KR100405398B1 - Alignment method - Google Patents

Abstract

The alignment of the search mark on the wafer is not restricted, and the wafer is aligned at high speed. The first alignment sensor is used to detect the positions of the first and second search marks 47A and 47B with respect to the first wafer, and the coordinate system based on the search marks is obtained based on the detection results. In the state where the first search mark 47A is detected by the sensor, the position of the street line region 70 detected by the second alignment sensor is stored. For the wafers after the second sheet, the position of the street line region 70 detected by the second alignment sensor is detected while the first search mark 47A is detected by the first alignment sensor. The coordinate system based on the search mark is obtained based on the amount of deviation from the stored position.

Description

Substrate positioning method {ALIGNMENT METHOD}

In the exposure apparatus used in the photolithography process for manufacturing a semiconductor element, an imaging element (CCD, etc.), a liquid crystal display element, or a thin film magnetic head, etc., for example, the pattern of a mask image is exposed on a photosensitive board | substrate. The present invention relates to a method for positioning a photosensitive substrate at the time, and is particularly suitable when applied to a rough positioning (prior alignment) of a photosensitive substrate on a stage of an exposure apparatus.

In an exposure apparatus such as an imaging exposure apparatus (stepper or the like) used for manufacturing a semiconductor element or a liquid crystal display element or the like, or an exposure apparatus of a proximity type, a circuit pattern formed on a reticle as a mask as a photosensitive substrate (or glass) In order to transfer to a photoresist layer on a plate or the like with high overlapping accuracy, it is desired to align (align) the reticle and the wafer with high accuracy.

As an alignment sensor for this purpose, as disclosed in Japanese Patent Laid-Open No. 5-21314, a laser beam is irradiated to an alignment mark on a dot string on a wafer, and the position of the mark using light diffracted or scattered by the mark. LSA (Laser Step Alignment) method for detecting an image, FIA (Field Image Alignment) method for image-processing and measuring image data of an alignment mark photographed by illuminating a light having a wide wavelength bandwidth using a halogen lamp as a light source, or diffraction on a wafer For example, an alignment such as a laser interferometric alignment (LIA) method in which a laser beam having a slightly changed frequency is irradiated to the alignment mark on the lattice in two directions, the two diffracted light beams are interfered, and the position of the alignment mark is measured from the phase. There is a sensor. The alignment method is roughly divided into a TTL (through the reticle) method for measuring the position of the wafer through the imaging optical system, and an off-axis method for directly measuring the position of the wafer without passing through the projection optical system.

By performing the position detection of at least two points of the wafer mounted on the wafer stage by such an alignment sensor, not only the translation direction but also the position (rotation angle) of the rotation direction is also detected. Examples of sensors used to measure the rotation angle of wafers include alignment sensors such as a laser interferometric alignment (LIA) method in a TTL method, a laser step alignment (LSA) method in a TTL method, or a field image alignment (FIA) method in an off-axis method. There is this.

With respect to the exposure apparatus, it is also desirable to align the reticle and the wafer with high accuracy according to the detection results of these alignment sensors, to shorten the time required for alignment, and to maintain high throughput (the number of wafers processed per unit time). It is becoming. Therefore, it is necessary to increase the processing efficiency at every step from the step of transferring the wafer to the wafer stage to the final exposure. Here, the operation | movement in the wafer transfer process before reaching final alignment in the conventional exposure apparatus is demonstrated with reference to FIG.

Fig. 1 shows a configuration around a wafer stage for explaining a wafer delivery mechanism in a conventional exposure apparatus. In this FIG. 1, the wafer conveyance apparatus (not shown) shows the state where the board | substrate, ie, the wafer 6, was received on the elevating device 19 provided with the expansion and contraction mechanism 20 on the X-stage 11 interposed. .

The elevating device 19 includes three support pins (two support pin portions of which are inserted into the sample table 9, the θ rotation correction mechanism 8, and the wafer holder 7 so as to be movable). 19a, 19b), and the three support pins move up and down in order to transfer the wafer 6 between the wafer holder 7 by the vertical movement of the expansion and contraction mechanism 20. It is supposed to be. Moreover, the front end surface of each support pin is made to adsorb | suck the lower surface of a wafer by an external vacuum pump, and when the lifting device 19 moves up and down, the wafer 6 does not come out.

In the related art, the lifting device 19 is lowered to mount the wafer 6 on the wafer holder 7, and then, for example, by pressing the side surface of the wafer 6 against a plurality of pins, Alignment was done. By this prealignment, after rough positioning in the rotational direction and the translational direction of the wafer 6 was made, the wafer 6 was vacuum-adsorbed on the wafer holder 7.

In this manner, after the wafer 6 is left standing on the wafer holder 7 by vacuum adsorption, alignment and marks formed at both ends of the surface of the wafer 6 by alignment sensors such as an LSA method or a FIA method. (Search mark) is detected and the detection signal is output.

For example, the wafer is obtained by obtaining the coordinates of the moving mirror 13 fixed to the end of the sample stand 9 and the sample stand 9 measured by an external laser interferometer when the detection signal becomes a peak. The lateral shift error and the rotation error of the wafer on the coordinate system of the stage system are calculated. Based on the results, the θ rotation correction mechanism (θ table) 8 on the sample table 9 is driven to eliminate the rotational error of the wafer 6, and to position the reticle and the rotation direction of the wafer 6 (search). Alignment).

As described above, in the related art, the θ rotation correction mechanism 8 for rotating the wafer between the sample stage 9 and the wafer 6 provided with the moving mirror 13 serving as a reference of the coordinate system of the wafer stage system is provided. Since the lateral displacement of the wafer 6 occurs when the suction force of the vacuum system of the wafer holder 7 which adsorbs the wafer 6 is weak, or the complicated mechanism is provided on the sample stand 9, There has been a defect that the rigidity of the whole stage is weakened or the weight of the entire stage is increased so that the stage control performance is not improved. Thus, for example, it is conceivable to arrange the θ rotation correction mechanism under the sample stage 9. When the θ rotation correction mechanism is driven to adjust the rotation angle of the wafer 6, the moving mirror on the sample stage 9 can be considered. Since the angle of the light beam from the laser interferometer incident on (13) changes, the rotation angle of the θ rotation correction mechanism 8 is limited, for example, if the degree of prealignment of the wafer is bad, it is inconvenient to correct it sufficiently. There was this.

In the conventional exposure apparatus, after adsorbing the wafer 6 onto the wafer holder 7, as described above, the alignment sensor of either the LSA method or the FIA method is used for two positions on the wafer 6. The position of the alignment mark (search mark) was detected, and the lateral shift error and the rotational error of the wafer were detected. However, when two alignment marks are detected by one alignment sensor as described above, it is necessary to move the wafer 6 so that each mark sequentially enters the detection area of the alignment sensor, and all wafers in one lot Repeating such an operation with respect to was a factor of lowering the throughput of the exposure step. In order to avoid this, arranging the two alignment sensors so as to detect two such alignment marks simultaneously, the position of the two alignment marks on the wafer is limited by the arrangement of the two alignment sensors provided in the exposure apparatus. In order to take it, there existed a drawback that it is difficult to respond to the wafer etc. which differ in size, for example.

In this regard, a mechanism for varying the distance between two alignment sensors can also be considered. Such a variable mechanism is complicated, and it is difficult to arrange around the wafer stage where various sensors and the like are arranged. There was also a flaw.

In addition, conventionally, after mounting a wafer on the wafer holder 7, approximate positioning of the wafer was made by using a contact-type alignment device. However, when pre-alignment is performed after mounting the wafer on the wafer holder 7 in this manner, there is a drawback that the throughput cannot be increased. However, even when a prealignment mechanism capable of obtaining a new high throughput is proposed instead, it is desirable to be able to match with another exposure apparatus already equipped with a contact-type prealignment mechanism.

An object of the present invention can simplify the configuration of a wafer stage, thereby improving the rigidity and lightening of the wafer stage, and consequently, when mounting a wafer from the wafer loader system to the wafer stage, for example. SUMMARY OF THE INVENTION An object of the present invention is to provide a substrate positioning method capable of positioning a wafer at high speed and with high accuracy.

Another object of the present invention is a position of a substrate that can be positioned at high speed without restriction of the alignment mark, for example, when positioning the wafer through the wafer stage based on the position of the alignment mark on the wafer. To provide a method of decision.

Another object of the present invention is to provide a substrate positioning method that can achieve high matching accuracy during pre-alignment, for example, with another exposure apparatus that performs contact-alignment.

A first positioning method of the present invention is a positioning method for positioning a substrate having a cutout portion on a part of its outer circumference on a two-dimensionally movable substrate stage,

(1) a step of conveying the substrate to a sorghum position above the substrate stage,

(2) non-contact measurement of the position of the measurement point in the said cutting part of the outer peripheral part of the said board | substrate and the position of the other measuring point of the outer peripheral part of the said board | substrate, respectively, at the said delivery position,

(3) calculating a rotational error of the substrate based on the measurement result.

In this case, as shown in FIG. 6B, when the cutout part of the outer peripheral part of the board | substrate is a notch part, it is preferable to perform the position measurement at the measurement point set in the notch part with a two-dimensional image processing apparatus. In this case, the two-dimensional position shift amount and the rotational error of the substrate are detected by performing one-dimensional position measurement at another point on the outer circumference of the substrate on which the notched portion is formed.

On the other hand, as shown in FIG. 6A, when the cutout part of the outer periphery of the board | substrate is an orientation flat part, you may perform only one-dimensional position measurement using an image processing apparatus in any measurement point of the outer periphery of the board | substrate. In the case of performing one-dimensional position measurement, however, the two-dimensional position shift amount and the rotation error of the substrate are detected by performing position measurement at two or more other points, that is, three or more measurement points in the orientation flat portion.

In these cases, the two-dimensional position shift amount of the substrate is corrected by adding it as an offset to a positioning target position in the search alignment, for example. By this invention, the rotation correction mechanism of a board | substrate is unnecessary on the board | substrate stage side, and the precision improves.

The second substrate positioning method of the present invention is a positioning method for positioning a substrate on a two-dimensionally movable substrate stage, comprising: (1) a first and a second showing a two-dimensional position on the substrate, respectively; Forming a search mark of (2), (2) detecting two-dimensional positions of the first and second search marks of the first substrate, and (3) the position detected in the step (2) Calculating a rotational error of the first substrate on the basis of the above, and (4) the first search mark on the first substrate in parallel with detecting the two-dimensional position of the first search mark. Detecting and storing at least one dimensional position of the pattern away from the predetermined distance relative to (5) detecting the two dimensional position of the first search mark on the second substrate on the substrate stage, For the first search With respect to greater detected positional deviations from the position stored in the step (4) with a predetermined distance away from the pattern, and a step of calculating a positioning error of the second substrate on the basis of the position displacement.

In this second positioning method, after setting the first search mark for the second substrate as the detection area of the predetermined first alignment sensor, the second alignment is separated from the first alignment sensor by a predetermined distance. The position of the pattern (street line, etc.) in the detection area of the sensor is compared with the position stored with respect to the first substrate, and a positioning error is obtained from this comparison result. Therefore, it is not necessary to detect the position of the second search mark in the second substrate, and the detection area under the second alignment sensor is detected while the first sensor mark is detected by the first alignment sensor. Since only the position of the pattern inside may be detected, the measurement time is shortened.

The positioning method of the 3rd board | substrate of this invention is a positioning method for positioning the board | substrate which has a cutout in a part of outer periphery on the two-dimensionally movable substrate stage, (1) The said board | substrate stage (2) the position of the measurement point in the cutout portion of the outer peripheral portion of the substrate and the position of the other measurement point in the outer peripheral portion of the substrate using the two-dimensional image processing system, respectively; Non-contact measurement; (3) setting a virtual position corresponding to a reference position when positioning the substrate on the substrate stage by a contact-type positioning method within the observation field of view of the two-dimensional image processing system; And (4) image the substrate by using the shift amount from the virtual position of the position of the measurement point measured by the two-dimensional image processing system. And a step of predicting the position of the substrate when mounted on the substrate stage.

According to the third positioning method of the present invention, since the rotational error of the substrate is detected at the receiving point of the substrate at a position away from the substrate stage, for example, the photosensitive substrate is removed from the receiving point. When mounted on the stage, the rotation error can be corrected through the substrate raising means. Therefore, it is not necessary to provide a rotation correction mechanism on the substrate stage side, and the structure of the substrate stage can be simplified, and the rigidity improvement and the weight reduction of the substrate stage can be thereby achieved. Substrate positioning when mounting a substrate on the substrate stage from the system (wafer loader system) can be performed at high speed and with high accuracy.

