Scanning probe measurement method and system with high signal-to-noise ratio
By using reference curves to set weighting coefficients and adjusting measurement time-consuming parameters in scanning probe measurement, the problem of signal-to-noise ratio increase and time-consuming increase in scanning measurement is solved, and the measurement data with high signal-to-noise ratio is achieved in a shorter range. Obtaining within time.
Patent Information
- Application Number
- CN202310492104.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-04
- Publication Date
- 2024-11-12
AI Technical Summary
In scanning measurement techniques, in order to improve the signal-to-noise ratio, existing methods usually require significantly increasing the measurement time, resulting in excessive time-consuming measurements, and methods that increase the signal-to-noise ratio by increasing the number of measurements or time can lead to data distortion.
By obtaining the reference curve of the sample to be tested, setting the weighting coefficients of each measurement point according to the reference curve, and adjusting its measurement time-consuming parameters to allocate longer measurement time in important areas, improve the signal-to-noise ratio, and shorten in flat areas. Measure time to achieve high signal-to-noise ratio measurement data within the same time.
Under the condition that meets certain signal-to-noise ratio requirements, the measurement time is significantly reduced, the overall signal-to-noise ratio of the measurement curve is improved, data distortion is avoided, and measurement efficiency is improved.
Smart Images

Figure CN118936304A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of measurement technology, and in particular to a scanning probe measurement method and system with a high signal-to-noise ratio. Background Art
[0002] In the measurement process, a suitable signal-to-noise ratio is always the purpose of data measurement. In order to obtain a sufficient signal-to-noise ratio, multiple sampling at each measurement point or long-term integration can be used to remove noise and interference in the measurement process.
[0003] The signal-to-noise ratio of optical imaging measurement is generally very high because all measurement points are measured simultaneously, which is a parallel measurement technology. In this way, all measurement points receive the same integration time at the same time.
[0004] However, this is not the case with scanning measurement technologies, such as scanning electron microscopes, scanning probe microscopes, and electron beam lithography. Scanning measurement requires setting a measurement time, which is allocated to all measurement points. For the case where the processing time for each measurement point is the same, the measurement time obtained for each measurement point is the total measurement time divided by the total number of measurement points. For example, if there are 1 million measurement points, compared with the parallel measurement technology of optical imaging measurement, the measurement time obtained for each measurement point is only one millionth of that of the parallel measurement technology.
[0005] Measurement always refers to measuring the measurement effect that meets the required signal-to-noise ratio within a certain period of time.
[0006] For example, when measuring the sample in the X direction, a measurement function F=F(x) of 100 points is set up between position X1 and position X2. The general practice is to assign the same measurement time to each measurement point and measure the measurement value of each measurement point. It is also possible to repeat the measurement N times (N is an integer greater than 1) from position X1 to position X2, and obtain the same measurement time for each measurement point each time. Then the measurement values of each measurement point are superimposed and averaged accordingly; thus, the improvement of the signal-to-noise ratio of each measurement point of this measurement function relative to a single measurement is N. 1 / 2 .like Figure 1 As shown in FIG. 1 , a measurement curve with noise is obtained by superimposing and averaging a measurement curve with noise removed after a single measurement, and the noise is 10 times smaller than that of a single measurement. The horizontal axis is the measurement time t, and the vertical axis is the measurement value Z.
[0007] The above measurement method of setting each measurement point to obtain the same measurement time or sampling number is the method used in most cases.
[0008] A more specific example is the use of probe technology to measure a three-dimensional topography image. Currently, commonly used scanning measurement methods include Figure 2As shown in the figure, scanning measurement methods include reciprocating measurement and Z-shaped measurement, both of which scan down one line at a time until the entire image is scanned and measured. The improvement of signal-to-noise ratio is generally implemented by measuring the time of each measurement point on the surface by the probe. If the measurement time of the entire image is 10 minutes, the measurement time is increased to 100 minutes. In theory, the signal-to-noise ratio is only increased by 3.3 times.
[0009] The above process is a method of greatly increasing the measurement time in exchange for the signal-to-noise ratio. After a complete measurement is carried out, the signal-to-noise ratio can be increased by increasing the number of scans of the entire image, that is, increasing the number of measurements of each measurement point. This is equivalent to increasing the measurement time of each point to increase the signal-to-noise ratio; people generally use this integration to improve the signal-to-noise ratio.
[0010] The problem is that the measurement time is too long to obtain a smooth measurement curve with a high signal-to-noise ratio.
[0011] In order to obtain a smooth measurement curve in a limited time, one way is to post-process the measurement data, such as the commonly used image processing algorithm. However, it must be pointed out that although many people use image processing technology to improve the smoothness of the curve, this is not real measurement data after all, which will lead to data distortion.
[0012] So the question is, with the same measurement time, how to use a suitable measurement method to improve the signal-to-noise ratio of the scanning measurement curve? Or how to reduce the measurement time at the specified signal-to-noise ratio target value? Based on the above technical problems, the applicant proposed the technical solution of this application. Summary of the invention
[0013] The object of the present invention is to provide a scanning probe measurement method and a scanning probe system with a high signal-to-noise ratio, which can maintain a high signal-to-noise ratio in areas of high importance and improve the overall curve signal-to-noise ratio of the obtained measurement curve; that is, measurement data with a high signal-to-noise ratio can be obtained within the same measurement time, that is, the measurement time can be reduced as much as possible under the condition of meeting certain signal-to-noise ratio requirements.
[0014] To achieve the above-mentioned objectives, the present invention provides a scanning probe measurement method with a high signal-to-noise ratio, comprising: obtaining a reference curve of a sample to be measured, and setting a weighting coefficient of each measurement point according to the reference curve, wherein the reference curve represents the reference measurement values of multiple measurement points on the surface of the sample to be measured obtained by at least one measurement; based on the weighting coefficient of each measurement point, adjusting a measurement time parameter of each measurement point, wherein the measurement time parameter represents the processing time required to measure the measurement point; and scanning and measuring the surface of the sample to be measured based on the adjusted measurement time parameter of each measurement point to obtain a measurement curve.
[0015] The present invention also provides a scanning probe system for executing the above-mentioned scanning probe measurement method with a high signal-to-noise ratio.
[0016] In one embodiment, adjusting the measurement time parameter of each measurement point based on the weighted coefficient of each measurement point includes: for each measurement point, taking the product of the current measurement time parameter of the measurement point and the weighted coefficient of the measurement point as the adjusted measurement time parameter of the measurement point.
[0017] In one embodiment, the reference curve and the measurement curve belong to the same measurement of the sample to be measured, the measurement time of the reference curve accounts for 10%-50% of the preset measurement time, and the measurement time of the measurement curve accounts for 90%-50% of the preset measurement time; the preset measurement time represents the processing time required for one measurement.
[0018] In one embodiment, before setting the weighting coefficient of each measurement point based on the reference curve of the sample to be measured, it also includes: performing a pre-scan measurement on the surface of the sample to be measured to obtain the reference curve, wherein during the pre-scan measurement process, the measurement time parameters of each measurement point are the same.
[0019] In one embodiment, the reference curve is a measurement curve obtained by scanning and measuring the surface of the sample to be measured last time.
[0020] In one embodiment, the measurement time parameter is measurement time or measurement times.
[0021] In one embodiment, setting the weight coefficient of each of the measurement points according to the reference curve includes: using the curvature of each of the measurement points on the reference curve as the weight coefficient of each of the measurement points.
[0022] In one embodiment, setting the weight coefficient of each of the measurement points according to the reference curve includes: using the gradient value of each of the measurement points on the reference curve as the weight coefficient of each of the measurement points.
[0023] In one embodiment, the weighting coefficient of each measurement point is set according to the reference curve, including: obtaining the fine structure area of the reference curve in each row and column respectively based on the structural gradient value of each measurement point on the reference curve; for each measurement point not located in any of the fine structure areas, setting the weighting coefficient of the measurement point to a constant; for each of the fine structure areas, setting the weighting coefficient of the measurement point located in the fine structure area based on the reference measurement data of the reference curve in the fine structure area and the reference measurement data of the reference curve in the fine structure area after noise elimination.
[0024] In one embodiment, each of the fine structure regions includes: a starting interval, a non-mutation interval, and an ending interval; the surface of the sample to be measured is scanned and measured based on the adjusted measurement time parameters of each measurement point to obtain a measurement curve, including: for each of the fine structure regions, based on the adjusted measurement time parameters of the measurement points in the fine structure region, measuring the starting interval and the ending interval in the fine structure region; for each of the fine structure regions, based on the measurement data of each of the measurement points in the starting interval and the ending interval in the fine structure region, interpolating the measurement values of each of the measurement points in the non-mutation interval of the fine structure region.
[0025] In one embodiment, after respectively acquiring the fine structure regions of the reference curve in each row and each column, the method further includes: for the row fine structure regions and the column fine structure regions covering the same measurement point, selecting fine structure regions with a small overall width from the row fine structure regions and the column fine structure regions for retention.
