Method and system for measuring depth of semiconductor high aspect ratio micro-size structure

Abstract

The invention discloses a semiconductor high aspect ratio micro-size structure depth measurement method and system, and relates to the technical field of semiconductors. The method comprises the following steps: dividing a view field area of a measurement system, and determining a position relationship between an area to be measured and an auxiliary area; based on the position relation between the to-be-detected area and the auxiliary area, the efficiency is improved by monitoring the light intensity fluctuation in the scanning process and automatically adjusting the step length. Meanwhile, state values are introduced in the method for step length adjustment, segmentation and matching of a plurality of local coherent signals are achieved, the accuracy of peak value positions is guaranteed, the scanning step length adjustment condition is recorded after collection is completed, and local coherent peaks of continuous sections of the state values are recognized and adjusted. And calculating a peak value position of the local coherent peak and a corresponding depth value by using an improved gravity center algorithm. According to the invention, on the premise that the precision is not reduced, the image acquisition amount can be effectively reduced, and the multi-peak ambiguity influence of interference signals can be eliminated at the same time.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and in particular to a method and system for measuring the depth of a semiconductor micro-size structure with a high aspect ratio. Background Art

[0002] High aspect ratio (HAR) structures at the micro-nano scale provide a breakthrough in three-dimensional heterogeneous integration for semiconductor technology, allowing researchers to integrate multiple structures in a limited space and effectively realize the micro-nanoization of functional structures. However, the small wire diameter and highly directional characteristics of the HAR structure pose challenges to the efficiency and accuracy of its vertical depth measurement. White light scanning interferometry technology is suitable for HAR structure depth measurement scenarios due to its non-contact and sub-nanometer longitudinal accuracy. The wide spectrum characteristics of the white light source determine the localization of the coherent fringe signal, especially in the long-distance scanning process, a considerable part of the out-of-focus signal acquisition has no interference fringes. Therefore, although the white light scanning interferometry technology shows reliable accuracy in the measurement of HAR structures, this mode of consistent step length scanning throughout the entire process limits the measurement efficiency.

[0003] In order to solve the problem of time-consuming redundant information in white light scanning interferometry technology, researchers added additional devices to the original interferometry system to assist in obtaining the coherence peak position, thereby driving the motion system to perform fine scanning near the target surface. However, these methods increase the complexity of the measurement system and increase the difficulty of integrating the measurement system and the threshold for implementation. With this in mind, some researchers integrated the autofocus algorithm into the optical system and used software-level algorithms to search for the optimal interference fringe area. However, these methods still cannot achieve a good balance between stability and efficiency. In terms of signal formation, the HAR structure is susceptible to numerical aperture limitations and interference from stray reflected light, resulting in the signal obtained by the white light scanning interferometry system containing two coherence peaks. In addition, the common coating structure in the semiconductor field will also affect the number of coherence peaks of the interference signal. This abnormal multi-peak signal will cause significant interference to the envelope extraction of the interference signal and the positioning of the coherence peak. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for measuring the depth of semiconductor high aspect ratio micro-size structures, which can effectively reduce the image acquisition amount without reducing the accuracy and eliminate the multi-peak ambiguity of the interference signal.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A method for measuring the depth of a semiconductor microstructure with a high aspect ratio, comprising:

[0007] Divide the field of view of the measurement system and determine the positional relationship between the area to be measured and the auxiliary area;

[0008] Based on the positional relationship between the area to be measured and the auxiliary area, the micro-displacement structure is driven to scan and collect with an initially set step length to obtain an initial adjacent frame microscopic interference pattern; the initial adjacent frame microscopic interference pattern is the first two interference patterns obtained through data collection;

[0009] Continue to drive the displacement structure to perform scanning acquisition, and calculate the light intensity fluctuation evaluation function of adjacent frames according to the n-1th interference pattern and the nth interference pattern; wherein n=[3, N]; N is the total number of interference patterns;

[0010] Determine whether to adjust the scanning step length according to the light intensity fluctuation evaluation function and the set light intensity fluctuation threshold, and record the scanning step length adjustment after the acquisition is completed;

[0011] Based on the scanning step size adjustment, identifying the local coherence peak of the continuous segment of the adjustment state value;

[0012] The improved centroid algorithm is used to calculate the peak position of the local coherence peak and the corresponding depth value.

