JP6087792B2 - Shape measuring device - Google Patents

Description

  The present invention relates to a technique for measuring the surface shape of a wafer using a Shack-Hartmann sensor.

  The present invention relates to a technique for measuring the surface shape of a wafer using a Shack-Hartmann sensor.

  When a device is manufactured from a wafer made of silicon, sapphire, or the like, a very fine wiring design of nanometer order is required. In order to realize this, there is an increasing demand for flatness of the surface shape of the wafer. Therefore, a very fine measurement accuracy is required also in a measuring apparatus for inspecting the flatness of a wafer.

  In the wafer measuring apparatus, it is necessary to measure the surface shape of the wafer, particularly a deviation shape from a plane in the vicinity of an edge called a roll-off portion (for example, a range of 1 to 5 mm from the edge end).

  FIG. 18 is a diagram for explaining a problem when a roll-off portion of a wafer is measured using an SH sensor that is a Shack-Hartmann wavefront sensor. The surface of the wafer is divided into a flat portion, a roll-off portion, and a chamfered portion from the center toward the edge.

  When measuring the surface shape of the roll-off part, it is necessary to determine the position coordinates of the roll-off part on the wafer. Usually, the coordinates are set on a line (radial direction) connecting the center and the edge of the wafer with reference to the edge position of the wafer. At this time, it is necessary to realize the positioning of the coordinates with high accuracy (for example, 10 μm or less).

  When this is realized, the following two problems arise. Problem 1) The SH sensor receives reflected light from the wafer of light source light and measures the surface shape. The reflected light from the roll-off part or the flat part can reach the imaging surface of the SH sensor. Therefore, the SH sensor can measure the surface shapes of the flat surface portion and the roll-off portion. However, since the reflected light reflected on the edge side from the chamfered portion is reflected toward the outside of the edge, it does not reach the imaging surface of the SH sensor, and the SH sensor cannot recognize the surface shape of this region. Therefore, there is a problem that the SH sensor cannot accurately determine the coordinates of the surface shapes of the flat surface portion and the roll-off portion when the edge position is used as a reference.

  Problem 2) The lens pitch of the lens array constituting the SH sensor is generally around 100 μm. Therefore, even if the SH sensor can detect the edge position, there is a problem that the edge position cannot be recognized with higher resolution. The reason why the lens pitch is about 100 μm is that the pixel pitch of the image sensor that constitutes the SH sensor is about 5 μm, and it is necessary to secure the lens pitch about 20 times or more with respect to the pixel pitch due to the principle of the SH sensor. Because there is.

  Therefore, in Patent Document 1, in the shape measuring apparatus that calculates the surface shape data of the wafer using the SH sensor, the edge position of the wafer to be measured is detected by an imaging device provided separately from the SH sensor. It is disclosed that surface shape data is obtained based on an edge position.

JP 2011-226989 A

  However, in the method of Patent Document 1, the surface shape data is calculated on the assumption that the distance between the measurement region where the SH sensor measures the surface shape and the edge position is known. However, the distance between the measurement region and the edge position is calculated. Varies from wafer to wafer. For this reason, the method of Patent Document 1 cannot accurately specify the distance from the edge position to the measurement area, and cannot accurately calculate the surface shape data of the measurement area based on the edge position. is there.

  An object of the present invention is to provide a shape measuring apparatus that accurately calculates surface shape data of a measurement region when the edge position of a wafer to be measured is used as a reference.

(1) A shape measuring apparatus according to an aspect of the present invention is a shape measuring apparatus that measures the surface shape of a predetermined measurement region on the surface of a wafer, and includes a light source that irradiates measurement light, and the wafer of the measurement light. A Shack-Hartmann-type SH sensor that receives the reflected light of the wafer and measures the surface shape of the wafer, an edge measurement unit that measures the edge of the wafer, and a reference pattern formed at a reference position that is a predetermined distance away from the edge. The surface shape of the reference sample is measured by the SH sensor, the reference position detection unit for detecting the reference position, the edge of the reference sample is measured by the edge measurement unit, and the edge position of the reference sample is the first position. An edge position detection unit that detects an edge position; and an edge measurement unit that measures an edge of a target wafer that is a measurement target. The second edge position is detected, the inter-edge distance calculation unit for calculating the inter-edge distance of the first and second edge positions, and the SH sensor is configured to measure the surface shape of the measurement region of the target wafer, A surface shape calculation unit that calculates surface shape data of the measurement region with a reference position as a reference coordinate, and a coordinate having a reference coordinate that is a position away from the reference position by a distance obtained by adding the distance between the edges to the specified distance A system including a coordinate conversion unit for converting the surface shape data.

  According to this configuration, the reference sample in which the reference pattern is formed at the reference position that is a predetermined distance away from the edge is measured by the SH sensor, and the reference position is detected. Further, the edge of the reference sample is measured by the edge measuring unit, and the first edge position that is the edge position of the reference sample is detected. This completes the calibration process.

  Next, the edge of the target wafer is measured by the edge measuring unit, the second edge position which is the edge position of the target wafer is detected, and the distance between the edges of the first and second edge positions is obtained. Next, the measurement area of the target wafer is measured by the SH sensor, and the surface shape data of the target wafer is calculated using the reference position of the measurement area as the reference coordinates.

  Next, the surface shape data is coordinate-transformed into a coordinate system in which a position that is separated from the reference position by a distance obtained by adding the distance between edges to the specified distance is a reference coordinate. Here, the distance obtained by adding the inter-edge distance to the specified distance represents the distance from the reference position to the second edge position, so this coordinate system is a coordinate system with the second edge position as the reference coordinate. Therefore, it is possible to accurately calculate the surface shape data of the measurement region when the edge position of the target wafer is used as a reference.

(2) The edge measurement unit is a camera installed on the opposite side of the wafer with respect to the light source, and the measurement is performed so that a part of light passes outside the edge of the wafer and is guided to the camera. reflect light, and passes through the reflected light from the wafer may further include a half mirror rather guiding the SH sensor.

  According to this configuration, the surface shape and the edge of the measurement region can be measured with one light source.

