Quantitative evaluation method of OPC model

Abstract

The present invention provides a quantitative evaluation method of an OPC model. Based on the OPC model, the light intensity distribution I of each mask pattern is obtained. The light intensity distribution I is composed of necessary items, auxiliary items, and coefficients for adjusting each item. The light intensity I is obtained by the light intensity distribution I of each mask pattern and the critical size of the wafer. meas , light intensity I meas The light intensity of each mask pattern at the location of the critical dimension endpoint is obtained by meas and light intensity I threshole The difference I err , taking all mask patterns as samples to obtain the difference I err The root mean square of the difference I err and the difference I err The root mean square of the OPC model is used to evaluate the optimization direction of the OPC model; the mask pattern is classified and the light intensity I of each classification of the OPC model is obtained respectively. meas and light intensity I threshole The difference I err and the difference I err The root mean square of the OPC model is used to achieve a detailed quantitative evaluation of the optimization direction of the OPC model. The evaluation results obtained by the present invention provide a clearer direction for the optimization of the OPC model, shorten the modeling cycle, and help obtain an OPC model that is closer to the requirements.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and in particular to a quantitative evaluation method of an OPC model. Background Art

[0002] Optical Proximity Correction (OPC) is an essential technology in the production of advanced-node semiconductor integrated circuits. The most widely used technology, model-based OPC (MOPC), uses a model to predict lithography results and then modifies the layout shape through a feedback mechanism to bring the predicted results close to the target. Therefore, the error in the model (OPC model) used by MOPC determines the quality of the OPC results.

[0003] At present, the general modeling method for the OPC model is to select a certain number of various types of mask patterns, use the lithography process to be modeled, collect the wafer CD (critical dimension on the wafer) of these mask patterns, and use the lithography parameters to establish an initial OPC model. The simulated light intensity I is calculated by the following formula. I consists of one required item and zero to multiple auxiliary items. The required item is calculated by the optical model established by the physical parameters of the lithography system. The auxiliary item is a correction item added during modeling to make the simulated CD (critical dimension) more consistent with the wafer CD.

[0004] For example, the simulated light intensity I is:

[0005]

[0006] After getting I, I and I threshold The distance between the intersection points is the simulated CD. Calculate the difference between the simulated CD of the OPC model and the wafer CD for these mask patterns (CD err ) and the root mean square (RMS) of the difference, take the value that makes CD err When the root mean square is the smallest, I threshold I is the current OPC model threshold Without changing the optical model, I can only be changed by adding / reducing auxiliary terms and adjusting the coefficient C of each term to achieve the purpose of optimizing the OPC model.

[0007] At present, the quantitative evaluation method for OPC model error is to calculate the difference between the simulated CD of the OPC model for the mask pattern and the wafer CD (CD err ) and the root mean square (RMS) of the difference, the error of the current OPC model can be evaluated from the distribution of the difference and the RMS. errAnd its root mean square cannot intuitively reflect where the OPC model needs to be optimized. For example, the calculation results of an OPC model for two different mask patterns A and B are as follows: Figure 2 As shown, A's CD err CD than B err Small, while A's I err I than B err If you download from CD err Judging by the size of the B graph feature, it can be concluded that the coefficient of the auxiliary term related to the B graph feature in the OPC model needs to be adjusted more than that of A. However, in fact, because of the I err I than B err The correct model optimization direction is that the coefficient of the auxiliary item related to the graphic feature A in the OPC model needs to be adjusted more than that of B. Therefore, from the perspective of OPC model optimization, a method based on I err A quantitative assessment method based on the benchmark.

[0008] In order to solve the above problems, it is necessary to propose a new quantitative evaluation method for the OPC model. Summary of the Invention

[0009] In view of the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a quantitative evaluation method of the OPC model to solve the problems of CD in the prior art. err And its root mean square cannot intuitively reflect the problems where the OPC model needs to be optimized.

