JPH07321018A - Positioning method of semiconductor aligner - Google Patents

Abstract

PURPOSE:To correct the error of demagnification factor of a reduction projection aligner. CONSTITUTION:In a reduction projection aligner, a part of exposure rays 1 is used, and marks 4 of at least two or more points on a reticle 3 are formed as images on the surface of a water stage 5, with a demagnification projection lens 2. By measuring the positions of the images, the rotation of shot, the error of demagnification factor, and the deviation of translation are corrected.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor exposure apparatus alignment mechanism and method. The miniaturization of devices has been remarkable in recent years, and in order to form fine resist patterns in multiple layers on a semiconductor substrate, a more precise mechanism is required in the alignment method of the semiconductor exposure apparatus.

[0002]

2. Description of the Related Art FIGS. 4 to 6 are explanatory views of a conventional alignment method. In the figure, 1 is an exposure light beam, 2 is a reduction projection lens, 3 is a reticle, 4 is a mark, 5 is a wafer stage, 6 is a shot, 7 is a fly-eye lens, 9 is a reticle blind, 10 is a wafer, 11 is light intensity. Sensor, 18 light source, 19 reflection mirror, 20 filter, 21 condenser lens, 25 shot area, 26 reflection mirror, 27 grating, 28 grating, 28 reflected light, 29 detector, 30 microscope, 31
Is a CCD camera, 32 is an input lens, 33 is a first relay lens, and 34 is a second relay lens.

Conventionally, a semiconductor exposure apparatus, in particular, a step-and-repeat type reduction projection exposure apparatus (hereinafter referred to as a stepper) has been used, as shown in FIG.
When the reticle pattern is reduced and projected for each shot 6 by the stepper on the upper resist film, it is necessary to accurately position the XY coordinates of the shot position to be projected.

The following two methods are mainly used as the positioning method used in this stepper. (1) Die-by-die alignment method.

With the alignment method in which exposure is performed on the wafer on a shot-by-shot basis, as shown in FIG. 4B, the alignment marks 4 arranged for each shot are detected and the position is corrected. After the exposure, the operation of moving the wafer stage to the position of the next shot 6 is repeated for each shot 6.

(2) Global alignment method. By detecting several marks 4 for alignment of each shot 6 to be exposed on the wafer 10 before exposure,
This is a method in which the array lattice of the shots 6 is obtained in advance, and the wafer stage is moved according to it to perform exposure.

In addition to the above method, as shown in FIG. 5C, the diffracted and reflected light 28 from the grating on the wafer 10
Is detected by the detector 29, and the mark 4 on the wafer 10 as shown in FIG.
There are various types of mark detection methods, such as a method of taking a picture with and sending it to an image memory.

[0008]

However, even if the alignment is carried out by the above-mentioned alignment method, the positional deviation actually remains inside the shot 6 due to the following factors.

(1) Position shift caused by telecentric shift. Generally, as shown in FIG. 6, an optical system designed so that the light source 18 is parallel to the optical axis on the surface (image surface) of the wafer 10 is called a telecentric optical system. In such an optical system, the size and position of the image do not change even if the image plane is slightly shifted in the optical axis direction.

The optical system of the stepper is also a telecentric optical system. However, the illumination optical system of the stepper is
The situation becomes complicated because the fly-eye lens (compound eye lens) 7 is used to smooth the light from the mercury lamp, which is 18.

As shown in FIG. 6, the direction of the light incident on an arbitrary point on the reticle 3 is the sum of the lights emitted from the cells of the fly-eye lens 7. Therefore,
The direction of light incident on an arbitrary point on the reticle 3 can be defined as a weighted average of direction vectors of light emitted from each cell of the fly-eye lens 7 at respective intensities. Since the processing accuracy of the fly-eye lens 7 is naturally limited, it is difficult to make the incident illumination light parallel to the optical axis on the entire surface of the reticle 3 in the meaning defined above.

The same applies to the surface of the wafer 10 which is conjugate with the surface of the reticle 3 via the reduction projection lens 2. Therefore, when the surface of the wafer 10 is displaced in the optical axis direction, each point of the image is displaced laterally. To do. As a result, the image of the shot 6 is distorted into an arbitrary shape, but especially translational deviation, reduction ratio error, and rotation error are remarkable.

