JP2014175330A - Alignment method for substrate - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an alignment method which, even when alignment marks of two flat substrates cannot be simultaneously imaged within the same screen with the same alignment scope, can correctly align the two flat substrates in a plane direction.SOLUTION: An alignment method includes: preparing a calibration substrate in which reference alignment marks corresponding to a correct position relationship of two substrates are drawn so as to form a pair between on the top surface and on the rear surface; imaging the reference alignment marks on the top surface and the rear surface of the calibration substrate with an alignment scope; setting a virtual reference coordinate system to determine positions of the pair of reference alignment marks on the reference coordinate system; imaging a pair of real alignment marks on the two substrates with the same alignment scope; and determining positions in a plane direction of the two substrates so as to match the positions of the pair of real alignment marks with the positions of the pair of reference alignment marks, respectively.

Description

  The present invention relates to an adjustment (alignment method) for positioning two flat substrates, for example, a wafer and a mask, or two wafers in a planar direction.

  In the manufacture of semiconductor devices, the mask and wafer (two substrates) must be positioned in the planar direction before being exposed to the wafer through the mask. Alignment marks are formed on both the mask and the wafer, and alignment is the operation of determining the position in the plane direction of the substrate so that both alignment marks are correctly positioned on a straight line perpendicular to the plane direction of the substrate. . The present invention is directed to an alignment method for determining the position of two substrates in the planar direction.

  Conventionally, such alignment is performed by placing the mask and the wafer alignment mark, which have been moved to a predetermined position, on the same screen of a single alignment scope and photographing them at the same time. At least one of them was moved (Patent Document 1).

JP 2007-512694 A

  This conventional method is an alignment method established on the premise that two alignment marks can be photographed simultaneously. However, there are cases where two alignment marks cannot be photographed simultaneously. For example, the alignment mark of the wafer is attached to the surface opposite to the mask facing surface, or the two substrates are opaque. In this case, conventionally, the alignment scope corresponding to each of the alignment scopes is separately moved along the optical axis and focused on the corresponding alignment mark to capture the image of the alignment mark. However, in this case, since the alignment scope is moved along the optical axis, an error has occurred in reading the position of the alignment mark.

  Therefore, the present invention correctly aligns the positions of the two flat substrates in the planar direction even when the alignment marks on the two flat substrates cannot be photographed simultaneously on the same screen with the same alignment scope ( An object is to obtain an alignment method capable of aligning). Another object of the present invention is to obtain a method capable of performing alignment with high accuracy without moving the alignment scope during the alignment step.

  The present invention prepares a calibration board on which an alignment mark that forms a pair of correct positional relations between two boards is prepared on the front and back, and images the alignment mark that forms the pair of the calibration board on the front and back with an alignment scope to obtain reference coordinates. By creating a system, the position of a pair of alignment marks on the reference coordinate system is determined, and then the alignment marks forming a pair of two substrates are photographed with the same alignment scope to form a pair on the reference coordinate system If it matches the mark, it has been completed based on the viewpoint that alignment can be performed correctly even if the alignment marks on the two substrates cannot be photographed simultaneously.

  That is, the present invention is a substrate alignment method for determining the position in the planar direction of two flat substrates each having at least one alignment mark for mutual alignment and arranged in parallel to each other. Preparing a calibration substrate having a reference alignment mark that forms a pair of correct positional relationship between the two substrates; photographing a reference alignment mark that forms a pair of front and back surfaces of the calibration substrate with an alignment scope to obtain a reference coordinate system; Making a pair of reference alignment marks on the reference coordinate system; removing the calibration substrate; photographing the actual alignment mark forming the pair of the two substrates with the same alignment scope; The position of the paired actual alignment mark is the position of the paired reference alignment mark on the reference coordinate system. It is characterized by having; a step determining the position of the plane direction of the two substrates to match.

  In a preferred aspect of the present invention, the calibration substrate is a transparent substrate, and the alignment scope is located on one side of the front and back of the calibration substrate and simultaneously focuses on the front and back reference alignment marks. Consists of.

