JP2006135104A - Alignment mark and its detection method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an alignment mark which can be detected in a relatively short time even under such a situation as the alignment mark is deviated from a recognition region, and to provide its detection method. <P>SOLUTION: The alignment mark 1 has a main pattern 1m being used for detecting a position, and a plurality of subpatterns 1s provided around the main pattern 1m in proximity thereto. The subpatterns 1s has a profile different from that of the main pattern 1m and the interval between the main pattern 1m and the subpattern 1s is set such that at least one subpattern 1s is located in a recognition region when the main pattern 1m is located on the outside of the recognition region due to positional shift so that a charged beam can be focused by the subpattern 1s. Consequently, shutdown of a measuring SEM due to error can be avoided and the position can be detected in a relatively short time. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an alignment mark and an alignment mark detection method, and more particularly to an alignment mark and an alignment mark detection method suitable for measuring a dimension of a fine pattern of a semiconductor device or the like.

  Along with miniaturization of semiconductor devices, a charged beam such as an electron beam is used for measuring a dimension of a resist pattern formed by a photolithography method or a metal pattern formed using the resist pattern.

  For example, a length measuring SEM (Scanning Electron Microscope), which is a dimension measuring apparatus that uses an electron beam as a charged beam, irradiates an electron beam onto a dimension measuring pattern formed on a semiconductor substrate. Then, secondary electrons generated on the surface of the dimension measurement pattern by the irradiation of the electron beam are detected, the edge of the dimension measurement pattern is extracted from the detected secondary electron signal waveform, and the dimension is measured from this edge interval. There is (for example, Patent Document 1).

  When such a dimension measurement pattern is measured, the dimension measurement pattern needs to be accurately aligned immediately below the charged beam.

  Hereinafter, alignment in a conventional length measurement SEM will be briefly described with reference to the drawings.

  FIG. 5 is a plan view schematically showing a semiconductor substrate S that is a target for dimension measurement. As shown in FIG. 5, a plurality of regions 2 including dimension measurement patterns are provided on the semiconductor substrate S so that the dimensions can be confirmed over the entire semiconductor substrate S. In addition, a plurality of alignment marks 11 for specifying the reference position are formed on the semiconductor substrate S so that the position of the region 2 including the dimension measurement pattern can be designated by a relative positional relationship from the reference position. .

  When measuring such a dimension measurement pattern of the semiconductor substrate S, the length measurement SEM transports the semiconductor substrate S to a measurement chamber having a predetermined degree of vacuum and can move in a plane direction in the measurement chamber. Install on top and align.

  In this alignment, first, the length measurement SEM moves the stage, and when the semiconductor substrate S is arranged on the stage without error, the position where the alignment mark 11 exists is positioned at the measurement position (directly below the electron beam). . Then, the planar shape on the semiconductor substrate S is acquired as image data through an optical microscope with a low magnification (for example, 100 times), and the shape of the alignment mark registered in advance in the length measurement SEM for the image data. The shape that matches is searched by pattern matching or the like.

  At this time, the detected position of the alignment mark 11 is stored as absolute coordinates of the stage used when controlling the movement of the stage. Such a search is executed for a plurality of designated alignment marks 11. Then, based on the acquired absolute coordinates of the alignment mark 11, a coordinate axis for designating a pattern on the semiconductor substrate S and a θ alignment for making the direction of the absolute coordinate axis of the stage coincide, for example, an axis orthogonal to the stage surface Is performed by rotating the stage with respect to.

  Next, the length measurement SEM switches to a secondary electron image (hereinafter referred to as SEM image) having a higher magnification (for example, 5,000 to 10,000 times) than the optical microscope, and changes the planar shape on the semiconductor substrate S to image data. (Hereinafter referred to as SEM image data), and the alignment marks 11 are detected and theta alignment is performed in the same manner as the alignment with the above-described optical microscope, thereby performing more exact alignment.

  When the alignment is completed as described above, the length measurement SEM measures the length of the dimension measurement pattern according to the following procedure. FIG. 6 is an enlarged plan view showing the region 2 including the dimension measurement pattern shown in FIG.

