JPS6135692B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a position detection method for detecting the position of a two-dimensional pattern formed on a semiconductor wafer, a mask, or the like.

[Background of the invention]

In the conventional position detection method, for example, an etched area provided as a position alignment target on a semiconductor wafer is superimposed on a transparent window provided as a position alignment target on a mask, and the amount of light reflected from the etched area is detected from the window. The reflected light intensity method detects the position of the semiconductor wafer with respect to the mask, and the other uses the method of forming a linear pattern on the mask and the semiconductor wafer as a target for position alignment. There is an optical scanning method in which the position of a semiconductor wafer with respect to a mask is determined by optically scanning an optical image of a pattern.

The former position detection method has problems such as the characteristics of the amount of reflected light on the smooth surface of the semiconductor wafer or the etched portion, and the non-uniformity of the amount of light due to uneven illumination.

In addition, in the latter position detection method, the scanning speed of the optical image by the slit and the mirror is uneven, so it is difficult to accurately detect the position of the semiconductor wafer with respect to the mask. Since this method has to be mechanically driven, there is a problem in that the reliability of the mechanical drive section is poor.

[Purpose of the invention]

An object of the present invention is to solve the above-mentioned problems, eliminate moving parts, and optically emphasize a linear or dot array pattern oriented perpendicular to the arrangement direction of one-dimensional solid-state image sensors with high reliability. An object of the present invention is to provide a position detection method that can detect the position of a linear or dotted pattern of levers with high sensitivity.

[Summary of the invention]

That is, in order to achieve the above object, the present invention compresses a two-dimensional pattern having a linear or dot-like pattern in the linear direction of the pattern using an optical system including a cylindrical lens to produce a one-dimensional pattern. The image is formed on a one-dimensional solid-state image sensor in which elements are arranged, and this compressed one-dimensional pattern is converted into a video signal by the one-dimensional solid-state image sensor, and the center of the mountain of the converted video signal is Alternatively, the present invention is a position detection method characterized by detecting the position of a linear or dot array pattern based on the address of the one-dimensional solid-state image pickup device that indicates a peak.

[Embodiments of the invention]

