JPH0669022B2 - Position detector - Google Patents

Description

Description: [Industrial field of use] The present invention relates to a position detection device for determining the position of an object.
This position detecting device is used for obtaining the positional relationship between the reticle and the wafer in, for example, a semiconductor printing apparatus, but is also suitable especially for signals with a lot of noise and distortion.

[Prior Art] Conventionally, there are an optical aligner and an X-ray aligner in a semiconductor printing apparatus. In these apparatuses, a mask and a wafer are aligned before printing.

In this alignment, an alignment mark, which is provided on the wafer and the mask and is capable of knowing the positional relationship between the mask and the wafer, is detected by light or an electron beam to obtain the positional relationship between the mask and the wafer.

The alignment mark on the wafer has a step structure formed on the wafer substrate, and at the time of alignment, a resist is applied thereon in a thickness of submicron to several microns. Therefore, the alignment mark on the wafer is detected through the resist layer.

[Problems to be Solved by the Invention] When detecting through a resist layer in this way, light (especially laser light) is reflected or refracted on the resist surface, is multiple-reflected on the resist layer, and is Light reflected from the back surface interferes. Further, the electron beam is scattered by the resist layer.
Due to these effects (hereinafter referred to as resist effect), the S / N of the detection signal, the contrast and the resolution are lowered.

Therefore, there is a problem that the accuracy of detecting the positional relationship between the mask and the wafer is lowered, and thus the positioning accuracy is lowered.

In order to solve this problem, the present inventors first performed a partial correlation process between a reference signal and a detection signal prepared in advance, and then detected the peak position of the signal obtained by the partial correlation process. A position detecting device for detecting is devised. However, this device has the drawbacks that (1) it takes a long time to perform the partial correlation processing, and (2) it is difficult to detect the peak position in a signal with large distortion.

Hereinafter, based on specific examples, a detailed description will be given of the alignment device according to the conventional example and the above-described proposal.

FIG. 6 is a diagram showing an example of a mark and a detection signal detected by the alignment apparatus according to the conventional example, the previous proposal, and the present invention. In the figure, 1 to 4 are alignment marks on the mask, 5 is an alignment mark on the wafer, and 6
Is a scanning line of a laser beam or an electron beam, 7 to 10 are detection signals generated from intersections of the marks 2, 5 and 4 and the scanning line 6, and 11 is a slice / binarization of the signals 7 to 10.
It is a level.

The procedure for obtaining the amount of deviation between the mask alignment mark and the wafer alignment mark by the conventional method will be described below with reference to FIG.

When detecting the positional relationship between the mask and the wafer, first, the alignment marks 1 to 4 on the mask and the alignment mark 5 on the wafer are scanned by laser light or an electron beam in two directions orthogonal to each other, and The detector detects the scattered light, reflected electrons, or secondary electrons at the intersection. This allows signals 7-10 for each scanning direction.
An electric pulse signal as shown in is obtained. Since the processing is exactly the same for each scanning direction, the case of scanning in the illustrated direction will be described here. In order to find the distance between the marks 2 and 4 on the mask and the mark 5 on the wafer, it is necessary to find the center position of each of the signals 7 to 10. A binarized rectangular signal is obtained, and the center position of each rectangular signal is obtained as the center position of each signal.

The distance between the center positions of the signal 7 from the mark 2 on the mask and the signal 8 from the mark 5 on the wafer is t1, and the distance between the center positions of the signals 9 and 10 is t2. If the mask is aligned at the center of the mark position, the shift amount d can be calculated by d = (t1-t2) / 2.

In order to improve the alignment accuracy in this conventional example,
It is necessary to improve the detection accuracy of the shift amount, in other words, the distance between the signals, in other words, the center position of each signal.
Then, a necessary condition for improving the detection accuracy of the center position of each signal is that the waveform of each signal is symmetrical about the peak position. Usually, the line width and spot diameter of the alignment mark are set so as to satisfy this condition.

However, as described above, at the time of alignment, a resist is applied on the alignment mark on the wafer in a thickness of submicron to several microns. For this reason, the alignment marks will be detected through the resist,
The signal is distorted due to the resist effect and the nonuniformity of the resist film thickness, and the symmetry of the signal waveform is broken. That is, as shown in FIG. 7, the distortion of the signal causes a detection error of the center position, which deteriorates the alignment accuracy.

