JPH04186716A - Alignment device - Google Patents

Abstract

PURPOSE:To detect a position of an alignment mark highly accurately while unnecessary noise light is eliminated by correcting an adjustable allowance based on an uneven shape on a resist surface in the vicinity of the alignment mark on a wafer surface by using Moire fringes obtained by two grating patterns different in pitch. CONSTITUTION:Grating patterns G1, G2 different in pitch from each other in a conjugation relation with a wafer W through an optical system are arranged so that their grating directions coincide with an alignment direction, while the pattern G1 is illuminated by a lighting system. Accordingly a light flux from the pattern G1 is incident to an alignment mark face Wa on the wafer surface perpendicularly to the wafer W surface on a plane parallel to a grating direction of the pattern G1 and slantly with a predetermined angle in a plane perpendicular to the grating direction of the pattern G1 to form an image of the pattern G1. The image is projected onto the pattern G2 surface, a signal from an image sensing device 110 is used to calculate a surface of a resist on the alignment mark Wa region by an arithmetic means and relative alignment of the wafer W and the image sending device 110 is performed.

Description

Detailed Description of the Invention (Industrial Application Field) The present invention relates to an alignment device suitable for manufacturing semiconductor devices, and in particular, to an alignment device suitable for manufacturing semiconductor devices, and in particular, to an alignment device that is provided on a wafer surface for relative alignment with an imaging means that is in a conjugate relationship with the wafer. Alignment mark 3
The present invention relates to a positioning device that can perform alignment with high precision by capturing the undulating shape of a resist/surface based on a dimensional structure.

(Prior Art) In recent years, with the demand for miniaturization of circuit patterns of semiconductor devices, exposure apparatuses for manufacturing semiconductor devices are required to have a high precision of relative positioning between a reticle and a wafer.

One of the alignment methods in an exposure apparatus is a so-called one-dimensional pattern match detection method, which is proposed in Japanese Patent Application No. 2-127005, for example.

Figure 8 shows 1 proposed in Japanese Patent Application No. 2-127005.
1 is a schematic diagram of a main part of an exposure apparatus using a dimensional pattern match detection method. In the figure, the positional relationship between the reticle R and the imaging device 86 has already been determined.

A light beam having a wavelength substantially equal to the wavelength of the exposure light emitted from the illumination device 84 is reflected by a mirror 83, condensed by an illumination optical system 82, and a polarized light beam oscillating in a predetermined direction is reflected by a polarization splitter 81. The light is then incident on the objective lens 80. The light beam incident on the objective lens 80 is reflected by a mirror 810, and the light beam that has passed through the transmission portion of the reticle R passes through the projection optical system 11 (also passing through the input/quarter plate 811).

) Alignment marks for positioning on the wafer W surface (
Also called "wafer mark." ) Illuminate Wa.

The wafer mark image (Wa') formed near the R surface of the reticle by the projection optical system 11 is transferred to the mirror 810 and the objective lens 8.
0, a polarized beam splitter, a scanner 81, and a detection optical system 85 to form an image on an imaging surface 86a of an imaging device 86. Note that 114 is an illumination system for exposure, which uniformly illuminates the R surface of the reticle during exposure.

FIG. 9 is an explanatory diagram of a two-dimensional electrical signal regarding a mark image obtained by the image pickup device 86 of FIG. 8.

In the figure, since the relationship between the position of the reticle and the position of the imaging device 86 is determined in advance, the position of the reticle and the wafer can be determined by determining the center position of the alignment mark image on the eight faces of the imaging device on the screen of the imaging device. relationships can be determined.

In FIG. 9(A), Wa' is a wafer mark image formed on the imaging surface 86a. After converting the two-dimensional electrical signal from the imaging device 86 into a two-dimensional discrete digital signal string corresponding to the XY address of the pixel by the A/D converter 87, the integrating device 88a converts the two-dimensional electrical signal into a two-dimensional window 90 of a predetermined size.
The pixels are integrated in the longitudinal direction (y direction) of the wafer mark image Wa' to obtain an integrated signal t(x) which is a one-dimensional discrete electrical signal as shown in FIG. 9(B). Then, the center position of the wafer mark Wa is calculated by the center value calculating device 88b as follows.

