JPH07134965A - Scanning electron microscope - Google Patents

Abstract

PURPOSE:To extract a periodic structure without any error so as to precisely detect a target shape even if a semiconductor pattern has a complex periodic structure by using a sample space period previously given in detection of a periodic structure as a constraint. CONSTITUTION:In a detection process, a sample 7 to be measured or to be examined is placed on a stage 8, and thereto electron beams 2 are radiated, and then, an image is displayed on an image display monitor 19. When instructions for detection and shift processing are given to a computer 21 and a possibility restricting process is carried out, a two-dimensional grid image is generated on the basis of the previously given space period, and then, a two-dimensional image, in which possibility positions detected by the computer 21 serving as a detecting means are arranged, is generated. Subsequently, a relative value provides a peak, when an intersection point of the grid image generated on the basis of the space period is overlapped with the right possibility among the image possibilities generated by the extracted possibilities. In this way, wrong possibilities are eliminated, while undetected possibilities due to noise or the like are interpolated, that reliability of a periodic structure detection can be improved.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a scanning electron microscope which scans a sample image with an electron beam to image secondary electrons generated.

[0002]

2. Description of the Related Art A conventional apparatus, as described in Japanese Patent Application No. 5-19866, obtains a plurality of positions on a two-dimensional image having a shape similar to a predetermined reference image, and uses them as guides. The one having the shortest distance to the specific position on the two-dimensional image is automatically selected. This method works reliably for samples with a simple periodic structure.

[0003]

In recent years, the process technology for semiconductor wafers has become more sophisticated, and the process rules have become finer or more complicated. In the above-mentioned conventional technique, since the periodic structure is detected using only the reference image for one period of the sample having the periodic structure, the structure becomes complicated and the number of highly similar partial regions increases. However, similar positions are detected as candidates. If the position is at the shortest distance from the guide position, it is judged to be a necessary periodic structure and automatically selected. In order to automate the process quality control of semiconductor wafers, such misjudgments must be avoided.

An object of the present invention is to provide a function of accurately extracting a target shape by accurately extracting a periodic structure of a semiconductor wafer pattern having a complicated periodic structure.

[0005]

Generally, in the process of designing a semiconductor wafer, process rules (line pattern width,
Design values such as the diameter of contact holes) and the cell array period of memory cells are set in advance. The purpose is to use the design value such as the cell periodic structure as a constraint condition when detecting the periodic shape, select a combination of candidates that best fits the design period from the extracted candidates, and select other candidates. Is achieved by excluding it as an erroneously detected location that deviates from the original periodic structure and limiting the candidates to correct candidates.

[0006]

By using the absolute value of the design period, determination of periodicity becomes much easier. If the determination is easy, errors are reduced and reliability is improved. In addition, a statistical method such as applying a design period is resistant to noise, and even if a part of the periodic structure is buried in noise and cannot be detected, it can be interpolated.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIG. The electron beam 2 emitted from the electron gun 1 is narrowed down by the objective lens 6 and applied to the sample 7. The objective lens 6 is excited by the objective lens power supply 11. Further, the deflection signal generated by the deflection signal generator 14 can change the scanning range and scanning position on the sample by the computer 21, the deflection coil 10 is excited by the deflection amplifier 10, and the electron beam 2 on the sample 7. Two-dimensional scanning.

Signals generated by the electron beam 2 incident on the sample 7 (secondary electron signals, reflected electron signals, etc.)
Is converted into an electric signal by the detector 12, converted from an analog signal into a digital signal by the A / D converter 15, and stored in the image memory 16. This image memory 16
The content of is always converted from a digital signal to an analog signal by the D / A converter 17, and added to the grid as a luminance signal of the image display monitor (cathode ray tube) 19. At this time, the A / D converter 15, the image memory 16, and the D / A converter 17 generate timing signals for A / D conversion, storage, and D / A conversion for image display.
Receive from 4 Deflection coil 2 of image display monitor 19
0 is excited by the deflection amplifier 18 based on the deflection signal of the deflection signal generator 14.

On the other hand, the stage 8 on which the sample 7 is placed is
It is moved by the stage drive circuit 13, whereby the scanning position on the sample 7 is changed by the electron beam 2 and the visual field is moved. The same visual field movement can also be performed by exemplifying the image shift coil 5 by the DC amplifier 9 and offsetting the scanning position of the electron beam 2 on the sample. The movement amount of these visual field movements is controlled by the computer 21.

The signal from the cursor signal generator 22 changes according to the signal from the trackball 24 or the computer 21 to change the display position of the cursor displayed on the image display monitor. The computer 21 can acquire the position information on the image display monitor 19 from the state of the cursor signal generator 22. Further, the computer 21 can read all or a part of the image data in the image memory 16, and by combining it with the cursor position information in the previous term, a part of the image data centering on the cursor position corresponds to the image memory. You can read from the position.