For this reason, in the alignment device according to the above proposal, the alignment accuracy is improved by performing partial correlation between the position-detected signal and the reference signal.

FIG. 8 shows the structure of a positioning device to which the previously proposed position detecting device is applied.

In the figure, 21 is a mask, 22 is a wafer, 23 and 24 are optical systems of light or electron beams for detecting signals from alignment marks on the mask and wafer, 25 is the X direction, Y
And a stage for holding the wafer 22, which can be moved in the rotational and rotational directions, 26, 27, and 28 are the stages 25, respectively.
Device for moving in the direction of rotation, Y direction and rotation direction, 29
Is a signal detection device, 30 is a signal processing device, 31 is a storage device, 32
Is a position detecting device, 33 is a cpu, and 34 is a control device for controlling the driving devices 26, 27 and 28.

FIG. 9 is a flowchart of the process of the apparatus shown in FIG.

In the figure, step 41 is signal detection processing, and step 42
Is a partial correlation process, step 43 is a peak position detection process of each signal, step 44 is a process of detecting the positional relationship between the mask and the wafer, and step 45 is a process of matching the mask and the wafer with a predetermined positional relationship.

The operation of the apparatus shown in FIG. 8 will be described in detail below with reference to the flow chart of the processing shown in FIG.

First, in the signal detection process (step 41), the signal detection device
29, two sets of optical systems 23 and 24 are used to perform scanning in two directions orthogonal to each other to detect signals from the alignment mark of the mask and the wafer. Since two positions are scanned in two directions, if the alignment marks shown in FIG. 6 are used, four sets of time series signals as shown by signals 7 to 10 in FIG. 6 can be obtained. In addition, noise in the high frequency region included in the signal is removed by a low pass filter.

Next, in the partial correlation processing (step 42), the output signal from the signal detection device 29 detected in the signal detection processing (step 41) and the reference signal stored in the storage device 31 are partially correlated in the signal processing device 30. To process.

FIG. 10 shows the state of the correlation processing at this time. In the figure, 50 to 53 are a set of signals among the signals subjected to the signal detection processing, and correspond to the signals 7 to 10 in FIG. Reference numeral 54 is a reference signal, and reference numerals 55 to 58 are output signals obtained by performing partial correlation processing on the signals 50 to 53 and the reference signal 54, respectively.

With respect to this output signal, in the peak position detection processing of each signal (step 43), the position detection device 32 detects the peak position of each of the signals 55 to 58, that is, the center position of the signals 50 to 53. Then, in the mask and wafer position detection processing (step 44), the signal 55, the signal 56, the signal 57, and the signal 58 are obtained from the peak position of each signal thus obtained by the cpu 33.
The respective distances t1 and t2 are calculated. At this time, if the alignment mark 5 shown in FIG. 6 is used and predetermined alignment is performed when the alignment mark 5 is at the center of the alignment marks 1 to 4, the amount of deviation between the mask and the wafer is dt = (t1− t2) / 2.

The above processing is performed on four sets of time-series signals obtained by scanning two alignment marks in two directions orthogonal to each other, and the shift amount dXr in the X and Y directions from the right alignment mark. And dYr are calculated, and the deviation amount dXl in the X and Y directions from the left alignment mark
And dYl.

From these results, the deviation amounts in the X direction, the Y direction, and the rotation direction are: dX = (dXr + dXl) / 2 (1) dY = (dYr + dYl) / 2 (2) dR = (dYr + dYl) / D (3) where , DR are shift amounts in the rotational direction, and D is obtained from the distance between the alignment marks.

In the alignment process (step 45), the stage 25 is moved by the drive devices 26, 27 and 28 under the control of the control device 34 so as to eliminate the deviation amount of the expression (1) to the expression (3), and the mask is removed. And the wafer alignment is completed.

However, according to the method using such partial correlation processing, although the position detection accuracy is improved, the partial correlation processing takes time as described above, and the partial correlation processing is performed for a large amount of noise or distortion. There is a drawback that it is difficult to detect the peak position from the correlation-processed signal.

An object of the present invention is to further solve these drawbacks in a position detecting device using partial correlation processing.

[Means and Action for Solving Problems] In order to solve the above problems, the present invention first detects the position of each detected pulse signal with relatively coarse accuracy, and then
The position of each pulse signal is detected with high accuracy by performing a partial correlation process between each pulse signal and the reference signal in the vicinity of each position.