  (3) It is preferable that the light source outputs incoherent light.

  In this case, it is possible to prevent the generation of an interference fringe pattern due to diffracted light generated at the edge portion of the microlens constituting the SH sensor. Therefore, when the configuration for obtaining the position of the center of gravity of the luminance distribution as an imaging point in the block in the imaging surface block corresponding to a certain microlens is obtained, the obtained imaging point can be obtained geometrically. It is possible to prevent deviation from the original image formation point.

  (4) The light source includes a light source element that outputs the incoherent light, a pinhole provided on an output side of the light source element, and a lens that converts light passing through the pinhole into parallel light. Also good.

  According to this configuration, the pinhole is provided in front of the lens. Thereby, since the light as if it was output from the point light source is incident on the lens, the lens can more accurately convert the light output from the light source element into parallel light.

  (5) You may further provide the condensing lens provided between the said light source element and the said pinhole.

  According to this configuration, the proportion of light that passes through the pinhole in the light output from the light source element can be increased. As a result, the conversion efficiency of the light output from the light source element into parallel light can be increased.

  (6) The light source includes a first light source for an SH sensor and a second light source for an edge measuring unit installed at a position different from the first light source, and the edge measuring unit includes the second light source. The camera may be installed on the opposite side of the wafer.

  According to this configuration, the first light source for the SH sensor is irradiated to the measurement region, the SH sensor is caused to measure the surface shape of the measurement region, and the second light source for the edge measurement unit is irradiated to the edge to the edge measurement unit. Edges can be measured.

  (7) The edge measurement unit may be a laser displacement meter.

  According to this configuration, since the edge is measured by the laser displacement meter, the edge can be measured with an accuracy equal to or lower than the resolution of the SH sensor.

  (8) The edge measurement unit may be a contact displacement meter.

  According to this configuration, since the edge is measured by the contact type displacement meter, the edge can be measured with an accuracy equal to or lower than the resolution of the SH sensor.

  (9) The reference position detection unit may detect the reference position with a resolution smaller than the resolution of the SH sensor.

  According to this configuration, the reference position can be detected with a resolution smaller than the resolution of the SH sensor.

  (10) The reference pattern is a mask pattern, and further includes a moving unit that moves the reference sample in a direction parallel to the surface, and the reference position detection unit has the reference sample shorter than the resolution of the SH sensor. Each time the distance is moved, the surface shape of the reference sample may be measured by the SH sensor, and the reference position may be detected based on the measurement result.

  According to this configuration, each time the reference sample is moved a shorter distance than the resolution of the SH sensor, the surface shape of the reference sample is measured by the SH sensor, so that the reference position is detected with a resolution lower than the resolution of the SH sensor. be able to.

  (11) The reference pattern is a periodically changing waveform pattern, and the reference position detection unit causes the SH sensor to measure the waveform pattern once and detects the reference position based on a measurement result. Also good.

  According to this configuration, since the waveform pattern that periodically changes is formed as the reference pattern in the reference sample, the reference position can be detected by only causing the SH sensor to measure the waveform pattern once.

  (12) The reference sample may be composed of a plate-like mirror object that simulates the shape of the wafer.

  According to this configuration, since the reference sample is created using a plate-like mirror object different from the wafer, the reference sample can be created at low cost.

  (13) The measurement area may be a roll-off part.

  According to this configuration, the surface shape of the roll-off part can be accurately detected.

  According to the present invention, it is possible to accurately calculate the surface shape data of the measurement region when the edge position of the target wafer is used as a reference.

It is a block diagram of the shape measuring apparatus by embodiment of this invention. It is a figure explaining the calibration process of the shape measuring apparatus by embodiment of this invention. It is explanatory drawing which shows the measurement principle of SH sensor. It is a figure for demonstrating the calculation formula of the wavefront angle in SH sensor. It is a figure for demonstrating the calculation method which calculates | requires the object wafer of a measuring object from the wavefront angle measured with SH sensor. It is a flowchart which shows operation | movement of the shape measuring apparatus by embodiment of this invention. It is a figure which shows the hardware constitutions of the shape measuring apparatus by embodiment of this invention. It is the figure which showed another example of the hardware constitutions of the shape measuring apparatus by embodiment of this invention. It is explanatory drawing of the method of detecting a reference position from the reference sample in which the mask area | region was formed. It is explanatory drawing of the method of detecting a reference position from the reference sample in which the waveform area | region was formed. It is a figure which shows an example of the hardware constitutions of the shape measuring apparatus at the time of employ | adopting a laser distance meter as an edge measurement part. It is a figure which shows an example of the hardware constitutions of the shape measuring apparatus at the time of employ | adopting a contact-type displacement meter as an edge measurement part. It is the figure which showed an example of the hardware constitutions of the shape measuring apparatus at the time of employ | adopting the aspect which detects a reference | standard position using the method of FIG. It is the figure which showed an example of the hardware a structure of the shape measuring apparatus at the time of employ | adopting the aspect which detects a reference position by finely moving a reference sample. It is the figure which showed an example of the structure of the light source in this Embodiment. It is the figure which showed the luminance distribution in the block on the imaging surface corresponding to a certain one microlens. It is the graph which showed the luminance distribution when the luminance distribution of FIG. 16 was cut by a cutting line. It is a figure explaining the problem at the time of measuring the roll-off part of a wafer using SH sensor which is a wavefront sensor of a Shack-Hartmann system.

  As shown in FIG. 18, the surface of the wafer is divided into a flat portion, a roll-off portion, and a chamfered portion from the center toward the edge. The flat portion is a flat region on the surface of the wafer. The roll-off portion is a region having a gentle slope toward the edge portion. The chamfered portion is a region having a larger slope than the roll-off portion toward the edge portion. When evaluating a wafer, not only the surface shape of the flat surface portion but also the surface shape of the roll-off portion is an effective evaluation index. Therefore, in this embodiment, a case where the surface shape of the roll-off part is measured will be described as an example. However, this is only an example, and a region other than the roll-off portion (for example, a flat portion) may be set as a measurement region, and the surface shape may be measured.

