JPH10260010A - Mark measuring method and its device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To measure shape of a periodic mark accurately in an processing method for periodic mark images such as a wedge mark. SOLUTION: By photographing an evaluation mark image WMI which consists of wedge marks arranged periodically with an image pickup element 15, extracting scanning lines R1 -R8 in area where wedge mark images 311 exist from scanning lines Sj in the image pickup element 15, summing up image signals of these scanning lines R1 -R8 in the periodic direction, and comparing the summed image signals with specified slice level, for example, the length of the mark is determined. As necessary, the evaluation mark image WMI and the image pickup element 15 are shifted only by one integer-th of pixel arrangement pitch and the operation which is the image pickup and summing of image signals is repeated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a mark measuring method and apparatus for measuring the shape of a periodic mark, and relates to a micro device such as a semiconductor device, a liquid crystal display device, an image pickup device (CCD) or a thin film magnetic head. It is suitable for use in measuring the shape of a periodic mark for calibration of a projection exposure apparatus used in a photolithography process for manufacturing.

[0002]

2. Description of the Related Art Conventionally, when manufacturing semiconductor devices and the like,
A projection exposure apparatus (stepper or the like) that projects a pattern image of a reticle (or a photomask or the like) as a mask onto a wafer (or a glass plate or the like) coated with a photoresist as a photosensitive substrate via a projection optical system. Is used. The reticle pattern image is always in high resolution,
In addition, when performing overlay exposure, in order to perform exposure while maintaining a predetermined overlay accuracy, always keep the imaging characteristics of the projection optical system of the projection exposure apparatus, the positioning accuracy of the stage system, and the like in the highest state. There is a need. Therefore, in this type of projection exposure apparatus, for example, self-diagnosis including calibration of the imaging characteristics and the like is generally performed periodically, for example.

In order to calibrate the position of the image plane of the projection optical system (best focus position) in the image forming characteristics, for example, an image of a specific test pattern is changed while the height of the wafer is changed stepwise. Conventionally, a method has been used in which a series of shot areas on a wafer are sequentially exposed to light, and the intervals and lengths of uneven resist patterns formed in each shot area by development are measured. Of these methods, the 39th (Spring 1992) Applied Physics-related Lecture 30p-
A so-called "wedge-shaped mark" formed by periodically repeating a diamond-shaped pattern elongated in a measurement direction in a direction perpendicular to the diamond-shaped pattern as a test pattern reported in NA-1 "Automatic Best Focus Measurement Method of Projection Exposure Apparatus" The method is
It is widely used because it is simple and has high measurement accuracy.

In this measurement method, wedge-shaped mark images are sequentially exposed to different positions on the wafer via a projection optical system while changing the focus position of the wafer, and each of the resist patterns formed on the wafer after development is formed. The length of the wedge-shaped mark in the longitudinal direction is measured, and the best focus position is detected by utilizing the fact that the wedge-shaped mark becomes the longest when the surface of the wafer is at the best focus position. In order to measure the length of the wedge-shaped mark, conventionally, a laser scanning sensor provided for alignment in a projection exposure apparatus has been used. The laser scan type sensor irradiates a laser beam having a width of several μm in the measurement direction and a length of about 50 to 60 μm in a direction orthogonal thereto in the vicinity of the alignment mark or the wedge-shaped mark on the wafer, At this time, it receives the diffracted light generated from the mark to be detected. At this time, for example, by driving the wafer stage on which the wafer is mounted in the measurement direction, the laser beam and the mark to be detected are relatively scanned in the measurement direction, and the photoelectric conversion signal of the diffracted light is, for example, The position of the mark to be detected or the length in the measurement direction is obtained from the range of the position of the wafer stage when the predetermined slice level is exceeded. The position of the wafer stage is, for example, 0.01 μm by a laser interferometer.
It is measured with about the same resolution.

Further, instead of exposing a single wedge-shaped mark image on a wafer, two lines and space pattern images having slightly different arrangement directions are superposed and exposed to form moire fringes. Forming a composite pattern substantially equivalent to the wedge-shaped mark, and measuring the length of the composite pattern to measure the amount of misalignment between the two line-and-space patterns, ie, the overlay accuracy. Has also been proposed.

