Measurement methods and measuring devices

Abstract

[Problem] To measure the thickness of each wafer while minimizing the influence of the thickness of the outer periphery of each wafer. [Solution] A measurement method is provided, comprising: a first memory step, wherein the thickness of each of a plurality of wafers held by a holding surface of a holding stage through a protective member is measured at multiple measurement points, and the thickness data of each wafer is stored, wherein the thickness data is obtained by establishing a correlation between the position of each of a plurality of measurement points in a predetermined plane corresponding to the back side of the wafer and the thickness at the plurality of measurement points; a second memory step, wherein the back side of the plurality of wafers is photographed, and the position of the center line of each slit groove in the predetermined plane is stored; a setting step, wherein a non-reference thickness data region is set around each wafer, wherein the non-reference thickness data region includes measurement points where the thickness data is not referenced when calculating the thickness of each wafer; and a calculation step, wherein the average thickness of the remaining measurement points in each of the plurality of wafers is calculated among the measurement points remaining after removing the measurement points contained in the non-reference thickness data region from the plurality of measurement points.

Description

Technical Field

[0001] This invention relates to a method for measuring the thickness of multiple wafers and a measuring apparatus for measuring the thickness of multiple wafers. Prior Technology

[0002] When manufacturing component wafers (hereinafter referred to as wafers) from wafers, for example, wafers are divided into wafers in component units. The wafers have multiple predetermined dividing lines (cutting tracks) set in a grid pattern on the front side, and components such as ICs (Integrated Circuits) are formed in each rectangular area divided by the multiple cutting tracks.

[0003] After dicing, the thickness of the wafer is measured to check whether the thickness of the wafer falls within the allowable range. For example, the thickness of each wafer after dicing is evaluated by measuring the thickness of each rectangular region of the wafer before dicing it into multiple wafers (see Patent Document 1).

[0004] Thickness measurement is performed on the wafer in a non-contact manner using an optical thickness gauge. For example, while the wafer is held by a chuck and a measurement light is shone onto the wafer from the thickness gauge, the chuck is rotated around a predetermined axis, and the spot of the measurement light is moved linearly from the outside to the inside along the radial direction of the wafer. In this way, the spot (i.e., the measurement point for thickness measurement) moves in a vortex pattern.

[0005] Since the positional information of each measurement point on the wafer can be obtained based on the rotation angle information of the chuck stage and the positional information of the thickness measuring device, thickness data is obtained by establishing a correlation between the positional information and the thickness information at each measurement point. Then, the average thickness obtained by measuring the thickness of a rectangular area at multiple measurement points is regarded as the thickness of the wafer.

[0006] However, the thickness of the dicing line in a wafer may differ from the thickness of the central portion of the rectangular region surrounded by the dicing line. For example, to improve the processing quality when dicing the dicing line, sometimes an insulating film that might degrade the processing quality is not formed at the dicing line.

[0007] Therefore, the thickness at the predetermined dividing line may be thinner than the thickness at the center of the rectangular region. Consequently, if the thickness of the predetermined dividing line is included in the thickness data when calculating the wafer thickness, there is a problem that the calculated thickness may differ from the actual wafer thickness that should be measured. [Known Technical Documents] [Patent Literature]

[0008] [Patent Document 1] Japanese Patent Application Publication No. 2017-199845 Summary of the Invention

[0009] [The problem that the invention aims to solve] The present invention was made in view of this problem and aims to measure the thickness of each wafer by reducing the influence of the thickness of the outer periphery of each wafer.

[0010] [Technical means to solve the problem] According to one aspect of the present invention, a measurement method is provided for measuring the thickness of a plurality of wafers by dividing the wafer into the plurality of wafers by forming dicing grooves within predetermined dicing lines on the wafer. The wafer has elements formed in each of a grid-like region divided by a plurality of predetermined dicing lines set on the front side and each having a predetermined width. The measurement method includes: a holding step in which the front side of the plurality of wafers is held by a holding surface of a holding stage and by a protective member; and a first memory step in which the thickness of each of the plurality of wafers held by the holding surface is measured at a plurality of measurement points, and the thickness data of each wafer is memoryed, wherein the thickness data is the result of the plurality of measurements taken at a predetermined plane corresponding to the back side of the wafer. The relationship between the position of each point and the thickness in the plurality of measurement points is established; the second memory step involves photographing the back side of the plurality of wafers held by the holding surface and memorizing the position of the center line of each slit groove in the predetermined plane; the setting step involves setting a non-reference thickness data region around each wafer based on the thickness data of each wafer, the non-reference thickness data region being located below a predetermined distance from the position of the center line of each slit groove and including measurement points where the thickness data is not referenced when calculating the thickness of each wafer; and the calculation step involves calculating the average thickness of the measurement points remaining in each of the plurality of wafers after removing the measurement points contained in the non-reference thickness data region from the plurality of measurement points.