  FIG. 1 is a block diagram of a shape measuring apparatus according to an embodiment of the present invention. The shape measuring apparatus includes a light source 101, an SH (Shack-Hartmann) sensor 102, an edge measuring unit 103, and a control unit 110.

  The light source 101 irradiates the wafer with measurement light. The SH sensor 102 is a Shack-Hartmann wavefront sensor that receives the reflected light of the measurement light from the wafer and measures the surface shape of the wafer. The edge measuring unit 103 measures the edge of the wafer.

  FIG. 7 is a diagram illustrating an example of a hardware configuration of the shape measuring apparatus according to the embodiment of the present invention. In FIG. 7, a camera 103 a is employed as the edge measurement unit 103. The light source 101 is provided on the upper side of the wafer W and irradiates the measurement light 701 in a direction parallel to the surface of the wafer W. A half mirror 104 is disposed above the edge of the wafer W. The half mirror 104 reflects the measurement light 701 so that part of the measurement light 701 reaches the wafer W and the rest of the measurement light 701 passes outside the edge.

  The SH sensor 102 is disposed immediately above the outer peripheral region of the wafer W. The camera 103a is disposed directly below the SH sensor 102 with the wafer W interposed therebetween. Here, the outer peripheral region is a region including a roll-off portion, a chamfered portion, and an edge.

  The measurement light 701 emitted from the light source 101 is reflected by the half mirror 104, deflected by 90 degrees, and irradiates the outer peripheral region of the wafer W. A part of the measurement light 701 reflected by the half mirror 104 is reflected by the wafer W. The reflected light 702 passes through the half mirror 104 and enters the SH sensor 102. Thereby, the SH sensor 102 measures the surface shape of the roll-off part.

  On the other hand, a part of the measurement light 701 reflected by the half mirror 104 is not reflected by the wafer W, passes through the outside of the wafer W, and directly enters the camera 103a. Thereby, the camera 103a measures an edge.

  FIG. 8 is a diagram showing another example of the hardware configuration of the shape measuring apparatus according to the embodiment of the present invention. FIG. 8 shows a state in which the surface of the wafer W is viewed from directly above. In the example of FIG. 8, two light sources 101a and 101b are provided. The light source 101a (an example of a first light source) is a light source for the SH sensor 102, and the light source 101b (an example of a second light source) is a light source for the camera 103a.

  In the example of FIG. 8, the arrangement of the light source 101a, the half mirror 104, and the SH sensor 102 is the same as that in FIG. 7, but the arrangement of the light source 101b and the camera 103a is different from that in FIG. The light source 101b and the camera 103a are at the same height as the wafer W and are arranged in parallel with the surface of the wafer W. Further, the light source 101b and the camera 103a are arranged to face each other with the wafer W as the center.

  The measurement light 701a emitted from the light source 101a is reflected by the half mirror 104, deflected by 90 degrees, reflects the outer peripheral area of the wafer W, and enters the SH sensor 102 as reflected light. Thereby, the SH sensor 102 measures the surface shape of the roll-off part.

  On the other hand, a part of the measurement light 701b emitted from the light source 101b is shielded by the wafer W and enters the camera 103a. Thereby, the silhouette image of the edge enters the camera 103a, and the camera 103a measures the edge. In the following description, it is assumed that the camera 103a is employed as the edge measuring unit 103.

  Returning to FIG. 1, the control unit 110 includes, for example, a microcontroller, and includes a light source control unit 111, a reference position detection unit 112, an edge position detection unit 113, an edge distance calculation unit 114, a surface shape calculation unit 115, and coordinates. A conversion unit 116 is included.

  The light source control unit 111 supplies power to the light source 101 and turns on the light source 101. The reference position detection unit 112 causes the SH sensor 102 to measure the surface shape of the reference sample in which the reference pattern is formed at the reference position that is separated from the edge by the specified distance a, and detects the reference position. FIG. 2 is a diagram for explaining calibration processing of the shape measuring apparatus according to the embodiment of the present invention. In the reference sample RW, a reference position R0 is set at a position a predetermined distance a from the edge. Here, the specified distance a is known and has, for example, a distance along the radial direction from the edge to an assumed position in the roll-off portion.

  The reference sample RW is created in advance by simulating the target wafer TW separately from the target wafer TW to be measured. The reference sample RW may be created using the same material as the target wafer TW, or may be created using a material different from the target wafer TW. As another material, for example, a specular object is adopted. In the reference sample RW, a mask region 901 or a waveform region 1001 as shown in FIGS. 9 and 10 is formed as a reference pattern at the reference position R0. For example, it is assumed that the measurement image P301 illustrated in FIG. It is assumed that the measurement image P301 has a coordinate system in which the y axis is set in the vertical direction and the x axis is set in the horizontal direction. In the measurement image P301, for example, the center coordinate in the vertical direction is set to y = 0, and the radial direction of the reference sample RW is positioned on the straight line y = 0. In the reference sample RW, a reference position R0 is formed on a straight line passing through a certain position on the x-axis and parallel to the y-axis. Therefore, the reference position R0 is detected by the reference position detection unit 112 as data having a value of only the x coordinate. Then, the reference position detection unit 112 sets the coordinate system of the SH sensor 102 so that the coordinate of the reference position R0 is the origin of the x coordinate.

  Returning to FIG. 1, the edge position detection unit 113 causes the edge measurement unit 103 to measure the edge of the reference sample RW and detects the first edge position E <b> 1 that is the edge position of the reference sample RW, as shown in FIG. 2. Here, it is assumed that the edge measurement unit 103 captures a measurement image P201 as shown in FIG. In this case, the edge position detection unit 113 first detects the edge contour E1a. As a method for detecting the contour E1a, for example, a method may be employed in which a region having a luminance equal to or higher than a specified value is extracted from the measurement image P201 and an edge portion of the extracted region is detected as the contour E1a. Or you may employ | adopt the method of detecting the edge outline E1a by differentiating the measurement image P201. Then, the edge position detection unit 113 may identify the peak position of the contour E1a and detect the position as the first edge position P1 (e1x, e1y). Here, as shown in FIG. 2, it is assumed that the measurement image P201 has a y-axis in the vertical direction and an x-axis in the horizontal direction. In this case, for example, the edge position detection unit 113 may detect the position where the value of the x component is maximum in the contour E1a as the first edge position P1 (e1x, e1y).