(2) Reduction rate error. The reduction ratio of the reduction projection exposure apparatus depends on the environment around the reduction projection lens 2, that is, the atmospheric pressure, the temperature, and the like. In order to always perform exposure at the same reduction ratio, the reduction projection exposure apparatus is equipped with a control mechanism such as a lens controller that constantly monitors the environment and corrects the reduction ratio. However, since this mechanism does not directly monitor the magnification, some reduction ratio error actually occurs.

In view of the above problems, it is an object of the present invention to develop a mechanism in which no alignment error occurs and to obtain a highly accurate alignment method means using the mechanism.

[0015]

FIG. 1 is a diagram for explaining the principle of the present invention, which is a system diagram of the alignment system of the present invention.

In the figure, 1 is an exposure light beam, 2 is a reduction projection lens, 3 is a reticle, 4 is a mark, 5 is a wafer stage, 10 is a wafer, 11 is a light intensity sensor, 12 is a laser length measuring device, and 13 is a CPU. , 14 is a register, 15 is a reticle stage control system, 16 is a wafer stage control system, and 17 is a lens controller.

In order to solve the above problems in the present invention,
In the stepper alignment method, the actual position of the reticle image is detected. At this time, it is necessary to detect at least two or more positions on the reticle. Further, in consideration of the influence of the telecentric shift, a part of the actual illumination light after the fly-eye lens is used as the measurement light.

For the measurement, as shown in FIG. 1, a light intensity sensor 11 for detecting the position of the reticle image is arranged on the wafer stage 5, and the wafer stage 5 is run to perform the measurement.
After that, the conventional alignment measurement is performed.

From the position of the reticle image, the reduction ratio, the rotation error, and the translational deviation are detected. The reduction ratio is sent to the control system of the lens controller 17, and the rotation error is sent to the reticle stage control system. The translational deviation value is added to the translational deviation value detected by the conventional alignment method.

That is, as shown in FIG. 1, an object of the present invention is to use at least two or more marks on a reticle on a stage surface with a projection lens by using a part of an exposure light beam in a reduction projection exposure apparatus. This is achieved by combining the images, measuring the position of the images, and correcting shot rotation, reduction error, and translational deviation.

[0021]

As described above, in the present invention, the reduction ratio, the rotation error, and the translational deviation are detected from the position of the reticle image. The reduction ratio is sent to the control system of the lens controller, and the rotation error is sent to the reticle stage control system. The translational deviation value is added to the translational deviation value detected by the conventional alignment method.

Therefore, it is possible to reduce the positional deviation inside the shot due to the illumination / projection optical system that remains in the conventional alignment method, and it is possible to perform highly accurate overlay exposure.

[0023]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is an explanatory view of an alignment mechanism of the present invention, and FIGS.

In the figure, 1 is an exposure light beam, 2 is a reduction projection lens, 3 is a reticle, 4 is a mark, 5 is a wafer stage, 7 is a fly-eye lens, 8 is a blind, 10 is a wafer, 11 is a light intensity sensor, 18 is a light source, 19 is a reflection mirror, 20
Is a filter, 21 is a condenser lens, 22 is a reference position, 23
Is a conventional alignment method, and 24 is a focus reference plane.

First, the peripheral elements constituting the present invention, which are not described in the above-mentioned means, will be described. (1) In order to use a part of the illumination light for position detection, at the time of measurement, as shown in FIG. 2, the alignment mark 4 on the reticle 3 is aligned with the position of the reticle blind 9 of the conventional example. A blind 8 is inserted so that the illumination light only hits the area. Since the pattern on the surface of the blind 8 forms an image on the reticle 3 once, the area including the mark 4 on the surface of the reticle 3 can be sharply illuminated. As the position detecting mark 4 on the reticle 3, a mark as shown in FIG. 3A is used.

(2) It is necessary to know in advance the relative positions of the light intensity sensor 11 and the conventional alignment system in order to use it together with the conventional alignment mechanism. Therefore, by providing the conventional alignment mark 23 used in the conventional alignment system in the vicinity of the light intensity sensor 11, and detecting the alignment mark 23 of the conventional system by the alignment system of the conventional system, Find its relative position. Note that the relative position between the light intensity sensor 11 and the conventional alignment mark 23 must be accurately checked in advance.