The following processing is performed by using the apparatus having the above-mentioned configuration.

The procedure for registering the target image and the reference image will be described with reference to FIGS. 1 and 2.

First, a sample including a target partial image is placed on the stage 8 and the electron beam 2 is irradiated to display an image on the image display monitor 19. Next, the cross hair cursor 27 is manually aligned with the target partial image 25 in the displayed image using the trackball 24 (FIG. 2), and the computer 21 is instructed to register. The computer 21 acquires the cursor position information from the cursor signal generator 22 and saves it, and reads a part of the image data centered on the cursor position from the corresponding position on the image memory 16,
Save as a reference image.

The above series of procedures is referred to as a reference image registration procedure.

Next, the detection process will be described with reference to FIGS. 1 and 3.

First, a sample to be measured or inspected is placed on the stage 8 and the electron beam 2 is irradiated to display an image on the image display monitor 19. Next, the computer 21 is instructed to perform detection and movement processing. The computer 21 compares the reference image registered in advance with the image data in the image memory 16, obtains a plurality of coordinates X, Y on the image display monitor in which the reference image has a similar shape, and respectively Xi, Y.
i.

For a sample having a simple periodic structure, the desired periodic structure can be extracted by the above procedure. However, for example, for a sample having a complicated shape as shown in FIG. 4, an erroneous position similar to the original periodic structure may be detected (a in the figure).

In order to prevent this, the following candidate limitation processing is carried out.

First, a two-dimensional lattice-shaped image as shown in FIG. 5 is generated with a predetermined spatial period. Next, as shown in FIG. 6, a two-dimensional image in which the candidate positions detected by the detecting means are arranged is generated. Then, a two-dimensional cross-correlation calculation is performed between the generated two images. The result of the cross-correlation calculation is also a two-dimensional distribution, but when the two images best match, that is, the intersection of the grid of the grid-shaped image generated from the spatial period and the image generated by the extracted candidate When the correct candidate among the above candidates overlaps, the correlation value gives a peak. Even if the erroneously detected candidate overlaps the intersection point of the lattice, the other candidates do not overlap, so that the correlation value does not increase. The two-dimensional distribution of the correlation values thus obtained is a periodic distribution having peaks only at correct candidate positions (FIG. 7). By using this result, not only erroneous candidates can be excluded, but also candidates that have not been detected due to deterioration due to noise can be interpolated, and the reliability of periodic structure detection is extremely high.

On the other hand, if the design dimensions given in advance are incorrect, the intersections of the lattice and the candidates do not overlap well, and it becomes impossible to confirm a significant peak in the distribution of the correlation values. In this case, it is possible to prevent human error by making the detection failure.

In the above example, the periodic structure is extracted by the two-dimensional cross-correlation calculation. However, in addition to this, for example, a function with the given vertical and horizontal periods as variables is assumed, and this function is A method is conceivable in which the optimum period is obtained by function fitting the extracted coordinate values of the candidate position. The least squares method or the like may be used for fitting.

In any of the methods, a combination of candidates that best fits the design size is obtained, so that it is resistant to noise and highly reliable.

[0023]

According to the present invention, not only erroneously detected candidates can be excluded, but also the positions of candidates that cannot be detected due to deterioration due to noise or the like can be extracted, so that all candidates can be reliably detected. Further, even if a design dimension is erroneously given, it is possible to easily determine an error in detection failure, so that it is possible to prevent human error.

[Brief description of drawings]

FIG. 1 is a block diagram of a scanning electron microscope.

FIG. 2 is a diagram showing an example of designation of a reference image using a crosshair cursor.

FIG. 3 is a diagram showing an example of detection of a periodic structure.

FIG. 4 is a diagram showing an example of detecting a periodic structure.

FIG. 5 is a diagram showing a lattice image generated from design values.

FIG. 6 is a diagram showing an image generated by the extracted periodic structure candidates.

FIG. 7 is a diagram showing a two-dimensional cross-correlation calculation result.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Electron gun, 2 ... Electron beam, 3 ... Deflection coil, 5 ... Image shift coil, 7 ... Sample, 15 ... A / D converter,
19 ... Monitor for image display, 21 ... Computer.

Claims (1)

[Claims] 1. A scanning electron having means for scanning a sample having a periodic spatial structure with an electron beam to obtain a two-dimensional image of the sample surface, and means for detecting the periodic structure on the obtained two-dimensional image. A scanning electron microscope characterized by using a spatial period of a sample given in advance as a constraint condition when detecting a periodic structure in the microscope.