  As shown in FIG. 2, the edge-to-edge distance calculation unit 114 causes the edge measurement unit 103 to measure the edge of the target wafer TW and detects the second edge position E2 that is the edge position of the target wafer TW. Here, assuming that the measurement image of the target wafer TW imaged by the edge measurement unit 103 is a measurement image P202, the edge-to-edge distance calculation unit 114 is similar to the edge position detection unit 113 in the same way as the second edge from the measurement image P202. The position P2 (e2x, e2y) may be detected.

  Then, the edge distance calculation unit 114 calculates the edge distance b between the first and second edge positions E1 and E2. Thereby, the deviation of the edge between the target wafer TW and the reference sample RW is known. Here, it is assumed that the reference sample RW and the target wafer TW are placed on the same sample support 1301 (see FIG. 13) and there is no deviation of the x axis. Therefore, the edge-to-edge distance calculation unit 114 calculates the difference between the x components of the first and second edge positions P1 and P2, converts this difference into a real space distance, and calculates the edge-to-edge distance b. Here, the edge distance calculation unit 114 may calculate the edge distance b using a predetermined coefficient that associates the coordinate system of the edge measurement unit 103 with the real space.

  The surface shape calculation unit 115 causes the SH sensor 102 to measure the surface shape of the roll-off part of the target wafer TW, and calculates the surface shape data of the roll-off part using the reference position R0 as reference coordinates. Here, the surface shape data has a data structure in which height data indicating the height of each position of the roll-off portion is arranged in a predetermined row × predetermined column. Specifically, as shown in FIG. 3A, since height data is obtained for each imaging point 42c from the measurement image P301, the surface shape data has a height corresponding to the imaging point 42c. It has a data structure in which data is arranged.

  As shown in FIG. 2, the coordinate conversion unit 116 obtains a distance c (= a + b) obtained by adding the inter-edge distance b to the specified distance a from the reference position R0. Here, since the specified distance a is known, the distance c (= a + b) represents the distance from the reference position R0 to the second edge position E2. Then, the coordinate conversion unit 116 converts the distance c (= a + b) into a distance in the coordinate system of the SH sensor 102 to obtain a coordinate offset.

  Then, the coordinate conversion unit 116 adds a coordinate offset to the coordinates of the surface shape data calculated by the surface shape calculation unit 115, and performs coordinate conversion of the surface shape data into a coordinate system based on the second edge position E2. For example, in FIG. 5, it is assumed that xA is set as a reference coordinate corresponding to the reference position R0. In this case, the coordinate conversion unit 116 sets a coordinate system having a coordinate obtained by adding a coordinate offset xC to xA as a reference coordinate, and performs coordinate conversion of the surface shape data into this coordinate system. Here, it is assumed that the coordinate offset xC has a value only for the x component and no value for the y component. Thereby, the coordinates of each height data constituting the surface shape data of the roll-off portion of the target wafer TW accurately represent the distance from the second edge position E2. As a result, it is possible to accurately obtain the surface shape data of the roll-off portion when the second edge position E2 is used as a reference.

  FIG. 3 is an explanatory diagram showing the measurement principle of the SH sensor 102. The SH sensor 102 includes a microlens array 41 and an image sensor 42 provided behind the microlens array 41. The microlens array 41 is a lens in which a plurality of microlenses 411 are arranged in an array of predetermined rows × predetermined columns. The image pickup element 42 is an image pickup element in which a plurality of pixels are arranged in an array of predetermined rows × predetermined columns. For example, a CCD image sensor or a CMOS image sensor is employed.

  The SH sensor 102 forms an image of incident light with the microlens 411 and causes the image sensor 42 to pick up an image of the obtained incident light. The imaging surface 42 a of the imaging element 42 is divided into a plurality of blocks 42 b corresponding to the microlenses 411. When the incident light is a plane wave, the incident light is imaged at the center of each block 42b as shown in FIG.

  On the other hand, when the incident light is distorted, as shown in FIG. 3 (B), the incident light is imaged with a deviation from the center in each block 42b. Therefore, in each block 42b, the deviation from the center of the imaging point 42c is obtained, whereby the angle of the wavefront (wavefront angle) with respect to the plane wave is known, and the inclination at each position of the wafer is known from the wavefront angle, and the inclination is integrated. Thus, height data at each position of the wafer is obtained.

  FIG. 4 is a diagram for explaining a calculation formula of the wavefront angle in the SH sensor 102. The microlens 411 has a dome-shaped cross section that is convex toward the image sensor 42 side. A reference image formation point P401 indicates an image formation point on the image pickup device 42 of the plane wave WV1 whose wavefront is parallel to the image pickup surface of the image pickup device 42. The reference imaging point P401 is an intersection between the optical axis 411a of the microlens 411 and the image sensor 42. An observation image formation point P402 indicates an image formation point on the image pickup device 42 of the observation wave WV2 whose wavefront is inclined from the image pickup surface of the image pickup device 42. In FIG. 4, the distance between the microlens array 41 and the image sensor 42 is F. Also, let L be the distance between the reference imaging point P401 and the observation imaging point P402.

From the geometrical relationship, the wavefront angle α is represented by tan α = L / F. Therefore, the surface shape calculation unit 115 obtains a shift amount (distance L) from the reference image formation point P401 of the observation image formation point P402 in each block 42b from the measurement image imaged by the image sensor 42 of the SH sensor 102. The wavefront angle α in each block 42b is calculated from α = tan ⁻¹ (L / F).

