JPH06283419A - Positioning device - Google Patents

Abstract

PURPOSE:To accurately recognize a mark for alignment by selecting, as a coordinate position for reference pattern, at least one coordinate among a plurality of pattern coordinates in which the space between adjoining patterns is almost equivalent to that between a plurality of specific patterns, in the coordinate positions of respective patterns. CONSTITUTION:On a wafer on which circuit patterns are formed, three marks YMa, YMb and YMc are formed as a Y mark parallel to each other. The space between marks YMa and YMb in a scanning direction of spot light YS is B and the space between marks YMb and YMc is C, further they are respectively placed at the specified positions against the circuit patterns. In addition, these three marks YMa, YMb an YMc have the same grid pattern, and when the spot light YS is applied thereto, diffracted light of the same intensity is similarly generated therefrom. Furthermore, when the spot light is scanned, a detection signal generates the peaks caused by the diffracted light from the pattern as well as the diffracted light from the marks.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a positioning apparatus suitable for an exposure apparatus for manufacturing a semiconductor device, and more particularly, when positioning a substrate having an alignment mark and other patterns, the marks and other patterns The present invention relates to a device for identifying, detecting, and positioning.

[0002]

2. Description of the Related Art Although miniaturization of large-scale integrated circuit (LSI) patterns is progressing year by year, reduction projection type exposure apparatuses have come into widespread use as circuit pattern printing apparatuses satisfying the requirements for miniaturization and having high productivity. There is. In these devices which have been conventionally used, a reticle pattern, which is several times (for example, five times) the pattern to be printed on a silicon wafer, is reduced and projected by a projection lens and printed in a single exposure. 21mm diagonal on the wafer
The area is smaller than the square. Therefore, in order to print a pattern on the entire surface of a wafer having a diameter of about 125 mm, a so-called step-and-repeat method is adopted in which the wafer is placed on a stage, moved for a certain distance, and then exposure is repeated.

In LSI manufacturing, patterns of several layers or more are sequentially formed on a wafer. However, if the overlay error (positional deviation) of patterns between different layers is not kept below a certain value, the conductivity or The insulation state becomes unintended, and the LSI function cannot be achieved. For example, for a circuit having a minimum line width of 1 μm, only a positional deviation of about 0.2 μm is allowed. The reduction projection exposure method is a method of superimposing patterns, that is, a method of superimposing a projected image of a pattern on a reticle and a pattern on a wafer that has already been formed.
There are two methods, an (Off-Axis) method and a through-the-lens (TTL) method.

[0004]

However, whichever method is used, the wafer is roughly positioned (pre-aligned) when it is mounted on the stage. If the precision of this pre-alignment is poor, the alignment mark formed on the wafer will be greatly displaced from the detection center position of the alignment microscope by the off-axis method or the observation system (detection system) by the TTL method. Therefore, there is a drawback that a pattern other than the mark is erroneously recognized and detected, and as a result, incorrect alignment is performed.

It is an object of the present invention to solve these drawbacks and to obtain an apparatus for accurately recognizing alignment marks and positioning a substrate even if the pre-alignment accuracy of the substrate (wafer) is poor. To do.

[0006]

In order to solve such a problem, in the present invention, a substrate stage (4) that holds a substrate (wafer 3) on which a predetermined pattern (circuit pattern or the like) is formed and moves two-dimensionally. ), Coordinate measuring means (X, Y interferometer blocks 9, 10) for measuring the coordinate position of the substrate stage in a predetermined stationary coordinate system (orthogonal coordinate system XY), and driving the substrate stage in the stationary coordinate system. And a moving unit (driving unit 13, 14) for positioning the substrate at a predetermined position (for example, the spot light irradiation position of the alignment microscope) in the stationary coordinate system, the substrate has a circuit pattern (61, 65). A plurality of specific patterns (marks YMa, YMb, Y) that are formed adjacent to each other and are arranged at predetermined intervals (B, C) in a predetermined direction (search direction).
A pattern in a search region which has a reference pattern of Mc) and which extends on the substrate by a predetermined distance including at least a part of the circuit pattern and the reference pattern in the search direction when the substrate stage is moved in the search direction. A pattern detection means (off-axis type alignment microscope or the like) that outputs a photoelectric signal according to the distribution, and the output photoelectric signal,
Position detecting means for detecting each coordinate position on the stationary coordinate system of the pattern in the search area based on the position information (coordinate value) output from the coordinate measuring means when the substrate stage is moved in the search direction. (A microcomputer or the like), and at least one of the coordinate positions of the plurality of patterns in which the intervals between the adjacent patterns substantially match the intervals of the plurality of specific patterns among the detected coordinate positions of the patterns, The determining means (microcomputer) for selecting the coordinate position of the reference pattern, and the moving means for controlling the moving means based on the selected coordinate position and the coordinate position measured by the coordinate measuring means to position the substrate at a predetermined position. A control means (microcomputer) is provided.

[0007]

In the present invention, the reference pattern for alignment is composed of a plurality of specific patterns, and by utilizing the intervals between the plurality of specific patterns, the coordinate position of each pattern detected by the position detecting means is selected. The coordinate position of the reference pattern is selected. Therefore, even if the reference pattern for alignment is formed adjacent to the circuit pattern or the like on the substrate, its coordinate position can be accurately detected.

