Measurement method and measurement device

Abstract

To measure the thickness of each chip reducing the influence of the thickness of peripheral parts of each chip.SOLUTION: Provided is a measurement method comprising: a first storing step for measuring the thickness of each of a plurality of chips held on a holding face of a holding table via a protective member and storing the thickness data of each chip in which the respective positions of a plurality of measurement points on a prescribed plane corresponding to the reverse side of a wafer and the thicknesses at the plurality of measurement points are associated; a second storing step for imaging the reverse side of the plurality of chips and storing the center line position of each divide groove in the prescribed plane; a setting step for setting a thickness data unreferenced region to the surrounding of each chip that includes a measurement point where the thickness data is not referred to when the thickness of each chip is calculated; and a calculation step for calculating the average value of thicknesses at measurement points in each of the plurality of chips, excluding the measurement points included in the thickness data unreferenced region from the plurality of measurement points.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to a measuring method for measuring the thickness of a plurality of chips and a measuring device for measuring the thickness of a plurality of chips. [Background technology]

[0002] When manufacturing device chips (hereinafter simply referred to as chips) from a wafer, for example, a wafer in which a plurality of planned division lines (streets) are set in a grid pattern on the surface and devices such as ICs (Integrated Circuits) are formed in each rectangular area partitioned by the plurality of streets is divided into chips on a device-by-device basis.

[0003] After division, the thickness of each chip is measured to check whether it is within a tolerance range. For example, before dividing the wafer into multiple chips, the thickness of each rectangular area on the wafer is measured to evaluate the thickness of each chip after division (see Patent Document 1).

[0004] The thickness is measured without contacting the wafer using an optical thickness measuring device. For example, while the wafer is irradiated with measurement light from the thickness measuring device, the chuck table holding the wafer by suction is rotated around a predetermined rotation axis, and the measurement light spot is moved linearly from the outside to the inside along the diameter of the wafer. This causes the spot (i.e., the measurement point where the thickness is measured) to move in a spiral pattern.

[0005] Based on the rotation angle information of the chuck table and the position information of the thickness gauge, the position information of each measurement point relative to the wafer is obtained, and thickness data is acquired that associates the position information and thickness information at each measurement point.Then, the average value of the thicknesses obtained by measuring the thickness of one rectangular area at multiple measurement points is regarded as the thickness of the chip.

[0006] However, the thickness of a wafer at the dividing lines may differ from the thickness of the central part of a rectangular region surrounded by the dividing lines. For example, in order to improve the processing quality when cutting the dividing lines, an insulating film or the like that may degrade the processing quality may not be formed on the dividing lines.

[0007] Therefore, the thickness at the dividing line may be thinner than the thickness at the center of the rectangular region, and if the thickness at the dividing line is incorporated into the thickness data when calculating the thickness of the chip, the calculated thickness may differ from the thickness of the chip that should be measured. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Japanese Patent Application Laid-Open No. 2017-199845 Summary of the Invention [Problem to be solved by the invention]

[0009] The present invention has been made in view of the above problems, and has as its object to measure the thickness of each chip while reducing the influence of the thickness of the outer periphery of each chip. [Means for solving the problem]

[0010] According to one aspect of the present invention, there is provided a method for measuring the thickness of a plurality of chips, the method comprising the steps of: a wafer on a surface of which a device is formed in each of areas partitioned in a grid pattern by a plurality of planned dividing lines, each having a predetermined width; the wafer being divided into the plurality of chips by forming dividing grooves in each planned dividing line; the method including a holding step of holding the front side of the plurality of chips on a holding surface of a holding table via a protective member; and measuring the thickness of each of the plurality of chips held on the holding surface at a plurality of measurement points, and correlating the positions of the plurality of measurement points in a predetermined plane corresponding to the back surface of the wafer with the thickness at the plurality of measurement points. a setting step of setting a thickness data non-reference area around each chip based on the thickness data of each chip, the thickness data non-reference area including measurement points that are located at a predetermined distance or less from the position of the center line of each division groove and whose thickness data is not referenced when calculating the thickness of each chip; and a calculation step of calculating, for each of the plurality of chips, an average thickness at the measurement points excluding the measurement points included in the thickness data non-reference area from the plurality of measurement points.