  The SH sensor 102 captures two-dimensional image data with the x axis in the radial direction of the target wafer TW and the y axis as the direction orthogonal to the x axis. For this reason, the amount of deviation from the observation image formation point P402 to the reference image formation point P401 includes an amount of deviation between the x component and the y component, but in this embodiment, the height data of the target wafer TW along the radial direction. The focus is on measuring change. Therefore, the surface shape calculation unit 115 obtains the deviation amount of the x component as the distance L shown in FIG. 4 and obtains the wavefront angle α. Note that the surface shape calculation unit 115 may obtain the position of the center of gravity of the luminance distribution in the block 42b and obtain the position as the observation imaging point P402. In this case, the surface shape calculation unit 115 may calculate the position of the center of gravity of the luminance at a subpixel level smaller than the resolution of the image sensor 42. Thereby, a wavefront angle can be calculated | required more correctly.

  FIG. 5 is a diagram for explaining a calculation method for obtaining the surface shape of the target wafer from the wavefront angle measured by the SH sensor 102. The upper curve shows the wavefront 501 measured by the SH sensor 102, and the lower curve shows the surface shape of the target wafer TW. In FIG. 5, the wavefront 501 and the surface shape of the target wafer TW in one line parallel to the x-axis are shown.

  Since the surface shape of the target wafer TW changes in nanometer order and the distance between the SH sensor 102 and the target wafer TW is about 100 mm, the distance D between the SH sensor 102 and the measurement target can be regarded as constant.

  The x coordinate of a certain position P501 on the wavefront 501 is assumed to be x0. The wavefront angle α at the position P501 represents information on the height of the intersection (position P502) between the normal to the tangent to the position P501 and the target wafer TW. In this case, the x coordinate of the position P502 is expressed by x = x0 + D · tan α. Further, the inclination φ (x) at the position P502 of the target wafer TW is represented by φ (x) = (1/2) · α.

  Therefore, the surface shape calculation unit 115 obtains φ (x) = (1/2) · α and x = x0 + D · tan α as the inclination of the position P502. Then, the surface shape calculation unit 115 integrates the inclination φ (x) with respect to x over the entire roll-off part in one line parallel to the x axis, and obtains height data Φ (x) (= ∫φ (x ) Dx). The surface shape calculation unit 115 performs this integration for all the lines parallel to the x-axis, and obtains surface shape data in the entire roll-off portion of the target wafer TW.

  FIG. 9 is an explanatory diagram of a method for detecting the reference position R0 from the reference sample RW in which the mask region 901 is formed. In the reference sample RW, a rectangular mask region 901 is formed at a position a predetermined distance a from the edge. Here, the mask region 901 is a region that has a significantly different reflectance from other regions and does not reflect light. Examples of the method for creating the mask region 901 include a method of forming a non-reflective surface by photolithography, and a method of printing a non-reflective object such as a non-reflective coating or a thin film sheet.

  When the measurement region D93 is set so as to partially include the mask region 901, the measurement region D93 is irradiated with the measurement light, and imaged by the SH sensor 102, the measurement image P9 in which the imaging points D9 are arranged in an array is obtained. can get. The imaging point D9 is divided into a bright area D91 and a dark area D92. The boundary between the bright region D91 and the dark region D92 indicates the boundary of the mask region 901. However, the SH sensor 102 cannot obtain a resolution less than the pitch of the microlens array 41.

  Therefore, the reference position detection unit 112 moves the reference sample RW stepwise in the radial direction (direction from the center toward the edge) with a resolution (for example, every 10 μm) equal to or less than the pitch of the microlens array 41. Then, each time the reference sample RW moves, the reference position detection unit 112 causes the SH sensor 102 to image the measurement region D93 and monitors a change in luminance at a certain imaging point D94.

  Then, a graph G901 is obtained. A graph G901 is a graph showing a change in luminance at the imaging point D94. The horizontal axis indicates the moving distance of the reference sample RW, and the vertical axis indicates the luminance. When the movement distance is 0, the imaging point D94 is not affected by the mask region 901 and has high luminance. Thereafter, the high luminance is maintained until the imaging point D94 is affected by the mask region 901. Then, when the moving distance becomes about 50 μm, the influence of the mask region 901 starts to appear at the imaging point 94, and thereafter the luminance gradually decreases as the reference sample RW moves. When the moving distance is about 125 μm, the image point D94 is completely affected by the mask region 901, and the luminance becomes zero.

  As described above, when the reference sample RW is finely moved to monitor the luminance at a certain imaging point D94, a graph in which the luminance changes stepwise is obtained as shown in the graph G901.

  Next, the reference position detection unit 112 differentiates the graph G901. Then, an upward convex bell-shaped graph G902 is obtained. Next, the reference position detection unit 112 detects the center of gravity with respect to the movement distance from the graph G902 as the reference position R0, and sets the coordinates of the reference position R0 as the reference coordinates of the SH sensor 102. As a result, the reference coordinates can be detected with an accuracy equal to or lower than the resolution of the SH sensor 102.

  FIG. 14 is a diagram illustrating an example of a hardware configuration of the shape measuring apparatus in a case where a mode in which the reference position R0 is detected by finely moving the reference sample RW is employed.

  The sample support 1301 is provided with a measurement object 1304 such as a target wafer TW or a reference sample RW. The sample support base 1301 is provided with a plurality of pins 1302. The pins 1302 include a support pin that supports the measurement object 1304, a positioning pin for positioning the measurement object 1304, and the like.

  The replacement unit 1303 is a device for replacing the measurement object 1304 currently installed on the sample support 1301 with another measurement object 1304. Here, as the replacement unit 1303, a robot mechanism that automatically replaces the measurement object 1304 may be employed, or a manual mechanism that manually replaces the measurement object 1304 may be employed.

  A moving unit 1403 is provided below the sample support 1301. A moving unit 1403, a stage 1401, and a fine movement mechanism 1402 are included. The stage 1401 is provided below the sample support and is moved in the direction of the arrow by the fine movement mechanism 1402. Thereby, the reference sample RW is finely moved stepwise. The fine movement mechanism 1402 is configured by a piezo element or a device that finely moves the stage 1401 with a mechanical type or a manual screw type.

  The fine movement mechanism 1402 can move the stage 1401 with a target resolution (10 μm) for detecting the reference position R0, and can also move the stage 1401 with a resolution exceeding the lens pitch (100 μm). Can be moved.

