CN110244527A - A Method for Optimizing Overlay Mark Shape and Measurement Conditions - Google Patents

Abstract

The invention discloses a kind of overlay mark pattern and measuring condition optimization methods, belong to field of lithography.Firstly, determining overlay mark appearance structure and materials optical constant according to photoetching process.Secondly, alignment optical characterisation curve can be calculated by the modeling method of parsing or numerical value.Then, the expression formula of alignment measurement reproducibility precision and accuracy is obtained according to Taylor's formula.Then, variable to be optimized and suitable multi-objective Algorithm are selected, alignment measurement reproducibility precision and accuracy are optimized, obtains multiple Pareto optimization results.Finally, it is verified that the robustness of multiple Pareto optimization results, selects the preferable result of robustness as final optimization pass result.The overlay mark pattern and measuring condition of eDBO method is optimized in method disclosed by the invention simultaneously, final optimization pass result has the characteristics that repeatability precision is high, accuracy of measurement is high, measurement robustness is good, meets the needs that overlay error measures in photoetching process.

Description

A Method for Optimizing Overlay Mark Shape and Measurement Conditions

technical field

The invention belongs to the field of photolithography, and more specifically relates to a method for optimizing the shape of overlay marks and measurement conditions.

Background technique

With the rapid development of integrated circuit technology, integrated circuits have developed from small scale to extremely large scale, and their critical dimensions (Critical Dimension, CD) have also been continuously reduced from micron level to today's 7nm node. The manufacturing process of integrated circuits includes multiple processes such as material preparation, photolithography, cleaning, etching, doping, and chemical mechanical polishing, among which the photolithography process is the most critical. The main indicators of the lithography process include resolution, depth of focus, key dimensions, alignment and overlay accuracy, etc. Among them, overlay refers to the alignment relationship between the current photolithography process layer and the previous layer process layer. Generally, the overlay error is required to be no more than 1/3 of the critical dimension. Therefore, the rapid measurement and accurate evaluation of the overlay error is to ensure the process and key to device performance.

Haiyong Gao et al. summarized in the literature "Comparison study of diffraction-based overlay and image-based overlay measurements on programmed overlay errors" that overlay measurement methods can be roughly divided into two categories: image-based overlay measurement (Image-Based Overlay, IBO) and Diffraction-Based Overlay (DBO). Wei Yayi mentioned in the document "Advanced Lithography Theory and Application of VLSI", in order to measure the overlay error of the front and back process layers, both methods need to be designed at the same position on the front and back process layers that are registered with each other. Overlay marks, and then measure the overlay marks with overlay measuring equipment, so as to obtain overlay error measurement results. These marks are usually at the edge of the exposed area, also known as the "scribeline" or "kerf" area.

Jie Li et al. mentioned in the document "Evaluating Diffraction-Based Overlay" that DBO methods mainly include eDBO and mDBO (model-based DBO). For the mDBO method, it is mentioned in the patent CN103472004B that it is difficult to meet the time requirement of actual measurement because it needs to construct a large number of forward optical characteristic models of overlay marks. For the eDBO method, the patent CN103454861B mentions that this method has the advantages of fast measurement speed and small sampling area, and at the same time eliminates many error items of traditional measurement methods, such as positioning error, focal plane error, aberration factors and mechanical vibration. In addition, Chinese patent CN200510091733, US patents US7173699B2, US7477405B2, US7428060B2 and US6985232B2 all disclose such methods.

The eDBO method extracts the overlay error based on the local linear relationship of the overlay optical characterization curve. At the origin O of the overlay optical characterization curve 200, the overlay offset δ is approximately linear in relation to the optical characterization quantity I:

I=Kδ (1)

Among them, the overlay offset δ and the measurement sensitivity K are both unknown quantities, and it is difficult to determine their values through a single measurement. For this reason, eDBO realizes the extraction of overlay error by differential method by introducing known overlay offset.

