An OCD measurement method based on 3D model relationship

Abstract

The present invention provides an OCD measurement method based on a 3D model relationship, comprising: collecting planar measurement results of a 3D measurement object at different heights using a microscope, each planar measurement result corresponding to a mapping number; establishing a 3D model relationship among the planar measurement results, the mapping number, and the 3D model of the 3D measurement object; and calculating the optical critical dimensions of the 3D measurement object based on the relationship among the planar measurement results, the mapping number, and the 3D model of the 3D measurement object. The key points of the present invention are: (1) simply using 2D imaging at different heights to construct a 3D model; (2) utilizing the different performance characteristics of the measurement object under G and B color light for data fusion; and (3) utilizing the convergence and fluctuation patterns of the data to infer the critical dimensions of interest. The advantages of the present invention are: verified by a large amount of data, the method has good stability and accuracy, and has a high degree of matching with CDSEM.

Description

Technical Field

[0001] The present invention relates to the technical field of wafers and display panels, and in particular to an OCD measurement method based on a 3D model relationship. Background Art

[0002] In complex process environments with multi-layer stacking, 3D shapes, and varying film thicknesses, the following problems arise when using planar measurement methods:

[0003] (1) Plane microscopes cannot stably focus on a certain layer or position;

[0004] (2) At sensitive locations, the focus height changes slightly, resulting in large differences in measured values;

[0005] (3) It is impossible to determine at which height the measurement is correct.

[0006] The above problems mean that the 2D measurement method has poor measurement stability in 3D environments, cannot guarantee accuracy, and has low reliability.

[0007] With the development of the semiconductor and display panel industries, integrated circuit manufacturing processes are constantly improving and becoming increasingly complex. To ensure process quality, precise monitoring is required during the manufacturing process to control quality and make timely process adjustments. OCD (Optical Critical Dimension) measurement is a very important monitoring method. For processes with relatively thin or flat films, 2D measurement methods are generally used. However, for processes with thicker films or 3D shapes, 2D measurement methods cannot meet the requirements and cannot provide accurate and stable measurement. Therefore, more accurate, stable, and efficient 3D measurement methods are needed.

[0008] The existing technologies generally include: (1) using multi-dimensional acquisition, which requires the integration of multiple collectors, data acquisition at different angles and orientations, and complex fitting of the data. Only when a complete three-dimensional stereo image is constructed can OCD measurement be performed. This method has a complex structure and is relatively cumbersome; (2) SEM REVIEW machines can measure more accurately, but generally only support the measurement requirements of one shot and one point, and cannot perform large-scale measurement. In addition, the measurement speed is relatively slow. This method needs to be improved before it can be applied, and the price is relatively expensive.

[0009] Existing 2D and 3D measurement methods have the following shortcomings when facing practical engineering problems:

[0010] 1. In the complex 3D process environment with multi-layer stacking, 3D shapes, and varying film thicknesses, 2D measurement methods cannot accurately measure OCD and have poor stability.

[0011] 2. The existing 3D measurement solutions have complex structures, cumbersome methods, and high costs;

[0012] 3. Among the existing 3D measurement methods, the measurement speed is relatively slow, and the measurement precision, accuracy and stability need to be verified. Summary of the Invention

[0013] The purpose of the present invention is to use 3D technology to simply, cost-effectively and efficiently solve the problem of unstable and unreliable measurement results faced by existing 2D measurement methods in 3D measurement environments, while being able to accurately and stably measure the OCD of 3D models.

[0014] The key points of the present invention are: (1) simply using 2D imaging at different heights to construct a 3D model; (2) using the different performance characteristics of the measured object under G and B color light to fuse data; and (3) using the convergence and fluctuation patterns of the data to infer the key dimensions of interest.

[0015] Specifically, the present invention provides an OCD measurement method based on a 3D model relationship, comprising:

[0016] Collect the plane measurement results of the 3D measurement object at different heights of the microscope. Each plane measurement result corresponds to a map number.