  FIG. 10 is an explanatory diagram of a method for detecting the reference position R0 from the reference sample RW in which the waveform region 1001 is formed. The waveform area 1001 is a rectangular area formed at a position a predetermined distance a from the edge in the reference sample RW. In the waveform area 1001, a waveform pattern WP having a periodically changing cross-sectional height is formed. In the example of FIG. 10, a sine curve is adopted as the waveform pattern WP. The waveform pattern WP is not limited to a sine curve, and for example, a triangular wave may be adopted. In the present embodiment, for example, a waveform pattern having a period of 0.5 mm and an amplitude of 0.05 μm is employed as the waveform pattern. As the period, a value larger than at least the lens pitch of the SH sensor 102 may be adopted.

  The width D11a of the measurement region D11 of the SH sensor 102 is set to a size larger than the width H1 of the waveform region 1001. The waveform pattern WP can be formed using a technique such as photolithography, for example.

  When the waveform region 1001 is irradiated with measurement light and imaged by the SH sensor 102, a measurement image P10 shown in FIG. 10 is obtained. In the measurement image P10, a plurality of imaging points D10 are arranged in a periodic light and dark pattern corresponding to the waveform pattern WP.

  A graph G1001 is a graph in which the luminance of the imaging point in the monitor region D12 in the measurement image P10 is plotted, the vertical axis indicates the luminance, and the horizontal axis indicates the x-axis coordinate of the monitor region D12. Since the imaging point located on the leftmost side in the monitor region D12 is located at 500 μm from the left end of the measurement image P10, in the graph G1001, 500 μm is plotted at the start position on the horizontal axis.

  Here, a certain position of the waveform pattern WP formed in the waveform area 1001 (here, the position of the first peak from the left end in the monitor area D12) is a position away from the edge of the reference sample RW by a specified distance a. The reference position detecting unit 112 has the information in advance. Therefore, the reference position detection unit 112 fits the known waveform pattern WP formed in the waveform region 1001 in the graph G1001, detects the position of the first peak from the left end as the reference position R0, and the reference position R0 The coordinates are set as reference coordinates for the SH sensor 102.

  As described above, according to the method of FIG. 10, the reference position R0 can be detected with a resolution equal to or less than the lens pitch by one imaging of the SH sensor 102.

  FIG. 13 is a diagram illustrating an example of a hardware configuration of the shape measuring apparatus in a case where an aspect in which the reference position R0 is detected using the method of FIG. 10 is employed. When the method of FIG. 10 is used, it is not necessary to finely move the reference sample RW. Therefore, in FIG. 13, the moving unit 1403 is omitted. The other configuration is the same as that in FIG. 14 and will not be described.

  FIG. 6 is a flowchart showing the operation of the shape measuring apparatus according to the embodiment of the present invention. First, the reference sample RW is placed on the sample support 1301, and the reference pattern is measured by the SH sensor 102 (S601). Next, the reference position detection unit 112 detects the reference position R0 from the measurement result of the SH sensor 102 (S602). In this case, the reference position detection unit 112 may detect the reference position R0 using the method of FIG. 9 or FIG.

  Next, the edge position detection unit 113 causes the edge measurement unit 103 to measure the edge of the reference sample RW (S603). Here, the edge measurement unit 103 captures an edge of the reference sample RW, thereby acquiring a measurement image in which the silhouette of the edge of the reference sample RW as shown in FIG. 2 appears.

  Next, the edge position detection unit 113 detects the first edge position E1 that is the edge position of the reference sample RW (S604). Note that the edge measurement unit 103 is configured by a camera having a resolution of about 5 μm, for example, and therefore can detect the first edge position E1 with a resolution equal to or less than the lens pitch of the microlens array 41.

  Thus, the calibration process using the reference sample RW ends.

  Next, the target wafer TW is placed on the sample support 1301, and the edge-to-edge distance calculation unit 114 causes the edge measurement unit 103 to measure the edge of the target wafer TW (S605).

  Next, the inter-edge distance calculation unit 114 detects the second edge position E2 from the measurement image captured by the edge measurement unit 103 (S606). Similarly to the first edge position E1, the second edge position E2 is detected with a resolution equal to or less than the lens pitch of the microlens array.

  Next, the edge distance calculation unit 114 obtains a difference between the first edge position E1 and the second edge position E2 and calculates the edge distance b (S607).

  Next, the surface shape calculation unit 115 causes the SH sensor 102 to measure the surface shape of the roll-off part (S608). Next, the surface shape calculation unit 115 calculates the surface shape data of the target wafer TW from the measurement image captured by the SH sensor 102 using the reference position R0 as the reference coordinates (S609).

  Next, the coordinate conversion unit 116 adds the specified distance a and the edge distance b, calculates a distance c (= a + b), and converts the distance c (= a + b) into a distance in the coordinate system of the SH sensor 102. Then, the coordinate offset xC is obtained (S610).

  Next, the coordinate conversion unit 116 adds the coordinate offset xC to the coordinates of each height data constituting the surface shape data of the target wafer TW, and converts the surface shape data into a coordinate system based on the second edge position E2. (S611).

  As described above, the surface shape data of the roll-off portion when the second edge position E2 is used as a reference is obtained.

  In the above description, the camera 103a is employed as the edge measuring unit 103, but the present invention is not limited to this. For example, a laser distance meter may be employed as the edge measurement unit 103. FIG. 11 is a diagram illustrating an example of a hardware configuration of the shape measuring apparatus when the laser distance meter 103 b is employed as the edge measuring unit 103. The laser distance meter 103 b is disposed at a position facing the edge of the measurement object 1304. The laser distance meter 103b measures the time from irradiating the laser beam toward the edge to receiving the reflected light from each position of the edge, so that the distance from its own position to each position of the edge is measured. Measure distance.

  The edge position detection unit 113 causes the laser distance meter 103b to measure the edge of the reference sample RW, and detects the position of the reference sample RW closest to the laser distance meter 103b as the first edge position E1. Further, the edge-to-edge distance calculation unit 114 causes the laser distance meter 103b to measure the edge of the target sample TW, and detects the position of the reference sample RW closest to the laser distance meter 103b as the second edge position E2. Then, the edge distance calculation unit 114 obtains the edge distance by taking the difference between the first edge position E1 and the second edge position E2.