[0008]

1 is a perspective view showing a schematic structure of a reduction projection type exposure apparatus to which an embodiment of the present invention is applied. A photomask or reticle (hereinafter represented by a reticle) 1 on which a predetermined circuit pattern is drawn is illuminated with light from an illumination light source (not shown). A reduction projection lens (hereinafter, simply referred to as a projection lens) 2 reduces the pattern image of the reticle 1 and projects it onto a wafer 3 coated with a photosensitive agent. The stage 4 which can be moved two-dimensionally is mounted on the orthogonal coordinate system X in the figure with the wafer 3 placed thereon.
Move according to Y. The stage 4 is provided with a θ table 4a that is capable of vacuum-adsorbing the wafer 3 and rotatable with respect to the stage 4. A mark plate (hereinafter simply referred to as a reference mark) 4b serving as a reference at the time of positioning and alignment is fixedly provided on the stage 4. In order to align the wafer 3 by the off-axis method, three alignment microscopes 5, 6, 7 are provided around the projection lens 2, and the optical axis of each microscope is arranged parallel to the optical axis 1 of the projection lens 2. To be done. Since the microscope 5 is used to detect the positional deviation of the wafer 3 in the Y direction, it will be referred to as the Y microscope 5 hereinafter, and the microscope 6 together with the Y microscope 5 will be referred to as the wafer 3.
Since the rotational deviation of the wafer 3 is detected, it will be referred to as a θ microscope 6 hereinafter, and the microscope 7 detects the positional deviation of the wafer 3 in the X direction.
Hereinafter, it will be referred to as an X microscope 7.

A helium neon laser light source (hereinafter, simply referred to as a laser light source) 8 generates laser light having a wavelength that does not expose the photosensitive agent on the wafer 3 to light. The optical member 8a has a beam expander and a cylindrical lens, and shapes the laser light from the laser light source 8 into a parallel light flux having an elongated elliptical cross section. The parallel light flux of the laser light from the optical member 8a is split into two by the beam splitter 8b, and one of the light fluxes is reflected by the mirror 8c and enters the X microscope 7. The other light beam split by the beam splitter 8b is reflected by the mirror 8d and then further split into two by the beam splitter 8e. One of the light beams enters the θ microscope 6 and the other light beam is reflected by the mirror 8f. It is reflected by and enters the Y microscope 5. The laser light flux incident on each of the microscopes 5, 6 and 7 is imaged on the wafer 3 into minute elongated elliptical spot lights YS, θS and XS by the action of the cylindrical lens provided on the optical member 8a. Spot light YS,
θS is determined to be elongated in the X direction of the movement direction of the stage 4, and the spot light XS is determined to be elongated in the Y direction. Moreover, the Y microscope 5 and the θ microscope 6 are arranged in a predetermined relationship so that the direction of the line segment connecting the spot lights YS and θS coincides with the X direction. These spot lights YS, θS, and XS are for detecting the alignment mark or the reference mark 4b formed on the wafer 3.

Now, the X interferometer block 9 includes the stage 4
The X moving mirror 11 having a reflection plane extending in the Y direction is irradiated with a laser beam having a stable frequency, and the reference mirror provided in the X interferometer block 9 is also irradiated with the laser beam. Irradiate. Then, the X interferometer block 9
Causes the reflected light from the X moving mirror 11 and the reference mirror to interfere with each other,
The position of the stage 4 in the X direction is detected by photoelectrically detecting the change in the interference fringe that occurs on the interference wavefront. Since this detection has good stability of the wavelength of the laser beam, for example, 0.
It is performed with a resolution of 02 μm. On the other hand, the Y interferometer block 10 is also constructed in exactly the same manner as the X interferometer block 9, and irradiates the laser beam on the Y moving mirror 12 extending in the X direction to set the position of the stage 4 in the Y direction to 0. Detection with a resolution of 0.02 μm. The drive unit 13 such as a motor is the stage 4
Is moved linearly in the X direction, and the drive unit 14 moves the stage 4 to the Y direction.
Move linearly in the direction. The stage 4 is two-dimensionally positioned at an arbitrary position by these drive units 13 and 14.

Now, an example of the wafer 3 placed on the stage 4 will be described with reference to FIG. Wafer 3 shown in FIG.
Is exposed by a step-and-repeat method on the pattern of the reticle 1 using, for example, the exposure apparatus shown in FIG.
The pattern of the layer is formed. In FIG. 2, only a specific row is shown among a plurality of patterns formed in a matrix on the wafer 3. The wafer 3 has a linear notch on a part of its circumference (hereinafter referred to as a flat).
F is formed, and when the wafer 3 is mounted on the stage 4, it is positioned (pre-aligned) by a pre-positioning device (not shown) so that the linear direction of the flat F coincides with the X direction in FIG. It The pattern of the first layer (geometrical shape formed by minute irregularities on the surface) is arranged in a matrix on the wafer 3 depending on the prealignment accuracy. Therefore, the coordinates of this matrix array are referred to as an array coordinate system X'Y ', and a plurality of rectangular patterns formed according to this array coordinate system X'Y' are referred to as chips. The array coordinate system X′Y ′ corresponds to the moving coordinate system XY of the stage 4, and the origin of the array coordinate system X′Y ′ is set at substantially the center of the wafer 3. Each of the plurality of chips along the X ′ axis of the array coordinate system X′Y ′ is formed with minute unevenness around the chips.
One mark (reference pattern) YM, θM, XM is formed. The marks YM and θM are linearly and alternately arranged along the X ′ axis, and the distance between the marks YM and θM in the X ′ direction is the same in each chip. Now, these plural marks YM,
Of θM and XM, a mark YM1 formed with the chip C1 located at the left end on the wafer 3, a mark XM2 formed with the chip C2 located at the center, and a mark formed with the chip C3 located at the right end. θM
3 and 3 are used to perform overall alignment (global alignment) of the wafer 3. Therefore, the distance between the marks YM1 and θM3 in the X ′ direction is equal to the spot light YS.
The distance between the marks YM and θM in one chip is designed in advance in accordance with the arrangement pitch of the chips and the size (size) of the chips so as to match the distance in the X direction between θS and θS. Although the mark XM2 coincides with the Y ′ axis in FIG. 2, the Y ′ axis may be set so as to pass through the mark XM1 at the other extreme. In FIG. 1, the rotation center of the θ table 4a on which the wafer 3 is placed is the wafer 3
It does not coincide with the center O of the mark, but is rather eccentric, and here it is set at the position of the mark YM1 of the chip C1 or in the vicinity thereof.