Embodiments Embodiments of the present invention will be described below with reference to the drawings.

First, the principle of this embodiment will be described.

Conventionally, the alignment is performed in two steps, that is, coarse alignment with an accuracy of about several microns and fine alignment with an accuracy of about 0.1 micron. This embodiment is for fine alignment. At this stage, the order of signals detected from the alignment marks is fixed. If the alignment marks shown in FIG. 6 are used, the signal 7 shown in FIG. Signal 8 and signal 9 are the wafer, and signal 10 is the signal from the mask. The distance between the signal 7 and the signal 8 is between a predetermined value and an error range of rough alignment accuracy. That is, the distance between the signal 7 and the signal 8 is approximately fixed.

Signal 7 and signal 10 and signal 8 and signal 9 are a pair of signals having a predetermined distance interval, that is, signal 7 and signal
Each of 10 and signals 8 and 9 can also be regarded as one signal.

From the above, it is possible to perform specific signal processing on a specific signal of signals 7 to 10, or on a signal paired with signal 7 and signal 10 and signal 8 and signal 9.

Conventionally, correlation processing is performed for two functions f (x) and g (x). However, for example, a part of the function f (x) and a part of the function g (x) that perform correlation processing and a part of the function f (x) that performs correlation processing on a specific part are called partial correlation processing.

FIG. 2 is a diagram showing an example of such correlation processing. In the figure, 80 is a reference signal, 81 is a detection signal, 82 is a correlation function between the reference signal 80 and the detection signal 81, 83 is a distorted detection signal, 84 is a correlation function between the reference signal 80 and the detection signal 83, Reference numeral 85 is a distorted detection signal, and 86 is a correlation function between the reference signal 80 and the detection signal 85.

As shown in the figure, the center position of the detection signal is obtained from the position of the peak of the correlation function, but when the distortion is large, the correlation function is also distorted, and the peak position detection processing becomes complicated and the processing takes time.

In this embodiment, rough position detection is performed before performing the correlation processing, and partial correlation processing is then performed around the obtained position to detect the peak position of the obtained correlation function. As a result, the processing of equations (1) and (4) can be performed within the range [x ₀ + a, x ₀ -a] (x ₀ is the position obtained by rough position detection), and therefore the time required for the correlation processing can be shortened. .

(2) Since the peak position of the correlation function in the narrow area is obtained, the processing is simple and the processing time can be shortened. (3) The position detection accuracy can be improved by obtaining the correlation function with a fine accuracy. , Solves the drawbacks of the previous and previous proposals.

Next, this embodiment will be specifically described.

FIG. 1 shows the arrangement of an alignment apparatus according to an embodiment of the present invention. In the figure, 21 is a mask, 22 is a wafer, 23 and 24 are optical systems of light or electron beams for detecting signals from the alignment marks on the mask 21 and the wafer 22, 25
Is a stage that holds the wafer 22 that can be moved in the X, Y, and rotational directions, 26, 27, and 28 are drive units that move the stage 25 in the X, Y, and rotational directions, and 29 is signal detection. Reference numeral 31 is a storage device, 32 is a position detecting device, 33 is a cpu, 34 is a device for controlling the driving devices 26, 27 and 28, 100 is a coarse position detecting device, and 101 is a signal processing device.

Here, the difference from the configuration of the alignment apparatus according to the previous proposal in FIG. 8 is that the coarse position detection apparatus 100 is added in the present embodiment, and the signal processing apparatus 101 for the signal processing apparatus to obtain the correlation function of the local region. Has changed to.

FIG. 3 is a flow chart of processing of the apparatus of FIG. In the figure, step 110 is signal detection processing, step 111 is coarse position detection processing, step 112 is partial correlation processing, step 1
13 is a peak position detection process for each signal, step 114 is a process for detecting the positional relationship between the mask and the wafer, and step 115 is a process for aligning the mask and the wafer with a predetermined positional relationship.

Below, a detailed description will be given of the parts different from the alignment device according to the above proposal.

If the alignment mark shown in FIG. 6 is used, the signal obtained in the signal detection process (step 110) will be as shown in FIG. In the figure, signals 120 to 123 are detected from the alignment marks 2, 5 and 4 on the mask and wafer, respectively.

In the coarse position detecting device 100, the center position of each signal of the signals 120 to 123 when the signal is binarized into a rectangular signal waveform is obtained (step 111). Let these be T1, T2, T3 and T4. When there is noise or distortion in the signal, the precision of these values is about 0.1 to 1.0 micron, which cannot be said to be sufficient. That is, the coarse position detection is performed.