[0011] According to another aspect of the present invention, a measuring apparatus is provided for measuring the thickness of a plurality of wafers by dividing the wafer into the plurality of wafers by forming dicing grooves in each predetermined dicing line on the wafer, wherein the wafer is formed with elements in each of a grid-like region divided by a plurality of predetermined dicing lines set on the front side and each having a predetermined width, the measuring apparatus comprising: a holding stage having a holding surface capable of holding the front side of the plurality of wafers through a protective member; a thickness measuring unit measuring the thickness of each of the plurality of wafers held by the holding surface at a plurality of measuring points; an imaging unit capturing images of the back side of the plurality of wafers held by the holding surface; a moving unit moving the thickness measuring unit and the imaging unit relative to the holding stage; and a control unit controlling at least the holding stage, the thickness measuring unit, and the imaging unit. The image unit and the moving unit, and the control unit having: a memory unit that stores: thickness data of each wafer, which is formed by establishing a correlation between the respective positions of a plurality of measurement points in a predetermined plane corresponding to the back side of the wafer and the thickness in the plurality of measurement points; and the position of a center line, which is the center line of each slit groove captured by the image unit and located in the predetermined plane; a setting unit that sets a non-reference thickness data region around each wafer based on the thickness data, the non-reference thickness data region being located at a predetermined distance from the position of the center line of the slit groove and including measurement points where the thickness data is not referenced when calculating the thickness of each wafer; and a calculation unit that calculates the average thickness of the measurement points obtained by removing the measurement points included in the non-reference thickness data region from the plurality of measurement points in each of the plurality of wafers.

[0012] [Invention Benefits] In the measurement method for one state and the measurement apparatus for another state of the present invention, a non-reference thickness data region is set around the wafer. The non-reference thickness data region is located at a predetermined distance below the position of the center line of the sizing groove and includes measurement points where the thickness data is not referenced when calculating the thickness of each wafer.

[0013] Furthermore, since the average thickness of the wafer is calculated from the remaining thicknesses among the measurement points after removing the measurement points in the area not referenced for thickness data, the thickness of the wafer, which is not included near the predetermined dividing line, can be calculated. This reduces the influence of the thickness of the outer periphery of each wafer, allowing for more accurate measurement of the thickness of each wafer. Simple Explanation of the Diagram

[0014] Figure 1 is a schematic three-dimensional view of a wafer. Figure 2(A) is a perspective view of the wafer unit, and Figure 2(B) is a cross-sectional view of the wafer unit. Figure 3 is a three-dimensional view of the measuring device. Figure 4 shows the flowchart of the measurement method. Figure 5 is a diagram illustrating the holding steps. Figure 6 is a schematic diagram of multiple measurement points that become the objects of memory in the first memory step. Figure 7 is a schematic diagram of the center lines of the multiple slots that become memory objects in the second memory step. Figure 8 is a schematic diagram showing the center lines of multiple measurement points and multiple slits in a wafer. Figure 9 is a chart showing the non-reference thickness data area set by the setting steps. Figure 10 shows the non-referenced thickness data area in the setting step and the multiple measurement points that become the calculation objects in the calculation step. Implementation

[0015] Referring to the accompanying drawings, one embodiment of the present invention will be described. First, the wafer 11 before it is divided into multiple wafers 21 (see FIG2(A)) will be described. FIG1 is a schematic perspective view of the wafer 11.

[0016] Wafer 11 is in the shape of a circular plate and has a single-crystal substrate formed of semiconductor materials such as silicon. However, wafer 11 is not limited to a single-crystal substrate of silicon, and may also have a single-crystal substrate formed of compound semiconductors such as silicon carbide (SiC) and gallium nitride (GaN).

[0017] The wafer 11 has a circular front side 11a and a circular back side 11b. A notch (cut) 11c indicating the crystal orientation of the wafer 11 is formed on a portion of the outer periphery. Alternatively, an orientation plane (cut) may be formed instead of the notch 11c.

[0018] On the front side 11a, multiple pre-defined dividing lines 13, each with a predetermined width 13a, are arranged in a grid pattern. For example, on a 300mm (12-inch) diameter wafer 11, multiple pre-defined dividing lines 13, each with a width 13a of 0.1mm, are arranged in a grid pattern. In this case, there are 968 rectangular areas of 8mm square on the front side 11a.

[0019] Furthermore, for example, a lattice of multiple predetermined dividing lines 13, each with a width of 0.07 mm, is provided on a 300 mm (12-inch) diameter wafer 11. In this case, there are 58,788 rectangular areas of 1 mm square on the front side 11a.

[0020] Each of the rectangular regions divided by the predetermined dividing line 13 has an IC, an LED (Light Emitting Diode) or other components 15 formed thereon. However, there are no particular restrictions on the size of the wafer 11, the type, number, shape, structure, size, or arrangement of the components 15.

