JPS6338697B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for inspecting optical characteristics such as distortion and magnification error, or so-called distortion, of a projection optical system of an IC projection exposure apparatus, and particularly to a method for inspecting optical characteristics of a projection optical system of an exposure apparatus. , relates to a method of inspecting marks on a mask by printing them onto a photosensitive substrate.

A conventional inspection mask substrate of this type (hereinafter simply referred to as a reticle) will be explained. Figure 1a
1 is a diagram of a conventional reticle. Marks 2, 3, 4, and 5 are arranged on the top, bottom, and left and right edges of the reticle 1, respectively. Each mark 2~
5 has a checkered pattern as shown in FIG. 1b, and constitutes a vernier with marks 2 and 5 as the main scale and marks 3 and 4 as the vernier scale. A vernier is formed by overlapping and printing the main scale and vernier patterns on a wafer, for example, and the optical characteristics of the projection optical system are inspected by measuring the amount of deviation.

This pattern will be explained using FIG. 1c. In addition, in Fig. 1a, patterns 2 and 3 placed above and below the reticle 1 measure the amount of deviation in the y direction, and patterns 4 and 5 placed on the left and right side measure the amount of deviation in the x direction. , now these patterns 2 and 3,
For example, let the interval between 4 and 5 be 80 mm. It is also assumed that the patterns 2 to 5 on the reticle 1 are printed onto a wafer using a 1/10 reduction projection exposure apparatus, and the optical characteristics of the 1/10 reduction projection lens system are measured.

First, a pattern chip 6 is printed on the wafer on the wafer mounting table, and then the wafer is measured based on the measurement value of a laser interferometric measuring device equipped on the wafer mounting table that can measure the length with an accuracy of about 4/100 μm or less. Drive the mounting table (stage) and move the wafer in the x direction.
Move each part accurately by 8mm in the y direction. And again patterns 2, 3 on reticle 1,
4 and 5 are printed on the wafer, the main vernier of the main pattern 5 of the chip 6 and the sub vernier of the vernier pattern 4 of the chip 7 overlap to form a vernier 8, as shown in FIG. Vernier pattern 3
The secondary vernier and the main vernier of the main scale pattern 2 of the chip 9 overlap to form a vernier 10. At this time, if there is no distortion or magnification error in the projection lens system, the amount of deviation between the main vernier and the sub-vernier of each of the verniers 8 and 10 will both be zero. However, if the projection lens system has distortion, magnification error, etc., the lattice fringes (so-called scale) of the vernier 8 or 10 will shift. By measuring this amount of deviation, it is possible to determine the positions of the verniers of patterns 2 to 5 on the reticle of FIG. The amount of distortion and magnification error of the projection lens system at the position) is measured.

However, the conventional method described above has the disadvantage that only the relative amounts of distortion, magnification error, etc. of the projection lens system can be known. This will be explained using FIG. 2. Figure 2 shows the area 11 projected onto the wafer when distortion, magnification error, etc. of the projection lens system are ideally zero, and the area of the projected image on the wafer deformed due to distortion etc. due to the projection lens system. 12, and in this figure, the center 13 of the projected image coincides with the optical axis of the projection optical system. In this illustrated example, the right side of region 12 in the x direction is deformed more than the left side, but conventional methods cannot detect the difference in the amount of deformation on the left and right, and the length in the x direction is shorter than that of region 11. All you need to know is that it is long (or long). That is, in the above-mentioned conventional system, when both the left and right sides of the region 12 in the x direction are deformed inwardly, the distance l between the patterns in the projected image corresponding to the patterns 4 and 5 on the reticle 1 is
becomes shorter than the original distance. For this reason, the vernier 8 shown in FIG.
Even if the amount of directional deviation is measured, it only shows that the x direction of region 12 is shorter than that of region 11. Furthermore, if the left and right sides of the area 12 are deformed to the same extent, for example, the right side is outward and the left side is deformed to the same extent, the deviation of the lattice stripes of the vernier 8 is canceled out, and the measurement is performed as if there is no distortion or the like.
Therefore, the absolute amount of distortion from the center 13 cannot be determined.

