JPH0142131B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention generally aligns the positional relationship between a first figure that has already been formed and a second figure to be newly formed thereon, and creates an optical image of the second figure. a device that exposes a first figure with a predetermined positional relationship;
More specifically, the present invention relates to a method for calibrating a projection exposure apparatus that prints circuit patterns on a wafer in an integrated circuit manufacturing process.

Conventionally, pattern exposure methods used for semiconductor integrated circuits and the like can be roughly divided into two methods: contact exposure method and projection exposure method. The former is a method that has been used for a long time and has advantages such as a simple configuration of the exposure equipment and productivity as it can expose the entire wafer at once, but it is difficult to bring the mask into complete contact with the entire wafer. Therefore, if a gap is formed, a fine circuit pattern cannot be transferred due to light diffraction phenomenon, and the wafer surface is easily damaged when the wafer is brought into close contact with the wafer. On the other hand, the latter method has the advantage of not damaging the wafer or even the mask since it can be exposed without contacting the wafer. However, in order to print a fine pattern by the projection exposure method, it is usually necessary to increase the reduction ratio due to the difficulty in manufacturing the projection lens, and the exposed area inevitably becomes smaller. For example, a high-resolution lens capable of resolving a pattern with a width of 1 μm has a maximum exposure range of about 14 mmφ at a magnification of 1/10. Therefore, in order to print a larger wafer, the wafer must be sequentially stepped and exposed several times. Conventional projection exposure equipment optically observes the positioning mark on the wafer and the positioning mark on the mask through a projection lens each time the wafer is fed one step, and usually moves the mask slightly to align the two marks. exposure was carried out. Therefore, in order to expose one wafer, alignment operations were required several to several dozen times, making the apparatus unproductive. In addition, when aligning both marks, it is necessary to use light with a wavelength different from the wavelength of the light used during exposure so as not to expose the photoresist coated on the wafer. The equipment itself, including the switching mechanism, was also complex. As a projection exposure apparatus that improves the shortcomings of the conventional projection exposure apparatus, that is, a projection exposure apparatus that completes alignment with only one operation for each wafer,
IBM Technical Disclosure Bulletin Volume 13, Issue 7
There is a method proposed in P1816. That is,
A microscope observation device dedicated to wafer positioning is installed outside the projection exposure optical system, and after positioning the wafer with respect to the optical axis of this microscope device, the wafer is moved a certain distance and guided into the exposure optical system. Align the mask and wafer.
After that, the wafers are transferred in a certain absolute amount at a time.
Precise step feeding in XY and exposure are performed to complete exposure of the entire surface of one wafer. With this method, the wafer only needs to be positioned once before exposure.
Although this is very productive, on the other hand, it is difficult to maintain a constant spatial absolute positional relationship between the microscope device and the exposure optical system over a long period of time, and poor relative positioning accuracy between the mask and the wafer is likely to occur.

The present invention has been made in order to eliminate the above-mentioned drawbacks of the conventional projection exposure apparatus, and it only requires one wafer positioning when exposing one wafer, and it also provides a highly accurate mask and wafer that is stable for a long period of time. The present invention provides a method for calibrating a projection exposure apparatus, which calibrates a projection exposure apparatus so as to perform relative positioning.

The prior art on which the present invention is based and the principle of operation of the present invention will be explained with reference to FIG. Although FIG. 1 shows only one axis, the same applies to the two XY axes. The purpose here is to form an optical image of the circuit pattern 2 on the mask 1 by the reduction lens 3 so as to accurately overlap the pattern 5 on the wafer 4. In the conventional apparatus, a mark detection optical system, for example, a microscope optical system, includes an objective lens 6, a reflector 7, an eyepiece 8, and an epi-illumination system (not shown), which are firmly fixed to the reduction lens 3. system (hereafter explained in terms of microscope optical system)
Observe the pattern 5 _on the wafer 4 with
Align. On the other hand, if the circuit pattern 2 on the mask 1 is on the central optical axis _l2 of the reduction lens 3, the pattern 5 on the wafer 4 is arranged so that it coincides with the intersection point 0 of the central optical axis _l2 and the plane P. By moving the wafer 4, that is, by moving the wafer 4 by a certain distance AO= _L1 , the optical image of the circuit pattern 2 taken by the reduction lens 3 can be made to match the pattern 5 on the wafer 4. In this case, in order to arrange the circuit pattern 2 on the mask 1 on the optical axis _l2 , a mask positioning mark 9 is provided on the mask 1 in advance, and the mark 9 is determined from the microscope optical system and the reduction lens 3. This can be done by aligning the mark 9 with the unique position 10 on the stationary coordinate system. However, in actual equipment, the distance between the reduction lens 3 and the mask 1 is large, about 300 to 600 mm, and even if the members that connect them are manufactured with high rigidity, temperature changes and external vibrations may cause the stationary coordinate system to The characteristic position 10 relative to the mask changes by about 5 to 10 μm over a period of several days, and it is difficult to perform mask alignment with an accuracy of 0.1 to 0.2 μm over a long period of time. However, since this displacement occurs gradually rather than abruptly, correct mask alignment can be performed by knowing the relationship of the unique position 10 with respect to the stationary coordinate system at any time and correcting this.

