JP2867065B2 - Position detection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to a pattern alignment method, for example, a semiconductor IC or
The present invention relates to a position detection method suitable for alignment for detecting a relative position between a wafer and a mask and controlling the position in an exposure apparatus for manufacturing an LSI.

[Prior Art] Recently, the integration density of semiconductor ICs and LSIs has been increasing more and more, and the dimensions of fine patterns of devices have been on the order of submicrons.

For example, Japanese Patent Application Laid-Open
There is a pattern matching detection method disclosed in JP-A-62-232504.

FIG. 4 is a schematic diagram of an apparatus using a two-dimensional pattern matching detection method. In the figure, the illumination light is a projection optical system.
12 illuminates the wafer pattern M, and the reflected light is
Image at 15. FIG. 5 is a diagram illustrating a two-dimensional electric signal obtained by the imaging device. In FIG. 5A, the wafer pattern is M '. The two-dimensional electric signal is
After being converted into a two-dimensional discrete digital signal sequence corresponding to the XY address of the pixel by the A / D conversion device 16, it is binarized by the binarization device 41, and the binary image is stored in the device in advance by the matching device 42. Two-dimensional pattern matching is performed to obtain the correlation of the pattern with the template, and the position having the highest degree of correlation is detected with an accuracy of about one or two pixels, and this is set as the temporary center. On the other hand, the multi-valued image is pixel-integrated in a direction perpendicular to the pattern in a two-dimensional window 21 of a predetermined size by the integrator 17 to obtain a one-dimensional discrete electric signal 51 as shown in FIG. . Next, in order to achieve the required accuracy, the position detector 43 increases the resolution of the one-dimensional discrete electric signal 51 by using interpolation means for the correlation value calculated by pattern matching in the vicinity of the temporary center, and then increases the resolution. the maximum value and outputs a position x _C indicating a, by the alignment control 112 is performed alignment by controlling the position of the wafer on the basis of the x _C.

[Problem to be Solved by the Invention] However, in the above-described conventional example, the signal becomes a discrete sequence by A / D conversion for processing the electric signal, and the detected mark position also takes a discrete value. In order to achieve this, it is necessary to use some kind of interpolation means, and an error due to the approximation at that time causes a detection error. further,
The image of the mark may be distorted due to the influence of noise such as uneven coating of the resist applied thereon or uneven illumination, and this may cause a reduction in detection accuracy due to deterioration in S / N.

The present invention has been made in view of the above-described drawbacks of the related art, and has as its object to provide a positioning method capable of detecting a position with high accuracy without being affected by noise.

[Means and Actions for Solving the Problems] According to the position detection method of the present invention, one of the images of the patterns arranged at equal intervals in the position detection direction is one of the images in the position detection direction.
A method of detecting a position of a pattern with respect to a predetermined reference point on an imaging surface where the pattern is captured from a one-dimensional video signal, wherein the one-dimensional video signal is converted into a spatial frequency domain signal by orthogonal transformation. The spatial frequency strength E (k) obtained from the spatial frequency domain signal in the spatial frequency domain signal is obtained from the spatial frequency range E (k) where E (k) = [｛Re (X (k))｝ ² + ｛Im ( X
(K)) { ^{2 1/2} X (k): the peak position Pi = {k | Max (E (k))} of the complex Fourier coefficient of the signal in the spatial frequency domain is obtained. Phase of frequency component with respect to reference point Θ
_pi = tan ^-1 {Im (X (k)) / Re (X (k))}
From the phase Θ _pi , based on the formula Δ _pi = (1 / 2π) · (N / Pi) · λ _s · Θ _pi , where N: number of sampling points, λ _s : sampling interval Deviation from point Δ
_pi is detected.

[Embodiment] FIG. 1 shows the configuration of the whole apparatus when a position detecting mechanism according to the present invention is applied to a wafer positioning apparatus. First
In the drawing, W is a semiconductor wafer, on which a pattern M for position detection is formed. Reference numeral 10 denotes an x, y stage for moving the wafer W in the x, y, and z directions, and reference numeral 11 denotes a projection optical system for projecting a circuit pattern on the reticle R surface onto the wafer W. Reference numeral 12 denotes an illumination system. Light emitted from the illumination system passes through a half mirror 13 and a projection optical system 11 to illuminate a pattern M. The reflected light from the pattern M passes through the projection optical system 11 and is guided to the detection optical system 14 by the half mirror 13. The detection optical system 14 forms an image of the pattern M on the imaging surface of the imaging device 15 at a predetermined magnification. In this apparatus, the relative positions of the reticle R with respect to the projection optical system 11, the detection optical system 14, and the imaging device 15 are determined in advance by appropriate detection means. Therefore, in order to determine the relative position of the wafer W with respect to the reticle R, an imaging device
What is necessary is just to detect the position of the pattern M on the 15 imaging surfaces. The imaging device 15 is, for example, a photoelectric conversion device such as an ITV or a two-dimensional image sensor, and converts a captured image into a two-dimensional electric signal.

