JPS61236117A - Pattern position detecting method - Google Patents

Abstract

PURPOSE:To enable to easily recognize the position of an alignment mark in a highly precise manner from the detected waveform of the alignment mark with a simple process by a method wherein a symmetry pattern matching is performed on the differential waveform of the detected waveform. CONSTITUTION:The alignment mark 2 inscribed on a wafer 1 and the alignment mark 4 on a mask 3, which is opposing to the wafer 1 across a gap (g) of several tens mum, are detected by a detecting optical system 5. The light reflected from the surfaces of the mask 3 and the wafer 1 is divided into two parts by a beam splitter 9, one beam is magnified by a relay lens 11, condensed by a cylinder lens 12 and forms an image on a linear sensor 13. When the detected waveform is differentiated, the value of the flat part of the waveform is reduced to almost zero, the waveform is greatly changed at the position of the alignment 2, and the result of differentiation is substantially changed. When a symmetry pattern matching treatment is performed on the differentiated waveform, the position (nc) of the alighment mark shows the maximum value, thereby enabling to easily recognize the position of the alignment mark.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a method for detecting alignment marks on a 6-mask or a wafer and determining their positions in an exposure device that prints a circuit pattern on a sea urchin. be.

[Background of the invention]

As described in Japanese Patent Publication No. 56-2284.

FIG. 3 shows an example of a calculation result of a detected waveform by the conventional symmetric pattern matching method.

FIG. 3 shows the detected waveform detected by imaging with the imaging device as it is, and the alignment mark appears at the position shown in ■.

Figure 3 (B) shows the result of symmetric pattern matching calculation processing of the detected waveform of 311iSI (Al). The position where the value is minimum is determined as the position of the alignment mark. However, in symmetric pattern matching, the place where this value becomes small is the flat part of the waveform, as shown in Figure 3 (B). Both end parts OO and center part O of the alignment mark
It appears in 03 places. Unable to recognize alignment mark position.

For this reason, in symmetric pattern matching, the alignment mark position has been recognized by setting it near the alignment mark within the calculation range. Therefore, preprocessing is required to limit the calculation range, making the processing complex and time consuming.

[Purpose of the invention]

The purpose of the present invention is to
To provide a pattern position detection method that allows the position of an alignment mark to be easily recognized with high precision and simple processing.

[Summary of the invention]

The gist of the present invention is that the detected waveform is differentiated.゛When the detected waveform is differentiated, the processing result becomes 0 in the portion where the detected waveform is almost flat. In areas where the waveform that detects the alignment mark changes significantly, the differential results also show significant values.

Moreover, the almost symmetrical detected waveform becomes an antisymmetrical waveform.

When this differential waveform is subjected to symmetric pattern matching processing, only the part of the waveform where the alignment mark is detected will show a large value, and the position showing the maximum value of the symmetric pattern matching result will be the position of the alignment mark. can be recognized correctly.

[Embodiments of the invention]

Hereinafter, one embodiment of the present invention will be specifically explained with reference to FIGS. 1 and 2.

Figure 1 (to) shows an example of an optical system for detecting alignment marks between a mask and a wafer of an X-ray exposure device, and shows alignment marks 2 and wafer 1 engraved on wafer 1, several tens of μm apart. The detection optical system 5 detects alignment marks 4 on the mask 3 facing each other with a gap (!1). In the configuration of the detection optical system 5, illumination light from an optical fiber 6 is collected by an optical system 7IIC, passes through a beam splitter 8, and enters an objective lens 9. Illumination light from the objective lens 9 is applied to three mask surfaces and one Ueno surface.

The light reflected from the 3rd mask surface and the 1st mask surface contains information on each alignment mark 2.3. This light passes through the objective lens 9 again, passes through the beam sub 11 and the beam 8, and is split into two parts by the beam splitter 9, one of which is imaged on a TV camera for visual viewing. The other light is magnified by a relay lens 11, and further by a cylindrical lens 12.
The light is focused and imaged onto the linear sensor.

FIG. 1(B) shows the arrangement of the alignment marks 2 and 4 on the mask l and the wafer 3, and FIG. 1(C) shows the signal intensity of the alignment marks 2 and 4 imaged on the linear sensor 13. This is what was shown.

The two convex parts on both ends are the mask 30 alignment mark 4
The concave waveform in the center is the detected waveform of the alignment mark 2 on the wafer 1.

As shown in the conventional example, FIG. 2 shows an example of the result of processing the detected waveform shown in FIG. 1 (0) using the conventional method.

FIG. 2 shows the results of processing using the method invented this time. Figure 2 (A> is a partial extraction of the alignment mark 2 on wafer 1 from the detected waveform in Figure 1 (Q). Figure 2 (13) is the detected waveform in Figure 2). (takes the difference between the signal waveforms). When differentiated, the value is almost O in the flat part of the waveform, and the waveform changes greatly at the flyment mark 2 position, so the differential result is Also, as mentioned in the overview, a substantially symmetrical detected waveform becomes a substantially antisymmetrical shape by being differentiated.

FIG. 2 (q shows the result of symmetric pattern matching processing of the differential waveform described above).

The symmetry pattern matching (Z(rL)) is first-order where Z(r): Symmetry pattern matching result D(rL):
Value P of differential waveform: Range for comparing symmetry As shown in Figure 2 (C), when the symmetry matching process is performed on the differentiated waveform, the maximum value is obtained at the position of the 2 alignment marks. The position of the alignment mark can be easily recognized.

Note that even if symmetrical pattern matching (Z(rL)) is used, the alignment mark position can be detected sufficiently, although the accuracy will drop to some extent.

〔Effect of the invention〕

As explained above, according to the present invention, alignment mark positions can be accurately determined with simple processing.

Moreover, it has the effect of being able to obtain the results at high speed.

[Brief explanation of drawings]

11% (A) is a perspective view showing the optical system for detecting alignment marks on the mask and wafer when used in the xijg optical system according to the present invention, and FIG. 1(B) is the alignment shown in FIG. Figure 1 shows a state in which marks are superimposed; A diagram showing an example, Figure 3, +13 is a diagram showing a conventional pattern matching processing method. 1... Wafer, 2... Mask. 9... Objective lens, 13... Linear sensor. .

Claims (1)

[Claims] 1. A video signal obtained by imaging a pattern with an imaging device is converted into a digital signal, the obtained digital signal is differentiated, and a symmetric pattern matching process is performed on the differentiated signal. , a pattern position detection method characterized in that the position where the processing result is maximum is recognized as the pattern position. 2. Pattern matching processing ▲ There are mathematical formulas, chemical formulas, tables, etc. ▼ or ▲ There are mathematical formulas, chemical formulas, tables, etc. ▼ Here [where, D(n) is the differential value, P is the comparison range, and n is the position . ] The pattern position detection method according to claim 1, characterized in that the pattern position detection method is carried out in the following steps.