[0013] Optionally, the auxiliary area is a peripheral surrounding area of ​​the area to be measured, and based on the field of view area and combined with the basic theory of partially coherent light interference, the following mathematical model equation is established for a general interference signal under the influence of thin film interference and high aspect ratio structural characteristics:

[0014]

[0015] Among them, S i (z,NA) represents the regional factor; Ω = ∑ i＝1,2,3 S i (z, NA), represents the regional factor and background light I involved in the formation of the local interference region signal bg Equal to the reference light path intensity I ref And the test light path intensity I surf and; z is the height of the surface of the topography to be measured; c is the speed of light; Δν and represent the spectral width and average frequency of the broadband light source, respectively, and NA represents the numerical aperture. When NA decreases or the z position moves up, the multi-peak characteristics of I(z) increase with S i The signal attenuates due to the decrease of (z, NA), and then degenerates into a resolved single-peak interference signal.

[0016] Due to the limitations of the experimental environment and the optical system, the interference fringes formed appear as band-like straight color fringes moving on the surface at the same height. Based on the mathematical model equation, the Gaussian envelope term in the expression I(z) It is maximum at the equal optical path z = 0 and decays rapidly with the increase of coherence distance. Therefore, the coordinates of points in the plane area of ​​the field of view meet the conditions:

[0017] max{d(∑1(x1,y1),∑2(x2,y2))}=FWHM

[0018] min{d(∑2(x2,y2),T(x,y))}=l max

[0019] Among them, ∑1(x1,y1),∑2(x2,y2) are points on the curves ∑1 and ∑2 respectively, and the maximum Euclidean distance between the curves is equal to the half-maximum full width FWHM of the Gaussian envelope curve of the interference signal; T(x,y) is a point in the measured area T, and the minimum Euclidean distance from the point on the curve ∑2 to the area T is equal to the maximum coherence length l of the wide-spectrum light source max .

[0020] Optionally, the calculation formula for setting the light intensity fluctuation threshold is:

[0021] η=k·std[I2(T1)-I1(T1)]

[0022] Wherein, std(·) is a mean square error function, which is multiplied by a coefficient k as a threshold η after calculating the light intensity fluctuation evaluation function of the initial adjacent frame microscopic interference images I1, I2 in the auxiliary area T1; k is set to 1.1-1.4.

[0023] Optionally, the expression of the light intensity fluctuation evaluation function is:

[0024] ψ i = std[I i (T1)-I i-1 (T1)],i=2···N

[0025] Where std(·) is the mean square error function, T1 represents the auxiliary area, I i represents the i-th interference pattern, I i-1 represents the i-1th interference pattern, and N is the total number of interference patterns.

[0026] Optionally, a light intensity fluctuation threshold is set according to the light intensity fluctuation evaluation function to determine whether to adjust the scanning step length, and the specific process includes:

[0027] The light intensity fluctuation evaluation function is compared with the set light intensity fluctuation threshold: when the light intensity fluctuation evaluation function is less than the set light intensity fluctuation threshold, the scanning step length is not adjusted, so that the micro-displacement structure maintains the initially set step length; when the light intensity fluctuation evaluation function is greater than or equal to the set light intensity fluctuation threshold, the scanning step length is adjusted to a small step length, and scanning is performed with the updated small step length; the small step length is a set value that is smaller than the initially set step length.

[0028] Optionally, the process of recording the scanning step length adjustment is as follows:

[0029] When the step size of the micro-displacement structure is adjusted during the scanning process, the monitoring state is recorded as a value of 0, and when the step size of the micro-displacement structure is not adjusted, the monitoring state is recorded as a random value within (0,1).

[0030] Optionally, the calculation formula of the improved center of gravity method is:

[0031] P=∑ m M(m)·m / ∑ m M(m)

[0032] M(m)=[I(m)-I(m-1)] 2 ,m=2···N s

[0033] Among them, N s represents the length of the local coherence peak signal segment, M(m) represents the square of the light intensity difference between adjacent frames, I(m) represents the mth interference pattern, and P represents the depth value corresponding to the peak position of the local coherence peak.

