阅读说明：本技术 分析光刻掩模的缺陷位置的方法与设备 (Method and apparatus for analyzing defect positions of a lithographic mask ) 是由 M.布达科 R.肖恩伯格 于 2018-03-01 设计创作，主要内容包括：本发明涉及一种分析光刻掩模(200、700)的至少一个缺陷位置(230、730)的方法,该方法具有下列步骤：(a)获得该光刻掩模(200、700)的至少一个缺陷位置(230、730)的测量数据；(b)从该光刻掩模(200、700)的计算机辅助设计(CAD)数据(300)决定该缺陷位置(230、730)的参考数据；(c)使用至少一个位置相关修正值来修正该参考数据；以及(d)通过比较该测量数据与该已修正参考数据来分析该缺陷位置(230、730)。(The invention relates to a method for analyzing at least one defect location (230, 730) of a lithography mask (200, 700), the method having the following steps: (a) obtaining measurement data of at least one defect location (230, 730) of the lithography mask (200, 700); (b) determining reference data for the defect location (230, 730) from Computer Aided Design (CAD) data (300) of the lithographic mask (200, 700); (c) correcting the reference data using at least one position-dependent correction value; and (d) analyzing the defect location (230, 730) by comparing the measurement data with the corrected reference data.)

1. A method for analyzing at least one defect location (230, 730) of a lithography mask (200, 700), wherein the method comprises the steps of:

a obtaining measurement data of at least one defect location (230, 730) of the lithography mask (200, 700);

b determining reference data for the defect location (230, 730) from Computer Aided Design (CAD) data (300) of the lithographic mask (200, 700);

c correcting the reference data using at least one position-dependent correction value; and

d analyzing the defect location by comparing the measurement data with the corrected reference data (230, 730).

2. A method for analyzing at least one defect location (230, 730) of a lithography mask (200, 700), wherein the method comprises the steps of:

a obtaining measurement data of at least one defect location (200, 730) of the lithography mask (200, 700);

b determining reference data for the defect location (230, 730) from Computer Aided Design (CAD) data (300) of the lithographic mask (200, 700);

c determining a profile (1030) of the at least one defect location (230, 730) from the measurement data and the reference data; and

d analyzing the defect location (230, 730) by correcting the contour (1030) of the at least one defect location (230, 730) using at least one location dependent correction value.

3. The method of claim 1 or 2, wherein obtaining measurement data of at least one defect location (230, 730) of the lithography mask (200, 700) comprises: the at least one defect location (230, 730) is scanned using a particle beam (1005).

4. The method of any one of the preceding claims, further comprising the steps of: a measurement data image (250) is generated from the measurement data.

5. The method of any one of the preceding claims, wherein determining the reference data comprises: a section is extracted from CAD data (300) including at least one defect location (230, 730) of the lithography mask (200, 700).

6. The method of any one of the preceding claims, wherein determining the reference data comprises: a reference image (400) is synthesized from the CAD data (300).

7. The method of claim 6, wherein synthesizing the reference image (400) comprises: consider the systematic variation of this reference image (400) experienced by CAD data (300) during the mask manufacturing process.

8. The method of claim 6 or 7, wherein synthesizing the reference image (400) further comprises: a light boundary (460) for at least one structuring element (420) in the reference image (400) is generated.

9. The method of claim 8, wherein the at least one structural element (420) comprises an element from the group consisting of: a pattern element (220, 720) of the lithography mask (200, 700), a marking of the lithography mask (200, 700) and a defect (240, 740) of at least one defect location (230, 730) of the lithography mask (200, 700).

10. The method of any of claims 6 to 9, wherein synthesizing the reference image (400) further comprises: parameters of a point spread function describing an exposure process of the lithographic mask (200, 700) during a mask manufacturing process are determined.

11. The method of any of the preceding claims, wherein the at least one position-dependent correction value takes into account a difference between a Critical Dimension (CD) (225) of the measurement data and the CAD data (300) at the defect position (230, 730).

12. The method of any one of claims 4 to 11, further comprising the steps of: for each of at least one defect location (230, 730) of the lithography mask (200, 700), the at least one location-dependent correction value is determined by minimizing a difference between a CD (225) of a structuring element (220) of the measurement data image (250) and a CD (425) of a structuring element (420) of the reference image (400).

13. The method of claim 12, wherein minimizing a difference between the CD (225) of the structure element (220) of the measurement data image (250) and the CD (425) of the structure element (420) of the reference image (400) is performed within an area (250, 750) of the lithographic mask (200, 700) around the at least one defect location (230, 730), wherein the area (250, 750) excludes the at least one defect location (230, 730), and wherein the area (250, 750) is larger than the at least one defect location (230, 730).

14. The method of claim 13, wherein minimizing a difference between a CD (225) of a structural element (220) of the measurement data image (250) and a CD (425) of a structural element (420) of the reference image (400) comprises: parameters of a point spread function within an area (250, 750) around the at least one defect location (230, 730) are determined such that a difference between the CDs of the structural elements of the measurement data image (250) and the reference image (400) becomes minimal.

15. The method according to any one of claims 12 to 14, wherein determining the at least one position-related correction value comprises: an algorithm is performed that minimizes the difference between the CD (225) of the structural element (220) of the measurement data image (25) and the CD (425) of the structural element (420) of the reference image (400).

16. The method of any one of claims 12 to 15, further comprising the steps of: determining a distribution of at least one pattern element (220) and/or marked CD (225) on the lithographic mask (200, 700).

17. The method of claim 16, further comprising the steps of: at least one position-dependent correction value is determined from the distribution of the at least one pattern element (220) and/or the marked CD (225), and a corrected reference image (800) is generated with the aid of the generated distribution of the CD (225).

18. The method of any of claims 6 to 17, wherein modifying the reference data comprises: a corrected reference image (800) is formed by correcting the reference image (400) using the at least one position-dependent correction value.

19. The method of any of claims 1 or 3 to 18, wherein analyzing the at least one defect location (230, 730) comprises: determining an outline (930) of the at least one defect (240) at the at least one defect location (230).

20. The method of claim 19, wherein determining the profile (930) comprises: the corrected reference data of the defect location (230, 730) and the measurement data of the defect location (230, 730) are superimposed and a difference between the superimposed corrected reference data and the measurement data is identified.