[0010] To achieve the above and other related objectives, the present invention provides a quantitative evaluation method for an OPC model, comprising:

[0011] Step 1: Based on the OPC model, the light intensity distribution I of each mask pattern is obtained. The light intensity distribution I consists of necessary items, auxiliary items, and coefficients for adjusting each item. The extension line of the position of the wafer critical dimension on the mask pattern is the measurement axis of each mask pattern. The light intensity I on the light intensity distribution I is threshole The position is the OPC model simulation of the corresponding pattern outline of the mask pattern on the wafer, and the distance between the intersection of the pattern outline and the measurement axis is the OPC model simulation of the critical dimension of the corresponding pattern of the mask pattern on the wafer;

[0012] Step 2: Obtain the light intensity I by using the light intensity distribution I of each mask pattern and the critical size of the wafer meas , light intensity I meas The light intensity of each mask pattern at the location of the critical dimension endpoint is obtained by meas and light intensity I threshole The difference I err , taking all mask patterns as samples to obtain the difference I errThe root mean square of the difference I err and the difference I err The root mean square of the OPC model is used to evaluate the optimization direction;

[0013] Step 3: Classify the mask pattern and obtain the light intensity I of each category of the OPC model meas and light intensity I threshole The difference I err and the difference I err The root mean square of the OPC model is used to achieve a detailed quantitative evaluation of the optimization direction of the OPC model.

[0014] Preferably, the difference I in step 2 err The larger the RMS is, the larger the optimization space of the OPC model is.

[0015] Preferably, the step 2 according to the difference I err and the difference I err The methods for evaluating the optimization direction of the OPC model by the root mean square include: light intensity I background is the background light intensity of each mask pattern along the measurement axis, and the light intensity I max is the maximum light intensity of each mask pattern along the measurement axis, light intensity I min is the minimum light intensity of each mask pattern along the measurement axis; for the mask with 100% light transmittance inside the closed pattern in the corresponding layout, I background Less than I threshole , when the difference I err When it is a negative value, calculate the I of each mask pattern err / (I background -I threshole ), when the difference I err When it is a positive value, calculate the I of each mask pattern err / (I max -I threshole ), denoted as r diff ; For the mask with non-100% light-transmitting area inside the closed pattern in the corresponding layout, I background Greater than I threshole , when the difference I err When it is a positive value, calculate the I of each mask pattern err / (I background -I threshole ), when the difference I err When it is a negative value, calculate the I of each mask pattern err / (I min -I threshole ), denoted as r diff . r diff When it is greater than 0.5, it means that the optimization of the OPC model is difficult for the layout graphics.

[0016] Preferably, step 2 calculates the r diff The percentage of measured axes with a value greater than 0.5 in the total number or the calculated r diff RMS, percentage and r diff The larger the RMS is, the more difficult it is to optimize the OPC model overall.

[0017] Preferably, if the OPC model is optimized by adding the auxiliary terms and adjusting the coefficients of each term in step 2, it is necessary to first reduce or adjust the existing auxiliary terms to prevent overfitting.

[0018] Preferably, in step three, the mask pattern is classified according to at least one of a distance interval, a mask critical dimension, and a mask pattern type.

[0019] Preferably, in step three, the mask patterns are classified according to the distance intervals by using calculable weights.

[0020] Preferably, in step three, the sample mask patterns are first screened to ensure that the sample set simultaneously satisfies the following conditions to ensure the accuracy of subsequent evaluation results: after classification according to the mask pattern type, the set of mask critical dimensions contained in each category is the same; after classification according to the mask pattern type and the mask critical dimension, the set of periods contained in each category is the same.

[0021] Preferably, in step 3, the mask patterns are classified according to the distance intervals and the difference values ​​I of each category are calculated respectively. err The larger the root mean square, the greater the amplitude by which the coefficient of the auxiliary item related to the mask pattern in the corresponding distance interval in the OPC model needs to be adjusted.

[0022] Preferably, in step 3, the mask patterns are classified according to the mask critical dimensions and the difference values ​​I of each category are calculated respectively. err The larger the root mean square, the greater the amplitude by which the coefficients of the auxiliary items related to the mask pattern corresponding to the critical size of the mask in the OPC model need to be adjusted.

[0023] Preferably, in step 3, the mask patterns are classified according to the mask pattern types and the difference values ​​I of each type are calculated respectively. err The larger the root mean square, the greater the amplitude by which the coefficient of the auxiliary item related to the mask pattern of the corresponding mask pattern type in the OPC model needs to be adjusted.

[0024] Preferably, step three further includes classifying the mask patterns according to pattern density.