[0034] The present invention also provides a semiconductor high aspect ratio micro-size structure depth measurement system, which is used to apply the method as described above, including: a light source, an aperture, a beam expansion and collimation module, a beam splitter plate, an imaging lens group, a camera, an absorption plate, a micro-displacement structure and an interference objective lens; a wide-spectrum light beam is emitted by the light source, and the light beam size is controlled by an aperture with a size adjustment function before being emitted, and the light beam is modulated into a uniformly distributed light beam after passing through the formed beam expansion and collimation module, and after entering the beam splitter plate, the projected split light beam is directly irradiated on the absorption plate and absorbed, and the reflected measurement split light beam enters the interference objective lens, wherein the interference objective lens is a Michelson structure, and is clamped by a micro-displacement structure to achieve positioning and micro-displacement drive, and the measurement light beam carries the morphology information of the object to be measured under the action of the interference objective lens and returns to the optical path system, and finally passes through the imaging lens group and is collected by the camera target surface as an interference light intensity map.

[0035] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0036] The present invention discloses a semiconductor high aspect ratio micro-size structure depth measurement method and system, the method comprising: dividing the field of view area of ​​the measurement system, determining the positional relationship between the area to be measured and the auxiliary area; based on the positional relationship between the area to be measured and the auxiliary area, improving efficiency by monitoring the light intensity fluctuation during the scanning process and automatically adjusting the step length. At the same time, the state value is introduced into the inventive method to adjust the step length, so as to achieve the segmentation and matching of multiple local coherent signals, thereby ensuring the accuracy of the peak position, and recording the scanning step length adjustment after the acquisition is completed, and identifying the local coherent peak of the continuous segment of the adjusted state value, and using the improved center of gravity algorithm to calculate the peak position and corresponding depth value of the local coherent peak. The present invention can effectively reduce the image acquisition amount without reducing the accuracy, while eliminating the multi-peak ambiguity of the interference signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0038] Figure 1 It is a schematic flow chart of a method for measuring depth of a semiconductor micro-structure with a high aspect ratio according to the present invention;

[0039] Figure 2 Schematic diagram of light beam propagation of the HAR structure in this embodiment;

[0040] Figure 3 This is a schematic diagram of the location division of the area to be tested and the monitoring area in this embodiment;

[0041] Figure 4 Schematic diagram of the variable speed scanning method for optimizing the sampling frequency in this embodiment;

[0042] Figure 5 Schematic diagram of the identification method of multiple local coherence peaks in this embodiment; wherein, part (a) is a schematic diagram of the inter-frame light intensity variance fluctuation distribution obtained by uniform speed step scanning, part (b) is a schematic diagram of the step length monitoring state value during the speed change process, and part (c) is a schematic diagram of the interference light intensity signal of a certain pixel point;

[0043] Figure 6 is a simplified diagram of the optical path system used in this embodiment;

[0044] Figure 7Schematic diagrams of examples of three-dimensional reconstruction results under different algorithms in this embodiment; wherein, part (a) is a schematic diagram of the reconstruction results of the invention method, part (b) is a schematic diagram of the reconstruction results of the improved centroid method, part (c) is a schematic diagram of the Fourier transform method, part (d) is a schematic diagram of the wavelet transform method, and part (e) is a schematic diagram of the cross-sectional morphology corresponding to different algorithms.

[0045] Reference numerals:

[0046] 1. Light source; 2. Aperture; 3. Beam expansion and collimation module; 4. Beam splitter; 5. Imaging lens group; 6. Camera; 7. Absorption plate; 8. Micro-displacement structure; 9. Interference objective lens; 10. Object to be measured. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] The purpose of the present invention is to provide a method and system for measuring the depth of semiconductor high aspect ratio micro-size structures, which can effectively reduce the image acquisition amount without reducing the accuracy and eliminate the multi-peak ambiguity of the interference signal.

[0049] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0050] like Figure 1 As shown, the present invention provides a method for measuring the depth of a semiconductor micro-size structure with a high aspect ratio, comprising:

[0051] Step 100: Divide the field of view of the measurement system to determine the positional relationship between the area to be measured and the auxiliary area.

[0052] Step 200: Based on the positional relationship between the area to be measured and the auxiliary area, drive the micro-displacement structure to scan and collect with an initially set step length to obtain an initial adjacent frame microscopic interference pattern; the initial adjacent frame microscopic interference pattern is the first two interference patterns obtained through data collection.