  In the substrate alignment method of the present invention, the front and back reference alignment marks are simultaneously photographed by the bifocal alignment scope, and the positions of the two pairs of reference alignment marks on the reference coordinate system are stored in a storage means; Photographing the actual alignment marks forming a pair of the two substrates sequentially with the bifocal alignment scope and storing the positions of the actual alignment marks on the reference coordinate system in the storage means; Determining the position in the planar direction of the two substrates so that the positions of the alignment marks on the reference coordinate system respectively match the positions of the paired reference alignment marks stored in the storage means on the reference coordinate system; It is practical to include.

  In another aspect of the present invention, the alignment scope may be composed of a pair of fixed alignment scopes that are located on both sides of the calibration board and that respectively photograph the front and back reference alignment marks. In this embodiment, the calibration substrate may be transparent or opaque.

  Further, in another aspect of the present invention, the reference alignment marks on the front and back sides are respectively photographed by the pair of fixed alignment scopes, and the positions of the two pairs of reference alignment marks on the reference coordinate system are stored in the storage means. The actual alignment mark forming the pair of the two substrates is photographed by the pair of fixed alignment scopes, and the position of the actual alignment mark forming the pair on the reference coordinate system is stored in the storage means. Determining a position in the plane direction of the two substrates so as to respectively match the positions of the paired reference alignment marks on the reference coordinate system.

  According to the present invention, in manufacturing a semiconductor device, a calibration substrate is prepared by drawing a reference alignment mark having a correct positional relationship between two substrates on the front and back, and the reference alignment mark on the front and back of the calibration substrate is photographed with a fixed alignment scope. Then, set the reference coordinate system to determine the position of the two reference alignment marks on the reference coordinate system, and then photograph the actual alignment marks formed on the two substrates with the same fixed alignment scope. The positions of the two substrates in the plane direction are determined so that the positions of the two actual alignment marks coincide with the positions of the two reference alignment marks on the reference coordinate system. Even if the image cannot be taken, the two substrates can be correctly positioned in the plane direction.

  In addition, according to the method of the present invention, the calibration substrate is placed in the field of alignment scope, and the paired alignment marks are photographed and removed. Then, the two substrates are placed in the field of alignment scope and aligned to align each other. There is no need to move the alignment scope until it is fixed, and alignment errors due to the movement of the alignment scope do not occur.

One Embodiment of the calibration board | substrate used for the method of this invention is shown, (A) is a top view, (B) is sectional drawing. The alignment mark of the two board | substrates to which this invention method is applied is shown, (A) is a top view, (B) is sectional drawing. It is a figure which shows typically the different process in 1st embodiment of the alignment method by this invention. It is the longitudinal cross-sectional view cut | disconnected by the surface which passes along the imaging | photography optical axis which shows the structure of the bifocal alignment scope used in the 1st embodiment. It is an optical system block diagram of the bifocal alignment scope of the first embodiment. (A) is a figure which shows the Example of an alignment mark of a calibration board | substrate with which this invention method is applied, and two board | substrates, (B) thru | or (E) show the mode of the alignment mark image image | photographed with the bifocal alignment scope. FIG. It is a figure which shows typically the different process in 2nd embodiment of the alignment method by this invention.

  FIG. 1 is a schematic view of an embodiment of a calibration substrate 10 used in the method of the present invention. The calibration substrate 10 is for gap exposure in which two aligned substrates are exposed with a predetermined distance in a direction orthogonal to the substrate plane. The calibration substrate 10 is a transparent plate (parallel plane glass substrate) composed of a precise parallel plane, and reference alignment marks 11 and 12 that are paired on the front and back are drawn by, for example, vapor deposition. In this embodiment, the reference alignment mark 11 is a circle (hollow circle), and the reference alignment mark 12 has a smaller diameter than the inner diameter of the reference alignment mark 11 when the calibration substrate 10 is vertically projected onto a plane parallel to the front and back surfaces. It consists of a spot (solid circle) whose center is located at its correct center. Since the reference alignment marks 11 and 12 are drawn by vapor deposition, the thickness thereof can be ignored (positioned within the depth of field of the alignment scope). The reference alignment marks 11 and 12 have shapes and sizes that can be clearly distinguished when displayed on the same screen, and it is sufficient that the reference point such as the center can be accurately identified with high accuracy. The shape is not limited to circles, solid circles, and small-diameter spots.