The object of the present invention is to overcome this drawback and provide a method for accurately testing the optical characteristics (distortion) of a projection optical system.

Embodiments of the present invention will be described below using the drawings from FIG. 3 onwards.

In the first embodiment shown in FIG. 3a, a main scale pattern 15 is provided in the central part of the reticle 14, that is, in the part through which the optical axis of the projection lens system passes when the reticle 14 is set in the exposure device. Vernier patterns 16a, 16b are located at positions, for example, ±40 mm away from 15 in the y direction, and vernier patterns 16c, 16d are located at positions, for example, ±40 mm away from the x direction.
are arranged.

In order to inspect the optical characteristics of, for example, a 1/10 reduction projection lens system using this embodiment, first, as shown in FIG. First, set the reticle 14 on the exposure device, and first set the reticle 14 as shown in the dotted line.
The image is printed onto a transfer target (photosensitive substrate) such as a wafer coated with photoresist.

Next, the wafer stage is moved -4 mm in the y direction and the image of the reticle 14 is printed based on the measurement value of a laser interferometer, etc., and a printed image 21 is created.
A vernier A is formed in which the sub-vernier, which is the image of a, and the main vernier, which is the image of the main scale pattern 15 printed for the second time, overlap.

Next, the wafer stage is moved -4 mm in the x direction from the first printing position and the image of the reticle 14 is printed. This time, a printed image 22 is created, which combines the secondary vernier of the vernier pattern 16c from the first printing and the third printing. Main scale pattern 15 by
Vernier B is formed by overlapping the main verniers.

Thereafter, the printed images 23 and 24 are transferred in the same manner,
A vernier C is formed in which a main vernier according to the main scale pattern 15 and a sub-vernier according to the vernier pattern 16b overlap, and a vernier D is formed in which a main vernier according to the main scale pattern 15 and a sub-vernier according to the vernier pattern 16d overlap.

Verniers A to D thus formed are observed using a microscope or the like. At this time, the main vernier, which is the image of the main scale pattern 15 through which the optical axis of the projection lens system passes, is located at the center of the printed image, and therefore generally serves as a reference for measuring distortion, magnification error, etc. Therefore, absolute distortion in the x and y directions can be determined by simply reading the amount of deviation of the lattice fringes of each vernier A to D with the optical axis of the projection lens system as a reference. That is, in the conventional example, only the relative positions of both edges of the projected image could be measured (in other words, even if there was distortion, it could not be measured as an absolute amount), but in this example, the optical axis Since the main scale pattern 15 serving as a reference was provided so as to pass through the
The absolute amount of distortion, magnification error, etc. of the projection lens system relative to the optical axis can be measured simply by directly reading the scale of each vernier. Note that the checkered stripes as the scale of the vernier and how to read them will be described in detail later.

Next, FIG. 4 shows a second embodiment. In this embodiment, the vernier patterns 16a to 16d are the reticle 14.
are provided at the four corners of each vernier pattern 16a.
16d are, for example, equidistant from the main scale pattern 15. The first printing image is shown by a dotted line, and from the first printing position, move the wafer stage 4 mm in the x direction and -4 mm in the y direction to obtain a printed image 25, and move it -4 mm in the x direction and -4 mm in the y direction. Printed image 26, x
Printed image 27 by moving -4 mm in the direction and 4 mm in the y direction,
A printed image 28 is formed by moving 4 mm in the x direction and 4 mm in the y direction. A main vernier based on the main scale pattern 15, a sub-vernier based on the vernier pattern 16a, a sub-vernier based on the vernier pattern 16b, a sub-vernier based on the vernier pattern 16c, a sub-vernier based on the vernier pattern 16c, and a sub-vernier based on the vernier pattern 16d.
Vernier A, B, and secondary vernier by
C and D are formed.