A top view of a typical eDBO overlay mark is shown at 300 , 301 and 304 are used to measure the overlay error in the X direction, and 302 and 303 are used to measure the overlay error in the Y direction. Taking the X direction as an example, the overlay error to be obtained is ε, and the known overlay offsets +D and -D are respectively introduced into the two overlay marking units 301 and 304, and the corresponding cross-sectional view is shown in Figure 3(b) and 3(c), the total overlay offset δ ^± of the two units is D+ε and -D+ε respectively. Therefore, according to the formula (1), the optical characterization quantities of the two overlay marking units are as follows:

I ⁺ ＝K(D+ε) (2)

I ^- ＝K(-D+ε) (3)

Then the overlay error ε is:

Therefore, only the overlay marking units 301 and 304 need to be measured separately to obtain I ⁺ , I ⁻ , and the overlay error ε can be obtained through the above formula. Considering the influence of noise △I in the actual overlay measurement, formulas (2) and (3) are rewritten as:

I ⁺ ＝K(D+ε)+△I ⁺ (5)

I ^- ＝K(-D+ε)+△I ^- (6)

Therefore, the overlay error value obtained by actual measurement is as follows:

Replace △I ⁺ +△I ^- in formula (7) with σ _m (the standard deviation of the instrument measurement noise that characterizes the optical quantity I overlaid), and ignore △I ⁺ -△I ^- in the denominator to obtain the repetition of the eDBO method Sexual measurement accuracy σ:

It can be seen that the measurement sensitivity K and the noise level of the instrument jointly determine the repeatability measurement accuracy of the overlay error.

The key to the eDBO method is that the overlay offset δ∈[-D+ε, D+ε] approximately satisfies the linear relationship assumption in formula (1). However, the overlay optical characteristic curve is a periodic odd function (see 201), so the nonlinearity of the characteristic curve will introduce the overlay offset characterization error Δδ ₁ (see 202), affecting the measurement accuracy μ. In addition, other asymmetric factors of the overlay mark will translate the overlay optical characteristic curve and introduce an overlay offset error Δδ ₂ (see 203), which will also affect the measurement accuracy μ.

The overlay error characterization curve not only depends on the shape of the overlay mark, but also is related to the measurement conditions (such as measurement wavelength, incident angle, azimuth angle), so in order to achieve high repeatability, high accuracy, and good robustness In order to measure the engraving error, it is necessary to study a method for optimizing the morphology and measurement conditions of eDBO overlay marks.

Contents of the invention

Aiming at the above defects or improvement needs of the prior art, the present invention provides a method for optimizing the shape of the overlay mark and the measurement conditions. Repeatability measurement accuracy σ and accuracy μ are two optimization objectives, and multiple Pareto optimal results with good robustness are obtained, thereby solving the problem of low repeatability, low accuracy, and robustness of existing overlay error measurement. Poor technical issues.

In order to achieve the above object, the present invention provides a method for optimizing the shape of overlay marks and measurement conditions, which is characterized in that it includes the following steps:

Step 101, determining the overlay mark topography and material optical constants;

Step 102, according to the parameters and material optical constants of the overlay mark determined in step 101, and the set measurement conditions, a single optical characteristic I is calculated by analytical or numerical modeling methods, and by continuously changing the overlay offset δ is calculated to obtain an overlay optical characterization curve; the overlay optical characterization curve represents the functional relationship between the optical characterization quantity I and the overlay offset δ;

Step 103, expanding the Taylor formula at the origin of the overlay optical characterization curve to obtain the expressions of the overlay measurement repeatability precision σ and accuracy μ;

Step 104, select the variables to be optimized from the parameters of the topographic structure of the overlay mark and the measurement conditions, take the repeatability precision σ and accuracy μ of the overlay measurement as optimization objectives, and use a multi-objective optimization algorithm to optimize the The variables to be optimized are selected iteratively, and Pareto optimization results including multiple sets of overlay measurement repeatability precision σ and accuracy μ are obtained.

Preferably, the method for optimizing the shape of the overlay mark and the measurement conditions further includes step 105, performing overlay error extraction simulation on a plurality of the Pareto optimization results, and verifying the repeatability accuracy and measurement accuracy of the corresponding overlay measurement degree and robustness, choose the optimal solution for measurement experiments.