First, as shown in FIG. 9(B), a one-dimensional window 91 for evaluation is prepared for the integrated signal t(X). One-dimensional pattern matching is performed on the integrated signal t(x) within the one-dimensional window 91, taking into consideration the cross-sectional shape of the wafer mark Wa. At that time, after minimizing the difference between the integrated signal and the template (the template using the quadratic function shown in Fig. 9(C)) using the least squares approximation method for each point, the integrated signal s (x) is The degree of correlation is defined as the product of the reciprocal of the sum within the one-dimensional window 91 of the absolute value of the difference with the template and the geometric information (absolute value of the quadratic coefficient of the quadratic function) used in the template and approximation. shall be. Accumulated signal s at this time
As for (x), for example, Fig. 10 (A), (B), (C)
, (D), and corresponding templates are shown in FIGS. 11 (A), ('B), and (C).

There is a part shown by the solid line in (D).

In the alignment shown in FIG. 8, the point with the highest correlation value is taken as the center position for wafer mark alignment. The positioning controller 89 drives the XY stage 10 to align the reticle R and the wafer W with respect to the center of the wafer mark thus determined.

(Problem to be solved by the invention) The image of the alignment mark (wafer mark) on the eight surfaces of the sea urchin is affected by the unevenness of the resist applied on that surface and the asymmetry of the three-dimensional structure of the alignment mark. It is known that the image is distorted when observed. This causes a measurement error, so-called sagging, when detecting the position of the alignment mark, which causes a decrease in alignment results.

In particular, in the process of applying resist to the surface of a sea urchin after a metal compound has been deposited by sputtering or the like in the semiconductor device manufacturing process, sagging caused by uneven resist coating near alignment marks is a serious problem. It's full.

On the other hand, the cause of dullness caused by the resist layer in the process of vapor-depositing the metal compound is the high reflectance of the metal compound, and compared to that caused by thin film interference or the wavelength dependence of the refractive index, the cause is geometrical optics due to uneven coating of the resist. The effect of natural phenomena is greater.

FIGS. 7(A) and 7(B) are optical path diagrams showing the phenomenon number at this time. FIG. 7(A) is a cross-sectional view of the vicinity of an alignment mark for positioning a wafer.

In the figure, 71 is a wafer, 70 is a resist, and 73 is an air layer. 72 indicates the optical path of one ray of the observation light.

75 is a groove portion of an alignment mark. The refraction direction of the light ray 72 changes significantly due to local changes (undulations) on the resist surface 70a.

When the reflectance of the sea urchin surface 71a is large, the interference effect between the light reflected from the resist surface 70a and the light reflected from the sea urchin surface 71a is considered to be sufficiently small compared to the geometrical optical effect.

FIG. 7(B) is an explanatory diagram of the distribution of the light intensity of observation light corresponding to FIG. 7(A). The alignment mark image is observed to be laterally shifted due to local changes in the resist surface 70a. This cannot be removed by whitening the illumination light or adjusting the wavelength of the illumination light, as in observing a spoon placed in water, for example. Therefore, it is important to measure the undulating shape of the resist surface 70a and predict the amount of undulation.

On the other hand, since the reflectance of metal compounds is high, the shape of a bright field alignment mark image for a metal compound can be determined by considering only the light distribution corresponding to FIG. 10 (C) described in the conventional technique. good.

Furthermore, a conventional alignment apparatus has been proposed in which alignment is performed after projecting a grating pattern onto a wafer surface. However, they were parallel to the longitudinal direction of the alignment mark on the wafer surface in order to observe the phase shift with the alignment mark on the wafer surface. Therefore, changes in resist film thickness were integrated and could not be separated from signals from alignment marks on the wafer surface. Therefore, it is not suitable as a means for measuring the required undulation shape of the resist surface.

Furthermore, although the conventional positioning apparatus uses two-dimensional information, it is ultimately reduced to one-dimensional information, and the SZN ratio is only improved by integrating other dimensions.