Therefore, in the present invention, a detection mark 11 is newly provided on the mask 1, and the detection mark 11 can be enlarged and observed using a reflecting mirror 12 and an eyepiece 13. Furthermore, wafer 4 is
When moving from the position to the B position, the reflected light from the pattern 5 on the wafer 4 is conversely magnified by the reduction lens 3 and formed into an image on the detection mark 11, and then by the magnification observation system consisting of the reflection mirror 12 and the eyepiece 13. , for example, a ten-shaped wafer 4 as shown in FIG.
An image 14 of the upper pattern 5 and an image 15 of the detection mark 11 surrounding it can be observed within the same field of view. In the following description, when the center of a ten-shaped pattern such as wafer pattern 5 and the center of a surrounding pattern such as detection mark 11 match, it is defined that the two patterns match. Now, the first
If the wafer pattern image 14 and the detection mark image 15 match at position B in the figure, since the distance between the circuit pattern 2 on the mask 1 and the detection mark 11 is known, the reduction ratio of the reduction lens 3 is determined by the distance. You can find out OB=L ₂ by multiplying . AB, that is, after the center of the wafer pattern 5 is determined by the microscope optical system, the enlarged image 14 of the wafer pattern becomes the detection mark image 15.
By measuring the distance that the wafer 5 has been moved until the wafer 5 matches , it becomes possible to know AO = L ₁ , and thereby, regardless of the change in the characteristic position 10, the circuit pattern on the wafer pattern 5 and the mask 1 2 mask alignment can be performed.

FIG. 3 is an overhead view showing a detailed embodiment of the present invention. The wafer 4, on which the positioning pattern 5 has been formed in the pre-processing process, is vacuum-adsorbed onto a movable stage 18 that is moved in XY directions by drive mechanisms 16 and 17. The amount of movement of this moving table 18 is the reflecting mirror 19, 2.
By the length measurement system consisting of 0 and laser interferometer 21,
Measured with an accuracy of 0.1 μm.

On the other hand, the mask 1 is vacuum-adsorbed by pulse motors 22 and 23 onto a fine movement table 24 that can move about ±50 μm in XY directions, and two vibrating slit photoelectron microscopes 25 and 26 are fixed to a stationary coordinate system.
Two mask positioning marks 2 on the central optical axis of
The fine movement table 24 is moved so that the centers of the positions 7 and 28 coincide, and the mask 1 is placed at a fixed position in the stationary coordinate system. Objective lenses 29, 30, reflecting mirrors 31, 32,
Eyepiece lenses 33, 34 and visual field cross mark 3
Two microscope optical systems consisting of 5 and 36 are provided to determine the center position of two wafer positioning marks 37 and 38 on the wafer 4. The moving stage 18 is moved so that the two wafer positioning marks 37 and 38 match the two field cross marks 35 and 36 of the microscope optical system, and when they match, the laser measuring device 21 is reset and the origin of the XY coordinates is set. shall be. After that, mask 1 is applied to the circuit pattern (not shown in the figure) on the wafer that has been processed in the previous process.
The movable table 18 is positioned at the predetermined coordinates to overlap the new circuit pattern 39 above, using the laser interferometer 21 as a reference. When the positioning is completed, the shutter 40 is opened, and the light from the mercury lamp 41 is converted into parallel light by the condenser lens 42 and irradiates the mask 1, exposing the photoresist on the wafer 4 through the reduction lens 3. Thereafter, the movable table 18 is positioned at new coordinates one after another, and exposure is repeated each time. Regarding the positioning of the moving table 18 with respect to the target coordinates,
Due to the exposure principle already proposed by the present inventors, very high accuracy is not required. In other words, if there is a positioning error of the moving table 18 with respect to the target coordinates, the fine movement table 24 moves in the opposite direction by an amount ten times the error according to the command from the control circuit 54, and the exposure optical system is equivalently Then, the positioning error is corrected and the circuit pattern 39 of the mask 1 is exposed at the correct position on the wafer 4.