  As shown in FIG. 6, in the region 2 including the dimension measurement pattern, a dimension measurement pattern 4 having a rectangular planar shape having a predetermined design width is formed, and the dimension measurement pattern 4 is present in the periphery. A distinctive pattern 3 (here, a character pattern) having no similar shape is formed so that it can be clearly distinguished.

  First, the length measurement SEM moves the stage so that the region 2 including one dimension measurement pattern specified by the relative positional relationship from the reference position specified by the alignment mark 11 is located immediately below the electron beam. The SEM image is acquired at the stage position.

  In the length measurement SEM, for example, the shape of the characteristic pattern region 5 constituted by a part of the characteristic pattern 3 and a part of the dimension measurement pattern 4 is registered in advance, and the registered shape And the acquired SEM image are compared to identify the position of the characteristic pattern region 5.

  Further, the length measurement SEM includes from a specific position of the characteristic pattern region 5 (the upper end of the characteristic pattern region 5 in FIG. 6) to the length measurement position 6 where the dimension measurement on the dimension measurement pattern 4 is performed. The distance L is also registered. After focusing on the dimension measurement pattern 4 at the measurement position 6, the electron beam is scanned in the width direction of the dimension measurement pattern 4.

The secondary electron signal waveform obtained at this time has peaks P1 and P2 corresponding to the two edges of the dimension measurement pattern 4 traversed by the electron beam, as shown in FIG. Therefore, the length measurement of the dimension measurement pattern 4 can be performed by measuring the interval between these peaks P1 and P2.
Japanese Patent Laid-Open No. 5-94942

  However, in the above-described alignment in the conventional dimension measurement, the alignment mark 11 is detected in the image data acquired through the optical microscope, and then the alignment mark 11 is detected in the SEM image through the optical microscope. In some cases, for example, a positional shift may occur between the image data acquired in this way and the SEM image due to the fact that the optical axis of the optical microscope and the emission axis of the electron beam do not completely match. Such a positional shift is unavoidable in terms of machine accuracy. Specifically, a positional shift of about several μm to 10 μm occurs.

  For this reason, even when the alignment mark 11 is positioned at the center based on the absolute coordinates acquired by the alignment by the optical microscope, the alignment mark 11 may not be displayed at the center of the SEM image when switching to the SEM image. . That is, as shown in FIG. 8, the alignment mark 11 is deviated in the SEM image 20, or it protrudes from the SEM image 20 as shown in FIG. 9, and further, as shown in FIG. A situation exists that exists outside of.

  In the conventional length measurement SEM, when the shape to be detected does not exist on the SEM image 20, the shape most similar to the shape registered in advance is searched. When there is no similar shape, the search range, that is, the acquisition area of the SEM image 20 is enlarged, and the alignment mark 11 is searched.

  However, since the length measurement SEM is configured to automatically focus on the assumption that the alignment mark 11 is present in the SEM image 20, the alignment mark 11 is not present in the SEM image 20 (FIG. In 10), since the pattern to be focused does not exist in the SEM image, it cannot be focused. Therefore, since an attempt is made to search for a registered shape in an out-of-focus state, an alignment mark recognition error occurs, the measurement operation is stopped, and work efficiency is reduced.

  As a countermeasure, it is possible to perform alignment at a low magnification using an SEM image. However, in order to acquire an SEM image, an electron beam is applied to the entire image acquisition region with a beam corresponding to the resolution. Since it is necessary to scan with a diameter, the time required to acquire the SEM image increases remarkably when the image acquisition region becomes large.

  The length measurement of the dimension measurement pattern is performed in order to confirm the processing accuracy in the process of forming the semiconductor device, and it takes time for the confirmation to reduce the throughput of the process, which is not preferable.

  The present invention has been proposed on the basis of the above-described conventional circumstances, and enables an alignment mark to be detected in a relatively short time when the alignment mark is out of the recognition area. And an alignment mark detection method.

  In order to achieve the above object, the present invention employs the following means. That is, the present invention presupposes an alignment mark formed on the substrate in order to specify a reference position for designating the position of each pattern on the substrate. The alignment mark according to the present invention is arranged so that the main pattern, which is a pattern specified by the reference position, is arranged around the main pattern without overlapping the main pattern, and has a plane different from the planar shape of the main pattern. A configuration including a plurality of sub-patterns having a shape is employed.