[0017] Establish the relationship between plane measurement results, drawing sequence numbers and 3D model of 3D measurement objects;

[0018] The optical critical dimension of the 3D measurement object is calculated according to the relationship between the plane measurement result, the image acquisition sequence number and the 3D model of the 3D measurement object.

[0019] Furthermore, the acquisition of plane measurement results of the 3D measurement object at different heights by the microscope, wherein each plane measurement result corresponds to a drawing sequence number, includes the following steps:

[0020] The wafer is placed on the motion platform, and the microscope focuses on the measurement point. Starting from the focus position, the microscope is moved vertically up or down along the Z axis to the height a or b position as the preparation position;

[0021] The microscope moves along the Z axis from position a\b to position b\a, while simultaneously starting image acquisition and plane calculation;

[0022] After reaching the designated position, the microscope stops moving and the image acquisition and calculation are turned off.

[0023] Furthermore, the acquisition of plane measurement results of the 3D measurement object at different heights by the microscope, wherein each plane measurement result corresponds to a drawing sequence number, includes the following steps:

[0024] (1) The microscope performs initial focusing, with the focus position O as the starting position;

[0025] (2) The microscope moves up and down along the Z axis by a preset distance to the a / b preparation position;

[0026] (3) The microscope moves downward or upward along the Z axis to b or a;

[0027] (4) When moving from a to b, switch to G color light, and take a picture and calculate the key dimensions of the plane every time it moves a step;

[0028] (5) When b moves to a, switch to color light B. Each time it moves a step, it takes a picture and calculates the key dimensions of the plane.

[0029] (6) Store the results into an array by number and save the image.

[0030] Furthermore, the 3D model relationship is as follows:

[0031] The microscope moves from position a\b to position b\a along the Z axis. Under G color light, the plane measurement results experience a process of divergence, convergence, and divergence. In the convergence stage, the data is stable and slowly increases or decreases.

[0032] The microscope moves from position a\b to position b\a along the Z axis. Under the B color light, the plane measurement results experience a process of divergence, convergence, and divergence. In the convergence stage, the data is stable and slowly decreases or increases.

[0033] The microscope moves from a\b to b\a along the Z axis. Under the mixing of G\B color light data, the plane measurement results experience a process of divergence, convergence, and divergence. During the convergence stage, the data is stable and almost unchanged.

[0034] Under the G\B color light data mixture, the most stable data is obtained in the convergence stage, the G\B color light mixed data is fused, the stability of the mixed data segment is evaluated, and the data in the most stable stage is extracted as the measurement result.

[0035] Furthermore, the calculating of the optical critical dimension of the 3D measurement object according to the relationship between the plane measurement result, the image acquisition sequence number and the 3D model of the 3D measurement object includes:

[0036] G\B color light data fusion;

[0037] Evaluate the stability of fused data;

[0038] Calculate the most stable data segment and output the results.

[0039] Furthermore, the G\B color light data fusion includes:

[0040] The data of G and B color light are fused by corresponding image acquisition sequence numbers to obtain G\B color light fusion data.

[0041] Furthermore, the evaluating the stability of the fused data includes:

[0042] The standard deviation is used to evaluate the stability and the mean of the fused data is solved at the same time. More than three consecutive measurement data are used for calculation to obtain the stability evaluation and mean data of the fused data.

[0043] Furthermore, the calculation of the most stable data segment and the output of the result include:

[0044] The one with the smallest standard deviation is regarded as the most stable one. Based on the stability evaluation of the fused data and the mean data, the minimum stability evaluation parameter is obtained, and the image number corresponding to the minimum stability evaluation parameter is recorded.

[0045] Using the mean data as the final measurement output result;

[0046] Output the picture corresponding to the picture number.