  On the other hand, as shown in FIG. 2, two substrates to be aligned, for example, the mask 20 and the wafer 30, have the same shape as the reference alignment marks 11 and 12 on one side (the same side), in the illustrated embodiment, on the back side. Real alignment marks 11X and 12X are drawn. When exposed, the mask 20 and the wafer 30 are arranged such that the distance between the actual alignment marks 11X and 12X (the distance in the direction orthogonal to the substrate) becomes a predetermined value. The thickness t of the calibration substrate 10 (that is, the distance in the direction perpendicular to the reference alignment marks 11 and 12) is the distance between the actual alignment mark 11X of the mask 20 and the actual alignment mark 12X of the wafer 30 (substrate) when exposed. Is set equal to the distance d) in the direction orthogonal to The calibration substrate 10 may be formed of glass or silicon material. The thickness of the wafer 30 is generally about 425 μm to 750 μm.

  Since at least two sets of actual alignment marks 11X and 12X are provided on one mask 20 and wafer 30, the reference alignment marks 11 and 12 correspond to at least the actual alignment marks 11X and 12X on the calibration substrate 10, respectively. Two sets are provided. The reference alignment marks 11 and 12 are preferably the same shape (congruent) as the actual alignment marks 11X and 12X, but may be different if the positional relationship in the direction orthogonal to the substrate is exactly the same.

  FIG. 3 shows a first embodiment of the alignment method according to the present invention. The alignment scope 40 is located on one of the front and back sides of the transparent calibration substrate 10 and includes a bifocal (double focus) optical system 40A that can photograph the reference alignment marks 11 and 12 simultaneously. Usually, two bifocal alignment scopes 40A are prepared and arranged on the same side of the front and back of the calibration substrate 10 at positions where the actual alignment marks 11X and 12X of the alignment target mask 20 and the wafer 30 can be photographed.

  FIG. 4 shows a specific configuration example of the bifocal alignment scope 40A, and FIG. 5 shows a lens configuration of two photographing optical systems having different focal points of the bifocal alignment scope 40A. The bifocal alignment scope 40 </ b> A includes two imaging optical systems with different object-side focal positions (focus positions) in the lens barrel 41. That is, the objective lens group L1, the first beam splitter BS1, the second beam splitter BS2, the second lens group L2, and the focus adjustment lens group L3 as the first imaging optical system in order from the measurement target (mask 20 and wafer 30) side. , A third beam splitter BS3 and an image sensor 42, and as a second imaging optical system, an objective lens group L1, a first beam splitter BS1, a first mirror M1, a second lens group L4, a second mirror M2, a third beam A splitter BS3 and an image sensor 42 are provided, and the images of the reference alignment marks 11 and 12 on the calibration substrate 10 are simultaneously formed on the image pickup surface 42A of the image sensor 42 by the first and second imaging optical systems. it can. The first and second imaging optical systems share the objective lens group L1, the first beam splitter BS1, the third beam splitter BS3, and the image sensor 42. FIG. 5A shows the lens configuration of the first imaging optical system, and FIG. 5B shows the lens configuration in which the first and second mirrors M1 and M2 of the second imaging optical system are omitted. The first photographing optical system is an inner focus lens provided with a focus adjusting mechanism that moves the focus adjusting lens group L3 in the optical axis direction and focuses the object to be measured. The second photographing optical system is a two-focus alignment scope 40A. This is a fixed focus lens that adjusts the focus by moving the entire lens in the optical axis direction of the objective lens unit L1. Illumination light enters the second beam splitter BS2 from outside the first imaging optical system. The illumination light is reflected by the second beam splitter BS2 in the direction of the first beam splitter BS1, reflected by the first beam splitter BS1 in the direction of the objective lens group L1, converged by the objective lens group L1, and then the calibration substrate 10 and the mask 20 The object to be measured such as the wafer 30 is illuminated.