In the signal processing device 101, the reference signal 124 and each signal 120, signal 1
21, partial correlation processing with the signals 122 and 123 is performed (step 112). Here, in the above proposal, the signals 120 to 123 are the function f (x) and the reference signal 124 is g.
If (x), The following calculation was obtained within the range of T _n >x> T _s , but in the present embodiment, Here, f ₁ , f ₂ , f ₃ and f ₄ represent the signals 120 to 123, h ₁ , h ₂ , h ₃ and h ₄ represent the signals 125 to 128, and a is a value about the width of each signal. Are calculated as T ₁ + a>x> T ₁ −a and T ₂ + a, respectively.
_{>X> T 2 -a, T} 3 + a>x> T 3 -a and _T 4 +
It is determined within the range of a>x> T ₄ −a.

Since this process has a narrow integration range and a narrow range taken by the variable x as compared with the previously proposed process, the time required for the process can be greatly reduced. Further, the detection accuracy can be improved by narrowing the sampling interval of the variable x.

[Modification of Embodiment] In the embodiment described above, as shown by the reference signal 124 in FIG. 4, the correlation processing with the detection signals 120 to 123 is performed using one reference signal. Reference signals 134 and 1 as shown
35 and reference signals 136 and 137 are used, and partial correlation processing is performed between the reference signal composed of the signals 134 and 135 and the signal composed of the detection signals 130 and 133, and the reference signal composed of the signals 136 and 137 and the detection signal. By performing the partial correlation processing with the signal composed of 131 and 132, the detection accuracy can be further increased.

In the rough position detection processing at this time, the center positions of the signals 130 to 133 are obtained by the same method as in the above-described embodiment, and the center position T ₅ of the signals 130 and 133 and the center position T ₆ of the signals 131 and 132 are determined. Ask.

In the partial correlation process (step 112) of FIG. 3, Here, g ₁ is the signal 134 and the signal 135, g ₂ is the signal 136 and the signal 137, h ₅ is the signal 138, h ₆ is the signal 139, and f is the signal 130-.
Computes representing 133.

From the distance t3 between the signal 138 and the signal 139 thus obtained, the shift amount is dt = t3.

[Effects of the Invention] As described above, according to the present invention, the following effects can be obtained.

(1) It is possible to detect a position by a signal with a poor S / N, for example, at the time of alignment in the process of baking aluminum wiring.

(2) Reduce the influence of distortion of the detected signal (3) Improve the detection accuracy by increasing the sharpness of the signal waveform (4) Can perform signal processing at high speed (5) Peak of the processed correlation The process can be detected accurately.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a configuration diagram of an alignment apparatus according to an embodiment of the present invention, FIG. 2 is a diagram showing an example of signal processing, and FIG. 3 is an apparatus of FIG. FIG. 4 and FIG. 5 are diagrams showing another example of signal processing, FIG. 6 is a diagram showing an example of an alignment pattern and its detection signal, and FIG. 7 is a waveform. FIG. 8 is a diagram showing an example of distortion and detection error, FIG. 8 is a configuration diagram of the alignment device according to the above proposal, FIG. 9 is a flow chart for explaining alignment processing of the device of FIG. 8, and FIG. It is a figure which shows an example of a signal processing. 21: mask, 22: wafer, 23, 24: optical system, 25: stage, 26, 27, 28: stage drive device, 29: signal detection device, 30: signal processing device, 31: storage device, 32: position detection Device, 33: cpu, 34: control device, 100: coarse position detection device, 101: signal processing device.

Claims (2)

[Claims] 1. A means for generating a reference signal, a signal detecting means for detecting a signal from an alignment mark on an object as an electric pulse signal train, and the position of each pulse signal in the pulse signal train is relatively rough. Partial correlation processing between the coarse detection means for detecting with accuracy and the reference signal generated by the reference signal generation means in the vicinity of the position of each pulse signal detected by the coarse detection means and each pulse signal detected by the signal detection means. By so doing, it is provided with a fine detection means for detecting the position of each pulse signal with high accuracy, and a position detection means for obtaining the object position based on the position of each pulse signal detected by the fine detection means. Position detection device. 2. The position detecting device according to claim 1, wherein the object is a mask or a wafer.