JP2775519B2 - Dual focus device using reference reticle - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention accurately determines a relative position between a first object and a second object, each of which is, for example, a mask and a wafer arranged at minute intervals in an optical axis direction. The present invention relates to a position detecting device for detecting the position.

[Prior Art] If the present invention is applied to only an alignment apparatus of an X-axis exposure apparatus, the position detection method is roughly classified into the following three methods.

Pattern measurement method Fresnel diffraction method Diffraction grating method In this classification, the present invention relates to a pattern measurement method. Therefore, a conventional technique of this pattern measurement method will be briefly described.

 First, the known literature is listed.

(1) "Study on X-ray exposure alignment method" August 1980 Precision Machinery, Vol. 46, No. 8, PP 83-88 (2) "Evaluation of alignment performance of X-ray exposure apparatus SR-1" 1985 Precision Machinery, No. Vol. 51, No. 5, pp. 156 to 161 (3) "Alignment pattern detection method for X-ray exposure apparatus using oblique imaging optical system" Sep. 1989 Precision Engineering Journal PP139 to 145 (4) "Chromatic aberration double focus apparatus" High-accuracy Position Detection by the First Report "1990 Fall Meeting of the Japan Society of Precision Engineering, PP 667-668 In the above document (1), a vibration type photoelectric microscope is used as a detection system as an alignment device. That is, in FIG. 8, in order to make the focal plane of the photoelectric microscope 30 coincide with the X-ray mask 1 and the wafer 2, the objective lens 31 is moved in the direction of the optical axis by the optical axis moving mechanism of the photoelectric microscope 30 (not shown).
The alignment marks on the axis mask 1 and the wafer 2 are separately detected. Specifically, first, the optical axis is moved so that the focal plane of the objective lens 31 coincides with the X-axis mask 1, and the reference slit 32 in the photoelectric microscope 30 and the enlarged image of the alignment mark (10 times) of the X-axis mask 1 are formed. After detecting the relative position, the optical axis is moved so that the focal plane of the lens coincides with the wafer 2, and the relative position between the reference slit 32 and the enlarged image of the alignment mark on the wafer 2 is detected.

Since the relative position between the X-ray mask 1 and the wafer 2 from the reference slit 32 is calculated by this sequence, the relative position between the X-ray mask 1 and the wafer 2 is obtained.

The disadvantage of this device is that, since the optical axis of the detection system is moved, the amount of movement of the optical axis in a plane perpendicular to the optical axis becomes a detection error. That is, the position of the optical axis when the wafer 2 is detected with respect to the position of the optical axis when the X-ray mask 1 is detected is (Δx,
Δy) If it moves by μm, the detection error will be (Δx, Δy)
Is added by that value.

The problem is the amount of (Δx, Δy), but the detection resolution that the X-ray exposure apparatus aims at is 0.01 μm or less.
Since the overall detection accuracy is 0.05 μm or less, at least (Δx, Δy) ≦ 0.01 μm must be achieved.

However, the optical axis of the moving objective lens is 0.01 μm
It is clear that moving below is extremely difficult.

It can be said that having a mechanical or electrical drive unit in the detection unit as in this example has a problem that must be avoided at least in an apparatus aiming at position detection on the order of 0.01 μm.

Next, FIG. 9 shows the apparatus of the above-mentioned document (2). In this example, an enlarged image of the X-axis mask 1 and the alignment mark on the wafer 2 are simultaneously detected by using a bifocal lens 33 utilizing the principle of birefringence as the objective lens of the alignment scope.

 The problems of this device are listed below.

In order to obtain a high-contrast image with the same resolution (resolution) as that of a normal lens, it is necessary to use the 2
Double numerical aperture (NA) is required.

By using the objective lens having the double numerical aperture as described above, the focal depth becomes extremely shallow, and it becomes difficult to detect the X-ray mask 1 and the wafer 2 in two focal planes.

The gap between the X-ray mask 1 and the wafer 2 cannot be set large. That is, in consideration of the refractive power of the birefringent surface and the aberration of the lens, it is extremely difficult to greatly separate the double focal points from the viewpoint of design and manufacturing. In FIG. 9, the distance between the X-axis mask 1 and the wafer 2 is large. It is set to 10 μm.

Production requires a high level of technology and skills, which increases costs.