In the first embodiment, distortions and the like were measured at positions 4 mm above, below, left and right from the center of the projected image, but in the second embodiment, distortions and the like at the four corners of the projected image can be measured. That is, according to the first and second embodiments, the vernier pattern can be placed at any position on the reticle as long as the placement with respect to the main pattern is known in advance, and thereby there is no absolute distortion at any position of the projected image. ,
Magnification errors can also be measured.

By the way, the main and vernier patterns in FIGS. 3 and 4 have an inverted T-shape, but this is because two plaid patterns extending in the x direction and the y direction are provided orthogonally to each other. FIG. 5 shows an enlarged example of this main scale pattern, and FIG. 6 shows an enlarged example of this vernier pattern.

In FIG. 5, a checkered pattern 120a is provided on the reticle 14 at a constant pitch in the x direction, and a checkered pattern 120b is provided in the pattern 120a.
Similarly, it is provided along the y direction. Each checkered pattern 120a, 120b is marked with numbers such as 2, 4, 6, and 8 in the positive and negative directions, with 0 at the center of the pattern. plaid pattern 50
a, 50b are plaid patterns 120a, 120b
It was installed as an auxiliary tool, and acts as a rough vernier.

In the vernier pattern of FIG. 6, this is a checkered pattern 130a, 130 provided at a slightly larger (or smaller) pitch than the pitch of the checkered pattern 120a, 120b of the main pattern.
b, and each lattice of the patterns 130a, 130b is shaped so as to be sandwiched between the lattices of the plaid patterns 120a, 120b, respectively. Checkered pattern 51a, 5 that acts as a rough vernier
1b is similarly provided with a pitch slightly larger (or smaller) than the pitch of the checkered patterns 50a, 50b of the main scale pattern, and each grid of the patterns 51a, 51b is provided with a pitch of the pattern 5, respectively.
It is shaped so as to be sandwiched between the grids 0a and 50b.

When the main vernier and the sub-vernier are printed one on top of the other, the lattice image of the main vernier corresponding to the lattice x ₀ in the center of the pattern 130a is sandwiched between the part of the main vernier corresponding to 0 of the checkered pattern 120a, and the checkered pattern When the lattice image of the sub-vernier corresponding to the lattice y ₀ at the center of the pattern 130b is sandwiched between the main vernier portion corresponding to 0 of the pattern 120b, the printed image will be x, Distortion and magnification error in the y direction become zero.

Figure 7 shows the plaid pattern 12 of Figures 5 and 6.
FIG. 7A is a more detailed enlarged view of an example of the gratings 0a and 130a, with FIG. 7a showing the grating of pattern 120a and FIG. 7b showing the grating of pattern 130a.

As shown in FIG. 7a, each lattice of the checkered pattern 120a on the main scale pattern side is provided with a stepped portion 100 and a tapered portion 101 symmetrically with respect to the central axis l, and the pitch P _M between the lattices is, for example, 10.0 μm. It is stipulated in Further, as shown in FIG. 7b, each grid in the vernier pattern has a width equal to the inter-grid spacing d of the main pattern, and a pitch _Ps of, for example, 10.1 μm. The stepped portions and tapered portions of these gratings in the main pattern are for making the vernier easier to read, and therefore may be provided in the grating on either side of the main pattern or the vernier pattern.

Figure 8 shows the checkered pattern of Figures 7a and b by 1/1.
This shows the state when the image is enlarged to 10 and overprinted, and a certain grating S ₁ of the sub-vernier is completely inserted between the gratings of the main vernier. This determination can be easily made by looking at the upper and lower staircase portions and tapered portions that are in contact with the grid _S1 . If we focus on this lattice S ₁ , the lattice S ₁ ,
Each position of the lattice S ₂ above, the lattice S ₃ above 2
For E ₁ , E ₂ , and E ₃ , the deviation is, for example, 0,
It is possible to determine +0.1 μm and +0.2 μm, and it is possible to measure overlapping polymerization (shift amount) with an accuracy of 0.1 μm or more.

