JPH0328809B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to improvement of a reduction projection exposure apparatus used in the manufacturing process of semiconductor devices, and in particular, in the production of high-precision semiconductor devices in which dimensional accuracy is strictly required in this type of device. The present invention relates to means for automatically correcting reduction magnification errors that occur.

A conventional reduction projection exposure apparatus is based on an optical system whose principle configuration is shown in FIG. 3, and projects the pattern of an original image 42 onto a substrate 44 using a reduction lens 43. If the distance between the reduction lens 43 and the substrate 44 is a, and the distance between the reduction lens 43 and the original image 42 is b, then as long as the refractive index distribution within the optical system is constant, the reduction magnification m is m=a/b, and the focal length f. is 1/
It is expressed as f=1/a+1/b. In a conventional reduction projection exposure apparatus, the distances a and b are fixed, so the refractive index may change due to changes in the conditions of the medium in the optical path of the optical system from the original image 42 to the substrate 44. When the distribution changes, the distances a and b on the optical system become a function of the refractive index, which changes the reduction magnification m, which has the drawback of deteriorating the dimensional accuracy of the pattern projected onto the substrate 44. .

For this reason, in the past, when adjusting the reduction magnification, as disclosed in the Proceedings of the 1981 Precision Machinery Society Autumn Conference Academic Lectures, pages 262 to 264, a standard pattern whose dimensions were known in advance was exposed on the wafer, developed, and the length was measured. I was adjusting the lens to find the error.

However, this does not take into consideration the point of automatically correcting the deterioration in the absolute dimensional accuracy of the pattern due to changes in the refractive index of the medium in the reduction projection optical system, resulting in the problem that it takes a lot of time to adjust the magnification of the exposure equipment. It was hot.

Therefore, an object of the present invention is to prevent a reduction magnification error caused by changes in the refractive index distribution of a medium in a reduction projection optical system, which deteriorates the absolute dimensional accuracy of a pattern reduced and projected onto a substrate. It is an object of the present invention to provide a reduction projection exposure apparatus that prevents this and makes it possible to always automatically produce semiconductor elements with highly accurate absolute dimensions.

In order to achieve the above object, the present invention automatically measures marks on a substrate through a reduction projection lens and a reticle, calculates a reduction magnification error, and slightly moves the reduction projection lens in the optical axis direction. The present invention is characterized in that the reduction projection exposure apparatus is configured to correct reduction magnification errors.

An embodiment of the present invention will be described below with reference to FIG.

A pattern on a reticle 22, which is an original image, is illuminated by condensing illumination light from a light source (not shown) with a condenser lens 21, and is reduced and projected onto a substrate 24 through a reduction lens 23. The shaded portion in FIG. 2 is a fixed and immovable reference member.

The substrate 24 is installed on a movable stage composed of a Z movable stage 30 movable in the optical axis direction and an XY movable stage 31 movable in a plane perpendicular to the optical axis, and the focal position imaged by the reduction lens 23 is An automatic focusing mechanism as shown in Japanese Patent Publication No. 55-18043 is provided so that the surface of the substrate 24 is always held. The relative position of the substrate 24 can be determined by the arithmetic circuit 35 by measuring the position of the movable table with a laser length measuring meter 39.

The reduction lens 23 is supported by a guide member 32 with respect to a reference member so as to be movable only in the optical axis direction, and can be moved slightly in the optical axis direction by a drive motor 33. The length is measured by the lens vertical control circuit 38 and fed back to the lens vertical control circuit 38.

Here, a method for measuring the reduction magnification will be described using FIGS. 1a, b, and c.

This reduction projection exposure device includes a condenser lens 1,
It is composed of a reticle 2, a reduction projection lens 3, a substrate 4, a first pattern detector 5, and a second pattern detector 6. Square light-transmissive windows 7 and 8 are provided on the reticle 2, which is the original picture, at positions separated by a predetermined distance L (see FIG. 1b). FIG. 1b shows an enlarged view of the windows 7, 8 mentioned above. When the mark 9 on the substrate 4 is viewed through the reduction lens 3 and the window 7 of the reticle 2, it looks like a mark 9'.
If you look through window 8 after moving by 8, you will see a mark 9'' (see Figure 1b).