  The bifocal alignment scope 40A is connected to an image processing device 43 such as a personal computer (FIG. 4). The image processing device 43 includes an image signal processing circuit 44 that performs a predetermined image signal processing on the image signal that has been imaged and output by the imaging element 42 of the bifocal alignment scope 40A, and converts the image signal into a predetermined signal. The monitor 45 for displaying the video by the video signal converted into the video signal by the above, the memory (storage means) 46 for storing the image data converted into the image data by the image signal processing circuit 44, and the image data as the data of the reference coordinate system An arithmetic circuit 47 that converts and calculates a shift of a specific image in the image data, a relative positional relationship, an alignment adjustment amount, and the like is provided. The bifocal alignment scope 40A can perform moving image shooting and still image shooting. The image processing device 43 can display a moving image captured by the bifocal alignment scope 40A on the monitor 45, and can capture a still image (store the image data in the memory 46) when switch means (not shown) is operated. .

  6A shows an example of the shape of the reference alignment marks 11 and 12 drawn on the calibration substrate 10, and FIG. 6B shows an image 11P of the reference alignment mark 11 photographed by the two-focus alignment scope 40A. An example of the shape (relative position) of the image 12P of the reference alignment mark 12 on the imaging surface 42A is shown. The reference alignment mark 11 drawn on the calibration substrate 10 is a correct circle centered on the center of the reference alignment mark 12. If the bifocal alignment scope 40A is an ideal optical system, the image 11P is a correct circular image centered on the center of the image 12P (FIG. 6B). On the other hand, the center of the image 11P of the reference alignment mark 11 and the center of the image 12P of the reference alignment mark 12 taken by the actual bifocal alignment scope 40A do not always coincide with each other. Rather, due to the error of the bifocal alignment scope 40A, there is always an error E in the plane direction of the substrate at the center position of the image 11P and the center position of the image 12P (FIG. 6C). In the conventional method, the alignment error occurs because the actual alignment mark 11X of the mask 20 and the actual alignment mark 12X of the wafer 30 are aligned without considering this error.

  In contrast, in the present embodiment, the error E is taken into account when the actual alignment mark 11X on the mask 20 and the actual alignment mark 12X on the wafer 30 are aligned. That is, the center position of the image 11XP when the real alignment mark 11X on the mask 20 is photographed by the bifocal alignment scope 40A and the center position of the image 12XP when the real alignment mark 12X on the wafer 30 is photographed are The position of the wafer 30 relative to the mask 20 is determined so as to include (cancel) the error E.

  In the present embodiment, a virtual XY orthogonal coordinate system having the intersection of the short side and the long side of the imaging surface 42A as the origin, the long side as the X axis, and the short side as the Y axis is set as the reference coordinate system, and the error is determined. A coordinate difference on the reference coordinate system is obtained by using E as a shift amount E between the center position (ideal center) of the image 11P and the center position (real center) of the image 12P. In this embodiment, it is assumed that Δx is shifted in the X-axis direction and Δy is shifted in the y-axis direction (FIG. 6C). These Δx and Δy constitute calibration data.

  Next, the real alignment marks 11X and 12X are photographed by the bifocal alignment scope 40A. FIG. 6D shows the positions of the images 11XP and 12XP of the actual alignment marks 11X and 12X on the imaging surface 42A. The center position of the image 12XP is shifted by Δx1 and Δy1 from the apparent center position of the image 11XP. Here, considering the error of the bifocal alignment scope 40A, the true center position of the image 11XP is shifted from the apparent center position by calibration data Δx and Δy. Therefore, the position of the wafer 30 is adjusted (aligned) so that the center position of the image 12XP matches the position shifted by the calibration data Δx and Δy from the apparent center position of the image 11XP ((FIG. 6E)).

  The setting of the reference coordinate system is not limited to this embodiment. For example, the center position of the image 11P is set as the origin, the X ′ axis is set in parallel with the longitudinal direction of the imaging surface 42A, and the Y ′ axis is set in parallel with the short side direction. Or a polar coordinate system, and an arbitrary coordinate system can be used.