21. The method of claim 2, wherein determining the contour (1030) comprises: the reference data of the defect location (230, 730) and the measurement data of the defect location (230, 730) are superimposed and a difference between the superimposed reference data and the measurement data is identified.

22. A computer program comprising instructions which, when executed by a computer system (1085), cause the computer system (1085) to carry out the method steps according to any one of claims 1 to 21.

23. An apparatus (1100) for analyzing at least one defect location (230, 730) of a lithography mask (200, 700), comprising;

a means for obtaining measurement data of at least one defect location (230, 730) of the lithographic mask (200, 700);

b determining means for determining reference data for the defect location (230, 730) from Computer Aided Design (CAD) data (300) of the lithographic mask (200, 700);

c correction means for correcting the reference data using at least one position-dependent correction value; and

d analyzing means for analyzing the defect location (230, 730) by comparing the measurement data with the corrected reference data.

24. An apparatus (1100) for analyzing at least one defect location (230, 730) of a lithography mask (200, 700), comprising;

a means for obtaining measurement data of at least one defect location (230, 730) of the lithographic mask (200, 700);

b determining means for determining reference data for the defect location (230, 730) from Computer Aided Design (CAD) data (300) of the lithographic mask (200, 700);

c determining means for determining a profile (1030) of the at least one defect location (230, 730) from the measurement data and the reference data; and

d analyzing means for analyzing the defect location (230, 730) by correcting the contour (1030) using at least one location dependent correction value.

Technical Field

The present invention relates to a method and apparatus for analyzing defect locations of a lithographic mask.

Background

Due to the ever-increasing integration density of the semiconductor industry, photolithographic masks must form smaller and smaller structures on a wafer. In order to image small feature sizes on wafers, increasingly complex processing procedures are therefore required. These must in particular ensure that the untreated semiconductor material is not altered accidentally and/or in an uncontrolled manner as a result of the treatment process.

In the case of lithography, the trend of increasing integration density is addressed by shifting the exposure wavelength of the lithography system to shorter wavelengths. Currently, ArF (argon fluoride) excimer lasers, which emit at a wavelength of about 193nm, are often used as light sources in lithography systems.

Lithography systems using electromagnetic radiation in the EUV (extreme ultraviolet) wavelength range, preferably in the range of 10nm to 15nm, are currently being developed. The EUV lithography system is based on a completely new beam guiding concept which uses reflective optical elements without exception, since no material is currently available which is optically transparent in the EUV range. The technical challenges faced in developing EUV systems are considerable and considerable development effort is required to bring the system to a level useful for industrial applications.

A photolithographic mask, exposure mask, photomask or simply mask has a significant impact on the imaging of smaller and smaller structures in the photoresist disposed on the wafer. With each increase in integration density, it becomes more important to improve the minimum feature size of the exposure mask. Consequently, the manufacturing process of the photolithographic mask becomes more and more complex and therefore more time consuming and eventually also more expensive. Due to the minute structural dimensions of the pattern elements, defects during mask production cannot be taken into account, which must be corrected as far as possible.

Before defects of a lithographic mask can be corrected, the defects must be located and analyzed. One form of defect analysis is to determine the profile of the defect. To determine the outline of a defect, the defective area of the photomask is typically compared to an equivalent area (equivalent region) of a mask that is not defective. The equivalent area corresponding to the defective area is a region of the photomask having the same arrangement of pattern elements as the defective area, and there is no defect in the equivalent area.

The partial image at the top of fig. 1 shows a section or image field recorded by a scanning electron microscope with defects extending over a plurality of pattern elements in the absorber structure. In the lower left peripheral area, a rectangle representing an equivalent area or a reference area is marked, which is larger in size than the defect but has no defect. The partial image shown as an arrow at the bottom of fig. 1 shows the superposition (superimposition) of the defect-free reference region and the defect region recorded by the scanning electron microscope. From the superposition of the two image sections, the complete contour of the defect, which is emphasized by the brighter color, can be determined.

In the example of fig. 1, a suitable reference region is advantageously present together with the defect in the same image field recorded by the scanning electron microscope. If this is not the case, the reference area may be searched and used at different locations on the photomask. However, manually searching for a suitable reference area is rather time consuming. Furthermore, an operator of a scanning electron microscope may erroneously select a reference region of the pattern element structure having a similar appearance. Furthermore, it is possible that there is no second area of the photomask having a pattern of the same pattern elements as the defective area, which may prevent inspection of the defect, or at least make it more difficult to inspect.

The patent documents mentioned below describe methods to alleviate the above problems: EP 983193 a1, US 2003/0174876 a1, US 5849440 and US 5916716.

However, these documents do not relate to repairing defects of the photolithographic mask.

It is therefore the problem underlying the present invention to specify an improved method and an improved apparatus for analyzing defect locations of a lithography mask.

Disclosure of Invention

According to an exemplary embodiment of the present invention, this problem is solved by a method as claimed in claim 1. In one embodiment, a method for analyzing at least one defect location of a lithography mask includes the steps of: (a) obtaining measurement data of at least one defect location of the lithographic mask; (b) determining reference data for the defect location from computer-aided design (CAD) data of the lithographic mask; (c) correcting the reference data using at least one position-dependent correction value; and (d) analyzing the defect location by comparing the measurement data with the corrected reference data.

Because systematic variations experienced by the CAD data during the mask manufacturing process are considered globally (globally) and locally (locally) during defect location inspection, enabling defect analysis to be performed more accurately. First, the false detection rate of defects is thus reduced, thereby reducing the complexity of mask inspection. Second, the identified defects can be repaired in a more precise manner, thereby improving mask productivity.

Furthermore, since defect inspection (die-to-database inspection) is performed with the aid of CAD data, the defect position of the lithography mask can also be inspected under no usable equivalent area or no reference area. Furthermore, manual searching of the reference area and erroneous comparison of the defect location with a mask portion of a pattern having similar pattern elements may be omitted.

Furthermore, since the above-mentioned manual search of the reference area has been eliminated, the use of CAD data for determining position-dependent correction values for correcting the reference data will contribute to a better automated analysis process.

Finally, production inaccuracies of the used reference areas (key points: "line edge roughness") and measurement inaccuracies of the data records relating to the used reference areas are omitted when comparing defect locations with CAD data. By using CAD data instead of reference areas, a lower variation of the inspection defect location results can be achieved.

In the present application, the "position-dependent" property means that the correction values are not constant over the area of the lithography mask, but depend on the observed mask position.