The present invention corrects the unevenness based on the undulating shape (three-dimensional shape) on the resist surface near the alignment mark on the wafer surface by using moiré fringes (stripe shape images) obtained from two lattice patterns with different pitches. The purpose of the present invention is to provide an alignment device that can remove unnecessary noise light and detect the position of a mark with a high degree of convergence, thereby enabling alignment with a high degree of convergence.

(Means for Solving the Problems) The alignment device of the present invention arranges a wafer on which a relief-like alignment mark is formed and an imaging means so that they are in a conjugate relationship via an optical system. When performing relative alignment, a resist is coated on the wafer surface, and two grating patterns G1 and G2 having different pitches are formed in a conjugate relationship with the wafer through the optical system, and their grating directions are The grating pattern G1 is illuminated by an illumination system, and the light beam from the grating pattern G1 is directed onto the alignment mark surface of the sea urchin surface through the optical system. An image of the lattice pattern G1 is formed by making the image of the lattice pattern G1 perpendicular to the surface of the sea urchin eight in a plane parallel to the lattice direction of the lattice pattern and obliquely incident at a predetermined angle in a plane perpendicular to the lattice direction of the lattice pattern. Then, the image of the grating pattern G1 is projected onto the grating pattern 02 surface via the optical system, the grating pattern G1 is imaged on an imaging device, and the signal from the imaging device is used to calculate the image of the grating pattern G1 by a calculation means. The present invention is characterized in that the surface shape of the resist in the alignment mark area on the wafer surface is determined, and relative positioning of the wafer and the imaging means is performed with reference to signals from the calculation means.

In particular, in the present invention, the calculation means is configured to
The method is characterized in that the local three-dimensional shape near the alignment mark on the eight sea urchin faces is determined based on the striped image formed on the surface.

(Embodiment) FIG. 1 is a schematic diagram of a main part of an embodiment of the present invention. In the figure, R is a reticle, and an electronic circuit pattern is formed on its surface. Further, the reticle R is supported by a reticle stage R1. Reference numeral 114 denotes an illumination system for exposure, which includes an ultra-high pressure mercury lamp, a condenser lens, a shutter, etc., and uniformly illuminates the surface of the reticle 8 during exposure. Reference numeral 11 denotes a projection lens (exposure optical system), which reduces and projects the pattern on the 8 surfaces of the reticle onto the wafer W surface (for example, 1
15 to 1/10 times). 1° is a wafer stage, on which the wafer W placed is driven in the x, y, and z directions.

In this embodiment, the pattern on the reticle 8 surface is projected and exposed onto the wafer W surface using the above-mentioned elements.

Incidentally, a relief-type wafer mark (alignment mark) Wa for alignment, which will be described later, is formed on the surface of the wafer W. Furthermore, the reticle R has already been positioned with respect to the main body of the exposure apparatus (for example, the imaging means 110).

Next, each element for aligning the wafer W and the exposure apparatus main body, for example, the imaging means 110 in this embodiment will be explained. In FIG. 1, the optical axis direction of the projection lens 11 is two axes, the left-right direction in the plane of the paper is the y-axis, and the direction perpendicular to the plane of the paper is the y-axis.

In this embodiment, alignment in the X direction is shown as an example, and the same is true for the X direction as well.

Reference numeral 18 denotes a light source means, which emits a light beam having a wavelength different from the wavelength of the exposure light. The light beam from the light source means 18 is reflected by a mirror 19 and condensed by an illumination optical system 16 to illuminate the grating pattern Gl.

As shown in FIG. 2(A), the grating pattern G1 consists of a plurality of gratings with a pitch of λ1, each consisting of a transparent region and a non-transparent region whose grating direction extends in the X direction, and is surrounded by a light shielding plate (not shown). . Reference numeral 14 denotes a correction optical system, which corrects aberrations based on the difference in wavelength used with the projection lens, so that the grating pattern G1 and the surface of the wafer W are in an optically conjugate relationship. Reference numeral 13 denotes a mirror for optical path conversion, which guides the light beam from the correction optical system 14 to the projection lens 11.

In this embodiment, the light beam from the light source means 18 is reflected by a mirror 19, and after condensing by the illumination optical system 16, the lattice pattern G
1 is illuminated. Then, the light flux from the grating pattern G1 is made incident on the wafer mark Wa surface of the wafer W via the correction optical system 14, the mirror 13, and the projection lens 11.