[0026] In one embodiment, the method of obtaining the fine structure region of the reference curve in each row and column based on the obtained structure gradient value of each measurement point on the reference curve comprises: for each target sequence of the reference curve, taking the current measurement point as the midpoint, sequentially selecting the structure gradient values of at least 4 measurement points with amplitude jumps from the target sequence to obtain the median, and selecting half of the median as the local threshold; wherein the target sequence is a row or a column of the reference curve; taking the current measurement point as the starting point, searching the target sequence for the first measurement point with a structure gradient value greater than or equal to the local threshold as the starting point of the starting interval of the found fine structure region, and searching the first measurement point with a structure gradient value less than or equal to the local threshold after the starting point of the starting interval as the end point of the starting interval; taking the next measurement point of the end point of the found starting interval of the fine structure region as the starting point, searching the target sequence for the first measurement point with a structure gradient value greater than or equal to the local threshold as the starting point of the ending interval of the fine structure region, and searching the first measurement point with a structure gradient value less than or equal to the local threshold after the starting point of the ending interval as the end point of the ending interval.
[0027] In one embodiment, the number of measurement curves is N, and each measurement curve includes M measurement points; after the surface of the sample to be measured is scanned and measured based on the adjusted measurement time parameters of each measurement point to obtain the measurement curve, it also includes: selecting one measurement curve from the N measurement curves as a reference curve, and obtaining the coordinate displacement of each measurement curve in the N measurement curves except the reference curve relative to the reference curve; superimposing and averaging the N measurement curves based on the coordinate displacement of each measurement curve relative to the reference curve to obtain a final target measurement curve.
[0028] In one embodiment, respectively obtaining the coordinate displacement of each measurement curve other than the reference curve among the N measurement curves relative to the reference curve includes: for each measurement curve other than the reference curve among the N measurement curves, obtaining a correlation function between the measurement curve and the reference curve; finding the minimum value of the correlation function with the coordinate displacement as a variable, and taking the coordinate displacement corresponding to the first minimum value of the correlation function as the coordinate displacement of the measurement curve relative to the reference curve.
[0029] In one embodiment, the number of measurement curves is N, and each measurement curve includes M measurement points; after the surface of the sample to be measured is scanned and measured based on the adjusted measurement time parameters of each measurement point to obtain the measurement curve, it also includes: for the Lth measurement curve among the N measurement curves, based on the superimposed average curve of the first L-1 measurement curves after displacement, the coordinate displacement of the Lth measurement curve is obtained; 1≤L≤N; based on the coordinate displacement of each measurement curve, the N measurement curves are superimposed and averaged to obtain the final target measurement curve.
[0030] In one embodiment, for the Lth measurement curve among the N measurement curves, the coordinate displacement of the Lth measurement curve is obtained based on the superposition and average of the first L-1 measurement curves after displacement, including: obtaining the correlation function between the Lth measurement curve and the superposition and average curve of the first L-1 measurement curves after displacement; finding the minimum value of the correlation function with the coordinate displacement as a variable, and taking the coordinate displacement corresponding to the first minimum value of the correlation function as the coordinate displacement of the Lth measurement curve.
[0031] In one embodiment, the measurement time parameter is any one of the following: measurement time, sampling times, or a measurement parameter proportional to the measurement time. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] Figure 1It is the measurement curve with noise from a single measurement and the measurement curve after denoising obtained by superimposing and averaging 100 measurements;
[0033] Figure 2 There are two existing scanning measurement methods;
[0034] Figure 3 is a specific flow chart of the scanning probe measurement method with high signal-to-noise ratio according to the first embodiment of the present invention;
[0035] Figure 4 is the isolated singular point p according to the first embodiment of the present invention. i,j Schematic diagram of a 5×5 neighborhood;
[0036] Figure 5 is a specific flow chart of a scanning probe measurement method with a high signal-to-noise ratio according to a second embodiment of the present invention;
[0037] Figure 6 is the measurement point p according to the second embodiment of the present invention i,j Schematic diagram of measurement points in the horizontal and vertical fields;
[0038] Figure 7 is a schematic diagram of a definition domain of a fine structure region to be found in a row in the horizontal direction according to a second embodiment of the present invention;
[0039] Figure 8 is a schematic diagram of a definition domain of a fine structure region to be found in a column in a vertical direction according to a second embodiment of the present invention;
[0040] Fig. 9 is a schematic diagram of a position where a mutation occurs in a second embodiment of the present invention, in which the position is located in both a fine structure region in the horizontal direction and a fine structure region in the vertical direction;
[0041] Fig.10 is a schematic diagram of three measurement curves obtained by measuring the surface of a sample to be measured three times according to the third embodiment of the present invention;
[0042] Fig.11 is a specific flow chart of the scanning probe measurement method with high signal-to-noise ratio according to the third embodiment of the present invention, wherein one measurement curve is selected from N measurement curves as a reference curve, and the coordinate displacement of each measurement curve other than the reference curve in the N measurement curves relative to the reference curve is obtained respectively;
[0043] Fig.12 It is a specific flow chart of the scanning probe measurement method with high signal-to-noise ratio according to the third embodiment of the present invention, wherein for the Lth measurement curve among N measurement curves, the coordinate displacement of the Lth measurement curve is obtained based on the superimposed average curve of the first L-1 measurement curves after displacement. Specific embodiments
[0044] The following will be described in detail with reference to the accompanying drawings to provide a clearer understanding of the purpose, features and advantages of the present invention. It should be understood that the embodiments shown in the accompanying drawings are not intended to limit the scope of the present invention, but are only intended to illustrate the essential spirit of the technical solution of the present invention.
[0045] In the following description, certain specific details are set forth for the purpose of illustrating various disclosed embodiments to provide a thorough understanding of the various disclosed embodiments. However, those skilled in the relevant art will recognize that the embodiments may be practiced without one or more of these specific details. In other cases, well-known devices, structures, and techniques associated with the present application may not be shown or described in detail to avoid unnecessarily obscuring the description of the embodiments.
[0046] Unless the context requires otherwise, throughout the specification and claims, the word "comprise" and variations such as "include" and "have" should be construed in an open, inclusive sense, ie, should be interpreted as "including, but not limited to."
[0047] References throughout the specification to "one embodiment" or "an embodiment" indicate that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, the appearances of "in one embodiment" or "in an embodiment" in various places throughout the specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any manner in one or more embodiments.
[0048] As used in this specification and the appended claims, the singular forms "a", "an", and "the" include plural referents unless the context clearly dictates otherwise. It should be noted that the term "or" is generally employed in its sense including "or / and" unless the context clearly dictates otherwise.
[0049] In the following description, in order to clearly show the structure and working mode of the present invention, many directional words will be used for description, but the words "front", "back", "left", "right", "outside", "inside", "outward", "inward", "up", "down", etc. should be understood as convenient terms and should not be understood as restrictive terms.
[0050] The first embodiment of the present invention relates to a scanning probe measurement method with a high signal-to-noise ratio, which is applied to a scanning probe system. A plurality of arrayed measurement points are arranged on the surface of a sample to be measured. The probe on the scanning probe system is used to perform scanning measurement on the surface of the sample to be measured, such as one-dimensional, two-dimensional or three-dimensional scanning measurement. The scanning measurement method includes but is not limited to: atomic force microscope probe scanning measurement, scanning electron microscope measurement, scanning beam measurement, scanning ion measurement and radar scanning measurement. A plurality of arrayed measurement points are arranged on the surface of the sample to be measured. The measurement area on the surface of the sample to be measured can be a one-dimensional space area, a two-dimensional space area or a three-dimensional space area. The sample to be measured is, for example, a wafer, a silicon wafer, a quartz wafer, etc.
[0051] The specific process of the high signal-to-noise ratio scanning probe measurement method of this embodiment is as follows: Figure 3 shown.
[0052] Step 101 , obtaining a reference curve of the sample to be tested, and setting a weighting coefficient of each measuring point according to the reference curve, wherein the reference curve represents reference measurement values of multiple measuring points on the surface of the sample to be tested obtained by at least one measurement.
[0053] Specifically, before measuring the surface of the sample to be tested, a reference curve of the sample to be tested is first obtained. The reference curve represents the reference measurement values of multiple measurement points on the surface of the sample to be tested obtained by at least one measurement, that is, the reference curve can be a curve obtained after multiple measurements of the surface of the sample to be tested.
[0054] Then, a weighting coefficient of each measuring point is set according to the reference measuring point of each measuring point on the reference curve. The weighting coefficient is related to the measurement time parameter of this measurement. The measurement time parameter characterizes the processing time required to measure the measuring point. It can be the measurement time of the measuring point or the number of measurements within a certain period of time, or other measurement parameters that are proportional to the measurement time.
[0055] In one example, a pre-scan measurement is performed on the surface of the sample to be tested to obtain a reference curve, wherein during the pre-scan measurement, the measurement time parameters of each measurement point are the same. That is, a pre-scan measurement is performed before each measurement of the sample to be tested, and the pre-scan measurement includes one or more measurements of the surface of the sample to be tested. If multiple measurements are performed, the measurement values of each measurement point obtained by the multiple measurements are superimposed and averaged to obtain a reference curve. Among them, during the pre-scan measurement, the measurement time parameters of each measurement point are the same, that is, the measurement time or the number of measurements is the same. For example, it is necessary to use an atomic force microscope to measure the three-dimensional morphology of the surface of the sample to be tested. Before the measurement, the three-dimensional morphology measured by an optical microscope can be used as a reference curve.