Now, the marks YM, θM, X on the wafer 3
M has a so-called lattice structure in which a plurality of minute line elements are regularly arranged in one direction at a constant pitch as shown in FIG. 3, and the minute line elements are each of the coordinate system X'Y '. It is assumed that it is inclined by 45 ° with respect to the axes X'and Y '. FIG. 3 shows the positional relationship between the mark YM1 and the spot light YS. The mark YM1 is formed by forming a plurality of minute line elements inclined at 45 ° in a line in the X ′ direction. Also, mark θ
M3 has the same structure as mark YM1, and mark X
M2 is formed by forming a plurality of minute line elements inclined at 45 ° in a line in the Y ′ direction. When this mark YM1 is irradiated with the spot light YS, diffracted light L ₀ is generated from the mark YM1 depending on the wavelength of the spot light YS, that is, the wavelength of the helium neon laser light and the grating period. In this embodiment, the width of the spot light is equal to the width of the mark, and the mark is longer than the spot light.

FIG. 4 shows marks YM, θM, X on the wafer 3.
3 is a block diagram of a detection system and a control system for photoelectrically detecting M to perform alignment of the wafer 3. FIG. Although only the detection system of the Y microscope 5 is shown here, the other θ microscope 6 and X microscope 7 are similarly configured. Although not shown in FIG. 1, the laser light beam emitted from the laser light source 8 and the optical member 8a is a vibrating mirror 20 that slightly vibrates at a constant amplitude.
Is deflected by a small angle. The laser light flux from the vibrating mirror 20 is reflected by the above-mentioned mirror 8f and is reflected by the Y microscope 5
The mirror 8f is a half mirror. Therefore, the spot light YS is Y which is orthogonal to the longitudinal direction.
It vibrates slightly in the direction. Then, the diffracted light L ₀ from the mark YM1 is incident back on the objective lens 5a of the Y microscope 5, is transmitted through the mirror 8f, is reflected by the dichroic mirror 21, and is incident on the diffracted light detector 22. The diffracted light detector 22 blocks the specularly reflected light L 'on the wafer 3 surface of the spotlight YS, a spatial filter which transmits only the diffracted light L _0, a lens for condensing the diffracted light L _0, collecting Diffracted light L ₀
And a photoelectric detector that outputs a photoelectric signal S ₁ according to the light intensity. The photoelectric signal S ₁ is amplified by the amplifier 23 by a predetermined amount and input to the phase-locked detection circuit (hereinafter referred to as PSD) 24. The PSD 24 inputs an oscillation signal of a frequency f for vibrating the oscillating mirror 20 and an oscillation signal of a frequency 2f which is twice the frequency f, and photoelectrically modulates the oscillation signal with the frequency f by either one of the oscillation signals. The signal S ₁ is synchronously detected. The detection output signal S _{2 obtained} by synchronously detecting the photoelectric signal S ₁ with the oscillation signal of the frequency 2f is input to the comparator 25 which compares it with a predetermined reference level V _L, and the comparator 2
5 binarizes the detection output signal S ₂ . The binarized signal S
₃ is input to the edge detection circuit 26 and outputs an edge signal according to the rising and falling of the binarized signal S ₃ .

The interferometer counter 27 generates a predetermined counting pulse based on the photoelectric signal corresponding to the interference fringes from the X interferometer block 9 and the Y interferometer block 10 and counts the counting pulse. This interferometer counter 2
7 is an X counter 27 for detecting the position of the stage 4 in the X direction.
It has x and a Y counter 27y for position detection in the Y direction. The stage drive circuit 28 includes the drive unit 13 of the stage 4,
14 drives a predetermined amount, and the θ table drive circuit 29 drives a motor that rotates the θ table 4a on the stage 4 by a predetermined amount. The control system and the detection system are microcomputers (hereinafter, μCOM
30). The μCOM 30 outputs the reference level V _L , the edge signal of the edge detection circuit 26, the detection output signal S ₄ of the oscillation signal of the frequency f of the PSD 24, and the interference through the interface circuit (hereinafter, IFC) 31. X of total counter 27
Input of count values (X position information and Y position information) from the counter 27x and the Y counter 27y, and the drive circuits 28 and 2
The drive information is output to 9. The Y microscope 5 is provided with an illumination light source 33 for observing a local area including the spot light YS, and the illumination light from this illumination light source 33 is reflected by the half mirror 34 and enters the objective lens 5a. ,
Illuminate the wafer 3. The TV camera 35 has an objective lens 5
An image observed in the visual field of a is taken and displayed on the television cathode ray tube. The detection system of the Y microscope 5 has been mainly described above, but the detection systems of the θ microscope 6 and the X microscope 7 are also described.
The diffracted light detection circuit, the comparison circuit, the PSD, etc. are configured in exactly the same manner, and are collectively controlled by the μCOM 30. The spot light θS slightly vibrates in the Y direction, and the spot light XS slightly vibrates in the X direction. The spot light YS, θ
The vibration amplitudes of S and XS are set to be substantially equal to the width of the mark, for example.