[0025] As described above, the quantitative evaluation method of the OPC model of the present invention has the following beneficial effects:

[0026] The evaluation results obtained by the present invention can provide a clearer direction for the optimization of the OPC model, shorten the modeling cycle, and help obtain an OPC model that is closer to the needs, laying a solid foundation for excellent OPC solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 Shown is a schematic diagram of establishing an OPC model in the prior art;

[0028] Figure 2 Shown is a schematic diagram of calculation results of an OPC model in the prior art for two different mask patterns A and B;

[0029] Figure 3 Shown is a schematic diagram of the quantitative evaluation method of the OPC model of the present invention;

[0030] Figure 4 Shown is a schematic diagram of establishing an OPC model according to the present invention;

[0031] Figure 5 The display shows that the present invention classifies the layout graphics according to the distance interval and calculates the I of each category respectively. err Root mean square diagram;

[0032] Figure 6 Shown is a schematic diagram of the root mean square calculated by classifying the layout graphics according to the distance interval of the present invention;

[0033] Figure 7 Shown is a schematic diagram of the root mean square calculated by classifying the layout graphics according to the critical dimensions of the mask according to the present invention;

[0034] Figure 8 Shown is a schematic diagram of the root mean square calculated by classifying layout graphics according to the layout graphic type of the present invention. DETAILED DESCRIPTION

[0035] The following describes the embodiments of the present invention through specific examples. Those skilled in the art will readily understand the other advantages and benefits of the present invention from the disclosure herein. The present invention may also be implemented or applied through various other specific embodiments, and the details in this specification may be modified or altered based on different viewpoints and applications without departing from the spirit of the present invention.

[0036] See also Figure 3 The present invention provides a quantitative evaluation method of an OPC model, comprising:

[0037] Step 1: Please refer to Figure 4Based on the OPC model, the light intensity distribution I of each mask pattern is obtained. The light intensity distribution I consists of necessary items, auxiliary items and coefficients for adjusting each item. The extension line of the position of the wafer critical dimension on the mask pattern is the measurement axis (gauge) of each mask pattern. The light intensity I on the light intensity distribution I threshole The position is the OPC model's simulation of the corresponding pattern outline of the mask pattern on the wafer, that is, the simulation contour, and the distance between the intersection of the pattern outline and the measurement axis is the OPC model's simulation of the critical dimension of the corresponding pattern on the wafer, that is, the simulation CD;

[0038] Step 2: Obtain the light intensity I by using the light intensity distribution I of each mask pattern and the critical size of the wafer meas , light intensity I meas The light intensity of each mask pattern at the location of the critical dimension endpoint is obtained by meas and light intensity I threshole The difference I err , taking all mask patterns as samples to obtain the difference I err The root mean square of the difference I err and the difference I err The root mean square of the OPC model is used to evaluate the optimization direction;

[0039] In some embodiments, the difference I in step 2 err The larger the RMS is, the larger the optimization space of the OPC model is. err Related.

[0040] In some embodiments, the difference I in step 2 err and the difference I err The methods for evaluating the optimization direction of the OPC model by the root mean square include: light intensity I background is the background light intensity of each mask pattern along the measurement axis, and the light intensity I max is the maximum light intensity of each mask pattern along the measurement axis, light intensity I min is the minimum light intensity of each mask pattern along the measurement axis; for the mask with 100% light transmittance inside the closed pattern in the corresponding layout, I background Less than I threshole , when the difference I err When it is a negative value, calculate the I of each mask pattern err / (I background -I threshole ), when the difference I err When it is a positive value, calculate the I of each mask pattern err / (I max -Ithreshole ), denoted as r diff ; For the mask with non-100% light-transmitting area inside the closed pattern in the corresponding layout, I background Greater than I threshole , when the difference I err When it is a positive value, calculate the I of each mask pattern err / (I background -I threshole ), when the difference I err When it is a negative value, calculate the I of each mask pattern err / (I min -I threshole ), denoted as r diff . r diff When it is greater than 0.5, it means that the optimization of the OPC model is difficult for the layout graphics.

[0041] In some embodiments, step 2 calculates r diff The percentage of measured axes with a value greater than 0.5 in the total number or the calculated r diff RMS, percentage and r diff The larger the RMS is, the more difficult it is to optimize the OPC model overall.

[0042] In some embodiments, if the OPC model is optimized by adding auxiliary terms and adjusting the coefficients of each term in step 2, it is necessary to first reduce or adjust the existing auxiliary terms to prevent overfitting.