At this time, since the grating pattern G1 and the surface of the wafer W are in a conjugate relationship, the image of the grating pattern G1 is formed to be superimposed on the wafer mark Wa surface -F.

FIG. 4 is a schematic diagram showing the incident optical path state of the illumination light beam to the wafer mark Wa on the wafer W surface.

As shown in the figure, a light beam 41 from the projection lens 11 (not shown) is transmitted in a plane parallel to the grating direction of the grating pattern G1 (xz
In the plane (y
(in the z-plane), it is obliquely incident at a predetermined angle θ.

The grating pattern G1 is projected onto the wafer mark Wa surface so that the grating direction is parallel to the X direction. Note that 42 schematically indicates specularly reflected light from the wafer mark Wa, and 43 schematically indicates the same phase plane of the light beam 41.

FIG. 3(A) shows the relationship between the wafer mark image Wa' and the image of the grating pattern G1 on the wafer W surface at this time. As shown in the figure, the image of the lattice pattern G1 in the area on the surface of the wafer mark Wa is generally formed to be deformed according to the undulating shape of the resist surface.

Returning to FIG. 1, the reflected light flux from the image of the grating pattern G1 projected onto the wafer mark Wa surface is reflected by the objective lens 15 via the projection lens 11 and mirror 12 onto the grating pattern 02 surface as shown in FIG. 3 (A). A lattice pattern image as shown in is formed. As shown in FIG. 2('B), the grating pattern G2 is made up of a plurality of gratings with a pitch λ2 different from the pitch λ1 of the grating pattern G1, which consists of a transparent region and a non-transparent region whose grating direction extends in the X direction. .

Reference numeral 17 denotes an imaging optical system, which forms an image of the grating pattern 02 on the imaging means 110. As a result, the wafer W surface and the imaging means 110 are connected to the optical system (projection lens 11, objective lens 15
Then, a conjugate relationship is established via the imaging optical system 17).

FIG. 3(B) shows moire fringes (scale-shaped) 33 and wafer mark image Wa' generated from the grating pattern G1 and the grating pattern G2 formed on the surface of the imaging means 110 at this time. As shown in the figure, moiré fringes 33 that are deformed in accordance with the difference in surface change (undulation shape) between points P1 and P2 on the resist surface on the wafer mark image Wa' surface are formed on the surface of the imaging means 110. It is formed.

In this embodiment, in order to obtain a focus signal on the surface of the image pickup means 110, the grating patterns G1, G2 are connected to each other λ1. λ2 is determined in consideration of optical magnification.

Here, moire fringes using moire topography will be briefly explained. For simplicity, let us assume that the magnification of the observation optical system is 1. A projected image of the lattice pattern in the vicinity P1 of the wafer mark Wa is simply expressed as . Here ρ. is the initial position, and ρ(x, y) is a modulation component, which is a function depending on the undulations of the resist. At this time, the projected image of the lattice pattern in the vicinity P2 of the wafer mark Wa can be written as. ρ0' is the initial position.

At this time, the width of the two waves is 1 (x)ρ. =ρo'=0 and note that λ1 cape λ2, ρ(x, y) are relaxed functions. Further, in general, the optical resolution on the surface of the imaging means 110 and the pixel resolution of the imaging means 110 are different.

It is a high frequency wave and is optically designed to be optically decomposed, but it does not need to be decomposed in the imaging means.

This embodiment uses such an imaging means. Therefore, its strength is proportional to . Furthermore, when one pixel is not coarse compared to the pitch λ1, a filter may be applied to drop components around the pitch of λ1. Here, what is the fringe (λ1)2/(
This refers to the component around the pitch of λ1-λ2). however,
This depends on the difference between the known λ1 and λ2 and the modulation component. The modulation component depends on the difference in change in resist undulation.

The moire image formed on the imaging means 110 is transferred to the A/D converter 1
11 and sent to an image analysis device 112. The image analysis device 112 sets a processing window 38 and integrates the data in the y-axis direction. FIG. 3(C) is an explanatory diagram of the integrated data at this time.