  The alignment method of this embodiment will be further described with reference to FIG. The image processing device 43 (mainly the arithmetic circuit 47) executes processing and control of the image signal captured by the image sensor 42 below. An exposure apparatus (alignment apparatus) (not shown) to which this embodiment is applied includes a holding mechanism that holds the calibration substrate 10, the mask 20, and the wafer 30 in and out of a predetermined position, and at least the wafer 30 in a direction parallel to the substrate plane. It is a known apparatus provided with a precision stage that can be translated in parallel.

[Step 1]
First, the calibration substrate 10 is inserted into the exposure apparatus, and the reference alignment marks 11 and 12 on the calibration substrate 10 are focused by the two-focus alignment scope 40A, and the images 11P and 12P of the reference alignment marks 11 and 12 are simultaneously obtained. An image is formed on the image sensor 42, and the images 11P and 12P are captured and displayed on the monitor 45, and converted into image data (coordinate data) on the reference coordinate system and stored in the memory 46 (FIG. 3A). ).

[Step 2]
Next, the calibration substrate 10 is removed without moving the bifocal alignment scope 40A, and then the mask 20 is carried in, and the actual alignment mark 11X is viewed on the monitor 45 while checking the actual alignment mark 11X (in the imaging screen). ). Then, the actual alignment mark 11X is photographed by the bifocal alignment scope 40A, the image data of the image 11XP of the actual alignment mark 11X is stored in the memory 46, and the arithmetic circuit 47 calculates the center position of the image 11XP on the reference coordinate system. And stored in the memory 46 (FIG. 3B).

[Step 3]
Next, the wafer 30 is loaded without moving the bifocal alignment scope 40A and the mask 20, and the actual alignment mark 12X is placed in the field of view (in the imaging screen) of the bifocal alignment scope 40A while being confirmed by the monitor 45. Then, the real alignment mark 12X is photographed by the two-focus alignment scope 40A, the image data of the image 12XP of the real alignment mark 12X is stored in the memory 46, and the arithmetic circuit 47 calculates the center position of the image 12XP on the reference coordinate system. And stored in the memory 46 (FIG. 3C).

  Then, the center position of the image 11XP of the actual alignment mark 11X previously stored is read from the memory 46, and the deviation amount of the center position of the image 12XP of the actual alignment mark 12X with respect to the center position of the image 11XP on the reference coordinate system is calculated. Then, the shift amount is corrected by the calibration data obtained in step 1, and the substrate plane at the center position of the image 12 XP of the actual alignment mark 12 X of the wafer 30 with reference to the center position of the image 11 XP of the actual alignment mark 11 X of the mask 20. The direction alignment adjustment amount is calculated.

  Images 11XP and 12XP of actual alignment marks 11X and 12X taken with the bifocal alignment scope 40A and images 11P and 12P of reference alignment marks 11 and 12 taken in step 2 and stored in the memory 46 are on the same reference coordinate system. You may display on the monitor 45 in the state piled up as an image. By this display, the user can visually recognize the shift state between the mask 20 and the wafer 30. At this time, the images 11XP and 12XP and the images 11P and 12P may be displayed with different colors or brightness so that they can be distinguished.

[Step 4]
Then, based on the calculated alignment adjustment amount, the precision stage is driven to perform alignment adjustment of the wafer 30 (FIG. 3D).

  When the alignment adjustment in step 4 is completed, the actual alignment mark 12X is imaged by the bifocal alignment scope 40A again in step 3, and the shift amount of the center position of the image 12XP relative to the center position of the image 11XP on the reference coordinate system is calculated. Then, this is calibrated with the calibration data, and the alignment adjustment amount is calculated. If the alignment adjustment amount is 0 or within the allowable error range, the alignment adjustment is terminated. If the alignment adjustment amount is not 0 or within the allowable error range, the processes in steps 3 and 4 are repeated within a predetermined number of times.

  When all of the above steps 1 to 4 are completed, the bifocal alignment scope 40A is removed, and the wafer 30 is exposed by a known exposure apparatus (not shown). The bifocal alignment scope 40A may not be removed unless it interferes with exposure.