In addition, in the contact positioning method within the observation field of the two-dimensional image processing system, a virtual position corresponding to the reference position when positioning the substrate on the stage is set, and the position from the corresponding virtual position of the measurement point on the photosensitive substrate. Since the substrate is positioned on the basis of the amount of misalignment, high matching at the time of rough positioning (at the time of the alignment) with other exposure apparatuses performing contact positioning (preliminary alignment). Precision can be obtained.

In this case, when the cut part of the outer peripheral part of the board | substrate is one cut part, it is preferable that the measurement point by the two-dimensional image processing system is set in one place in a cut part, and two places in the outer peripheral part of the board | substrate other than that. This means that the cutout is a recess such as a cutout. In this case, by detecting the position of the substrate at such three measurement points, the rotation angle and the two-dimensional position of the substrate can be specified.

On the other hand, when the cut part of the outer peripheral part of a board | substrate is one cut part with a flat edge, it is preferable that the measurement point by these two-dimensional image processing systems is set in two places in a cut part and one place in the outer peripheral part of the other board | substrate. This means that the cut portion is a flat cut portion such as an orientation flat portion, in which case by detecting the position of the substrate at these three measurement points, the rotation angle and the two-dimensional position of the substrate can be specified. In addition, when estimating the position of the substrate when the substrate is mounted on the substrate stage, the rotational error compared with the case where the substrate is positioned by the contact positioning method when the substrate is mounted on the substrate stage through the substrate raising means as it is. And the amount of misalignment, the rotational error is corrected when the substrate is mounted on the substrate stage by the substrate raising means, the substrate is mounted on the substrate stage, and then the amount of misalignment is corrected through the substrate stage. Do. Thereby, the structure of a board | substrate stage can be simplified and positioning of a board | substrate can be performed at high speed.

1 is a schematic structural diagram showing a conventional wafer transfer structure used in a projection exposure apparatus.

2 is a flowchart showing a prealignment process which is an example of an embodiment of a positioning method according to the present invention.

3 is a flowchart showing a search alignment step and a fine alignment step as examples of embodiments of the alignment method according to the present invention.

4 is a schematic configuration diagram showing an example of a projection exposure apparatus for implementing the alignment method of FIGS. 1 and 3.

5A to 5D are diagrams showing a wafer conveying apparatus, a wafer receiving mechanism, a turntable mechanism, and the like used in the projection exposure apparatus of FIG.

6A is a plan view of a wafer having an orientation flat portion;

6B is a plan view showing a wafer having a notch;

Fig. 7 is a structural diagram cut away showing an example of a two-dimensional image processing apparatus 50 for detecting an edge portion of a wafer.

8A and 8B are views used for explaining an example of a method for detecting a notched portion of a wafer;

Fig. 9A is a plan view showing the case where the first fine mark 47A is within the viewing field of view of the FIA microscope 5A.

Fig. 9B is a plan view showing the case where the second-degree fine mark 47B is in the field of view of the FIA microscope 5A.

10A is a diagram showing an observation image of the first search mark 47A of the first wafer in one example of the embodiment.

Fig. 10B is a waveform diagram showing an image pickup signal obtained by scanning Fig. 10A in the Y direction.

Fig. 10C is a waveform diagram showing an image pickup signal obtained by scanning Fig. 10A in the X direction.

Fig. 11A is a view showing an observation image of the first search mark 47A of the second and subsequent wafers.

11B is a view showing an image observed with a θ microscope 5B at that time.

FIG. 12A is a waveform diagram illustrating an image pickup signal corresponding to the image of FIG. 11A. FIG.

FIG. 12B is a waveform diagram illustrating an image pickup signal corresponding to the image of FIG. 11B. FIG.

Fig. 13 is a structural diagram cut away showing a different embodiment of the image processing apparatus 50;

14 is an explanatory view of various notches and orientation flat portions of a wafer;

15A to 15D are explanatory views of a method for detecting the edge position of the outer circumference of the wafer 6 using a line sensor or a two-dimensional image processing apparatus.

16A to 16D are explanatory diagrams of a method of detecting edge positions while matching a contact alignment prealignment mechanism with respect to the wafer 6N having notches.

17A to 17D are explanatory diagrams of a method for detecting edge position while matching a contact alignment prealignment mechanism with respect to a wafer 6 having an orientation flat portion.

18 is a flowchart showing an example of an operation of determining whether to remount a wafer and an operation of the reload;

Explanation of symbols on the main parts of the drawings

15: alignment 16: stage control system

18: central control system

Hereinafter, an example of embodiment of the positioning method by this invention is demonstrated with reference to FIGS. This example applies the present invention when the wafer is loaded and aligned in a stepper projection optical apparatus in which the reticle pattern is reduced and projected onto each shot area on the wafer through the projection optical system.

4 shows a schematic configuration of a projection exposure apparatus that performs the positioning method of this example. In Fig. 4, under the illumination light IL from the illumination optical system IA including a light source made of mercury lamp, fly eye lens, condenser lens, etc., the pattern on the reticle 1 is exemplified through the projection optical system 3. For example, it is reduced to 1/4 or 1/5 and is projected and exposed to each shot region of the wafer 6 to which the photoresist is applied. Hereinafter, the Z axis is held parallel to the optical axis AX of the projection optical system 3 of FIG. 4, the X axis is parallel to the ground of FIG. 3 in a plane perpendicular to the Z axis, and the Y axis is perpendicular to the ground of FIG. 3. I will explain.

The reticle 1 is held on the reticle stage 32 mounted on the reticle mount 31. The reticle stage 32 is configured to allow translation in the XY plane and rotation in the θ direction (rotation direction) by a reticle drive system (not shown). Moving mirrors 33a and 33b are provided on the edge portion of the reticle stage 32 in both the X and Y directions, and are reticle stages by the laser interferometers 34a and 34b fixed on the moving mirror and the reticle mount 31. Positions in the X-direction and Y-direction at (32) are always detected with a resolution of about 0.01 µm, for example, and the rotation angle of the reticle stage 32 is also detected at the same time.

The measurement values of the laser interferometers 34a and 34b are sent to the stage control system 16, and the stage control system 16 controls the reticle drive system on the reticle mount 31 based on the information.

Moreover, the information of the measured value of the laser interferometers 34a and 34b is supplied from the stage control system 16 to the central control system 18, and the central control system 18 controls the stage control system 16 based on the information. It is.

On the other hand, the wafer 6 is held by vacuum suction on the wafer holder 30 arranged on the sample stage 29 on the X stage 11. The sample stage 29 is moved in the Z direction (in this example, three Z tilt drivers) for correcting the position and tilt of the optical axis AX direction (Z direction) of the projection optical system 3. And a Z tilt drive unit 10 are provided on the X stage 11. Further, the X stage 11 is mounted on the Y stage 12, and the Y stage 12 is disposed on the wafer base 14, and can move in the X direction and the Y direction through a wafer stage drive system (not shown), respectively. It is configured to. Further, the movable mirror 13b is fixed to the edge portion of the sample stage 29, and the sample stage 29 is provided by the laser interferometer 17b disposed in a direction opposite to the movement mirror 13b and the movement mirror 13b. The coordinates and rotation angles in the X direction are detected. Moreover, the moving mirror 13b orthogonal to the interferometer 17b is provided in the other edge part of the sample stand 29, and the coordinate of the Y direction of the sample stand 29 is detected by the other laser interferometer 17a. The coordinate system defined by the coordinates X and Y measured by the laser interferometers 17a and 17b is called the coordinate system (stage coordinate system) X and Y of the wafer stage.

Measurement values of the laser interferometers 17a and 17b are sent to the stage control system 16, and the stage control system 16 controls the wafer stage drive system based on the information. In addition, measurement value information of the laser interferometer 17 is supplied from the stage control system 16 to the central control system 18, and the central control system 18 is configured to control the stage control system 16 based on the information. . In addition, a wafer transfer device 39 (see FIG. 5A) for receiving a wafer is arranged near the wafer stage, and a wafer transfer mechanism is provided in the wafer stage, which will be described later in detail.

In addition, the projection exposure apparatus of this example includes an alignment sensor 4 of a TTL method for performing alignment between the reticle 1 and the wafer 6, and two alignment sensors of an FIA (imaging method) method in an off-axis method. 5a and 5b). In the alignment sensor 4 of this example, an alignment sensor of the LSA (Laser Step Alignment) method and an alignment sensor of the LIA (Laser Interferometric Aligment) method are assembled in parallel, and either method is used depending on the required alignment accuracy and the like. . In the case of alignment, the position of the alignment mark formed on the wafer 6 or the position of a predetermined pattern is detected by any of these alignment sensors 4, 5a, and 5b, and based on the detection result, the said wafer The pattern formed in the previous step and the pattern position on the reticle are precisely matched to each shot region of (6). The detection signals from these alignment sensors 4, 5a and 5b are processed by the alignment control system 15, and the alignment control system 15 is controlled by the central control system 18. Moreover, on the sample stage 29, a standard mark member 43 having a surface having the same height as the surface of the wafer 6 is fixed, and a mark as a reference for alignment is formed on the surface of the reference mark member 43. It is.

As described above, the stage control system 16 and the alignment control system 15 are controlled by the central control system 18, and the central control system 18 collectively controls the entire projection exposure apparatus so that the exposure operation is performed in a constant sequence. It is a structure which consists of.

Next, in this example, three off-axis two-dimensional image processing apparatuses 50, 51, and 52 are disposed in the vicinity of the end portion on the wafer side of the projection optical system 3. Each of these image processing apparatuses 50 to 52 captures an image of the edge portion of the outer peripheral portion of the wafer when the wafer is transferred to the loading position (receiving position) above the wafer holder 30 as described below. The imaging signals from the image processing apparatuses 5b to 52 are supplied to the alignment control system 15, and the alignment control system 15 calculates the horizontal misalignment error and the rotational error of the wafer at the receiving position from the supplied imaging signal. . The arrangement and configuration of the image processing apparatuses 50 to 52 will be described later.

Next, the transfer mechanism of the wafer on the wafer transfer system and the wafer stage will be described with reference to FIGS. 5A and 5B. In addition, the wafer stage generically refers to the wafer holder 30, the sample stage 29, the Z tilt driving unit 10, the X stage 11, the Y stage 12, and the wafer base 14.

5A is a plan view of the wafer transfer system and the wafer stage peripheral configuration of the present example, and FIG. 5B is a side view. 5A and 5B, a wafer transfer device 39 for receiving a wafer is disposed above the -X direction of the wafer stage. The wafer conveying apparatus 39 includes conveying arms 21 and 22 arranged in series in the X direction, a slider 23 for sliding those conveying arms 21 and 22 to a predetermined position, and conveying arms 21 and 22. It consists of the arm drive system which is not shown in figure which drives the. Moreover, the slider 23 is arrange | positioned independently of the exposure apparatus main body, and the vibration at the time of the drive of the slider 23 is not transmitted to the exposure apparatus main body side. In addition, the two transfer arms 21 and 22 both have a U-shaped flat plate portion, and the wafers are mounted on the upper surfaces thereof. These two transfer arms 21 and 22 are capable of unloading (exporting) the wafer after exposure and loading the next wafer. In addition, since the structure of this wafer conveyance apparatus is a well-known thing, detailed description of this is abbreviate | omitted.

That is, the conveying arms 21 and 22 move along the slider 23 to the loading position where the wafer is received by the wafer stage system based on the instruction from the loader control device 24, and the conveying arms 22 are moved to the conveying arm 22. The wafer 6a before exposure is carried out. Thereafter, the wafer 6 exposed next by the transfer arm 21 is moved on the wafer stage and mounted on the lift table 38. The 5th b has shown the state in which the exposed wafer 6a was mounted in the conveyance arm 22 on the slider 23, and the wafer 6 was passed by the conveyance arm 21 to the front-end | tip of the platform 38. As shown in FIG.

The lifting platform 38 is supported by the expansion and contraction mechanism 35 provided on the X stage 11, and triaxial support pins 38a to 38c which are regrettable in the opening of the sample stage 29 and the wafer holder 30, respectively. When the expansion and contraction mechanism 35 moves in the vertical direction (Z direction), the three-axis support pins 38a to 38c move the wafer up and down, and the wafer is delivered. Adsorption holes for vacuum adsorption are formed at the ends of the triaxial support pins 38a to 38c, respectively, and their ends are moved to a height as high as possible between the carrier arms 21 and 22 when the wafer is delivered. When the wafer is mounted on the wafer holder 30, the wafer moves to a position lower than the surface of the wafer holder 30. In addition, the wafer is vacuum-adsorbed from the upper ends of the support pins 38a to 38c so that the wafer does not come off when the lifting platform 38 is moved up and down.