[0047] The advantages of this invention lie in its ability to utilize 3D technology to simply, cost-effectively, and efficiently address the unstable and unreliable results of existing 2D measurement methods in 3D environments. It also enables accurate and stable measurement of the OCD of 3D models. Extensive data validation demonstrates the method's excellent stability and accuracy, and its high degree of compatibility with CDSEM. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0049] Figure 1 A schematic diagram of a visual model according to an embodiment of the present invention is shown.

[0050] Figure 2 A schematic diagram of a map acquisition process according to an embodiment of the present invention is shown.

[0051] Figure 3 A schematic diagram showing the relationship between the plane measurement results under G color light, the image acquisition sequence number, and the 3D measurement object.

[0052] Figure 4 A schematic diagram showing the relationship between the plane measurement results under B color light, the image acquisition sequence number, and the 3D measurement object.

[0053] Figure 5 A schematic diagram showing the relationship between the plane measurement results under GB color light mixing, the image acquisition sequence number, and the 3D measurement object. DETAILED DESCRIPTION

[0054] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0055] When dealing with high-rise, multi-layered, or thickly coated products, the laser focus stability of a microscope cannot be guaranteed. Furthermore, the dimensions vary significantly at different heights, making it difficult to accurately determine the results. Therefore, this method proposes an OCD measurement method based on 3D model relationships to solve this problem.

[0056] Specifically, the present invention provides an OCD measurement method based on a 3D model relationship, comprising:

[0057] Collect the plane measurement results of the 3D measurement object at different heights of the microscope. Each plane measurement result corresponds to a map number.

[0058] Establish the relationship between plane measurement results, drawing sequence numbers and 3D model of 3D measurement objects;

[0059] The optical critical dimension of the 3D measurement object is calculated according to the relationship between the plane measurement result, the image acquisition sequence number and the 3D model of the 3D measurement object.

[0060] Figure 1 A schematic diagram of a visual model according to an embodiment of the present invention is shown. Figure 2 A schematic diagram of a map acquisition process according to an embodiment of the present invention is shown.

[0061] like Figure 1 、 2 As shown in the figure, this method moves the microscope along the Z axis while taking images with the microscope. As the height changes, the measured object also presents different performances, accompanied by changes in the critical dimensions of the plane. This method is centered on the Z-axis focus position O and moves within the range of offset = ±|ab| / 2. By collecting images of green (G) and blue (B) light at different heights, the data of the two colors are calculated separately and the two sets of data are fused. The test data shows that the stability of the plane measurement results and the change of the data have the following 3D model relationship:

[0062] (1) Figure 3 As shown in the figure, the microscope moves from position a\b to position b\a along the Z axis. Under G color light, the plane measurement results experience a process of divergence, convergence, and divergence. In the convergence stage, the data is stable and slowly increases and decreases.

[0063] (2) Figure 4 As shown, the microscope moves from position a\b to position b\a along the Z axis. Under the B color light, the plane measurement results experience a process of divergence, convergence, and divergence. In the convergence stage, the data is stable and slowly decreases and increases.

[0064] (3) Figure 5 As shown in the figure, the microscope moves from position a\b to position b\a along the Z axis. Under the mixing of G\B color light data, the plane measurement results experience a process of divergence, convergence, and divergence. During the convergence stage, the data is stable and almost unchanged.

[0065] Through the above data analysis, it is concluded that when G / B color light data is mixed, the most stable data is obtained during the convergence phase, and the result is close to the correct size. Therefore, this method fuses the G / B color light data, evaluates the stability of the mixed data segment, and extracts the most stable stage data as the measurement result. The specific method is as follows:

[0066] 1. G\B color light data collection

[0067] The wafer is placed on the motion platform, and the microscope performs initial focus on the measurement point. Starting from the focus position (or near) O, the microscope moves up and down along the Z axis by offset to the a\b position to prepare the position; the microscope moves along the Z axis from the a\b position to the b\a position, with the distance between a and b being 2 times the offset, and the microscope image acquisition and plane calculation are started at the same time; after reaching the specified position, the microscope stops moving and the microscope image acquisition and calculation process is closed, and data acquisition is completed. Specific process:

[0068] (1) The microscope performs initial focusing, with the focus position O as the starting position;

[0069] (2) The microscope moves up and down along the Z axis by a preset distance to the a / b preparation position;

[0070] (3) The microscope moves downward or upward along the Z axis to b or a;

[0071] (4) When moving from a to b, switch to G color light, and take a picture and calculate the key dimensions of the plane every time it moves a step;

[0072] (5) When b moves to a, switch to color light B. Each time it moves a step, it takes a picture and calculates the key dimensions of the plane.

[0073] (6) Store the results into an array by number and save the image.

[0074] During the data acquisition process, initial focusing is performed to ensure clear coverage of the microscope as it moves along the Z axis. The microscope moves between positions a and b along the Z axis while simultaneously switching between G / B color light, capturing images, and performing planar calculations. This ensures high synchronization, thereby improving system efficiency. The image acquisition interval and Z-axis position do not have a linear relationship, so the images are numbered in the order they are taken, and the corresponding calculation results and images are recorded.

[0075] Planar key dimension measurement is the key dimension measurement of 2D images, which includes various measurement methods such as line width, overlay, circle diameter, etc. This 3D imaging processing method is based on this method.

[0076] In the above, two sets of key size arrays are obtained. The following data is processed and calculated.

[0077] 2. Data Processing

[0078] The data is processed by fusion, stability assessment, and mean calculation. The specific contents are as follows:

[0079] (1) G\B color light data fusion

[0080] The value of the object measured under G color light is represented by g_cd(n), the value of the object measured under B channel is represented by b_cd(n), and the value of the object measured after G\B color light fusion is represented by gb_cd(n); perform data fusion, the fusion formula is:

[0081] gb_cd(n)=(g_cd(n)+b_cd(n)) / 2

[0082]

[0083]

[0084] The data of G and B color light are fused by corresponding serial numbers to obtain the fused data of G\B color light;

[0085] (2) Evaluate the stability of fused data

[0086] This method uses standard deviation to evaluate stability and solves the corresponding mean at the same time. It uses three or more consecutive measurement data for calculation. gb_avg(n) represents the mean of the fused data, and gb_avg(n) represents the standard deviation of the fused data. The calculation using three data is as follows:

[0087] gb_avg(n)=avg(gb_cd(n),gb_cd(n+1),gb_cd(n+2))

[0088] gb_cd(n)=Sigma(gb_cd(n),gb_cd(n+1),gb_cd(n+2))

[0089] Serial number B_CD GB_Average GB_Sigma 1 gb_cd(1) gb_avg(1) gb_sigma(1) 2 gb_cd(2) gb_avg(2) gb_sigma(2) … … … … n-2 gb_cd(n-2) gb_avg(n-2) gb_sigma(n-2) n-1 gb_cd(n-1) n gb_cd(n)

[0090] In the above table, GB_Average represents the mean of the fused data, and GB_Sigma represents the standard deviation of the fused data.

[0091] The above obtains the stability evaluation and mean data of the fusion data.

[0092] (3) Calculate the most stable data segment and output the results

[0093] The one with the smallest standard deviation is the most stable. Using the above data, we can find the minimum stability evaluation parameter: MIN(g_sigma(1), g_sigma(2)…g_sigma(n-2)). The image number minNum_g corresponding to the minimum value is recorded.

[0094] In order to achieve better filtering effect and adaptability, the average value is used as the final result. By querying the above data, the measurement output result is obtained: g_avg(minNum_g);

[0095] The output image is the serial number corresponding to the minNum_g-th image.