Preferably, the expressions of the overlay measurement repeatability precision σ and accuracy μ are,

Among them, σ _m is the instrument measurement standard deviation of the optical characterization, D is the set overlay offset, K is the overlay sensitivity, ε is the overlay error value to be measured, and Δ _a is the impact of other asymmetric factors on the optical characterization Quantitative impact,

In the formula, I″(ξ ⁺ ) is the second derivative value of the overlay optical characteristic curve at δ=ξ ⁺ , I″( ^{ξ- )} is the second derivative value of the overlay optical characteristic curve at δ=ξ- , ξ ⁺ ∈ [0, D + ε], ξ ^- ∈ [0, -D + ε].

Preferably, step 103 is specifically,

According to Taylor's formula, I ⁺ and I ^- are corrected as the following formula:

Among them, ΔI ⁺ and ΔI ^- are the measurement noise of the instrument, and I(0) is the value of the overlay optical characteristic curve at the origin;

According to formula (13) and formula (14), the measurement error △ε of eDBO method is shown in formula (15):

Using an optical characterization instrument to measure the standard deviation σ _m to replace △I ⁺ + △I ^- in formula (15), the expression (9) and expression (9) and expression ( 10).

Preferably, in step 102, the parameters of the topography of the overlay mark include the line width CD ₁ of the top grating, the line width CD ₂ of the bottom grating, the period Pitch of the top grating and the bottom grating, and the wall height H ₁ of the top grating , the wall height H ₃ of the bottom grating, the sum H ₂ of the thickness of the middle film layer and the wall height of the bottom grating, the overlay offset δ, the left wall angle LSWA of the bottom grating and the right wall angle RSWA of the bottom grating; the material optics The constant refers to the complex refractive index of the material.

Preferably, in step 102, the measurement conditions include measuring the incident angle θ, the azimuth angle Measuring one or more combinations of wavelength λ and polarization angle Ψ; the optical characterization quantity I refers to reflectivity, ellipsometric parameters or Mueller matrix.

Preferably, in step 104, the selected variables to be optimized are any combination of all parameters in the topographic structure parameters of the overlay mark and the measured incident angle θ, azimuth angle Any combination of one or more of wavelength λ and polarization angle Ψ is measured.

Preferably, the analytical or numerical modeling method uses rigorous coupled wave analysis, finite element method, boundary element method or finite time domain difference method.

Preferably, the multi-objective optimization algorithm adopts a multi-objective particle swarm optimization algorithm, a multi-objective evolutionary algorithm or a multi-objective genetic algorithm.

Generally speaking, compared with the prior art, the above technical solutions conceived by the present invention can achieve the following beneficial effects:

1. The optimization method based on eDBO provided by the present invention calculates and obtains the overlay optical characterization curve representing the functional relationship between the optical characterization quantity I and the overlay offset δ through analytical or numerical modeling methods, and obtains the overlay optical characterization curve through the overlay optical characterization curve and Taylor's formula is expanded to obtain the expressions of overlay measurement repeatability precision σ and accuracy μ, which take into account the influence of nonlinearity of overlay optical characterization curve, instrument measurement noise and other asymmetric factors on eDBO method. Then, taking the overlay measurement repeatability precision σ and accuracy μ as optimization objectives, a multi-objective optimization algorithm is used to iteratively select the variables to be optimized to obtain Pareto optimization results. The present invention simultaneously considers the overlay mark shape optimization and measurement condition optimization configuration, so that the overlay error measurement has a better performance in terms of precision and accuracy, and realizes high repeatability measurement accuracy, high measurement accuracy and robustness. Overlay error measurement with good stickiness.

2. The present invention provides multiple Pareto optimization results according to the two indicators of measurement repeatability precision σ and accuracy μ. Considering the influence of the actual processing and measurement process, the optimization results can be further verified by overlay measurement repeatability accuracy, measurement accuracy and robustness through simulation, and a solution that meets the needs and good robustness can be selected to ensure that the optimization results are consistent with the actual measurement results. effectiveness in the process.

3. The multi-objective optimization algorithm iteratively selects the two optimization objectives of overlay measurement repeatability accuracy σ and accuracy μ at the same time, and finally obtains multiple sets of Pareto optimization results of the overlay measurement repeatability accuracy σ and accuracy μ . The optimization results also take into account the nonlinearity of overlay optical characterization curves, instrument measurement noise and other asymmetric factors on the eDBO method, and the repeatability measurement accuracy and accuracy are better, and the robustness is better.