[0053] Step 300: Continue to drive the displacement structure to perform scanning acquisition, and calculate the light intensity fluctuation evaluation function of adjacent frames according to the n-1th interference pattern and the nth interference pattern; wherein n=[3, N]; N is the total number of interference patterns.

[0054] Step 400: Determine whether to adjust the scanning step length according to the light intensity fluctuation evaluation function and the set light intensity fluctuation threshold, and record the scanning step length adjustment after the acquisition is completed.

[0055] Step 500: Based on the scanning step size adjustment, identify the local coherence peak of the continuous segment of the adjustment state value.

[0056] Step 600: Calculate the peak position and corresponding depth value of the local coherence peak using an improved centroid algorithm.

[0057] As a specific implementation method, the above steps are described in detail:

[0058] HAR structures play an important role in semiconductor fields such as manufacturing and designing microelectronic devices. As the size of semiconductor devices continues to shrink, especially entering the micro-nano process node, the depth value measurement of HAR structures has ushered in challenges. White light scanning interferometry technology shows certain advantages in the measurement of HAR structures, especially its non-contact, nanoscale resolution and large-range measurement characteristics, making it a tool for measuring and characterizing complex microstructures in the semiconductor field. However, this technology relies on scanning one position at a time and collecting a large number of interference images. When it comes to three-dimensional structures of large size or complex structures, a more detailed scanning step is required to ensure the accuracy of the measurement, resulting in the measurement efficiency of white light scanning interferometry technology being limited. In addition, under the influence of thin film multi-beam interference and numerical aperture, the interference signal at a single pixel of a small-diameter HAR structure is usually affected by the superposition of stray reflected light around the point to be measured, so that the coherent signal presents multiple interference peaks, which makes the application effect of the existing peak location algorithm worse. Therefore, the present invention mainly embodies a white light automatic speed-changing scanning interferometry system and a method for processing multi-peak ambiguous interference signals. It is intended to improve the measurement efficiency of white light scanning interferometry technology and the measurement accuracy of HAR structures. Specifically, the improved centroid method is selected as the peak location algorithm of the local coherence peak in the method of the present invention, and the slot and hole structures with different periods and line diameters are taken as the objects in the experiments of the specific implementation cases.

[0059] The specific invention method is as follows Figure 1 The method flow is shown in .

[0060] Limited by the relatively large focused spot size, when WLSI measures the bottom plane of the HAR microstructure, the stray reflection signal from the non-focal surface will interfere with the collected interference signal. Figure 2 As shown in the test beam propagation process in the white light interferometer system, the reflected light from the surface and bottom layers of the film is collected in the bottom signal of the HAR structure that propagates through the optical path to the CCD target surface. This microstructure coating and high aspect ratio feature cause the interference signal formed by the final superposition to have multiple coherent peaks, which in turn affects the desired height information solution.

[0061] Based on this, combined with the basic theory of partially coherent light interference, the following mathematical model equation is established for the multi-peak interference signal considering the influence of the film layer and HAR structure characteristics:

[0062]

[0063] Among them, S i (z,NA) represents the regional factor, the regional factor involved in the formation of the local interference regional signal and Ω = ∑ i＝1,2,3 S i (z,NA), background light I bg Equal to the reference light path intensity I ref And the test light path intensity I surf z is the height of the surface to be measured, c is the speed of light, Δν and Represent the spectral width and average frequency of the broadband light source respectively. Different from the conventional partially coherent light field distribution, the regional factor is defined in the formula, S i (z,NA) represents the weight of the multi-layer surface in the process of forming the interference signal. This term is mainly related to the numerical aperture NA and the scanning longitudinal position z. When NA decreases or the z position moves up (i.e., measuring the surface morphology), the multi-peak characteristics of I(z) decrease with S i The signal decays with the decrease of (z, NA) and then degenerates into a single-peak interference signal which can be directly solved by the algorithm.

[0064] Furthermore, considering the oscillating light intensity distribution near the peak of the broad light interference signal, a surrounding area is divided around the area to be measured to monitor the equal optical path working distance position during the scanning process. The light intensity fluctuation feature in the area is selected to realize the local interference signal judgment and sampling fine adjustment. Figure 3 The detailed parameter basis for setting the monitoring area is shown in the figure. The purple area is the HAR structure scale information that needs to be obtained, and the area surrounded by the two yellow dotted lines is the monitoring area set to evaluate the light intensity fluctuation. The height distribution of the HAR structure on the wafer is a discontinuous step type, and the interference fringes appear as strip-shaped straight color fringes moving on the surface at the same height. From the mathematical model, it can be seen that the Gaussian envelope term in the formula It is maximum at the equal optical path z=0 and decays rapidly with the increase of coherence distance.