[0043] For example, based on the OPC model A, the light intensity distribution I of each mask pattern is obtained, and the I of each mask pattern is obtained according to the critical dimension of the wafer. meas and I err =I meas -I threshole Calculation I err The rms of the reticle is 0.014. This can be used to compare with other OPC models. err / (I background -I threshole )(I err is a negative value) or I err / (I max -I threshole )(I err is a positive value), denoted as r diff Calculate r diff The rms of the OPC model is 0.393. The overall optimization space of this OPC model is large, and the overfitting risk is moderate. The existing auxiliary terms in OPC model A can be retained.

[0044] Step 3: Classify the mask pattern and obtain the light intensity I of each category of the OPC modelmeas and light intensity I threshole The difference I err and the difference I err The root mean square of the OPC model is used to achieve a detailed quantitative evaluation of the optimization direction of the OPC model.

[0045] In some embodiments, the mask pattern is classified according to the distance interval, the mask critical dimension, and the mask pattern type in step 3. It should be noted that in other embodiments, step 3 also includes classifying the mask pattern according to aspects such as pattern density.

[0046] In some embodiments, in step three, the mask patterns are classified according to the distance intervals through the use of calculable weights.

[0047] For example, see Figure 5 , with the position of the measured critical dimension on the measurement axis as the center, the mask plane is divided into areas with concentric circles of radius R1, R2, R3, and R4 respectively, that is, the distance is divided into (0, R1], (R1, R2], (R2, R3], (R3, R4], (R4, ∞) intervals. In order to distinguish between the classification based on distance interval and the classification based on mask critical dimension, the value of R1 must be greater than the minimum pitch (period) and the maximum mask critical dimension in all mask patterns, and when comparing the root mean square of different distance intervals, (0, R1] is not in the comparison range.

[0048] Figure 5 The mask pattern in is numbered i, and there are n distance intervals in total. When calculating the RMS of the distance interval (R1, R2], the weight wi of the mask pattern numbered i (R1,R2] yes

[0049]

[0050] Among them, di (R1,R2] yes Figure 5 The ratio of the area occupied by the middle mask pattern in the (R1, R2] region, d in is the ratio of the area occupied by the mask pattern outside (R1, R2], W (R1,R2] is the maximum value of the weight of the local mask pattern in (R1, R2] (must be a positive value), W n It is the maximum value of the weight of the local mask pattern outside (R1, R2]) (must be a positive value).

[0051] by Figure 5 For example, the length of the measuring axis (gauge) in the overlapping area between the mask pattern and (R1, R2] is recorded as G eff , requiring that G effWhen the ratio of R2-R1 is less than r, the weight of the mask pattern in (R1, R2] is less than or equal to 0. The maximum weight of the local mask pattern in (R1, R2] should be 1 / r times the maximum weight of the local mask pattern in other intervals, that is, W (R1,R2] / W n =1 / r.

[0052] If there are m measurement axes in total, then the root mean square (rms) of the distance interval (R1, R2] is

[0053]

[0054] Measurement axes with weights less than or equal to 0 do not participate in the RMS calculation.

[0055] Similarly, the rms of the distance interval (R2, R3] is

[0056]

[0057] The design of the maximum value W of the local weight of the mask pattern in each distance interval does not need to stick to the previous algorithm. However, care should be taken to avoid the situation where the weights of all mask patterns are less than or equal to 0 (this situation will occur when the design of W is unreasonable).

[0058] For example, considering the objective rule that the closer the surrounding light-transmitting points are to the sampling point, the greater the influence on the light intensity of the sampling point is, a cone curve (linear function) or a bell curve (such as a Gaussian function) with the center point as the axis of symmetry can be used as the constant value of W.

[0059] For another example, when the distance intervals are designed to be much larger than half of the critical dimension of the mask, W of all distance intervals may be set to the same value.

[0060] In some embodiments, in step three, the sample mask patterns are first screened to ensure that the sample set simultaneously meets the following conditions to ensure the accuracy of subsequent evaluation results: after classification by mask pattern type, the set of mask critical dimensions contained in each category is the same; after classification by mask pattern type and mask critical dimension, the set of periods contained in each category is the same.

[0061] In some embodiments, in step 3, the mask patterns are classified according to the distance intervals and the difference values ​​I of each category are calculated. err The larger the root mean square, the greater the amplitude by which the coefficient of the auxiliary item related to the mask pattern in the corresponding distance interval in the OPC model needs to be adjusted.