[0021] Furthermore, no element 15 is formed in the annular region (the so-called peripheral remaining region) of the wafer 11, which extends only a predetermined distance (e.g., 3 mm) along the radial direction of the wafer 11 from the outer periphery end.

[0022] The wafer 11 follows predetermined processing procedures such as DBG (Dicing Before Grinding) and SDBG (Stealth Dicing Before Grinding) to form dicing grooves 13b in each predetermined dicing line 13 (see Figure 2(A)), thereby being diced into multiple wafers 21.

[0023] In the DBG process, firstly, grooves (so-called half-cut grooves) that do not reach the predetermined depth of the back side 11b are formed on each predetermined dividing line 13 on the front side 11a of the wafer 11. Then, the wafer 11 is divided into multiple wafers 21 by grinding the back side 11b.

[0024] Furthermore, in the SDBG process, firstly, a modifier layer (not shown) is formed at a predetermined depth along each predetermined dicing line 13 on the wafer 11, and then the back side 11b is ground. In this way, by causing cracks to extend from the modifier layer to the front side 11a and the back side 11b, the wafer 11 is diced into multiple wafers 21.

[0025] Thus, when the wafer 11 is divided into multiple chips 21 by grinding, the back side 11b side is ground while the front side 11a side is covered with a resin film (protective member) 17 (see Figure 2(A)).

[0026] Therefore, after dicing the wafer 11, a wafer unit 23 is formed by integrally holding multiple wafers 21 with a film 17 (see Figure 2(A)). Figure 2(A) is a perspective view of the wafer unit 23, and Figure 2(B) is a cross-sectional view of the wafer unit 23.

[0027] As shown in Figures 2(A) and 2(B), the front side 11a of wafer 11 corresponds to the front side 11a of wafer 21, and the back side 11b of wafer 11 corresponds to the back side 11b of wafer 21.

[0028] The thickness of the plurality of wafers 21 in the wafer unit 23 is measured using the measuring device 2 (see Figure 3) and following the measurement method described later (see Figure 4). Therefore, the measuring device 2 will be described first.

[0029] Figure 3 is a perspective view of measuring device 2. In Figure 3, for ease of explanation, functional blocks represent some of the constituent elements. The X-axis, Y-axis, and Z-axis directions shown in Figure 3 are mutually orthogonal. The +X and -X directions are opposite to each other and are both parallel to the X-axis direction.

[0030] Similarly, the +Y and -Y directions are opposite to each other and both are parallel to the Y-axis. Furthermore, the +Z and -Z directions are opposite to each other and both are parallel to the Z-axis. Additionally, in this specification, the +Z direction may sometimes be referred to as "up" and the -Z direction as "down."

[0031] The measuring device 2 has a circular plate-shaped chuck stage (holding stage) 4. As shown in Figure 5, the chuck stage 4 has a circular plate-shaped frame 4b made of non-porous ceramic. A circular plate-shaped recess 4c is formed on the upper surface of the frame 4b.

[0032] A porous plate 4d made of porous ceramic is fixed in the recess 4c. Grooves 4e and holes 4f are formed at the bottom of the recess 4c to transmit negative pressure to the porous plate 4d. When negative pressure is applied to the porous plate 4d through the grooves 4e and holes 4f from a suction source (not shown) such as a vacuum pump or ejector, negative pressure will be generated on the upper surface of the porous plate 4d.

[0033] The upper surfaces of the frame 4b and the perforated plate 4d function as holding surfaces 4a for attracting and holding the wafer units 23 (i.e., multiple wafers 21). The multiple wafers 21 are held with their back surfaces 11b exposed on top, and their front surfaces 11a are attracted and held by the holding surfaces 4a through the adhesive film 17.

[0034] Here, returning to Figure 3, a cylindrical column 6 is connected to the lower part of the chuck table 4. A driven pulley (not shown) is arranged concentrically with the column 6. A rotary drive unit 8, including a drive pulley, a motor, an endless belt, etc. (all not shown), is arranged near the column 6.

[0035] An endless belt is mounted on the driven pulley and drive pulley. When the rotary drive unit 8 is driven, power is transmitted to the column 6 through the endless belt, and the chuck table 4 rotates with the column 6 as the rotation axis 6a. The chuck table 4 can rotate without limitation regardless of whether the rotation angle is clockwise or counterclockwise when viewed from above.

[0036] An angle measuring unit 10 is provided on the lower side of the column 6. The angle measuring unit 10 is also referred to as a rotary encoder. The angle measuring unit 10 in this embodiment includes a circular scale (not shown) and an optical reader (not shown) provided on the lower surface of the scale for reading the scale.

[0037] The scale is fixed concentrically to the column 6, and the scale graduations are formed at equal intervals along the circumference of the scale. The reader is disposed on the lower surface of the scale and includes a light-emitting part such as an LED to illuminate the measurement light.