[0065] Based on this, the coordinates of points in the plane area should satisfy the conditions:

[0066] max{d(∑1(x1,y1),∑2(x2,y2))}=FWHM

[0067] min{d(∑2(x2,y2),T(x,y))}=l max

[0068] Among them, ∑1(x1,y1),∑2(x2,y2) are points on the curves ∑1 and ∑2 respectively, and the maximum Euclidean distance between the curves is equal to the full width at half maximum (FWHM) of the Gaussian envelope curve of the interference signal. T(x,y) is a point in the measured area T, and the minimum Euclidean distance from the point on the curve ∑2 to the area T is equal to the maximum coherence length l of the wide-spectrum light source max . The first distance condition in the formula maximizes the inclusion of the discrete features of the light intensity near the interference peak into area T1 to capture the light intensity fluctuation. The second distance condition eliminates the redundant information of the out-of-focus area to the greatest extent. When the interference fringes enter area T1 to trigger the monitoring state, the area to be tested T just enters the coherent range and the PZT step distance is reduced at the same time to obtain refined sampling data. When the interference fringes leave area T1 to release the monitoring state, the area to be tested T leaves the coherent range and adjusts the large step sampling frequency of the PZT to reduce data redundancy.

[0069] Furthermore, the trigger conditions of area T1 are set as follows: First, the operation starts with the set PZT initial step size, and the initial two images I1 and I2 are collected. The light intensity fluctuation evaluation function ψ is established based on the mean square error of the light intensity in area T1. i , whose expression is:

[0070] ψ i = std[I i (T1)-I i-1 (T1)],i=2···N

[0071] Among them, std(·) is the mean square error function. After calculating the light intensity fluctuation evaluation function ψ2 corresponding to I1 and I2, it is multiplied by the coefficient k as the threshold η. The value of k is between 1.1 and 1.4. If k is too small, it means that interference factors such as noise and stray light are likely to cause misjudgment of the monitoring of area T1 and collect useless interference information. If k is too large, it means that area T1 is not easily triggered for monitoring, thus missing the information of some coherent areas.

[0072] Furthermore, after the threshold is determined, the PZT is driven to continue to acquire the image I3, ψ3 is calculated and compared with the threshold η, and when ψ3 is less than η, the PZT maintains the original large step, and when ψ3 is greater than or equal to η, the PZT reduces the step size to perform a refined scan of the next position. This process is repeated until the Nth step completes the white light variable step size scanning interferometry process.

[0073] Based on the established mathematical model, the interference signal obtained after the PZT scanning is completed does not have a maximum light intensity at only one equal optical path. The light beam with a larger aperture angle will receive part of the reflected light at the local edge to be measured. The reflected interference light beam on this part of the surface will produce redundant local interference peaks in the signal, which will affect the accuracy of solving the equal optical path position.

[0074] Furthermore, in order to correctly separate the desired local interference region signal from the multi-phase coherence peak signal, a monitoring state value is introduced in the step adjustment link to mark the coherence peak position. Figure 4 As shown, the sampling rhythm is optimized by the variable step scanning interference system, and the interference signal area is adjusted to fine sampling. For the defocused area, only sparse sampling is used to monitor the light intensity fluctuation of area T1. When the step size of the PZT is adjusted during the scanning process, a monitoring state is recorded in the form of a 0 value, and when the PZT step size is not adjusted, the monitoring state is recorded as a random value within (0,1).

[0075] Based on this, in order to verify the effectiveness of this setting, Figure 5 Part (a) shows the light intensity fluctuation value of area T1 during the traditional constant step scanning process. Every time the area to be measured T enters the coherence range, ψ on T1 i The value will increase rapidly accordingly and determine the assignment of the monitoring status.