  According to this configuration, when the alignment mark is detected based on the image data acquired by irradiating the predetermined region on the substrate with the charged beam, the main pattern is included in the image data due to the positional deviation caused by the detection device. Even if it does not exist, a sub-pattern can be present in the image data. For this reason, the charged beam can be focused using the sub-pattern. The sub-patterns are preferably arranged at a distance from the main pattern based on a positional deviation amount caused by a detection device used for detecting the main pattern.

  On the other hand, from another viewpoint, the present invention can provide an alignment mark detection method suitable for the detection of the alignment mark. First, the present invention is based on image data acquired by irradiating a predetermined area on a substrate with a charged beam, in order to specify a reference position of a relative positional relationship with each pattern on the substrate. It is premised on an alignment mark detection method for detecting the formed pattern. The alignment mark detection method according to the present invention includes a step of searching for a main pattern that is a pattern for specifying the reference position from image data, and when the main pattern is not detected in the search, Of the plurality of sub-patterns arranged around, the step of focusing the charged beam on any one of the sub-patterns included in the image data, and obtaining the original image data in the state of the focus A step of searching for the main pattern from an area larger than the area obtained.

  In this way, even when the main pattern exists outside the image data acquired by irradiation of the charged beam due to the positional deviation caused by the apparatus, the alignment mark is searched again in the focused state. As a result, the alignment mark can be detected in a relatively short time compared to the prior art. When focusing on the sub-pattern, the focus may be adjusted after detecting the sub-pattern from the image data.

  According to the present invention, it is possible to improve the recognition rate of the alignment mark when a positional deviation caused by the apparatus occurs, and it is possible to detect the alignment mark in a relatively short time under this situation. Become.

  Hereinafter, an alignment pattern and an alignment pattern detection method according to the present invention will be described with reference to the drawings.

  FIG. 1 is a plan view showing an example of an alignment mark 1 according to the present invention. FIG. 2 is a plan view showing a state in which the alignment mark 1 is arranged on the semiconductor substrate S.

  As shown in FIG. 1, an alignment mark 1 according to the present invention is used to specify a reference position on a substrate that serves as a reference when relatively specifying the position of each pattern formed on the substrate. A main pattern 1m and a plurality of sub-patterns 1s arranged around the main pattern 1m are provided. The sub-pattern 1s only needs to have a difference between a different planar shape and a recognizable planar shape when performing pattern matching with the main pattern 1m, and the shape is not particularly limited. In the example of FIG. 1, the main pattern 1m of the alignment mark 1 is formed in the same cross pattern as the conventional alignment mark, and each sub-pattern 1s is formed in a rectangular pattern.

  The main pattern 1m and the sub-pattern 1s are formed as patterns (same layer) processed in the same processing step on the semiconductor substrate S, for example, so that the focal positions of the charged beams are equivalent to each other. It is configured as follows. Further, it is preferable that no other pattern is formed at the position where the alignment mark 1 is formed and in the vicinity thereof, including the layer existing below the alignment mark 1. By doing so, the alignment mark 1 is formed on a flat surface in a state where there is no pattern that interferes with detection around the flat surface, so that the alignment mark 1 can be detected more reliably.

  The size of the main pattern 1m is based on the size of a region irradiated with a charged beam when irradiating a charged beam (here, an electron beam) and acquiring image data (the size of a region acquired as an SEM image). It can be designed as appropriate. That is, any size may be used as long as the main pattern 1m can be reliably recognized when searching for the main pattern 1m by pattern matching or the like based on the image data. For example, when the size of the charged beam irradiation region is 20 μm × 20 μm, the outer shape of the main pattern 1 m may be about 10 μm × 10 μm.

  Further, the size and the number of arrangement of each sub-pattern 1s are not particularly limited, but here, for the reasons described later, a rectangular pattern of 2 μm × 2 μm is changed to each center of the pattern constituting the cross of the main pattern 1 m. Three sub-patterns 1s are arranged on each side of the quadrilateral formed by straight lines separated from the line by a predetermined distance d.

  Note that the distance d between the central portion of the main pattern 1m and the central portion of the sub-pattern 1s is designed based on the value of the amount of misalignment caused by a device that detects the main pattern 1m (here, the length measurement SEM). preferable. For example, when the positional deviation due to the length measurement SEM is expected to be about 10 μm at the maximum, the distance d between the central portion of the main pattern 1 m and the central portion of the sub pattern 1 s may be set to 10 μm.

