waveform interval interference degree;
[0016] The step of obtaining the optimal filtering window length of each waveform interval according to the optimized waveform interval interference degree comprises:
[0017] Obtaining a filter window adjustment factor for each waveform interval according to the optimized waveform interval interference degree;
[0018] For each adjustment, the product of the window length of the previous adjustment and the filter window adjustment factor corresponding to the window length of the previous adjustment is calculated, and recorded as the second product; the rounded-up value of the second product is used as the estimated length increment of the window length of each adjustment;
[0019] The twice of the estimated length increment of the window after each adjustment is used as the length increment of the window length adjusted each time; the window length adjusted each time is the sum of the window length after the previous adjustment and the length increment;
[0020] The window length adjustment process is repeated until the filter window adjustment factor is less than the first preset threshold or the number of adjustments reaches the second preset threshold; the window length obtained by the last adjustment is used as the optimal filter window length.
[0021] Furthermore, the echo signal is divided into waveform intervals according to the distribution of zero points in the excitation signal and the echo signal at each sampling point, including:
[0022] Obtain each zero point in the echo signal, calculate the absolute value of the difference between each zero point and the acquisition time corresponding to the next zero point, and record it as the time sequence distance between each zero point and the next zero point;
[0023] Obtaining the standard waveform signal interval length of the excitation signal, calculating the absolute value of the difference between the standard waveform signal interval length and the timing distance, and recording it as the first absolute value of the difference; the timing distance difference of each zero point is positively correlated with the first absolute value of the difference;
[0024] The average value of the timing distance difference of all zero points in the echo signal is taken as the interference threshold; the zero point whose timing distance difference is less than or equal to the interference threshold is recorded as the timing zero point; the echo signal contained in the closed interval between any two adjacent timing zero points is recorded as a waveform interval.
[0025] Further, the step of obtaining the center point of each waveform interval and obtaining the waveform interference degree according to the difference between echo signal points symmetrical about the center point includes:
[0026] Record the echo signal point located at the center of the waveform interval as the center point; obtain the echo signal point that is symmetrical about the center point of each echo signal point in the waveform interval, and record it as the central symmetrical point of each echo signal point;
[0027] Obtain the minimum absolute value of the difference between the signal strength values of each echo signal point and other echo signal points in the waveform interval, and record it as the adjacent signal strength difference value of each echo signal point;
[0028] For each echo signal point in the waveform interval, a preset initial filtering window centered on the echo signal point is matched one by one with the echo signal points in the initial filtering window centered on the central symmetric point to obtain each group of corresponding points; the absolute value of the difference between the signal strength values of two echo signal points in each group of corresponding points is calculated, recorded as the strength corresponding difference value; the absolute value of the difference between the adjacent signal strength difference values of two echo signal points in each group of corresponding points is calculated, recorded as the adjacent corresponding difference value;
[0029] The intensity corresponding difference values and the adjacent corresponding difference values of all groups of corresponding points are positively correlated with the deviation degree of the echo signal point;
[0030] The sum of the deviations of all echo signal points within the waveform interval is taken as the waveform interference degree of the waveform interval.
[0031] Further, the obtaining of the film thickness attenuation characteristic value comprises:
[0032] The maximum absolute value of the signal strength values of all echo signal points within the waveform interval is taken as the amplitude of the waveform interval;
[0033] The amplitude of the first waveform interval in the echo signal at each sampling point is used as the characteristic signal amplitude of each sampling point;
[0034] The sampling points on the line starting from the center point of the semiconductor substrate surface and ending at the sampling points on the edge of the semiconductor substrate surface form a sampling point sequence;
[0035] The sequence composed of the characteristic signal amplitudes of all sampling points in each sampling point sequence is recorded as the characteristic signal amplitude sequence of each sampling point sequence;
[0036] Calculate the correlation between each sampling point sequence and the characteristic signal amplitude sequence of other sampling point sequences; obtain the first-order difference sequence of the characteristic signal amplitude sequence of each sampling point sequence; calculate the mean of all elements in the first-order difference sequence, recorded as the attenuation amplitude of each sampling point sequence; calculate the absolute value of the difference between the attenuation amplitude of each sampling point sequence and other sampling point sequences, recorded as the attenuation difference between each sampling point sequence and other sampling point sequences;
[0037] The film thickness attenuation characteristic value of each sampling point sequence is positively correlated with the correlation and negatively correlated with the attenuation difference.
[0038] Further, the obtaining of each concentric circle and obtaining the film thickness measurement concentricity according to the film thickness attenuation characteristic value of the sampling point sequence to which the sampling points located on the same concentric circle belong include:
[0039] Obtain concentric circles with the center point of the semiconductor substrate surface as the center and the Euclidean distance from each sampling point to the center point of the semiconductor substrate surface as the radius; record the sampling points located on the same concentric circle as each sampling point as concentric sampling points of each sampling point;
[0040] The absolute value of the difference between the characteristic signal amplitude of each sampling point and its concentric sampling point is calculated, which is recorded as the concentric characteristic difference between each sampling point and its concentric sampling point; the correlation between the echo signal of each sampling point and its concentric sampling point is calculated, which is recorded as the signal similarity between each sampling point and its concentric sampling point;
[0041] The film thickness attenuation characteristic value of the sampling point sequence to which each sampling point belongs and the signal similarity are both positively correlated with the film thickness measurement concentricity of each sampling point, and the concentricity characteristic difference is negatively correlated with the film thickness measurement concentricity of each sampling point.