[0076] Correspondingly, when the variable step-size scanning method is adopted, we get Figure 5 The monitoring state value of part (b) is plotted on Figure 5 In this case, the invention method designs an algorithm for extracting continuous values ​​to realize the identification of continuous segments of the monitoring state value, and realizes the matching and positioning of multiple local coherent peaks according to the length and distribution of the continuous segments. Figure 5 In the continuous segments ①②③④ of part (c), area T1 monitors four local interference signals, and it can be determined that segments ①②③④ correspond to thin film surface interference, thin film bottom layer interference, thin film multi-beam interference and structure bottom interference respectively.

[0077] Furthermore, in order to ensure that the computer can select the corresponding local interference signal to solve the peak position when reconstructing the surface and bottom layer of the HAR structure, it is necessary to distinguish the different surfaces to be measured within T. The microscopic images collected by the scanning interferometer system are significantly affected by stray light and noise. If the surface and bottom layer areas of the HAR structure are directly binarized and then morphologically processed, the final effect is not ideal. Therefore, the maximum light intensity ratio between area ① and area ④ in the signal at each point in T can be calculated, and then binarized to obtain a clear edge contour. In the inventive method, the local coherence peaks of different areas are selected according to the top and bottom of the structure, and then the improved center of gravity method is used to quickly and effectively determine the equal optical path position P. The calculation process is shown in the following formula:

[0078] P=∑ m M(m)·m / ∑ m M(m)

[0079] M(m)=[I(m)-I(m-1)] 2 ,m=2···N s

[0080] Among them, N s represents the length of the local coherence peak signal segment, and M(m) represents the square of the light intensity difference between adjacent frames, which is brought into the centroid algorithm to eliminate the influence of background light intensity.

[0081] Comparative Example 1:

[0082] In order to verify the performance of the invention method in actual measurement scenarios and demonstrate its superiority, the applicant conducted experimental verification. Figure 6 As shown in the figure, a white light vertical scanning interferometry system was built in the comparative example, and supporting software with variable step length function was developed. This system uses a Michelson-type 5x micro-interference objective lens, and a white light LED lamp is selected as the broad spectrum light source. After the light beam propagates through the optical system, the interference image is collected by a color camera with a resolution of 2592×1944. The scanning process of the objective lens in the vertical direction is driven by a PZT controller with a minimum resolution step of 14nm, and the object to be measured is placed on a horizontal translation stage driven by a two-axis ball screw. Unless otherwise specified in the example, the collected interference patterns are all obtained at a scanning step of 20nm.

[0083] The measurement area of ​​this group of comparative experiments is a groove structure with a width of 9μm and a period of 27μm on the wafer. The reference measurement depth value obtained by the manual calibration method is 35.2264μm. Subsequently, the interference data was collected in a variable step scanning mode under the same experimental environment, in which the large step length of the coarse scan was set to 320nm and the small step length of the fine scan was set to 20nm. The height line difference between the upper and lower surfaces of the morphology reconstructed by the proposed method was used as the variable step scanning measurement depth result of 35.2458μm, which only deviated by 0.0551% from the manual calibration result. At the same time, the number of image frames collected was 24.4% of the uniform step scanning method. The example results show that the proposed invention method can greatly improve the measurement efficiency while ensuring accuracy.

[0084] Comparative Example 2:

[0085] In order to demonstrate the applicability of the inventive method to different deep groove sizes and deep hole units, a comparative example 2 was conducted. The properties of the measured objects and the measurement results in this group of experiments are presented in Table 1. The measurement depth becomes shallower as the structure size decreases, which is consistent with the result of the deep silicon etching process error. In addition, the data in Table 1 show that the proposed method not only achieves stable morphology reconstruction results in different periods, different widths and deep hole structures, but also has an average compression of 75.27% in the number of collected image frames, which improves the measurement efficiency by about 4 times on this group of microstructure samples.

[0086] Table 1 Measurement object attributes and measurement results

[0087]

[0088] Comparative Example 3:

[0089] In comparative example 3, a set of interference signal processing algorithm comparisons are considered. In order to verify the processing effect of multi-peak signals, the proposed invention method, the improved centroid method (Ai), the Fourier transform method (FT) and the wavelet transform method (WT) are used to reconstruct the three-dimensional morphology of the sample to be tested. The reconstructed groove structure is shown in FIG. Figure 7 As shown, Figure 7 Parts (a) to (d) use the same range of color bars, and different colors correspond to different height values. From a three-dimensional perspective, the longitudinal scale of the reconstructed morphology by the proposed invention method is the largest and closest to the reference value, while the depth results obtained by the other three types of algorithms are shallow. Figure 7 In part (e), directly using the Ai algorithm to process the multi-peak signal results in significant errors in morphology, while the FT and WT solution results result in obvious depth loss due to interference from the oxide film layer.