Description of drawings

Fig. 1 is a flow chart of an overlay mark topography and a method for optimizing measurement conditions in a preferred embodiment of the present invention;

Fig. 2 (a) is the local linear relationship between typical overlay optical characteristic curve and overlay offset in a preferred embodiment of the present invention;

Fig. 2 (b) is the impact of the non-linearity of the overlay optical characterization curve on the overlay error characterization in a preferred embodiment of the present invention;

Fig. 2 (c) is the impact of other asymmetric factors of the overlay mark in the overlay optical characterization curve on the overlay error characterization in a preferred embodiment of the present invention;

Figure 3(a) is a top view of the topography of a typical eDBO overlay mark in a preferred embodiment of the present invention;

Fig. 3 (b) is a sectional view of the overlay marking unit after introducing the set overlay offset + D in the preferred embodiment of the present invention;

Figure 3(c) is a cross-sectional view of the overlay marking unit after introducing the set overlay offset -D in a preferred embodiment of the present invention;

Fig. 4 is a schematic cross-sectional structure diagram of each overlay marking unit in a preferred embodiment of the present invention;

Fig. 5 is a schematic diagram of overlay error measurement conditions in a preferred embodiment of the present invention;

Fig. 6 (a) is the flow chart that adopts multi-objective optimization algorithm to seek Pareto optimal result in the preferred embodiment of the present invention;

Fig. 6 (b) is the measurement repeatability precision and accuracy of the overlay error before optimization in a preferred embodiment of the present invention;

Fig. 6 (c) is the measurement repeatability precision and accuracy of overlay error after optimization in the preferred embodiment of the present invention;

Fig. 7 (a) is the robustness of the repeatability measurement precision σ of Pareto optimization result in the preferred embodiment of the present invention;

Fig. 7(b) is the robustness of the measurement accuracy μ of the Pareto optimization result in the preferred embodiment of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not constitute a conflict with each other.

The invention provides an empirical relationship-based diffraction overlay error measurement method (Empirical Diffraction-Based Overlay, eDBO), which provides an eDBO method-based method for optimizing the shape of overlay marks and measurement conditions. This method takes the overlay error repeatability measurement precision σ and accuracy μ as two optimization objectives, and obtains multiple Pareto optimal results. Then, the robustness of the optimization results can be verified by simulation for screening. After screening, The result is the final optimization result of overlay mark morphology parameters and measurement conditions.

First, according to the semiconductor manufacturing process, it is necessary to determine the topology of the eDBO overlay mark, that is, the shape of each layer structure (such as a film or grating), and the optical constants (refractive index n and extinction coefficient k) of the corresponding materials. On this basis, through modeling methods such as RCWA, FEM, BEM, FDTD, etc., the forward optical characteristic model can be established, and the overlay optical characteristic curve can be calculated. Secondly, according to the Taylor formula, the expressions of the overlay error repeatability measurement accuracy σ and measurement accuracy μ of the eDBO method can be obtained, and the approximate values of the two can be calculated according to the overlay optical characterization curve. Since repeatability measurement precision and accuracy are very important indicators in overlay measurement, and the correlation between these two indicators is weak, it is necessary to use a multi-objective optimization algorithm to optimize the two indicators at the same time. Select the shape parameters and measurement conditions of the overlay mark as the variables to be optimized, and use the multi-objective optimization algorithm as the iterative strategy to obtain the Pareto optimal measurement repeatability precision σ and accuracy μ. The corresponding overlay of these two indicators The marked morphology parameters and measurement conditions are the optimization results.

Finally, it is also possible to perform overlay error extraction simulation verification on multiple optimization results. The simulation includes simulating the processing error in the semiconductor process, the uncertainty of the measurement conditions, and the random noise of the measurement to verify the performance and robustness of different optimization results, and select the overlay mark shape and measurement conditions for measurement according to specific needs experiment.

The method for optimizing the overlay mark appearance and measurement conditions provided by the present invention will be further described in detail below with reference to the accompanying drawings and examples.