[0006]

In the prior art as described above, in the laser scan type sensor, since the detection light is a laser beam (monochromatic light), when the alignment mark on the wafer is detected, the laser beam is detected in the photoresist. It is easily affected by multiple interference, and there is a possibility that the detection accuracy may be reduced due to uneven coating of the photoresist or the asymmetry of the mark shape. Also, when detecting the resist pattern, the detection accuracy may be reduced due to the asymmetry of the mark shape. Thus, as an alignment sensor, an imaging sensor using a broadband light source is becoming mainstream. Recently, an exposure apparatus equipped only with an imaging sensor and not equipped with a laser scan sensor has begun to be used.

In the case of a laser scan type sensor, the resolution of a measurement position is almost the resolution of a laser interferometer, for example, as small as about 0.01 μm, whereas the resolution of an imaging sensor is about the resolution of an image sensor. There is a limit of about 0.1 μm or more even in consideration of the imaging magnification. Therefore, simply performing normal image processing on an image signal using an imaging sensor may result in insufficient resolution for measuring the length of the test pattern.

In a normal imaging sensor, an image signal obtained from an image sensor is integrated in a non-measurement direction (a direction orthogonal to the measurement direction) and processed as a one-dimensional signal. Therefore, in a pattern having periodicity in the non-measurement direction such as the wedge-shaped mark, image information of the mark portion is buried in an image signal from the non-mark portion, and it is difficult to accurately measure the mark length.

In view of the above, the present invention provides a mark measuring method capable of determining the shape of a periodic mark with high accuracy by a method of image processing a periodic mark image such as a wedge-shaped mark. With the goal. Still another object of the present invention is to provide a mark measuring device capable of performing such a mark measuring method.

[0010]

According to the mark measuring method of the present invention, an image of a periodic mark (WM) formed on a substrate (W) to be processed is picked up via a predetermined image pickup device (15). In a mark measuring method for processing an image signal obtained by this imaging to measure the shape of the periodic mark, an image (WM) of the periodic mark is included in the image signal.
In the period direction (y direction) of I), the image signal of the portion having the image of the periodic mark is extracted, and the image signal thus extracted is integrated in the period direction of the image of the periodic mark. The mark length in the measurement direction (x direction) intersecting with the periodic direction of the image of the periodic mark is obtained from the image signal integrated in the above.

According to the present invention, the image signal of the non-mark portion (background portion) having no image of the periodic mark is excluded from the integration target, and only the image signal of the portion having the image of the periodic mark is included. Determines the mark length. Therefore, the SN ratio of the used image signal is improved, and the mark length can be obtained with high accuracy. In this case, when extracting an image signal of a portion having the image of the periodic mark in the periodic direction of the image of the periodic mark (WMI), the image signal is extracted at a predetermined pitch in the periodic direction of the image of the periodic mark. in divided into image signals (S _j) from the pixel of arrayed plurality of rows, the contrast between the maximum and minimum levels within the image signals from the pixels of plural rows (S _j) exceeds a predetermined threshold value It is desirable to extract signals or signals whose signal level is within a predetermined slice level range over a predetermined length. Since the level of the image signal changes in the portion where the image of the periodic mark is present as compared with the non-mark portion, the portion where the image of the periodic mark exists can be accurately detected by comparing with the contrast or a predetermined slice level. .

The image of the periodic mark and the image sensor (15) are sequentially arranged in at least one of the periodic direction of the periodic mark or the measurement direction intersecting the periodic direction. The operation of picking up an image of the periodic mark is repeated n times while being relatively shifted by 1 / n (n is an integer of 2 or more) of the pitch, and the period is calculated from n sets of image signals obtained by this repetition. It is desirable to extract an image signal of a portion where the image of the periodic mark exists in the periodic direction of the image of the target mark. Thereby, as shown in FIG. 4A, even if the image of the periodic mark is at the boundary of the pixel column of the image sensor (15), the image of the periodic mark and the image sensor (15) Is shifted by の of the pixel pitch, the image of the periodic mark moves to the center of a predetermined pixel row of the image sensor (15) as shown in FIG. The mark length can be accurately detected by using the signal.