In addition, the expansion and contraction mechanism 35 is engaged with the drive shaft 37 which is supported by a rotation material on the XY plane about the central axis 35Z and rotated by the rotation drive system 36 provided on the X stage 11. Thus, it is possible to rotate to the desired angle by the command from the central control system 18 that controls the rotation drive system 36. The rotation system consisting of the rotation drive system 36, the drive shaft 37, and the expansion / retraction mechanism 35 has sufficient angle setting resolution capability, for example, clockwise, when viewed from the top of the wafer 6 with an accuracy of 20 μrad, or It can be rotated in either counterclockwise direction. The expansion and contraction mechanism 35, the rotation drive system 36 and the drive shaft 37 constitute a drive mechanism for the platform.

5C shows the turntable 60 of the wafer transfer system. In this FIG. 5C, the wafer 6 on the turntable 60 is passed to the platform 38 via the transfer arm 21 of FIG. 5B. In the vicinity of the turntable 60, a light transmitting portion 61a for irradiating the slit-shaped light beam to the outer peripheral portion of the wafer 6 and a light receiving portion 61b for receiving and photoelectrically converting the light beam passing through the outer peripheral portion of the wafer 6. ), An eccentric sensor 61 is disposed, and the detection signal S1 from the light receiving portion 61b is supplied to the central control system 18 in FIG. 5B. In addition, although the light receiving part 61b of this example consists of one photodiode, you may detect the position of the outer peripheral part of a wafer directly, for example using a one-dimensional line sensor etc. In this case, as shown in FIG. 5A, the wafer 6 of the present example has a circular shape and a flat orientation flat portion FP is formed in a part of the outer peripheral portion.

Therefore, in FIG. 5C, when the wafer 6 is rotated while being sucked and held by the turntable 60, the wafer passes through the eccentric sensor 61 due to the eccentricity of the wafer 6 and the presence of the orientation flat portion. The width of (6) changes. And as shown in FIG. 5D, the output signal S1 output from the light receiving part 61b with respect to the rotation angle (phi) of the turntable 60 is a sine wave form, and the part 62 corresponding to the orientation flat part FP is shown. Change to a low level at. In the central control system 18, the rotation angle Φ _F when the orientation flat portion is located at the center of the eccentric sensor 61 from the detection signal S1 and the rotation angle Φ of the tin table 60, and the wafer 6. The amount of eccentricity is obtained, and the turntable 60 is stopped with the orientation flat portion in the predetermined direction. In addition, the central control system 18 adjusts the position of the sample stage 29 for the wafer when the wafer 6 is received at the loading position based on the information of the amount of eccentricity.

In the central control system 18, as shown in Fig. 5D, the detection signals S1 at the rotation angles Φ _A , Φ _B , and Φ _C corresponding to the measurement points by the three image processing apparatuses 50 to 52 described above. ), and placed in mind the detection signal (S1) at a predetermined rotation angle Φ _D as digital data. In this regard, for example, even when it is necessary to measure the position of the wafer 6 at the measurement points corresponding to the rotation angles Φ _A , Φ _B , and Φ _C on the specifications of the exposure apparatus, the rotation is performed on the structure of the wafer stage of the exposure apparatus. It is difficult to arrange the image processing apparatus 52 in the position corresponding to each phi _C , and it may be necessary to arrange the image processing apparatus 52 in the position corresponding to the rotation angle phi _D. Furthermore, there may be a case where only two image processing devices can be arranged instead of three. First, in the former case, in the central control system 18, for example, the value of the detection signal S1 at the rotation angles Φ _A , Φ _B , Φ _C , Φ _D , and the rotation angles Φ _A , Φ _B , Φ _D The measured value at the measurement point corresponding to the rotation angle Φ _C is estimated from the measured value at the position of the outer circumferential portion of the wafer 6 at the measurement point corresponding to, and the lateral shift amount and the rotation error of the wafer 6 are calculated using the estimated value. Calculate. In the latter case, the central control system 18 corresponds to the value of the detection signal S1 at the rotation angles Φ _A , Φ _B , Φ _C , Φ _D , and, for example, the rotation angles Φ _B , Φ _D. From the measured values of the position of the outer peripheral part of the wafer 6 at the measuring point, the measured values at the measuring points corresponding to the rotation angles Φ _A and Φ _C are estimated. Calculate.

In general, such a calculation method has a position corresponding to the rotation angles Φ _A , Φ _B , Φ _C , and Φ _D when the mechanical prealignment mechanism using the positioning pin is mounted in another exposure apparatus. It is used for matching when the reference position is set. However, in the present example, as described later, the position of the wafer 6 is adjusted through the two-dimensional image processing apparatus 50 to 52 at the same measurement point as in the case of prealignment using a contact-type prealignment mechanism. Since it can measure, it is necessary to match with the contact-type pre-alignment mechanism using the eccentric sensor 61 shown in FIG. 5C. This is a case where the observation visual field by the image processing apparatus 50-52 cannot be set mainly in the vicinity of the position of the pin of a contact-type prealignment mechanism.

Next, the arrangement and configuration of the image processing apparatuses 50 to 52 will be described in detail. First, FIG. 6A shows the wafer 6 in the loading position. In this FIG. 6A, the observation visual fields 50a, 51a, 52a of the three image processing apparatuses 50, 51, 52 of FIG. 4 are set in three edge parts of the outer periphery of the wafer 6, respectively. In addition, although the object (observation field of view) of actual image processing is a rectangular area, it is shown by the circular area for the convenience of description. In this case, two observation fields 50a and 51a are set on the orientation flat part FP, and the other observation field 52a is set on the circumference. In this way, by detecting the positions of the three edge portions of the outer circumference of the wafer 6, the positional shift amount (lateral shift amount) in the X direction and the Y direction of the wafer 6 in a moment after passing the wafer 6, and Detection of rotational error, that is, detection for prealignment, is performed.

When such a lateral shift amount and a rotation error are detected, the correction of the position shift in the X direction and the Y direction is detected at the time of search alignment which will be described later performed after the wafer 6 is mounted on the wafer holder 30. By adjusting the position. On the other hand, correction of the rotation error is performed by rotating the platform 38 through the rotation drive system 36 before the platform 38 descends and the wafer 6 contacts the wafer holder 30 in FIG. 5B.

In addition, in the wafer, instead of the orientation flat portion, as shown in FIG. 6B, there is also a wafer 6N in which a V-shaped notch NP is formed in a part of the circular outer circumference. With respect to such a wafer 6N, in these three observation fields 50a to 52a, one observation field 51a covers the notch portion NP, and the other two observation fields 50a and 52a are circular. It is set to cover the edge portion of the outer circumference. By this arrangement, the amount of lateral shift and the rotational error of the wafer 6N having the notched portion NP are immediately detected after the transfer of the wafer 6.

7 shows a configuration of an example of the image processing apparatus 50 of this example. In FIG. 6, illumination light of a wavelength band in which photoresist from a light source 58 such as a lamp or a light emitting diode is hardly exposed to light is collected at one end of the light guide 57. Then, the illumination light emitted from the other end of the light guide 57 passes through the collimator lens 56, the half prism 54, and the objective lens 53, and the loading position on the tip end of the three support pins 38a to 38c. Is irradiated to the edge portion of the outer circumference of the wafer 6 in the wafer. The reflected light from the edge portion forms an image of the edge portion on the imaging surface of the imaging element 59 made of a two-dimensional CCD or the like via the objective lens 53, the half prism 54, and the imaging lens 55. The imaging signal from the imaging element 59 is supplied to the alignment control system 15, and the alignment control system 15 is configured to obtain the position of the edge of the detection target of the wafer 6 from the imaging signal.

13 shows another example of the structure of the image processing apparatus 50 of this example. In FIG. 13, illumination light of a wavelength band in which photoresists from a light source not shown, such as a lamp or a light emitting diode, is hardly exposed to light, is collected at one end of the light guide 72.

And the illumination light emitted from the other end of the light guide 72 is bent by the deflection mirror 73, and passes the illumination light which passed the opening 75 to the wafer holder 30A arrange | positioned on the sample stand 29A. The cut part 74 for making it is provided. It is comprised so that the illumination light which passed the opening part 75 and the cutting | disconnection part 74 may be irradiated to the edge part of the outer periphery of the wafer 6 in the loading position on the front-end | tip of three spindle parts 38a-38c. And the illumination light which permeate | transmitted the edge part vicinity forms the edge part image on the imaging surface of the imaging element 59 which consists of a two-dimensional CCD etc. through the objective lens 53A and the imaging lens 55A. An image of the edge portion is formed on the imaging surface of the imaging signal 59 from the imaging element 59. The imaging signal from the imaging element 59 is supplied to the alignment control system 15, and the alignment control system 15 is configured to obtain the position of the edge which is the detection target of the wafer 6 by the imaging signal.

Whatever the above image processing apparatus is used, since the wafer 6 is mounted on the lift table 38 (support pins 38a to 38c), the outer periphery of the wafer 6 as indicated by the dashed-dotted line in FIG. The edge portion is bent slightly downward (-Z direction). In addition, since the amount of warpage varies depending on the variation in the thickness of the wafer 6, the imaging optical system composed of the objective lens 53 and the imaging lens 55 has a numerical aperture NA having a telecentric optical system and a large depth of focus. There is a need. If the wavelength of the illumination unit is λ, the depth of focus is almost proportional to λ / NA ^2. Therefore, by decreasing the numerical aperture NA, a large depth of focus can be obtained. As a result, the largest depth of warp in the wafer 6 is obtained. Even the edge part of a part can be detected correctly. For example, when the wavelength λ of the irradiation light is 0.633 μm, when the numerical aperture NA is about 0.03, a focal depth of 0.5 mm or more can be obtained, and a resolution of about 20 μm is obtained. In general, since about 1/10 of the resolution capability becomes the detection capability, the detection capability becomes about 2 μm, which enables highly accurate alignment.

In addition, for example, in FIG. 6A, in particular, when it is not necessary to exactly match with the non-contact prealignment mechanism, in order to detect the position of the orientation flat portion FP or the normal edge portion of the outer circumference of the wafer, it is necessary. It is not necessary to perform two-dimensional image processing, and the imaging signal from one-dimensional imaging element such as a line sensor that uses the normal direction of the edge portion as the measurement direction, or the imaging tube (ITV) whose normal direction is the scan direction, respectively. You may process it. This is because these position detection results are one-dimensional, for example, in the case of FIG. 6A, the position shift amounts and rotational errors in the X and Y directions of the wafer 6 can be obtained from the three-dimensional position detection results. Because there is.

On the other hand, when it is necessary to match exactly with a non-contact prealignment apparatus, and when a line sensor is used as an image processing apparatus for edge part detection of a wafer, when the conveyance degree of a wafer to a loading position is bad, The outer periphery position of the wafer to be matched from the detection area of the line sensor may deviate, and the matching error may increase. This will be described with reference to FIG. 15.

First, in FIG. 15A, linear observation visual fields (detection areas) 50a1 and 51a1 of two line sensors are set so as to cross the orientation flat portion FP of the wafer 6, and the circumferential image of the wafer 6 is set. The case where the linear observation visual field 52a1 of one line sensor is set in the figure is shown. 15C shows an enlarged view of the observation field 50a1 within these observation fields. In this FIG. 15C, the observation visual field 50a1 on the outer periphery 6a of the wafer 6 is, for example, a measurement point at which the reference pin 81A for positioning of the contact-type prealignment mechanism has contacted in the last exposure step. It is located in a position away from the non-measurement direction (direction perpendicular to the measurement direction) from (A). At this time, the straight line 82A passing through the measurement point A and parallel to the measurement direction represents the matching position. In this way, when the viewing field of view 50a1 deviates from the matching position, the unevenness of the outer peripheral portion 6a of the wafer 6 causes the gap between the position at the measurement point 83A and the position at the measurement point A in the observation field of view 50a1. There is a fear that a large error (matching error) DELTA 1 occurs.

In this case, as described above, using the turntable 60 and the eccentric sensor 61 of FIG. 5C, as shown in FIG. 5D, at predetermined rotation angles Φ _A , Φ _B , Φ _C , and Φ _D The eccentricity of the wafer can be measured and corrected using this measured value.

As another correction method, there is also a method in which the width in the non-measurement direction of the observation field (detection area) 50a1 of the line sensor is widened to an accuracy of conveyance. In this method, as shown to FIG. 15C, the line sensor which has the observation visual field 50a1 'widened to the extent including the measurement point A which the pin 81a for positioning contacted in the non-measurement direction is used.

In this case, since the image in the observation visual field 50a1 'is integrated optically or electrically with respect to the non-measurement direction, the position of the outer periphery 6a of the wafer 6 in the observation visual field 50a1' is the average line ( 84A). Since the average line 84A is closer to the measurement point A to be matched in the measurement direction than the measurement point 83A when the slit-shaped observation field of view 50a1 is used, the position of the average line 84A and the measurement point A ), The matching error Δ2 between positions is smaller than the above error Δ1. In other words, by widening the observation field (detection area) of the line sensor in the non-measurement direction to have an averaging effect, it is possible to reduce the matching error with the contact-type alignment device.