In another embodiment, a method for analyzing at least one defect location of a lithography mask includes the steps of: (a) obtaining measurement data of at least one defect location of the lithographic mask; (b) determining reference data for the defect location from computer-aided design (CAD) data of the lithographic mask; (c) determining a profile (contour) of the at least one defect location from the measurement data and the reference data; and (d) analyzing the defect location by modifying the profile of the at least one defect location using at least one location dependent correction value.

In a second embodiment of the method according to the invention, one or more position-dependent correction values are not used for correcting the reference data, but for correcting a defect position profile or a profile of a photomask defect, which profile has been determined from the measurement data and the reference data. Thus, the second embodiment applies a (global) bias (bias) to the reference data and one or more position-dependent correction values to the determined profile for the defect or defect position.

The advantages of the method according to the invention described above also apply at least in part to the second exemplary embodiment. This second embodiment is particularly useful when it is not possible to directly acquire the reference data.

Obtaining measurement data of at least one defect location of the lithographic mask may include scanning the at least one defect location using a particle beam.

Obtaining the measurement data may further include: coordinates of the at least one defect location are obtained from an inspection tool for the lithographic mask. The particle beam may comprise an electron beam.

On the basis of the coordinates obtained from the inspection tool, a particle beam, preferably an electron beam, can be used to analyze the defect locations of the lithography mask and the area around the defect locations in detail.

The method according to the invention may further comprise the steps of: a measurement data image is generated from the measurement data. The measurement data image may include a two-dimensional representation of the measurement data on a monitor.

Determining the reference data may include: a section including at least one defect location of the lithographic mask is extracted from the CAD data.

For retrieving the section of the CAD data, the coordinates of the defect location supplied by the inspection tool may be used. The required CAD data segments may also be determined based on the photomask area scanned by the particle beam.

Determining the reference data may include synthesizing a reference image from the CAD data.

The method according to the invention may be performed on the basis of data, i.e. measurement data, reference data and corrected reference data, or on the basis of images, i.e. measurement data images, reference images and corrected reference images.

Generating or synthesizing reference images from CAD data is commonly referred to in the art as "rendering".

Synthesizing the reference image may include accounting for reference image system variations experienced by the CAD data during the mask manufacturing process.

During transfer to the photomask, several factors cause changes in the CAD data: an exposure step, development of the photoresist, etching of the pattern elements of the absorber structure, and imaging or measurement of the resulting pattern elements. This list is not exhaustive. For example, the first three factors may result in rounded corners in the pattern elements generated on the mask. The above-mentioned factors are further responsible for the generation of deviations, i.e. dimensional deviations of the structural elements of the CAD data and the measurement data.

Compositing the reference image may further include generating an optical boundary for at least one structuring element of the reference image.

The images produced by exposure with the electron beam have material contrast and topographical contrast (topographies) that preferably arises at the edges and/or corners of the structural elements. For this reason, the structural elements in Scanning Electron Microscope (SEM) images typically have optical boundaries that are better used to align the two images during overlay. For the reasons described above, the light edges of the imaged structuring elements do not correspond to the edges of the structuring element edges generated from the CAD data.

The at least one structural element may comprise an element from the group: a pattern element of a lithography mask, a mark of a lithography mask, and a defect of at least one defect location of a lithography mask.

The marks of the mask may be used to adjust the beam relative to the defect location.

The synthetic reference image may further comprise parameters determining a point spread function describing a lithographic mask exposure process during a mask manufacturing process.

Point Spread Function (PSF) is often used to describe the interaction of a particle beam with the absorbing and/or phase shifting material of a photoresist and/or photomask during exposure.

The point spread function may comprise an addition of at least two gaussian distributions. The first gaussian distribution may describe particle forward scattering of particle beams in a photoresist disposed on and/or within the lithographic mask and the at least one second gaussian distribution may describe particle backward scattering of particle beams in a photoresist disposed on and/or within the lithographic mask.

The point spread function may further be designed to take into account bias asymmetry with respect to a first direction and a second direction in the plane of the lithographic mask, wherein the first direction and the second direction preferably form a right angle. The combination of two or more gaussian distributions described above may take into account this bias asymmetry between the structuring element of the CAD data and the photomask.

This design of the point spread function makes it possible to take into account local asymmetry variations of the CAD data during the manufacture of the pattern elements of the mask in the defect location analysis.

The at least one location dependent correction value may take into account a difference between the measured data and a Critical Dimension (CD) of the CAD data at the location of the defect.

Since the one or more correction values are not considered constant on the mask but depend on the position on the photomask, local adaptation can be performed to the above-mentioned systematic variations of the CAD data during transfer onto the photomask. Therefore, the correspondence between the measurement data of the defective area and the corrected reference data generated for the defective area can be improved.

The method for analyzing defect locations of a lithography mask may further comprise the steps of: at least one position-dependent correction value is determined by minimizing the difference between the CD of the structural elements of the measurement data image and the CD of the structural elements of the reference image for each of at least one defect position of the lithography mask.

Minimizing the difference between the CDs of the structural elements of the measurement data image and the reference image may be performed within a region of the lithographic mask around at least one defect location, wherein the region excludes the at least one defect location, and wherein the region is larger than the at least one defect location.

The defective location area having the defect can be excluded from the image comparison. This may be advantageous because there are already design deviations in the defect area and the defect may be a disturbance of the image comparison.

Minimizing the difference between the CDs of the structural elements of the measurement data image and the reference image may comprise determining parameters of a point spread function in the area around the at least one defect location, whereby the difference between the CDs of the structural elements of the measurement data image and the reference image becomes minimal.

Since the parameters of the point spread function describing the mask manufacturing process are set to be location dependent and determined locally for each defect location individually, the correspondence between the measured data image and the associated reference image for each defect is optimized. Thus, defects can be accurately inspected and repaired.

In the illustrated exemplary embodiment, the determination of the position-related correction values is a first step in the defect position verification process.

Determining the at least one position-dependent correction value may comprise performing an algorithm that minimizes the difference between the CD of the structural element in the measurement data image and the CD of the structural element in the reference image.

If a defect location is too large to fill the largest area to the extent that an undisturbed CD decision for the structural elements of the measurement data image is no longer possible or difficult to predict, the area next to the defect location is used to minimize the CD difference between the structural elements of the measurement data image and the structural elements of the reference image. The maximum area is determined, for example, by the maximum number of image points or pixels stored in the image memory. This maximum region can be further limited by the maximum deflection of the particle beam of the apparatus for analyzing the lithography mask when observing the prescribed imaging aberrations.