[0042] Further, the waveform interference degree of each waveform interval is optimized by combining the film thickness attenuation characteristic value with the film thickness measurement concentricity to obtain the optimized waveform interval interference degree, including:
[0043] Calculate the product of the film thickness attenuation characteristic value of the sampling point sequence to which each sampling point belongs and the film thickness measurement concentricity, and record it as the first product; take the sum of the first products of all sampling points included in each waveform interval as the interference degree optimization factor of each waveform interval;
[0044] The waveform interval interference degree after optimization of each waveform interval is positively correlated with the waveform interference degree of each waveform interval, and is negatively correlated with the interference degree optimization factor.
[0045] Further, obtaining the filter window adjustment factor of each waveform interval according to the optimized waveform interval interference degree includes:
[0046] The echo signal is filtered using a preset initial filter window to obtain a smoothed signal strength value of each echo signal point;
[0047] For each waveform interval, the absolute value of the difference between the signal strength value of each echo signal point in the waveform interval and the smoothed signal strength value is calculated, and recorded as the filtering strength of each echo signal point;
[0048] The interference degree of each waveform interval after optimization and the filtering strength of all echo signal points in the waveform interval are positively correlated with the filter window adjustment factor of each waveform interval.
[0049] Furthermore, filtering the echo signal at each sampling point according to the optimal filtering window length to obtain a filtered echo signal includes: taking the maximum value of the optimal filtering window length of all waveform intervals contained in the echo signal at each sampling point as the length of the filtering window in the filtering algorithm, filtering the echo signal, and obtaining a filtered echo signal.
[0050] Further, obtaining the film thickness of the semiconductor substrate coating according to the filtered echo signals at all sampling points and the waveform interval interference degree includes:
[0051] For each sampling point, take time as the horizontal axis and the signal strength value of the echo signal point in the echo signal as the vertical axis to obtain the echo signal curve after filtering; obtain the peak point of the echo signal curve; the time corresponding to the intersection of the straight line obtained by fitting all the peak points and the horizontal axis is recorded as the first stage coating reflection time;
[0052] Record the waveform interval with the largest waveform interval interference as the second stage characteristic waveform interval; obtain the mutation point in the second stage characteristic waveform interval; record the minimum value of the time corresponding to all the mutation points as the second stage substrate reflection time; and take the difference between the second stage substrate reflection time and the first stage coating film reflection time as the characteristic propagation time;
[0053] Obtaining the propagation speed of the ultrasonic wave in the semiconductor substrate; the coating thickness at each sampling point is the product of the characteristic propagation time and the propagation speed;
[0054] The film thickness of the semiconductor substrate coating is the average of the coating thickness at all sampling points.
[0055] This application has at least the following beneficial effects:
[0056] The present application divides the echo signal into different waveform intervals by adaptively analyzing the echo signal when measuring the film thickness of the semiconductor substrate, and adaptively constructs the interference degree of the waveform interval based on the changes in the echo signal in different waveform intervals, which preliminarily reflects the interference degree at the echo signal point; further, the different effects of noise on the interference degree at different echo signal points are analyzed, and the waveform interval interference degree is adaptively optimized based on the center accumulation effect generated in the semiconductor spin coating process and the spin coating characteristics, eliminating the influence of different noises generated by uneven spin coating on the calculation accuracy of the waveform interval interference degree; a filter window adjustment factor is constructed based on the waveform interval interference degree, and the filter window is adaptively adjusted to obtain the optimal window according to the echo signal characteristics at each sampling point, thereby improving the filtering effect, retaining the original signal characteristics while removing the noise at the same sampling point in a targeted manner, thereby calculating the film thickness and improving the film thickness measurement accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work, among which:
[0058] Figure 1 A flow chart of the steps of a method for accurately measuring the film thickness of large-size spin coating provided in this application;
[0059] Figure 2 The block diagram for obtaining the interference degree of the optimized waveform interval;
[0060] Figure 3 This is a schematic diagram of peak point fitting. DETAILED DESCRIPTION
[0061] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0062] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
[0063] The following is a detailed description of a specific method for accurately measuring the film thickness of large-size spin coating provided by the present application in conjunction with the accompanying drawings.
[0064] An embodiment of the present application provides a method for accurately measuring the thickness of a film for large-size spin coating. Specifically, the following method for accurately measuring the thickness of a film for large-size spin coating is provided. Please refer to Figure 1 , the method comprises the following steps:
[0065] S10, collecting the excitation signal applied to the ultrasonic transducer and the echo signal at each sampling point on the surface of the semiconductor substrate.