Further, by using the two-dimensional image processing apparatuses 50 to 52 instead of the line sensor, matching errors can be reduced. FIG. 15D shows an observation field corresponding to FIG. 15C when the two-dimensional image processing apparatus is used. FIG. In this Fig. 15D, the scanning direction of the pixel is set in the Y direction (measurement direction) with respect to the image in the observation field of view. In this case, first, the average line 84A 'of the outer periphery 6a of the wafer 6 within the observation field of view is obtained, and the position (average position) of the average line 84A' is obtained. Similarly, the average position of the wafer 6 is also found in the other two places, and the position in the X direction, the position in the Y direction, and the rotation angle of the wafer 6 are obtained from the average positions at the three places. Then, by the position and rotation angle of the wafer 6, a row of pixels 85 including a straight line 82A parallel to the Y direction past the measurement point A to be matched is specified. The position of the outer periphery 6a of the wafer 6 at the measurement point substantially the same as the measurement point A in the contact-type alignment device is obtained from the image pickup signals from the specified one-row pixels 85. By this method, the matching error can be reduced.

As described above, in the case of the wafer 6 having the orientation flat portion FP, the position detection of the orientation flat portion FP can also be performed using a one-dimensional imaging element. However, as shown in FIG. 6B, the notch portion NP of the notch portion NP of the wafer 6N having the notch portion NP needs to be detected in the X and Y directions. Needs to perform position detection in a two-dimensional image processing apparatus. Here, the detection method of the notch part NP is demonstrated with reference to FIG. 8A and 8B.

First, FIG. 8A is an enlarged view of the notch portion NP of the wafer 6N. In Fig. 8A, in order to position the wafer 6N on the wafer holder, a circumferential reference pin of a predetermined diameter d is inserted into the notch portion NP. Therefore, the standard of the shape of the notch part NP was decided based on the shape of the reference | standard pin. Therefore, when the area on the notch part NP which is conjugate with the imaging surface of the two-dimensional imaging element is called the observation field 63, it contacts with the two edges of the notch part NP from image data in the observation field 63 as an example. Assuming that a virtual reference pin 64 having a diameter d is assumed, the X coordinate and Y coordinate of the center (0) of the virtual reference pin 64 are detected.

As another example, as shown in FIG. 8B, the coordinates of the intersection point P of the two edges 65A and 65B of the notch portion NP and the one edge 65B from the image data in the observation field view 63. There is also a method for obtaining the coordinates of the intersection 65C with the outer circumference of the wafer. In this case, it is assumed that an intersection 65D is virtually provided at a position symmetrical with the intersection 65C on the edge 65A, and a triangle having three intersections P, 65A, 65B as a vertex. Then, according to the proportional distribution with respect to the intervals of the intersections 65C and 65D, which are the bases, the positions of the triangles are determined so that the bases have the interval d. The X coordinate and the Y coordinate of 0) may be obtained.

Next, another example of a wafer and an example of a detection system suitable for them will be described with reference to FIGS. 14A and 14B. First, as the notch portion, as shown in Fig. 13A, there is a wafer 6M having any one of the notch portion NP1 at the 6 o'clock position and the notch portion NP2 at the 3 o'clock position. You need to detect it correctly.

Therefore, in FIG. 14A, the wafer holder 30B under the wafer 6M is generally used as cutting portions 30Ba and 30Bb for irradiating the notch portions NP1 and NP2 and mechanical prealignment. Cutting portions 30Bc to 30Be for irradiating the position of the reference pins are formed. These cut | disconnected parts 30Ba-30Be are respectively illuminated from the bottom face side by the image processing apparatus of FIG. The detection of the position and rotation angle when the notch portion NP1 in the 6 o'clock position is provided on the wafer 6M is performed by a two-dimensional image processing system having a circular observation field 51a2, and a linear observation field 50a2 and 52a1) and the first and second line sensors are used. On the other hand, the detection of the position and rotation angle when the notch portion NP2 in the 3 o'clock direction is present on the wafer 6M is performed by a two-dimensional image processing system having an observation field 51a1, and a linear observation field 52a1 and Using the second and third line sensors with 50a1). In other words, the five sensors made up of two two-dimensional image processing systems and three line sensors have a configuration in which both types of wafers can be used for prealignment.

Next, also as a kind of orientation flat part, as shown to FIG. 14B, there exists the wafer 6A which has either the orientation flat part FP1 of the 6 o'clock direction, or the orientation flat part FP2 of the 3 o'clock direction, Both of these need to be detected correctly. Therefore, in FIG. 14B, three cutting parts 30Ca-30Cc corresponding to the position of one orientation flat part FP1, and another orientation flat part (in the wafer holder 30C of the base of the wafer 6A) are arranged. Three cut parts 30Cd to 30Cf corresponding to the position of FP2 are formed, and these cut parts 30Ca to 30Cf are illuminated by the image processing apparatus shown in FIG. 13 from the bottom. And the detection of the position and rotation angle when the orientation flat part FP2 of the 3 o'clock direction exists in the wafer 6A is performed by the line sensor which has observation visual field 52a2, 51a2, 50a2, respectively. The detection of the position and rotation angle when the orientation flat portion FP1 in the 6 o'clock direction is provided on the wafer 6A is performed by a line sensor having observation fields 52a1, 51a1, and 50a1, respectively. These have shown the case where an optical detection system can be arrange | positioned on the reference pin position of a mechanical prealignment system on a wafer stage. However, if this is difficult, as described above, the outer shape of the wafer may be obtained using the measurement results on the turntable 60 shown in FIG. 5C, and may be replaced with the position measurement results using the reference pins.

Next, the operation of the case where the wafer position and the rotation angle are detected while matching with the contact-type prealignment mechanism using the three two-dimensional image processing apparatuses 50 to 52 of this example. An example will be described with reference to FIGS. 15, 16A-16D, and 17A-17D. As the image processing apparatuses 50 to 52 in this case, as shown in FIG. 13, a transmission illumination type image processing apparatus is used.

First, FIG. 16A shows a state where the wafer 6N having the notch NP in the 6 o'clock direction is conveyed in the loading position. In this FIG. 16A, the observation visual field of the actual process target of the three image processing apparatuses 50, 51, and 52 of FIG. 4 is made into the rectangular observation fields 86A, 86B, 86C, respectively. Moreover, in the area | region corresponding to observation visual field 86a-86C, for example, the reference pin of the contact-type prealignment mechanism used in the previous exposure process is set as the virtual pin 82A-82C, respectively, and a virtual line is used as a virtual line. In the same figure, in order to make it easy to see, the virtual pins 82A-82C are enlarged and displayed by the position outside the corresponding observation visual field 86A-86C, respectively. In the circumferential observation fields 86A, 86B of the wafer 6N, the notch portion (A, B) is used as the virtual position A, B, where the virtual pins 82A, 82B and the outer circumference of the wafer are in contact with each other. The center of the virtual pin 82C in the observation visual field 86C on NP) is called a virtual position E. FIG.

In order to obtain these virtual positions A, B, and E, for example, a reference wafer on which a reference mark indicating a virtual position is formed is mounted on the wafer holder 30 of the wafer stage of FIG. 4, and the laser interferometer 17 The wafer stage may be driven based on the measured values based on the design positions of the three reference pins of the contact-type alignment device, and the reference marks may be sequentially moved into the observation fields 86A to 86C. The virtual positions A, B, and E are set by detecting the positions of the reference marks in the observation visual fields 86A to 86C in the corresponding image processing apparatuses 50 to 52 and storing the detected positions. As the reference wafer, a wafer with a groove on the outer circumference of the wafer, a glass wafer with a pattern or the like may be used.

Next, as demonstrated with reference to FIG. 5B, the platform 38 of this example is supported so that it may rotate by the rotation drive system 36 centering on the central axis 35Z. As such, when the platform 38 is rotated, the positions of the virtual positions A, B, and E, which are the contact points between the wafers and the virtual pins 82A to 82C, change within the observation fields 86A to 86C of FIG. 16A. Therefore, the above-mentioned reference wafer is actually sucked on the platform 38 to measure the positions of the virtual positions A, B, and E. Then, the platform 38 is rotated by a predetermined angle, and the virtual positions A, B, and E are again. The position of is measured and the coordinate in the stage coordinate system of the platform 38 and the rotation center 0 'is calculated | required from the fluctuation amount of the measurement result. Thereafter, a new coordinate system (X, Y) is a new coordinate system using the origin of the coordinate system (stage coordinate system) (X, Y) of the wafer stage determined based on the measured value of the laser interferometer 17 as the rotation center (0 '). 0 '). That is, in this new coordinate system (X, Y; 0 '), the coordinate of the rotation center (0') is (0,0).

The virtual positions A and B are arranged along a straight line parallel to the X axis, and the intervals in the X direction of the virtual positions A and B are set to L. FIG. Further, in the new coordinate system (X, Y; 0 '), the coordinates of the virtual positions A and E are (x ₁ , a) and (x ₂ , a), respectively, and the coordinates of the virtual position E are ( The relative coordinates in the observation visual fields 86A to 86C of the virtual positions A, B, and E are stored as (x ₁ + x ₂ ) / 2, b). If the Y coordinate in the new coordinate system (X, Y; 0 ') is y, as shown in Fig. 16A, the virtual positions A and B are located on a straight line (y = a). And the following equation holds.

x ₂ -x ₁ = L

At this time, under the illumination light from the bottom surface side of the wafer 6N, the images of the pattern in the observation fields 86A to 86C are each as shown in Figs. 15B to 15D. Therefore, the outer circumferential edge of the wafer 6N can be detected as the boundary between the dark portion and the roll.

Next, a method of detecting the position and rotation angle of the wafer 6N when the wafer 6N actually exposed is set to the loading position as shown in FIG. 16A. First, in two observation visual fields 86A and 86B on the circumference of the outer circumference of the wafer 6N, as shown in FIGS. 16B and 16C, the outer circumference corresponding to the boundary between the wrist Bi and the arm Dk, respectively. The edge is almost straight. Therefore, in the new coordinate system (X, Y; 0 '), the X coordinate is x and the Y coordinate is y, and the edge in the observation field 86A is represented by a function (y = f (x)), and the observation field 86B ) Is represented by the function (y = g (x)). For this purpose, two coefficients of the first-order functions f (x) and g (x) by the least-squares method may be determined based on the edge detection positions in the observation visual fields 86A and 86B. In addition, since the outer circumference of the wafer is almost circular, instead of the first-order functions f (x) and g (x), a function of an arc represented by (xx ₀ ) ² + (yy ₀ ) ² = r ² may be used. .

After that, the coordinate in the new coordinate system (X, Y; 0 ') of the measurement point C on one function (y = f (x)) is (x ₃ , y ₃ ), and the other function (y The coordinate in the new coordinate system of the measuring point D on = g (x)) is (x ₄ , y ₄ ), and the straight line passing through the measuring points C and D is parallel to the X axis and Set their coordinates so that the interval in the X direction is L. This means to find the coordinates x ₃ and x ₄ to satisfy the following equation.

g (x ₄ ) = f (x ₃ ), x ₄ -x ₃ = L

In addition, Y coordinate y ₃ is determined by f (x ₃ ) and Y coordinate y ₄ is determined by g (x ₄ ) using the coordinates x ₃ and x ₄ determined in this way.

In this case, a straight line (function x = (x ₄ + x ₃ ) / 2 that is parallel to the Y axis passing through the center point (X coordinate of (x ₄ + x ₃ ) / 2) of these two measuring points (C and D) Is passed through the center 0 of the wafer 6N. Thus, a straight line represented by the function x = (x ₄ + x ₃ ) / 2 and a straight line perpendicular to the straight line represented by the function (y = f (x) past the measurement point C (x ₃ , y ₃ ) By finding the coordinates of the intersection of (i.e., center 0) and finding the distance between the center 0 and the measurement point C, the radius r from the center (0) to the measurement point C. Similarly, the function x = (x ₄ the intersection of the straight line represented by + x ₃ ) / 2 and the straight line perpendicular to the straight line represented by the function (y = g (x)) past the measurement point D (x ₄ , y ₄ ) ) And the distance between the center 0 and the measuring point D, the radius r 'from the center 0 to the measuring point D can be obtained. When using the function of, the intersection with the straight line orthogonal to the tangent of the circle becomes the center (0).

In addition, it is necessary to determine the center coordinates of the reference pin when the reference pin is inserted into the notch part NP from the image data of the observation field of view 86C on the notch part NP. Therefore, the space | interval LM of two virtual contact points between the virtual pin 82C and the notch part NP is calculated | required beforehand. And as shown to FIG. 16D, the function (y = h (x)) corresponding to the edge of the left side of the notch part NP by the least-squares approximation method based on the image data in 86 C of observation visual fields, and Find the function (y = i (x)) corresponding to the right edge. In addition, h (x) and i (x) are linear functions. Thereafter, the center of the virtual pin 82C when the virtual pin 82C is inserted into the notch portion NP is the measurement point F, and the measurement point F in the new coordinate system X, Y; The coordinates are called (x ₅ , y ₅ ).