[0038] The reader further includes a light-receiving unit such as a light sensor that receives reflected light from the scale markings. The reader has predetermined circuitry for signal processing. The predetermined circuitry counts the number of times the reflected light is received by the light-receiving unit and transmits information about the rotation angle of the chuck stage 4 as an electrical signal to the control unit 26, which will be described later.

[0039] The control unit 26 can determine the rotation angle of the chuck stage 4 based on this electrical signal. However, the angle measuring unit 10 is not limited to the aforementioned reflective optical type. As long as the angle measuring unit 10 can obtain information on the rotation angle of the chuck stage 4, it can also be a transmissive optical type, or a magnetic or electromagnetic induction type.

[0040] A ball screw-type Y-axis moving unit 12 is provided above the chuck table 4. The Y-axis moving unit 12 causes the moving block 14 to move along the Y-axis. On the -X direction side of the moving block 14, there is a pair of guide rails (not shown) extending along the Y-axis.

[0041] The movable block 14 is slidably mounted on a pair of guide rails. A nut (not shown) is provided on the side of the movable block 14 in the -X direction, and a ball screw (not shown) extending along the Y-axis direction is rotatably connected to the nut.

[0042] A drive source such as a stepper motor (not shown) is provided at one end of the ball screw in the Y-axis direction. When the drive source is activated, the moving block 14 moves along the Y-axis direction. Two arms (first arm 16 and second arm 20) are provided on the side of the moving block 14 in the +X direction in a state of separation along the Y-axis direction.

[0043] A thickness measurement unit 18 is provided in the first arm 16. The thickness measurement unit 18 is a reflective spectrophotometer (also simply called a film thickness gauge) using a microscope, and can measure the thickness of a small area (measurement point PA shown in FIG. 6) on the wafer 21 by means of optical interferometry.

[0044] In this embodiment, the symbol PA is used to conveniently indicate one of the plurality of measurement points in each wafer 21, and does not indicate only a specific measurement point. Furthermore, as described later, for convenience, each of the plurality of measurement points PA that is not used in calculating the average thickness is marked as measurement point PB.

[0045] The head 18a of the thickness measuring unit 18 is configured to face the holding surface 4a in the Z-axis direction. A condenser lens (not shown) with its optical axis arranged along the Z-axis direction is provided on the head 18a.

[0046] Measurement light (e.g., infrared light) derived from a light source such as an LED and capable of penetrating the wafer 21 in a predetermined wavelength band is directed from the head 18a onto each wafer 21 held by the holding surface 4a in a focused manner near the back surface 11b. The head 18a receives reflected light from the back surface 11b of the wafer 21 and reflected light from the front surface 11a of the wafer 21.

[0047] The interference light system of reflected light from the back side 11b and the front side 11a is detected by an optical detection element (not shown) such as an optical sensor in the thickness measurement unit 18. When the wavelength of the measurement light is set as the horizontal axis and the reflectivity of the interference light as the vertical axis, the spectrum of the interference light changes according to the optical path difference of the reflected light. Based on this spectrum, the thickness at each measurement point PA on each wafer 21 is measured.

[0048] A camera unit 22 is provided on the second arm 20. The camera unit 22 is an optical microscope, and it obtains an image including the dividing groove 13b formed on the dividing predetermined line 13 by photographing the back side 11b of the chip 21.

[0049] The head 22a of the camera unit 22 is configured to face the holding surface 4a in the Z-axis direction. A condenser lens (not shown) with its optical axis arranged along the Z-axis direction is provided inside the head 22a.

[0050] The camera unit 22 has a light source (not shown) such as an LED that can illuminate measurement light (e.g., white light), and illuminates the chip 21 from the head 22a through a focusing lens in a manner that focuses the light near the back surface 11b.

[0051] The measurement light beam illuminating the chip 21 is reflected by the back surface 11b and then guided through a condenser lens to a line sensor (image sensor) located in the image unit 22 (not shown).

[0052] Line sensors, also known as linear image sensors, have multiple light-emitting diodes or transistors arranged in a straight line, and constitute a CCD (Charge-Coupled Device) image sensor or a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor.

[0053] The heads 18a and 22a are sufficiently separated from each other to the extent that they do not affect the measurement. For example, the optical axes of the condenser lenses of the heads 18a and 22a are separated from each other by more than the radius of the wafer 11.

[0054] In addition, between heads 18a and 22a, in order to prevent interference between the measurement light from head 18a and the measurement light from head 22a, a partition plate (not shown) that has undergone a predetermined surface treatment or has formed a light absorption film on the surface can also be provided.

[0055] The rotary drive unit 8 and the Y-axis moving unit 12 function as a moving unit 24, which causes the thickness measuring unit 18 and the camera unit 22 to move relative to the chuck stage 4.