[0066] The surface of a semiconductor substrate is often circular. In this embodiment, N=10 sampling points are set at equal intervals on the edge of the semiconductor substrate surface. M=9 sampling points are set at equal intervals on a line starting from the center point of the semiconductor substrate surface and ending at each sampling point, so there are M+1 sampling points on each sampling line; the M+1 sampling points on each sampling line are sorted in ascending order according to the Euclidean distance from the center point of the semiconductor substrate surface to obtain a sequence of sampling points. It should be noted that, through the sampling points set at equal intervals, the sampling points with the same sequence number in the N sampling point sequence have the same Euclidean distance to the center point of the semiconductor substrate surface, that is, they are located on the same concentric circle centered on the center point of the semiconductor substrate surface; based on this, the sampling points located on the same concentric circle as each sampling point are recorded as concentric sampling points of each sampling point.
[0067] When ultrasonic wave is used to measure the film thickness of semiconductor substrate, it is usually measured through reflection between the coating surface and the substrate. After applying the excitation signal to the ultrasonic transducer, the ultrasonic transducer generates an ultrasonic signal from the excitation signal. Specifically, the ultrasonic signal first passes through the coating and finally reaches the surface of the semiconductor substrate. At the junction of the coating and the substrate, part of the ultrasonic wave will be reflected back by the coating, while the other part will penetrate the coating and continue to propagate downward and be reflected on the surface of the substrate.
[0068] At each sampling point, the excitation signal and echo signal applied to the ultrasonic transducer are collected. According to the process of ultrasonic measurement of semiconductor substrate film thickness, the ultrasonic signal is reflected twice on the coating surface and the substrate surface, so the echo signal received by the ultrasonic transducer is divided into two stages. The echo signal of the first stage is generated by the reflection of the coating surface, and the echo signal of the second stage is the superposition of the echo signals generated by the reflection of the coating surface and the reflection of the substrate surface.
[0069] S20, dividing the echo signal into waveform intervals according to the excitation signal and the distribution of zero points in the echo signal at each sampling point.
[0070] Since ultrasonic measurement of film thickness may be affected by various noises from the environment, equipment or the material itself, such as electromagnetic interference, mechanical vibration or background noise, it is often necessary to use a filtering algorithm to remove noise interference from the echo signal when measuring film thickness.
[0071] Median filtering is a commonly used nonlinear filtering method. Its window size has an important influence on the filtering effect. The traditional median filtering window is often a fixed window set in advance. However, if the window size is too large, it may cause signal loss or blurring. Although the window size is too small, more signal details are retained, but its ability to remove noise also decreases. Therefore, determining the optimal median filtering window for the echo signal data point can effectively improve the quality of the echo signal and measure the film thickness more accurately. In this embodiment, the initial filter window size is set to K×K, K=3.
[0072] The echo signals collected at each sampling point are analyzed. Since the echo signals are usually a series of waveforms, the echo signals are preliminarily segmented in this embodiment to divide the echo signals into different waveform intervals, and the zero point of the echo signal is obtained, that is, the data point where the signal strength value is 0. Since slight fluctuations will occur in the initial stage when the ultrasonic transducer receives the echo signal, the obtained zero point needs to be screened to obtain the waveform interval.
[0073] Specifically, since the echo signal is generated by the excitation signal, the corresponding signal interval should be nearly consistent, and the excitation signal is a pre-set signal. The excitation signal can be regarded as a sine wave signal, so the corresponding standard waveform signal interval length L can be directly obtained, which is half the period length of the sine wave. Zero points are adaptively screened based on the standard waveform signal interval length and the distance between zero points. For the zero point corresponding to the acquisition time t in the echo signal, the time sequence distance between the zero point corresponding to the acquisition time t and the next zero point is obtained. ,in, is the acquisition time of the next zero point of the zero point corresponding to the acquisition time t. Further, the time sequence distance difference of each zero point is determined according to the time sequence distance between each zero point and its next zero point; the absolute value of the difference between the time sequence distance d and the standard waveform signal interval length L is calculated, and recorded as the first difference absolute value; then the time sequence distance difference of the zero point corresponding to the acquisition time t is positively correlated with the first difference absolute value.
[0074] It should be noted that positive correlation indicates a directional change relationship in which the dependent variable increases with the increase of the independent variable and decreases with the decrease of the independent variable. The positive correlation can specifically be a multiplication relationship, an addition relationship, and a normalized result, etc., which is determined by actual application and is not particularly limited in this application.
[0075] In this embodiment, the temporal distance difference of the zero point corresponding to the acquisition time t is the normalized value of the absolute value of the first difference.
[0076] The smaller the time series distance difference, the more likely the zero point is a waveform zero point. The average value of the time series distance difference of all zero points is used as the interference threshold. When the time series distance difference corresponding to the zero point is greater than the interference threshold, it is a noise interference zero point. The zero point whose time series distance difference is less than or equal to the interference threshold is recorded as a time series zero point. The echo signal contained in the closed interval between any two adjacent time series zero points in the echo signal is recorded as a waveform interval.