[0011] According to another aspect of the present invention, there is provided a measuring device for measuring the thickness of a plurality of chips, the measuring device comprising: a wafer having a surface divided into a grid pattern by a plurality of planned dividing lines, each having a predetermined width, and having devices formed in each of the areas; the wafer being divided into the plurality of chips by forming dividing grooves in each planned dividing line; the measuring device comprising: a holding table having a holding surface capable of holding the front side of the plurality of chips via a protective member; a thickness measuring unit that measures the thickness of each of the plurality of chips held on the holding surface at a plurality of measurement points; an imaging unit that images the back side of the plurality of chips held on the holding surface; a moving unit that moves the thickness measuring unit and the imaging unit relative to the holding table; and a measuring device including at least the holding table, the thickness measuring unit, the imaging unit, and and a control unit for controlling the moving unit, the control unit having: a memory section for storing thickness data for each chip, in which the positions of the plurality of measurement points in a predetermined plane corresponding to the back surface of the wafer are associated with the thickness at the plurality of measurement points; and the center line of each division groove imaged by the imaging unit, the position of the center line in the predetermined plane; a setting section for setting a thickness data non-reference area around each chip based on the thickness data, the thickness data non-reference area including measurement points that are located within a predetermined distance from the position of the center line of the division groove and whose thickness data is not referenced when calculating the thickness of each chip; and a calculation section for calculating, for each of the plurality of chips, an average thickness at the measurement points excluding the measurement points included in the thickness data non-reference area from the plurality of measurement points. [Effects of the Invention]

[0012] In a measurement method and measurement device according to one aspect of the present invention, a thickness data non-reference area is set around the chip, which area is located within a predetermined distance from the center line of the dividing groove and includes measurement points for which thickness data is not referenced when calculating the thickness of each chip.

[0013] The thickness of each chip is calculated by averaging the thicknesses at multiple measurement points, excluding those included in the thickness data non-reference area, so that the thickness of the chip can be calculated without including the thickness near the planned division line. This reduces the influence of the thickness of the outer periphery of each chip, allowing for more accurate measurement of the thickness of each chip. [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 2 is a perspective view schematically showing a wafer. [Figure 2] FIG. 2A is a perspective view of the chip unit, and FIG. 2B is a cross-sectional view of the chip unit. [Figure 3] FIG. [Figure 4] FIG. 1 is a flow diagram of a measurement method. [Figure 5] FIG. [Figure 6] FIG. 10 is a diagram schematically showing a plurality of measurement points to be stored in a first storage step. [Figure 7] 10 is a diagram schematically showing center lines of a plurality of dividing grooves to be stored in a second storing step. FIG. [Figure 8] FIG. 10 is a schematic diagram showing a plurality of measurement points and the center lines of a plurality of division grooves in one chip. [Figure 9] 10 is a graph showing thickness data non-reference regions and the like set in the setting step. [Figure 10] 10 is a diagram showing a thickness data non-reference area in the setting step and a plurality of measurement points that are the subject of calculation in the calculation step; FIG. DETAILED DESCRIPTION OF THE INVENTION

[0015] An embodiment according to one aspect of the present invention will be described with reference to the accompanying drawings. First, a wafer 11 in a state before being divided into a plurality of chips 21 (see FIG. 2(A)) will be described. FIG. 1 is a perspective view schematically showing the wafer 11.

[0016] The wafer 11 has a disk shape and has a single crystal substrate made of a semiconductor material such as silicon. However, the wafer 11 is not limited to a single crystal silicon substrate, and may have a single crystal substrate made of a compound semiconductor such as silicon carbide (SiC) or gallium nitride (GaN).