Next, an example of the wafer alignment operation in this apparatus will be described with reference to the flow charts of FIGS. First, in steps 100 and 101, the wafer 3 as shown in FIG. 2 is positioned by a pre-alignment mechanism (not shown) such that the flat F is parallel to the X axis of the coordinate system of the stage 4. Then step 1
In 02, the wafer 3 is transferred to the θ table 4a of the stage 4 while keeping the state, and the stage 4 is moved to move the Y microscope 5
Then, the wafer 3 is carried under the θ microscope 6. At this time, due to an error according to the pre-alignment accuracy of the wafer 3, the wafer 3
The upper mark YM1 (Y mark) and the mark θM3 (θ mark) are aligned in a predetermined positional relationship with the spot lights YS and θS, respectively. FIG. 7 is an arrangement diagram of the wafer 3 and the spot lights YS and θS showing the state. In FIG. 7, the wafer 3 is placed on the stage 4 with a rotation error and an error in the XY directions. Next, the μCOM 30 executes laser scanning of the Y and θ mark portions in step 103. this is,
This is performed by moving the stage 4 (wafer 3) in the Y direction and relatively scanning the mark YM1 (θM3) with the spot light YS (θS). The amount of movement is determined by the pre-alignment accuracy of the wafer 3 and the mark YM1 (θM3).
Is set to a range (exploration region) that is sufficiently longer than the sum of the lengths of the pattern inhibition regions provided before and after the scanning direction.

FIG. 8 shows how the spot light YS scans a certain range on the wafer 3 including the mark YM1 in step 103. Actually, the stage 4 (wafer 3) moves in the Y direction, but for convenience of explanation, it is assumed that the spot light YS moves (scans) on the wafer 3 in the direction of arrow 66 (search direction) in the figure. The pattern prohibited area is defined by the circuit patterns 61 and 65 before and after the mark YM1 in the scanning direction.
Is a region in which the formation of is prohibited, and is secured only in a predetermined range in design. In this embodiment, the pattern prohibited area is defined as the areas A and B before and after the mark YM1.
Range A has a shorter distance in the scanning direction than range B. In FIG. 8, chips and scribe lines 64 for partitioning the chips are formed between the mark YM1 and the circuit pattern 65.

Now, when step 103 is executed, μ
The COM 30 is the wafer 3 according to the flowchart of FIG.
The mark position of the wafer 3 is detected based on the diffracted light from the. This operation will be described with reference to the time chart of FIG. First, in step 200, the spot light YS as shown in FIG.
The stage 4 starts moving in the Y direction at a predetermined speed so that the scanning is performed from the state of being positioned at the left end toward the right end. When the spot light YS moves in the direction of the arrow 66, the detection output signal S ₂ becomes the circuit pattern 61 as shown in FIG.
There are peaks 72 and 73 at the position, a large peak 74 at the mark YM1, a small peak 74 at the scribe line 64, and peaks 76, 77 and 78 at the circuit pattern 65. Such a detection output signal S ₂
Is compared with the reference level V _L by the comparator 25, and the binarized signal S ₃ is generated so that it becomes the theoretical value “H” when the detection output signal S ₂ is larger than the reference level V _L. In the detected output signal S ₂ , peaks 72 and 77 that are smaller than the reference level V _L are ignored. Next, the μCOM 30 detects the edge signal from the edge detection circuit 26 in step 201, and latches the value Y ₁ of the Y counter 27y of the interferometer counter 27 in step 202 in the case of a rising edge. Then, again in step 201, it is detected whether or not it is a rising edge. Next, the μCOM 30 monitors the edge signal from the edge detection circuit 26 in step 203, and in the case of the falling edge, latches the value Y ₂ of the Y counter 27y in step 204. Then, the μCOM 30 performs step 2
In 05, the absolute value of the difference between the values Y ₁ and Y ₂ , that is, the movement amount of the stage 4 in the Y direction while the detection output signal S ₂ is larger than the reference level V _L is obtained, and the absolute value | Y ₁ −
It is determined whether Y ₂ | is a value between a predetermined value WA and WB. The width of the mark YM1 in the Y direction is known in advance by design. Therefore, if there is an absolute value | Y ₁ −Y ₂ | between the value WA which is slightly smaller than the width of the mark and the value WB which is slightly larger than the width of the mark, it receives the diffracted light from the pattern that seems to be the mark YM _1. I judge that I did. Those that do not satisfy this condition cannot be marks, so
The process is repeated from step 201 again. If the condition is satisfied, the μCOM 30 proceeds to step 206.
The position of the rising edge latched in step ₁ , that is, the value Y ₁ is stored and the operation from step 201 is repeated again. Figure 9
(C) shows the binarized signal S depending on the condition of step 205.
Reference numeral ₃ is a binarized signal that represents the equivalently filtered state. The pulse 79 in the binarized signal S ₃ of FIG.
Since the widths of 80 and 80 are narrow and do not satisfy the above condition, they are erased as shown in FIG. 9C. And step 206
The value Y ₁ stored in FIG. 9C is as shown in FIG.
4 positions 8 corresponding to the rising edge of the binary signal of
Represents 1, 82, 83, 84. If the μCOM 30 does not detect the falling edge in step 203, the process proceeds to step 207. ΜCOM in step 207
30 determines whether or not the spot light YS has moved to the right end in FIG. 8 by a predetermined distance, and if the movement has not been completed, repeats the operation from step 201.