[0077] S30, obtaining the center point of each waveform interval, and obtaining the waveform interference degree according to the difference between the echo signal points symmetrical about the center point.
[0078] The waveform intervals are divided according to the above steps. The echo signals in each waveform interval are further analyzed. First, the echo signal point located at the center of the waveform interval is recorded as the center point o. The echo signals in the waveform area are often symmetrical. q Take point as an example, and obtain the echo signal point symmetrical about the center point o, which is recorded as the central symmetric point p q,o .
[0079] Get the echo signal point p in the waveform interval q The minimum absolute value of the difference between the signal strength values of the echo signal points and other echo signal points is recorded as the echo signal point p q The difference in adjacent signal strength is q The echo signal point in the initial filtering window centered at p q,o The echo signal points in the initial filtering window centered at are matched one by one in time order to obtain the echo signal point p qK groups of corresponding points; calculate the absolute value of the difference between the signal strength values of two echo signal points in each group of corresponding points, recorded as the strength corresponding difference value; calculate the absolute value of the difference between the adjacent signal strength difference values of two echo signal points in each group of corresponding points, recorded as the adjacent corresponding difference value.
[0080] The intensity corresponding difference values and the adjacent corresponding difference values of all groups of corresponding points are all the same as the echo signal point p q The deviation is positively correlated.
[0081] As an embodiment of the present application, the echo signal point p q The degree of deviation is the product of the sum of the intensity corresponding difference values of all groups of corresponding points and the sum of the adjacent corresponding difference values.
[0082] The greater the intensity corresponding difference value of all groups of corresponding points of the echo signal point and the adjacent corresponding difference value, the more inconsistent the local signal of the echo signal point and its symmetrical point is. The greater the deviation, the greater the degree of destruction of the symmetry at the echo signal point, and the greater the possibility of interference.
[0083] The sum of the deviations of all echo signal points in the waveform interval is further taken as the waveform interference degree of the waveform interval.
[0084] The greater the waveform interval interference, the greater the interference to the echo signal point in the waveform interval, and the larger the corresponding median filter window adjustment factor should be.
[0085] S40, optimizing the waveform interference degree according to the difference between the waveform intervals at the sampling points at different positions.
[0086] Since the waveform interval interference degree is constructed by the deviation degree of the data points, when the noise interference is distributed on both sides of the center point, that is, the data points on both sides of the center point are interfered by the noise, the data points may be interfered with but the deviation degree is low. Therefore, in this application, the waveform interval interference degree optimization factor is constructed by changing the echo signal of the sampling point and the echo signal of the associated sampling point, which specifically includes:
[0087] S41, obtaining a film thickness attenuation characteristic value according to the echo signals of all sampling points in each sampling point sequence obtained by arranging all sampling points.
[0088] When the semiconductor substrate is spin-coated, a central accumulation effect often occurs, that is, the closer to the center area of the coating application point, the thicker the coating thickness is. This is because the coating diffuses outward under the centrifugal force around the application point during the spin coating process, causing the coating to accumulate in the center and form a thicker coating. The farther away from the application point, the thinner the coating tends to be. The centrifugal force of the spin coating on the coating at the sampling points on the same concentric circle is consistent, so the coating thickness should be similar, that is, the echo signals received are also close to the same. Therefore, the sampling points on the semiconductor substrate are close to the echo signals at their concentric sampling points.
[0089] The maximum absolute value of the signal strength values of all echo signal points within the waveform interval is taken as the amplitude of the waveform interval. Since the amplitude of the waveform interval can reflect the film thickness to a certain extent, in the first stage, since the echo signal is reflected from the coating surface, the greater the film thickness, the more reflected signals there are, which increases the amplitude of the echo signal. Therefore, the change in the signal amplitude of the first waveform interval of the sampling point can be used to reflect the central accumulation effect of the film thickness, and the amplitude of the first waveform interval in the echo signal at each sampling point is taken as the characteristic signal amplitude of each sampling point.
[0090] For the sampling points on the surface of the semiconductor substrate, the film thickness corresponding to the sampling points in the sampling point sequence should show a trend of decreasing thickness from the center to the edge, and the attenuation trend between different sampling point sequences is consistent, otherwise it is subject to greater interference. Based on this, the film thickness attenuation characteristic values of each sampling point sequence can be obtained, including:
[0091] The sequence composed of the characteristic signal amplitudes of all sampling points in each sampling point sequence is recorded as the characteristic signal amplitude sequence of each sampling point sequence;
[0092] Calculate the correlation between each sampling point sequence and the characteristic signal amplitude sequence of other sampling point sequences; obtain the first-order difference sequence of the characteristic signal amplitude sequence of each sampling point sequence; calculate the mean of all elements in the first-order difference sequence, recorded as the attenuation amplitude of each sampling point sequence; calculate the absolute value of the difference between the attenuation amplitude of each sampling point sequence and other sampling point sequences, recorded as the attenuation difference between each sampling point sequence and other sampling point sequences;
[0093] The film thickness attenuation characteristic value of each sampling point sequence is positively correlated with the correlation and negatively correlated with the attenuation difference.