[0017] Each wafer 11 has a circular front surface 11a and a circular back surface 11b. A notch (notch) 11c indicating the crystal orientation of the wafer 11 is formed in a part of the outer periphery of the wafer 11. Note that an orientation flat (notch) may be formed instead of the notch 11c.

[0018] On the front surface 11a, a plurality of dividing lines 13, each having a predetermined width 13a, are set in a grid pattern. For example, on a wafer 11 having a diameter of 300 mm (12 inches), a plurality of dividing lines 13, each having a width 13a of 0.1 mm, are set in a grid pattern. In this case, there are 968 rectangular regions, each 8 mm square, on the front surface 11a.

[0019] Also, for example, a plurality of planned dividing lines 13, each having a width 13a of 0.07 mm, are set in a grid pattern on a wafer 11 having a diameter of 300 mm (12 inches). In this case, there are 58,788 rectangular regions, each 1 mm square, on the front surface 11a side.

[0020] Devices 15 such as ICs and LEDs (Light Emitting Diodes) are formed in each of the rectangular regions defined by the dividing lines 13. However, there are no particular limitations on the size of the wafer 11, the type, number, shape, structure, size, arrangement, etc. of the devices 15.

[0021] In addition, no devices 15 are formed in a band-shaped annular region (so-called peripheral excess region) extending from the outer peripheral edge of the wafer 11 to a position a predetermined distance (e.g., 3 mm) inward along the radial direction of the wafer 11.

[0022] The wafer 11 is divided into multiple chips 21 by forming division grooves 13b (see Figure 2(A)) within each planned division line 13 according to a predetermined processing process such as a DBG (Dicing Before Grinding) process or an SDBG (Stealth Dicing Before Grinding) process.

[0023] In the DBG process, first, grooves (so-called half-cut grooves) of a predetermined depth that do not reach the back surface 11b are formed on each planned dividing line 13 on the front surface 11a of the wafer 11, and then the back surface 11b side is ground to divide the wafer 11 into multiple chips 21.

[0024] In the SDBG process, a modified layer (not shown) is first formed along each of the division lines 13 at a predetermined depth in the wafer 11, and then the back surface 11b is ground. As a result, cracks are propagated from the modified layer to the front surface 11a and the back surface 11b, dividing the wafer 11 into a plurality of chips 21.

[0025] In this way, when dividing the wafer 11 into multiple chips 21 by grinding, grinding is performed on the back surface 11b side with a resin tape (protective member) 17 attached to the front surface 11a side (see Figure 2(A)).

[0026] Therefore, after dividing the wafer 11, a chip unit 23 is formed in which a plurality of chips 21 are integrally held by the tape 17 (see FIG. 2(A)). FIG. 2(A) is a perspective view of the chip unit 23, and FIG. 2(B) is a cross-sectional view of the chip unit 23.

[0027] As shown in FIGS. 2A and 2B, the front surface 11a of the wafer 11 corresponds to the front surface 11a of the chip 21, and the back surface 11b of the wafer 11 corresponds to the back surface 11b of the chip 21.

[0028] The thicknesses of the plurality of chips 21 in the chip unit 23 are measured using a measuring device 2 (see FIG. 3) in accordance with a measuring method (see FIG. 4) described below. First, the measuring device 2 will be described.

[0029] Fig. 3 is a perspective view of the measurement device 2. For ease of explanation, Fig. 3 shows some of the components as functional blocks. The X-axis direction, Y-axis direction, and Z-axis direction shown in Fig. 3 are directions perpendicular to one another. The +X-direction and -X-direction are opposite to one another and are both parallel to the X-axis direction.

[0030] Similarly, the +Y direction and the -Y direction are opposite to each other and are both parallel to the Y axis direction. The +Z direction and the -Z direction are opposite to each other and are both parallel to the Z axis direction. In this specification, the +Z direction may be referred to as upward and the -Z direction may be referred to as downward.

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

[0032] A porous plate 4d made of porous ceramics is fixed in the recess 4c. Grooves 4e, holes 4f, etc. are formed in 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 from a suction source (not shown) such as an ejector via the grooves 4e, holes 4f, etc., negative pressure is generated on the upper surface of the porous plate 4d.