The present invention will be specifically described below based on embodiments shown in the drawings. FIG. 1 shows a mask, and FIG. 2 shows a semiconductor wafer.
The mask 1 shown in the figure includes a circuit pattern 2 to be exposed and printed on a predetermined position 4 of a semiconductor wafer 6, and two first identification patterns for positional alignment at both ends for positional alignment with the semiconductor wafer 6. Target patterns 3a and 3b are formed. On the other hand, on the semiconductor wafer 6, two position matching target patterns 5a and 5b, which are second identification patterns, are formed at positions corresponding to the position matching target patterns 3a and 3b. The target patterns 3a and 3b for positional alignment of the mask are as follows:
For example, as shown in FIG. 1, there is a device in which L-shaped opaque lines are arranged in a cross shape in a transparent area to form a transparent band having a constant width in the shape of a cross. This transparent band is an identification area, and one end of each band is the starting end and the other end is the ending end. Further, as the target patterns 5a and 5b for positional alignment of a semiconductor wafer, for example, as shown in FIG. 2, there is a pattern formed of a cross-like thin line or a linear pattern having a different reflectance from the surrounding area. Next, the mask position alignment device of the present invention will be explained based on FIGS. 3 to 6. 7 is a rotary table on which a semiconductor wafer 6 is placed. 8
is an X-axis moving table, rotatably supports a rotary table 7, and a motor 9 for rotating the rotary table 7 is attached. Reference numeral 10 designates a Y-axis moving table, on which an X-axis moving table 8 is placed slidably in the X-axis direction, and an X-axis motor 11 is attached to move the X-axis moving table 8 in the X-axis direction. Furthermore, the Y-axis drive table 10 is mounted on a base (not shown).
It is mounted slidably in the axial direction and connected to a Y-axis motor 12 that moves in the Y-axis direction. On the other hand, the mask 1 is placed at a small distance from the semiconductor wafer 6 and is attached to a holding member (not shown) attached to a base (not shown). 13a
and 13b are optical systems provided opposite to the alignment targets 3a and 3b, respectively, and are constructed as shown in FIG. That is, the optical systems 13a and 13b include an objective lens 14 and a semi-transparent mirror 1 for enlarging the optical image of the superimposed target pattern for positional alignment.
5. A condenser lens 16, a light source 17, a semi-transparent mirror 18, and a cylindrical lens 20 whose axial center is arranged in the vertical direction and reduces the width of the light image reflected by the semi-transparent mirror 18 in the X-axis direction. and a cylindrical lens 21 whose axial center direction is arranged in the X-axis direction and reduces the width of the optical image passing through the semi-transparent mirror 18 in the Y-axis direction. That is, the semi-transparent mirror 18 and the cylindrical lens 20 directly project the Y component of the optical image magnified by the objective lens 14 in the vertical direction onto the photodiode arrays 22a, 22b, and reduce the X component to the photodiode arrays 22a, 22b. It is arranged to project within the width of the Further, the semi-transparent mirror 18 and the cylindrical lens 21 project the X component of the enlarged optical image as it is in the X direction of the photodiode arrays 23a, 23b, and reduce the Y component and project it within the width of the photodiode arrays 23a, 23b. It is arranged like this. 22a and 22b are Y-axis photodiode arrays in which a large number of photodiode elements are arranged in a line, and the longitudinal direction is arranged in the axial direction of the cylindrical lens 20.
It receives an optical image that has passed through zero, and outputs an electrical signal in series from each photodiode element in accordance with a scan signal. 23a and 23b are X-axis photodiode arrays in which a large number of photodiode elements are arranged in a line, and the longitudinal direction is arranged in the axial direction of the cylindrical lens 21, and receives the light image that has passed through the cylindrical lens 21 and converts it into a scan signal. Accordingly, an electrical signal is output in series from each photodiode element. 2
Reference numeral 4 designates a scan signal generating circuit, and output scan signals 25a and 25b are connected to photodiode arrays 22a and 22b and photodiode arrays 23a and 23b. Reference numeral 28a denotes a processing circuit, which uses a signal 26a output from the photodiode array 22a to detect the alignment target 3a on the mask 1 and the alignment target 5a on the semiconductor wafer 6 in the Y-axis direction, as shown in FIG. This circuit calculates the deviation Δy ₁ from the reference distance Δ ₂ . 28a is a processing circuit which uses a signal 26b outputted from the photodiode array 22b to calculate the deviation Δy 2 between the position matching target 3b and the position matching target 5b with respect to the reference distance Δ ₂ in the X-axis direction, as shown in _FIG . This is the circuit you are looking for. 29a is a processing circuit which uses a signal 27a outputted from the photodiode array 23a to detect the position matching target 3a and the position matching target 5a as shown in FIG.
This circuit calculates the deviation Δx from the reference distance _Δ1 in the X-axis direction. 29b is a processing circuit and a signal 27 outputted from the photodiode array 23b.
At b, position alignment target 3 is placed as shown in FIG.
This circuit calculates the deviation Δx ₂ from the reference distance Δ ₁ in the X-axis direction between the alignment target 5b and the position matching target 5b. Reference numeral 30 denotes an arithmetic circuit which receives a signal of Δy ₁ output from the processing circuit 28a, a signal of Δy ₂ output from the processing circuit 28b, a signal of Δx 1 output from the processing circuit 29a, and a signal of Δy ₁ output from the processing circuit 29b. Performs arithmetic processing based on the output Δx ₂ signal,
The relative displacement amount Δx of the mask 1 and the semiconductor wafer 6 in the X-axis direction, the relative displacement position Δy of the mask 1 and the semiconductor wafer 6 in the Y-axis direction, and the relative displacement amount Δy of the mask 1 and the semiconductor wafer 6 in the Y-axis direction. This is a circuit for determining the relative displacement amount Δθ in the rotational direction of the rotation direction of the rotation direction. 31
is a drive circuit, which drives the X-axis motor 11 until the Δx signal output from the arithmetic circuit 30 becomes zero, and drives the Y-axis motor 12 until the Δy signal output from the arithmetic circuit 30 becomes zero. and arithmetic circuit 3
The motor 9 is driven until the signal of Δθ outputted from zero becomes zero.