[0094] It should be noted that the calculation method of the correlation between the two sequences can be Pearson correlation coefficient, cosine similarity, etc., which is determined by actual application and is not specifically limited in this application. Reverse correlation indicates an inverse change relationship in which the dependent variable decreases as the independent variable increases and increases as the independent variable decreases. The reverse correlation can specifically be a division relationship, a subtraction relationship, and reverse normalization, etc., which is determined by actual application and is not specifically limited in this application.
[0095] It should be noted that in order to ensure that the calculation results are meaningful, when the reverse correlation is a division relationship, if the denominator is 0, it is necessary to add a parameter adjustment factor greater than 0 to the denominator to prevent the fraction from being meaningless. The value of the parameter adjustment factor is set by the implementer according to actual conditions, and this application does not impose any special restrictions.
[0096] As an embodiment of the present application, the correlation between the characteristic signal amplitude sequences of two sampling point sequences is the Pearson correlation coefficient between the characteristic signal amplitude sequences of the two sampling point sequences; the specific calculation process of the Pearson correlation coefficient is a well-known technology and will not be repeated here.
[0097] As an embodiment of the present application, the film thickness attenuation characteristic value of each sampling point sequence is the ratio of the sum of the correlations between each sampling point sequence and all sampling point sequences to the attenuation amplitude.
[0098] The greater the correlation between the characteristic signal amplitude sequences of each sampling point sequence and other sampling point sequences, the closer the received echo signals are, and the smaller the attenuation difference is, the more consistent the attenuation trend between each sampling point sequence and other sampling point sequences is, the smaller the film thickness attenuation characteristic value is, and the less interference is received.
[0099] S42, obtaining each concentric circle, and obtaining the film thickness measurement concentricity according to the film thickness attenuation characteristic value of the sampling point sequence to which the sampling points located on the same concentric circle belong.
[0100] As for the concentric sampling points on the concentric circles, the influence of spin coating during the spin coating process is similar, and the excitation signal of the ultrasonic transducer is also consistent, so the received echo signals are also close to the same, otherwise they are subject to greater interference. Based on this, the concentricity of the film thickness measurement of each sampling point is constructed. The specific process is as follows:
[0101] The absolute value of the difference between the characteristic signal amplitude of each sampling point and its concentric sampling point is calculated, which is recorded as the concentric characteristic difference between each sampling point and its concentric sampling point; the correlation between the echo signal of each sampling point and its concentric sampling point is calculated, which is recorded as the signal similarity between each sampling point and its concentric sampling point;
[0102] As an embodiment of the present application, the correlation between the echo signals is the inverse of the DTW distance between the echo signals; the calculation process of the DTW distance is a well-known technology and will not be described in detail.
[0103] The film thickness attenuation characteristic value of the sampling point sequence to which each sampling point belongs and the signal similarity are both positively correlated with the film thickness measurement concentricity of each sampling point, and the concentricity characteristic difference is negatively correlated with the film thickness measurement concentricity of each sampling point.
[0104] As an embodiment of the present application, the film thickness measurement concentricity of the sampling point s , among which SJ s is the film thickness attenuation characteristic value of the sampling point sequence to which the sampling point s belongs; XS s is the cumulative sum of the signal similarities between sampling point s and all its concentric sampling points; CY s It is the cumulative sum of the concentric feature differences between sampling point s and all its concentric sampling points.
[0105] The concentricity of the sampling point is represented by the signal change in the waveform interval between the sampling point and the concentric sampling point. The greater the concentricity, the more accurate the signal received by the sampling point.
[0106] S43, optimizing the waveform interference degree of each waveform interval by combining the film thickness attenuation characteristic value and the film thickness measurement concentricity to obtain an optimized waveform interval interference degree.
[0107] Furthermore, the product of the film thickness attenuation characteristic value of the sampling point sequence to which each sampling point belongs and the concentricity of the film thickness measurement is calculated and recorded as the first product; the sum of the first product of all sampling points contained in each waveform interval is used as the interference optimization factor of each waveform interval; the waveform interval interference after optimization of each waveform interval is positively correlated with the waveform interference of each waveform interval, and is negatively correlated with the interference optimization factor.
[0108] In this embodiment, the waveform interference degree of each waveform interval after optimization is the ratio of the waveform interference degree of each waveform interval to the interference degree optimization factor.
[0109] The block diagram for obtaining the optimized waveform interval interference degree is as follows: Figure 2 shown.
[0110] S50, obtaining the optimal filtering window length of each waveform interval according to the optimized waveform interval interference degree.