The pattern detector 5 detects the position of the mark 9' in the window 7 of the reticle 2 and the pattern detector 6
can measure the length of each mark 9'' within the window 8 of the reticle 2, and the two marks 9'
It is possible to calculate the amount ΔL by which the relative distance between the windows 7 and 9″ of the reticle 2 deviates from the distance L between the windows 7 and 8 of the reticle 2.

Now, the reduction magnification of the reduction projection optical system is m (m<1)
Then, if the mark 9 on the substrate 4 is shifted by ΔL even though it is moved so that l=mL, then the relationship l=(m+Δm)(L+ΔL) is satisfied. The reduction magnification error is Δm, which means that the optical system has reduced the image too much.

On the other hand, as shown in FIG. 1c, when the distance b between the reticle 2 and the reduction lens 3 is increased by Δb in the reduction projection optical system, when the mark 9 is moved by l as described above, , ΔL/
Only ΔL that satisfies the relationship Δb=L/b is marked 9'
This results in a deviation of 9″.

Therefore, from the positional deviation amount ΔL of the mark 9 detected by the first and second pattern detectors 5 and 6, Δb=
Reduce lens 3 by the amount that satisfies the relationship ΔL・b/L
If the relative distance between the reticle 2 and the reticle 2 is narrowed, the reduction magnification error Δm becomes zero. When ΔL is a negative value, the distance between the reduction lens 3 and the reticle 2 may be increased by Δb.

As an example, in the case of a 1/10 reduction projection exposure apparatus, m = 0.1, f = 49.2 mm, a = 54.127 mm, b
= 54.27 mm, l = 10 mm. At this time, Δm
If there is a reduction magnification error that is larger by =0.000001 (10 ^-4 %), the pattern with L ₀ = 100mm will be L ₀ (m + Δm)
(10 mm - 0.1 μm), resulting in a small dimensional error of 0.1 μm. When such a state is measured by the first and second pattern detectors 5 and 6, the length l=10 mm is L+ΔL=l/(m+Δm),
ΔL=1μm is detected larger than L=100mm.

At this time, if the distance between the reduction lens 3 and the reticle 2 is narrowed by Δb=5.4 μm, the reduction magnification error will disappear.

As can be seen from the above description, the reduction magnification error can be determined by simply measuring the positional deviation amount ΔL of the mark 9 on the substrate 4 that has been moved by a predetermined amount l through the reduction lens 3 from above the reticle 2 at a predetermined distance L. By finely adjusting the distance between the reduction lens 3 and the reticle 2 so as to satisfy the relationship Δb=ΔL·b/L, the reduction magnification error can be corrected.

By the way, in the conventional reduction projection exposure apparatus, once the distance b between the reticle 2 and the reduction lens 3 is accurately fixed, the reduction magnification does not change.

However, in recent years, as semiconductor devices have become more precise, in order to prevent dust from adhering to the original image, that is, the two surfaces of the reticle, the reticle 2 is covered with a cover glass, and air is left between the reduction lens 3 and the substrate 4. Reducing projection exposure apparatuses have been developed that operate using media other than those described above. In this way, when a medium other than air is interposed in the reduction projection optical system, as long as the refractive index and thickness of the medium remain constant, the difference in refractive index However, it is usually extremely difficult to maintain the refractive index and thickness of the medium constant, and this results in a reduction magnification error.

Therefore, in the present invention, even if there is a medium having an arbitrary refractive index or thickness between the reduction exposure optical system as described above, the reduction magnification is always increased by using the pattern detectors 5 and 6 as described above. The feature is that it allows errors to be detected and corrected.

The present invention will now be explained in more detail by returning to the embodiment shown in FIG. The pattern detector 25 passes through the reticle 22 and the reduction lens 23 to the substrate 24.
The position of the upper mark 29 is detected, the detection signal is sent to the detection circuit 37 and the calculation circuit 36, and the reduction lens 23 is slightly moved in the optical axis direction by the lens vertical control circuit 38 and the driving system such as the motor 33. will be done.

Although not shown, a medium having an arbitrary refractive index or thickness may be present anywhere between the reticle 22, the reduction lens 23, and the substrate 24, and the functions of the present invention can be utilized.