The method according to the invention may further comprise the steps of: a CD profile of at least one pattern element and/or mark on the lithographic mask is determined.

This process represents an alternative to minimizing the CD of the structural elements of the measurement data image and the reference image. It is necessary to measure the CD distribution over the photomask before inspecting the mask for the defect location(s). On the other hand, this alternative allows the distribution of the correction values to be confirmed before the analysis processing of the defect mask is performed.

Furthermore, the method according to the invention comprises the following steps: determining at least one position-dependent correction value from the CD distribution of at least one pattern element and/or mark; and generating a corrected reference image by means of the generated distribution of CDs.

Determining at least one position-dependent correction value may include determining a position-dependent parameter of a point spread function.

Correcting the reference data may include correcting the reference image by using at least one position-dependent correction value to create a corrected reference image.

The position-dependent correction values can be described analytically as two-dimensional functions. However, this is usually sufficient to divide the photomask into suitable regions, for example into rectangles and squares, and to assume that the correction values within the defined regions are constant. For example, the critical dimension distribution of pattern elements and/or marks on a photomask may be measured prior to analyzing the photomask for defect locations. The mask area can thus be defined on the basis of the determined CD distribution, wherein the CD and thus the one or more position-dependent correction values are considered to be constant. Thus, for example, correction values for different areas of the photomask are calculated and stored in a table. At the start of a defect position inspection program for a photomask, one or more correction values which are locally valid are retrieved from the table and used to correct the reference data or the reference image, respectively.

Analyzing the defect location may include overlaying the measurement data image with the corrected reference image and identifying differences between the overlaid images. Identifying the difference between the superimposed images may include generating a difference image by subtracting the corrected reference image from the measurement data image. The overlay of the measurement data image and the corrected reference image may further comprise an area of at least one defect determining the location of the at least one defect in a pixel-wise manner.

Analyzing the at least one defect location may include determining an outline of the at least one defect at the at least one defect location. Analyzing the at least one defect location further may include determining a repair shape of the at least one defect at the at least one defect location.

Determining the profile may include overlaying reference data for the defect location and measurement data for the defect location and identifying a difference between the overlaid reference data and measurement data. Identifying the difference between the superimposed reference data and the measured data may include generating a difference image by subtracting the reference image from the measured data image. Superimposing the measurement data image with the reference image may further comprise determining at least one defect in a pixel-wise manner (pixel-wise manner) of the location of the at least one defect. Correcting the contour may include correcting the contour in a pixel-wise manner using at least one position-dependent correction value.

In this embodiment, the one or more position-dependent correction values are not applied to the reference data, but to the contour of the defect position or to the contour of the defect.

Determining at least one position-related correction value may include: defining a quality factor (figure of merit) between the measured data of at least one defect location and the reference data of the defect location; and minimizing the quality factor by changing the measurement data of the at least one defect location.

The figure of merit may include a Critical Dimension (CD) of the measurement data and a Critical Dimension (CD) of the reference data, and minimizing the figure of merit may include adapting the CD of the measurement data to the CD of the reference data.

The quality factor may comprise a CD of the at least one structural element of the measurement data image and a CD of the at least one structural element of the reference data image, and minimizing the quality factor may comprise adapting the CD of the at least one structural element of the measurement data image to the CD of the at least one structural element of the reference data image.

The figure of merit may include spacing (spacing) between structural elements of the measurement data image and structural elements of the reference data image, and minimizing the figure of merit may include maximizing superposition of the structural elements of the measurement data image and the structural elements of the reference data image.

The quality factor may comprise parameters of a point spread function, and minimizing the quality factor may comprise changing the parameters of the point spread function such that the measurement data maintains a correspondence with reference data in an area surrounding the at least one defect location.

Determining at least one position-related correction value may include: minimizing a quality factor within a region of the lithographic mask around at least one defect location, wherein the region excludes the at least one defect location, and wherein the region is larger than the at least one defect location.

The method according to the invention may further comprise the steps of: at least one defect at a defect location is repaired using the profile determined for the at least one defect or using the profile corrected with the at least one location dependent correction value. Repairing the at least one defect at the at least one defect location may include: depositing an absorbing and/or phase shifting material onto a substrate of a lithographic mask or onto a cladding of a lithographic mask for the Extreme Ultraviolet (EUV) wavelength range, and/or removing an absorbing and/or phase shifting material from a substrate of a lithographic mask or from a cladding of a lithographic mask for the Extreme Ultraviolet (EUV) wavelength range. Repairing at least one defect at least one defect location may further include providing at least one particle beam and at least one precursor gas (precursor) at the defect location. Furthermore, the at least one particle beam may comprise an electron beam, and the electron beam is used not only for repair, but also for scanning the at least one defect location.

According to various aspects, a computer program comprises instructions which, when executed by a computer system, cause the computer system to perform the method steps of the above-specified aspects.

Furthermore, in a different exemplary embodiment, the above-mentioned problem is solved by an apparatus according to claim 23. In one embodiment, an apparatus for analyzing at least one defect location of a lithographic mask includes: (a) obtaining means for obtaining measurement data of at least one defect location of the lithographic mask; (b) determining means for determining reference data for the defect location from Computer Aided Design (CAD) data of the lithographic mask; (c) correction means for correcting the reference data using at least one position-dependent correction value; and (d) analyzing means for analyzing the defect location by comparing the measurement data with the corrected reference data.

Finally, in a further exemplary embodiment, an apparatus for analyzing at least one defect location includes: (a) obtaining means for obtaining measurement data of the at least one defect location; (b) determining means for determining reference data for the defect location from Computer Aided Design (CAD) data of the lithographic mask; (c) determining means for determining a profile (930) of the at least one defect location from the measurement data and the reference data; (d) and an analyzing means for analyzing the defect position by correcting the profile using at least one position-dependent correction value.

Drawings

The presently preferred exemplary embodiments of the invention are described in the detailed description below with reference to the drawings, in which:

the top half image in FIG. 1 illustrates a cross-section of a Scanning Electron Microscope (SEM) image of a photolithographic mask in which defect locations are located extending over a plurality of pattern elements and having defect-free reference regions represented by rectangles in the bottom left portion; and

the lower half image represents the defect locations of the lithographic mask where the reference areas (represented by the arrow symbols) of the upper half image overlap the upper half image.