As described above, it is determined whether the level of the detection output signal S ₂ is equal to or higher than a predetermined value (level determination) and whether it is close to the mark width (width determination). Positions 81 and 8 that are candidates for the Y-direction position of the mark YM1
When 2, 83 and 84 are obtained, the μCOM 30 stops moving the stage 4. The above operation is performed in the same manner for the spot light θS, and the position in the Y direction of the pattern that looks like the mark θM3 is stored in the μCOM 30.

Next, the μCOM 30 executes step 104 in FIG.
To execute. In step 104, based on the mark forming condition, that is, there is no pattern in the ranges A and B before and after the mark YM1, a position satisfying the condition is found among the positions detected as candidates. . The μCOM 30 has a mark Y as shown in FIG.
The positions 81, 82, 83, 84 that are candidates for M1 are stored. Therefore, the μCOM 30 first checks the position 81 for any pattern within the ranges A and B (distances A and B) before and after the position 81. There is no pattern in the range A before the position 81, that is, the left end in FIG. However, after position 81 (on the right side in FIG. 9D), there is a candidate pattern (position 82) in range B. Therefore, the μCOM 30 determines that the pattern existing at the position 81 is not the mark YM1. Next μC
The OM 30 also looks at position 82. At this time,
Since the pattern existing at the position 82 satisfies the mark forming condition, the μCOM 30 sets the pattern to the mark YM.
Judge as 1. It should be noted that the same determination is performed for the positions 83 and 84 to confirm that the position 82 is the position of the mark YM1.

Next, the μCOM 30 returns the stage 4 in the Y direction so that the measurement position of the Y counter 27y of the interferometer counter 27 coincides with the stored position 82 at the target Y mark positioning in step 105, and marks the spot YS. Overlap with YM1. Normally, if the stage 4 is positioned based on the position 82 determined in step 104, the spot light YS and the mark YM1 coincide with each other without error. During this operation, the μCOM 30 selects a position satisfying the condition that no pattern exists in the pattern prohibited area (range A and range B) among the plurality of positions that are candidates for the mark θM3, as in step 104. To detect.

When the spot light YS and the mark YM1 overlap, a strong diffracted light L ₀ is generated from the mark YM1. Since the spot light YS vibrates slightly in the Y direction and its amplitude is set to be substantially equal to the width of the mark YM1, for example, the detection signal S ₄ synchronously detected by the PSD 24 at the frequency f is as shown in FIG. It becomes an S-curve signal. In FIG. 10, the vertical axis represents the magnitude of the detection output signal S ₄ , and the horizontal axis represents the deviation amount ΔD of the center position of the width of the mark YM1 with respect to the vibration center position of the spot light YS. When this deviation amount ΔD becomes zero, that is, the detection output signal S
_{When 4} becomes zero, it means that the vibration center of the spot light YS and the center of the width of the mark YM1 coincide with each other, and precise alignment is achieved. Therefore, the μCOM30 is step 1
In the Y mark precision positioning of 06, the position of the stage 4 in the Y direction is precisely positioned by servo control using the detection output signal S ₄ . Specifically, based on the detection output signal S ₄ from the PSD 24, the stage drive circuit 28 is servo-controlled (feedback control) so that the magnitude of the detection output signal 24 is zero or within a predetermined range including zero. To do. As a result, the mark YM1 and the spot light YS exactly match.

Along with this operation, the μCOM 30 determines the rotation direction of the wafer 3 in the rotation direction determination in step 107, that is, the X'axis of the coordinate system X'Y 'of the wafer 3 with respect to the X axis of the coordinate system XY of the stage 4. Detect which direction the is tilted. This is done by calculating the deviation between the Y-direction position 82 of the mark YM1 detected in the previous step 104 and the Y-direction position of the mark θM3, and determining whether the difference value is positive or negative. Required by For example, when the deviation is positive, the wafer 3 rotates counterclockwise with respect to the coordinate system XY of the stage 4 as shown in FIG. 7, and when the deviation is negative, the wafer 3 rotates clockwise. Next, μCOM30
Rotates the θ table 4a in the direction opposite to the detected rotation direction by the θ table rotation in step 108. At this time, the diffracted light generated by the spot light θS is photoelectrically detected to obtain a detection output signal which is synchronously detected by the oscillation signal of frequency 2f, and based on the binarized signal obtained by binarizing the detection output signal, The μCOM 30 simultaneously performs the detection operation of the pattern that looks like the mark θM3.