[0111] According to the above steps, the optimized waveform interval interference degree is obtained, the local data signal strength distribution of the echo signal data point is analyzed, and the filter window adjustment factor is adaptively constructed:
[0112] The echo signal is filtered using a preset initial filter window to obtain a smoothed signal strength value of each echo signal point. As an embodiment of the present application, the echo signal is filtered using a median filter algorithm; it should be understood that as other implementations, the implementer may choose other filter algorithms with a filter window, such as a mean filter, and the present application does not impose any special restrictions on this.
[0113] For each waveform interval, the absolute value of the difference between the signal strength value of each echo signal point in the waveform interval and the smoothed signal strength value is calculated, and recorded as the filtering strength of each echo signal point;
[0114] The interference degree of each waveform interval after optimization and the filtering strength of all echo signal points in the waveform interval are positively correlated with the filter window adjustment factor of each waveform interval.
[0115] In this embodiment, the filter window adjustment factor of each waveform interval is the product of the average filter strength of each waveform interval and the optimized waveform interval interference degree; wherein the average filter strength of the waveform interval is the sum of the filter strengths of all echo signal points in the waveform interval.
[0116] Furthermore, the filter window adjustment factor is normalized, and the larger the value is, the larger the window needs to be increased.
[0117] According to the above steps, each waveform interval has a corresponding filter window adjustment factor. In order to obtain a better filtering effect, the window size is adjusted multiple times based on the filter window adjustment factor:
[0118] Taking the calculation of the window length adjusted for the i-th time as an example, the product of the window length adjusted for the i-1th time and the filter window adjustment factor corresponding to the window length adjusted for the i-1th time is calculated, and recorded as the second product; the rounded-up value of the second product is used as the estimated length increment of the window length adjusted for the i-th time.
[0119] In order to ensure the normal filtering, the filtering window length should be an odd number. Therefore, twice the estimated length increment of the window after the i-th adjustment is used as the length increase value of the window length adjusted for the i-th time; then the window length adjusted for the i-th time is the sum of the window length after the i-1-th adjustment and the length increase value.
[0120] According to the above steps, the window size is continuously adjusted until the filter window adjustment factor is less than the preset threshold value 0.6 or the number of adjustments reaches 20 times, and the window length obtained by the last adjustment is used as the optimal filter window. The maximum value of the optimal filter window of all waveform intervals contained in the echo signal at each sampling point is used as the length of the filter window in the filtering algorithm, and the echo signal is filtered to obtain the filtered echo signal.
[0121] S60, filtering the echo signal at each sampling point according to the optimal filter window length to obtain a filtered echo signal; and obtaining the film thickness of the semiconductor substrate coating according to the filtered echo signals at all sampling points and the waveform interval interference.
[0122] Due to the slight fluctuation at the time of receiving the echo signal, it is impossible to accurately obtain the moment when the ultrasonic transducer receives the echo signal. Instead, a straight line fitting is often performed through the peak point of the echo signal, and the characteristic point of the echo signal is determined by the intersection of the fitted straight line and the horizontal axis, i.e., the time axis. The characteristic point is the moment when the ultrasonic transducer receives the echo signal, and the time difference between the characteristic point and the excitation moment of the excitation signal is the characteristic propagation time of the ultrasonic signal.
[0123] For each sampling point, take time as the horizontal axis and the signal strength value of the echo signal point in the echo signal as the vertical axis to obtain the echo signal curve after filtering; obtain the peak point of the echo signal curve; Figure 3 As shown, all peak points are fitted with a straight line using the least square method, and the time corresponding to the intersection of the fitted straight line and the horizontal axis is recorded as the first stage coating reflection time. The specific implementation process of peak point detection and the least square method are both well-known technologies and will not be repeated here.
[0124] Further screening of the waveform intervals shows that since the two echo signals in the second stage are excited by the same excitation signal, the superimposed echo signals in the second stage are echo signals with equal periods and larger amplitudes. At the beginning of the second stage, assuming point b, the waveform interval where point b is located is often the largest in terms of waveform interval interference because the echo signals before point b are reflected by the coating, and the echo signals after point b are reflected by the coating and the surface of the semiconductor substrate.
[0125] Based on this, the waveform interval with the largest waveform interval interference is recorded as the second stage characteristic waveform interval. The echo signal points in the second stage characteristic waveform interval are detected using the wavelet transform algorithm to obtain the mutation points, and the minimum value of the time corresponding to all the mutation points is recorded as the second stage substrate reflection time; the difference between the second stage substrate reflection time and the first stage coating reflection time is used as the characteristic propagation time. The wavelet transform is a well-known technology and will not be described here.
[0126] According to the above steps, the characteristic propagation time is obtained, and the propagation speed of the ultrasonic wave in the semiconductor substrate coating is obtained based on the material of the semiconductor substrate coating; the coating thickness at each sampling point is the product of the characteristic propagation time and the propagation speed. Finally, the film thickness of the semiconductor substrate coating is the average of the coating thickness at all sampling points.