FIG. 2 presents a plan view of an SEM recording cross-section of a lithographic mask having defect locations;

FIG. 3 presents a CAD data plan view of the portion of the photolithographic mask profile of FIG. 2 that includes the location of the defect;

FIG. 4 shows a reference image synthesized from the CAD data of FIG. 3;

FIG. 5 illustrates the principle of the mask manufacturing process;

FIG. 6 schematically represents a variation of Critical Dimensions (CDs) over a lithographic mask;

FIG. 7 presents a cross-section from a photomask illustrating defect locations having defects in the form of a missing absorbing material;

FIG. 8 shows the reference image of FIG. 4 having been corrected with position-dependent correction values;

FIG. 9 depicts an overlay of the corrected reference image of FIG. 8 and the measured data image of FIG. 2;

FIG. 10 depicts an overlay of the reference image of FIG. 4 and the measured data image of FIG. 2;

FIG. 11 shows a diagrammatic cross-section through an apparatus whereby defect locations of a lithographic mask may be analyzed and repaired;

FIG. 12 depicts a flow chart of a first embodiment of a method according to the invention; and

fig. 13 shows a flow chart of a second exemplary embodiment of a method according to the present invention.

Detailed Description

Preferred embodiments of the method according to the invention and of the apparatus according to the invention are explained in more detail below. These embodiments are discussed with reference to an example of opaque defect analysis of a binary transmissive lithography mask. However, the application of the method according to the invention and the application of the apparatus according to the invention are not limited to binary transmissive lithographic masks. Rather, they can be used to analyze any reflective and transmissive mask. Furthermore, the methods described herein and the apparatus described herein may be used to inspect template defects for imprint lithography. Furthermore, the apparatus according to the invention and the method according to the invention can be used for analyzing defects on a substrate if the substrate has at least one structural element in the vicinity of the defect location and the inspection of the defect location is performed on the basis of the design data.

FIG. 2 shows a cross-section of a Scanning Electron Microscope (SEM) image of a lithography mask 200. The mask 200 includes a substrate 210 and a pattern element 220. The mask 200 may be any transmissive or reflective lithography mask. The pattern elements 220 may include an absorbing material, such as chromium or aluminum. Alternatively, pattern element 220 may comprise a phase-shifting material, such as quartz. The pattern elements 220 further may include materials that not only shift the phase of the actinic light by a defined angle, but also absorb a defined percentage of the incident radiation. For example, a molybdenum-doped silicon oxynitride layer exhibits such properties.

The pattern elements 220 of the photomask 200 have a Critical Dimension (CD) 225. The cd describes the maximum lateral deviation from the specified set point values for which pattern element 220 must still meet the specified specifications.

The photolithographic mask 200 further has defect locations 230 having defects 240. In the example illustrated in fig. 2, defect location 230 is the same size as defect 240. This is due to the defect locations 230 being substantially rectangular. For analysis purposes, defect 240 has been intentionally deposited on photomask 200 and is significantly different from the outer profile of the defect location illustrated in FIG. 1.

The defects 240 of the defect sites 230 have the same material composition as the pattern elements 220 of the photolithographic mask 200. Furthermore, the height of defect locations 230 within the example of fig. 2 generally corresponds to the height of pattern elements 220. These conditions may not be satisfied in order to apply the methods defined herein to analyze the defined locations 230 of the photolithographic mask 200.

In the example illustrated in fig. 2, defect location 230 is adjacent to plurality of pattern elements 220. This constitutes an additional complexity in analyzing the defect location 240. However, the methods described in the present application may also be used to inspect isolated defect locations or defect locations adjacent to only one pattern element 220 (not illustrated in FIG. 2). The method discussed may be used to analyze the defect location 230 for any material composition and profile.

The illustrated example of fig. 2 shows a defect of excess absorbing material. This defect is referred to in the art as an opacity defect. The methods presented within this application can also be used to analyze defects that lack an absorbing material (known in the art as definite defects) (not illustrated in fig. 2). Furthermore, the method can also be used to analyze mask substrates (not illustrated in fig. 2) for defects.

In the example of FIG. 2, the electron beam scans an area 250 (i.e., the image field shown in FIG. 2, or the indicated image area) around the defect location 230 to obtain measurement data for the defect 240. Hereinafter, the region 250 is also referred to as a measurement data image 250. Similar to the situation of fig. 1, the electron beam generates a light boundary 260 along the boundary of the pattern element 220 and the defect location 230. The light boundaries are the result of image formation within a Scanning Electron Microscope (SEM). Due to the topographical contrast of the electron beam, the edges or corners of the imaged structure are reinforced by the optical boundaries. This also applies to the marks that may be located on the photomask 200 to find the defect structures 230 (not illustrated in fig. 2).

By scanning the defect location 230 with a laser beam (not illustrated in fig. 2), measurement data of the defect location can be obtained. In addition, the defect site 230 may be scanned with an Atomic Force Microscope (AFM) to obtain measurement data (not illustrated in fig. 2) of the defect site 230.

FIG. 3 presents CAD data 300 for a partial cross-sectional area of the photolithographic mask 200 shown in FIG. 2. Fig. 3 illustrates the substrate 210 of the photomask 200 by a black area 310. The pattern elements 220 of the photomask 200 are represented in fig. 3 by white structures 320. The pattern element 320 has a CD 325. CAD data 300 does not have defect location 230 or defect 240 of fig. 2.

Fig. 4 shows a reference image 400 synthesized or "rendered" from the CAD data 300 of fig. 3. The synthesis process converts the structural elements or pattern elements 320 of the CAD data 300 into structural elements or pattern elements 420 as fabricated on the substrate 210 of the lithography mask 200 using a mask fabrication process. Pattern element 420 has rounded corners 440, unlike pattern element 320. In the reference image 400 of fig. 4, the pattern element 420 has a CD 425. In addition, an optical boundary 460 is added to the pattern elements 420 of the reference image 400 to simulate the imaging process using an electron beam.