Then, the distance between two points where a straight line parallel to the X axis past the measurement point F intersects the straight line defined by the function (y = h (x)) and the function (y = i (x)) is LM. , The coordinates (x ₅ , y ₅ ) are determined so that the measurement point F is the midpoint of these two points. For this purpose, first, the X coordinates (x ₅₁ , x ₅₂ ) satisfying the following equation are determined.

h (x ₅₁ ) = i (x ₅₂ ), x ₅₂ -x ₅₁ = LM

Then, using the X coordinates (x ₅₁ , x ₅₂ ) determined as above, the coordinates (x ₅ and y ₅ ) may be obtained by the following equation. In this way, the coordinates of the measurement point F are determined.

x ₅ = (x ₅₁ + x ₅₂ ) / 2, y ₅ = h (x ₅₁ )

Next, in FIG. 16A, it corresponds to a straight line (a straight line represented by the function x = (x ₃ + x ₄ ) / 2) parallel to the Y axis past the center point of the measurement points C and D, and the observation field ( Since the X coordinate which is a straight line passing through the inside of 86C) is determined by (x ₁ + x ₂ ) / 2 which is the X coordinate of the virtual position E, when the center (0) of the wafer 6N is the rotation center, The rotation error Δθ from the virtual position E of the measurement point F of the notch portion NP in the observation field 86C is obtained by the following equation.

Δθ = (x _5- (x ₁ + x ₂ ) / 2} / (r + r ')

However, since this is a rotation error when the center of rotation 6N is used as the rotation center, the wafer 6N has been rotated around the platform rotation center 0 '(0,0) as an axis. In this case, the rotation error Δθ is common and can be corrected, but the offset of the measured value occurs at each measurement point C, D, or F. Accordingly, the rotational error Δθ is calculated based on the new coordinate system X, Y; 0 ', and in FIG. 16A, the rotational center 0' is offset so that the wafer 6N is offset by the rotational error Δθ. Assume that you have rotated around). At this time, the functions (y = f (x)) and (y = g (x)) are changed into the functions (y = f '(x)) and (y = g' (x)), respectively. The functions f '(x) and g' (x) after the change are obtained.

Since the rotation error (Δθ) has already been corrected, the measurement point on the straight line determined by the function (y = f '(x)) is changed to C' (x ₃ ', y ₃ ') and the function (y = g ' The measurement points on the straight line determined by (x)) are D '(x ₄ ', y ₄ '), and the measurement points C' and D 'are arranged in a straight line parallel to the X-axis, and their spacing is L Find the coordinates x ₃ ', x ₄ ' to be. That is, the coordinates x ₃ ', x ₄ ' are determined to satisfy the following equation.

g '(x ₄ ') = f '(x ₃ '), x ₄ '-x ₃ ' = L

Moreover, y ₃ 'and y ₄ ' which are Y coordinates are represented by f '(x ₃ ') and g '(x ₄ '), respectively. As a result, the virtual positions A (x ₁ , a) and B (x ₂ , a), and the measured points C '(x ₃ ', y ₃ ') and D' (x ₄ ', y ₄ ') after rotation correction The average difference DELTA X, DELTA Y is given by the following equation.

ΔX = (x ₃ '+ x ₄ ' -x ₁ -x ₂ ) / 2

△ Y = (y ₃ '+ y ₄ ' -2a) / 2

= {f '(x ₃ ') + g '(x ₄ ') -2a} / 2

The differences DELTA X and DELTA Y are positional shift amounts in the X-direction and Y-direction of the wafer 6N with respect to the position when the alignment is performed by the contact-type prealignment mechanism.

In this example, the position shift amount is regarded as an offset during search alignment, and is corrected by controlling the amount of movement of the wafer stage during search alignment.

In the above-described example, it is assumed that the radius r (or r ') of the wafer 6N fluctuates greatly. However, in general, the error between the wafers of the outer diameter of the wafer is about ± 0.1 µm, so that the radius r ) May be fixed.

In the case where the variation in the outer diameter of the wafer is large, the contact points of the virtual pins 82A and 82B are different in FIG. 16A, and therefore, the distance L between the two virtual positions A and B cannot be determined as a fixed value. In this case, the wafer outer periphery from the first-order functions f (x) and g (the slope of x (also the slope of the tangent to the arc when using a function of the arc) determined by the edges in the observation fields 86A and 86B. The contact between the virtual pin and the virtual pin may be calculated and used as the distance between these contacts as L. Also, when the difference between the wafers is the same with respect to the shape in the notch portion NP, two of the virtual pins 82C are used. The contact surface spacing LM can be handled as a variable, and when the shape is stable, the spacing LM can be fixed. In addition, it is good also as a quadratic curve approximation. Moreover, the calculation so far is a case where the light receiving system of a two-dimensional image processing apparatus is installed in parallel with respect to the straight line (y = a), In some cases, the coordinate system rotates. , When there is a rotating coordinate system, it is necessary to correct a software ever.

By the above method, the rotational error Δθ and the offsets ΔX and ΔY of the wafer 6N can be accurately obtained with only one correction calculation. As described above, when the functions f '(x) and g' (x) after rotation correction are calculated, and the coordinates x ₃ ', y ₃ ' and the like of the measurement point after rotation correction are obtained, as shown in Fig. 15D, Similarly, correcting the approximation error in calculating the linear functions f '(x) and g' (x) as the point where the virtual pin touches actually improves the matching accuracy with the contact prealignment mechanism.

Next, FIG. 17A shows a state where the wafer 6 having the orientation flat portion FP is conveyed in the loading position in the 6 o'clock direction. Also in FIG. 17A, within the areas corresponding to the observation visual fields 86A, 86B and 86C of the actual processing target of the three image processing apparatuses 50, 51 and 52 of FIG. 4 and the observation visual fields 86A to 86C. For example, the virtual pins 82A-82C which represent the reference pin of the contact-type prealignment mechanism used in the previous exposure process by a virtual line are displayed. However, for easy viewing, the virtual pins 82A to 82C are enlarged and displayed at positions outside the corresponding observation visual fields 86A to 86C, respectively. In the observation fields 86A and 86B on the orientation flat portion FP of the wafer 6, the points where the virtual pins 82A and 82B and the outer circumference of the wafer come into contact with each other are assumed to be the virtual positions A1 and B1. Let the point where the virtual pin 82C in the circumferential observation field 86C contact is made into the virtual position E1.

In order to obtain these virtual positions A1, B1, and E1, for example, a reference wafer having a reference mark indicating a virtual position may be used as in the above example. Moreover, also in this example, the coordinate in the stage coordinate system of the rotation center 0 'of the center up 38 of FIG. 5B is calculated | required. Thereafter, the origin of the stage coordinate system (X, Y) is set to the rotation center 0 ', and a new coordinate system in which the straight line joining the virtual positions A1 and B1 is parallel to the X axis is a new coordinate system (X, Y; 0 ').

And the space | interval of the X direction of virtual position A1 and B1 is set to L1. In the new coordinate system (X, Y; 0 '), the coordinates of the virtual positions A1 and B1 are (-L1 / 2, c) and (L1 / 2, c), respectively, and the coordinates of the virtual position E1 are set. As (x8, y8), the relative coordinates in the observation visual fields 86A to 86C of the virtual positions A1, B1, and E1 are stored. If the Y coordinate in the new coordinate system (X, Y; 0 ') is y, as shown in Fig. 16A, the virtual positions A1 and B1 are located on a straight line (y = c).

At this time, under the illumination light from the bottom surface side of the wafer 6, the image of the pattern in the observation visual fields 86A to 86C is each as shown in Figs. 17B to 17D, and the wafer portion is a dark portion such as an oblique line, and the transparent portion is uneven. Therefore, the edge of the outer circumference of the wafer 6 can be detected as the boundary between the dark portion and the roll.

Next, a method of detecting the position and rotation angle of the wafer 6 when the wafer 6 actually exposed is set to the loading position, as shown in FIG. 17A. First, in the two observation fields 86A and 86B on the orientation flat portion FP of the wafer 6, as shown in Figs. 17B and 17C, the edges of the outer circumference corresponding to the boundary between the roll and the dark portion are located on the X axis. It is parallel almost straight. Thus, in the new coordinate system (X, Y; 0 '), the intersection between the straight line passing through the virtual positions A1 and B1 and perpendicular to the edge of the orientation flat portion FP and the edge is measured point C1 (-L1 / 2, y ₆ ) and (L 1/2 / 2y ₇ ). Finding the these measurement points C1, D1, by the average position of the edge of the predetermined range in the vicinity of the measurement points (C1 and D1) in respective coordinates (y ₆ and y _7), the detection accuracy is improved.

In the observation visual field 86C, the intersection of the straight line passing through the virtual position E1 and parallel to the X axis and the edge of the outer circumference of the wafer 6 is referred to as measurement point F1 (x ₉ , y ₈ ). Coordinates (x ₉ , y ₈ ) are coordinates in the new coordinate system, and thus, the edge of the outer circumference of the wafer 6 in the observation field 86C is represented by a straight line (x = x ₉ ).

Next, since the distance between the two measurement points C1 and D1 on the orientation flat part FP is (L1), the rotation error Δθ of the wafer 6 is as follows.

Δθ = (y ₇ -y ₆ ) / L1

Since the rotation error Δθ of the wafer 6 is determined by the orientation flat portion FP, the rotation center 0 'of the center up 38 is not known here even though the position of the center 0 of the wafer 6 is not known. The rotational error is corrected by correcting the rotational error? However, when the rotation center 0 '(0,0) is rotated about the axis, the coordinates (x ₉ , y ₈ ) of the measurement point F1 in the observation field 86C have an operation. Therefore, in the new coordinate system when the new coordinate system (X, Y; 0 ') is rotated by Δθ, the repair line and the edge of the edge of the wafer 6 corresponding to the edge of the corresponding wafer 6 are again from the virtual positions A1, B1, and E1. Find the intersections C1 ', D1', and F1 ', and find the coordinates of these intersections C1', D1 ', and F1' (-L / 2, y ₆ '), (L / 2, y ₇ '), (x ₉ ', recalculate y ₈ ). Then, the position shift amount (offset) (DELTA) X and (DELTA) Y of the wafer 6 to an X direction and a Y direction are computed from following Formula.

ΔX = x ₉ '-x ₈ ,

△ Y = (y ₆ '+ y ₇ ') / 2-c

These offsets [Delta] X, [Delta] Y may be corrected by adjusting the position of the wafer stage during search alignment as in the example shown in FIGS. 16A to 16D. In addition, as described in the position detection of the wafer having the notched portion shown in FIG. 16A, when there is an error between the edge of the wafer 6 obtained from the average value of the image data and the contact point between the actual virtual pin and the edge, As shown in Fig. 15D, data to be finally used in the image data may be found.

As described above, in the present example, after the detection of the edge position of the outer circumference of the wafer, the rotational error is generated by the platform 38 so that the rotational error is not corrected and redetection for confirmation is not necessary. Calculation is performed in advance by predicting the error to be made. Therefore, since there is no need for redetection, the wafer can be aligned in a non-contact manner at high speed.

However, when the lifting platform 38 is rotated and the wafer is mounted on the wafer holder 30, the vibration is at the start of vacuum adsorption, the error of the parallelism between the surface of the wafer and the surface of the wafer holder 30, and the like. The offset of the position and the offset of the rotation angle may be added. On the other hand, for example, the difference between the edge position of the wafer detected beforehand using the reference wafer or the like and the edge position of the actually mounted wafer during the alignment of the wafer after mounting the wafer is, for example, the alignment of the off axis system. It is good to detect by a sensor etc. and make it the system offset, and to raise to the calculation result of the offset (DELTA) X, (DELTA) Y, and rotation error (DELTA) (theta) of the position shift amount of a wafer.

Next, an example of the entire flow of the positioning operation in the projection exposure apparatus of this example will be described with reference to the flowcharts of FIGS. 1 and 2.