Then, after the rotation of the θ table 4a is started, it is checked how many marks θM3-like patterns are detected. The operation will be described with reference to FIG. Figure 1
1 is a diagram schematically showing the state of the wafer 3 before the rotation of the θ table 4a. A line l ₁ is a line segment that connects both vibration centers of the spot light YS and θS, and is parallel to the X axis of the coordinate system XY of the stage 4. The Y mark position information is the positions 81, 82, 8 of the pattern like the mark YM1 shown in FIG.
3 and 84 are represented in the Y direction, and the θ mark position information is the positions 85, 86, and 8 of the pattern that looks like the mark θM3.
7 and 88 are represented in the Y direction. And mark Y
The positions of M1 and θM3 are determined to be the position 82 and the position 87, respectively, in step 104 described above. Now, since the rotation center of the θ table 4a is in the vicinity of the mark YM1 and the Y-direction servo is operating on the stage 4 so that the spot light YS and the mark YM1 coincide with each other,
The mark YM1 and the spot light YS are not displaced by the rotation of the table 4a, and the mark θM3 and the spot light θS are relatively scanned in the Y direction. μC
The OM30 has a mark θ for the position 82 of the mark YM1.
It is determined in which direction each of the positions 85 to 89 of the M3-like pattern is in the Y direction. When the θ table 4a is rotated clockwise as indicated by arrow 90,
5 does not cross the line l ₁ . Therefore, the μCOM 30 detects the second position of the signal generated after starting the rotation of the θ table 4a, that is, the position 87 next to the position 86 of the mark θM3-like pattern except for the position 85. At this time, the μCOM 30 detects the mark-like pattern of the mark θM3 by the level determination and the width determination based on the photoelectric signal of the diffracted light detected by the θ microscope 6. When determining the width at this time, the Y counter 2 of the interferometer counter 27 is used.
Since the measured value of 7y cannot be used, the rotation speed of the θ table 4a is used for calculation.

When the θ table 4a starts rotating, the μCOM 30 detects a rising edge and a falling edge from the edge signal of the edge detection circuit 26θ based on the photoelectric signal detected by the θ microscope 6. Then, after obtaining the time interval between the rising edge and the falling edge, the θ table 4a
The widths of the rising edge and the falling edge are calculated from the relative moving speeds of the mark θM3 and the spot θS in the Y direction due to the rotation. When the width WS detected as described above satisfies the width determination condition (WA ≦ WS ≦ WB), the edge signal is detected as the first signal (position 86 in FIG. 11). Similarly, when the μCOM 30 detects the second signal (position 87 in FIG. 11), the μCOM 30 stops the rotation of the θ table 4a. At that time, in FIG. 11, the line l ₂ connecting the positions 82 and 87, that is, the array coordinate system X′Y ′
The X'axis of and the line l ₁ coincide.

Next, the μCOM 30 executes step 10 in FIG.
9 θ mark precise positioning is performed. This is based on the detection output signal (S curve signal) obtained by synchronously detecting the photoelectric signal detected by the θ microscope 6 with the oscillation signal of the frequency f.
The rotation drive of the θ table 4a is servo-controlled so that the center of the M3 in the Y direction and the vibration center of the spot light θS coincide with each other. This control is also performed so that the S-curve signal becomes zero as in step 107. Of course this step 10
During 9, the servo control of the stage 4 in the Y direction is also continuously performed, and the marks YM1 and θM3 are simultaneously positioned. As described above, the center (width center) of the mark YM1 in the Y direction and the vibration center of the spot light YS match, the center (width center) of the mark θM3 in the Y direction and the vibration center of the spot light θS match, and the stage The rotational deviation between the coordinate system XY of No. 4 and the arrangement coordinate system X'Y 'of the wafer 3 is accurately corrected. After that, the μCOM 30 presets in the Y counter 27y of the interferometer counter 27 a value corresponding to the distance in the Y direction from the optical axis 1 of the projection lens 2 to the line l ₁ connecting the spot lights YS and θS. This defines the position of the wafer 3 in the Y direction with respect to the optical axis 1 of the projection lens 2.

Next, the μCOM 30 sets X in step 110.
In alignment, the mark XM2 and the spot light XS are made to coincide with each other in exactly the same way as the detection and positioning of the marks YM1 and θM3, and more precise X-direction alignment is performed by synchronous detection. After completing the alignment, μCOM
Reference numeral 30 denotes an X counter 27x of the interferometer counter 27, and a spot light XS from the optical axis l of the projection lens 2 from the X microscope 7.
The value corresponding to the distance in the X direction up to is preset. As a result, the X of the wafer 3 with respect to the optical axis 1 of the projection lens 2 is
The directional position is defined. This completes the global alignment of the wafer by the off-axis method.
Then, the μCOM 30 executes the exposure operation of the next step 111, and steps the stage 4 in the X direction and the Y direction based on the measurement value of the interferometer counter 27 to expose the circuit pattern of the reticle 1 onto the wafer 3. Repeat what you do.

According to the above operation, even if the pre-alignment accuracy of the wafer 3 is poor, whether the photoelectric signal corresponding to the intensity of the diffracted light is higher than the reference level V _L , and the width thereof satisfies the predetermined condition. It is determined whether or not the value is close to the width of the mark, and the condition for forming the mark, that is, that there is no pattern within the range A, B of the predetermined distance before and after the mark, the true mark Is recognized, and the rotational deviation of the wafer 3 and the positional deviation in the X and Y directions are detected extremely accurately.

In step 108, the width of the edge signal is obtained from the rotation speed of the θ table 4a.
If there is an encoder or the like that accurately detects the rotation amount of a, the width of the edge signal may be obtained from the measurement signal of the encoder. If the encoder is also provided, then in step 107 the Y at positions 82 and 87
After obtaining the deviation in the direction and calculating the rotation amount of the θ table 4a according to the deviation, while rotating the θ table 4a by the rotation amount while reading the measurement signal of the encoder,
You may make it implement step 109.