[0062] The length of the measuring axis in the overlapping area between the mask pattern and any distance interval is recorded as G. eff, requiring that G eff When the ratio of the distance interval is less than r, the weight of the mask pattern in the distance interval is less than or equal to 0. Therefore, when calculating the I err When the rms is , the maximum weight of the current distance interval is W0, and the maximum weights of other intervals are r*W0.

[0063] Then Figure 5 Take the mask pattern (numbered i) as an example, di (R1,R2] 、di (R2,R3] 、di (R3,R4] 、di (R4,∞] They are Figure 5 The ratio of the area occupied by the mask pattern in the region (R1, R2], (R2, R3], (R3, R4], (R4,,∞], is used in the rms calculation:

[0064] w i(R1，R2] =d i(R1,R2] *w0-(d i(R2，R3] +d i(R3,R4] +d i(R4,∞] ))*r*W0

[0065]

[0066] w i(R2，R3] =d i(R2，R3] *W0-(d i(R1，R2] +d i(R3，R4] +d i(R4,∞] )*r*W0

[0067]

[0068] w i(R3,R4] =d i(R3，R4] *W0-(d i(R1,R2] +d i(R2，R3] +d i(R4,∞] )*r*W0

[0069]

[0070] w i(R4,∞] =d i(R4，∞] *W0-(d i(R1，R2] +d i(R2，R3] +d i(R3，R4] )*r*W0

[0071]

[0072] For example, Figure 6 Shown is a schematic diagram of the root mean square calculated by classifying the layout graphics according to the distance interval of the present invention.

[0073] In some embodiments, in step 3, the mask patterns are classified according to the mask critical dimensions (mask CD) and the difference values ​​I of each category are calculated. err The larger the root mean square, the greater the amplitude by which the coefficients of the auxiliary items related to the mask pattern corresponding to the critical size of the mask in the OPC model need to be adjusted.

[0074] For example, Figure 7 The diagram shows the root mean square (RMS) calculated by classifying the layout patterns according to the critical dimensions of the mask according to the present invention.

[0075] In some embodiments, in step 3, the mask patterns are classified according to the types of the mask patterns and the difference values ​​I of each type are calculated. err The larger the root mean square, the greater the amplitude by which the coefficient of the auxiliary item related to the mask pattern of the corresponding mask pattern type in the OPC model needs to be adjusted.

[0076] For example, Figure 8 Shown is a schematic diagram of the root mean square calculated by classifying layout graphics according to the layout graphic type of the present invention.

[0077] Based on the above results, the optimization space of the OPC model is large, the overfitting risk is moderate, and the simulation of the optical effect in the (R2, R3] interval is mainly err The subsequent optimization can be done by adding strongly correlated auxiliary items in the (R2, R3] interval to the OPC model, or adjusting the coefficients of the existing strongly correlated auxiliary items.

[0078] It should be noted that the illustrations provided in this embodiment are only used to schematically illustrate the basic concept of the present invention. Therefore, the illustrations only show components related to the present invention and are not drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component can be changed at will, and the component layout type may also be more complicated.

[0079] In summary, the evaluation results obtained by this invention can provide a clearer direction for optimizing OPC models, shorten the modeling cycle, and help obtain OPC models that are more tailored to requirements, laying a solid foundation for excellent OPC solutions. Therefore, this invention effectively overcomes the shortcomings of existing technologies and has high industrial application value.

[0080] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by one of ordinary skill in the art without departing from the spirit and technical principles disclosed herein are intended to be covered by the claims of the present invention.

Claims

1. A quantitative evaluation method for an OPC model, characterized in that: At least: Step 1: Based on the OPC model, the light intensity distribution I of each mask pattern is obtained. The light intensity distribution I consists of necessary items, auxiliary items, and coefficients for adjusting each item. The extension line of the position of the wafer critical dimension on the mask pattern is the measurement axis of each mask pattern. The light intensity I on the light intensity distribution I is threshole The position is the OPC model simulation of the corresponding pattern outline of the mask pattern on the wafer, and the distance between the intersection of the pattern outline and the measurement axis is the OPC model simulation of the critical dimension of the corresponding pattern of the mask pattern on the wafer; Step 2: Obtain the light intensity I by using the light intensity distribution I of each mask pattern and the critical size of the wafer meas , light intensity I meas The light intensity of each mask pattern at the location of the critical dimension endpoint is obtained by meas and light intensity I threshole The difference I err , taking all mask patterns as samples to obtain the difference I err The root mean square of the difference I err and the difference I err The root mean square of the OPC model is used to evaluate the optimization direction; Step 3: Classify the mask pattern and obtain the light intensity I of each category of the OPC model meas and light intensity I threshole The difference I err and the difference I err The root mean square of the OPC model is used to achieve a detailed quantitative evaluation of the optimization direction of the OPC model.