Further, the mark measuring apparatus according to the present invention provides a periodic mark (W) formed on a substrate (W) to be processed.
M) having an imaging optical system (13, 14) for forming an image and an image sensor (15) for picking up an image of the periodic mark. In a mark measuring device for measuring the shape of a target mark, the image of the periodic mark and the image sensor (15) are moved relative to at least one of the periodic direction of the image of the periodic mark or the measurement direction intersecting the periodic direction. A moving relative moving device (4, 7) and an image of the periodic mark on the image pickup device (15) each time the image of the periodic mark and the image pickup device (15) are relatively moved via the relative moving device. A control unit (6) for capturing an image of the image and an image signal of a portion having the image of the periodic mark in the periodic direction of the image of the periodic mark are extracted from the image signals obtained by the image sensor (15).
An arithmetic unit (16) that integrates the extracted image signal in the periodic direction of the image of the periodic mark and obtains a mark length of the image of the periodic mark in the measurement direction from the integrated image signal. It is provided.

According to the mark measuring device of the present invention,
The image of the periodic mark and the image sensor are moved, for example, 1 / n of the pixel pitch in the periodic direction through the relative moving device.
The mark measurement method of the present invention can be implemented by using the n sets of image signals obtained by shifting each other.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. In this embodiment, the present invention is applied to a case where calibration of a best focus position is performed by a projection exposure apparatus. FIG. 5 shows a schematic configuration of a projection exposure apparatus used in this example. In FIG. 5, at the time of exposure, exposure light IL (for example, i-line of a mercury lamp, excimer laser light, or the like) from an illumination optical system 1 for exposure is used. ) Illuminates the pattern area of the reticle R, and the exposure light IL
, The pattern of the reticle R is projected and exposed through the projection optical system PL onto each shot area of the wafer W coated with the photoresist at a projection magnification β (β is, for example, ４ ，, ５, etc.). You. In the following description, the Z axis is taken parallel to the optical axis AX of the projection optical system PL, the X axis is taken parallel to the plane of FIG. 5 on a plane perpendicular to the Z axis, and the Y axis is taken perpendicular to the plane of FIG. I do.

The reticle R is an optical axis AX of the projection optical system PL.
The reticle stage 2 is mounted on a reticle stage 2 which is two-dimensionally movable and rotatable in a plane perpendicular to the reticle stage. On the other hand, the wafer W is held on the Z stage 3 via a wafer holder (not shown). The Z stage 3 is a position (focus position) of the wafer W in the Z direction.
, For example, in a range of several μm to several tens μm. The Z stage 3 is fixed on an XY stage 4, and the XY stage 4 positions the Z stage 3 (wafer W) in the X and Y directions. XY measured by moving mirror 5m and laser interferometer 5 on Z stage 3
The position of the stage 4 is supplied to the main control system 6, and the main control system 6
Controls the operation of the XY stage 4 via a drive system 7 of, for example, a feed screw system based on the position information. The Z stage 3 and the XY stage 4 constitute a wafer stage.

Further, a non-photosensitive detection light is applied to a photoresist for forming a slit image or the like on the surface of the wafer W obliquely to the optical axis AX on the side surface of the projection optical system PL. An irradiation optical system 8a and a light receiving optical system 8b that receives reflected light from the slit image or the like, re-images the slit image on, for example, a vibration slit, and receives detection light passing through the vibration slit. And an oblique incidence type focus position detection system (hereinafter, referred to as an “AF sensor”) 8. From the light receiving optical system 8b, a focus signal that changes in proportion to the amount of deviation of the focus position on the surface of the wafer W from the image plane (best focus position) within a predetermined range is output, and this focus signal is output to the main control system. 9.