  Here, the laser distance meter 103b can determine the distance between each position of the edge with a resolution of the order of 10 μm. Therefore, even when the laser distance meter 103b is used, the same effect as that obtained when the camera 103a is used can be obtained.

  Further, as the edge measuring unit 103, a contact displacement meter may be employed. FIG. 12 is a diagram illustrating an example of a hardware configuration of a shape measuring apparatus when a contact displacement meter 103 c is employed as the edge measuring unit 103. The contact-type displacement meter 103c abuts the probe 1201 at each position of the edge and detects the change of the probe 1201, thereby measuring the shape of the edge.

  The edge position detection unit 113 causes the contact displacement meter 103c to measure the edge of the reference sample RW, and detects the position of the reference sample RW closest to the contact displacement meter 103c as the first edge position E1. Further, the edge distance calculation unit 114 causes the contact displacement meter 103c to measure the edge of the target sample TW, and detects the position of the reference sample RW closest to the contact displacement meter 103c as the second edge position E2. Then, the edge distance calculation unit 114 obtains the edge distance by taking the difference between the first edge position E1 and the second edge position E2.

  Here, the contact-type displacement meter 103c can also determine the distance between each position of the edge with a resolution of the order of 10 μm. Therefore, even when the contact displacement meter 103c is used, the same effect as that obtained when the camera 103a is used can be obtained.

  Next, the light source 101 used in the shape measuring apparatus of the present embodiment will be specifically described. The SH sensor 102 measures the angle of the wavefront of parallel light. For this reason, as the light source 101 of the SH sensor, it is preferable to employ the light source 101 that outputs light having a uniform wavefront. Therefore, in general, a laser light source that outputs monochromatic light having a uniform wavefront is employed as the light source of the SH sensor 102.

  However, in this shape measuring apparatus, it is sufficient if the traveling direction of the reflected parallel light reflected on the surface of the wafer is known. Therefore, there is no necessity to use a laser light source as the light source of the SH sensor, and any light source that can output parallel light can be used. Any light source may be employed.

  In the SH sensor 102 shown in FIG. 3A, the light that has passed through the microlens array 41 is mainly condensed by the microlens 411 to become a point image (image formation point 42c), but some of the light is microscopic. The light is diffracted by the edge portion of the lens 411 and the like, and enters the adjacent microlens 411 to become unnecessary diffracted light. When the diffracted lights or the diffracted lights and the imaging point 42c overlap, interference fringes 1602 (see FIG. 16) are generated on the point image and around the point image.

  FIG. 16 is a diagram showing a luminance distribution in the block 42b on the imaging surface corresponding to one microlens 411. As shown in FIG. FIG. 17 is a graph showing the luminance distribution when the luminance distribution of FIG. 16 is cut along the cutting line 1601. The vertical axis indicates the luminance, and the horizontal axis indicates the position on the cutting line 1601. Note that FIG. 16 shows that the luminance is higher as the density is lower (closer to white), and the luminance is lower as the density is higher (closer to black).

  In the block 42b, the luminance distribution on the cutting line 1601 when the cutting line 1601 is set so as to pass through the imaging point 42c is geometrically optically symmetrical about the imaging point 42c such as a sinc function. It is thought that it has a simple shape. However, when the coherent light is irradiated, the above-described interference fringes 1602 are generated, and the luminance distribution on the cutting line 1601 does not have a symmetrical shape about the imaging point 42c. That is, the luminance distribution in the block 42b deviates from the original luminance distribution predicted geometrically.

  Therefore, if the position of the center of gravity of the luminance distribution is obtained in the block 42b and the position is adopted as the image formation point, there arises a problem that the image formation point is deviated from the original image formation point.

  In addition, when coherent light is used, if the incident angle of the light beam incident on the SH sensor changes minutely, the way in which the diffracted light overlaps changes according to the change, and the shading pattern of the interference fringes also changes. The position of the calculated center of gravity deviates from the original existence point due to slight fluctuations.

  Therefore, in this embodiment, an incoherent light source that outputs incoherent light is used as the light source. FIG. 15 is a diagram illustrating an example of the configuration of the light source 101 in the present embodiment.

  The light source 101 includes a light source element 1501, a condenser lens 1502, a pinhole 1503, a convex lens 1504, and a housing 1505. The light source element 1501 is configured by a light source element that outputs incoherent light, such as a light emitting diode, and is disposed in the housing 1505 so that the light emitting surface is located on the opening side of the housing 1505.

  The condensing lens 1502 is provided so that the main surface is perpendicular to the optical axis L15 at a position between the light source element 1501 and the pinhole 1503. The pinhole 1503 is configured by a through hole formed in a partition plate 1503a provided in the housing 1505 so as to be orthogonal to the optical axis L15. The partition plate 1503a is provided in the housing 1505 so that the pinhole 1503 is positioned on the optical axis L15.

  The convex lens 1504 is provided on the opening side of the housing 1505 with respect to the pinhole 1503 with the main surface orthogonal to the optical axis L15. The convex lens 1504 converts light that has passed through the pinhole 1503 into parallel light. The light converted into parallel light by the convex lens 1504 is output from the opening of the housing 1505 and irradiates the measurement object.

  Here, a pinhole 1503 is provided in front of the convex lens 1504. As a result, light as if it is output from a point light source is incident on the convex lens 1504, so that the convex lens 1504 can more accurately convert the light output from the light source element 1501 into parallel light. A condensing lens 1502 is provided between the light source element 1501 and the pinhole 1503. Accordingly, it is possible to increase a ratio of light passing through the pinhole 1503 out of light output from the light source element 1501. As a result, the conversion efficiency of the light output from the light source element 1501 into parallel light can be increased.

  Thus, since the light source 101 shown in FIG. 15 is a coherent light source, generation of an interference fringe pattern due to diffracted light generated at the edge portion of the microlens 411 can be prevented as in the case where a laser light source is employed. As a result, the position of the center of gravity of the luminance distribution in the block 42b is suppressed from deviating from the original image formation point, and the original image formation point position can be obtained accurately.