In this figure, positioning is performed by conventionally known one-dimensional pattern matching in order to accurately provide the reference X coordinate. The position is assumed to be Xo as a temporary center. In this embodiment, after alignment, the undulation shape of the resist surface near the wafer mark Wa is determined by the following method.

That is, the temporary center position X of the wafer mark Wa. Processing windows 34, 35, 36, and 37 are set for the surroundings, and integration is performed in the X-axis direction. At this time, a high frequency cut filter may be applied in the y-axis direction. FIG. 3(D) is an explanatory diagram of the integrated data at this time.

FIG. 5 (D), (C), (B), and (A) are each processing window 3.
4, 35, 36, and 37.

Here, the processing windows 34, 35, and 36 are near the wafer mark Wa, and the processing window 37 is an area where the grating pattern is considered to be approximately stable in the X direction at a location sufficiently layered from the wafer mark Wa. The waveforms obtained in this way (Fig. 5 (A) to (
D) is subjected to image processing, such as fast Fourier transform, to determine the phase change.

5 for the phase of FIG. 5(A) corresponding to the processing window 37.
The change in phase [0°2π] in FIGS. (B), (C), and (D) is the change in the undulation shape of the resist surface in the X direction.

In reality, the processing window is divided finely and the relationship between the phase and the X direction is investigated. The results at this time are shown in FIG. However, there is no data in the area corresponding to area 39 in FIG. 3(C).

Next, a method for correcting the temporary center position Xo of the wafer mark Wa obtained from FIG. 3(C) based on the data obtained from FIG. 6 will be described.

As mentioned above in Fig. 7, sagging occurs due to the curvature and differential absolute value of the resist surface on the XZ plane, so the measured values obtained are interpolated with a spline function as shown in Fig. 6, and the curve V (x) (takes the value of [0°2π]).

Then, with respect to that value, there is a quadratic effect in the area @60 in FIG. Perform function approximation, and let g be the quadratic coefficient at this time.

At this time, g/I g I indicates the direction of Tamasare. In other words, look at the curvature. Also, take the differential absolute value of V (x) and set it as d (x). The value of d(x) is as shown in the curve shown in FIG. 6(C). d appropriately in (x)
(x) above and below (e.g. 80 for the maximum and minimum in the area
% or more and 20% or less), calculate the center of gravity, and d
o, then dx=α・g/I g I ・do −・”(+
＞ is the amount of tamasare. At this time, x 1 = xo-dx '' (2) is the center position of the wafer mark Wa without any damage. Here, xl is converted into the positional relationship with the reticle, and the control bag W113 is
Then, drive the xyz stage 1° to complete the alignment. After that, exposure is performed. However, α is a parameter determined experimentally, and the tentative center position x0 is a center value using the primary pattern matching described in the conventional example using <C> in FIG. The qualitative image of α is as follows. In Fig. 7, the incident light is
perpendicular to the plane). When the angle of resist coating unevenness θ<<1 (rad), xl θ# tan θ”'x d(x)・tan θ・
−−−−−(3) Corresponds to 4 π. θ is the angle of incidence in FIG. λ1 is the value of the effective pitch λ1 of the grating Gl on the scale of eight sea urchins (taking into account the multiplication factor).

The amount of Tamasare in Figure 7 is approximately 1 koku (1--) ・dx −-XΔ, −tanθ −-
--(4) n2π. Here, n is the refractive index of the resist.

(n clause 1.68) Δ is the resist thickness given in FIG. Since it is calculated like this for each point] After 6 clauses (1--)・Δ,・tanθ□ ・・・・(
5) It becomes n2π. However, since it actually involves many parameters such as the effect of the step thickness Δ2 of the wafer and the NA of the observation optical system (objective lens 15 and imaging optical system), it is more practical to determine it by experiment. The determining method is to determine dx based on optical information from the wafer mark after resist coating and from the previous wafer mark, and to calculate it from the relationship of equation (1) using the above do after resist coating.

In general, if the alignment marks are similar in shape and the devices are the same, the following relationship (6) will apply. Therefore, if α is determined under certain conditions, other resist thicknesses Δ1. The case where the step thickness to the sea urchin is Δ2 can also be easily calculated.