[0056] In this embodiment, with the spot of the measurement light irradiated from the thickness measurement unit 18 and the camera unit 22 positioned near the back surface 11b of the wafer 21 in the Z-axis direction, the moving block 14 is moved in the +Y direction while the chuck stage 4 is rotated counterclockwise when viewed from above.

[0057] This allows for the measurement of the thickness of each wafer 21 at multiple measurement points PA, and the acquisition of an overall image of the back side 11b. The movements of the chuck stage 4, angle measurement unit 10, thickness measurement unit 18, camera unit 22, and movement unit 24 are controlled by the control unit 26.

[0058] The control unit 26 in this embodiment is constructed by a computer, which has: a processor (processing device) represented by a CPU (Central Processing Unit); and a memory (memory device) including a main memory device and an auxiliary memory device.

[0059] The main memory devices include DRAM (Dynamic Random Access Memory), SRAM (Static Random Access Memory), ROM (Read Only Memory), etc., while the auxiliary memory devices include flash memory, hard disk drives, solid-state drives, etc.

[0060] The auxiliary memory device stores software containing a predetermined program. By following this software, the processing device and other components are operated, thereby realizing the function of the control unit 26. Part of the auxiliary memory device functions as a memory unit 28, which stores predetermined information obtained by measurement using the thickness measuring unit 18 and the camera unit 22.

[0061] The memory section 28 stores the thickness data 28a of each wafer 21. The thickness data 28a is formed by establishing the relationship between the respective positions (i.e., coordinates) of multiple measurement points PA in a predetermined plane (e.g., the plane corresponding to the back surface 11b of the wafer 11 before or after dicing) with the notch 11c as a reference and the thickness in the multiple measurement points PA.

[0062] In starting the thickness measurement, the position of the moving block 14 and the chuck stage 4 is adjusted in such a way that the spot of the measuring light illuminated from the head 18a will be located in the gap 11c . Therefore, the position of each measurement point PA is automatically calculated based on the initial position of the light spot, the rotation speed of the chuck stage 4 and the movement speed of the moving block 14 and by resorting to the control unit 26 .

[0063] In this embodiment, the chuck table 4 has a rotational speed of 300 degrees / s and the moving block 14 has a moving speed of 9,308 μm / s, with multiple measuring points PA become swirlingly configured on the back side 11b (refer to Fig. 6 ).

[0064] Furthermore, in the direction of travel of the moving block 14 shown in FIG. 3 (+Y direction), the head 22a of the camera unit 22 is configured in the reverse direction (-Y direction) only a predetermined distance from the head 18a of the thickness measurement unit 18 . Accordingly, the shooting of the back 11b begins after a predetermined time calculated by dividing the predetermined distance by the moving speed of the moving block 14 .

[0065] Image data 28b taken by the camera unit 22 are also remembered in the memory section 28 . By executing a first program of memory in an auxiliary memory device with a processing device, the image data 28b is synthesized into one image of a camera range having a predetermined size (e.g., a square with a side length of about 335 mm).

[0066] Further, the composite image is image processed by means of a second program memorized in the auxiliary memory device, and in each of the plurality of segmentation slots 13b the position of the centerline 13b 1 (referring to FIG. 7 ) in the width direction of the segmentation slot 13b , the width direction of the segmentation slot 13b is orthogonal to the length direction of the segmentation slot 13b .

[0067] For example, first, by resorting to image processing such as edge detection on the synthesized image, the coordinates of the two edges specific to the width direction of a segmentation slot 13b . Then, by calculating the central position of the coordinates of these two edges, the position of the center line 13b 1 that divides the width direction of the slot 13b can be specified.

[0068] The position of the center line 13b 1 of each split slot 13b is also memorized in the memory portion 28 . The position of centerline 13b 1 of this embodiment is also positioned (i.e., coordinates) within a predetermined plane (e.g., the plane corresponding to the backside 11b of the split wafer 11 ) with the gap 11c as the position of the respective measurement point PA similarly.

[0069] For example, for one of the chips 21 shown in Figure 7, the positions of one center line 13b 1 parallel to the Y-axis direction (X=c) and another center line 13b 1 parallel to the Y-axis direction (X=d) are memorized.

[0070] Furthermore, for this chip 21, the positions of one center line 13b 1 parallel to the X-axis direction (Y=e) and the position of another center line 13b 1 parallel to the X-axis direction (Y=f) are stored.

[0071] The auxiliary memory device stores a third program. The third program is executed by the processing device to perform the function of the setting unit 30, which sets a region Q (see Figure 10) on the outer periphery (around) of each chip 21 that does not refer to the thickness data based on the thickness data 28a of each chip 21.

[0072] The non-referenced thickness data area Q is located at a predetermined distance R from the position of the center line 13b 1 of a dividing groove 13b (see Figure 10), and includes the measurement point PB where the thickness data 28a is not referenced when calculating the thickness of the wafer 21.