As described above, this device has no driving parts compared to the above-described device (1), so that the accuracy is improved accordingly. However, in order to further increase the accuracy, it is necessary to increase the lens resolution. However, since the principle of birefringence is used, the resolution of the lens is reduced to less than half of the normal value, and it is extremely difficult from the viewpoint of design and manufacture.

Next, the device of the above (3) is shown in FIG. This apparatus uses an objective lens with a low numerical aperture and detects the X-ray mask 1 and the wafer 2 from an oblique direction to obtain an X-ray.
The line mask 1 and the wafer 2 can be observed at the same time. Objective lens by detecting diagonally
The X-ray mask 1 and the wafer 2 farther than the depth of focus 34 can be observed at the same time.

 The problems of this device are listed below.

To obtain sufficient resolution, the numerical aperture of the objective lens (N.
A.) may be increased, but the wafer 2
If a sufficiently wide gap corresponding to the mask 1 is to be obtained, the numerical aperture (NA) of the objective lens 34 must be reduced, and the resolution and the gap cannot be satisfied simultaneously.

The focal plane is extremely narrow due to oblique image formation, and a mark formed in a sufficiently wide range cannot be detected due to image blur.

Due to the oblique image formation, the image formation pattern is inclined, resulting in a measurement error.

The tilt error between the illumination optical axis and the detection optical axis directly becomes the detection error.

In summary, by tilting one focal plane of the lens defined by Rayleigh's equation, the X-ray mask and the wafer, which are object planes that are far apart, are simultaneously imaged, so that the resolution and the gap setting are mutually different. They interfere, making it impossible to satisfy both at the same time.

Next, an optical path diagram showing the principle of the device (4) is shown in FIG.
Shown in the figure. In this apparatus, an optical system having two focal planes M and M 'and one imaging plane B which forms images at two different wavelengths on the same optical axis is realized by utilizing the axial chromatic aberration of the objective lens 35. I have. In these two focal planes M and M ', the definition of Rayleigh is independently established, so that the required lens resolution is X
Since it can be freely selected irrespective of the gap δ between the line mask 1 and the wafer 2, it is ideal as a detection optical system.

Further, the present inventor has proposed a "bicolor illumination method and a band light illumination method in a dual focus detection device utilizing chromatic aberration" (see Japanese Patent Application No. 63-261379). Can be detected with band light,
This is a very effective means because the alignment apparatus is not affected by the wafer process such as a resist film at all.

The disadvantage of this device is that it has no technical problems including design and manufacturing, but requires some development cost. This would be the best method if an optical system having a sufficient and sufficient dual focus for measurement can be realized at very low cost with general optical components.

[Problems to be Solved by the Invention] As described above, various devices have been devised for the alignment device of the X-ray exposure apparatus. In each case, an optical system is used for the detection system. In the case where a lens is used for this detection system, as shown in FIG. 6, when the X-ray mask surface 1 and the wafer surface 2 separated by, for example, 40 μm are simultaneously detected by the objective lens 3, the numerical aperture NA of the lens is 0.4. And the focusing range, that is, the depth of focus is 3.5μ from Rayleigh's equation.
X-ray mask surface 1 and wafer surface 2 which are 40 μm apart
Cannot be detected simultaneously. Generally, the characteristics of a lens are expressed by resolution and depth of focus, which are Rayleigh's equation, that is, Is defined by

As is apparent from these equations (1) and (2), the depth of focus and the resolution are determined by the numerical aperture (NA) of the lens if the wavelength λ is constant.

FIG. 7 is a graph showing the relationship between the depth of focus and the numerical aperture (NA). From this graph, the numerical aperture (NA) and the depth of focus are inversely related. When the numerical aperture is increased to increase the resolution, the depth of focus becomes shallower. Conversely, when the depth of focus is increased, the numerical aperture (NA) becomes larger. It can be seen that the lens resolution decreases and the lens resolution decreases.

That is, the X-ray mask surface 1 which is far away in FIG.
In order to simultaneously detect and the wafer surface 2, the depth of focus must be increased, and since the numerical aperture (NA) is set low, the resolution of the lens is greatly reduced.

If the depth of focus is set to 40 μm in order to simultaneously detect the X-ray mask surface 1 and the wafer surface 2, the NA becomes extremely small as 0.08 from the above formula (1), and the resolution is expressed by formula (2). 4 μm.