FIG. 2A shows an image of the wafer area including the position detection mark M formed on the image pickup device 15 when detecting the position in the x direction. F is an effective field of view of the imaging device 15, and M is an example of an image of a position detection pattern. When the position detection in the x direction in FIG. 2A is performed, the pattern M 'is a pattern in which a plurality of rectangular patterns having the same shape are arranged at equal intervals λp in the x direction.

The pattern image converted into a two-dimensional electric signal by the imaging device 15 is converted into a two-dimensional discrete electric signal sequence corresponding to an address in the xy direction of the imaging surface by the A / D conversion device 16 in FIG. The sampling interval λ _S of this discrete signal is
It is determined by the optical magnification of the projection optical system 11 and the detection optical system 14, and the pixel interval on the imaging surface. Reference numeral 17 in FIG. 1 designates a two-dimensional window 21 including a pattern M 'shown in FIG. 2 and performs pixel integration in the y direction in FIG. The one-dimensional discrete signal sequence S (x) in the x direction shown in b) is output. In FIG. 2 (b), each point of the discrete signal sequence S (x) is connected by a straight line for easy viewing.

Reference numeral 18 in FIG. 1 denotes an FFT operation device, and the input signal train S
(X) is converted into a signal in the spatial frequency domain by a discrete Fourier transform, and a Fourier coefficient is calculated at high speed. The method is a well-known (eg, Science and Technology Publishing Company “Fast Fourier Transform”, E, ORAN BRIGHAM, Chapter 10, Fast Fourier Transform) Fast Fourier Transform of N points (N = 2 ^r )
FFT) and the sampling frequency fs
Is normalized to 1, the complex Fourier coefficient X (k) of the spatial frequency f (k) ≡k / N is Becomes At this time, the intensity E of the spatial frequency f (k)
(K) and phase Θ (k) are E (k) = [｛Re (X (k))｝ ² + ｛Im (X (k)) ² ] ^1/2 (k) = tan ^-1 {Im (X (k)) / Re (X (k))} (where Re (X (k)) and Im (X (k)) are the real part of complex number X (K), respectively. (Representing the imaginary part). FIG. 3A shows the vicinity X _S of the center of the pattern M ′ which has been roughly determined in advance by an appropriate detecting means.
Represents the distribution of the spatial frequency intensity E (k) when the FFT is performed so as to include the entire pattern with reference to. Due to the periodicity of the pattern, the spatial frequency unique to the pattern and its harmonic f (hi) (where hi is the value) can be obtained from the design value of the pattern appearing in the one-dimensional discrete signal S (x) of the pattern.
In the case of ≡i · (λ _S / λp) · N (i = 1, 2,...), The signal intensity becomes large, and a peak occurs on the graph of FIG.

Reference numeral 19 in FIG. 1 denotes a frequency intensity detection device, and a pattern-specific spatial frequency f (hi) (i = 1, 2, 3,
...), The peak position Pi and the peak frequency f (Pi) are detected.

Pi (｛K | max (E (k), f (hi) · N−α <K <f (hi) · N + α, K = 0, 1, 2,..., N−1｝) where α Is a positive integer, and is determined so that Pi becomes significant with respect to the pattern pitch fluctuation. On the other hand, if the optical magnification fluctuates for any reason within the range of α,
By detecting the frequency f (Pi) showing this peak, it is possible to correct the fluctuation of the optical magnification from the frequency unique to the pattern. For peak detection, interpolation means (for example, least square approximation,
The frequency resolution may be increased using a weighted averaging process or the like.

Reference numeral 110 in FIG. 1 denotes a phase detector, which is a reference point of a spatial frequency intensity peak f (Pi) and a frequency component near the peak f (Pi).
The phase for X _{S is} detected according to the equation. Reference numeral 111 in FIG. 1 denotes a shift amount detection device, which is obtained from a phase based on an equation. It calculates a shift amount delta _K of the real space by. FIG. 3 (b)
Is a plot of the deviation amount delta _K for the previous spatial frequency domain. In general, as shown in FIG. 3 (b), the deviation amount delta _K of the mark natural frequency near, each calculated to show a certain stable value over each harmonic component delta
Against _K, performs a weighted average processing by the frequency intensity, is detected in accordance with weighted averaging (equation) a deviation amount delta _C from the reference point X _S pattern center.

In the equation, l may be selected to be a required harmonic order. FIG. 3C is a diagram showing the result of the weighted averaging process.

Δ in place of the weighted average processing when determining the _C, P the peak position obtained using the interpolation means Pi neighborhood determined by the formula _C
And the shift amounts Δ [P _C ] and Δ [P _C +1] as shown in the equations.
Accordingly, the deviation amount and first-order approximation, on the approximate line, P _C
The deviation amount corresponding to or as delta _C.