[0090] In addition, the time consumed for the reconstruction of the three-dimensional morphology of the local area is compared in the example, and all the image signal processing processes are run on a desktop computer with a 2.50GHz i5-12400FCPU in Matlab software. The proposed invention method takes a total of 2.20 seconds from image input to reconstruction completion, second only to the 1.28 seconds of the Ai method, while the FT and WT algorithms take 3.78 seconds and 372.33 seconds respectively due to the frequency domain transformation involved. The reason why the proposed method takes a little longer than the Ai method is that there is an additional region segmentation process, but this process is the key to solving the problem of incorrect restoration of the morphology by the Ai algorithm. Therefore, from a general point of view, the proposed invention method not only reduces most of the time in variable-speed scanning interferometry measurement, but also has the advantage of short algorithm processing time. The most critical thing is that it achieves accurate measurement results of white light scanning interferometry.

[0091] The above takes a certain measurement object as an example to introduce in detail a white light automatic variable speed scanning interference system and a method for processing multi-peak ambiguous interference signals of the present invention. In the content of the present invention, a piezoelectric ceramic driver is used as a mechanical mechanism to generate micro-displacement, and an LED light source is used as a wide-spectrum light source for illumination. This specific interference measurement system is intended to help relevant industry insiders better understand the idea and execution process of the method of the present invention.

[0092] At the same time, for specific application measurement objects and algorithm processing flows, for example, the area to be measured and the auxiliary detection area are set to other shapes or divided according to other distance limits, and the improved centroid peak positioning algorithm of the local coherent signal is replaced by the existing peak positioning algorithm. Improvements based on the above similar methods or application classes should be deemed to be within the scope covered by the method of the present invention, and these changes should all fall within the scope of protection of the claims attached to the present invention.

[0093] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0094] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for measuring the depth of a semiconductor microstructure with a high aspect ratio, characterized in that: include: Divide the field of view of the measurement system and determine the positional relationship between the area to be measured and the auxiliary area; Based on the positional relationship between the area to be measured and the auxiliary area, the micro-displacement structure is driven to scan and collect with an initially set step length to obtain an initial adjacent frame microscopic interference pattern; the initial adjacent frame microscopic interference pattern is the first two interference patterns obtained through data collection; Continue to drive the displacement structure to perform scanning acquisition, and calculate the light intensity fluctuation evaluation function of adjacent frames according to the n-1th interference pattern and the nth interference pattern; wherein n=[3, N]; N is the total number of interference patterns; Determine whether to adjust the scanning step length according to the light intensity fluctuation evaluation function and the set light intensity fluctuation threshold, and record the scanning step length adjustment after the acquisition is completed; Based on the scanning step size adjustment, identifying the local coherence peak of the continuous segment of the adjustment state value; The improved centroid algorithm is used to calculate the peak position of the local coherence peak and the corresponding depth value.

2. The method for measuring the depth of a semiconductor high aspect ratio micro-scale structure according to claim 1, characterized in that: The auxiliary area is the peripheral surrounding area of ​​the area to be measured, and based on the field of view area and combined with the basic theory of partially coherent light interference, the following mathematical model equation is established for the general interference signal under the influence of thin film interference and high aspect ratio structural characteristics: Among them, S i (z,NA) represents the regional factor; Ω = ∑ i＝1,2,3 S i (z, NA), represents the regional factor and background light I involved in the formation of the local interference region signal bg Equal to the reference light path intensity I ref And the test light path intensity I surf and; z is the height of the surface of the topography to be measured; c is the speed of light; Δν and represent the spectral width and average frequency of the broadband light source, respectively, and NA represents the numerical aperture. When NA decreases or the z position moves up, the multi-peak characteristics of I(z) increase with S i The signal attenuates due to the decrease of (z, NA), and then degenerates into a resolved single-peak interference signal. Due to the limitations of the experimental environment and the optical system, the interference fringes formed appear as band-like straight color fringes moving on the surface at the same height. Based on the mathematical model equation, the Gaussian envelope term in the expression I(z) It is maximum at the equal optical path z = 0 and decays rapidly with the increase of coherence distance. Therefore, the coordinates of points in the plane area of ​​the field of view meet the conditions: max{d(∑1(x1,y1),∑2(x2,y2))}=FWHM Among them, ∑1(x1,y1),∑2(x2,y2) are points on the curves ∑1 and ∑2 respectively, and the maximum Euclidean distance between the curves is equal to the half-maximum full width FWHM of the Gaussian envelope curve of the interference signal; T(x,y) is a point in the measured area T, and the minimum Euclidean distance from the point on the curve ∑2 to the area T is equal to the maximum coherence length l of the wide-spectrum light source max .