With the above configuration, when the semiconductor wafer 6 is first placed on the rotary table 6 and the mask 1 is placed on a holding member (not shown), the target patterns 5a and 5b for positional alignment of the semiconductor wafer 6 and the target for positional alignment of the mask 1 are aligned. Patterns 3a and 3b are shown in Figure 5a,
They are arranged in an overlapping manner as shown in b. Next, light source 1
When 7 is turned on, the light emitted from the light source 17 is
The target pattern 3 for position alignment is converted into parallel light by the condenser lens 16, reflected by the semi-transparent mirror 15, and superimposed through the objective lens 14.
A, 5a and 3b, 5b are irradiated. The light reflected from the position alignment target pattern 5a passes through the transparent area of the position alignment target pattern 3a and is formed as an optical image. This optical image is magnified by the objective lens 14, passes through the semi-transparent mirror 15, and reaches the semi-transparent mirror 18. The light image reflected by the semi-transparent part 18 of this light image is reflected by the cylindrical lens 2
0, the light image is reduced in the X-axis direction from WX to the width LX of the photodiode array 22a and formed on the light receiving surface of the photodiode array 22a, as shown in FIG. 21, move WY in the Y-axis direction as shown in Figure 6.
The image is then reduced to the width LY of the photodiode array 23a and formed on the light receiving surface of the photodiode array 23a. On the other hand, the light reflected from the position matching target pattern 5b passes through the transparent area from the position matching target pattern 3b and is formed as an optical image. This light image is transmitted to the semi-transparent mirror 1 as before.
8, the light image reflected by the semi-transparent mirror 18 is formed on the light receiving surface of the photodiode array 22b, and the light image passing through the semi-transparent mirror 18 is reflected by the diode array 23.
The image is formed on the light receiving surface b. That is, the photodiode elements of the photodiode arrays 22a, 23a and 22b, 23b are formed by superimposed alignment patterns 5a, 3a and 5b, 3, respectively, as shown in FIG.
The optical image b is magnified by the objective lens 14 and is divided into picture elements, that is, detection regions, which are divided into photodiode elements at regular intervals in the X-axis direction, and is further divided into picture elements (which are divided into photodiode elements at regular intervals in the Y-axis direction). street address)
That is, the detection area is imaged. Next, the photodiode arrays 22a and 22b are each scanned in the scan direction shown in FIG. 6 by a scan signal output from the scan signal generation circuit 24, and a signal is output from each photodiode element (address) as shown in FIG. Signals 26a and 26b in the form of a series of are output. The processing circuits 28a and 28b perform pit processing on the signals 26a and 26b, respectively, and calculate the distances ya ₁ and ya ₂ between the address indicating the first peak value and the address indicating the second peak value, that is, the time tya. ₁ , tya ₂ and the distance yb ₁ , yb ₂ between the third peak value indicating the second peak value, that is, the time tyb ₁ , tyb ₂ ,
Determine the deviations Δy ₁ and Δy ₂ from equations (1) and (2) shown below.

Δy ₁ =Δh・(tyb ₁ −tya ₁ )/k・Δt =K・(tyb ₁ −tya ₁ ) ...(1) Δy ₂ =Δh・(tyb ₂ −tya ₂ )/k・Δt =K・(tyb ₂ −tya ₂ ) ...(2) where Δh: interval between diode elements (bits) Δt: time required to scan adjacent diode elements k: magnification rate of image by objective lens 14 K: K= Conversion constant for time and displacement found from Δh/t・Δt Similarly, the photodiode arrays 23a and 23b are also scanned, and as shown in FIG. , 27b are output. The processing circuits 29a and 29b respectively bit-process the signals 26a and 26b to determine the distance xa ₁ between the address indicating the first peak value and the address indicating the second peak value.
Find xa ₂ , that is, times tx ₁ and tx ₂ , and the _distances xb ₁ and xb 2 , that is, times txb ₁ and txb ₂ between the address indicating the second peak value and the third peak value, and then Determine the deviations Δx ₁ and Δx ₂ from equations (3) and (4) shown in