[0073] Furthermore, a fourth program is stored in the auxiliary memory device. The fourth program is executed by the processing device to function as a calculation unit 32, which calculates the average thickness of the multiple measurement points PA remaining after removing the measurement points PB contained in the non-reference thickness data region Q from the multiple measurement points PA.

[0074] Next, referring to Figures 4 to 10, a method for measuring the thickness of multiple wafers 21 using the measuring device 2 will be described. Figure 4 is a flowchart of the measurement method.

[0075] First, as shown in FIG5, the front side 11a of the plurality of wafers 21 formed by the diced wafer 11 is held by the holding surface 4a through the adhesive film 17 (holding step S10). FIG5 is a diagram showing the holding step S10.

[0076] In the holding step S10, the wafer unit 23 is attracted and held when the imaginary straight line W1 (see Figure 6) connecting the radial center position W0 of the back surface 11b and the notch 11c is parallel to the Y-axis direction and the notch 11c is located on the -Y direction side.

[0077] After holding step S10, the thickness measurement is performed by the thickness measurement unit 18 and the segmentation groove 13b is photographed from the back side 11b by the imaging unit 22 (measurement and imaging step S20).

[0078] In the measurement and imaging step S20, with the light spot of the measuring light from the head 18a of the thickness measuring unit 18 positioned in the notch 11c, the chuck stage 4 begins to rotate in a predetermined direction, and the moving block 14 begins to move in the predetermined direction at the same time.

[0079] In this embodiment, the chuck stage 4 rotates counterclockwise when viewed from above, and the moving block 14 moves in the +Y direction. In the measurement and imaging step S20, thickness data 28a is obtained using the thickness measurement unit 18, and image data 28b is obtained using the imaging unit 22.

[0080] After the measurement and imaging steps S20, the memory unit 28 stores the aforementioned thickness data 28a (first memory step S30). Figure 6 is a schematic diagram of the multiple measurement points PA that become the objects of memory in the first memory step S30.

[0081] In this embodiment, the number of measurement points PA is approximately 800,000. In Figure 6, for ease of reading and convenience, a much smaller number of measurement points PA than 800,000 is depicted. In thickness measurement, the thickness at each measurement point P2, P3, ... PZ is measured sequentially from measurement point P1 (z is a natural number greater than 3, approximately 800,000 in this example).

[0082] Therefore, when there are 968 rectangular areas on wafer 11 where components 15 are formed, the average number of measurement points PA per wafer 21 is approximately 826. Furthermore, when there are 58,788 rectangular areas on wafer 11 where components 15 are formed, the average number of measurement points PA per wafer 21 is approximately 14.

[0083] As described above, the position of each measurement point PA is defined within a predetermined plane (the plane corresponding to the back surface 11b of the wafer 11 before or after dicing) with the notch 11c as the reference. After the measurement and imaging step S20, the memory unit 28 stores the position of the center line 13b 1 of each dicing groove 13b obtained by processing the image data 28b (second memory step S40).

[0084] Furthermore, the first memory step S30 and the second memory step S40 can be performed either first or simultaneously. Figure 7 is a schematic diagram of the center lines 13b 1 of the plurality of dividing slots 13b that become memory objects in the second memory step S40 in a chip 21. In Figure 7, for ease of explanation, several measurement points PA are omitted.

[0085] As mentioned above, the position of each centerline 13b 1 is also defined in a predetermined plane (the plane corresponding to the back surface 11b of the wafer 11 before or after slitting) with the notch 11c as the reference.

[0086] After the first memory step S30 and the second memory step S40, based on the thickness data 28a of each chip 21, the setting unit 30 sets a region Q around each chip 21 that does not refer to the thickness data (setting step S50).

[0087] Figure 8 is a schematic diagram showing multiple measurement points PA in a wafer 21 and the center lines 13b 1 (X=c), (X=d), (Y=e), and (Y=f) of multiple segments 13b surrounding the wafer 21. Figure 9 is a graph showing the non-reference thickness data area Q set in the setting step S50.

[0088] The horizontal axis of Figure 9 represents the distance (μm) from one of the adjacent center lines 13b 1 (X=c) in the -X direction of the wafer 21 shown in Figure 8 to each measurement point PA located within the wafer 21. The origin of the horizontal axis corresponds to the position of the center line 13b 1 (X=c). Furthermore, the vertical axis of Figure 9 represents the thickness (μm) of the wafer 21 at the multiple measurement points PA.

[0089] In Figure 9, the average thickness and standard deviation at measurement point PA are calculated in each of several unit regions divided into 10 μm units from a centerline 13b 1. Furthermore, the line graph shown in Figure 9 is plotted by connecting the average values ​​of the unit regions.