Next, an embodiment of the present invention will be described with reference to FIGS. In this embodiment, the marks (YM, θ
Unlike the previous operation example, the formation conditions of (M, XM) are such that a plurality of marks are arranged at a predetermined interval in the relative scanning direction with respect to the spot light, and the intervals are recognized. FIG. 12 is a plan view showing the arrangement of the marks and the circuit pattern on the wafer 3. In FIG. 12, three marks YMa, YMb, and YMc are formed in parallel as Y marks on the wafer 3 on which the same circuit pattern as in FIG. 8 is formed. The distance between the marks YMa and YMb in the scanning direction of the spot light YS is B, and the distance between the marks YMb and YMc is C, both of which are predetermined values in design and are provided at predetermined positions with respect to the circuit pattern. Has been. Also,
The three marks YMa, YMb, and YMc have the same grid pattern, and when the spot light YS is irradiated, they all generate diffracted light of the same intensity. Such marks are formed in the same manner for the θ mark and the X mark.

Now, the alignment of the wafer 3 is performed by the same procedure as that of the flow chart of FIG.
However, the mark condition determination in step 104 is different. When the spot light YS is scanned as indicated by the arrow 66 in FIG. 12 in step 103, the detection output signal S is obtained as shown in FIG.
_In No. 2, peaks 74a, 74b, and 74c due to the strong diffracted light from the marks YMa, YMb, and YMc occur together with the peak due to the diffracted light from the pattern. When this detection output signal S ₂ is binarized, a binarized signal S ₃ as shown in FIG. 13 (b) is obtained, and a signal as shown in FIG. 13 (c) is obtained by the width determination. Rising edge position 8
1, 82a, 82b, 82c, 83, 84 are detected.
Then, in the mark condition determination in step 104, μCOM3
0 is the position 82a and the position 82b of the plurality of positions.
Is B, and the succeeding positions 82b and 82c are
Position 82a, 82b,
It is determined that 82c is the position of the marks YMa, YMb, YMc, respectively.

At this time, determination based on the pattern prohibited area may be added. For example, before the position 82a, that is, there is no pattern in the range A before the spot light YS irradiates the mark YMa in FIGS. 12 and 13, and in the range D after the spot light YS irradiates the mark YMc. Determine that there is no pattern. In addition, in FIG. 12, range D
Although there is a scribe line 64 in the inside, since the scribe line 64 does not have a structure that actively generates diffracted light, the peak 75 generated in the detection output signal S ₂ is small and its width is narrow, and the mark condition is finally set. Figure 1
It rarely remains as information like 3 (d). In such a case, the scribe line 64 may be present in the pattern prohibited area (area B), but strong diffracted light is also generated from the scribe line 64 depending on the wafer manufacturing process, so erroneous detection is avoided. Therefore, it is preferable that the scribe line is also outside the pattern prohibited area. Then, in the next step 105 of Y mark positioning,
For example, the stage 4 is moved to the position 82a so that the mark YMa and the spot light YS match. Of course, the other marks YMb and YMc may be used, which can be appropriately determined in advance. As described above, according to the present embodiment, the detection rate by the mark condition determination in step 104, that is, the recognition rate is dramatically improved, and the mark is surely detected.

Next, a modified example of the above embodiment will be described.
In FIG. 4, the reference level V applied to the comparator 25
_L is a μCO using a digital-analog converter, for example.
It can be arbitrarily set to any value by the command of M30. Therefore, a peak hold circuit 40 for holding the maximum value of the detection output signal S ₂ and an analog-digital converter 41 for the μCOM 30 to read the held peak value via the interface 31 are added to the circuit block of FIG. To do. Then, in step 103 of FIG. 5, scanning with spot light in a certain range including the mark is performed twice, that is, one reciprocation. During the first scan, the peak hold circuit 40 is operated to hold the maximum peak value of the detection output signal S ₂ . Then, the μCOM 30 reads the peak value and controls so that the reference level _VL is set to a value of about 70 to 80% of the peak value. After that, the μCOM 30 scans the wafer 3 and the spot light in the opposite direction, and similarly detects the edge position of the binarized signal S ₃ . At this time, since the signal obtained by binarizing the detection output signal S ₂ shown in FIGS. 9 and 13 has a high intensity of diffracted light from the mark, a peak component due to weak diffracted light from other patterns is not detected. , Only the pulse due to the mark appears. However, when the pre-alignment accuracy is low and the circuit pattern is fine, the pattern may generate diffracted light at the same level as the diffracted light by the mark. Therefore, similar to the previous embodiment, the mark forming condition (step 104) to determine whether the mark is a true mark. According to this example, since the photoelectric signal (detection output signal S ₂ ) corresponding to the diffracted light having the largest possible intensity such as the diffracted light from the mark is binarized, the mark can be recognized more easily and reliably. Has the advantage that

Although the embodiment of the present invention has been described above, various other modifications are possible. In this embodiment, the spot light Y
Although S, θS, and MS are each made to vibrate slightly, they may be stationary. In that case, the photoelectric signal corresponding to the diffracted light has the same waveform as the detection output signal S ₂ . In addition, when the spot light is stationary, the phase-locked detection circuit cannot be used, and thus precise alignment is difficult as it is. Therefore, when scanning the spot light on the wafer 3, a counting pulse (for example, 0.0
It is preferable that the photoelectric signal corresponding to the diffracted light is sequentially sampled at every 2 μm), converted into a digital value, and then sequentially stored in the memory circuit. Then, the waveform of the photoelectric signal is temporarily stored and the position of the mark is detected by a process of extracting what appears to be the waveform of the mark, so-called waveform processing, whereby alignment according to the measurement resolution of the laser interferometer can be performed. Of course, the mark forming conditions are also determined at the same time.