Δx ₁ =Δh・(txb ₁ −txa ₁ )/k・Δt =K・(txb ₁ −txa ₁ )……(3) Δx ₂ =Δh・(txb ₂ −txa ₂ )/k・Δt =K・(txb ₂ −txa ₂ ) ...(4) Note that when the position matching target patterns 5a, 5b are significantly tilted with respect to the position matching target patterns 3a, 3b as shown in FIG. 6, the signals 26a, The address of the center of the peaks 26b, 27a, 27b or the address indicating the peak value may be set as the center position of the position matching target patterns 5a, 5b. Furthermore, the exercise circuit 30 performs the following calculation based on the signals Δx ₁ , Δx ₂ , Δy ₁ , Δy ₂ obtained by the processing circuit, and calculates the relative displacement position Δx between the mask 1 and the semiconductor wafer 6. ,Δy,Δ
Find Θ. That is, as shown in FIG. 7, position alignment target patterns 3a, 3b of the mask 1 and position alignment target patterns 5a, 5b of the semiconductor wafer 6 are aligned around the rotation axis p of the rotary table 7.
The relative displacement amount ΔΘ in the rotational direction with respect to

−LsinΔΘ=−(Δy ₁ −Δy ₂ ) ...(5) From the relationship of the above equation (5), the following equation (6) is established.

∴ΔΘ≒1/L (Δy ₁ - Δy ₂ ) = Θ ₀ (Δy ₁ - Δy ₂ ) ...(6) However, L is the position matching target pattern 5a,
5b, and Θ ₀ is a constant corresponding to the value 1/L. The relationship between ΔΘ and Δy ₁ and Δy ₂ is due to the position matching target pattern 5.
This is because two a and 5b are located in the X-axis direction. Further, the relative displacement amount 4x in the X-axis direction between the position alignment target patterns 3a, 3b of the mask 1 and the position alignment target patterns 5a, 5b of the semiconductor wafer 6 is calculated by the following equations (7) and (8). have the following relationship.

Δx=Δx ₁ -Δα, Δx=Δx ₂ +Δα...(7) ∴Δx=1/2(Δx ₁ -Δx ₂ )...(8) Also, the alignment target patterns 3a, 3b of the mask 1 and the semiconductor wafer The relative displacement amount Δy in the Y-axis direction with respect to the target patterns 5a and 5b for position alignment of No. 6 has the relationship shown in the following equations (9) and (10).

Δy=Δy ₁ +Δβ, Δy=Δy ₂ −Δβ ...(9) ∴Δy=1/2(Δy ₁ +Δy ₂ ) ...(10) Next, the relative values of Δx, Δy, and ΔΘ obtained as above A signal representing the amount of displacement is transmitted to the drive circuit 31.
The drive circuit 31 drives the X-axis motor 11 and Y-axis motor 1 in the opposite direction by the determined values of Δx, Δy, and ΔΘ.
2. Drive the motor 9, and when the above Δx, Δy, and ΔΘ all become zero, remove the mask 1 and the semiconductor wafer 6.
are precisely aligned.

As the photodiode array, there is already one in which 1024 photodiode elements are arranged in a line and has a length of 25.4 mm. Further, the target pattern for positional alignment is formed with a square of 1 to 2 mm. Therefore, even if the entire surface of the target pattern for position alignment is imaged by the photodiode array, the positioning accuracy between the mask and the semiconductor wafer is 1 μm.
~2 μm can be obtained.