3. The method for measuring the depth of a semiconductor high aspect ratio micro-scale structure according to claim 1, characterized in that: The calculation formula for setting the light intensity fluctuation threshold is: η=k·std[I2(T1)-I1(T1)] Wherein, std(·) is a mean square error function, which is multiplied by a coefficient k as a threshold η after calculating the light intensity fluctuation evaluation function of the initial adjacent frame microscopic interference images I1, I2 in the auxiliary area T1; k is set to 1.1-1.

4.

4. The method for measuring the depth of a semiconductor high aspect ratio micro-scale structure according to claim 1, characterized in that: The expression of the light intensity fluctuation evaluation function is: ψ i =std[I i (T1)-I i-1 (T1)],i=2···N Where std(·) is the mean square error function, T1 represents the auxiliary area, I i represents the i-th interference pattern, I i-1 represents the i-1th interference pattern, and N is the total number of interference patterns.

5. The method for measuring the depth of a semiconductor high aspect ratio micro-scale structure according to claim 1, characterized in that: The light intensity fluctuation threshold is set according to the light intensity fluctuation evaluation function to determine whether to adjust the scanning step length. The specific process includes: The light intensity fluctuation evaluation function is compared with the set light intensity fluctuation threshold: when the light intensity fluctuation evaluation function is less than the set light intensity fluctuation threshold, the scanning step length is not adjusted, so that the micro-displacement structure maintains the initially set step length; when the light intensity fluctuation evaluation function is greater than or equal to the set light intensity fluctuation threshold, the scanning step length is adjusted to a small step length, and scanning is performed with the updated small step length; the small step length is a set value that is smaller than the initially set step length.

6. The method for measuring the depth of a semiconductor high aspect ratio micro-scale structure according to claim 1, characterized in that: The recording process of the scanning step length adjustment is as follows: When the step size of the micro-displacement structure is adjusted during the scanning process, the monitoring state is recorded as a value of 0, and when the step size of the micro-displacement structure is not adjusted, the monitoring state is recorded as a random value within (0,1).

7. The method for measuring the depth of a semiconductor high aspect ratio micro-scale structure according to claim 1, characterized in that: The calculation formula of the improved center of gravity method is: P=∑ m M(m)·m / ∑ m M(m) M(m)=[I(m)-I(m-1)] 2 ,m=2···N s Among them, N s represents the length of the local coherence peak signal segment, M(m) represents the square of the intensity difference between adjacent frames, I(m) represents the mth interference pattern, and P represents the depth value corresponding to the peak position of the local coherence peak.

8. A semiconductor high aspect ratio micro-structure depth measurement system, used for applying the method according to any one of claims 1 to 7, characterized in that: include: Light source, aperture, beam expansion and collimation module, beam splitter, imaging lens group, camera, absorption plate, micro-displacement structure and interference objective lens; A broad-spectrum light beam is emitted from the light source, and then it is emitted after the size of the beam is controlled by an aperture with an adjustable size function. After passing through the formed beam expansion and collimation module, it is modulated into a uniformly distributed light beam. After entering the beam splitter plate, the projected split light is directly irradiated on the absorption plate and absorbed. The reflected measurement split light enters the interference objective lens, where the interference objective lens is a Michelson structure and is clamped by a micro-displacement structure to achieve positioning and micro-displacement drive. The measurement light beam carries the morphology information of the object to be measured under the action of the interference objective lens and returns to the optical path system. Finally, after passing through the imaging lens group, it is collected by the camera target surface as an interference light intensity map.

Images