First, in step 101 of FIG. 2, the wafer 6 is conveyed by the transfer arm 21 along the slider 23 of FIG. 5B, and the vacuum adsorption of the transfer arm 21 is released while the loading position 21 is released. The lifting platform 38 is lifted by the expansion and contraction mechanism 35, and the wafer 6 is received on the lifting platform 38. At this time, the vacuum suction of each support pin 38a-38c of the platform 38 turns on (step 102). In addition, until this step, the wafer 6 is roughly aligned in the X direction, the Y direction, and the rotational direction (θ direction) based on the external standard through a mechanism including the turntable 60 of FIG. 5C. The wafer 6 has only an error of about 1 to 2 mm in the X direction, the Y direction, and about 5 ° in the rotational direction. At this time, the rotational error is corrected by the rotation of the turntable 60, and the error in the X-direction and the Y-direction is the loading when the wafer 6 is received from the transfer arm 21 by the lifting table 38. The position is corrected by adjusting the position in the X and Y directions. Even after such rough pre-alignment, a relatively large position shift amount and rotation error remain that the exposure apparatus main body, the mechanism including the turntable 60, and the slider 23 do not transmit mutual vibration. This is because the mechanism including the independently installed. In this case, the position shift occurs when the wafer is picked up from the slider 23 to the center up 38 due to different shaking conditions between the exposure apparatus main body portion and the mechanism including the slider 23. Error).

Next, in step 103, the edge position measurement of the outline of the wafer 6 is performed using the three two-dimensional image processing apparatuses 50 to 52 of FIG. In this case, as described with reference to FIGS. 17A to 17D, two observation fields 86A and 86B along the orientation flat portion FP of the wafer 6 and one observation field 86C of a circular outer circumference. In each case, the edge position is measured. In this regard, conventionally, in order to perform the alignment of the wafer 6, the wafer 6 is placed on three reference pins on the wafer holder 30. That is, the conventional prealignment was made by the contact system. On the other hand, the prealignment of this example can be called a non-contact system.

The three observation visual fields 86A to 86C of the present example are set to include virtual positions A1, B1 and E1 which are the same as the positions of the contact points with the three reference pins that are used in the contact manner, respectively. Therefore, even if the layer of the photosensitive material immediately before the wafer 6 of FIG. 17A (the layer just below the photosensitive material layer from now on) is exposed by the exposure apparatus which performs positioning by a contact system, There is an advantage that the matching, that is, the amount of positional displacement is small. In addition, even when it is impossible to set observation visual field 86A-86C to the position containing a virtual position, for example by arrangement | positioning of various sensors etc. around the projection optical system 3, the turn table of FIG. 5C is already demonstrated. The edge position of the wafer at the position of the reference pin can be accurately estimated from the measured value using the measurement results of the outline of the wafer using the 60 and the eccentric sensor 61.

Similarly, even in the case of the wafer 6N having the notch NP as shown in Fig. 16A, as described with reference to Figs. 16A to 16D, the virtual positions A, B, and E are processed by image processing. ) And the contact method can be matched by determining the amount of shift between the outer edge of the wafer 6N and the peripheral edge of the wafer 6N.

In addition, when it is not necessary to match with the exposure apparatus which pre-aligns by the contact system with respect to the wafer 6N which has the notch part NP, you may carry out using the method demonstrated by FIG. 8B, for example. . That is, in Fig. 8B, the position data is found in all ranges of the two edges 65A and 65B of the notch portion NP, and the virtual notch shape (by two approximation straight lines) is obtained by calculating a least-squares approximation from these position data. The V-shape prescribed | regulated) is calculated | required, and let the notch detection apparatus make the intersection of these two approximated straight lines. Thereby, the position detection of a wafer can be performed with high precision, regardless of the shape error of the notch part NP.

Then, in step 104, the position shift amount ΔX in the X direction and the position shift amount ΔY in the Y direction and the rotation error Δθ in the X direction based on the measurement result in step 103. To calculate. In this case, the rotational error Δθ is also regarded as a large amount of positional shift. In this example, as described with reference to FIGS. 17A to 17D, the observation visual fields 86A and 86B along the orientation flat part FP. ), The rotational error Δθ can be obtained from the detection result in the above), and the position shift amounts ΔX, ΔY in the X and Y directions can be obtained from the detection results in the observation visual fields 86A to 86B.

On the other hand, as shown in FIG. 16A-FIG. 16D, when the positioning object is the wafer 6N which has the notch part NP, the observation field 86C on the notch part NP, and two other observation fields ( By processing the detection result in 86A, 86B), the position shift amount (DELTA) X, (DELTA) Y, and rotation error (DELTA) (theta) in the X direction and a Y direction can be calculated | required.

Subsequently, in step 105, whether or not the calculated rotational error Δθ is a permissible range that can be corrected by the rotation of the platform 38 and the positional shift amounts ΔX and ΔY are vacuum-adsorbed by the wafer holder 30. Investigate whether it is a possible allowable range and go to step 109 if something is outside the allowable range. If any of the position shift amounts ΔX, ΔY, Δθ is out of the allowable range in the first prealignment, the process proceeds to step 110 again to perform the rough prealignment described above. Then, the wafer 6 is received from the lift table 38 to the slider 23 (conveying arm 21), the wafer 6 is returned on the turn table 60 of FIG. 5C, and rough alignment is performed. After that, the process returns to step 101 and the operation up to step 105 is repeated.

However, in step 105, when any one of the position shift amounts? X,? Y, ?? is out of the allowable range, it is recognized that something other than a simple position shift has occurred, and step 109 The process then proceeds to Step 111, where error information is issued and the instruction waits from the operator.

On the other hand, in step 105, if all of the position shift amounts ΔX, ΔY, and Δθ are within the allowable range, the platform 38 is lowered and the rotational error Δθ of the wafer is corrected (step 106). . Almost simultaneously with the contact of the wafer 6 with the wafer holder 30, the vacuum adsorption of the spindle portions 38a to 38c is turned off, and the vacuum adsorption on the wafer holder 30 is turned on. The wafer 6 is mounted on the wafer (step 107). Subsequently, a series of pre-alignment sequences are completed by adding the wafer shift amounts ΔX and ΔY as offsets to the search alignment positions described later and driving the wafer stage to move the wafer (step 108). . The process then proceeds to the sequence of alignment (search alignment and fine alignment) in FIG.

Comparing the above-described prealignment operation of the present example and the conventional prealignment operation described with reference to FIG. 1, in the conventional method of FIG. 1, after measuring the rotational error of the wafer 6 by, for example, a contact method, The rotation of the wafer is corrected by the θ rotation correction mechanism 8 of the wafer on the sample stage 9. Since the time required for this is 1 to 2 seconds, the rotation correction is performed in advance in the configuration of the present embodiment so as to lower the platform 38 and fall within an allowable error, and such time does not occur. However, it takes time to reload a plurality of wafers on the wafer holder near the head of the lot.The learning effect of averaging and correcting the error amount increases the number and number of wafers in the lot. The effect of the present invention is increased.

Here, the reloading of the wafer will be described. After mounting the wafer in the wafer holder 30 (step 107), if the alignment of the prealignment is basically matched between the exposure apparatuses, an error such as not finding a search mark at the time of the later alignment is not generated. Do not. Therefore, in a normal operation, you may move from step 108 of FIG. 2 to step 112 of FIG.

However, under the following conditions (1) to (3), the amount of position shift during prealignment becomes large, and there is a possibility that an error may occur during search alignment, for example.

① When the parallelism between the upper surface of the platform 38 and the upper surface of the wafer holder 30 deteriorates due to some reason, and rotation or offset error occurs when the wafer is delivered to the wafer holder 30.

(2) When there is no matching of prealignment with other exposure apparatus, a certain rotation or offset error (rotation or offset between the outline of the wafer and the search mark) occurs in units of lots.

(3) After exposure to the first layer of the wafer, "defects" occur in the measuring device of the external appearance of the wafer, and rotation and offset errors occur in units of wafers.

Therefore, when there is such a concern, the process shifts from step 108 of FIG. 2 to step 130A of FIG. 18 to check whether or not an error of search alignment has occurred, and when an error occurs, the wafer is wafered. It is preferable to remount the wafer returned to the wafer holder 30 after moving from the holder 30 to the lifting table 38 and correcting the rotation. Hereinafter, the reloading operation of the wafer of FIG. 18 will be described.

First, in step 130A of FIG. 18, a high-order search alignment is performed. As shown to FIG. 5A, the 1st search alignment 47A and the 2nd search alignment 47B for search alignment are formed on the wafer. In the high magnification search alignment, when the first search alignment 47A and the second search alignment 47B on the wafer are detected by the FIA alignment sensor 5A of FIG. 4, the magnification of the imaging optical system is set to be high magnification. Processes image information of only one screen. Next, in step 130B, it is determined whether or not the offset? X in the X direction, the offset? Y in the Y direction, and the rotation error? In the Y direction can be measured by the search alignment. This is judged whether or not the two search marks 47A and 47B can be detected by the image information for each screen obtained in step 130A. If the search marks 47A and 47B can be detected, the operation moves on to step 112 of FIG. Therefore, normally, since it moves to step 112 only by performing high magnification search of one screen of step 130A, the throughput hardly decreases.

On the other hand, when the search marks 47A and 47B are not detected in step 130B, the operation is performed in step 131 (mode 1), step 132 (mode 2), or step 133 (mode 3). Move to either. Mode 1, mode 2, and mode 3 are selected in advance. First, in step 131 of the mode 1, the search marks 47A and 47B are detected by setting the magnification of the alignment sensor 5A to low magnification and making the field of view wide. This is a low magnification image processing search alignment.

Next, in step 132 of the mode 2, the alignment sensor 5A has a high magnification, and the search marks 47A and 47B are detected while stepping the wafer stage to connect the screen. In step 133 of mode 3, for example, the coordinates (F _X1 , F _Y1 ) of the first search mark 47A are measured by the operator's assist (manual assist), and the offsets δX and δY are obtained. Subsequently, the Y coordinate F _Y2 of the second search mark 47B is measured, and the rotation error δθ is obtained from the difference between the two Y coordinates. After the steps 131 and 132, as shown in step 134, it is again determined whether or not the offset? X,? Y and the rotational error ?? can be detected, that is, whether the search marks 47A and 47B can be detected. When it cannot be detected, the process proceeds to step 133 of the mode 3, and when it is possible to detect δX, δY, δθ, the process proceeds to step 137.

Moreover, in step 135 following step 133, it is determined whether the obtained rotation error (delta) (theta) is the range which can be corrected in the platform 38. As shown in FIG. If it cannot be corrected (over the rotation allowable value), the process proceeds to step 136 and an error display is displayed. If corrected, the process proceeds to step 137. In step 137, the platform 38 is moved together with the wafer after the vacuum adsorption of the platform 38 and the vacuum adsorption of the wafer holder 30 and the vacuum adsorption of the wafer holder 30 are turned off. It raises and the rotation error (delta) (theta) of a wafer is correct | amended. This is the case of the first lot of wafers of one lot, and for the wafers of the second and subsequent wafers, the rotation error δθ is corrected in addition to the step 106 of FIG. Then, the vacuum suction of the wafer holder 30 is turned on, and the lifting platform 38 is lowered to turn off the vacuum suction of the lifting table 38. This causes the wafer to be remounted on the wafer holder 30. Next, in step 138, the search alignment position is corrected only by the offsets δX and δY for the first wafer. In the second and subsequent wafers, for example, correction of the offsets δX and δY is added to the search alignment position in step 108 of FIG. Thereafter, the operation proceeds to step 112 of FIG.

In addition, in the case of (1) and (3) described above, since they are changed in units of wafers, the correction in step 106 of Fig. 2 is not performed (error processing for the second and subsequent wafers is performed). On the other hand, in the case of ②, since it can be determined as the first wafer, the wafer processing after the second sheet in step 137 of FIG. However, when the cases of ①, ②, and ③ are combined, there is a possibility that an error is generated only with the first wafer. Therefore, if the measured result is fed back after the second sheet, the search may not be possible. In such a case, it is preferable to process by a learning function as mentioned above.

In the conventional projection exposure apparatus shown in FIG. 1, the θ rotation correction mechanism 8 of the wafer as a drive system exists between the sample table 9 on which the moving mirror 13 is mounted and the wafer 6. In the example, since there is no drive system between the moving mirror 13 and the wafer 6, the stability of the stepping accuracy is increased.

Next, in the sequence of alignment of FIG. 3, search alignment is performed first. However, it is a case where the detection range (capture range) of the alignment sensor to be used is large, and when pre-alignment precision is favorable, search alignment can be skipped and it can enter fine alignment. For example, in the LSA and FIA alignment sensors, the detectable range is wide, for example, up to ± 25 μm. On the other hand, the detectable range of the LIA type alignment sensor is only about 1 to 2 m. Therefore, if the alignment accuracy is ± 25 µm or less, when the LSA method or the FIA type alignment sensor is used, it is possible to shift to fine alignment without search alignment.

Therefore, in step 112 of Fig. 3, the type of alignment sensor to be used is determined, and when using the LIA alignment sensor, the procedure shifts to the search alignment sequence of step 103 or less, and the alignment of the LSA method or the FIA method. When using the sensor, the flow advances to step 121 to determine whether the pre-alignment accuracy is outside the detectable range of the alignment sensor to be used, that is, whether or not to perform the search alignment. When the search alignment is performed, the process proceeds to step 113, and when the search alignment is not performed, the process proceeds to step 126. FIG.