[0090] The bars shown in Figure 9, arranged parallel to the vertical axis, schematically represent the standard deviation of the thickness of the measurement point PA in a unit area, with the upper end of the bar representing (mean + σ) and the lower end representing (mean - σ).

[0091] When creating the chart in Figure 9, firstly, the thickness data 28a of multiple measurement points PA located within the range enclosed by the four center lines 13b 1 (X=c), (X=d), (Y=e), and (Y=f) is extracted.

[0092] Then, in the data containing the distance from the centerline 13b 1 (X=c) to each measurement point PA and the thickness at each measurement point PA, outliers in the thickness are removed by known processing methods such as M-estimation. This enables robust data analysis.

[0093] As shown in Figure 9, the thickness deviation at measurement point PA is relatively large near the distance of 0 μm (i.e., center line 13b 1). The thickness deviation near center line 13b 1 is considered to reflect the thickness in the predetermined dividing line 13 where no insulating film or the like has been formed.

[0094] In contrast, as one moves away from the center line 13b 1, the average value becomes approximately constant, and the magnitude of the standard deviation (i.e., the length of the strip) decreases. This change in the average value and standard deviation suggests that the thickness of the wafer 21 becomes more constant the closer one gets to the center of the back surface 11b of the wafer 21.

[0095] Therefore, in this embodiment, firstly, following predetermined judgment conditions, the moving average of the standard deviation and the average thickness of each unit area is judged from near to far from the center line 13b 1 (that is, the difference of the average values ​​in adjacent unit areas).

[0096] In this embodiment, the predetermined judgment condition is that the standard deviation is less than 0.5 μm and the moving average of the average thickness is less than 0.1 μm. The setting unit 30 sets the distance R from the center line 13b 1 to the unit area that first meets this judgment condition.

[0097] Then, the area farther than the distance R from the center line 13b 1 is set as the effective area T (see Figures 9 and 10). The effective area T contains the measurement point PA used when calculating the thickness of the wafer 21. That is, the setting unit 30 sets the area from the center line 13b 1 to the distance R as the non-reference thickness data area Q. The non-reference thickness data area Q contains the measurement point PB that is not used for the average value of the thickness calculation.

[0098] In the example shown in Figure 9, the region from a center line 13b1 to a position of 500 μm does not meet the above criteria and is therefore designated as the non-reference thickness data region Q. In contrast, the region closer to the +X direction than the 500 μm position becomes the valid region T.

[0099] In Figures 8 and 9, the non-referenced thickness data region Q and the effective region T are explained with reference to a center line 13b 1 (X=c) of a wafer 21 adjacent to the -X direction.

[0100] However, in setting step S50, following the above judgment conditions, the distance R1 is further calculated using the center line 13b1 (X=d) as the reference, the distance R2 is further calculated using the center line 13b1 (Y=e) as the reference, and the distance R3 is further calculated using the center line 13b1 (Y=f) as the reference.

[0101] Thus, a region Q without reference thickness data is set in a rectangular annular area, which includes: a region extending from the center line 13b1 (X=c) in the +X direction to a distance R; a region extending from the center line 13b1 (X=d) in the -X direction to a distance R1; a region extending from the center line 13b1 (Y=e) in the -Y direction to a distance R2; and a region extending from the center line 13b1 (Y=f) in the +Y direction to a distance R3.

[0102] Then, the area enclosed by the non-reference thickness data area Q is set as the effective area T. The setting unit 30 performs the extraction of the corresponding thickness data 28a, the removal of thickness outliers, the calculation of the distance R according to the judgment conditions, and the setting of the non-reference thickness data area Q in this way.

[0103] Furthermore, the setting unit 30 similarly sets a non-reference thickness data region Q (i.e., an effective region T) in each of the plurality of wafers 21 in the wafer 11.

[0104] Figure 10 is a diagram showing the non-reference thickness data area Q in the setting step S50 and the multiple measurement points PA located in the effective area T and becoming the calculation objects in the calculation step S60.

[0105] After setting step S50, the calculation unit 32 calculates the average thickness of the multiple measurement points PA remaining after removing the measurement points PB contained in the non-reference thickness data region Q from the multiple measurement points PA (calculation step S60).

[0106] In calculation step S60, the average thickness of the multiple measurement points PA remaining after removing the measurement points PB contained in the non-reference thickness data region Q is calculated for a single wafer 21. Subsequently, in the wafer 21 picking and other processing procedures, each average value is processed as the corresponding thickness of the wafer 21.

[0107] Thus, the calculation unit 32 calculates the average thickness at multiple measurement points PA present in the effective region T. Furthermore, the calculation unit 32 similarly calculates the average thickness at each of the multiple wafers 21 in the wafer 11.

[0108] In this embodiment, since the thickness of each wafer 21 can be calculated without including the thickness near the dividing line 13, the influence of the thickness of the outer periphery of each wafer 21 can be reduced compared to the case where the thickness in the dividing line 13 is taken into account as the thickness of the wafer 21, and the thickness of each wafer 21 can be measured more accurately.