For example, in the initial state, adjustment is performed so that the focus signal becomes 0 while the surface of the wafer W is focused on the image plane, and the main control system 9 maintains the focus signal at 0. The Z stage 3 is driven as described above. Thus, the surface of the wafer W is always focused on the image plane by the autofocus method. In the light receiving optical system 8b, a parallel flat glass (plane parallel) whose angle changes according to a command from the main control system 9 is provided, and by changing the angle of the parallel flat glass, the focus signal is changed. Can be provided with a desired offset.
At the time of exposure, when the exposure of the pattern image of the reticle R to one shot area on the wafer W is completed in such a state where the autofocus operation is performed, the next shot area on the wafer W is moved by the stepping of the XY stage 4. Step-and-
Exposure is repeated in a repeat mode.

In the projection exposure apparatus of this embodiment, the wafer W
An off-axis imaging type alignment sensor 9 for detecting the position of an alignment mark (wafer mark) formed in each of the above shot areas is provided. In the alignment sensor 9, the illumination light AL that is insensitive to the photoresist from a broadband light source 10 such as a halogen lamp passes through a condenser lens 11, a half prism 12 and an objective lens 13, Light the mark. Then, the illumination light AL from the test mark passes through an objective lens 13, a half prism 12, and an imaging lens 14, and a two-dimensional image sensor 15 such as a CCD.
The image of the test mark is formed on the imaging surface of
Is supplied to the alignment signal processing system 16.

At the time of normal alignment, the alignment signal processing system 16 processes the image signal S to determine the amount of displacement of the test mark with respect to a predetermined reference position, and supplies the amount of displacement to the main control system 6. . In the main control system 6,
By adding the coordinates of the Z stage 3 (wafer W) measured by the laser interferometer 5 to the positional deviation amount, the position of the test mark can be obtained, and from this result, the wafer W with respect to the pattern image of the reticle R can be obtained. Can be accurately positioned. Further, in this example, the alignment sensor 16 is also used to measure the evaluation mark when calibrating the best focus position of the projection optical system PL. The block diagram in the alignment signal processing system 16 of FIG. 5 is a functional block diagram when measuring the evaluation mark (details will be described later).

As described above, in this embodiment, the use of the AF sensor 8 allows the surface of the wafer W to be focused on the image plane of the projection optical system PL by the autofocus method. However, the AF sensor 8
There is a possibility that a deviation may occur between the best focus position detected at (2) (the position where the focus signal becomes 0) and the actual best focus position. Therefore, for example, calibration of the best focus position is periodically performed as follows. That is, a test reticle on which a predetermined evaluation mark is formed is used as the reticle R in FIG.
A series of shot areas on an unexposed silicon wafer (hereinafter referred to as a wafer W) coated with a photoresist are sequentially exposed to an image of an evaluation mark of the test reticle.
At this time, based on the detection result of the AF sensor 8, the focus position of the wafer W is sequentially changed in a predetermined step for each shot area. Thereafter, when the wafer W is developed, a resist pattern of the image of the evaluation mark is formed in each shot area, and the wafer W after the development is mounted on the Z stage 3 in FIG. 5 again.

FIG. 6 shows the wafer W placed on the Z stage 3 after development. In FIG. 6, the surface of the wafer W has shot areas 21 at predetermined pitches in the X and Y directions.
A, 21B, 21C,..., 21U are arranged, and as typically shown by a shot area 21, a resist pattern of an image of an evaluation mark (hereinafter, referred to as “shot pattern”)
These are also called “evaluation marks”) WM and WMA are formed. However, since the shot areas 21A to 21U are exposed at different focus positions, the evaluation marks WM and WMA are exposed at a higher resolution as the shot area is exposed closer to the best focus position. One evaluation mark WM has a basic pattern of Y
Marks arranged in the same direction, and the other evaluation mark WM
A is a mark obtained by rotating the evaluation mark WM by 90 °. The difference between the best focus positions obtained for the two evaluation marks becomes astigmatism. Hereinafter, the former case of measuring the evaluation mark WM will be described.