2. The quantitative evaluation method of the OPC model according to claim 1, characterized in that: The difference I in step 2 err The larger the RMS is, the larger the optimization space of the OPC model is.

3. The quantitative evaluation method of the OPC model according to claim 1, characterized in that: The difference I in step 2 err and the difference I err The methods for evaluating the optimization direction of the OPC model by the root mean square include: light intensity I background is the background light intensity of each mask pattern along the measurement axis, and the light intensity I max is the maximum light intensity of each mask pattern along the measurement axis, light intensity I min is the minimum light intensity of each mask pattern along the measurement axis; for the mask with 100% light transmittance inside the closed pattern in the corresponding layout, I background Less than I threshole , when the difference I err When it is a negative value, calculate the I of each mask pattern err / (I background -I threshole ), when the difference I err When it is a positive value, calculate the I of each mask pattern err / (I max -I threshole ), denoted as r diff ; For the mask with non-100% light-transmitting area inside the closed pattern in the corresponding layout, I background Greater than I threshole , when the difference I err When it is a positive value, calculate the I of each mask pattern err / (I background -I threshole ), when the difference I err When it is a negative value, calculate the I of each mask pattern err / (I min -I threshole ), denoted as r diff , r diff When it is greater than 0.5, it means that the optimization of the OPC model is difficult for the layout graphics.

4. The quantitative evaluation method of the OPC model according to claim 3, characterized in that: Step 2: Calculate the r diff The percentage of measured axes with a value greater than 0.5 in the total number or the calculated r diff RMS, percentage and r diff The larger the RMS is, the more difficult it is to optimize the OPC model overall.

5. The quantitative evaluation method of the OPC model according to claim 4, characterized in that: If the OPC model is optimized by adding the auxiliary terms and adjusting the coefficients of each term in step 2, it is necessary to first reduce or adjust the existing auxiliary terms to prevent overfitting.

6. The quantitative evaluation method of the OPC model according to claim 1, characterized in that: In step three, the mask pattern is classified according to at least one of a distance interval, a mask critical dimension, and a mask pattern type.

7. The quantitative evaluation method of the OPC model according to claim 6, characterized in that: In step three, the mask patterns are classified according to the distance intervals by using the calculable weights.

8. The quantitative evaluation method of the OPC model according to claim 6, characterized in that: In step three, the sample mask patterns are first screened to ensure that the sample set meets the following conditions at the same time to ensure the accuracy of subsequent evaluation results: after classification by mask pattern type, the set of mask critical dimensions contained in each category is the same; after classification by mask pattern type and mask critical dimension, the set of periods contained in each category is the same.

9. The quantitative evaluation method of the OPC model according to claim 8, characterized in that: In step 3, the mask patterns are classified according to the distance interval and the difference I of each category is calculated. err The larger the root mean square, the greater the amplitude by which the coefficient of the auxiliary item related to the mask pattern in the corresponding distance interval in the OPC model needs to be adjusted.

10. The quantitative evaluation method of the OPC model according to claim 6, characterized in that: In step 3, the mask patterns are classified according to the key dimensions of the mask and the difference I of each category is calculated. err The larger the root mean square, the greater the amplitude by which the coefficients of the auxiliary items related to the mask pattern corresponding to the critical size of the mask in the OPC model need to be adjusted.

11. The quantitative evaluation method of the OPC model according to claim 1, characterized in that: In step 3, the mask patterns are classified according to the mask pattern type and the difference I of each type is calculated. err The larger the root mean square, the greater the amplitude by which the coefficient of the auxiliary item related to the mask pattern of the corresponding mask pattern type in the OPC model needs to be adjusted.

12. The quantitative evaluation method of the OPC model according to claim 1, characterized in that: Step three also includes classifying the mask patterns according to pattern density.

Images