(However, [] is a Gaussian symbol.) In the first embodiment, the FFT is used for the calculation of the Fourier coefficient. However, in the following second embodiment, the spatial frequency unique to the pattern does not fluctuate greatly. Is used to directly calculate the Fourier coefficient of the required frequency component from the equation. Due to the direct calculation of the Fourier coefficients, the FFT has a limit of the number of sample points N = 2 ^r (r = 1, 2,...), Whereas N
Can be set freely, and considering that the frequency component by the discrete Fourier transform is expressed by K / N (K = 0, 1, 2,..., N−1), K / N is calculated. The number of sampling points N to minimize the error of the frequency component to be
Is selected, the spatial frequency unique to the pattern can be obtained more accurately. Even when the pattern pitch fluctuation is relatively large, N can be optimally set again according to the frequency indicating the peak of the detected frequency intensity, and a Fourier operation can be performed to enable highly accurate position detection.

In addition, by limiting the frequency components to be calculated, the amount of calculation can be reduced. After the calculation of the Fourier coefficients, in the same manner as in the first embodiment, it detects the shift amount delta _C pattern.

[Effects of the Invention] As described above, the present invention converts a one-dimensional video signal in the position detection direction of an image of a plurality of patterns arranged at equal intervals in the position detection direction into a signal in the spatial frequency domain by orthogonal transformation. From the phase _{pi pi} of the substantial frequency component of the peak position Pi obtained from the spatial frequency vicinity range unique to the pattern in the signal in the spatial frequency domain with respect to the reference point of the pattern, the pattern of the pattern in the real space based on the above equation is obtained. Deviation Δ from reference point
By detecting _pi , it is possible to find the position of the pattern without depending on the sampling pitch, and it is possible to remove unnecessary noise components by selectively adopting effective frequencies. , Thereby enabling highly accurate position detection. As will be readily apparent, the present invention has been obtained by some means.
It goes without saying that the present invention can also be applied to finding the spatial position of a discrete signal sequence that can be expected to have periodicity in advance.

[Brief description of the drawings]

FIG. 1 is a schematic view of a main part of an embodiment of the present invention, FIG. 2 (a) is an explanatory view on a surface of a wafer W imaged by an imaging device, FIG. 2 (b) is an explanatory view of integrated data of a pattern, FIG. 3 is an explanatory diagram of a spatial frequency domain according to the embodiment of the present invention. FIG. 3 (a) is an explanatory diagram of a frequency intensity around a peak, FIG. 3 (b) is an explanatory diagram of a shift amount, and FIG. FIG. 4 is an explanatory diagram of weighted statistical processing, FIG. 4 is a schematic diagram of a main part of an example of a conventional technology, FIG. 5 (a) is an explanatory diagram of a pattern according to the conventional technology,
FIG. 5B is an explanatory diagram of integrated data of a pattern according to the related art. W: wafer, M: pattern, R: reticle, F: imaged image, M ': imaged pattern, 10: wafer stage, 11: projection optical system, 12: illumination system, 13: half mirror, 14: Detection optical system, 15: imaging device, 16: A / D converter, 17: integrating device, 18: FFT operation device, 19: symmetry evaluation calculation and maximum value detection device, 110: alignment control device, 111: deviation Amount detection device, 112: alignment control device, 21: window, 41: binarization device, 42: matching device, 43: position detection device.

──────────────────────────────────────────────────の Continued on the front page (72) Inventor Makoto Sato 53, Imaiue-cho, Nakahara-ku, Kawasaki-shi, Kanagawa Prefecture Within the Kosugi Office (72) Inventor Shigeyuki Uzawa 53, Imaiue-cho, Nakahara-ku, Kawasaki-shi, Kanagawa Canon Inc. Kosugi Plant (56) References JP-A-63-275200 (JP, A) JP-A-63-157435 (JP, A) JP-A-62-232504 (JP, A) Jpn. U) (58) Field surveyed (Int.Cl. ⁶ , DB name) G05D 3/12

Claims (1)

(57) [Claims] 1. A reference point set in advance on an imaging plane where an image of a plurality of patterns arranged at equal intervals in a position detection direction is taken from a one-dimensional video signal in the position detection direction. A spatial frequency obtained by converting the one-dimensional video signal into a spatial frequency domain signal by orthogonal transform, and obtaining a spatial frequency obtained from the pattern-specific spatial frequency vicinity range in the spatial frequency domain signal. Intensity E (k) where E (k) = [{Re (X (k))} ² + {Im (X (k))} ² ] ^1/2 X (k): complex Fourier of spatial frequency domain signal A peak position Pi = {k | Max (E (k))} of the coefficient is obtained, and the phase {} of the substantial frequency component of the peak position Pi with respect to the reference point is obtained.
_{pi where} Θ _pi = tan ^-1 ｛Im (X (k)) / Re (X (k))｝, and from the phase Θ _pi , the equation Δpi = (1 / 2π) · (N / Pi) · λ _s · Θ _pi where N is the number of sample points, λ, is the sampling interval, and the amount of deviation Δ from the reference point of the pattern in real space
A position detection method comprising detecting _pi .