[0056] In one example, the reference curve is a measurement curve obtained by scanning and measuring the surface of the sample to be measured last time. That is, if the surface of the sample to be measured needs to be measured multiple times, after the first measurement is completed, the measurement curve obtained by the previous measurement is used as the reference curve for this measurement. The multiple measurements can be for the same scanning area on the surface of the sample to be measured, or for different scanning areas on the surface of the sample to be measured. The reference curve is a measurement curve obtained by scanning and measuring the scanning area adjacent to the scanning area of the current measurement.
[0057] In the first measurement, the measurement points in the measurement area from position X1 to position X2 on the surface of the sample to be measured are scanned and measured to obtain the measurement values of each measurement point and form a reference curve. The function of fitting the reference curve can be expressed as F=F(X i ), where X i Indicates the i-th measurement point. The measurement time parameters of each measurement point are the same, for example, the measurement time is the same, which is △T. It can also be understood that each measurement point is sampled N times with equal time, and the time of each sampling is △t, then △T=N*△t.
[0058] For example, let the scanning mode when using the original step scanning be S1, and the scanning adopts the equal interval sampling mode. The number of measuring points in the horizontal direction and the vertical direction are n and m respectively, and the total number of measuring points is Q = n*m. If the sampling time of each measuring point is Δt, the time required to complete an S1 scanning is T = Q*Δt.
[0059] When scanning by enlarging the horizontal and vertical step lengths, the step length enlargement ratio is set to e, and this scanning method is recorded as S e , then the number of measurement points becomes [n / e]*[m / e], where the brackets indicate rounding operations.
[0060] If the sampling time of each measurement point is Δt, then one S e The time required for scanning is T e =[n / e]*[m / e]*Δt=[[Q*Δt / e] / e], where the brackets indicate rounding operations.
[0061] From the above, we can see that e*e times S can be completed within the time T of scanning using S1 mode. e Therefore, the measurement area is scanned e*e times S by enlarging the horizontal and vertical step sizes. e Scanning in this way can obtain e*e curves, which can be obtained by superposition and averaging to obtain a reference curve with a high signal-to-noise ratio. e In the scanning mode, the measuring point p i,j The measured value can be expressed as Fk (p i,j ), where i = 0, 1, 2, ..., [n / e]–1; j = 0, 1, 2, ..., [m / e]-1. After e*e scans are completed, the measurement point p i,j The accumulated measurement value on Taking the mean value, we can get F(p i,j )=S(p i,j ) / e 2 , that is, e*e times S e After scanning, the measuring point p i,j The measured value is F(p i,j ).
[0062] After obtaining the measurement values of each measurement point, if there are isolated bright spots or dark spots, that is, isolated singular points with excessive measurement value deviation, it can be considered that these isolated singular points are too greatly affected by noise and need to be eliminated. The noise elimination method is as follows:
[0063] For an isolated singular point p i,j , get the isolated singular point p i,j D×D neighborhood Ω, D is an integer greater than 1, and Ω is defined as: Ω={p x,y |x=i,i±1,…,i±D;y=j,j±1,…,j±D},such as Figure 4 As shown, taking D = 5 as an example, for the measured value {F(p i,y )|y=j,j±1,j±2} and the measured value of the jth column {F(p x,j )|x=i,i±1,i±2} are sorted and the value in the middle position is taken to replace the original F(p i,j ), as the noise elimination p i,y The measured value F(p i,j ).
[0064] After the above measurement process, a reference curve with a high signal-to-noise ratio can be obtained. Any of the following methods can also be used to obtain a reference curve with a high signal-to-noise ratio, such as the Fourier transform method (to remove the high-frequency part of the processing line in the Fourier space, to smooth the measurement line, and to obtain the derivative after smoothing), the derivative method, and the salt and pepper filtering method.
[0065] In one example, the reference curve may also be an artificially drawn non-measurement curve. Alternatively, a design drawing of the surface of the sample to be measured, such as a grating, may be set as the reference curve.
[0066] In one example, the reference curve and the measurement curve belong to the same measurement of the sample to be measured, the measurement time of the reference curve accounts for 10%-50% of the preset measurement time, and the measurement time of the measurement curve accounts for 90%-50% of the preset measurement time; the preset measurement time represents the processing time required for one measurement. That is, one measurement is divided into two times, and the same scanning area on the surface of the sample to be measured is measured twice. Let the total time of one measurement be T, the first measurement accounts for 10%-50% of the total time T, and this time period is used to measure and obtain the reference curve of the surface of the sample to be measured. At this time, the measurement time parameters of each measurement point can be set to be the same; the second measurement occupies the total time T minus the remaining measurement time after the first measurement, that is, it accounts for 90%-50% of the total time T, and this time period is used to measure and obtain the required measurement curve of the surface of the sample to be measured. Among them, between the pre-scan measurement and the actual scanning measurement, there is also a process of setting the measurement time parameters based on the weighting coefficient, which is generally short and can be ignored.
[0067] In one example, the curvature of each measurement point on the reference curve is used as the weighting coefficient of each measurement point. The curvature of the reference curve can be used to distinguish between flat areas and fine structure areas. Areas with small curvature represent flat areas, and areas with large curvature represent fine structure areas. The signal-to-noise ratio of the large curvature part, that is, the fine structure area, has a greater impact on the signal-to-noise ratio of the entire curve. Therefore, the weighting coefficient of the measurement point is set based on the curvature. The weighting coefficient of the measurement point with curvature can be set to be greater than 1, ensuring that more measurement time or measurement times can be allocated to the large curvature part in the subsequent measurement process, which can improve the signal-to-noise ratio of the fine structure area; for measurement points with smaller curvature, the weighting coefficients of these measurement points can be set to be less than or equal to 1, so that the measurement time of the measurement point remains unchanged or even reduced.
[0068] In another example, the gradient value of each measurement point on the reference curve is used as the weight coefficient of each measurement point. That is, on the reference curve, for each measurement point, the difference between the measurement value of the measurement point and the average of the sum of the measurement values of one or more adjacent measurement points is calculated, and the difference is used as the weight coefficient of the measurement point.
[0069] In this embodiment, for each measurement point, the weighting coefficient of each measurement point is in positive proportional relationship with the absolute value of the rate of change of the measurement value of the measurement point with the measurement value of one or more adjacent measurement points in the reference curve. Further, the weighting coefficient of the measurement point is in positive proportional relationship with the absolute value of the rate of change of the measurement value of each measurement point with the measurement value of one or more adjacent measurement points in the reference curve.
[0070] For each measuring point, the weighting coefficient of the measuring point is positively correlated with the average value of the measured values of the measuring points adjacent to the measuring point in the reference curve, including the closest, second closest and / or third closest measuring points, or the weighted average value and the absolute value of the rate of change of the measured value of the measuring point.
[0071] For each measuring point, the weighting coefficient of the measuring point is positively correlated with the average value of the measured values of the measuring points adjacent to the measuring point in the reference curve, including the closest, second closest and / or third closest and / or next closest measuring points, or the weighted average value and the absolute value of the rate of change of the measured value of the measuring point.
[0072] For each measuring point, the weighting coefficient of the measuring point is positively correlated by the average value or weighted average value of the measuring values of the measuring points adjacent to the measuring point in the reference curve, including the closest, second closest and / or third closest measuring points, and the absolute value of the rate of change of the measuring value of the measuring point. The positive relationship is a proportional relationship.
[0073] For each measuring point, the weighting coefficient of the measuring point is positively correlated by the average value or weighted average value of the measuring values of the measuring points adjacent to the measuring point in the reference curve, including the closest, second closest and / or third closest measuring points, and the absolute value of the rate of change of the measuring value of the measuring point. The positive relationship is a square relationship.
[0074] Step 102: adjusting a measurement time parameter of each measurement point based on a weight coefficient of each measurement point, where the measurement time parameter represents a processing time required to measure the measurement point.
[0075] Specifically, the weighting coefficient is related to the measurement time parameter of this measurement. The weighting coefficient can be expressed as the different speeds at which the measuring sensor passes through the measured measuring point or the different measurement time it stays for. A large weighting coefficient means that the measuring sensor stays at the measuring point for a long time, that is, the speed of passing the measuring point is slow. That is, the measurement time that the measuring sensor stays at each measuring point is proportional to the weighting coefficient, that is, the number of sampling times of the measuring sensor at each measuring point is proportional to the weighting coefficient.
[0076] For each measurement point, the current measurement time parameter of the measurement point is adjusted according to the weighting coefficient of the measurement point to obtain a new measurement time parameter, which is the measurement time parameter of the measurement point for this measurement; thus, the measurement time parameter of each measurement point in this measurement can be obtained, and the measurement time parameter represents the processing time required to measure the measurement point. Taking the measurement time parameter as the measurement time as an example, the larger the weighting coefficient, the longer the measurement time of this measurement obtained after adjustment; conversely, the smaller the weighting coefficient, the shorter the measurement time of this measurement obtained after adjustment. Among them, the current measurement time parameter of each measurement point can be the measurement time parameter of each measurement point indicated by the reference curve.