FIG. 1 shows a processing flow 100 of an overlay mark and measurement configuration optimization method provided by the present invention. The embodiment of the present invention will take the three-layer overlay mark 400 as an example to specifically illustrate the specific operation process of the method provided by the present invention:

(1) Determine the overlay mark structure and material optical constants

The top view of the overlay mark structure provided by this embodiment is shown in FIG. 304 is used to measure the overlay error in the X direction. Wherein, the cross-sectional view of each unit vertical to the grating direction is shown in FIG. 4 . The bottom layer is a Si grating layer (trapezoidal protrusions), the middle layer is a _SiO2 thin film layer, the top layer is a photoresist grating (square protrusions), and the base is Si. Among them, the parameters of the cross-sectional morphology of the overlay mark are as follows, CD ₁ represents the line width of the top photoresist grating, CD ₂ represents the line width of the bottom Si grating layer, Pitch represents the period of the grating, H ₁ represents the top layer photoresist grating Wall height, H ₃ represents the wall height of the bottom Si grating layer, H ₂ represents the sum of the thickness of the middle film layer and the wall height of the bottom grating layer, δ represents the overlay offset, LSWA and RSWA represent the left and right side wall angles of the bottom grating, respectively. The parameters of the cross-sectional topography of the overlay mark are not limited to the parameters listed above, and those skilled in the art can set them according to actual needs, and they are not exhaustively listed here. In addition, the optical constant of the material needs to be obtained by measurement means. The material optical constant refers to the complex refractive index of the material, that is, the refractive index and the extinction coefficient.

(2) Calculation method for determining overlay optical characterization curve

According to the overlay mark morphology parameters and material optical constants obtained in step (1), and the incident angle θ and azimuth angle measured in 500 in Fig. 5 By measuring the wavelength λ and the polarization angle Ψ, a single optical characteristic I can be calculated by analytical or numerical modeling methods. The overlay optical characterization curve can be calculated by continuously changing the overlay offset δ.

Available modeling methods include Rigorous Coupled Wave Analysis (RCWA), Finite Element Method (FEM), Boundary Element Method (BEM) or Finite Difference Time Domain (FDTD). The optical characterization quantity I includes any one of reflectivity, ellipsometric parameters, and Mueller matrix.

(3) Obtain the repeatability and accuracy of overlay measurement based on Taylor's formula

As shown in FIG. 2( b ), the overlay optical characteristic curve 201 is a periodic odd function, so its non-linearity will lead to inaccurate measurement (see 202 in FIG. 2( b )). Secondly, because the overlay error is essentially one of the asymmetric factors, other asymmetric factors (such as the left wall angle LSWA≠right wall angle RSWA) of the overlay mark in the process will also make the entire overlay optical The translation of the characteristic curve makes the overlay error measurement result inaccurate (see 203 in Fig. 2(c)). Both of these factors can cause measurement accuracy problems with the eDBO method. Considering the nonlinear and other asymmetrical factors of the aforementioned overlay optical characteristic curve, we correct I ⁺ and I ^- to the following formula based on the Taylor expansion:

Among them, ε is the overlay error value to be measured, D is the set overlay offset, Δ _a is the influence of other asymmetric factors on the optical representation of overlay, ΔI ⁺ and ΔI ^- are the measurement noise of the instrument, I( 0) is the value of the overlay optical characteristic curve at the origin (corresponding to the overlay offset of 0), generally 0, and K is the overlay sensitivity. I″(ξ ⁺ ) and I″(ξ ^- ) are the second-order derivative values of the overlay optical characteristic curve at ξ ⁺ , ξ ^- respectively. According to the Taylor expansion theorem, ξ ⁺ ∈[0,D+ε], ξ ^－ ∈[0,-D+ε].

According to formula (13) and formula (14), the measurement error △ε of eDBO method is shown in formula (15):

N ⁺ and N ^- in formula (15) are as follows:

By replacing △I ⁺ -△I ^- in formula (15) with σ _m (instrumental measurement standard deviation of overlay optical characterization quantity), the repeatability measurement precision σ and accuracy μ of overlay error can be obtained:

Since N ⁺ and N ⁻ are not definite values, in this embodiment, only their average values are selected to calculate σ and μ.

The advantages of the expressions (9) and (10) are that the influence of the nonlinearity of the overlay characteristic curve, the measurement noise of the instrument and other asymmetric factors on the eDBO method is considered at the same time.