FIG. 1A shows an evaluation mark WM formed in the shot area 21 of FIG.
In (a), the evaluation mark WM is a mark in which the same mark group 32A, 32B, 32C is formed at a pitch PL in the Y direction, and each mark group 32A, 32B, 32C is in the X direction as a basic pattern. This is formed by forming four elongated rhombic marks (hereinafter referred to as “wedge-shaped marks”) 31 at a pitch PS in the Y direction. That is, the evaluation mark WM is a periodic mark in which the mark groups 32A, 32B, and 32C in which the wedge-shaped marks 31 are arranged in the Y-direction are further arranged in the Y-direction. L is, for example, about 5 to 15 μm at the best focus position, and the width of the wedge-shaped mark 31 in the short direction (Y direction in FIG. 1A) is about 0.5 μm. The pitch PS of the arrangement of the wedge-shaped marks 31 is about 1 μm, and the pitch PL of the arrangement of the mark groups 32A, 32B, 32C is about 8 μm. However, the number of the mark groups 32A, 32B, 32C is not limited to three, but may be seven to eight, or even more.

As described above, the evaluation mark WM is composed of a plurality of mark groups 32A, 32B, 32C because the diffraction light is generated from the evaluation mark WM by the periodicity of the pitch PL, so that the conventional laser WM is used. This is to make it possible to detect even with a scan type sensor. Therefore, if the detection is performed only by the imaging type alignment sensor 9 of the present example, there is no need to arrange a plurality of mark groups 32A, 32B and the like, and a large number of wedge-shaped marks 31 are simply formed at a pitch PS. An evaluation mark formed of books (for example, several tens) can be used.

When the wedge-shaped mark 31 is exposed in the vicinity of the best focus position, its length in the longitudinal direction becomes L1 as shown by the solid line in FIG. If the exposure is performed at
As shown by the dotted line mark 31D, the length in the longitudinal direction is L2.
By measuring the length L in the longitudinal direction, an accurate best focus position under the exposure light can be obtained. Therefore, in FIG.
The periodic direction of the evaluation mark WM is the Y direction, and the measurement target is the average value of the length L of each wedge-shaped mark 31 in the longitudinal direction. Therefore, the measurement direction is the X direction.

Therefore, the image of the evaluation mark WM is picked up by the image pickup device 15 in the alignment sensor 9 shown in FIG. FIG. 2 shows an image WMI of the evaluation mark WM imaged in this manner. The X-direction (measurement direction) and the Y-direction (period) in FIG. direction)
Are defined as an x direction and a y direction, respectively. Further, on the imaging surface of the imaging element 15, it is assumed that a number of pixels are arranged at a pitch EX in the x direction and a pitch EY in the Y direction, and the electrical scanning direction when reading an image signal from each pixel is x direction ( Measurement direction). A series of many pixel rows arranged in the x direction are divided into scanning lines S ₁ , S ₂ ,.
Then, the respective scanning lines S _j (j = 1, 2,...) Of the image sensor 15 are arranged in parallel to the longitudinal direction (measurement direction) of the image 31I of each wedge-shaped mark 31 constituting the evaluation mark image WMI. become.

Next, returning to FIG. 5, a method of processing the image signal S from the image sensor 15 will be described. First, the main control system 6
Issues a command to the control unit 20 in the alignment signal processing system 16 to take in the image signal S of the image sensor 15 from the alignment signal processing system 16.
Is stored in the image memory 17. Thereafter, the image signal S in the image memory 17 is supplied to the mark extracting unit 18 for each image signal of each scanning line _{Sj in} FIG.
Each 8, for each image signal of the scanning line S _j determines whether parts are present wedge-shaped mark image 31I.

The dotted curve in FIG. 3 shows an example of the image signal of the scanning line _Sj . In FIG. 3, the horizontal axis represents the position of the image pickup surface in the x direction, and the vertical axis represents the level of the image signal. Is shown. In this example, the image signal of the portion where the wedge-shaped mark image exists is small (dark), and the image signal of the portion where the wedge-shaped mark image is not present is large (bright). _Assuming that the minimum level of the image signal is S _jmin and the maximum level is S _jmax , the mark extraction unit 18 calculates the contrast C _{j of the} image signal as follows, for example.