Actually, when the main and sub verniers shown in FIGS. 5 and 6 are overprinted, checkered patterns 120a, 12
Find out numerically where position E ₁ is in the image of 0b. For example, checkered pattern 120
Plaid pattern 13 at image 6 on the positive side of a
Assuming that one lattice image of 0a is safely inserted, the amount of distortion in the x direction can be found to be +0.6 μm. At this time, if one lattice image of the lattice pattern 130b is completely sandwiched between the two images on the negative side of the lattice pattern 120b, then
The amount of distortion in the y direction is found to be -0.2 μm.

As described above, the amount of distortion of the projection lens at the position where the main and sub vernier overlap on the wafer is measured as a vector quantity having components of +0.6 μm in the x direction and −0.2 μm in the y direction.

By the way, in the above description, the usage of "main" and "sub" in terms of the main vernier, sub-vernier, etc. is completely relative, and either one may be named "main" or "sub".

In addition, although the amount of deviation was measured using a vernier in the above embodiment, instead of using a vernier, for example, two types of rectangular patterns with different lengths etc. are prepared, and one type of rectangular pattern is set at the center of the reticle. Alternatively, other types of rectangular patterns may be placed on the reticle at the desired measurement positions and then printed over the wafer to measure the amount of deviation.

As described above, according to the present invention, since one of the plurality of marks on the mask substrate is provided at a position through which the optical axis of the projection optical system passes, the absolute amount of deviation at any position of the printed image can be measured. The optical characteristics of the projection optical system can be accurately measured.

[Brief explanation of the drawing]

FIG. 1 is a diagram for explaining a conventional example, FIG. 2 is a diagram for explaining measuring the optical characteristics of a projection optical system from the deviation of a printed image, and FIG. 3 is a diagram for explaining a first embodiment of the present invention. 4 is a diagram for explaining the second embodiment, FIG. 5 is an enlarged view showing an example of the main scale pattern, and FIG. 6 is an enlarged view showing an example of the vernier pattern. FIG. 7 is an enlarged view showing an example of the shape of each grid of the main scale pattern and the vernier pattern, and FIG. 8 is a diagram showing the state when the grids shown in FIG. 7 are overprinted. [Explanation of symbols of main parts], Mask substrate...1
4. Mark...15, 16a to 16d, 50a,
50b, 51a, 51b, 120a, 120b,
130a, 130b, step part...100, taper part...101.

Claims (1)

[Scope of Claims] 1. A method for inspecting distortion of a projection optical system for projecting and exposing a pattern formed on a mask substrate onto a photosensitive substrate, comprising: a first mark approximately in the center; a reference mask having second marks at a plurality of positions spaced apart from each other by a predetermined distance from the reference mask such that the first mark substantially coincides with the optical axis of the projection optical system; and placing the photosensitive substrate at a predetermined first position. a first step of projecting and exposing a mark image of the reference mask onto the photosensitive substrate; 2
a second step of projecting and exposing the mark image of the reference mask onto the photosensitive substrate exposed in the first step; detecting the amount of positional deviation between the image of the first mark and the image of the second mark exposed on the photosensitive substrate;
A distortion inspection method comprising the step of determining an amount of distortion at the position of the second mark. 2. The method according to claim 1, wherein each of the first mark and the second mark is formed as two checkered patterns in which grids are regularly arranged in directions perpendicular to each other. 3. The method according to claim 2, wherein each lattice of the lattice pattern constituting either the first mark or the second mark is provided with a stepped portion. 4. Claim 2, characterized in that each lattice of the checkered pattern constituting either the first mark or the second mark is provided with a tapered portion.
The method described in section. 5. The method according to claim 2, wherein the pitch of the checkered pattern of the first mark and the pitch of the checkered pattern of the second mark are slightly different and set in a vernier relationship.