Next, the search alignment will be described. On the wafer, marks for search alignment are formed. Also on the wafer 6 of this example, as shown in FIG. 5A, the search mark 45X of the X-axis which consists of the line-and-space pattern formed by the predetermined pitch in the X direction, and the line-and-space pattern formed by the predetermined pitch in the Y direction The first search mark 47A for the FIA method which combined the search mark 45Y of the Y-axis which consists of these is formed. Further, the second search mark 47B for the FIA method combining the search mark 44X on the X-axis and the search mark 44Y on the Y-axis at a position spaced apart from the first search mark 47A by a predetermined distance in the Y-direction. Is formed. In this example, in order to detect the positions of the two search marks 47A and 47B, the FIA alignment sensor 5A of FIG. 4 is used. As will be described later, in order to detect the rotation angle of the wafer 6, the alignment sensor 5B of the FIA method is detected. Therefore, in order to distinguish two alignment sensors, the alignment sensor 5A is called "FIA microscope 5A" below, and the alignment sensor 5B is called "theta microscope 5B".

In addition, wafer marks for fine alignment (hereinafter, referred to as " fine marks ") are also formed in all shot regions on the wafer 6, respectively. Specifically, in FIG. 5A, in one shot region SA representative of all the shot regions on the wafer 6, the fine marks 46X of the X axis on the point sequence extending in the Y direction and the Y-axis fine extending in the X direction are illustrated. The mark 46Y is also formed. These fine marks 46X and 46Y are marks detected by the LSA alignment sensor in the TTL alignment sensor 4 in FIG. 4. Moreover, as a fine mark actually used, there are also a mark for the LIA method and a mark for the FIA method depending on the process.

9A and 9B show the mark arrangement on the wafer 6. 9A and 9B, the first search mark 47A is in a street line area surrounded by four short areas 48A to 48D, and the four search areas 47A to 49D are also different from the second search mark 47B. It is in the street line area surrounded by). Further, the circular observation field 5Aa is the effective observation field of the FIA microscope 5A of FIG. 4, and the observation field 5Ba at a position away from the X direction is the effective observation of the θ microscope 5B of FIG. 4. Vision.

Subsequently, in order to perform the search alignment, the wafer stage is driven in step 113 to move the first search mark 47A within the effective observation field of view 5Aa of the FIA microscope 5A, as shown in FIG. 9A. do. In this state, there is no second search mark 47B in the observation field 5Ba of the θ microscope 5B, the end portions of the shot regions 49A and 49B, and the street line region 70. Subsequently, it is determined whether or not the wafer 6 to be exposed is the first wafer in this lot (step 114). In the case of the first wafer, the process proceeds to step 115 and the FIA microscope 5A is used. 1 Coordinates F _X1 and F _{Y1 in} the X and Y directions of the search mark 47A are detected.

Here, an example of the detection method in step 115 is demonstrated with reference to FIGS. 10A-10C. FIG. 10A shows the detection range 68 actually picked up by the imaging element within the observation field of view of the FIA microscope 5A. In this FIG. 10A, two independent indicator marks 66X1 and 66X2 corresponding to the X direction and two independent indicator marks 66Y1 and 66Y2 corresponding to the Y direction are displayed in the detection range 68. These indicator marks 66X1, 66X2, 66Y1, 66Y2 are arranged on the surface and the conjugate surface of the wafer in the FIA microscope 5A of FIG. 4, and are illuminated with illumination light independent of the illumination light for detecting the marks on the wafer. have. In addition, in the FIA microscope 5A, an imaging device for the X-axis scanning in the direction corresponding to the X direction and an imaging device for the Y-axis scanning in the direction corresponding to the Y direction are provided in parallel. The imaging element scans in the direction crossing the indicator marks 66X1 and 66X2 and outputs the imaging signal SX1 shown in FIG. 8C.

The signal portion 67X in Fig. 8C corresponds to the search mark 45X on the X-axis, and the image pickup signal SX1 is subjected to analog / digital (A / D) change and image processing, thereby making reference to the indicator marks 66X1 and 66X2. The X coordinate of the first search mark 47A is detected.

Similarly, the imaging element for Y-axis scans in the direction crossing the index marks 66Y1 and 66Y2, and outputs the imaging signal SY1 shown in FIG. 8B. The signal portion 67Y in Fig. 8B corresponds to the search mark 45Y on the Y-axis, and the image processing of the imaging signal SY1 makes the Y of the first search mark 47A based on the indicator marks 66Y1 and 66Y2. Coordinates are detected.

However, instead of the indicator mark, for example, when using a predetermined pixel or image pickup tube of the image pickup device in the FIA microscope 5A, the position detection may be performed based on the scanning start point or the like.

Thereafter, the wafer stage is driven to move the second search mark 47B in the observation field of view 5Aa of the FIA microscope 5A (step 116), as shown in Fig. 9B, to give the FIA microscope 5A. Then, the coordinates F _X2 and F _{Y2 in} the X-direction and the Y-direction of the second search mark 47B are detected (step 117). Subsequently, in step 108, the rotation angle θ, and the offset ((F _X1 + F _X2 ) / 2, in the wafer stage coordinate system (X, Y) with reference to the positions of the two search marks 47A, 47B. A new coordinate system (hereinafter referred to as "XYθ transform coordinate") (Xp, Yp) to which (F _Y1 + F _Y2 ) / 2) is introduced is introduced. The rotation angle θ in this case is represented by the following equation with L as the interval between the two search marks 47A and 47B.

The new XYθ transformation coordinates Xp and Yp are expressed by the following equations with respect to the coordinate systems X and Y of the wafer stage.

Next, in step 119, the wafer stage is driven along the XYθ conversion coordinates Xp and Yp, and as shown in FIG. 9A, the first search mark in the effective observation field of view 5Aa of the FIA microscope 5A. Move 47A again. In step 120 which follows, the pattern (street line etc.) which exists in the observation visual field 5Ba of (theta) microscope 5B in the state where step 119 is complete | finished, and memorize | stored the image | photographed image, Remember the features. This operation will be described with reference to FIGS. 11A and 11B.

11A shows an image of the detection area 68 of the FIA microscope 5A in the state where step 119 is finished. FIG. 12A shows an imaging signal SY1 obtained by scanning the image of FIG. 11A in the Y direction with an imaging device along the Y axis. Thus, when it moves along XY (theta) conversion coordinate, the center of the 1st search mark 47A is set to the center of the detection area 68 of FIA microscope 5A. In addition, corresponding to the search mark 45Y for the Y axis of the first search mark 47A, the imaging signal SY1 in FIG. 12A is peaked downward at positions Y1, Y2, and Y3. Thus, the Y coordinate YA obtained by (Y1 + Y2 + Y3) / 3 is detected as the position in the Y direction of the first search mark 47A.

In contrast, FIG. 11B shows an image in the detection area 69 of the θ microscope 5B in the state where the step 119 is finished. 11B, the pattern 71A in the shot region 49A and the shot region 49B, respectively, above and below the street line region 70 between two edge portions 70a and 70b extending in the X direction. There is an inner pattern 71B. And since the straight line which the value of the Xp axis in XY (theta) transformation coordinate, ie, the coordinate Yp becomes 0, exists in the street line area | region 70 from the determination method of XY (theta) transformation coordinates Xp and Yp of this example, the straight line Denotes an imaginary straight line 70c of a single-dot chain line. In this example, the image pickup signal SY2 shown in FIG. 12B is obtained by scanning the image of FIG. 11B in a direction corresponding to the Y direction (which is considered to be almost Yp direction). In this FIG. 12B, two peak positions SR1 and SR2 in the downward direction correspond to the Y coordinates of the edge portions 70a and 70b of FIG. 11B, respectively. Thus, the position YB (that is, the position at which Tp = 0) on FIG. 12B corresponding to the virtual straight line 70c of FIG. 11B is obtained.

In this example, the two edges of the virtual straight line 70c and the street line area 70 in Fig. 11B are detected by detecting the respective intervals ΔSR1 and ΔSR2 between the position YB and the positions SR1 and SR2 on both sides. The interval in the Y direction with the sections 70a and 70b is obtained and stored in the central control system 18 of FIG. Again, the signal strength at the positions SR1, SR2 and other patterns in the imaging signal SY2 in FIG. 12B so that the edge portions 70a, 70b and other patterns (patterns 71A, 71B, etc.) can be correctly identified. The characteristics of the signal strength at the portion corresponding to and the distance between the positions SR1 and SR2 and other patterns are obtained and stored in the central control system 18. It is assumed that the positional relationship between the two search marks 47A and 47B and the street line area 70 is the same for the wafers in one lot, and for the wafers after the second sheet, the position SR1, from the image pickup signal SY2 in FIG. SR2) is identified and the position YB at which the coordinate Yp becomes zero is detected based on the positions SR1 and SR2.

In the above-described example, the positions SR1 and SR2 of the two edge peaks are identified from the imaging signal SY2 in FIG. 12B. The waveforms of the imaging signal SY2 are subjected to A / D conversion and stored in the following. A method of detecting the position YB at which the coordinate Yp becomes 0 by correlating with the waveform of the image pickup signal obtained from the wafer can also be considered. In the above-described example, as shown in FIG. 11A, the center of the first search mark 47A is set to the center of the detection area 68 of the FIA microscope 5A. As shown in FIG. 11B, the coordinate Yp is shown. The virtual straight line 70c which becomes 0 may set the new coordinate system which becomes the center of the detection area 69 of the (theta) microscope 5B.

Next, the process proceeds to step 122 and fine alignment is performed by detecting the position of the fine marks 46X and 46Y placed in the predetermined shot area on the wafer 6. Here, fine alignment is performed, for example, by the enhanced global alignment (hereinafter referred to as "EGA") method disclosed in Japanese Patent Laid-Open No. 61-44429. In other words, by driving the wafer stage based on the XYθ transformation coordinates, the alignment marks 4 are used for the fine marks of the X-axis and Y-axis laid on the predetermined number of shot regions (sample shots) selected on the wafer 6. The coordinates are detected, and the detection results are statistically processed to calculate the array coordinates in the XYθ transformation coordinates of all the shot regions on the wafer 6.

Next, in step 123, the wafer stage is sequentially driven based on the array coordinates of the respective shot regions calculated in the fine alignment, and the respective shot regions on the wafer 6 are positioned at the exposure positions to respectively reticle 1 ) Is projected and exposed. At this time, in the final position adjustment, the stage on the reticle 1 side may be driven while the wafer stage is stopped to correct the relative positional displacement between the reticle and the wafer. Accordingly, after the exposure to the wafer 6 is finished and the wafer 6 is unloaded, the steps 101 to 108 in FIG. 2 are performed on the wafer to be exposed next in this lot, and prealignment is performed. Thereafter, steps 112 and 113 of FIG. 3 are performed on the wafer to proceed to step 114.

Since the current wafer is the second or later in the lot, the operation shifts from step 114 to step 124, and the FIA microscope 5A shows the first search mark 47A in the state as shown in Fig. 9A. The X coordinates and the Y coordinates F _X1 and F _Y1 are detected, and the Y coordinates SR1 and SR2 of both edge portions of the street line region 70 are detected by the θ microscope 5B. At this time, the image data stored in step 120 identifies the edge portions of both sides of the street line area 70 and other patterns. In addition, from these Y coordinates SR1 and SR2, the Y coordinate YB when the value of the Yp axis which is a new coordinate system becomes 0 is obtained.

In step 125, new XYθ transformed coordinates Xp and Yp, which are given rotation angles θ and offsets Ox and Oy, are introduced into the coordinate systems X and Y of the wafer stage. The rotation angle θ in this case is expressed by the following equation using the distance L 'between the detection center of the FIA microscope 5A and the detection center of the θ microscope 5B, and the measurement values described above.

θ = arctan {(YB-F _Y1 ) / L '}

Also, if the interval L between two search marks 47A and 47B is used, the coordinates F _X2 and F _Y2 of the second search mark 47B are substantially (F _X1 + L, F _Y2 + θ) in the first approximation. L). Thus, using the offsets Ox and Oy as the coordinates of the center points of the two search marks 47A and 47B, the new XYθ conversion coordinates Xp and Yp are described above with respect to the coordinate systems X and Y of the wafer stage. It is represented by (2). Next, alignment and exposure of the wafer are performed at steps 122 and 123. In this case, the second and subsequent wafers are simultaneously measured by FIA microscope 5A and θmicroscope 5B at the time of search alignment, and the position and rotation angle in the X and Y directions are determined at one time. This shortens the throughput.