[0127] Through the above description of the implementation method in combination with the accompanying drawings, technical personnel in the relevant field can understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.
[0128] The above contents are only specific implementation methods of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A method for accurately measuring the film thickness of large-size spin coating, characterized in that: The method comprises the following steps: S10, collecting an excitation signal applied to the ultrasonic transducer and an echo signal at each sampling point on the surface of the semiconductor substrate; S20, dividing the echo signal into waveform intervals according to the excitation signal and the distribution of zero points in the echo signal at each sampling point; S30, obtaining the center point of each waveform interval, and obtaining the waveform interference degree according to the difference between the echo signal points symmetrical about the center point; S40, optimizing the waveform interference degree according to the difference between the waveform intervals at the sampling points at different positions; the optimization process includes: S41, obtaining a film thickness attenuation characteristic value according to the echo signals of all sampling points in each sampling point sequence obtained by arranging all sampling points; S42, obtaining each concentric circle, and obtaining the film thickness measurement concentricity according to the film thickness attenuation characteristic value of the sampling point sequence to which the sampling points located on the same concentric circle belong; S43, optimizing the waveform interference degree of each waveform interval by combining the film thickness attenuation characteristic value and the film thickness measurement concentricity to obtain an optimized waveform interval interference degree; S50, obtaining the optimal filtering window length of each waveform interval according to the optimized waveform interval interference degree; S60, filtering the echo signal at each sampling point according to the optimal filter window length to obtain a filtered echo signal; obtaining the film thickness of the semiconductor substrate coating according to the filtered echo signals at all sampling points and the waveform interval interference degree; The step of obtaining the optimal filtering window length of each waveform interval according to the optimized waveform interval interference degree comprises: Obtaining a filter window adjustment factor for each waveform interval according to the optimized waveform interval interference degree; For each adjustment, the product of the window length of the previous adjustment and the filter window adjustment factor corresponding to the window length of the previous adjustment is calculated, and recorded as the second product; the rounded-up value of the second product is used as the estimated length increment of the window length of each adjustment; The twice of the estimated length increment of the window after each adjustment is used as the length increment of the window length adjusted each time; the window length adjusted each time is the sum of the window length after the previous adjustment and the length increment; The window length adjustment process is repeated until the filter window adjustment factor is less than the first preset threshold or the number of adjustments reaches the second preset threshold; the window length obtained by the last adjustment is used as the optimal filter window length.
2. A method for accurately measuring film thickness for large-size spin coating as claimed in claim 1, characterized in that: The method of dividing the echo signal into waveform intervals according to the distribution of zero points in the echo signal at each sampling point and the excitation signal includes: Obtain each zero point in the echo signal, calculate the absolute value of the difference between each zero point and the acquisition time corresponding to the next zero point, and record it as the time sequence distance between each zero point and the next zero point; Obtaining the standard waveform signal interval length of the excitation signal, calculating the absolute value of the difference between the standard waveform signal interval length and the timing distance, and recording it as the first absolute value of the difference; the timing distance difference of each zero point is positively correlated with the first absolute value of the difference; The average value of the timing distance difference of all zero points in the echo signal is taken as the interference threshold; the zero point whose timing distance difference is less than or equal to the interference threshold is recorded as the timing zero point; the echo signal contained in the closed interval between any two adjacent timing zero points is recorded as a waveform interval.
3. A method for accurately measuring film thickness for large-size spin coating as claimed in claim 1, characterized in that: The step of obtaining the center point of each waveform interval and obtaining the waveform interference degree according to the difference between echo signal points symmetrical about the center point includes: Record the echo signal point located at the center of the waveform interval as the center point; obtain the echo signal point that is symmetrical about the center point of each echo signal point in the waveform interval, and record it as the central symmetrical point of each echo signal point; Obtain the minimum absolute value of the difference between the signal strength values of each echo signal point and other echo signal points in the waveform interval, and record it as the adjacent signal strength difference value of each echo signal point; For each echo signal point in the waveform interval, a preset initial filtering window centered on the echo signal point is matched one by one with the echo signal points in the initial filtering window centered on the central symmetric point to obtain each group of corresponding points; the absolute value of the difference between the signal strength values of two echo signal points in each group of corresponding points is calculated, recorded as the strength corresponding difference value; the absolute value of the difference between the adjacent signal strength difference values of two echo signal points in each group of corresponding points is calculated, recorded as the adjacent corresponding difference value; The intensity corresponding difference values and the adjacent corresponding difference values of all groups of corresponding points are positively correlated with the deviation degree of the echo signal point; The sum of the deviations of all echo signal points within the waveform interval is taken as the waveform interference degree of the waveform interval.