The graph 500 of fig. 5 schematically illustrates a mask production process. Using a mask manufacturing process, a defined Critical Dimension (CD) CD having CAD data 300 is generated on a substrate 210 of a photomask 200_D325 pattern element320. To this end, a photoresist layer is applied to the absorber layer of photomask 200. The photoresist layer is then exposed using an electron beam. The top half image of fig. 5 presents the dose distribution 510 applied to the photoresist layer by the electron beam. The threshold 520 of the dose profile 510 defines the dose at which the photoresist begins to produce pattern elements 220 on the substrate 210 of the photomask 200. In the example illustrated in FIG. 5, the CD on the photomask 200 is based on the design of the pattern element 320_M225 smaller than CD_D325. CD of pattern element 320 according to design_D325 with the actually realized CD of the pattern elements 220 of the photomask 200_MThe difference between 225 is referred to as bias 550 as described above.

Simulations of the mask production process are typically performed using a Point Spread Function (PSF). Often, the point spread function is described by adding two gaussian distributions (see EP 2983193 specified in the first section):

where r denotes the radial position of a point relative to the centre of the electron beam, α denotes the full width half maximum of the incident electron beam, β denotes the full width half maximum of the backscattered electrons, and η denotes the intensity ratio of the incident and reflected radiation distributions.

If the deviation 550 is not rotationally symmetric in the mask plane, a superposition of the two gaussian distributions may be chosen. A first equation (1) in this specification describes the deviation 550 in the x direction, and a second equation (1) may describe the deviation 550 in the y direction of the reference image 400. The two axes are preferably perpendicular to each other, although this is not absolutely necessary. The orientation of the x-axis and y-axis relative to the reference image may be selected in any manner. If desired, three or more functions of equation (1) may be combined to account for the asymmetry of the offset 550.

In order for the reference image 400 to be able to be superimposed with the measurement data image 250, the reference image 400 needs to be corrected with the offset 550. Exemplary embodiments will be described below to illustrate how this is achieved.

For the synthesis of the reference image 400 from the CAD data 300, the parameters α, β and η of the gaussian distribution of the point spread function of equation (1) must be determined with sufficient accuracy, otherwise the structuring element 420 of the synthesized reference image 400 will have a different deviation from the structuring element 220 of the photomask 200. The overlay of the reference image with the deviation not adapted to the deviation 550 of the photomask 200 may result in a wrong localization of the defect location 230 and a wrong localization of the defect 240 of the lithographic mask 200. Any repair of defect 240 based on the localization that has been determined in this manner produces poor results.

However, for the case of registering the reference image 400 with the measured data image 250, it is generally not sufficient to correct the reference image 400 with a single global bias (single global bias) 500. Fig. 6 shows a plan view of the entire lithography mask 200. Curve 610 of fig. 6 illustrates that photomask 200 has a systematic variation of CD 225 for structuring element 220 over the entire mask area. In the area of the line designated with "0", the photomask 200 has the size of the pattern element 320 according to the design, i.e., the CD_D＝CD_M. In the example illustrated in FIG. 6, the bias 550 in the lower half of the photomask 200 increases gradually with a negative sign in the lower left-hand corner direction. In the upper right hand corner direction of the mask 200, the bias 550 increases more strongly than in the opposite direction. The addition of the offset 550 in this direction additionally has a positive sign. The numerical values of FIG. 6 indicate CD in nanometers_MAnd (4) changing. The variation in CD across photomask 200 is typically caused by one or more imperfect processes during mask production.

Except for the CD in the pattern element 220 of the photomask 200 in FIG. 6_MIn addition to a comparable systematic variation of, the CD_MIt is also possible to have a relatively small area of random variation over the photomask 200 (not illustrated in figure 6).

To improve the correspondence of the reference image 400 to the measurement data image 250, local or position-dependent correction values are determined to take account of the CD during production of the corrected reference image for a specific defect position 230 on the photomask 200_MA change in (c).

The position dependent correction values can be determined in two ways. First, as shown in FIG. 6, the CD variation or more accurate CD across the mask 200 may be measured_MVariations in. This may be done, for example, using SEM. Then, from CD_MDetermines a distribution of one or more position-dependent correction values. The position dependent correction values may be stored in a table in a non-volatile memory of the computer system. At the beginning of the analysis process for a particular defect location 230, the corresponding correction value or values are retrieved from the table and a corrected reference image is generated from the reference image 400 based on the correction values. In the exemplary embodiment in question, the position-dependent parameters of the point spread function, i.e., α (x, y), β (x, y), and η (x, y), are position-dependent correction values of the reference image 400. The deviation values δ α (x, y), δ β (x, y) and δ η (x, y) of the parameters α (x, y), β (x, y) and η (x, y) are derived from the average of the generated reference image 400<α>、<β>And<η>for generating a corrected reference image. In an alternative embodiment, these deviations are used to determine one or more position-dependent correction values, thus correcting the profile of defect location 230 determined from reference image 400 and measurement data image 250.

As illustrated in fig. 6, the position-dependent correction value is generally changed gradually over the photomask 200. For this reason, in general, it is sufficient to divide the photomask 200 into regions in which the parameters of the point spread function can be assumed to be constant, i.e., α (x, y) - > α (i), β (x, y) - > β (i), and η (x, y) - > η (i). This simplification reduces the complexity of generating a corrected reference image or correcting the contour of the defect location 230.

In the second exemplary embodiment, one or more position-dependent correction values are determined by the correction value that has been determined from the most likely correspondence of the reference image 400 and the measurement data image 250. To this end, the difference between the CD 425 of the pattern elements 420 of the reference image 400 and the CD 225 of the measurement data image 250 is minimized. This optimization process provides local parameters α (i), β (i) and η (i) for the point spread function of equation (1).

FIG. 7 shows a SEM recorded profile 760 or image field 760 of lithographic mask 700. Photomask 700 includes substrate 710 and pattern element 720. The pattern element 720 has defect sites 730, the defect sites 730 including defects 740 that lack the absorbing material.

To be used for acupuncturePosition-dependent or local correction values are determined for the reference image and a region 750 is selected which completely surrounds the defect position 730 and is significantly larger than the defect position 730 within the image field 760. The region 750 of the image field 760 is also referred to as a measurement data image 750. By minimizing the CD of the pattern elements 720 of the photomask 700 in the local metrology data image 750_M725 and reference picture CD_D425, the local correction value of the reference image may be determined. In this optimization process, the defect location 730 is not inspected. Defect location 730 includes a deviation from the intended layout of pattern element 720 in photomask 700. Defect 740 may affect CD_M725 are the result of the minimization process during the CD comparison with the pattern elements in the reference image.