[0077] Taking any measurement point as an example, the current measurement time parameter of the measurement point is the measurement time T1, and the weighting coefficient of the measurement point is K1. The product of the measurement time T1 and the weighting coefficient K1 is calculated as the adjusted measurement time parameter of the measurement point, that is, as the measurement time parameter for this measurement at the measurement point, thereby extending or shortening the measurement time of the measurement point.
[0078] Step 103 , scanning and measuring the surface of the sample to be measured based on the adjusted measurement time-consuming parameters of each measurement point to obtain a measurement curve.
[0079] Specifically, a scanning measurement is performed on the surface of the sample to be measured, and each measurement point on the sample surface is measured according to the measurement time parameter adjusted in step 102, that is, different measurement times or measurement times are used for different measurement points. In this way, a larger measurement time parameter can be set in the fine structure area of the sample surface, and the measurement time in the fine structure area can be increased, which helps to improve the signal-to-noise ratio in the fine structure area. In the flat area of the sample surface, since it has less impact on the overall sample surface, a smaller measurement time parameter can be set to shorten the measurement time in the flat area.
[0080] After completing the current measurement and obtaining the measurement curve, you can reselect the reference curve to repeat the above process and start the next measurement.
[0081] In this embodiment, before measuring the sample to be tested, a reference curve of the sample to be tested is first obtained, and the reference curve represents the reference measurement values of multiple measurement points on the surface of the sample to be tested obtained by at least one measurement. Then, the reference curve can reflect the importance of each measurement point to a certain extent, and a larger weighting coefficient is set for the measurement points in the area with high importance, and a lower weighting coefficient is set for the measurement points in the area with low importance. Then, based on the weighting coefficient, the measurement time parameter of each measurement point is adjusted respectively, and the measurement time parameter represents the processing time required for measuring the measurement point. The measurement point with a large weighting coefficient is allocated more measurement time, and the measurement point with a small weighting coefficient is allocated less measurement time. Therefore, when the surface of the sample to be tested is scanned and measured based on the adjusted measurement time parameters of each measurement point, a higher signal-to-noise ratio can be maintained in the area with high importance, thereby improving the overall curve signal-to-noise ratio of the obtained measurement curve; that is, measurement data with a high signal-to-noise ratio can be measured and obtained within the same measurement time, that is, the measurement time can be reduced as much as possible under the condition of meeting certain signal-to-noise ratio requirements.
[0082] The second embodiment of the present invention relates to a scanning probe measurement method with a high signal-to-noise ratio. The main improvement of this embodiment compared to the first embodiment is that this embodiment provides a specific method for setting weighting coefficients and measuring fine structure areas on the sample surface.
[0083] The specific process of the high signal-to-noise ratio scanning probe measurement method of this embodiment is as follows: Figure 5 shown.
[0084] Step 201 obtains a reference curve of the sample to be tested and sets a weighting coefficient for each measuring point according to the reference curve, including the following sub-steps:
[0085] Sub-step 2011, obtaining a reference curve of the sample to be tested, and obtaining the fine structure area of the reference curve in each row and column based on the structure gradient value of each measurement point on the obtained reference curve.
[0086] Specifically, for each target sequence of the reference curve, the structure gradient values of at least 4 measurement points with amplitude jumps are selected from the target sequence with the current measurement point as the midpoint to find the median, and half of the median is selected as the local threshold; wherein the target sequence is a row or a column of the reference curve. It should be noted that this embodiment and subsequent embodiments are all described by taking this as an example, and it can also be that for each target sequence of the reference curve, the structure gradient values of at least 4 measurement points with amplitude jumps are selected from the target sequence with the current measurement point as the midpoint to find the average, and half of the average is selected as the local threshold.
[0087] Taking the current measurement point as the starting point, the first measurement point whose structural gradient value is greater than or equal to the local threshold is searched on the target sequence as the starting point of the starting interval of the found fine structure area, and the first measurement point whose structural gradient value is less than or equal to the local threshold is searched after the starting point of the starting interval as the end point of the starting interval.
[0088] Taking the next measurement point at the end point of the starting interval of the fine structure region as the starting point, the first measurement point whose structure gradient value is greater than or equal to the local threshold is found on the target sequence as the starting point of the ending interval of the fine structure region, and the first measurement point whose structure gradient value is less than or equal to the local threshold is found after the starting point of the ending interval as the ending point of the ending interval. It should be noted that the rows and columns mentioned in this embodiment and the following embodiments refer to the rows and columns of the array formed by the measurement points on the surface of the sample to be tested.
[0089] The following is a detailed description of how to find fine structure regions on the rows and columns of the reference curve.
[0090] After obtaining the reference curve, the function of fitting the reference curve can be expressed as F = F (X i ), for the reference curve, all measurement points are traversed in the horizontal direction and the vertical direction. Assuming that the reference curve includes X rows and Y columns, each row includes Y measurement points and each column includes X measurement points.
[0091] In the horizontal direction:
[0092] For a certain measurement point p i,j , whose measurement value is F(p i,j ), select the measurement point p i,j There are A measurement points in the horizontal neighborhood, namely, measurement point p i,j The first A measurement points and the last A measurement points, where A is a positive integer.
[0093] Please refer to Figure 6 , taking A=2 as an example, the selected measurement point p i,j The measurement points in the horizontal neighborhood include: measurement point p i,j-2 , measuring point p i,j-1 , measuring point p i,j+1 , measuring point p i,j+2 , the measured values are: F(p i,j-2 )、F(p i,j-1 )、F(p i,j+1 )、F(p i,j+2 ); The selected measurement point p i,j The measurement points in the vertical neighborhood include: measurement point p i-2,j , measuring point p i-1,j , measuring point p i+1,j , measuring point p i+2,j, the measured values are: F(p i-2,j )、F(p i-1,j )、F(p i+1,j )、F(p i+2,j ).
[0094] In the horizontal direction, the measured values F(p i,j-2 )、F(p i,j-1 )、F(p i,j )、F(p i,j+1 )、F(p i,j+2 ) select the maximum value F max1 (p i,j )=max{F(p i,j+k )} and the minimum value F min1 (p i,j )=min{F(p i,j+k )}, where k = -2, -1, 0, 1, 2, and then the measurement point p can be calculated i,j The structural gradient value G h (p i,j )=F max1 (p i,j )-F min1 (p i,j ).
[0095] By repeating the above process and traversing each row of the reference curve, the structural gradient value of each measuring point in each row in the horizontal direction can be calculated.
[0096] Then, based on the structural gradient values of the measurement points in each row, the fine structure region within each row is found.
[0097] For the i-th row, calculate the mean of the structural gradient values of the measurement points in the row. There are Y measurement points in the i-th row, and the mean of the structural gradient values of the measurement points in the row is
[0098] Take 1 / 2 of the mean value of the structural gradient of the i-th row as the reference upper limit R Hh , that is, R H =1 / 2G havgi , take 1 / 4 of the mean value of the structure gradient of the i-th row as the reference upper limit R Lh , that is, R Lh =1 / 4G havgi .
[0099] In the i-th row, the structural gradient will produce two amplitude jumps in a fine structure area; for a measurement point, if the structural gradient value G ≥ R Lh And the structural gradient value of the adjacent measurement points G≤R Lh , it is determined that an amplitude jump occurs at this measurement point.
[0100] Starting from the first measurement point in the i-th row, select at least four measurement points where the amplitude jump occurs, and obtain the median value G of the structural gradient values of these selected measurement points where the amplitude jump occurs midh , take the median value G midh 1 / 2 of the local threshold, that is, the local threshold T Gh =1 / 2G midh .
[0101] Then, the local threshold T Gh , first find the starting interval of the fine structure area from left to right in the i-th row, starting from the first measurement point p in the i-th row i,1 To begin, select the first structure with a gradient value greater than or equal to T Gh The measurement point p i,x1 , and the first structural gradient value after the measurement point is less than or equal to T Gh The measurement point p i,x2 , measuring point p i,x1 From start to measuring point p i,x2 The end is taken as the starting interval of a fine structure area. Then, the definition domain for setting the weighting coefficient is found in the starting interval, and the measurement point p with the largest structure gradient is first selected in the starting interval. i,xmax1 , x1≤xmax1≤x2, the measurement point p i,xmax1 As the candidate position for generating the initial mutation (maximum curvature) in the starting interval; then calculate the mean value T of the structural gradient value of all the measurement points in the starting interval GK1 , and in the starting interval [p i,x1 , p i,x2 ] from left to right to find the first structure whose gradient value is greater than or equal to the mean T GK1 The measurement point is denoted as measurement point p i,xk1 , determine the measurement point p i,xk1 With measuring point p i,xmax1 The relative position relationship between them, if the measurement point p i,xk1 With measuring point p i,xmax1 The same measuring point or at the measuring point p i,xmax1 On the right side, set the starting measurement point p of the domain i,xs1 is the measuring point p i,xmax1-1 ; If the measurement point p i,xk1 Located at measuring point p i,xmax1 , then set the starting measurement point p of the domain i,xs1 is the measuring point p i,xk1 , then starting from the starting measurement point of the domain, in the starting interval [p i,x1 , p i,x2 ] to find the first structural gradient value less than the mean T GK1The measurement point is denoted as measurement point p i,xe1 , thus we get the definition domain of setting weight coefficients in the starting interval [p i,xs1 , p i,xe1 ], the width of the domain is q hk , such as Figure 7 As shown, Figure 7 The vertical dimensions in the figure have nothing to do with reality.
