KR20090071736A - Opc method by using etch bias modeling - Google Patents

Abstract

An optical proximity effect compensation method using the etch bias modeling is provided to previously design the layout considering the etching effect by using a secured etch bias model. The line width difference of wafer pattern(701) about the target line width is linearly in proportion according to the left and right spaces of wafer pattern. The etch bias density is linearly proportionate to the etch bias. The target layout, which is pattern-transferred to the wafer by the exposure and etching process, is designed. The etch bias model for predicting the etch bias in the etching process is constructed. The etch bias to be used for the etching process is predicted by using the etch bias model. The target layout is corrected by applying the predicted etch bias.

Description

Optical proximity effect correction method using etch bias modeling {OPC method by using etch bias modeling}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to semiconductor devices, and more particularly, to an optical proximity correction (OPC) method using etch bias modeling.

In order to integrate a semiconductor device, such as a memory device, onto a wafer, a lithography process of designing a target layout of a circuit pattern to be integrated on the wafer and transferring the pattern onto the wafer is performed. Is being performed. In order to more induce a wafer pattern to be transferred and formed on the wafer by the lithography process to match the designed target layout, an optical proximity effect correction (OPC) for correcting the target layout is performed.

As semiconductor devices are designed to be smaller and smaller, a resolution limit of an exposure process is generated, and resolution enhancement technology (RET) is introduced to overcome such extreme resolution. The introduction of this RET technology improves the resolution of certain parts of the actual pattern, while the optical proximity effect (OPE) is relatively large. To overcome this OPE, the OPC process is more precisely performed.

The process of forming the designed layout into a wafer pattern includes exposing a photoresist layer, developing the exposed portion to form a photoresist pattern, and selectively etching the etching target layer using the photoresist pattern as an etching mask. It is carried out including. Thus, factors to be considered in the OPC process are related to each of these pattern transfer processes. Therefore, the related optical proximity effect can be considered divided into a substantially pure optical proximity effect and a non-optical proximity effect. Pure optical proximity effects allow for relatively accurate modeling when illumination conditions and wavelengths of the exposure light sources are determined. In addition, among the factors causing the non-optical proximity effect, the elements related to the photoresist can be relatively predictable modeling due to the development of relatively many experiments and modeling methods.

On the other hand, the etching process, which is another factor that causes the non-optical proximity effect, is a process using plasma, which requires a great variety of variables to be modeled. Moreover, accurate prediction of the cross sectional portion, which is the three-dimensional pattern portion, is quite difficult, and also the portion regarding the two-dimensional pattern line width (CD) is difficult to accurately predict. Accordingly, a method of fitting using a mathematical black box model when modeling a photoresist without constructing a model of an actual etching process is used. In other words, the difference between the CD of the resist pattern and the CD of the etched pattern is simply applied by applying a Bessel polynomial and a kernel function. This mathematical approach uses a simple mathematical fitting function with no physical meaning, resulting in low predictive power on the extrapolation and relatively low fitting accuracy.

Accordingly, in order to apply the OPC to the target line width CD of the wafer pattern formed after etching, an approach through a hybrid OPC method has been introduced. This hybrid OPC is not a purely model-based OPC, but a combination of rule-based OPC and model-access OPC. In other words, the hybrid OPC uses the accuracy of the resist model, first corrects the target layout by an etch bias with a rule, and then corrects the corrected layout with the target of the resist pattern CD. This hybrid OPC has a weakness that requires a priori approach to the accuracy of the results. Accordingly, there is a limit to faster feed back and also a limitation to improvements. Moreover, there remains an inaccurate rule biasing issue for the unintended part of the rule approach OPC. Therefore, it is difficult to match the correct CD of the entire transistor.

The present invention is to obtain a model for estimating the etching bias by grasping the regularity of the etching phenomenon, and to propose a method of performing optical proximity effect correction using the obtained etching bias model.

의 수식으로 구축하는 단계; 상기 식각 바이어스 모델을 이용하여 상기 식각 과정에 수반될 식각 바이어스를 예측하는 단계; 및 상기 예측된 식각 바이어스를 적용하여 상기 목표 레이아웃을 보정하는 단계를 포함하는 식각 바이어스 모델링을 이용한 광학적 근접 효과 보정 방법을 제시한다. One aspect of the invention, the step of designing a target layout (target layout) to be transferred to the pattern on the wafer in the process of exposure and etching; An etch bias model is used to predict an etch bias that will be involved in the etching process.

Building with a formula; Predicting an etching bias to be involved in the etching process using the etching bias model; And correcting the target layout by applying the predicted etching bias.

The etch bias density (EB density) is a value proportional to the etch bias, the G (r) is a Gaussian kernel function representing a distribution of a polymer generated by the etching within the radius r during the etching, A (r, θ) may be set to an area of a radius r and azimuth angle θ that the polymer effectively affects, and L may be set to a total length of pattern edges to which the polymer is redeposited. .

According to an embodiment of the present invention, a model for predicting an etching bias may be secured and a method of performing optical proximity effect correction using the secured etching bias model may be provided.