FIG. 8 shows a corrected reference image 800 including a substrate 810, corrected pattern elements 820, and a corrected CD 825. The corrected reference image 800 corresponds to the reference image 400 in fig. 4, which has been corrected with local correction values assigned to the defect locations 230. The corrected reference image 800 is used in place of the reference area given in FIG. 1 as a comparison criterion for analyzing the defect locations 230 of the photomask 200. For this purpose, a corrected reference image 800 is superimposed on the measurement data image 250.

Diagram 900 of fig. 9 shows an outline 930 of defect location 230 including defect 240 of lithography mask 200. The defect location 230 or the outline 930 of the defect 240 is the result of the overlay of the corrected reference image 800 with the measured data image 250. The outer boundary 940 of the outline 930 of the defect 240 is clearly defined within the image section 950. The determined contour 930 of the defect location 230 may be used directly to repair the defect 240 of the defect location 230. If desired, the height distribution of the profile 930 of the defect 240 may be determined. The height distribution of the profile 930 of the defect may be measured, for example, using an Atomic Force Microscope (AFM).

Diagram 1000 of fig. 10 presents an outline 1030 of defect location 230 containing defect 240 of lithographic mask 200. In a second exemplary embodiment of a method for analyzing defect 230 (illustrated in FIG. 10) discussed herein, defect location 230 or outline 1030 of defect 240 is the result of superimposing reference image 400 with measured data image 250. The position dependent correction values have been used in a second step to correct the outer boundary 1040 of the contour 1030 of the defect 240. After this correction step, the outline 1030 of the defect location 230 may be used to correct the defect 240 of the defect location 230. Similar to the statements above in the scope of fig. 9, the height distribution of the corrected profile 1030 of the defect 240 (not shown in fig. 10) may be determined, if desired.

By etching the excess material of the defect 240, the defect 240 may be eliminated. This may be performed, for example, with a Focused Electron Beam Induced Etching (FEBIE) process. The defect 740 lacking the absorber material at the defect location 730 may be repaired using a deposition process performed with an electron beam and one or more precursor gases.

Fig. 11 presents a cross section through an apparatus 1100 in which the method according to the invention can be performed. The apparatus 1100 of FIG. 11 shows a modified SEM 1100. The modified SEM 1100 includes a particle gun 1102 and column 1107 as the basic components, in which an electron-optical device or beam-optical device 1110 is deployed. The electron gun 1102 generates an electron beam 1105, and an electron or beam optical device 1110 focuses the electron beam 1105 and directs it output from the cylinder 1112 onto the mask 200 or 700, thus acting as a beam shaping device 1110. The column 1107 further comprises a scanning unit 1115 designed to scan the electron beam 1105 over the surface of the mask 200, 700. As such, the scanning unit 1115 satisfies the function of the beam guide device 1115. By scanning the electron beam 1105 over the defect locations 230, 730, measurement data for the defects 240, 740 at the defect locations 230, 730 may be obtained.

The masks 200, 700 are disposed on an object stage or sample stage 1103. As indicated by the arrows in fig. 11, the sample stage 1103 can move in three spatial directions relative to the electron beam 1105 of the SEM 1100.

The apparatus 1100 comprises a detector 1120 for detecting secondary and/or backscattered electrons generated by the incident electron beam 1105 at a measurement point 1117. The detector 1120 is controlled by a control device 1180. Furthermore, the control means 1180 of the device 1100 receives measurement data of the detector 1120. The control device 1180 may generate a measurement data image 250 from the measurement data, which is presented on the monitor 1190. Alternatively and/or additionally, the apparatus 1100 may have detectors arranged annularly around the electron beam 1105 inside the column 1107 to detect secondary electrons and/or electrons backscattered by the mask 200, 700 (not shown in fig. 11).

Furthermore, the apparatus 1100 may include an ion source that provides low energy ions in the region of the measurement point 1117 that prevent the mask 200, 700 or a surface thereof from accumulating to form negative surface charges (not shown in FIG. 11) during the measurement process, the etching process, or the deposition process of the defect location 230, 730. By means of the ion source, it is possible to reduce the negative charge of the mask 200, 700 in a local and controlled manner and thus counteract the lateral spatial resolution reduction of the electron beam 1005.

The e-beam 1105 of the apparatus 1100 may additionally be used to analyze the defects 230, 730 before, during, and after performing a partial etch process or a deposition process.

The control device 1180 includes a computer system 1185. Computer system 1185 includes an interface 1187. The computer system 1185 may be connected to an inspection tool (not shown in fig. 11) via this interface. The computer system 1185 may receive via the interface 1187 the location or coordinates of the defects 240, 740 that have been measured with the inspection tool. The computer system 1185 or the control device 1180 may control the electron beam 1105 according to the received coordinates of the defect locations 230, 730. The computer system 1185 may further receive, via the interface 1187, CAD data 300 of the photolithographic mask 200, 700 that may be stored within a non-volatile memory (not shown in fig. 11) of the computer system 1185.

The computer system 1185 or the control device 1180 is designed to scan the electron beam 1105 over the mask 200, 700 using the scanning unit 1115. The scanning unit 1015 controls deflection elements within the cylinder 1107 of the modified SEM 1100, which is not shown in fig. 11. The computer system 1185 or the control device 1180 further includes a setting unit to set and control a number of parameters of the modified SEM 1000. The parameters that the setting unit can set may be, for example: magnification, focus of the electron beam 1105, one or more set values of an astigmatism correction device (Stigmator), beam displacement, position of the electron source, and/or one or more diaphragms (not shown in fig. 11).

The computer system 1185 may determine the reference data or reference image 400 from the CAD data 300 by synthesizing or rendering (rendering) the CAD data 300. The computer system 1185 may further determine position-related corrections to the reference data or reference image 400. The computer system 1185 may generate a corrected reference data or corrected reference image 800 from the position-related correction value and the reference data or reference image 400. Furthermore, the computer system 1185 is designed to confirm the profile 930 of the defect 240 by superimposing the corrected reference image 800 and the measured data image 250. The defects 240, 740 may be repaired based on the identified contours 930. The computer system 1185 may also determine the defect location 230 or the outline 1030 of the defect 240 from the reference image 400 and the measured data image 250. In addition, the computer system 1185 of the device 1100 may use one or more position-related corrections to correct the determined contour 1030.