[0102] After the starting interval of the current fine structure area is determined, the ending interval of the fine structure area is found after the starting interval, and the first structure gradient value greater than or equal to T is selected from left to right after the end measurement point of the starting interval. Gh The measurement point p i,x3 , and the first structural gradient value after the measurement point is less than or equal to T Gh The measurement point p i,x4 . Measuring point p i,x3 From start to measuring point p i,x4 The end is taken as the end interval of a fine structure area. Then, the definition domain for setting the weighting coefficient is found in the end interval, and the measurement point p with the largest structure gradient is first selected in the end interval. i,xmax2 , x3≤xmax2≤x4, the measurement point p i,xmax2 As the candidate position for the initial mutation (maximum curvature) in the end interval; then calculate the mean value T of the structural gradient value of all the measurement points in the end interval GK2 , and at the end interval [pi, x3 , p i,x4 ] from left to right to find the first structure whose gradient value is greater than or equal to the mean T GK2 The measurement point is denoted as measurement point p i,xk2 , determine the measurement point p i,xk2 With measuring point p i,xmax2 The relative position relationship between them, if the measurement point p i,xk2 With measuring point p i,xmax2 The same measuring point or at the measuring point p i,xmax2 On the right side, set the starting measurement point p of the domain i,xs2 is the measuring point p i,xmax2-1 ; If the measurement point p i,xk2 Located at measuring point p i,xmax2 , then set the starting measurement point p of the domain i,xs2 is the measuring point p i,xk2 , then starting from the starting measurement point of the domain, in the end interval [p i,x3 , p i,x4 ] to find the first structural gradient value less than the mean T GK2 The measurement point is denoted as measurement point p i,xe2, thus obtaining the definition domain of setting weighted coefficients in the end interval [p i,xs2 , p i,xe2 ], the width of the domain is q hw , such as Figure 7 As shown, Figure 7 The vertical dimensions in the figure have nothing to do with reality.
[0103] For a fine structure region, its overall width is Q h , in the domain of the starting interval [p i,xs1 , p i,xe1 ] and the domain of the end interval [p i,xs2 , p i,xe2 The area between ] is a non-mutation interval with small curvature and its width is Q hs .
[0104] After the current row is traversed based on the above process, the fine structure region contained in the row can be obtained. Similarly, after all rows are traversed, the fine structure region contained in each row can be obtained.
[0105] In the vertical direction:
[0106] For a certain measurement point p i,j , whose measurement value is F(p i,j ), select the measuring point p i,j B measurement points in the vertical neighborhood, namely, measurement point p i,j The upper B measurement points and the lower B measurement points, where B is a positive integer.
[0107] Please refer to Figure 6 , taking B=2 as an example, the selected measurement point p i,j The measurement points in the vertical neighborhood include: measurement point p i-2,j , measuring point p i-1,j , measuring point p i+1,j , measuring point p i+2,j , the measured values are: F(pi-2, j )、F(p i-1,j )、F(p i+1,j )、F(p i+2,j ).
[0108] In the vertical direction, the measured values F(p i-2,j )、F(p i-1,j )、F(p i,j )、F(pi+1, j )、F(p i+2,j ) select the maximum value F max2 (p i,j )=max{F(p i+k,j )} and the minimum value Fmin2 (p i,j )=min{F(p i+k,j )}, where k = -2, -1, 0, 1, 2, and then the measurement point p can be calculated i,j The structural gradient value G v (p i,j )=F max1 (p i,j )-F min2 (p i,j ).
[0109] By repeating the above process and traversing each column of the reference curve, the structural gradient value of each measurement point in each column in the vertical direction can be calculated.
[0110] Then, based on the structural gradient values of the measurement points in each column, the fine structure region within each column is found.
[0111] For the jth column, calculate the mean of the structural gradient values of the measurement points in the column. There are X measurement points in the jth column. The mean of the structural gradient values of the measurement points in the column is
[0112] Take 1 / 2 of the mean value of the structural gradient of the jth column as the reference upper limit R Hv , that is, R Hv =1 / 2G vavgj , take 1 / 4 of the mean value of the structure gradient of the jth column as the reference upper limit R Lv , that is, R Lv =1 / 4G vavgj .
[0113] In the jth column, the structural gradient will produce two amplitude jumps in a fine structure region; for a measurement point, if the structural gradient value G ≥ R Lv And the structural gradient value of the adjacent measurement points G≤R Lv , it is determined that an amplitude jump occurs at this measurement point.
[0114] Starting from the first measurement point in the jth column, select at least four measurement points where the amplitude jump occurs, and obtain the median value G of the structural gradient values of these selected measurement points where the amplitude jump occurs midv , take the median value G midv 1 / 2 of the local threshold, that is, the local threshold T Gv =1 / 2G midv .
[0115] Then, the local threshold T Gv , first find the starting interval of the fine structure area from top to bottom in the jth column, starting from the first measurement point p in the jth column 1,jTo begin, select the first structure with a gradient value greater than or equal to T Gv The measurement point p y1,j , and the first structural gradient value after the measurement point is less than or equal to T Gv The measurement point p y2,j , measuring point p y1,j From start to measuring point p y2,j The end is taken as the starting interval of a fine structure area. Then, the definition domain for setting the weighting coefficient is found in the starting interval, and the measurement point p with the largest structure gradient is first selected in the starting interval. ymax1,j , y1≤ymax1≤y 2, the measurement point p ymax1,j As the candidate position for generating the initial mutation (maximum curvature) in the starting interval; then calculate the mean value T of the structural gradient value of all the measurement points in the starting interval GV1 , and in the starting interval [p y1,j , p y2,j ] from top to bottom to find the first structure whose gradient value is greater than or equal to the mean T GV1 The measurement point is denoted as measurement point p xv1,j , determine the measurement point p xv1,j With measuring point p ymax1,j The relative position relationship between them, if the measurement point p xv1,j With measuring point p ymax1,j The same measuring point or at the measuring point p ymax1,j , then set the starting measurement point p of the definition domain ys1,j is the measuring point p ymax1-1,j ; If the measurement point p xv1,j Located at measuring point p ymax1,j , then set the starting measurement point p of the definition domain ys1,j is the measuring point p xv1,j , then starting from the starting measurement point of the domain, in the starting interval [p y1,j , p y2,j ] to find the first structural gradient value less than the mean T GV1 The measurement point is denoted as measurement point p ye1,j , thus we get the definition domain of setting weight coefficients in the starting interval [p ys1,j , p ye1,j ], the width of the domain is q vk , such as Figure 8 As shown, Figure 8 The vertical dimensions in the figure have nothing to do with reality.
[0116] After the starting interval of the current fine structure region is determined, the ending interval of the fine structure region is found after the starting interval, and the first structure gradient value greater than or equal to T is selected from top to bottom after the end measurement point of the starting interval. GvThe measurement point p y3,j , and the first structural gradient value after the measurement point is less than or equal to T Gv The measurement point p y4,j . Measuring point p y3,j From start to measuring point p y4,j The end is taken as the end interval of a fine structure area. Then, the definition domain for setting the weighting coefficient is found in the end interval, and the measurement point p with the largest structure gradient is first selected in the end interval. ymax2,j , y3≤y max2≤y4, the measurement point p ymax2,j As the candidate position for the initial mutation (maximum curvature) in the end interval; then calculate the mean value T of the structural gradient value of all the measurement points in the end interval Gv2 , and at the end interval [p y3,j , p y4,j ] from top to bottom to find the first structure whose gradient value is greater than or equal to the mean T Gv2 The measurement point is denoted as measurement point p yk2,j , determine the measurement point p yk2,j With measuring point p ymax2,j The relative position relationship between them, if the measurement point p yk2,j With measuring point p ymax2,j The same measuring point or at the measuring point p ymax2,j , then set the starting measurement point p of the definition domain ys2,j is the measuring point p ymax2-1,j ; If the measurement point p yk2,j Located at measuring point p ymax2,j , then set the starting measurement point p of the definition domain ys2,j is the measuring point p yk2,j , then starting from the starting measurement point of the domain, in the end interval [p y3,j , p y4,j ] to find the first structural gradient value less than the mean T Gv2 The measurement point is denoted as measurement point p ye2,j , thus obtaining the definition domain of setting weighted coefficients in the end interval [p ys2,j , p ye2,j ], the width of the domain is q vw , such as Figure 8 As shown, Figure 8 The vertical dimensions in the figure have nothing to do with reality.
[0117] For a fine structure region, its overall width is Qv, in the domain of the starting interval [p ys1,j , p ye1,j ] and the domain of the end interval [p ys2,j , p ye2,jThe area between ] is a non-mutation interval with small curvature and its width is Q vs .
[0118] After the current column is traversed based on the above process, the fine structure region contained in the column can be obtained. Similarly, after all columns are traversed, the fine structure region contained in each column can be obtained.
[0119] After traversing the horizontal and vertical directions, the distribution of the structural gradient in the horizontal and vertical directions, that is, the curve shape, can be obtained.