Contrast C _j = (S _jmax −S _jmin ) / (S _jmax + S _jmin ) (1) A wedge-shaped mark image exists on the scanning line S _j where the contrast C _j is equal to or more than a predetermined value C _th. Then, it is determined that no wedge-shaped mark image exists for the other scanning lines _Sj . When such a determination is made, only the scanning lines R ₁ , R ₂ ,..., R ₈ where the wedge-shaped mark image 31I exists are extracted from the scanning lines S _j as shown in FIG. The mark part extraction unit 18 outputs the scanning line R _i (i = 1 to 1) thus extracted.
The image signal of 8) is supplied to the integration measuring unit 19 of FIG.

As another method for extracting the scanning line R _{i in} which the wedge-shaped mark image exists, the image signal shown in FIG. 3 is divided at a predetermined slice level SL lower than the maximum level S _jmax . the image signals may be obtained length L _j of the x-direction of the lower range than the slice level SL. The length L _j calculated for all the scan lines S _j
Of the maximum value and L _max, for example k · L _max or more scan lines S _j with length L _j is a predetermined coefficient k (k is, for example, 0.8) is present wedge mark image, it Otherwise, it may be determined that the wedge-shaped mark image does not exist.

Next, the integrated measuring unit 19, by integrating the image signals of all the scanning lines R _i of the wedge-type mark image is present in the evaluation mark images WMI period direction (y direction in FIG. 2), integration An image signal ΣR is generated, and the integrated image signal ΣR is divided at a predetermined slice level SL. The solid line curve in FIG. 3 shows the integrated image signal ΔR (however, the value on the vertical axis is divided by the number of integrations and normalized).
The integrated measuring unit 19 is configured to output the maximum level TL of the integrated image signal ΣR.
(In this case, the signal level at a position where there is no wedge-shaped mark image) is obtained, and the slice level SL is set at, for example, 70% of the maximum level TL. Then, x in a range where the integrated image signal ΣR is equal to or less than the slice level SL.
The length LM in the direction is obtained as the length of the mark image. The image sensor 15 of the alignment sensor 9 shown in FIG.
LM · α obtained by multiplying the wafer W by the magnification α
Is substantially equal to the mark length L of the evaluation mark WM in FIG.

However, the measured value LM · α does not exactly coincide with the length L of the evaluation mark WM, and is determined by the slice level SL and a predetermined number of optical systems of the imaging type alignment sensor 9. The value is shifted by the offset ΔL.
In order to determine the offset ΔL, in this example, the length L of the evaluation mark WM is first or periodically measured using, for example, a scanning electron microscope (SEM) or the like. The difference from the mark length LM · α measured is calculated. And the offset Δ
L is stored in the integration measuring unit 19. Integrated measurement unit 1
9, the mark length LM measured from the integrated image signal ΣR
The length obtained by subtracting the offset ΔL from α is supplied to the control unit 20 as a mark length, and the control unit 20 supplies this mark length to the main control system 6. By doing so, it goes without saying that the measurement time is significantly shortened as compared with the case where an SEM or the like is used for each measurement.

The evaluation marks W formed in the respective shot areas 21A to 21U on the wafer W in FIG.
The mark length of M is measured and sequentially supplied to the main control system 6,
The main control system 6 determines, for example, the focus position in the shot area having the longest mark length as the best focus position, and performs adjustment so that the focus signal of the AF sensor 8 becomes 0 at this best focus position. by this,
The calibration of the best focus position of the projection optical system PL is completed.

As described above, in the present embodiment, the scanning line R _{i on} which the wedge-shaped mark image exists is extracted, and the mark length is measured based on the image signal obtained by integrating these scanning lines R _i .
In FIG. 2, the image signal in the range where the wedge-shaped mark image 31I does not exist is removed, and the S / N ratio of the integrated image signal to be detected is improved. Therefore, the mark length can be measured with high accuracy, and the calibration of the best focus position can be performed with high accuracy. This makes it possible to use an evaluation mark composed of a wedge-shaped mark even in an exposure apparatus that is not equipped with a laser scan sensor, and to eliminate the laser scan sensor from the exposure apparatus. Space can be saved.