According to the experiment of the present inventor, since a detection resolution of 0.01 μm is obtained when NA ≧ 0.3, using a general lens defined by the above-mentioned Rayleigh equations (1) and (2), It is impossible to simultaneously detect the X-axis mask and the wafer with sufficient resolution.

The invention of this application has been made in view of such a point, and, after satisfying the basic properties of the lens, detects the X-ray mask and the wafer far apart with sufficient resolution, and measures the relative position between the two. It is to realize a bifocal optical system which is necessary and sufficient for the camera very easily.

Means for Solving the Problems According to the present invention, in a position detecting device for detecting a relative position of a first object and a second object separated by a small distance in an optical axis direction in a direction orthogonal to the optical axis, Mark corresponding to the minute distance between the first object and the second object, having a thickness determined from the magnification of the objective lens, and forming marks necessary for detecting the positions of the first object and the second object on both surfaces. And a reference reticle provided at two image forming positions of an objective lens.

Further, in the present invention, in order to further enlarge the images of the first object and the second object formed on both surfaces of the reference reticle, the two relay lenses are split into optical paths so that the focal points coincide on both surfaces of the reference reticle. This is a bifocal device using a reference reticle, which is characterized in that the reference reticle is arranged.

Embodiment An embodiment of the present invention will be described below with reference to the drawings.
FIG. 1 is an optical path diagram showing a basic configuration of a dual focus device using a reference reticle. In order to simultaneously detect the X-axis mask 1 and the wafer 2 disposed at a gap of δ μm with the objective lens 14, the half mirror 12 is provided on the right side of the objective lens 14.
It is arranged in the optical path at an angle of 45 degrees, reflects illumination light from a white light source 11 such as a halogen lamp from above, and performs epi-illumination through an objective lens 14. The illuminated images of the mask mark 3 and the wafer mark 2 on the X-axis mask 1 pass through the half mirror 12 and are formed separately at points A and B on the optical axis by the objective lens 14. The interval between the points A and B is δ · N ² where N is the magnification of the objective lens 14. At this position, the thickness δ ・ N ²
The reference reticle 16 of the parallel plane plate is inserted, and the wafer reference mark 15 is formed on the optical axis at the center of the point A on the left side, and the mask reference mark 17 is formed above and below the point B on the right side using a reticle such as chrome. .

Accordingly, the wafer mark 2 forms an image magnified N times by the objective lens 14 at the point A, and this image, together with the wafer reference mark 15, is lowered by the half mirror 18 provided at an angle of 45 degrees on the optical axis. Reflected, first relay lens
19 forms an image on a line sensor 20. On the other hand, the image of the mask mark 3 on the X-axis mask 1 is formed at the point B by the objective lens 14, this image passes through the half mirror 18 together with the mask reference mark 17, and is transmitted to the line sensor 22 by the second relay lens 21. An image will be formed.

In the pattern 30 formed on the line sensor 20, the image of the wafer mark 2 is projected in parallel with the image of the central wafer reference mark 15 as shown in FIG. 2 (A).
The pattern 32 formed on the line sensor 22 is a pattern in which the image of the mask reference mark 17 is projected in parallel with the image of the central mask mark 3 interposed therebetween as shown in FIG. 3 (A). . Therefore, the waveforms detected by these line sensors 20 and 22 are the waveforms 31 and 33 shown in FIGS. 2B and 3B, respectively.

In the wafer imaging pattern 30, the relative position α between the wafer reference mark 15 and the wafer mark 2 can be measured. In the mask imaging pattern 32, the relative position β between the mask mark 3 and the mask reference mark 17 can be measured.

Since the wafer reference mark 15 and the mask reference mark 17 are formed on both surfaces of the reference reticle 16, the relative positions of the mark reference position and the mask mark 3 are constant (α−β). ). This relative position (α-β) is not affected at all by the movement of the reference reticle 16 in the optical axis direction.
The relative position when the reference reticle 16 moves by Δx in the direction perpendicular to the optical axis is α−β = (α + Δx) − (β + Δx) = (α−β) (3), which is also completely affected. Nothing.

From the above description, by providing the reference reticle 16 at the image forming position of the objective lens 14, the separate positions A,
It can be seen that the relative position between the wafer mark 1 and the mask mark 3 formed on B can be measured.

That is, the optical system shown in FIG. 1 has a focus on the X-ray mask 1 and the wafer mark 2 respectively, and can measure a relative position in a plane orthogonal to the optical axis between the focal plane and the object. It is possible. Since this optical system satisfies the Rayleigh equations (1) and (2) at each focal plane, the required lens resolution, that is, the numerical aperture (NA), can be freely set irrespective of the gap δ. It is.