[0109] Furthermore, the structure and method of the above-described embodiments can be appropriately modified and implemented as long as they do not depart from the scope of the present invention. For example, in the case where an oriented plane is formed to replace the notch 11c, the central position of the straight line formed by the oriented plane in the length direction performs the same function as the notch 11c.

[0110] However, the average thickness (i.e., the thickness) of each of the multiple chips 21 can also be displayed to the operator via a display device (not shown) along with the overall view of the wafer 11 shown in FIG. 6. Furthermore, chips 21 whose average thickness exceeds a predetermined tolerance can also be displayed to the operator in a conspicuous manner. This makes it easier for the operator to identify good or defective chips 21.

[0111] 2: Measuring device 4: Chuck stage (holding stage) 4a: Maintain surface 4b: Frame 4c: concave part 4d: Perforated plate 4e: slot 4f: Hole 6: Column 6a: Rotation axis 8: Rotary drive unit 10: Angle Measurement Unit 12: Y-axis direction moving unit 14: Moving Block 11: Wafer 11a: Front 11b: Back 11c: Notch (cut) W 0: Center position W 1: Straight line 13: Pre-defined dividing line 13a: Width 13b: Dividing slot 13b 1: Centerline 15: Components 17: Adhesive film (protective components) 21: Chip 23: Chip Unit 16: First Arm 18: Thickness Measurement Unit 18a: Head 20: Second Arm 22: Camera Unit 22a: Head 24: Moving Unit 26: Control Unit 28: Memory Department 28a: Thickness data 28b: Image data 30: Setting Department 32: Computing Department P1, P2, P3, PA, PB, PZ: Measurement points Q: Areas without reference to thickness data R, R1, R2, R3: Distance T: Effective area S10: Maintain the steps S20: Measurement and Photography Procedures S30: First Memory Step S40: Second Memory Step S50: Setup Steps S60: Calculation Steps

Claims

1. A measurement method for measuring the thickness of a plurality of wafers, wherein the wafer is divided into the plurality of wafers by forming dicing grooves in each predetermined dicing line on the wafer, the wafer having elements formed in each of a grid-like region divided by a plurality of predetermined dicing lines set on the front side and each having a predetermined width, the measurement method being characterized by comprising: a holding step, wherein the front side of the plurality of wafers is held by a holding surface of a holding stage and by a protective member; a first memory step, wherein the thickness of each of the plurality of wafers held by the holding surface is measured at a plurality of measuring points and the thickness data of each wafer is stored, the thickness data being generated by establishing a correlation between the position of each of the plurality of measuring points in a predetermined plane corresponding to the back side of the wafer and the thickness at the plurality of measuring points; and a second memory step, wherein the back side of the plurality of wafers held by the holding surface is photographed and the position of the center line of each dicing groove in the predetermined plane is stored. The process includes a setting step, which sets a non-reference thickness data region around each wafer based on the thickness data of each wafer. The non-reference thickness data region is located below a predetermined distance from the center line of each slit groove and includes measurement points where the thickness data is not referenced when calculating the thickness of each wafer. The process also includes a calculation step, which calculates the average thickness of the remaining measurement points in each of the plurality of wafers after removing the measurement points contained in the non-reference thickness data region from the plurality of measurement points.

2. A measuring apparatus for measuring the thickness of a plurality of wafers, wherein the wafer is diced into the plurality of wafers by forming dicing grooves in predetermined dicing lines on the wafer, the wafer having elements formed in each of a grid-like region divided by a plurality of predetermined dicing lines set on the front side and each having a predetermined width, the measuring apparatus being characterized by comprising: a holding stage having a holding surface capable of holding the front side of the plurality of wafers by means of a protective member; a thickness measuring unit measuring the thickness of each of the plurality of wafers held by the holding surface at a plurality of measuring points; a camera unit capturing images of the back side of the plurality of wafers held by the holding surface; a moving unit moving the thickness measuring unit and the camera unit relative to the holding stage; and a control unit controlling at least the holding stage, the thickness measuring unit, the camera unit, and the moving unit, the control unit comprising: The memory unit stores: thickness data for each wafer, which is established by relating the positions of multiple measurement points in a predetermined plane corresponding to the back side of the wafer to the thickness at those multiple measurement points; and the position of the center line, which is the center line of each slit groove captured by the camera unit and located in the predetermined plane; a setting unit sets a non-reference thickness data region around each wafer based on the thickness data, the non-reference thickness data region being located at a predetermined distance below the position of the center line of each slit groove and including measurement points where the thickness data is not referenced when calculating the thickness of each wafer; and a calculation unit calculates the average thickness of the measurement points remaining after removing the measurement points included in the non-reference thickness data region from the multiple measurement points in each of the multiple wafers.

Images