Incidentally, depending on the positional relationship between each scanning line _Sj and the wedge-shaped mark image 31I in FIG.
The center in the short direction of 1I may be located at the boundary between two scanning lines, causing an error in the measured value of the mark length. FIG. 4A shows such a state, in which the center of one wedge-shaped mark image 31I has two scanning lines S _j and S in the y direction.
It is located at the boundary of _{j + 1} . In consideration of such a case, in the main control system 6 in FIG. 5, the evaluation mark image WMI is once imaged by the image sensor 15 in the state of FIG. after performing the integration, by driving the XY stage 4 as shown in FIG. 4 (b), the array pitch EY scan lines S _j and the evaluation mark image WMI and the imaging element 15 in the y direction (the period direction) Relative to each other. Then, the main control system 6 causes the image pickup device 15 to capture the evaluation mark image WMI even in the state of FIG. 4B and causes the image signal to be extracted and integrated as described above. The mark length is calculated based on the two integrated image signals ΔR. Thereby, the mark length can always be measured with high accuracy. Note that the image signal of the mark portion may be extracted and integrated with respect to the two original image signals after the two imagings.

The sequence of the relative movement and the imaging is not limited to two times, but the relative movement and the imaging are divided into n times (n is a natural number). The use mark image WMI may be moved by 1 / n of the arrangement pitch EY of the scanning lines. On the other hand, particularly when a CCD is used as the image sensor 15,
In FIG. 2, each pixel of the CCD and the evaluation mark image WMI
May cause the edge of the wedge-shaped mark image 31I in the measurement direction (x direction) to coincide with the boundary of the immediately adjacent pixel, thereby causing an error of about one pixel in the length measurement value. There is also. In order to avoid this, the XY stage 4 is driven by the command of the main control system 6 and the wafer W is moved in the measurement direction (X direction) in the same manner as described above. Pitch E of each pixel
The mark length may be calculated from the combined signal by moving the image by の of X, again performing the imaging and the integration of the image signal. Also in this case, the imaging and the integration are not limited to two times, and the imaging and the integration are divided into m times (m is a natural number), and the image sensor 15 and the XY stage 4 are set to the pixel pitch of 1 at each relative movement. / M.

In the above-described embodiment, the present invention is applied to the case where the calibration of the best focus position is performed. Inclined line
The present invention can also be applied to a case in which an image of an and space pattern is superposed and exposed, and the length of a wedge-shaped mark formed as a moire fringe as a result of this exposure is measured. Also in this case, since the mark length can be measured with high accuracy, the overlay accuracy can be evaluated with high accuracy as a result.

It should be noted that the present invention is not limited to the above-described embodiment, and is applicable to a case where mark measurement is performed by a scanning exposure type projection exposure apparatus such as a step-and-scan method. Of course, various configurations can be taken without departing from the scope.

[0039]

According to the mark measuring method of the present invention, the image signal of the non-mark portion having no periodic mark image is excluded from the object of integration, and the mark length is obtained only by the image signal of the mark portion. Therefore, there is an advantage that the SN ratio of the used image signal is improved, and the mark length can be obtained with high accuracy by the image processing method. This makes it possible to measure the length of a periodic mark such as a wedge-shaped mark with high accuracy even using an imaging sensor, which was previously possible only with a laser scan sensor. It becomes possible.

Further, when extracting the image signal of the portion having the image of the periodic mark in the periodic direction of the image of the periodic mark, the image signal is extracted by a plurality of pixels arranged at a predetermined pitch in the periodic direction of the image of the periodic mark. Split into image signals from the pixels in the column,
A signal in which the contrast between the maximum level and the minimum level among the image signals from the pixels in the plurality of columns exceeds a predetermined threshold,
Alternatively, when extracting a signal whose signal level is within a predetermined slice level range over a predetermined length, it is possible to quantitatively and easily extract an image signal of the mark portion.

Further, the image of the periodic mark and the image pickup device are sequentially arranged in at least one of the periodic direction of the periodic mark and the measurement direction intersecting with the periodic direction by 1 / n of the arrangement pitch of the pixels of the image pickup device (n is The operation of picking up an image of a periodic mark is repeated n times while being relatively shifted by an integer of 2 or more). From the n sets of image signals obtained by this repetition, the image in the periodic direction of the image of the periodic mark is obtained. When extracting the image signal of the portion having the image of the periodic mark, even if the contour of the periodic mark image is located at the boundary between the pixels of the image sensor at the time of the first imaging, the image is finally obtained with high accuracy. Mark length can be measured. According to the mark measuring device of the present invention, the mark measuring method of the present invention can be implemented.