In the above mechanism, if the two photoelectron microscopes 25 and 26 corresponding to the unique point 10 do not change with respect to the static coordinate system based on the reduction lens 3 and the laser interferometric measurement system, accurate mask alignment is always achieved. However, as mentioned above, since changes occur in reality, it is necessary to calibrate them. During calibration, the illumination system consisting of a mercury lamp 43, a condenser lens 44, and an optical fiber 45 is moved to the moving table 1.
8, and the target mark 49 is illuminated with light guided by the fiber through a reflecting mirror 47 and a condensing lens 48. This target mark 49 is made by a commonly used photomask making process,
The metallic chromium film is attached only to the cross pattern part with a width of 2 μm. On the other hand, a detection mark 50 to be aligned with the target mark 49 is placed on the periphery of the mask 1.
is provided. The position of this detection mark 50 is provided at a position common to all masks with respect to the mask positioning marks 27 and 28. Detection mark 5
A magnifying optical system consisting of a reflecting mirror 51, an objective lens 52, and an eyepiece lens 53 is provided above the lens 0. Now, while illuminating the target mark 49 with the light guided by the optical fiber 45,
The moving table 18 is moved so that the target mark 49 is directly under the microscope optical system including the objective lens 29, and the image of the target mark 49 is made to coincide with the cross mark 35 within the field of view of the microscope optical system. The XY coordinate position of the moving table at this time is determined by the laser interference measuring device 21.
Read and record it. Next, the moving table 18 is moved again so that the target mark 49 is located below the reduction lens 3, and this time the magnification optical system including the objective lens 52 is observed. Since the detection mark 50 on the mask 1 is observed together with the cross image of the target mark 49 within the field of view, the moving table 18 is slightly moved so that they coincide, and the XY coordinate position of the moving table 18 at the time of coincidence is again set. Record.

Through the operations described above, it was possible to know the distance L ₁ - L ₂ for each X and Y axis, and from the position of the detection mark 50 on the mask 1, the distance L ₁ for each X and Y axis was determined. Since we have done this before, we can easily know L ₁ for each of the X and Y axes. As a result, it was possible to calibrate the spatial positional relationship of the mask 1 with respect to the stationary coordinate system based on the reduction lens 3 and the laser interferometric measurement system.

In the above embodiment, there is one target mark 49, but it is clearly possible to use two marks separated by the distance between the centers of the objective lenses 29 and 30. However, in this case the fiber illumination system becomes somewhat more complex. It is also possible to perform similar calibration without providing these dedicated target marks on the moving table 18 and using a reference wafer on which wafer positioning marks 37 and 38 are formed, but in this case, the illumination of the target marks Naturally, the light becomes reflected illumination through the reduction lens 3, and high target image contrast cannot be obtained, resulting in a disadvantage that the accuracy of matching with the detection mark 50 on the mask 1 is reduced.

As explained above, according to the present invention, the relative positioning of the wafer and the mask can be performed once per wafer.
Since the process can be completed in just one time and the accuracy of alignment can be easily confirmed, highly productive reduction projection exposure becomes possible.

[Brief explanation of drawings]

Fig. 1 is a diagram showing the principle of the present invention as well as the operating principle of the conventional device, Fig. 2 is a diagram showing a microscope field image when the wafer pattern and the detection mark on the mask match, and Fig. 3 is a diagram showing the implementation of the present invention. It is an example overhead view. 1...Mask, 3...Reducing lens, 4...Wafer, 6...Objective lens, 8...Eyepiece lens, 12
... Reflector, 13 ... Eyepiece, 18 ... Moving table, 54 ... Control circuit.

Claims (1)

[Claims] 1. When exposing a series of mask figures to a desired position on a photosensitive substrate via a projection lens in an exposure optical system, the series of masks is always placed in the same relative position with respect to the projection lens; Performing optical relative positioning of the series of masks and the photosensitive substrate by detecting the center of a reference mark provided on the photosensitive substrate with a mark detection optical system fixed outside the lens optical path;
an exposure optical system including an enlarging optical system for simultaneously observing a real image of the alignment mark in the focal plane of the projection lens formed on the mask surface by the projection lens and alignment figures of the series of masks; and a mark detection optical system, the method comprises: detecting the center of the alignment mark by the mark detection optical system, and detecting the position corresponding to the optical axis center of the mark detection optical system; The position of the alignment mark is determined, and the position of the alignment mark is determined in a state where the alignment mark is moved and the matching of the real image of the alignment mark and the alignment figure is detected by the enlarging optical system, and the position of the alignment mark is determined, The position of the alignment mark and the exposure in a state where the matching of the real image of the alignment mark and the alignment figure is detected by the enlarging optical system based on the distance from the center to the alignment figure and the magnification of the projection lens. A method for calibrating an exposure apparatus, characterized in that the spatial relationship between the exposure optical system and the mark detection optical system is determined and calibrated by determining the distance between the exposure optical system and the center position of the optical system.