Further, since the image of each focal plane is formed by the white illumination light, the image is ideal without any influence on the wafer process. Further, since the white illumination is used, there is no multiple reflection interference between the X-ray mask 1 and the wafer mark 2 and there is no influence of the fluctuation of the gap δ.

Furthermore, since an image is formed in white also on the X-ray mask, there are no restrictions on the thickness, the refractive index, the spectral transmittance, and the like of the mask, which is extremely convenient.

Therefore, this optical system can be said to be an optical system having two focal points necessary for detecting the relative positions of two objects separated by a small distance in the optical axis direction in the direction orthogonal to the optical axis. Also,
In terms of cost, since no special lens or optical component is required and all the components can be constituted by commercial products, the cost is extremely low.

The reference reticle 16 is not a special part, but can be formed easily and extremely inexpensively by depositing one or two layers of chrome or the like on a parallel flat glass plate.

Next, a supplementary explanation will be given on the detailed points in the above embodiment. First, the half mirror 18 and the first and second relay lenses
19 and 21. Half mirror 18 is the reference reticle 1
It is desirable to divide the light path from 6 into two, and to divide the light amount into two at a reflectance and transmittance of 50:50. Therefore, the amount of illumination light from the white light source 11 needs to be twice as large as that of a normal epi-illumination microscope. This increase in the amount of light can be easily handled by increasing the output of the white light source 11. Also,
This optical path branch is caused by the numerical aperture (N.
A.) has no effect.

Next, the reference reticle 16 will be described. Since both surfaces A and B of the reference reticle 16 are the image forming surfaces of the wafer mark 2 and the mask mark 3 on the X-axis mask 1 by the objective lens 14, it is preferable to coat an antireflection film.

For the wafer reference mark 15 and the mask reference mark 17 formed on both sides, general single-layer chromium or double-layer chromium is sufficient.

Next, the magnification error will be described. Although very slight, a magnification error occurs between the wafer mark 1 and the mask mark 3 formed by the objective lens 14. However, this magnification error is caused by the final image of the wafer pattern.
The line sensors 20 and 22 for detecting the mask pattern 30 and the mask pattern 32 can easily make correction. Specifically, after quantizing the electric signals obtained by the line sensors 20 and 22, it is only necessary to adjust a scaling coefficient (corresponding to an optical magnification) when performing digital processing. This is the method already applied in the literature (4).

Next, the resolution of the lens will be described. In the optical system shown in FIG. 1, the optical resolution of the detected wafer mark 2 and the mask mark 3 on the X-ray mask 1 is determined by the numerical aperture (NA) of the objective lens 14. This numerical aperture is set irrespective of the gap δ between the wafer mark 2 and the mask mark 3.

The set value is set in the range of 0.3 ≦ NA ≦ 0.5 (4) from the relationship between the depth of focus of the lens determined by the numerical aperture and the step of the wafer mark 2 and the control accuracy of the gap δ.

The wafer mark 2 which is the object plane of the objective lens 14
The resolution of the mask mark 3 at a position shifted from the wafer mark 2 by a gap δ is very small in design but lower than the resolution of the wafer mark 2, but this is a negligible amount.

Next, signal processing for relative position detection will be described. In the optical system shown in FIG.
The images respectively detected by are converted into electric signals, and become a wafer signal waveform 31 and a mask signal waveform 33. By quantizing these signal waveforms and performing digital signal processing, the relative positions of the wafer mark 1 and the mask mark 3 are obtained. Here, the flow of the digital signal processing is shown in FIG. 4 and FIG.

FIG. 4 corresponds to the quantized wafer signal waveform 31 and mask signal waveform 33. The digital waveform is subjected to a first-order differentiation process to obtain a digital waveform as shown in FIG. The relative position is calculated based on the first derivative data shown in FIG. Is obtained by a correlation equation called similarity pattern matching defined by

The similarity pattern matching is a function that is extremely sensitive to the width W of the left and right patterns, and the correlation value A (W,
The width W of the left and right patterns at which P) is the maximum value is defined as the distance between the marks. This process has already been filed as a patent application as a "method of detecting the position of an alignment mark" by the present inventor. (See Japanese Patent Application No. 63-242598) Next, the thickness and resolution of the reference reticle will be described. Wafer mark 2 generated on both sides of reference reticle 16
The image of the mask mark 3 is an intermediate image formed by the objective lens 14 and has a thickness δ ·
N ² does not affect the resolution of the first relay lens 19 and the second relay lens 21.