In this manner, in this embodiment, the undulations of the resist surface near the surface of the alignment mark provided on the wafer are determined, and alignment measurements are corrected based on the obtained results to achieve highly accurate alignment.

FIG. 12 shows a flowchart regarding alignment in this embodiment. In the embodiment shown in FIG. 1, the light beam for positioning is passed through the projection lens, but it may be performed without going through the projection lens.

In addition, this example shows an example using the moiré topography method, or uses other interferometry methods to determine the y of the wafer mark.
The alignment measurement value may be corrected by forming interference fringes substantially parallel to the X direction on the direction side, and determining the undulation shape of the resist surface from the distribution of the interference fringes.

For example, coherent light is divided into reference light and test light, and as shown in Figure 4 of this embodiment, the test light is made incident and then interfered with the reference light, and the resist is applied based on the change in the phase of the interference fringe due to the optical path difference of the light. You may also measure the unevenness.

Further, in this embodiment, the wafer position detection and the resist coating unevenness detection were performed simultaneously, but even if they are separated in time, their effectiveness is not affected. Further, coating unevenness detection and wafer position detection may be performed using different optical systems.

(Effects of the Invention) According to the present invention, unnecessary noise light is eliminated by utilizing the above-mentioned elements, especially two grating patterns with different pitches. Can be removed or
A positioning apparatus capable of highly accurate positioning of the wafer and the main body of the exposure apparatus can be achieved.

[Brief explanation of drawings]

Figure 1 is a schematic diagram of the main parts of an embodiment of the present invention, Figure 2 (A)
, (B) is a schematic diagram of a part of FIG. 1, FIG. 3 is an explanatory diagram of a wafer mark image according to the present invention, FIG. 4 is an explanatory diagram of the optical path of the light beam incident on the wafer in FIG. 1, and FIG. FIG. 6 is an explanatory diagram of integrated data of a moiré image using a lattice pattern according to the present invention, and FIG. 6 is an explanatory diagram of correction of the center position of a wafer mark according to the present invention.
FIG. 7 is an explanatory diagram of the optical path of the incident light beam on the wafer in FIG. 1, FIG. 8 is a schematic diagram of the main parts of a conventional alignment device, and FIG.
The figure is an explanatory diagram of signal processing using a conventional template. FIG. 11 is an explanatory diagram of integrated signals and signals related to templates, and FIG. 12 is a flow chart diagram according to the present invention. In the figure, R is a reticle, W is a wafer, Wa is a wafer mark, G1 and G2 are grating patterns, 11 is a projection lens,
12.13 is a mirror, 14 is a correction optical system, 15 is an objective lens, 16 is an illumination optical system, 17 is an imaging optical system, 18 is a light source means, 114 is an illumination system, 110 is an imaging device, 70 is a resist, and 70a is a The resist surface, 71 is a wafer, and 73 is an air layer. Patent applicant Canon Co., Ltd. 7wo

Claims (1)

[Claims] (1) When arranging a wafer on which a relief-like alignment mark is formed and an imaging means in a conjugate relationship via an optical system, and performing relative positioning of the two, the wafer is Two grating patterns G1 and G2 coated with a resist and having different pitches are arranged in a conjugate relationship with the wafer through the optical system so that their grating directions are in the same direction as the alignment direction. The grating pattern G1 is illuminated by an illumination system, and the light beam from the grating pattern G1 is transmitted through the optical system onto the alignment mark surface of the wafer surface on a plane parallel to the grating direction of the grating pattern with respect to the wafer surface. An image of the grating pattern G1 is formed by vertically entering the grating pattern G1 and obliquely incident at a predetermined angle in a plane perpendicular to the grating direction of the grating pattern.
is projected onto the lattice pattern G2 surface through the optical system, the lattice pattern G2 surface is imaged on an imaging device, and a calculation means uses a signal from the imaging device to form an alignment mark on the wafer surface. A positioning apparatus characterized in that the surface shape of a resist in a region is determined, and relative positioning of the wafer and the imaging means is performed by referring to a signal from the calculation means. (2) The calculation means is the lattice pattern G.
2. The alignment apparatus according to claim 1, wherein a local three-dimensional shape near the alignment mark on the wafer surface is determined based on striped images formed on two surfaces.