If desired, the repair shape of the defect 240, 740 may be determined. For either the contour 930 or the modified contour 1030, its height profile is determined for this purpose. The height distribution can be measured using AFM as described above. For this purpose, the modified SEM may include one or more AFMs (not shown in fig. 11). The repair shape indicates the length of time, number of times, and time interval that the electron beam 1105 is applied to the location of the defect 240,740. The repair shape additionally contains information about the gas flow rates of the one or more etching gases or the one or more deposition gases provided at the respective locations of the defects 240, 740.

The computer system 1185 may use the repair shape to control the gas flow rate of the activated particles or electron beam 1105, as well as many of the gas components used for defect repair. This means that the computer system 1185 may control the design of the repair shape. In an alternative embodiment, the design of the repair shape is implemented outside of device 1100 (not shown in FIG. 11).

The apparatus 1100 for analyzing the defect locations 230, 730 may also be used to correct or repair the defects 240 and 740. To this end, the apparatus 1100 preferably includes a plurality of different storage vessels for storing different gases or precursor gases. Three containers 1140, 1150, and 1160 are shown in the exemplary device 1100 of fig. 11. However, the apparatus 1100 may have only two or more storage containers 1140, 1150, 1160 for processing the masks 200 and 700.

A first storage containerThe tool 1140 stores a precursor (precursor) gas or a deposition gas that may be used to deposit an absorbing material on the clear defects 740 of the mask 700 in conjunction with the electron beam 1105 of the modified SEM 1100. For example, first storage vessel 1140 may have a metal carbonyl, such as molybdenum hexacarbonyl (Mo (CO)₆) Or chromium hexacarbonyl (Cr (CO)₆) Precursor gases in the form of.

The second storage container 1150 contains a first etching gas. For example, the second storage vessel 1050 may comprise xenon difluoride (XeF)₂) Or chlorine-containing etching gases, e.g. nitrosyl chloride (NOCl), nitroxyl chloride (NO)₂Cl) or chlorine nitrate (ClNO)₃)。

The third storage container 1160 in the example shown in fig. 11 stores a gas that can be added to the etching gas in the second storage container 1150. The gas in third storage container 1160 may comprise, for example, water vapor (H)₂O) or hydrogen peroxide (H)₂O₂)。

Each storage vessel 1140, 1150, 1160 is provided with its own valve 1142, 1152, 1162 to control the amount of gas particles provided per time unit or the flow rate of gas at the location 1117 where the electron beam 1105 is incident on the surface of the mask 200, 700 to be repaired. The valves 1142, 1152, 1162 may be designed in the form of gas flow or mass flow controllers.

In addition, all three storage containers 1140, 1150, 1160 have dedicated gas delivery devices 1144, 1154, and 1164 terminating in nozzles 1146, 1156, and 1166 near the point 1117 of incidence of the electron beam 1105 on the mask 200, 700. In the apparatus 1100 shown in the example of fig. 11, the valves 1142, 1152, 1162 are mounted adjacent to the storage containers 1140, 1150, 1160. In an alternative embodiment, the valves 1142, 1152, 1162 may be disposed adjacent to the corresponding nozzles 1146, 1156, 1166 (not shown in fig. 11). Each storage container 1140, 1150, 1160 may have dedicated elements for individual temperature setting and control. The temperature setting helps cool and heat each stored precursor material. In addition, each of the plenums 1144, 1154, 1164 may also have dedicated components (also not shown in FIG. 11) for setting and monitoring the temperature of the gases provided at the reaction site 1117.

FIG. 11 is a schematic view showingThe apparatus 1100 has a pumping system 1170 to create and maintain the desired vacuum within the reaction chamber 1175. The reaction chamber 1175 has a vacuum range (< 10) prior to performing the localized E-beam induced etching process^-6mbar).

In addition, the apparatus 1100 may include a suction (suction) type extraction apparatus (not shown in fig. 11). The pump-down extraction apparatus is coupled with a pumping system 1170 such that fragments (fragments) or components generated during the decomposition of the etching or deposition gas and not required for the local chemical reaction can be substantially extracted from the reaction or vacuum chamber 1175 of the apparatus 1100 at the beginning. Because the unwanted gas components are locally pumped out of the vacuum chamber 1175 of the apparatus 1100 at the point of incidence 1117 of the electron beam 1105 on the mask 200, 700 before diffusing and entering, contamination of the vacuum chamber 1175 is largely prevented by initiating local chemical etching reactions.

FIG. 12 presents a flow chart 1200 for analyzing the defect locations 230, 730 of the lithography masks 200, 700 according to the first embodiment of the invention. The method begins at step 1210. At step 1220, measurement data for the defect locations 230, 730 of the lithography masks 200, 700 is obtained. The measurement data may be obtained by scanning the electron beam 1105 of the apparatus 1100 over the defect locations 230, 730. The apparatus 1100 or the computer system 1185 may receive the location or coordinates of the defect location 230, 730 from the inspection tool.

In a next step 1230, reference data or reference image 400 is determined from the Computer Aided Design (CAD) data 300 of the lithography mask 200, 700. This step may be performed by computer system 1185 of device 1100. CAD data 300 as required herein may be stored in non-volatile memory of computer system 1185. Additionally, the computer system 1185 may receive the CAD data 300 from an external source via the connection 1187.

At step 1240, the reference data or the reference image 400 is corrected using the at least one position-related correction value. This step may also be performed by computer system 1185 of device 1100. The electron beam 1105 passing through the apparatus 1100 may determine a position dependent correction, wherein the apparatus 1100 is controlled by the control device 1180 or the computer system.

Next, at step 1250, the defect locations 230, 730 are analyzed by comparing the measured data or measured data image 250 with the corrected reference data or corrected reference image 800. The analysis process provides an outline 930 of the defect 240. This step may also be performed by computer system 1185 of device 1100. The method terminates at step 1260.

Finally, fig. 13 shows a flowchart 1300 for analyzing defect locations 230, 730 of a lithography mask 200, 700 according to a second exemplary embodiment of the present invention. The method begins at step 1310. Both steps 1320 and 1330 are the same as both steps 1220 and 1230 in fig. 12.

At step 1340, an outline 1030 of the defect location 240 is generated from the reference data or reference image 400 and the measured data or measured data image 250. This step may also be performed by computer system 1185 of device 1100.

Next, at step 1350, an analysis is performed by correcting the profile 1030 of the at least one defect location 240 using the one or more correction values. The position dependent correction values may be determined as explained in the context of fig. 12. The method then terminates at step 1360.