(3) Considering the influence of the telecentric shift, a light intensity sensor for measuring the position of the reticle image.
11 height and the height of the wafer 10 surface when actually performing the exposure,
That is, it is necessary to make the heights fitted to the focal plane of the reduction projection lens 2 equal by the autofocus mechanism provided in the stepper.

Therefore, as shown in FIG. 3 (b), a flat surface about the shot area is provided near the light intensity sensor 11, and when measuring the position of the reticle image, this surface is autofocus mechanism. Is used to adjust the focal plane, that is, the wafer 10 plane at the time of exposure. Actually, the height of the flat surface for height measurement and the height of the light detection surface of the light intensity sensor 11 are slightly different, so the difference should be accurately investigated beforehand, and at the time of measurement, the difference from the focal plane should be measured. Just move further.

(4) As the light intensity sensor 11 incorporated in the wafer stage 5, a light intensity detector having sensitivity to the exposure wavelength, such as Photomul, is used. The measurement is performed by moving the wafer stage 5 to a reference point as shown in FIG.
The wafer stage 5 is scanned in the XY directions to detect the intensity signal. The position of the image of the mark 4 on the reticle 3 is obtained by the signal peak position.

FIG. 1 shows a stepper incorporating the alignment mechanism of the present invention. The sequence until exposure is as follows. 1) Move the light intensity sensor 11 to the measurement position of the conventional alignment mark 23 in the vicinity, read the position,
Write to the control system register 14.

2) The height reference plane is moved to a position directly below the autofocus measurement light, and the wafer stage 5 is adjusted to the focal plane using the autofocus mechanism. Furthermore, the register
The height is displaced by the difference between the reference surface and the light detection surface written on 14.

3) Replace the reticle blind 9 with the reticle image position measuring blind 8. 4) The wafer stage 5 is moved to measure the position of the reticle image.

5) From the measured value of the position of the reticle image, the reduction ratio, rotation, and image center position shift are calculated and written in the register 14. 6) Perform conventional alignment. At the time of exposure, the wafer stage 5 is moved by adding the translational displacement value of the shot 6 obtained here with the displacement of the shot image center position written in the register 14. The reduction rate is added to the correction value of the lens controller 17, and the rotation value is given to the reticle stage to rotate the reticle 3.

[0034]

As described above, according to the present invention,
In the stepper, highly accurate alignment can be realized, and a semiconductor element such as an LSI having a fine pattern can be manufactured with high accuracy.

[Brief description of drawings]

FIG. 1 is a positioning mechanism of the present invention.

FIG. 2 is an explanatory diagram of an embodiment of the present invention (No. 1)

FIG. 3 is an explanatory diagram of an embodiment of the present invention (No. 2)

FIG. 4 is an explanatory diagram of a conventional example (No. 1)

FIG. 5 is an explanatory diagram of a conventional example (No. 2)

FIG. 6 Telecentric optical system

[Explanation of symbols]

 In the figure 1 exposure light beam 2 reduction projection lens 3 reticle 4 mark 5 wafer stage 7 fly eye lens 8 blind 10 wafer 11 light intensity sensor 12 laser length measuring machine 13 CPU 14 register 15 chicle stage control system 16 wafer stage control system 17 lens controller 18 Light source 19 Reflection mirror 20 Filter 21 Condenser lens 22 Reference position 23 Conventional alignment mark 24 Focus reference plane

Claims (3)

[Claims] 1. A reduction projection exposure apparatus, an exposure light beam
Use part of (1) and reticle with reduction projection lens (2)
(3) Connect at least two or more marks (4) on the wafer stage (5) as an image, measure the position of the image,
A method for aligning a semiconductor exposure apparatus, which comprises correcting a rotation of a shot (6), a reduction ratio error, and a translational shift. 2. A method for aligning a semiconductor exposure apparatus according to claim 1, wherein a part of the exposure light beam (1) uses the exposure light beam (1) transmitted through the fly-eye lens (7). . 3. A blind for irradiating only an area including the two or more marks (4) with a part of the exposure light beam (1).
3. The method for aligning a semiconductor exposure apparatus according to claim 1, wherein (8) is used by replacing it with an ordinary reticle blind (9).