In an embodiment of the present invention, by constructing an etch bias model, predicting a local bias difference with respect to the global bias variation associated with etching, and applying a local bias difference predicted during OPC progress on the layout How to do it. The global etch bias effect can be applied to the OPC at a constant value if the density of the wafer pattern is maintained at a specific specification.

In contrast, the generation of the line width (CD) difference due to the micro-loading effect on the individual patterns may have different specific trends depending on the etching method using plasma. The etching bias phenomenon is physically dependent on the etchant gas and the characteristics of the material layer to be etched and the thickness of the layer. In addition, the type of plasma generating electrode constituting the etching equipment, the applied power and the height of the electrode may act as variables affecting the etching bias phenomenon. However, it is practically difficult to obtain an etch model that reflects all of the variables depending on the characteristics of the equipment or the properties of the material due to the limitation of simulation time and software. Moreover, the OPC process requires rapid simulation and layout correction, making it difficult to obtain an etch model that reflects all variables. In an embodiment of the present invention, a phenomenologically matching factor is extracted and applied to modeling.

In an embodiment of the present invention, as an element affecting the etching bias, the effect of depending on the amount of distribution of polymers and the amount of re-deposition of these polymers according to the distance between adjacent patterns is considered.

1 and 2 are diagrams illustrating generation of an etched polymer and re-deposition of a polymer according to an embodiment of the present invention. 1 and 2, a resist pattern 200 is formed on an etching target layer 100, such as a conductive layer on a wafer, using an etch mask, and the region 101 exposed by the resist pattern 200 is formed. Plasma arrives at) to cause etching.

At this time, the etching polymer is generated and distributed by the etching reaction within the micro-loading range affecting the etching bias. As illustrated in FIG. 2, redeposition of the etch polymer generated in a region in the space S between wafer patterns 102, which is a pattern of an etch target layer formed by etching, is performed. In this etching method, the etching bias difference, that is, the line width difference of the wafer pattern 102 with respect to the target line width is linearly proportional to the left and right spaces of the wafer pattern 102. That is, considering the line and space pattern shape, the etching bias also increases in proportion to the size of the space S. However, when considering an etching process in which a lot of polymers are generated, a polymer redeposition phenomenon occurs around the pattern 102, and the polymer redeposition phenomenon has an effect of suppressing an etching bias. As such, the distribution of polymers affecting the etching bias follows a Gaussian distribution (201).

3 to 6 are diagrams for explaining the factors affecting the etching bias according to an embodiment of the present invention. Referring to FIG. 3, the etching bias at the calculation point for which the etching bias on the first wafer pattern 301 is to be predicted is influenced by the amount of polymer generated in the space with the directly adjacent neighboring second pattern 302. However, the etching bias tends to increase in proportion to the amount of such polymer.

In consideration of this tendency, the range in which the amount of polymer generated during etching affects the etching bias may be considered as the microloading range r because the microloading effect is affected. The distribution range of the polymer, which can be considered to affect the etch bias as such, can be taken into account in the area within the circle 305 with the radius of this mask loading range r. In this case, the distribution of the polymer may be considered as the Gaussian distribution 201 as shown in FIG. 2, and the amount of the polymer is increased as the polymer is generated near the calculation point. In other words, considering that the closer to the calculation point in the area within the circle 305 at the calculation point to calculate the effect of the distribution of the polymer has a greater effect, the influence on the etching bias by the polymer is Gaussian The kernel function can be applied.

As shown in FIG. 4, considering the curved third pattern 401 and the neighboring fourth pattern 402, the polymer that affects the etch bias at the calculation point on the third pattern 401 may be a polymer. Considering the linearity of the influence, it may be considered as the amount distributed in the effective etching area 407 in which the portion 406 that does not affect the etching bias is removed. At this time, the area of the effective area 407 may be obtained by integrating the area function A (r, θ) of the radius r and the azimuth angle θ. In addition, as shown in FIG. 5, considering the curved fifth pattern 501 and the neighboring sixth pattern 502, the polymer affecting the etching bias at the calculation point on the fifth pattern 501 may be In consideration of the effect of being blocked by the bent portion 511 of the fifth pattern 501, that is, a blocking effect, it may be considered as an amount distributed in the effective area 507.

With this in mind, considering the ratio of the area of the effective region 407 or 507 to the microloading range r as the etch bias density, the etch bias density is linearly proportional to the etch bias. do. That is, as the area is increased, the amount of polymer tends to increase, so that the value of the etching bias also increases. This tendency faithfully represents the phenomenon that the etching bias increases in proportion to the size of the space in the line and space patterns.

However, in addition to the blocking effect, as shown in FIG. 6, when considering the curved seventh pattern 601 and the neighboring eighth pattern 602, the etching bias at the calculation point on the seventh pattern 601 is affected. In addition to the amount of the polymer, the factor that affects the degree of consideration can be taken into account. As shown in FIG. 5, as the blocking effect represents the action of reducing the etching bias, as shown in FIG. 6, the more pattern edges 611 within the same area, the greater the amount of redeposited polymer produced. As a result, the tendency of the etching bias is reduced by the layer 630 of the deposited polymer. The etching bias reduction due to the redeposition of this polymer is consistent with the fact that at comparable pattern densities, the etching bias occurs relatively smaller in the pattern having a smaller line width (CD).