  According to the first embodiment, the reference alignment marks 11 and 12 forming a pair corresponding to the correct positional relationship between the mask 20 to be aligned and the actual alignment marks 11X and 12X of the wafer 30 (two substrates) are drawn on the front and back sides. The calibration substrate 10 is prepared, and the reference alignment marks 11 and 12 on the front and back of the calibration substrate 10 are photographed with a two-focus alignment scope 40A (alignment scope), and a virtual reference coordinate system is set to establish two pairs. An error of the images 11P and 12P of the reference alignment marks 11 and 12 forming the reference coordinate system on the reference coordinate system, a step of removing the calibration substrate 10, a mask 20 and a wafer 30 (two substrates) pair The actual alignment marks 11X and 12X forming the same bifocal alignment scope 40A (alignment scanner) The center positions of the images 11XP and 12XP of the actual alignment marks 11X and 12X that were photographed first are stored in the memory 46 (storage means), and the centers of the stored images 11XP and 12XP on the reference coordinate system are stored. The position and the center position of the image 11XP, 12XP of the next photographed actual alignment mark 11X, 12X match the positional relationship of the images 11P, 12P of the reference alignment mark 11, 12 that make a pair on the reference coordinate system. As described above, the step of moving at least one of the two substrates in the plane direction to determine the position in the plane direction is used, so that the actual alignment marks 11X and 12X on the two substrates cannot be photographed simultaneously with the same alignment scope. Even if it exists, accurate alignment becomes possible. In addition, the first embodiment is a second focus alignment scope 40A obtained by photographing the reference alignment marks 11 and 12 on the calibration substrate 10, and thereafter the actual alignment marks 11X on the mask 20 and the wafer 30 are not adjusted and moved. Since 12X is photographed, no error occurs due to the focus adjustment and movement of the alignment scope.

  The above is the order of alignment by one two-focus alignment scope 40A. Actually, two sets of reference alignment marks 11, 12 and actual alignment marks 11X, 12X separated from each other are photographed by two two-focus alignment scopes 40A. Then, by calculating two sets of calibration data, not only the amount of deviation parallel to the X-axis and Y-axis directions of the wafer 30 but also the rotational direction deviation about the axis perpendicular to the substrate plane of the wafer 30 is obtained. The wafer 30 is aligned.

  FIG. 7 shows a second embodiment of the alignment method according to the present invention. This embodiment is particularly suitable when the calibration substrate 10 is made of an opaque material. The alignment scope 40 according to this embodiment includes a pair of fixed alignment scopes 40B and 40C that are positioned on both the front and back sides of the calibration substrate 10 and photograph the reference alignment marks 11 and 12 simultaneously from above and below. One fixed alignment scope 40B is an alignment scope having a known straight optical system that does not have a bending optical system (prism), and the other fixed alignment scope 40C is a known bending optical system having a bending optical system (prism). For example, the alignment scope includes the first imaging optical system of the first embodiment, and any alignment scope is known. The fixed alignment scopes 40B and 40C are connected to the same image processing apparatus 43. As with the first embodiment, at least two sets of fixed alignment scopes 40B and 40C are installed, and each is connected to the same image processing apparatus 43. The steps of the alignment method of the second embodiment will be described with reference to FIG. In practice, the two sets of fixed alignment scopes 40B and 40C are used for alignment. However, in order to simplify the description, one set of fixed alignment scopes 40B and 40C will be described.

[Step 2-1]
In the second embodiment, first, the calibration substrate 10 is inserted into the exposure apparatus, and the images 11P and 12P of the reference alignment marks 11 and 12 on the calibration substrate 10 are respectively captured by the fixed alignment scopes 40B and 40C. The image 11P and the image 12P are converted into image data and stored in the memory 46 (FIG. 7A). As in the first embodiment, the reference coordinate system is an orthogonal coordinate having a specific point (pixel) on the imaging surface of the image sensor 42 as an origin and an orthogonal axis parallel to two orthogonal sides of the imaging surface as an XY axis. System. Then, the arithmetic circuit 47 calculates the shift amount of the center position of the image 11P with respect to the center position of the image 12P on the reference coordinate system from the image data of the images 11P and 12P, and stores it in the memory 46 as calibration data on the reference coordinate system. To do.