Next, the case where the alignment sensor to be used is the LSA method or the FIA method and the search alignment is not performed will be described. In this case, the operation shifts from step 121 to step 126, where priority mode selection is made. In this example, there are two modes, rough mode and fine mode. In this example, the precision of the prealignment shown in Fig. 2 is about 20 µm, for example, three times the standard deviation (3σ), so that the precision at the start of the fine alignment is sufficient (rough mode) and insufficient (fine Mode). Therefore, in the rough mode, the process proceeds to step 122, where the fine alignment of the EGA method is performed by measuring the position of a predetermined number of sample shots with the accuracy of prealignment.

In the fine mode, on the other hand, using the designated alignment sensor, the coordinates of the fine marks on the X-axis and the Y-axis laid in two separated shot regions on the wafer are measured (step 127). Similarly, the XYθ converted coordinates are obtained (step 128). After that, the wafer stage is driven along the XYθ conversion coordinates, and the measurement is performed by setting the center of the fine mark to almost the center of the detection area of the alignment sensor after the third sample shot or the first shot area, respectively. (Step 129) and exposure is performed in step 123 after completion | finish of a measurement.

In general, since the detectable range is wide in the LSA method and the FIA method, a rough mode that is advantageous in terms of throughput is selected. However, for example, when it is required to perform measurement with high accuracy without removing the influence of the intra-distortion and magnification error of the FIA type image processing system, the screen distortion is predicted and corrected in advance, or fine mode is preferable. do. In addition, although the mode selection may be selected in advance according to the process, the mode selection may be automatically performed depending on whether the alignment accuracy is good or bad. In addition, in step 124 of FIG. 3, when a pattern does not exist in the detection range of (theta) microscope 5B, it determines with no pattern to be detected, and after performing steps 115-118 performed with respect to the head wafer, The sequence to be moved to step 122 is automatically selected.

In addition, in embodiment mentioned above, it was presupposed that it can shift to search alignment or fine alignment after completion of prealignment. However, for example, the exposure to the first photosensitive material on the wafer is performed by another exposure apparatus, and the second photosensitive material is exposed to the second photosensitive material by the projection exposure apparatus of FIG. In the case where the alignment position of the alignment sensor and the like are not matched, the search marks 47A and 47B are so large that the search marks 47A and 47B do not exist in the observation field of the alignment sensor even if the wafer is accurately positioned on the basis of the outline. 47B) may be out of the X-, Y-, and rotational directions. In such a case, after the completion of the second prealignment on the first wafer in the lot, the operator may wait for the instruction of the operator and perform manual measurement of the positions of the search marks 47A and 47B by the operator. Based on the results, the offset of the rotation angle for the center-up rotation mechanism and the offset of the search alignment position in the X direction and the Y direction are calculated and corrected. After the prealignment, it is possible to automatically shift to the search alignment or the fine alignment in FIG. 3.

Moreover, in the above-mentioned aspect, although the two-dimensional image processing apparatus as an edge position sensor of a wafer is provided in the wafer loading position, it may be difficult to arrange | position an edge position sensor in a loading position depending on a wafer size. In such a case, the wafer may be moved to a position where the external shape can be measured by driving the wafer stage with the wafer mounted on the platform 38, or the wafer stage is driven after the wafer is once placed on the wafer holder 30. It may move to the position which can be measured externally.

In addition, the present invention can be applied to not only a step-and-repeat type exposure apparatus but also an exposure apparatus such as a step-and-repeat type exposure apparatus.

Thus, this invention is not limited to the aspect of embodiment mentioned above, A various structure can be taken in the range which does not deviate from the summary of this invention.

According to the positioning method of the present invention, immediately after the photosensitive substrate (wafer) is received, the edge measurement of the outer peripheral portion of the photosensitive substrate is performed in a non-contact manner using a two-dimensional image processing apparatus, so that the photosensitive substrate is dropped onto the substrate stage. In parallel with this, the rotation error of the photosensitive substrate can be corrected (pre-alignment). Therefore, the time required for prealignment can be shortened. In addition, since there is no need to provide a rotation mechanism on the substrate stage side, the structure of the substrate stage (wafer stage) can be simplified, and the rigidity and weight reduction of the substrate stage can be thereby achieved, for example. There is an advantage that the photosensitive substrate can be positioned at a high speed and with high accuracy when mounting the photosensitive substrate from the loader system of the substrate to the substrate stage.

In addition, a virtual position corresponding to a reference position when positioning the photosensitive substrate on the substrate stage is set by a contact positioning method within the observation field of the two-dimensional image processing system, and the measurement point on the photosensitive substrate is determined from the corresponding virtual position. Since the photosensitive substrate is positioned on the basis of the amount of position shift, high matching at the time of performing rough positioning (at the time of the alignment) with other exposure apparatuses performing contact positioning (prior alignment). Precision can be obtained.

In this case, when the cutout portion of the outer peripheral portion of the photosensitive substrate is one wedge-shaped cutout portion, when the measurement points by the two-dimensional image processing system are set at one cut portion and two outer peripheral portions of the photosensitive substrate, By detecting the position of the photosensitive substrate at these three measurement points, the rotation angle and the two-dimensional position of the photosensitive substrate can be specified.

On the other hand, when the cutout portion of the outer peripheral portion of the photosensitive substrate is one cutout portion having a flat edge, when the measurement points by the two-dimensional image processing system are set at two cutout portions and one outer peripheral portion of the other photosensitive substrate, By detecting the position of the photosensitive substrate at three measurement points, the rotation angle and the two-dimensional position of the photosensitive substrate can be specified.

In the case of predicting the position of the photosensitive substrate when the photosensitive substrate is mounted on the substrate stage, the photosensitive substrate is positioned by the contact type positioning method when the photosensitive substrate is directly mounted on the substrate stage through the substrate raising means. The rotational error and the amount of position shift are obtained, and the rotational error is corrected when the photosensitive substrate is mounted on the substrate stage through the substrate raising means. After mounting the photosensitive substrate on the substrate stage, when correcting the positional shift amount through the substrate stage, the configuration of the substrate stage can be simplified and positioning of the photosensitive substrate can be performed at high speed.

Claims (24)

A positioning method for positioning a substrate having a cutout on a portion of its outer circumference on a two-dimensionally movable substrate stage, (1) transferring the substrate to a sorghum position above the substrate stage; (2) non-contact measurement of the position of the measurement point in the said cutting part of the outer peripheral part of the said board | substrate and the position of the other measuring point of the outer peripheral part of the said board | substrate, respectively, at the said delivery position; And (3) A step of calculating the rotational error of the substrate based on the measurement result. The method of claim 1, And rotating the substrate to offset the calculated rotational error before mounting the substrate on the substrate stage. The method of claim 1, A positioning method of a substrate, characterized by detecting the position measurement of the measurement point by a two-dimensional image processing method. A positioning method for positioning a substrate on a two-dimensionally movable substrate stage, (1) forming first and second search marks representing two-dimensional positions on the substrate, respectively; (2) detecting the two-dimensional positions of the first and second search marks on the first substrate, respectively; (3) calculating a rotational error of the first substrate based on the position detected in the step (2); (4) In parallel with detecting the two-dimensional position of the first search mark, at least one-dimensional position of a pattern separated by a predetermined distance from the first search mark on the first substrate is detected and stored. Step to do; And (5) The step (4) of the pattern spaced apart from the first search mark by a predetermined distance in parallel with detecting the two-dimensional position of the first search mark on the second substrate on the substrate stage. Detecting a position shift amount from the position stored in the step S, and calculating a positioning error of the second substrate based on the position shift amount. The method of claim 4, wherein A step for positioning a substrate, wherein the steps (2) to (5) are performed for the first substrate in one lot. A positioning method for positioning a substrate having a cutout on a portion of its outer circumference on a two-dimensionally movable substrate stage, (1) transferring the substrate to a sorghum position above the substrate stage; (2) measuring the position of the measurement point in the cutout portion of the outer peripheral portion of the substrate and the position of another measurement point in the outer peripheral portion of the substrate at a non-contact position, respectively, using a two-dimensional image processing system; (3) setting a virtual position corresponding to a reference position when positioning the substrate on the substrate stage in a contact-type positioning method within the observation field of the two-dimensional image processing system; And (4) a step of predicting the position of the substrate when the substrate is mounted on the substrate stage by using the shift amount from the virtual position of the position of the measurement point measured by the two-dimensional image processing system; Positioning method of a substrate comprising a. The method of claim 6, The prediction of the position of the substrate in the step (4) is such that not only the amount of deviation from the virtual position of the position of the measurement point measured by the two-dimensional image processing system, but also the desired angle by rotating the substrate at a desired angle. And the coordinates of the rotation center previously measured with respect to the substrate lifting means for setting on the substrate stage from the substrate. The method of claim 6, And positioning the substrate at the same position as when the substrate is positioned on the substrate stage by the contact positioning method based on the predicted result. The method of claim 6, The cut portion of the outer peripheral portion of the substrate is a cut portion of one wedge shape, The measurement method of the board | substrate of Claim 2 characterized by setting the said measuring point by the said two-dimensional image processing system at two places of the said cutting part and other outer peripheral parts of the said board | substrate. The method of claim 6, The cut portion of the outer peripheral portion of the substrate is one cut portion with a flat edge, The measuring point of the substrate according to the two-dimensional image processing system is set at one position at the cutout portion and at two other peripheral portions of the substrate. The method of claim 6, When the substrate is mounted on the substrate stage, when the position of the substrate is predicted, when the substrate is positioned on the substrate stage through the substrate lifting means as it is, by the contact positioning method. Calculating a rotational error and an amount of position shift of the; And Correcting the rotational error when mounting the substrate on the substrate stage through the substrate lifting means, mounting the substrate on the substrate stage, and correcting the positional displacement amount through the substrate stage. Positioning method of a substrate comprising a. The method of claim 7, wherein When the substrate is mounted on the substrate stage, when the position of the substrate is predicted, when the substrate is positioned on the substrate stage through the substrate lifting means as it is, by the contact positioning method. Calculating a rotational error and an amount of position shift of the; And Correcting the rotational error when mounting the substrate on the substrate stage through the substrate lifting means, mounting the substrate on the substrate stage, and correcting the positional displacement amount through the substrate stage. Positioning method of a substrate comprising a. In the positioning method of positioning the board | substrate which has a cut part in a part of outer periphery with a board | substrate stage, Conveying the substrate to a first position above the substrate stage; Detecting a position of the cut portion of the substrate and a position of an outer peripheral portion different from the cut portion of the substrate at the first position; Detecting a position error of the substrate at the first position based on the detection result; And And correcting the positional error of the substrate while moving the substrate from the first position to the substrate stage. The method of claim 13, And the step of detecting the position of the cut portion of the substrate and the position of the outer peripheral portion different from the cut portion of the substrate is performed in a non-contact manner. The method of claim 13, And the positional error of the substrate is a rotational error of the substrate. In the exposure method which positions a board | substrate which has a cut part in a part of outer periphery by a board | substrate stage, and exposes a pattern to the said board | substrate, Conveying the substrate to a first position above the substrate stage; Detecting a position of the cut portion of the substrate and a position of an outer peripheral portion different from the cut portion of the substrate at the first position; Detecting a position error of the substrate at the first position based on the detection result; Correcting a positional error of the substrate while moving the substrate from the first position to the substrate stage; And And exposing said pattern to said substrate. The method of claim 16, The step of detecting the position of the cut portion of the substrate and the position of the outer peripheral portion different from the cut portion of the substrate is performed in a non-contact manner. The method of claim 16, The positional error of the substrate is a rotational error of the substrate. A positioning device for positioning a substrate having a cutout in a portion of its outer circumference with a substrate stage Conveying means for conveying the substrate to a first position above the substrate stage; First detecting means for detecting a position of the cut portion of the substrate and a position of an outer peripheral portion different from the cut portion of the substrate at the first position; Second detection means for detecting a positional error of the substrate at the first position based on a detection result of the first detection means; And And correction means for correcting a positional error of the substrate while moving the substrate from the first position to the substrate stage. The method of claim 19, And the first detecting means has a non-contact detecting device. The method of claim 19, And the second detecting means detects a rotational error of the substrate. In the exposure apparatus which has a board | substrate stage which mounts the board | substrate which has a cutting part in a part of outer periphery, and exposes a pattern to the said board | substrate, Conveying means for conveying the substrate to a first position above the substrate stage; First detecting means for detecting a position of the cut portion of the substrate and a position of an outer peripheral portion different from the cut portion of the substrate at the first position; Second detection means for detecting a positional error of the substrate at the first position based on the detection result; Correction means for correcting a positional error of the substrate while moving the substrate from the first position to the substrate stage; And And an exposure means for exposing the pattern to the substrate. The method of claim 22, The first detecting means has a non-contact detecting device. The method of claim 22, And the second detecting means detects a rotational error of the substrate.