4. A method for accurately measuring film thickness for large-size spin coating according to claim 1, characterized in that: The obtaining of the film thickness attenuation characteristic value comprises: The maximum absolute value of the signal strength values of all echo signal points within the waveform interval is taken as the amplitude of the waveform interval; The amplitude of the first waveform interval in the echo signal at each sampling point is used as the characteristic signal amplitude of each sampling point; The sampling points on the line starting from the center point of the semiconductor substrate surface and ending at the sampling points on the edge of the semiconductor substrate surface form a sampling point sequence; The sequence composed of the characteristic signal amplitudes of all sampling points in each sampling point sequence is recorded as the characteristic signal amplitude sequence of each sampling point sequence; Calculate the correlation between each sampling point sequence and the characteristic signal amplitude sequence of other sampling point sequences; obtain the first-order difference sequence of the characteristic signal amplitude sequence of each sampling point sequence; calculate the mean of all elements in the first-order difference sequence, recorded as the attenuation amplitude of each sampling point sequence; calculate the absolute value of the difference between the attenuation amplitude of each sampling point sequence and other sampling point sequences, recorded as the attenuation difference between each sampling point sequence and other sampling point sequences; The film thickness attenuation characteristic value of each sampling point sequence is positively correlated with the correlation and negatively correlated with the attenuation difference.
5. A method for accurately measuring film thickness for large-size spin coating as claimed in claim 4, characterized in that: The step of obtaining each concentric circle and obtaining the film thickness measurement concentricity according to the film thickness attenuation characteristic value of the sampling point sequence to which the sampling points located on the same concentric circle belong comprises: Obtain concentric circles with the center point of the semiconductor substrate surface as the center and the Euclidean distance from each sampling point to the center point of the semiconductor substrate surface as the radius; record the sampling points located on the same concentric circle as each sampling point as concentric sampling points of each sampling point; The absolute value of the difference between the characteristic signal amplitude of each sampling point and its concentric sampling point is calculated, which is recorded as the concentric characteristic difference between each sampling point and its concentric sampling point; the correlation between the echo signal of each sampling point and its concentric sampling point is calculated, which is recorded as the signal similarity between each sampling point and its concentric sampling point; The film thickness attenuation characteristic value of the sampling point sequence to which each sampling point belongs and the signal similarity are both positively correlated with the film thickness measurement concentricity of each sampling point, and the concentricity characteristic difference is negatively correlated with the film thickness measurement concentricity of each sampling point.
6. A method for accurately measuring film thickness for large-size spin coating according to claim 1, characterized in that: The method of optimizing the waveform interference degree of each waveform interval by combining the film thickness attenuation characteristic value and the film thickness measurement concentricity to obtain the optimized waveform interval interference degree includes: Calculate the product of the film thickness attenuation characteristic value of the sampling point sequence to which each sampling point belongs and the film thickness measurement concentricity, and record it as the first product; take the sum of the first products of all sampling points included in each waveform interval as the interference degree optimization factor of each waveform interval; The waveform interval interference degree after optimization of each waveform interval is positively correlated with the waveform interference degree of each waveform interval, and is negatively correlated with the interference degree optimization factor.
7. A method for accurately measuring film thickness for large-size spin coating according to claim 1, characterized in that: The step of obtaining the filter window adjustment factor of each waveform interval according to the optimized waveform interval interference degree includes: The echo signal is filtered using a preset initial filter window to obtain a smoothed signal strength value of each echo signal point; For each waveform interval, the absolute value of the difference between the signal strength value of each echo signal point in the waveform interval and the smoothed signal strength value is calculated, and recorded as the filtering strength of each echo signal point; The interference degree of each waveform interval after optimization and the filtering strength of all echo signal points in the waveform interval are positively correlated with the filter window adjustment factor of each waveform interval.
8. A method for accurately measuring film thickness for large-size spin coating as claimed in claim 1, characterized in that: The method of filtering the echo signal at each sampling point according to the optimal filtering window length to obtain the filtered echo signal includes: taking the maximum value of the optimal filtering window length of all waveform intervals contained in the echo signal at each sampling point as the length of the filtering window in the filtering algorithm, filtering the echo signal, and obtaining the filtered echo signal.
9. A method for accurately measuring film thickness for large-size spin coating according to claim 1, characterized in that: The method of obtaining the film thickness of the semiconductor substrate coating according to the filtered echo signals at all sampling points and the waveform interval interference degree comprises: For each sampling point, take time as the horizontal axis and the signal strength value of the echo signal point in the echo signal as the vertical axis to obtain the echo signal curve after filtering; obtain the peak point of the echo signal curve; the time corresponding to the intersection of the straight line obtained by fitting all the peak points and the horizontal axis is recorded as the first stage coating reflection time; Record the waveform interval with the largest waveform interval interference as the second stage characteristic waveform interval; obtain the mutation point in the second stage characteristic waveform interval; record the minimum value of the time corresponding to all the mutation points as the second stage substrate reflection time; and take the difference between the second stage substrate reflection time and the first stage coating film reflection time as the characteristic propagation time; Obtaining the propagation speed of the ultrasonic wave in the semiconductor substrate; the coating thickness at each sampling point is the product of the characteristic propagation time and the propagation speed; The film thickness of the semiconductor substrate coating is the average of the coating thickness at all sampling points.