[Step 2-2]
Next, without moving the fixed alignment scopes 40B and 40C, the calibration substrate 10 is removed, the mask 20 and the wafer 30 are carried in, and the actual alignment mark 11X is placed in the field of view of the fixed alignment scope 40B while checking with the monitor 45. The actual alignment mark 12X is placed in the field of view of the fixed alignment scope 40C. Then, the real alignment marks 11X and 12X are photographed with the fixed alignment scopes 40B and 40C (FIG. 7B), and the arithmetic circuit 47 calculates the reference coordinate system from the image data of the real alignment marks 11X and 12X and the image data 11XP and 12XP. The center positions of the upper images 11XP and 12XP are calculated. Further, the arithmetic circuit 47 calculates a deviation amount of the center position of the image 12XP with respect to the center position of the image 11XP, calibrates this deviation amount with the calibration data read from the memory 46, and based on the calibrated deviation amount of the wafer 30. An alignment adjustment amount in the substrate plane direction is calculated.

  Images 11XP and 12XP of the actual alignment marks 11X and 12X photographed by the fixed alignment scopes 40B and 40C and the images 11P and 12P of the reference alignment marks 11 and 12 photographed and stored in Step 2-1 are used in the same reference coordinate system. The image can be displayed on the monitor 45 in a state of being overlaid as the upper image. By this display, the alignment state of the mask 20 and the wafer 30 can be visually recognized. At this time, the images 11XP and 12XP and the images 11P and 12P may be displayed with different colors or brightness so that they can be distinguished.

[Step 2-3]
Based on the calculated alignment adjustment amount, the memory 46 drives and controls the precision stage to execute alignment adjustment of the wafer 30 (FIG. 7C).

  When the alignment adjustment in step 2-3 is completed, the real alignment marks 11X and 12X are photographed with the fixed alignment scopes 40B and 40C again by the processing in step 2-2, and the center positions of the images 11XP and 12XP on the reference coordinate system are determined. A calculation is performed to calculate a shift amount in which the center positions of the images 11XP and 12XP are shifted from the center positions of the images 11P and 12P. Based on the shift amount, an alignment adjustment amount in a direction parallel to the substrate plane of the wafer 30 is calculated. Calculate. The processes in steps 2-2 and 2-3 are repeated within a predetermined number of times until the alignment adjustment amount is 0 or within an allowable error range.

  When the above alignment adjustment is completed, the fixed alignment scope 40B and, if necessary, the fixed alignment scope 40C are retracted and exposed by the exposure apparatus.

  According to the second embodiment, the reference alignment marks 11 and 12 forming a pair corresponding to the correct positional relationship between the mask 20 to be aligned and the actual alignment marks 11X and 12X of the wafer 30 (two substrates) are drawn on the front and back sides. The prepared calibration substrate 10, and the reference alignment marks 11 and 12 on the front and back of the calibration substrate 10 are photographed with the fixed alignment scopes 40B and 40C (alignment scope), and a virtual reference coordinate system is set, so that two pairs The positions of the center positions of the image 11P and the image 12P of the reference alignment marks 11 and 12 forming the reference coordinate system, the step of removing the calibration substrate 10, the mask 20 and the wafer 30 (two substrates) The real alignment marks 11X and 12X forming a pair of the same fixed alignment scopes 40B and 40C Image), and the positional relationship of the center positions of the images 11XP and 12XP of the real alignment marks 11X and 12X forming a pair of the images 11P and 12P of the reference alignment marks 11 and 12 forming a pair on the reference coordinate system. Since there is a step of moving the at least one of the mask 20 and the wafer 30 (two substrates) in the plane direction so as to match the positional relationship of the center position, the position in the plane direction is determined. Even when the actual alignment marks 11X and 12X on the substrate) cannot be photographed simultaneously with the same alignment scope, accurate alignment is possible.