[0120] In one example, after respectively obtaining the fine structure regions of the reference curve in each row and each column, the method further includes: for the row fine structure regions and column fine structure regions covering the same measurement point, selecting fine structure regions with a small overall width from the row fine structure regions and the column fine structure regions for retention. That is, the fine structure region is optimized by synthesizing the structure gradient distribution (curve shape) in the horizontal direction and the vertical direction, thereby reducing the subsequent measurement time. The specific method is as follows:
[0121] Please refer to Fig. 9 , read the structure gradient distribution in the horizontal and vertical directions at the measurement point of the reference curve at the same time. If the position where the mutation occurs (the position with larger curvature) is also located in the fine structure area in the horizontal direction (whose overall width is Q h1 ) and the fine structure region in the vertical direction (whose overall width is Q v1 ), determine the extension direction of the mutation (larger), and retain the fine structure area with a smaller overall width min{Q h1 , Q v1}, removing another fine structure area reduces the repeated measurement part in the scanning measurement, and subsequently avoiding spending too much measurement time in the removed fine structure area, thereby improving the scanning measurement efficiency.
[0122] Sub-step 2012: for each measurement point that is not located in any fine structure region, a weighting coefficient of the measurement point is set to a constant.
[0123] Sub-step 2013, for each fine structure region, based on the reference measurement data of the reference curve in the fine structure region and the reference measurement data of the reference curve in the fine structure region after noise elimination, set the weight coefficient of the measurement point in the fine structure region.
[0124] Specifically, for each measurement point, when setting the weighting coefficient of the measurement point, if the measurement point is located outside the fine structure area, that is, the measurement point is not in any fine structure area, then the weighting coefficient of the measurement point is set to a constant, for example, the weighting coefficient is set to 1, then when subsequently setting the measurement time parameters of the measurement point, the measurement time parameters of the measurement point remain unchanged.
[0125] If the measurement point is located in a fine structure area, obtain the reference interval [p ik1,jk1 ,p ik2,jk2 ], if the fine structure area is in the horizontal direction, then the reference interval [p ik1,jk1 ,p ik2,jk2 ]=[p i,kl ,p i,kr ]+Q hs +[p i,wl ,p i,wr ], where [p i,kl ,p i,kr ] represents the domain within the starting interval of the fine structure region, Q hs represents the width of the interval without mutation in the fine structure region, [p i,wl ,p i,wr ] represents the domain within the end interval of the fine structure region. If the fine structure region is in the vertical direction, then the reference interval [p ik1,jk1 ,p ik2,jk2 ]=[p ku,j ,p kb,j ]+Q vs +[p wu,j ,p wb,j ], where [p ku,j ,p kb,j ] represents the domain within the starting interval of the fine structure region, Q vs represents the width of the interval without mutation in the fine structure region, [p wu,j ,p wb,j ] represents the domain within the end interval of the fine structure region. Then, the calculation is performed in the reference interval [p ik1,jk1 ,p ik2,jk2 ], calculate the reference measurement value F obtained by e*e measurements contained in the reference curve k (p i,j ) k (p i,j ) and the variance σ(p i,j ) σ =σ a (p i,j ) / σ(p i,j), the ratio r σ It can be used as the weighting coefficient of the measurement point. The weighting coefficients of the measurement points located in the same fine structure area are the same.
[0126] Based on the above process, the weighting coefficient of each measurement point can be set.
[0127] Step 203: adjusting the measurement time parameter of each measurement point based on the weight coefficient of each measurement point, where the measurement time parameter represents the processing time required to measure the measurement point.
[0128] Step 204 , scanning and measuring the surface of the sample to be measured based on the adjusted measurement time-consuming parameters of each measurement point to obtain a measurement curve.
[0129] Specifically, after the weighting coefficient of each measurement point is set, for the measurement points outside the fine structure area, the weighting coefficients of these measurement points are 1, and the measurement time consumption parameters of these measurement points remain unchanged.
[0130] For each measuring point in the fine structure area, taking the reference curve as an example, the surface of the sample to be measured is measured e*e times, then the number of measurements of the measuring point after adjustment is min{e*e,e*r σ}, r σ Indicates the number of measurements at this measuring point. The reference curve is used as the surface of the sample to be measured for a duration of T p Taking the measurement of as an example, the measurement time of the measurement point after adjustment is Among them, the measurement time parameters of the measurement points located in the same fine structure area are the same after adjustment.
[0131] When scanning and measuring the measuring points outside the fine structure area, each measuring point is measured based on the adjusted measurement time consumption parameters of each measuring point.
[0132] For scanning measurement in the fine structure area, for each fine structure area, the starting interval and the ending interval in the fine structure area are measured based on the adjusted measurement time parameters of the measurement points in the fine structure area; for each fine structure area, based on the measurement data of each measurement point in the starting interval and the ending interval in the fine structure area, the measurement values of each measurement point in the non-mutation interval of the fine structure area are interpolated.
[0133] Taking a fine structure region as an example, if the reference interval corresponding to the fine structure region [p ik1,jk1 ,p ik2,jk2 ]=[p i,kl ,p i,kr ]+Q hs +[p i,wl ,p i,wr ], then first calculate the interval [pi,kl ,p i,kr ], Q hs 、[p i,wl ,p i,wr ] is scaled and transformed into an interval [p I,kL ,p I,kR ], Q Hs 、[p I,wL ,p I,wR ], where, I=e*i, kL=e*kl, kR=e*kr, Q Hs =e*Q hs 、wL=e*wl、wR=e*wr、e 2 Indicates the number of measurements corresponding to the reference curve. This scan measurement is only in the interval [p I,kL ,p I,kR ] and [p I,wL ,p I,wR ], in the interval [p I,kL ,p I,kR ] and [p I,wL ,p I,wR ] The number of measurements Z in the internal scan measurement h is min{e*e,e*r σ}, that is, for the interval [p I,kL ,p I,kR ] and [p I,wL ,p I,wR Repeat the measurement points within Z h Scan the measurement again, take each measurement point at Z h The average value after the measurement is taken as the final measurement value of each measurement point. Hs No scanning measurement is performed within [p I,kL ,p I,kR ] and [p I,wL ,p I,wR The final measured values of each measuring point in the interval are interpolated by cubic spline for the non-mutation interval Q Hs Assign values to each measuring point.
[0134] If the reference interval corresponding to the fine structure area [p ik1,jk1 ,p ik2,jk2 ]=[p ku,j ,p kb,j ]+Q vs +[pwu, j ,p wb,j ], then first calculate the interval [p ku,j ,p kb,j ], Q vs 、[p wu,j ,p wb,j ] is scaled and transformed into an interval [pkU,J ,p kB,J ], Q Vs 、[p wU,J ,p wB,J ], where J=e*j, kU=e*ku, kB=e*kb, Q Vs =e*Q vs 、wU=e*wu、wB=e*wb、e 2 Indicates the number of measurements corresponding to the reference curve. This scan measurement is only in the interval [p kU,J ,p kB,J ] and [p wU,J ,p wB,J ], in the interval [p kU,J ,p kB,J ] and [p wU,J ,p wB,J ] The number of measurements Z in the internal scan measurement v is min{e*e,e*r σ}, that is, for the interval [p kU,J ,p kB,J ] and [p wU,J ,p wB,J Repeat the measurement points within Z v Scan the measurement again, take each measurement point at Z v The average value after the measurement is taken as the final measurement value of each measurement point. Vs No scanning measurement is performed within [p kU,J ,p kB,J ] and [p wU,J ,p wB,J The final measured values of each measuring point of ] are interpolated by cubic spline on the non-mutation interval Q Vs Assign values to each measuring point.
[0135] Repeat the above process to scan and measure the measurement points in all rows and all columns to obtain the final measurement value of each measurement point after this measurement.
[0136] The third embodiment of the present invention relates to a scanning probe measurement method with a high signal-to-noise ratio. The main improvement of this embodiment compared to the first embodiment is that the coordinates of the measurement points on the measurement curve obtained in this measurement are corrected in the lateral direction to solve the problem of coordinate drift. That is, when optimizing the signal-to-noise ratio of the measurement curve, it is not only necessary to optimize the measurement time or the number of measurements at each measurement point, but also necessary to optimize the coordinate drift between each measurement curve. This drift is particularly manifested in the coordinate movement between measurement curves of different times during repeated measurements. If the corresponding measurement points of these measurement curves are directly superimposed and averaged, the actual coordinate points will not correspond, resulting in a situation where the coordinate resolution is greatly reduced after superposition according to the coordinate axis, that is, the signal-to-noise ratio is reduced according to the coordinate axis; such as Fig.10 As shown in FIG. 1 , three measurement curves obtained by three measurements of the surface of the sample to be measured are schematically drawn, and the same measurement point X on the three measurement curves is marked. k It can be seen that there is a drift in the x-coordinate between the three measurement curves. The horizontal axis is the x-coordinate and the vertical axis is the measurement value Z. Fig.10 Only the x-coordinate is compared, not the measured value.
[0137] In this embodiment, there are two specific methods for correcting the coordinates of the measuring points in the horizontal direction:
[0138] The first method: The specific process of the high signal-to-noise ratio scanning probe measurement method is as follows Fig.11 shown.
[0139] Step 301, obtaining a reference curve of the sample to be tested, and setting a weighting coefficient of each measuring point according to the reference curve. This is substantially the same as step 101 in the first embodiment, and will not be described in detail here.