  Now, as shown in FIG. 2, a plurality of the alignment marks 1 described above are formed at predetermined positions on the semiconductor substrate S in the same manner as the conventional alignment mark 11 described above. The example of FIG. 2 shows a state in which alignment marks 1 are formed at four locations on the circumference concentric with the center on the semiconductor substrate S. The formation position and the number of alignment marks 1 are not particularly limited, and a desired number may be formed at a desired position according to the diameter of the semiconductor substrate S or the like.

  On the semiconductor substrate S, a region 2 including a dimension measurement pattern is formed at a position specified by a relative positional relationship from a reference position specified by the alignment mark 1. Since the region 2 including the dimension measurement pattern is the same as the conventional one, description thereof is omitted here.

  The reference position refers to a position specified by the alignment mark 1, and it does not matter whether it is a specific point (origin) or a plurality of points on the semiconductor substrate S. For example, one point such as the center point of the semiconductor substrate S specified by the plurality of alignment marks 1 may be set as the reference position, and one point on each alignment mark 1 (for example, the center of the alignment mark) is set as the reference position. Also good.

  Next, a method for detecting an alignment mark in the semiconductor substrate S on which the alignment mark 1 is formed will be described.

  Similar to the description in the above-described prior art, in the alignment mark 1 according to the present invention, after alignment with a low-magnification optical microscope, the alignment is performed by switching to an image with a charged beam, that is, an SEM image. Is called. At this time, when the above-described misalignment due to the length measurement SEM occurs, the alignment mark 1 is not displayed at the center of the SEM image 20 as shown in FIGS.

  For example, as shown in FIG. 3B, in a situation where the main pattern 1m is completely outside the SEM image 20, the length measurement SEM is a planar shape of the main pattern 1m registered in the length measurement SEM in advance. Even if the SEM image 20 is searched based on the above, the main pattern 1m cannot be detected.

  However, since the alignment mark 1 of the present invention has the sub-patterns 1s arranged around the main pattern 1m as described above, even when a positional deviation of 10 μm due to the length measurement SEM occurs, for example, at least Any one of the sub-patterns 1s is displayed in the SEM image 20. Therefore, for example, by adjusting the focus position of the charged beam so that the contrast (strength) of the secondary electron signal corresponding to the pattern edge of the sub pattern 1s is maximized, the charged beam is applied to the sub pattern 1s. Can be focused.

  In this manner, after focusing on the sub-pattern 1s, an SEM image is obtained by irradiating a charged beam to a wider area than the area from which the original image data was obtained in this focused state. This makes it possible to acquire an SEM image including the main pattern 1m in a focused state.

  The main pattern 1m that has not been displayed in the initially acquired SEM image is obtained by searching the planar shape of the main pattern 1m previously registered in the length measurement SEM with respect to the newly acquired SEM image. Can be reliably detected. That is, the recognition rate of the alignment mark can be improved as compared with the conventional alignment mark detection method. Further, since the alignment mark 1 is searched again in the focused state, the time required for detecting the alignment mark is shortened compared to the conventional case.

  Furthermore, in the present embodiment, as shown in FIG. 1, the sub-pattern 1s having a size smaller than that of the main pattern 1m is arranged in all the vertical, horizontal, and diagonal methods in plan view. Even if the positional deviation to occur occurs in any direction, the above-described actions and effects can be achieved.

  When focusing the charged beam on the sub-pattern 1s, the planar shape of the sub-pattern 1s may be registered in advance in the length measurement SEM, and the sub-pattern 1s may be detected from the image data 20. . In this way, since a pattern different from the sub-pattern 1s is not focused, the charged beam can be focused more reliably on the sub-pattern 1s.

  As described above, according to the present invention, when the alignment mark does not exist in the SEM image due to the positional deviation caused by the apparatus, the measurement operation of the length measuring SEM is stopped as an alignment mark recognition error. It is possible to detect the alignment mark even in a situation where the occurrence of the above has occurred.