 Next, a specific design example in FIG. 1 is shown.

Gap δ = 30 μm Objective lens 14 NA = 0.35 Working distance (WD) = 20 mm Magnification = 10 × Reference reticle 16 Thickness δ · N ² = 3 mm Double-sided anti-reflection film coated Wafer reference mark 15 and mask reference mark 17 = 2
Layer chrome, thickness = 0.3μmt Mark width = 7μm (7/10 = 0.7μm on wafer) First and second relay lenses 19,21 NA = 0.2 Working distance (WD) = 20mm Magnification = 10 × Line sensor 20 , 22 Number of elements = 2048 Element spacing = 14 µm Element width = 200 µm Clock frequency = 12 MHz (maximum) White light source 11 50 W halogen lamp Half mirror 12, 18 Reflectivity and transmittance = 50:50 An aperture must be provided at an appropriate position in the detection optical path so that the reflected light on the inner surface of the lens barrel does not cause a flare, and the flare does not cover the imaged light flux and lower the contrast.

 Finally, the features of the above embodiment are summarized and shown.

(1) Using a normal lens, double focus detection using white light is made possible by using a reference reticle in the optical path.

(2) The detection accuracy of the relative position in the direction orthogonal to the focal plane is:
It is not affected at all by the displacement of the reference reticle.

(3) The distance between the dual focal points can be freely changed by changing the thickness of the reference reticle.

(4) In each of the double focal planes, the definition of Rayleigh in white light is independently established, so that the required lens resolution can be set freely irrespective of the distance between the double focal points.

(5) The white light illumination does not affect the wafer process, the gap variation and the variation of the X-ray mask membrane thickness at all.

(6) Since no special optical parts are used, the apparatus can be configured at extremely low cost.

(7) Since no special optical technology is used, there is no technical problem.

[Effects of the Invention] As described above, the dual focus apparatus using the reference reticle of the present invention is an X-axis mask in an X-axis stepper which is promising for exposure of a next-generation LSI having a minimum pattern line width of 0.5 μm or more. Applied to an alignment device that accurately measures the relative position of the wafer and the wafer, and the relative position in the direction perpendicular to the optical axis of the wafer and the X-axis mask, which are separated by about 30 μm, is not affected by the wafer process. Detection can be performed with high accuracy (on the order of 0.01 μm).

The position signal from the alignment device is transmitted to the driving device as a feedback signal of an X-axis mask or an ultra-precision stage for moving the wafer, and is accurately controlled so that the relative position between the X-axis mask and the wafer has a predetermined positional relationship. It becomes possible.

[Brief description of the drawings]

FIG. 1 is an optical path diagram showing a configuration of a dual focus device using a reference reticle of the present invention. FIGS. 2 (A) and 3 (A) are diagrams showing an image forming pattern on a line sensor. 2 (B) and 3 (B) are waveform diagrams thereof, FIGS. 4 and 5 are waveform diagrams showing signal processing, and FIG. 6 is a positional relationship between the objective lens, the wafer and the mask. FIG. 7 is a graph showing a function of a depth of focus and a numerical aperture; FIGS. 8 to 10 are side and perspective views of a conventional alignment device; FIG. 11 is an optical path of a bifocal lens; FIG. 1 X-ray mask 2 Wafer mark 3 Mask mark 11 White light source 12, 18 Half mirror 15 Wafer reference mark 16 Reference reticle 17 Mark reference mark 19, 21 Relay lens 20,22 …… Line sensor

Claims (2)

(57) [Claims] 1. A position detecting device for detecting a relative position of a first object and a second object separated by a small distance in an optical axis direction in a direction orthogonal to an optical axis, wherein a position between the first object and the second object is determined. A reference reticle having a thickness determined by the magnification of the objective lens and having marks necessary for detecting the positions of the first object and the second object on both surfaces. A bifocal device using a reference reticle provided at two image forming positions. 2. The optical system according to claim 1, wherein the first and second objects formed on both sides of the reference reticle are further optically branched so that the focal points coincide on both sides of the reference reticle. A dual focus apparatus using a reference reticle according to claim 1, wherein the apparatus is arranged.