[0140] Step 302, based on the weight coefficient of each measuring point, adjust the measurement time parameter of each measuring point, the measurement time parameter represents the processing time required to measure the measuring point. It is substantially the same as step 102 in the first embodiment, and will not be described in detail here.
[0141] Step 303, based on the adjusted measurement time parameters of each measurement point, scan and measure the surface of the sample to be measured to obtain a measurement curve. This is substantially the same as step 103 in the first embodiment and will not be described in detail here.
[0142] Step 304: select a measurement curve from the N measurement curves as a reference curve, and obtain the coordinate displacement of each measurement curve other than the reference curve in the N measurement curves relative to the reference curve.
[0143] Step 305 : superimpose and average the N measurement curves based on the coordinate displacement of each measurement curve relative to the reference curve to obtain a final target measurement curve.
[0144] Specifically, this scanning measurement measured the surface of the sample N times, and obtained N measurement curves, each of which includes M measurement points. The Kth measurement curve can be expressed as U k =U k (x i ), where x i represents the i-th measurement point among the M measurement points on the K-th measurement curve.
[0145] An rth measurement curve is selected from the N measurement curves as a reference curve, and for each measurement curve except the reference curve among the N measurement curves, a correlation function between the measurement curve and the reference curve is obtained.
[0146] For the Kth measurement curve, its coordinate displacement relative to the rth measurement curve (reference curve) is △x Kr , the Kth measurement curve can be expressed as: U k =U k (x i +△x Kr ), calculate the correlation function of the square of the difference between the Kth measurement curve and the rth measurement curve (reference curve), and its expression is:
[0147]
[0148] Among them, the measurement points used are located in the middle section of the measurement curve, that is, the measurement values of the m measurement points in the middle measurement value area of the measurement curve are summed, the number of measurement points m used for summing the measurement curve is less than the number of measurement curve points M, and the measurement points used are located in the middle section of the measurement curve. If it is necessary to use a measurement value beyond the measurement point range in the calculation of the correlation function, a value can be set artificially, such as a fixed value, which can be the same measurement value as the nearest measurement point.
[0149] Then, the minimum value of the correlation function is obtained with the coordinate displacement as a variable, and the coordinate displacement corresponding to the first minimum value of the correlation function is used as the coordinate displacement of the measurement curve relative to the reference curve.
[0150] In the above formula, Cor kr (Δx kr ) with Δx kr Find the minimum value of the function for the variable, take the value away from Δx kr =0 for the first minimum value of Δx kr =Δx krmin , which is the optimal coordinate displacement of the Kth measurement curve relative to the rth measurement curve (reference curve). At this time, the Kth measurement curve can be expressed as: U k =Uk (x i +△x Krmin ).
[0151] By repeating the above process, the coordinate displacement of each curve except the reference curve in the N measurement curves relative to the rth measurement curve (reference curve) can be obtained. Based on the coordinate displacement of each measurement curve, each measurement curve is coordinate-displaced. At this time, each measurement curve is aligned with the rth measurement curve (reference curve). Then, the N measurement curves after coordinate displacement are superimposed and averaged to obtain the final target measurement curve. Its expression is:
[0152] Among them, K=1, 2, ..., N.
[0153] The second method: The specific process of the high signal-to-noise ratio scanning probe measurement method is as follows Fig.12 shown.
[0154] Step 401, obtaining a reference curve of the sample to be tested, and setting a weighting coefficient of each measuring point according to the reference curve. This is substantially the same as step 101 in the first embodiment, and will not be described in detail here.
[0155] Step 402, based on the weight coefficient of each measurement point, adjust the measurement time parameter of each measurement point, the measurement time parameter represents the processing time required to measure the measurement point. It is substantially the same as step 102 in the first embodiment, and will not be described again.
[0156] Step 403, based on the adjusted measurement time parameters of each measurement point, scan and measure the surface of the sample to be measured to obtain a measurement curve. This is substantially the same as step 103 in the first embodiment and will not be described in detail here.
[0157] Step 404, for the Lth measurement curve among the N measurement curves, based on the superimposed average curve of the first L-1 measurement curves after displacement, obtain the coordinate displacement of the Lth measurement curve; 1≤L≤N.
[0158] Step 405 : superimpose and average the N measurement curves based on the coordinate displacement of each measurement curve to obtain a final target measurement curve.
[0159] Specifically, this scanning measurement measured the surface of the sample N times, and obtained N measurement curves, each of which includes M measurement points. The Lth measurement curve can be expressed as U L =U L (x i ), where x i Represents the i-th measurement point among the M measurement points on the L-th measurement curve.
[0160] For the Lth measurement curve among the N measurement curves, obtain the correlation function between the Lth measurement curve and the superimposed average curve of the first L-1 measurement curves after displacement, then find the minimum value of the correlation function with the coordinate displacement as the variable, and use the coordinate displacement corresponding to the first minimum value of the correlation function as the coordinate displacement of the Lth measurement curve.
[0161] The specific instructions are as follows:
[0162] Starting from the first measurement curve among the N measurement curves, the coordinate displacement of the second measurement curve relative to the first measurement curve is expressed as △x 12 , then the second measurement curve can be expressed as: U2 = U2 (x i +△x 12 ), calculate the correlation function of the square of the difference between the second measurement curve and the first measurement curve, and its expression is:
[0163]
[0164] Among them, the measurement points used are located in the middle section of the measurement curve, that is, the measurement values of the m measurement points in the middle measurement value area of the measurement curve are summed, the number of measurement points m used for summing the measurement curve is less than the number of measurement curve points M, and the measurement points used are located in the middle section of the measurement curve. If it is necessary to use a measurement value beyond the measurement point range in the calculation of the correlation function, a value can be set artificially, such as a fixed value, which can be the same measurement value as the nearest measurement point.
[0165] In the above formula, Cor 12 (Δx 12 ) with Δx 12 Find the minimum value of the function for the variable, take the value away from Δx 12 =0 for the first minimum value of Δx 12 =Δx 12min , which is the optimal coordinate displacement of the second measurement curve relative to the first measurement curve. At this time, the second measurement curve can be expressed as:
[0166] U2=U2(x i +△x 12min ).
[0167] Then the superposition average of the first measurement curve and the second measurement curve after coordinate shift is calculated.
[0168] U2'(x i )=1 / 2(U2(x i +△x 12min )+U1(x i +△x 11min )), where △x11min =0.
[0169] The correlation function of calculating the square of the difference between the third measurement curve and the second measurement curve obtained by superposition and averaging is expressed as:
[0170]
[0171] In the above formula, Cor 23 (Δx 23 ) with Δx 23 Find the minimum value of the function for the variable, take the value away from Δx 23 =0 for the first minimum value of Δx 23 =Δx 23min , which is the optimal coordinate displacement of the third measurement curve relative to the second measurement curve obtained by superposition and averaging. At this time, the third measurement curve can be expressed as:
[0172] U3=U3(x i +△x 23min ).
[0173] Then, the superposition average of the first measurement curve and the second and third measurement curves after coordinate displacement is calculated.
[0174] U3'(x i )=1 / 3(U3(x i +△x 23min )+U2(x i +△x 12min )+U1(x i +△x 11min )),
[0175] where △x 11min =0.
[0176] The correlation function of calculating the square of the difference between the fourth measurement curve and the third measurement curve obtained by superposition and averaging is expressed as:
[0177]
[0178] In the above formula, Cor 34 (Δx 34 ) with Δx 34 Find the minimum value of the function for the variable, take the value away from Δx 34 =0 for the first minimum value of Δx 34 =Δx 34min , which is the optimal coordinate displacement of the fourth measurement curve relative to the third measurement curve obtained by superposition and averaging. At this time, the fourth measurement curve can be expressed as:
[0179] U4=U4(x i +△x 34min ).
[0180] Then, the superposition average of the first measurement curve and the second measurement curve, the third measurement curve, and the fourth measurement curve after coordinate displacement is calculated.
[0181] U4'(x i )=1 / 4(U4(x i +△x 34min )+U3(x i +△x 23min )+U2(x i +△x 12min )+U1(x i +△x 11min )), where △x 11min =0.
[0182] The above process is repeated until the superposition average of the first measurement curve and the second measurement curve, the third measurement curve, the fourth measurement curve, ..., the Nth measurement curve after coordinate displacement is calculated.
[0183] U N '(x i )=1 / N(U N (x i +△x NN-1min )+U N-1 (x i +△x N-1N-2min )+……+U3(x i +△x 23min )+U2(x i +△x 12min )+U1(x i +△x 11min )), where △x 11min =0.
[0184] The above U N '(x i ) is the final target measurement curve required.
[0185] In this embodiment, the problem of coordinate drift in each scanning measurement is solved, thereby improving the coordinate resolution and measurement signal-to-noise ratio of the scanning measurement.
[0186] The fourth embodiment of the present invention relates to a scanning probe system, including: a probe array, a carrying drive device and a controller, wherein the probe group is fixed on the carrying drive device, and each probe group includes at least two probes; the controller is respectively connected to each probe and the carrying drive device for communication. The probe is any one of the following types: a metal probe, a semiconductor probe, an insulator probe, an electron beam probe, a photon beam probe and an ion beam probe. The scanning probe system adopts the scanning probe measurement method with a high signal-to-noise ratio in this embodiment to achieve the splicing of at least one of the following situations: scanning probe microscopy imaging, scanning probe atomic force microscopy imaging, scanning probe electron beam lithography, and scanning probe ion beam positioning and ion implantation.