The spot light was scanned in the direction orthogonal to the longitudinal direction, but it is not always necessary, and the same applies even if the spot light and the mark are inclined at a predetermined angle (for example, 45 °) with respect to the scanning orbit. The effect of is obtained. Further, the mark has a grid pattern in which minute line elements inclined at 45 ° are gathered in a line, but the inclination at 45 ° is not always necessary.

Further, when a plurality of the same marks are provided as in the embodiment of the present invention, the peaks 74a, 74 corresponding to the marks in the detection output signal S ₂ as shown in FIG. 13A.
By utilizing the fact that the waveforms (shapes) of b and 74c are the same, the waveforms of a plurality of peaks obtained from the mark-like pattern are compared with each other, for example, by a correlation operation, and the peak waveforms determined to be the same You may make it recognize a mark from. According to this method, there is an advantage that the mark recognition rate is high even when each of a plurality of marks changes to the same due to a process such as development and etching of the wafer. Further, the spot light may be a simple circle instead of the elongated elliptical shape, and the mark may be a single stripe pattern instead of a grid pattern or a stripe pattern in which a plurality of stripes are arranged in the scanning direction.
In this case, scattered light generated at the edge of the stripe pattern may be photoelectrically detected. Further, instead of detecting the mark by using the spot light obtained by condensing the laser light, the so-called vibration slit type photoelectric microscope that photoelectrically detects the observation image of the mark by the objective lens through the vibration slit has the same effect. Is obtained. Further, even when the present invention is applied to an exposure apparatus having a structure for observing and detecting a mark on a wafer by a through-the-lens (TTL) method, the same effect can be obtained.

[0036]

As described above, according to the present invention, since the photoelectric signal (edge signal) satisfying the conditions such as the interval between the plurality of marks for alignment is selected, the signal from the mark is transmitted from another pattern. It can be distinguished from the signal, and the effect that the substrate (wafer) can be accurately positioned even if the pre-alignment accuracy is poor is obtained. Further, since the recognition rate of the mark is improved, the prohibited area around the mark can be narrowed, and the effect of increasing the area of the circuit pattern area of the IC or LSI is also obtained.

[Brief description of drawings]

FIG. 1 is a perspective view showing a schematic configuration of a reduction projection type exposure apparatus to which an embodiment of the present invention is applied.

FIG. 2 is a plan view showing chips and marks formed on a wafer.

FIG. 3 is a plan view showing a positional relationship between marks and spot lights.

FIG. 4 is a block diagram showing a configuration of a detection system and a control system for alignment.

FIG. 5 is a flowchart illustrating an overall operation for aligning a wafer.

FIG. 6 is a flowchart illustrating a mark-like pattern detecting operation.

FIG. 7 is a plan view showing an arrangement relationship between a wafer and spot light after prealignment.

FIG. 8 is a plan view showing the arrangement of patterns and marks on a wafer and spot light.

9 is a time chart diagram showing the states of various signals obtained when the spot light shown in FIG. 8 is scanned.

FIG. 10 is a waveform diagram showing an example of a detection output signal (S curve signal).

FIG. 11 is a diagram schematically showing a rotation deviation of a wafer.

FIG. 12 is a plan view showing an arrangement relationship of marks according to an embodiment of the present invention.

FIG. 13 is a time chart diagram showing states of various signals obtained by scanning the pattern and the mark shown in FIG. 12 with spot light.

[Explanation of symbols]

 1 ... Reticle 2 ... Reduction projection lens 3 ... Wafer 4 ... Stage 4a ... θ table 5 ... Y microscope 6 ... θ microscope 7 ...・ X microscope YS, θS, XS ... Spot light YM, θM, XM ... Mark

Claims (2)

[Claims] 1. A substrate stage that two-dimensionally moves while holding a substrate on which a predetermined pattern is formed, coordinate measuring means for measuring the coordinate position of the substrate stage within a predetermined stationary coordinate system, and the stationary coordinates. A device for moving the substrate stage in a system, which positions the substrate at a predetermined position in the stationary coordinate system, wherein the substrate is formed adjacent to the pattern and is arranged in a predetermined direction. A reference pattern composed of a plurality of specific patterns arranged at predetermined intervals; at least a part of the pattern in the predetermined direction on the substrate when the substrate stage is moved in the predetermined direction; Pattern detection means for outputting a photoelectric signal according to a pattern distribution in a search region extending by a predetermined distance and including a reference pattern; and the output photoelectric signal, Position detection for detecting each coordinate position on the stationary coordinate system of the pattern in the search region based on the position information output from the coordinate measuring means when the substrate stage is moved in the predetermined direction. A coordinate pattern of the reference pattern, at least one of coordinate positions of a plurality of patterns in which an interval between adjacent patterns substantially matches an interval of the plurality of specific patterns among coordinate positions of the detected patterns. Determination means for selecting as a position; and control means for controlling the moving means based on the selected coordinate position and the coordinate position measured by the coordinate measuring means to position the substrate at the predetermined position. A positioning device characterized by being provided. 2. The reference pattern includes three linear patterns extending in a direction orthogonal to the predetermined direction as the plurality of specific patterns, and the intervals of the three linear patterns are different from each other. An apparatus according to claim 1, characterized in that