[Brief description of the drawings]

FIG. 1A is an enlarged plan view showing an evaluation mark WM to be measured in an example of an embodiment of the present invention, and FIG. 1B shows a state in which the mark length of a wedge-shaped mark 31 changes depending on a focus position. FIG.

FIG. 2 is a diagram showing a state in which an image of an evaluation mark WM in FIG. 1 is projected on an image sensor 15.

FIG. 3 is a diagram showing an image signal of a scanning line _Sj and an integrated image signal ΔR of FIG. 2;

FIG. 4 is an explanatory diagram of a case in which an image of an evaluation mark WM and an image sensor 15 are repeatedly shifted in a periodic direction and image capturing is repeated.

FIG. 5 is a schematic configuration diagram showing a projection exposure apparatus used in an example of an embodiment of the present invention.

FIG. 6 is a plan view showing an arrangement of shot areas and evaluation marks exposed on a wafer W;

[Explanation of symbols]

R Reticle PL Projection optical system W Wafer 4 XY stage 6 Main control system 8 AF sensor 9 Imaging type alignment sensor 15 Image sensor S _j scan line 16 Alignment signal processing system 18 Mark extraction unit 19 Integration measurement unit WM evaluation mark 31 wedge-shaped mark 32A, 32B, 32C mark group WMI evaluation mark image

Claims (4)

[Claims] 1. An image of a periodic mark formed on a substrate to be processed is imaged via a predetermined image sensor, and an image signal obtained by the imaging is processed to measure the shape of the periodic mark. In the mark measurement method, among the image signals, an image signal of a portion where the image of the periodic mark is present in the periodic direction of the image of the periodic mark is extracted, and the extracted image signal is extracted from the periodic mark. A mark measuring method, wherein the mark length is integrated in a period direction of an image, and a mark length in a measurement direction intersecting with the period direction of the image of the periodic mark is obtained from the integrated image signal. 2. The mark measuring method according to claim 1, wherein, when extracting an image signal of a portion where the image of the periodic mark is present in a periodic direction of the image of the periodic mark, the image signal is extracted. The image is divided into image signals from a plurality of columns of pixels arranged at a predetermined pitch in the period direction of the image of the periodic mark, and the contrast between the maximum level and the minimum level among the image signals from the plurality of columns of pixels is predetermined. A mark measurement method comprising: extracting a signal exceeding a threshold value or a signal whose signal level is within a predetermined slice level over a predetermined length or more. 3. The mark measuring method according to claim 1, wherein the image of the periodic mark and the image sensor are arranged in at least one of a periodic direction of the periodic mark or a measurement direction intersecting the periodic direction. Sequentially shifted relative to each other by 1 / n (n is an integer of 2 or more) of the arrangement pitch of the pixels of the image sensor,
The operation of picking up the image of the periodic mark is repeated n times. From the n sets of image signals obtained by the repetition, the image signal of the portion having the image of the periodic mark in the periodic direction of the image of the periodic mark is obtained. A mark measurement method characterized by extracting a mark. 4. An image forming optical system that forms an image of a periodic mark formed on a substrate to be processed, and an image sensor that captures an image of the periodic mark, wherein an image from the image sensor is provided. In a mark measuring device that processes a signal to measure the shape of the periodic mark, a periodic direction of the image of the periodic mark and the imaging device, or a measurement direction that intersects the periodic direction of the image of the periodic mark. A relative movement device that relatively moves to at least one of the following: and control for causing the image sensor to capture the image of the periodic mark each time the image of the periodic mark and the image sensor are relatively moved via the relative movement device. Part, from among the image signals obtained by the imaging device, extracts an image signal of a portion where the image of the periodic mark is present in the periodic direction of the image of the periodic mark, and converts the extracted image signal to the period Target A mark measuring device comprising: a calculating unit that integrates in the periodic direction of a mark image and obtains a mark length of the periodic mark image in the measurement direction from the integrated image signal.