  Further, since the light source 101 shown in FIG. 15 is composed of relatively inexpensive materials such as a light source element 1501 composed of an LED or a light bulb, a condenser lens 1502, a pinhole 1503, and a convex lens 1504, a laser light source is adopted. Cost can be reduced compared to the case. Furthermore, since the light source 101 shown in FIG. 15 is an incoherent light source, the light source 101 can be made more compact than when a laser light source is employed. Moreover, when LED is employ | adopted as an incoherent light source, an incoherent light source has a long lifetime compared with the laser light source whose lifetime is about 5000 to 10,000 hours. As a result, the number of replacements of the light source element 1501 is reduced, and maintenance can be improved.

  Note that light output from the light source element 1501 preferably has a wavelength with small absorption with respect to the measurement target. Since silicon, sapphire, and the like are assumed as measurement objects in this embodiment, light having a wavelength band with less absorption with respect to silicon, sapphire, and the like is employed as light output from the light source element 1501. Is preferred. In the present embodiment, since the light output from the light source 101 passes through optical elements such as a half mirror (beam splitter) and a lens, light in a wavelength band that is less absorbed by these optical elements is employed. Is preferred. Visible light is an example of light that is less absorbed by silicon, sapphire, a half mirror, a lens, and the like, and the light source element 1501 preferably outputs visible light.

  As the light source element 1501, a light emitting diode is employed, but other than this, a general light source such as a light bulb may be employed. Here, the light bulb is obtained by enclosing a filament or a discharge element in a glass bulb shell. However, in consideration of directivity, luminance, lifetime, cost, versatility, and the like, it is preferable to employ a light emitting diode as the light source element 1501 as compared with a light bulb.

  In the light source 101 shown in FIG. 15, the condenser lens 1502 may be omitted. Further, as shown in FIG. 7, the light source 101 shown in FIG. 15 may be employed as the light source 101 that is shared by the SH sensor 102 and the camera 103a. Further, as shown in FIG. 8, when the light source 101a of the SH sensor 102 and the light source 101b of the camera 103a are individually provided, the light source 101 shown in FIG. 15 may be applied to the light source 101a. Further, as shown in FIGS. 11 and 12, when the edge measuring unit 103 is configured by a laser distance meter 103b or a contact displacement meter 103c, the light source 101 shown in FIG. 15 is applied to the light source 101 of the SH sensor 102. Just do it.

101, 101a, 101b Light source 102 SH sensor 103 Edge measurement unit 103a Camera 103b Laser distance meter 103c Contact displacement meter 104 Half mirror 110 Control unit 111 Light source control unit 112 Reference position detection unit 113 Edge position detection unit 114 Edge distance calculation unit DESCRIPTION OF SYMBOLS 115 Surface shape calculation part 116 Coordinate conversion part 901 Mask area | region 1001 Waveform area | region 1403 Moving part a Specified distance b Edge distance WP Waveform pattern

Claims (13)

A shape measuring device for measuring the surface shape of a predetermined measurement region on the surface of a wafer,
A light source that emits measurement light;
A Shack-Hartmann-type SH sensor that receives reflected light from the wafer of the measurement light and measures the surface shape of the wafer;
An edge measuring unit for measuring the edge of the wafer;
A reference position detector for detecting the reference position by causing the SH sensor to measure the surface shape of a reference sample in which a reference pattern is formed at a reference position separated from the edge by a specified distance;
An edge position detection unit that causes the edge measurement unit to measure an edge of the reference sample and detects a first edge position that is an edge position of the reference sample;
The edge measurement unit measures the edge of the target wafer that is the measurement target, detects the second edge position that is the edge position of the target wafer, and calculates the inter-edge distance between the first and second edge positions A distance calculator;
A surface shape calculation unit that causes the SH sensor to measure the surface shape of the measurement region of the target wafer, and calculates surface shape data of the measurement region using the reference position as a reference coordinate;
A shape measuring device comprising: a coordinate conversion unit for converting the surface shape data into a coordinate system having a reference coordinate at a position separated from the reference position by a distance obtained by adding the distance between edges to the specified distance. The edge measuring unit is a camera installed on the opposite side of the wafer with respect to the light source,
A part of the light passes through the external edge of the wafer to reflect the measurement light so guided to the camera, and, a half mirror rather guiding the SH sensor passes through the reflected light from the wafer The shape measuring device according to claim 1 further provided.   The shape measuring apparatus according to claim 1, wherein the light source outputs incoherent light. The light source is
A light source element that outputs the incoherent light;
A pinhole provided on the output side of the light source element;
The shape measuring apparatus according to claim 3, further comprising: a lens that converts light that has passed through the pinhole into parallel light.   The shape measuring apparatus according to claim 4, further comprising a condensing lens provided between the light source element and the pinhole. The light source includes a first light source for an SH sensor and a second light source for an edge measuring unit installed at a position different from the first light source,
The shape measuring apparatus according to claim 1, wherein the edge measuring unit is a camera installed on the opposite side of the wafer with respect to the second light source.   The shape measuring apparatus according to claim 1, wherein the edge measuring unit is a laser displacement meter.   The shape measuring apparatus according to claim 1, wherein the edge measuring unit is a contact displacement meter.   The shape measurement apparatus according to claim 1, wherein the reference position detection unit detects the reference position with a resolution smaller than a resolution of the SH sensor. The reference pattern is a mask pattern;
A moving unit for moving the reference sample in a direction parallel to the surface;
The reference position detector causes the SH sensor to measure the surface shape of the reference sample each time the reference sample moves a distance shorter than the resolution of the SH sensor, and detects the reference position based on the measurement result. Item 10. The shape measuring apparatus according to Item 9. The reference pattern is a waveform pattern that changes periodically;
The shape measurement apparatus according to claim 9, wherein the reference position detection unit causes the SH sensor to measure the waveform pattern once and detects the reference position based on a measurement result.   The shape measuring apparatus according to claim 1, wherein the reference sample is configured by a plate-like mirror object that simulates the shape of the wafer.   The shape measurement apparatus according to claim 1, wherein the measurement region is a roll-off unit.