In the above embodiment, two photodiode arrays in which photodiode elements are arranged in a row are used as optical image detection elements, but the semitransparent mirror 18 and cylindrical lenses 20 and 21 of the optical systems 13a and 13b are removed, and a net-like image is formed. By arranging a photodiode matrix in which photodiode elements for receiving an image of a detection area are arranged vertically and horizontally in a two-dimensional plane at the positions of the photodiode arrays 23a and 23b, the same effects as in the previous embodiment can be achieved. When using a photodiode matrix, when determining the relative displacement between the position matching target patterns 5a and 5b in the X-axis direction and the position matching target patterns 3a and 3b, the signals of the diode elements arranged in the Y-axis direction are used. At the same time as scanning, the signals of the diode elements arranged in the X-axis direction are used to obtain the relative displacement of the position matching target patterns 5a, 5b and the position matching target patterns 3a, 3b in the Y-axis direction. By simultaneously scanning and electrically integrating the light, the same effect as that of the cylindrical lens of the above embodiment in which the optical image is reduced in the width direction can be obtained.

Further, in the above embodiment, the target pattern for aligning the position of the mask is formed of a cross-shaped transparent band, and the target pattern for aligning the position of the semiconductor wafer is formed of thin cross lines, that is, a linear pattern, having a reflectance different from that of the surrounding area. In addition, as shown in FIGS. 8a to 8d, for example, patterns surrounded by transparent or opaque shapes, L shapes, or marks to form predetermined identification areas may be used. Or a target pattern for mask alignment formed with a square transparent pattern, etc.
As shown in FIGS. 9a to 9c, a semiconductor wafer alignment target pattern formed of marks, squares, cross bands, etc. having different reflectance or transmittance from the surroundings is combined and superimposed. Since the same effects as those of the above-mentioned embodiments can also be achieved, the target pattern for position alignment can be selected in various shapes and is not limited to the above-mentioned embodiments.

Furthermore, in the above embodiments, the X and Y directions are orthogonal, but are not necessarily limited to orthogonal coordinates.

[Effects of the Invention] As explained above, according to the present invention, the position of a linear or dot array pattern is detected based on the address of an element constituting a one-dimensional solid-state image sensor. In comparison, the unevenness of the amount of reflected light,
Accurately detects the position of linear or point array patterns without being affected by uneven illumination, noise from other patterns, etc. Moreover, since there is no mechanical scanning mechanism, it is compact and highly reliable. It has the effect that it can do.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing a mask, FIG. 2 is a diagram showing a semiconductor wafer, FIG. 3 is a schematic diagram showing an embodiment of the mask position alignment device of the present invention, and FIG. 4 is a diagram showing a semiconductor wafer. 3 is a diagram showing details of the optical system shown in FIG. Figure 3 shows the light image received by the photodiode array and the signal waveform obtained from the photodiode array, and Figure 7 shows the relative displacement between the target pattern for mask alignment and the target pattern for semiconductor wafer alignment. Diagram showing the relationship between quantities, No. 8
The figure shows an example of a target pattern for aligning a mask, and FIG. 9 is a diagram showing an example of a target pattern for aligning a semiconductor wafer. 5a, 5b... Target pattern for position alignment, 6... Semiconductor wafer, 13a, 13b... Optical system, 14... Objective lens, 17... Light source, 18
... Semi-transparent mirror, 20, 21 ... Cylindrical lens, 22a, 22b, 23a, 23b ... Photodiode array, 24 ... Scan signal generation circuit, 28a, 28b, 29a, 29b ... Processing circuit, 29 ... Arithmetic Circuit, 30...drive circuit.

Claims (1)

[Claims] 1 A two-dimensional pattern having a linear or dot-like pattern is compressed in the linear direction of the pattern using an optical system including a cylindrical lens, and a one-dimensional pattern is formed on a one-dimensional solid-state image sensor in which elements are arranged one-dimensionally. the compressed and imaged one-dimensional pattern is converted into a video signal by the one-dimensional solid-state imaging device, and the address of the one-dimensional solid-state imaging device indicating the center or peak of the mountain of the converted video signal; A position detection method characterized by detecting the position of a linear or dot array pattern.