(4) Optimizing overlay mark morphology and measurement conditions

It can be seen from the foregoing that σ and μ are related to the overlay optical characterization curve, and the overlay optical characterization curve is calculated from the overlay mark structure morphology parameters and measurement conditions through the optical model, that is, each set of overlay mark structure shape parameters and measurement conditions corresponding to each set of determined σ and μ. In this embodiment, we select Pitch, CD ₁ , CD ₂ in the overlay mark section view (see 400 in Figure 4), and measure the incident angle θ and azimuth angle in 500 of the overlay measurement condition schematic diagram 5 The measurement wavelength λ is used as an optimization variable, and σ and μ are used as optimization targets. The optimization variables in this embodiment are only used to explain the present invention, and other optimization variables include θ, Any combination of λ, Ψ, Pitch, CD ₁ , CD ₂ .

Since the optimization objective includes both σ and μ and there are many variables to be optimized, it is necessary to use a multi-objective optimization algorithm with strong global search capability for iterative selection of optimization variables. Common multi-objective optimization algorithms include multi-objective particle swarm optimization (MOPSO), multi-objective evolutionary algorithm (MOEA), and multi-objective genetic algorithm (MOGA). In this embodiment, the multi-objective evolutionary algorithm is selected as the optimization algorithm, and the multi-objective optimization algorithm is a global optimization algorithm, which has a better optimization effect than the local optimization algorithm. The multi-objective optimization algorithm iteratively selects the two optimization objectives of overlay measurement repeatability accuracy σ and accuracy μ at the same time, and finally obtains multiple sets of Pareto optimization results of the overlay measurement repeatability accuracy σ and accuracy μ.

As shown in Figure 6(a), before the iteration of the multi-objective evolutionary algorithm starts, multiple sets of initial values of the optimization variables are first generated, and then the overlay optical characterization curves are calculated according to the optical model to obtain the σ and μ corresponding to the initial values of the optimization variables. Then, the optimization variables are selected according to the optimization strategy of the multi-objective evolutionary algorithm (generally, the optimization variables with smaller values of σ or μ are preferred). In order to ensure the number of optimized variables, some mutations are usually performed on the basis of offspring to generate new individuals to maintain the number of individuals and individual diversity in each generation. After each iteration is completed, it is judged whether the optimization objectives σ and μ have reached the optimum, and if they are optimal, the optimization algorithm is stopped.

More specifically, for the overlay mark structure 400 in this example, Fig. 6(b) and Fig. 6(c) respectively represent the values of σ and μ corresponding to the variables to be optimized before and after optimization. It can be seen that in this example, after the iterative selection of the optimization algorithm, the size of σ and μ is one-tenth of the original size. In Figure 6(c), the optimization variables corresponding to σ and μ (θ, λ, Pitch, CD ₁ , CD ₂ ) are the obtained optimization results.

(5) Simulation verification of optimization results

In the semiconductor manufacturing process, it is difficult to guarantee the overlay mark morphology parameters Pitch, CD ₁ , CD ₂ and measurement conditions θ, λ is an ideal value. Therefore, it is also necessary to consider the influence of the actual processing and measurement process to verify the repeatability accuracy, measurement accuracy and robustness of the optimization results to ensure its effectiveness in the actual measurement process.

More specifically, for the overlay mark structure 400 in this example, R in Fig. 7(a) and Fig. 7(b) characterizes the repeatability measurement accuracy σ and measurement accuracy μ of 50 Pareto optimization results respectively robustness. It can be seen that some optimization results are difficult to resist the complex influence of processing technology and actual measurement conditions. Therefore, in practical applications, it is necessary to select the result with good robustness from the Pareto optimization results as the final optimization result for overlay error measurement experiment.

It is easy for those skilled in the art to understand that the above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention, All should be included within the protection scope of the present invention.