[0096] The foregoing description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be readily conceived by a person skilled in the art within the technical scope disclosed herein are intended to be encompassed within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. An OCD measurement method based on 3D model relationship, characterized in that: include: Collect the plane measurement results of the 3D measurement object at different heights of the microscope. Each plane measurement result corresponds to a map number. Establish the relationship between plane measurement results, drawing sequence numbers and 3D model of 3D measurement objects; Calculating the optical critical dimension of the 3D measurement object according to the plane measurement result, the image acquisition sequence number and the 3D model relationship of the 3D measurement object; The 3D model relationship is as follows: The microscope moves from position a\b to position b\a along the Z axis. Under G color light, the plane measurement results experience a process of divergence, convergence, and divergence. In the convergence stage, the data is stable and slowly increases or decreases. The microscope moves from position a\b to position b\a along the Z axis. Under the B color light, the plane measurement results experience a process of divergence, convergence, and divergence. In the convergence stage, the data is stable and slowly decreases or increases. The microscope moves from a\b to b\a along the Z axis. Under the mixing of G\B color light data, the plane measurement results experience a process of divergence, convergence, and divergence. During the convergence stage, the data is stable and almost unchanged. Under the G\B color light data mixture, the most stable data is obtained in the convergence stage, the G\B color light mixed data is fused, the stability of the mixed data segment is evaluated, and the data in the most stable stage is extracted as the measurement result.

2. The OCD measurement method based on 3D model relationship according to claim 1, characterized in that: The acquisition of the plane measurement results of the 3D measurement object at different heights using a microscope, wherein each plane measurement result corresponds to a map acquisition sequence number, includes the following steps: The wafer is placed on the motion platform, and the microscope focuses on the measurement point. Starting from the focus position, the microscope is moved vertically up or down along the Z axis to the height a or b position as the preparation position; The microscope moves along the Z axis from position a\b to position b\a, while simultaneously starting image acquisition and plane calculation; After reaching the designated position, the microscope stops moving and the image acquisition and calculation are turned off.

3. The OCD measurement method based on 3D model relationship according to claim 2, characterized in that: The acquisition of the plane measurement results of the 3D measurement object at different heights using a microscope, wherein each plane measurement result corresponds to a map acquisition sequence number, includes the following steps: (1) The microscope performs initial focusing, with the focus position O as the starting position; (2) The microscope moves up and down along the Z axis by a preset distance to the a / b preparation position; (3) The microscope moves downward / upward along the Z axis to b / a; (4) When moving from a to b, switch to G color light, and take pictures and calculate the key dimensions of the plane every time it moves a step; (5) When b moves to a, switch to color B. Each time it moves a step, it takes a picture and calculates the key dimensions of the plane. (6) Store the results into an array by number and save the image.

4. The OCD measurement method based on 3D model relationship according to claim 1, characterized in that: Calculating the optical critical dimension of the 3D measurement object according to the relationship between the plane measurement result, the image acquisition sequence number and the 3D model of the 3D measurement object includes: G\B color light data fusion; Evaluate the stability of fused data; Calculate the most stable data segment and output the results.

5. The OCD measurement method based on 3D model relationship according to claim 4, characterized in that: The G\B color light data fusion includes: The data of G and B color light are fused by corresponding image acquisition sequence numbers to obtain G\B color light fusion data.

6. The OCD measurement method based on 3D model relationship according to claim 5, characterized in that: The evaluation of the stability of the fused data includes: The standard deviation is used to evaluate the stability and the mean of the fused data is solved at the same time. More than three consecutive measurement data are used for calculation to obtain the stability evaluation and mean data of the fused data.

7. The OCD measurement method based on 3D model relationship according to claim 6, characterized in that: The calculation of the most stable data segment and the result output include: The one with the smallest standard deviation is regarded as the most stable one. Based on the stability evaluation of the fused data and the mean data, the minimum stability evaluation parameter is obtained, and the image number corresponding to the minimum stability evaluation parameter is recorded. Using the mean data as the final measurement output result; Output the picture corresponding to the picture number.

Images