[0187] The controller is used to execute the scanning probe measurement method with high signal-to-noise ratio in any one of the first to third embodiments.
[0188] Since the first to third embodiments correspond to this embodiment, this embodiment can be implemented in conjunction with the first to third embodiments. The relevant technical details mentioned in the first to third embodiments are still valid in this embodiment, and the technical effects that can be achieved in the first to third embodiments can also be achieved in this embodiment. In order to reduce repetition, they are not repeated here. Accordingly, the relevant technical details mentioned in this embodiment can also be applied in the first to third embodiments.
[0189] Preferred embodiments of the present invention have been described above in detail, but it should be understood that aspects of the embodiments can be modified, if necessary, to employ aspects, features and concepts of the various patents, applications and publications to provide further embodiments.
[0190] These and other changes can be made to the embodiments in light of the above detailed description.In general, in the claims, the terms used should not be considered limited to the specific embodiments disclosed in the specification and the claims, but should be construed to include all possible embodiments along with the full scope of equivalents to which these claims are entitled.
Claims
1. A scanning probe measurement method with high signal-to-noise ratio, characterized in that: include: Obtaining a reference curve of the sample to be tested, and setting weighting coefficients of each of the measuring points according to the reference curve, wherein the reference curve represents reference measurement values of multiple measuring points on the surface of the sample to be tested obtained by at least one measurement; Based on the weighting coefficient of each of the measuring points, adjusting a measurement time parameter of each of the measuring points, wherein the measurement time parameter represents a processing time required to measure the measuring point; The surface of the sample to be measured is scanned and measured based on the adjusted measurement time-consuming parameters of each measurement point to obtain a measurement curve.
2. The scanning probe measurement method with high signal-to-noise ratio according to claim 1, characterized in that: The adjusting the measurement time parameters of each measurement point based on the weighted coefficient of each measurement point includes: For each of the measuring points, the product of the current measurement time parameter of the measuring point and the weighting coefficient of the measuring point is used as the adjusted measurement time parameter of the measuring point.
3. The scanning probe measurement method with high signal-to-noise ratio according to claim 1, characterized in that: The reference curve and the measurement curve belong to the same measurement of the sample to be measured, the measurement time of the reference curve accounts for 10%-50% of the preset measurement time, and the measurement time of the measurement curve accounts for 90%-50% of the preset measurement time; the preset measurement time represents the processing time required for one measurement.
4. The scanning probe measurement method with high signal-to-noise ratio according to claim 1, characterized in that: Before setting the weighting coefficient of each of the measuring points based on the reference curve of the sample to be measured, the method further includes: A pre-scan measurement is performed on the surface of the sample to be measured to obtain the reference curve, wherein during the pre-scan measurement process, the measurement time parameters of each measurement point are the same.
5. The scanning probe measurement method with high signal-to-noise ratio according to claim 1, characterized in that: The reference curve is a measurement curve obtained by scanning and measuring the surface of the sample to be measured last time.
6. The scanning probe measurement method with high signal-to-noise ratio according to claim 1, characterized in that: The measurement time-consuming parameter is the measurement time or the number of measurements.
7. The scanning probe measurement method with high signal-to-noise ratio according to claim 1, characterized in that: The step of setting a weighting coefficient of each of the measurement points according to the reference curve comprises: The curvature of each measuring point on the reference curve is used as a weighting coefficient of each measuring point.
8. The scanning probe measurement method with high signal-to-noise ratio according to claim 1, characterized in that: The step of setting a weighting coefficient of each of the measurement points according to the reference curve comprises: The gradient value of each measuring point on the reference curve is used as a weighting coefficient of each measuring point.
9. The scanning probe measurement method with high signal-to-noise ratio according to claim 1, characterized in that: The step of setting a weighting coefficient of each of the measurement points according to the reference curve comprises: Based on the obtained structural gradient value of each measuring point on the reference curve, respectively obtaining the fine structure area of the reference curve in each row and each column; For each of the measurement points that is not located in any of the fine structure regions, setting a weighting coefficient of the measurement point to a constant; For each of the fine structure regions, a weighting coefficient of the measurement point in the fine structure region is set based on reference measurement data of the reference curve in the fine structure region and reference measurement data of the reference curve in the fine structure region after noise elimination.
10. The scanning probe measurement method with high signal-to-noise ratio according to claim 9, characterized in that: Each of the fine structure regions includes: a starting interval, a non-mutation interval, and an ending interval; The scanning measurement of the surface of the sample to be measured based on the adjusted measurement time-consuming parameters of each of the measurement points to obtain a measurement curve includes: For each of the fine structure regions, based on the adjusted measurement time parameters of the measurement points in the fine structure region, measuring the start interval and the end interval in the fine structure region; For each of the fine structure regions, based on the measurement data of each of the measurement points in the starting interval and the ending interval in the fine structure region, the measurement values of each of the measurement points in the non-mutation interval of the fine structure region are interpolated.
11. The scanning probe measurement method with high signal-to-noise ratio according to claim 9, characterized in that: After respectively acquiring the fine structure regions of the reference curve in each row and each column, the method further includes: For the row fine structure region and the column fine structure region covering the same measuring point, a fine structure region with a small overall width is selected from the row fine structure region and the column fine structure region to be retained.
12. The scanning probe measurement method with high signal-to-noise ratio according to claim 9, characterized in that: The obtaining of the structural gradient value of each measuring point on the reference curve, respectively obtaining the fine structure area of each row and each column of the reference curve, comprises: For each target sequence of the reference curve, taking the current measurement point as the midpoint, sequentially select at least 4 structural gradient values of the measurement points with amplitude jumps from the target sequence to find the median, and select half of the median as the local threshold; wherein the target sequence is a row or a column of the reference curve; Taking the current measurement point as the starting point, searching the measurement point on the target sequence whose first structure gradient value is greater than or equal to the local threshold as the starting point of the starting interval of the found fine structure region, and searching the measurement point whose first structure gradient value is less than or equal to the local threshold after the starting point of the starting interval as the end point of the starting interval; Taking the next measurement point of the end point of the starting interval of the fine structure region found as the starting point, searching the measurement point whose first structure gradient value is greater than or equal to the local threshold on the target sequence as the starting point of the ending interval of the fine structure region, and searching the measurement point whose first structure gradient value is less than or equal to the local threshold after the starting point of the ending interval as the end point of the ending interval.
13. The scanning probe measurement method with high signal-to-noise ratio according to claim 1, characterized in that: The number of measurement curves is N, and each measurement curve includes M measurement points; After scanning and measuring the surface of the sample to be measured based on the adjusted measurement time parameters of each measurement point to obtain a measurement curve, the method further includes: Selecting one of the N measurement curves as a reference curve, and respectively obtaining the coordinate displacement of each of the N measurement curves except the reference curve relative to the reference curve; The N measurement curves are superimposed and averaged based on the coordinate displacement of each measurement curve relative to the reference curve to obtain a final target measurement curve.
14. The scanning probe measurement method with high signal-to-noise ratio according to claim 13, characterized in that: The step of respectively obtaining the coordinate displacement of each of the N measurement curves except the reference curve relative to the reference curve comprises: For each of the N measurement curves except the reference curve, obtaining a correlation function between the measurement curve and the reference curve; The minimum value of the correlation function is obtained by taking the coordinate displacement as a variable, and the coordinate displacement corresponding to the first minimum value of the correlation function is used as the coordinate displacement of the measurement curve relative to the reference curve.
15. The scanning probe measurement method with high signal-to-noise ratio according to claim 1, characterized in that: The number of measurement curves is N, and each measurement curve includes M measurement points; After scanning and measuring the surface of the sample to be measured based on the adjusted measurement time parameters of each measurement point to obtain a measurement curve, the method further includes: For the Lth measurement curve among the N measurement curves, based on the superimposed average curve of the first L-1 measurement curves after displacement, the coordinate displacement of the Lth measurement curve is obtained; 1≤L≤N; The N measurement curves are superimposed and averaged based on the coordinate displacement of each measurement curve to obtain a final target measurement curve.
16. The scanning probe measurement method with high signal-to-noise ratio according to claim 15, characterized in that: For the Lth measurement curve among the N measurement curves, obtaining the coordinate displacement of the Lth measurement curve based on the superposition and average of the first L-1 measurement curves after displacement includes: Obtaining a correlation function between the Lth measurement curve and the superimposed average curve of the first L-1 measurement curves after displacement; The minimum value of the correlation function is obtained by taking the coordinate displacement as a variable, and the coordinate displacement corresponding to the first minimum value of the correlation function is used as the coordinate displacement of the Lth measurement curve.
17. The scanning probe measurement method with high signal-to-noise ratio according to claim 1, characterized in that: The measurement time-consuming parameter is any one of the following: measurement time, sampling times, or a measurement parameter proportional to the measurement time.
18. A scanning probe system, characterized in that: A scanning probe measurement method for performing high signal-to-noise ratio according to any one of claims 1 to 17.