Claims (9)

1. A method for overlay mark topography and measurement condition optimization, is characterized in that, comprises the steps: Step 101, determining the overlay mark topography and material optical constants; Step 102, according to the parameters and material optical constants of the overlay mark determined in step 101, and the set measurement conditions, a single optical characteristic I is calculated by analytical or numerical modeling methods, and by continuously changing the overlay offset δ is calculated to obtain an overlay optical characterization curve; the overlay optical characterization curve represents the functional relationship between the optical characterization quantity I and the overlay offset δ; Step 103, expanding the Taylor formula at the origin of the overlay optical characterization curve to obtain the expressions of the overlay measurement repeatability precision σ and accuracy μ; Step 104, select the variables to be optimized from the parameters of the topographic structure of the overlay mark and the measurement conditions, take the repeatability precision σ and accuracy μ of the overlay measurement as optimization objectives, and use a multi-objective optimization algorithm to optimize the The variables to be optimized are selected iteratively, and Pareto optimization results including multiple sets of overlay measurement repeatability precision σ and accuracy μ are obtained. 2. A method for optimizing overlay mark topography and measurement conditions according to claim 1, further comprising step 105, performing overlay error extraction simulation on a plurality of Pareto optimization results, and verifying that each Corresponding overlay measurement repeatability accuracy, measurement accuracy and robustness, select the optimal solution for measurement experiments. 3. A method for optimizing the shape and measurement conditions of an overlay mark according to claim 1 or 2, wherein the expressions of the overlay measurement repeatability precision σ and accuracy μ are: Among them, σ _m is the instrument measurement standard deviation of the optical characterization, D is the set overlay offset, K is the overlay sensitivity, ε is the overlay error value to be measured, and Δ _a is the impact of other asymmetric factors on the optical characterization Quantitative impact. In the formula, I″(ξ ⁺ ) is the second derivative value of the overlay optical characteristic curve at δ=ξ ⁺ , I″( ^{ξ- )} is the second derivative value of the overlay optical characteristic curve at δ=ξ- , ξ ⁺ ∈ [0, D + ε], ξ ^- ∈ [0, -D + ε]. 4. A method for optimizing the shape and measurement conditions of an overlay mark according to claim 1 or 2, characterized in that step 103 is specifically, modifying I ⁺ and I ^- according to the Taylor formula into the following formula: Among them, ΔI ⁺ , ΔI ^- are the measurement noise of the instrument, and I(0) is the value of the optical characterization quantity of the overlay optical characterization curve at the coordinate origin δ=0; According to formula (13) and formula (14), the measurement error △ε of eDBO method is shown in formula (11): Using an optical characterization instrument to measure the standard deviation σ _m to replace △I ⁺ + △I ^- in formula (15), the expression (9) and expression (9) and expression ( 10). 5. A method for optimizing the overlay mark topography and measurement conditions according to claim 1 or 2, characterized in that in step 102, the parameters of the topographic structure of the overlay mark include the line width CD ₁ of the top grating , the line width CD ₂ of the bottom grating, the period Pitch of the top grating and the bottom grating, the wall height H ₁ of the top grating, the wall height H ₃ of the bottom grating, the sum of the thickness of the middle film layer and the wall height of the bottom grating H ₂ , overlay Offset δ, the left wall angle LSWA of the bottom grating and the right wall angle RSWA of the bottom grating; the material optical constant refers to the complex refractive index of the material. 6. A method for optimizing overlay mark topography and measurement conditions according to claim 5, characterized in that in step 102, the measurement conditions include measuring incident angle θ, azimuth angle Measuring one or more combinations of wavelength λ and polarization angle Ψ; the optical characterization quantity I refers to reflectivity, ellipsometric parameters or Mueller matrix. 7. A method for optimizing the shape of an overlay mark and measurement conditions according to claim 6, wherein in step 104, the selected variables to be optimized are any of all parameters in the shape and structure parameters of the overlay mark Combine and measure incident angle θ, azimuth angle Any combination of one or more of wavelength λ and polarization angle Ψ is measured. 8. A method for optimizing the shape of an overlay mark and measurement conditions according to claim 1 or 2, wherein the analytical or numerical modeling method adopts strict coupled wave analysis, finite element method, boundary element method or Finite time-domain difference method. 9. A method for optimizing the topography and measurement conditions of an overlay mark according to claim 1 or 2, wherein the multi-objective optimization algorithm adopts a multi-objective particle swarm optimization algorithm, a multi-objective evolutionary algorithm or a multi-objective genetic algorithm .