  Therefore, by adopting the present invention, the length measurement SEM can detect any of the alignment marks 1 and perform alignment, and then can perform any dimension measurement pattern formed on the semiconductor substrate S as in the conventional case. The dimension measurement pattern 4 can be measured by moving the stage so that the region 2 including the region 2 is located immediately below the charged beam. For this reason, in the length measurement SEM, it is possible to remarkably improve the throughput when all the operations from wafer alignment to dimension measurement pattern detection / length measurement are automatically performed.

  By the way, in the above example, in order to provide a margin for the positional deviation caused by the length measurement SEM, the main pattern 1m is set to a size that is considerably smaller than the size of the SEM image 20, similarly to the conventional alignment mark 11. explained. However, according to the present invention, even when the main pattern 1m exists outside the SEM image 20, it can be detected with certainty, so there is no need to provide such a margin. That is, the main pattern 1m of the alignment mark 1 according to the present invention can adopt a larger dimension with the size of the SEM image 20 as an upper limit.

  As described above, according to the alignment mark 1 according to the present invention, since the main pattern 1m having a larger size than that of the conventional alignment mark 11 can be adopted, in particular, higher magnification (the size of the SEM image 20 is smaller). In the alignment mark used for the alignment in (), the effect that the pattern recognition can be improved can be obtained.

  In the above description, the apparatus for detecting the alignment mark 1 has been described as the length measurement SEM. However, the present invention is applicable to all apparatuses that perform alignment by placing the semiconductor substrate S on a stage that is movably provided. Is possible.

  Moreover, the shape of the alignment mark 1 demonstrated above shows the specific example, and does not limit the technical scope of this invention. That is, the alignment mark 1 of the present invention only needs to include a plurality of sub-patterns 1s around the main pattern 1m, and can be designed arbitrarily. For example, as shown in FIG. 4A, the sub-pattern 1s is circular, and as shown in FIG. 4B, the sub-pattern 1s is on the circumference centered on the central portion of the main pattern 1m. Of course, the same effect can be obtained even if they are arranged.

  The present invention detects an alignment mark formed on a test object in a relatively short time when a plane shape of the test object is acquired as image data by irradiating the test object with a charged beam. For example, it is useful for automatic alignment in dimension measurement.

The top view which shows an example of the alignment mark of this invention. The top view which shows the alignment mark of this invention arrange | positioned on the semiconductor substrate. Explanatory drawing which shows the alignment mark detection method of this invention. The top view which shows the other example of the alignment mark of this invention. The top view which shows the conventional alignment mark arrange | positioned on the semiconductor substrate. Explanatory drawing which shows the conventional automatic dimension measuring method. Explanatory drawing which shows the conventional automatic dimension measuring method. Explanatory drawing which shows the conventional alignment mark detection method. Explanatory drawing which shows the conventional alignment mark detection method. Explanatory drawing which shows the conventional alignment mark detection method.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Alignment mark 1m Main pattern 1s Sub pattern 2 Area including dimension measurement pattern 3 Characteristic pattern 4 Dimension measurement pattern 5 Characteristic pattern area 6 Measurement position 7 Secondary electron signal waveform 11 Alignment mark 20 SEM image S Semiconductor substrate

Claims (4)

In the alignment mark that is formed on the substrate and specifies the reference position for designating the position of each pattern on the substrate,
A main pattern which is a pattern for specifying the reference position;
A plurality of sub-patterns arranged around the main pattern without overlapping the main pattern and having a planar shape different from the planar shape of the main pattern;
An alignment mark characterized by comprising:   The alignment mark according to claim 1, wherein each of the sub-patterns is arranged at a distance based on a positional deviation amount caused by a detection device used for detecting the main pattern with respect to the main pattern. Detecting an alignment mark that detects a pattern that specifies a reference position for designating the position of each pattern formed on the substrate based on image data acquired by irradiating a predetermined region on the substrate with a charged beam In the method
Searching for a main pattern which is a pattern for identifying the reference position from the image data;
When a main pattern is not detected in the search, focus of the charged beam is applied to any sub-pattern included in the image data among a plurality of sub-patterns arranged around the main pattern. The step of matching
In the state of the focus, searching the main pattern from a wide area compared to the area from which the original image data was acquired;
A method for detecting an alignment mark, comprising: In the step of focusing the charged beam,
The alignment mark detection method according to claim 3, wherein the sub-pattern is detected from the image data, and the charged beam is focused on the detected sub-pattern.

Images