  In the above embodiment, the alignment adjustment is automatic adjustment, but it can be manually adjusted. In the case of manual adjustment, there is a method in which the calculated alignment adjustment amount is displayed on the monitor 45 and the operator manually operates the precision stage while viewing the alignment adjustment amount. As yet another manual adjustment method, the images 11P and 12P stored in the memory 46 are displayed on the monitor 45, and the captured images 11XP and 12XP are converted into the image 11XP and the image 11XP so that the image 11XP matches the image 11P. There is a method in which the 12XP is moved and displayed on the monitor 45, and the user operates the precision stage so that the image of the image 12XP matches the image of the image 12P while viewing the display on the monitor 45. The images 11P and 12P and the images 11XP and 12XP displayed on the monitor 45 are preferably displayed with different colors or brightness so that they can be identified on the screen of the monitor 45.

  The above embodiment is applied to an exposure apparatus used for manufacturing a semiconductor element. However, the present invention relates to alignment of a surface in manufacturing, assembly, inspection, etc. of a wafer, a liquid crystal panel, a touch panel, and a flat electronic component. Can be applied to devices that require As a combination of two members to be aligned, there are a substrate and a mask, glass and glass, a wafer and wafer, a wafer and glass, a wafer and mask, a glass and mask, and the like.

  As described above, according to the alignment method of the present invention, even if the two substrates to be aligned are opaque, the optical error of the alignment scope can be corrected by the calibration substrate, so that accurate alignment is possible. In the alignment method of the present invention, since the alignment scope is not moved after the calibration substrate is photographed by the alignment scope, an error accompanying the movement does not occur, and refocus adjustment is unnecessary, so that the alignment process time can be shortened. .

10 calibration substrate 11 12 reference alignment mark 11P 12P image 11X 12X actual alignment mark 11XP 12XP image 20 mask (substrate)
30 Wafer (substrate)
40 Alignment scope 40A Two-focus alignment scope (alignment scope)
40B 40C Fixed alignment scope (alignment scope)
42 Image sensor 42A Imaging surface 43 Image processing device 44 Image signal processing circuit 45 Monitor 46 Memory (storage means)
47 Arithmetic circuit

Claims (5)

A substrate alignment method for determining a planar position of two flat substrates each having at least one alignment mark for mutual alignment and arranged parallel to each other,
Preparing a calibration board on which a reference alignment mark that forms a pair of a correct positional relationship between the two boards is drawn on both sides,
Photographing the reference alignment marks on the front and back of the calibration board with an alignment scope, setting a virtual reference coordinate system, and determining the positions of the two pairs of reference alignment marks on the reference coordinate system;
Removing the calibration substrate;
The actual alignment marks forming the pair of the two substrates are photographed with the same alignment scope so that the positions of the actual alignment marks forming the pair coincide with the positions of the paired reference alignment marks on the reference coordinate system. Determining the planar position of the two substrates;
A substrate alignment method characterized by comprising: 2. The alignment method according to claim 1, wherein the calibration substrate is a transparent substrate, and the alignment scope is positioned on one side of the front and back of the calibration substrate and simultaneously focuses on the front and back reference alignment marks. An alignment method for a substrate made of a scope. The alignment method according to claim 2, wherein the front and back reference alignment marks are simultaneously photographed by the bifocal alignment scope, and the positions of the two pairs of reference alignment marks on the reference coordinate system are stored in a storage means; ,
The actual alignment marks forming the pair of the two substrates are sequentially photographed by the bifocal alignment scope, the position of the actual alignment mark on the reference coordinate system is stored in the storage means, and the actual alignment forming the pair Determining the position of the two substrates in the plane direction so that the position of the mark on the reference coordinate system matches the position of the pair of reference alignment marks on the reference coordinate system stored in the storage means. Alignment method. 2. The alignment method according to claim 1, wherein the alignment scope is located on both sides of the calibration substrate and includes a pair of fixed alignment scopes that respectively photograph the front and back reference alignment marks. According to a fourth aspect of the present invention, the front and back reference alignment marks are respectively photographed by the pair of fixed alignment scopes, and the positions of the two pairs of reference alignment marks on the reference coordinate system are stored in the storage means. Steps,
The reference coordinate system in which the actual alignment marks forming the pair of the two substrates are photographed by the pair of fixed alignment scopes, and the positions of the actual alignment marks forming the pair on the reference coordinate system are stored in the storage means An alignment method including a step of determining a position in a plane direction of two substrates so as to respectively coincide with the positions of the upper pair of reference alignment marks.

Images