As described above, in the exemplary embodiment of the present invention, an etch bias density proportional to the etch bias is determined, and a blocking effect in which the influence of the polymer is limited by the amount of the etched polymer affecting the calculation point and other patterns is determined. In consideration of the effective area calculated in consideration and the length of the edge (edge) of the pattern that the polymer is redeposited, it can be calculated by Equation 1.

Here, the etching bias density (EB density) means a value proportional to the etching bias, and reflects the effective etching area (effective etching area) in consideration of the blocked area (blocked Area). The length L also reflects the length of the total edge over which the polymer is redeposited. The distribution of the polymer is based on the Gaussian kernel G (r) at the effective area according to the azimuth angle θ with the microloading range r as the radius. The model of the etching process is constructed by Equation 1.

7 and 8 are views provided to explain an example of applying an etching bias model according to an embodiment of the present invention. FIG. 7 illustrates the application of the etching bias model according to the embodiment of the present invention in a relatively high density pattern layout, and FIG. 8 illustrates the application of the etching bias model in a relatively low density pattern layout. The area influencing according to the microloading range r at the calculation point on the pattern 701 of FIG. 7 can be considered as the circle 702, and the effective area 703 in which the effect of the actual polymer is effective can be considered. have. In addition, the length of the edge 704 representing the redeposition of the polymer may also be considered. These factors are applied to Equation 1 to calculate the etch bias density at the calculation point.

In comparison, the area of influence of the microloading range r at the calculation point on the relatively low density pattern 801 of FIG. 8 can be considered as the circle 802, and the effective area where the effect of the actual polymer is effective. 803 may be considered. In addition, the length of the edge 804 representing the redeposition of the polymer may also be considered. These factors are applied to Equation 1 to calculate the etch bias density at the calculation point. 7 and 8, in the case of FIG. 8, which is a low density pattern layout, the length of the edge 804 may be smaller than that of FIG. 7. Accordingly, the etching bias density calculated by Equation 1 is calculated as a larger value in FIG. 8, which substantially corresponds to the tendency of larger etching bias in the low density pattern experimentally.

In an embodiment of the present invention, a model is constructed by dividing factors depending on etching characteristics into long range and short range effects. Narrow range effects depend on the polymers occurring at the distance between the nearest patterns and the amount of redeposition of these polymers to determine the effect on the etching bias. Therefore, the specific distance affecting the etching bias, that is, the pattern loading at the microloading range and the etching bias calculation point, determines the etching bias characteristics. In addition, the direction in which the etching bias is limited due to the presence of the neighboring pattern or the polymer generated in the surrounding area is reduced as the length of the adjacent pattern edge increases, and the directivity of the polymer is also considered in calculating the etching bias. In the present invention, this effect is constructed as a model using a Gaussian kernel function. The widespread effect is a factor that is substantially dependent on the wafer open area, which substantially determines the total polymer amount. In the exemplary embodiment of the present invention, an etching bias model may be constructed in consideration of all these effects, and thus a more accurate and stable etching model may be implemented.

In an embodiment of the present invention, the local etching bias may be calculated using the etching model constructed as described above, and the calculated etching bias may be reflected in the OPC process. Accordingly, a more accurate etching bias can be calculated, thereby enabling OPC for FI (Field Inspection) CD, which is the measurement line width for the wafer pattern. In addition, a simulation contour using an optical model of a generally known exposure process and a resist model of a resist and an etch bias calculation value obtained in an embodiment of the present invention are used. Modeling can be applied. Using this model, accurate pattern formation can be predicted in advance without experimental implementation. In particular, by considering the etching effect due to the pattern, it is possible to design a layout in consideration of the etching effect in advance in the design stage.

1 is a view for explaining the generation of etched polymer (etched polymer) according to an embodiment of the present invention.

2 is a view provided to explain the redeposition of the etching polymer according to an embodiment of the present invention.

3 to 6 are diagrams for explaining the factors affecting the etching bias according to an embodiment of the present invention.

7 and 8 are views provided to explain an example of applying an etching bias model according to an embodiment of the present invention.

Claims (1)

Designing a target layout to pattern-transfer on the wafer by an exposure and etching process; Constructing an etch bias model to be obtained by Equation 2 to predict an etch bias to be involved in the etching process; Predicting an etching bias to be involved in the etching process using the etching bias model; And And correcting the target layout by applying the predicted etch bias.

The etch bias density (EB density) is a value proportional to the etch bias, the G (r) is a Gaussian kernel function representing a distribution of a polymer generated by the etching within the radius r during the etching, Where A (r, θ) is the area of radius (r) and azimuth (θ) that the polymer effectively affects, and L is the total length of the pattern edges that the polymer re-deposits.

Images