In the embodiment of the present invention, one mark 9, 29 on the substrate 4, 24 is moved by a predetermined distance l, and the windows 7, 8 of the reticle 2, 22 are moved by a distance L.
Although the positions of the marks 9 on the substrates 4 and 24 are measured, it is also possible to detect patterns of two marks separated by a predetermined distance l at the same time. In addition, the marks 9 on the substrates 4, 24 are detected at the predetermined positions of the two predetermined windows 7, 8 without simultaneously detecting the pattern and measuring the amount of deviation ΔL of the reticle 2, 22 with respect to the windows 7, 8. As an example of application, it is also possible to control the reduction lenses 3 and 23 by moving them up and down so that they match. Furthermore, reticle 2,2
The positions of the two windows 7 and 8 on 2 are set at a predetermined distance L.
As long as it is known, it can be placed anywhere, and as long as the origin of the two pattern detectors 5 and 6 has good accuracy, the substrate 4,
The mark 9 on the reticle 2, 22 is not detected with respect to the windows 7, 8 on the pattern detector 5, 6.
Detection may also be performed with respect to the origin.

Moreover, instead of providing two pattern detectors 5 and 6 independently, one pattern detector can be moved by a predetermined distance L, or such a function or detection range can be provided in one pattern detector. It is also possible.

In the embodiment of the present invention, the relative distance between the reduction lenses 3, 23 and the reticle 2, 22 is adjusted by moving the reduction lenses 3, 23 up and down, but it is sufficient to adjust the relative distance between the two. , similarly to the reduction lenses 3 and 23, it is also possible to achieve the objective by moving the reticle 2 and 22 up and down using a known guide drive means that is movable only in the optical axis direction.

Note that by adjusting the reduction magnification, the distance a to the optimum focus position of the imaging plane changes, and the reduction magnification changes slightly accordingly, but normally the reduction lenses 3 and 2
3 uses a telecentric wire-connected optical lens, so the reduction magnification error and defocus caused by changes in distance a are slight. Therefore, first, when the medium changes, the marks 9 and 29 on the substrates 4 and 24 are focused, and then the reduction magnification error is corrected using the method described above. If accurate correction is required, it is easy to think that it is sufficient to perform focusing again and then repeat correction of the reduction magnification.

In addition, as an application example of the present invention, it is possible to install a pattern detector at the position of the board and measure marks at predetermined intervals on the original reticle to adjust the reduction magnification. The method for correcting this can be easily deduced by considering it upside down through a reduction projection lens.

According to the present invention, it is possible to detect reduction magnification errors in an original image pattern that is reduced and projected onto a substrate, which is caused by changes in the distance of the projection exposure optical system or by the presence of a medium having an arbitrary refractive index or thickness. , can be automatically corrected, making it possible to manufacture semiconductor devices with higher precision than conventional ones using the reduction projection exposure apparatus of the present invention.

[Brief explanation of the drawing]

Figures 1a, b, and c are explanatory diagrams for explaining the principle of the present invention, in which figure a is a schematic configuration diagram of a reduction projection exposure device and a pattern detector, and figure b is an original reticle. FIG. 2 is a schematic diagram of the marks on the substrate as seen from the upper window; FIG. FIG. 3 is a diagram explaining the principle of the reduction projection optical system. 1, 21... Condenser lens, 2, 22, 4
2... Reticle, 3, 23, 43... Reduction lens, 4, 24, 44... Substrate, 5, 6, 25...
Pattern detector, 7, 8... window, 9, 9', 9'',
29...Pattern, 30...Z moving table, 31...
XY moving table, 32...Guide member, 33...Drive motor, 34...Differential coil, 35, 36...Arithmetic circuit, 37...Detection circuit, 38...Lens vertical control circuit.

Claims (1)

[Claims] 1. A reduction lens for reducing and projecting a pattern on a reticle, which is an original image, onto a substrate; a detection means for detecting a predetermined mark through the reduction lens; calculation means for calculating the error in the reduction magnification of the original image pattern based on the mark detected by the detection means; and an optical calculation means for calculating the error in the reduction magnification of the original pattern based on the mark detected by the detection means; 1. A reduction projection exposure apparatus comprising: a mechanism for finely adjusting relative spacing. 2. The reduction projection exposure apparatus according to claim 1,
As the detection means, first and second pattern detectors are installed with a predetermined distance apart in order to detect the position of a predetermined mark provided on the substrate, and the calculation means detects the position of the predetermined mark provided on the substrate. a first position of the substrate when the predetermined mark provided on the substrate is detected by the first pattern detector; and a second position of the substrate when the predetermined mark provided thereon is detected by the second pattern detector; A reduction projection exposure apparatus characterized in that the reduction projection exposure apparatus is a means for calculating the reduction magnification from .