[0033] The upper surfaces of the frame 4b and the porous plate 4d function as a holding surface 4a that suction-holds the chip unit 23 (i.e., the plurality of chips 21). The plurality of chips 21 are suction-held on the holding surface 4a via the tape 17, with the front surface 11a side thereof exposed upward.

[0034] Returning now to Fig. 3, a cylindrical pillar 6 is connected to the lower part of the chuck table 4. A driven pulley (not shown) is provided concentrically with the pillar 6. A rotation drive unit 8 including a drive pulley, a motor, an endless belt, etc. (none of which are shown) is disposed near the pillar 6.

[0035] An endless belt is wound around the driven pulley and the driving pulley. When the rotation drive unit 8 is operated, power is transmitted to the column 6 via the endless belt, and the chuck table 4 rotates around the column 6 as a rotation axis 6a. The rotation angle of the chuck table 4 can be either clockwise or counterclockwise when viewed from above, with no limit to the rotation angle.

[0036] An angle measurement unit 10 is provided on the underside of the column portion 6. The angle measurement unit 10 is also called a rotary encoder. The angle measurement unit 10 of this embodiment includes a disk-shaped scale (not shown) and an optical reader (not shown) that reads the graduations provided on the underside of the scale.

[0037] The scale is fixed concentrically to the column 6, and the graduations of the scale are formed at equal intervals along the circumferential direction of the scale. The reader is disposed on the underside of the scale and includes a light-emitting unit such as an LED that emits measurement light.

[0038] The reader further includes a light receiving unit such as a photosensor that receives light reflected by the graduations of the scale. The reader has a predetermined circuit that performs signal processing, etc. The predetermined circuit counts the number of reflected lights received by the light receiving unit, and transmits information on the rotation angle of the chuck table 4 as an electrical signal to a control unit 26, which will be described later.

[0039] Based on this electrical signal, the control unit 26 can determine the rotation angle of the chuck table 4. However, the angle measurement unit 10 is not limited to the reflective optical type described above. The angle measurement unit 10 may be of a transmissive optical type, a magnetic type, or an electromagnetic induction type, as long as it can acquire information about the rotation angle of the chuck table 4.

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

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

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

[0043] The first arm 16 is provided with a thickness measurement unit 18. The thickness measurement unit 18 is a reflection spectroscopic film thickness meter (also simply called a film thickness meter) that uses a microscope, and measures a minute area (measurement point P shown in FIG. 6) of the chip 21 by optical interferometry. A ) thickness can be measured.

[0044] In this embodiment, symbol P A is used for convenience to indicate one of the multiple measurement points on each chip 21, and does not indicate a specific measurement point. A Among these, one or more measurement points that are not used to calculate the average thickness are conveniently designated as measurement points P B It is written as follows.

[0045] The head portion 18a of the thickness measurement unit 18 is disposed so as to face the holding surface 4a in the Z-axis direction. The head portion 18a is provided with a condenser lens (not shown) whose optical axis is disposed along the Z-axis direction.

[0046] From head unit 18a, measurement light (e.g., infrared light) of a predetermined band that is guided from a light source such as an LED and that can transmit through chip 21 is irradiated onto each chip 21 that is sucked and held by holding surface 4a so that the light is collected near back surface 11b. Head unit 18a receives light reflected from back surface 11b of chip 21 and light reflected from front surface 11a of chip 21.

[0047] The interference light of the reflected light from the rear surface 11b and the front surface 11a is detected by an optical detection element (not shown) such as a photosensor provided in the thickness measurement unit 18. When the wavelength of the measurement light is plotted on the horizontal axis and the reflectance of the interference light is plotted on the vertical axis, the spectrum of the interference light changes depending on the optical path difference of the reflected light. Based on this spectrum, the reflectance of each measurement point P of each chip 21 is calculated. A The thickness at is measured.

[0048] The second arm 20 is provided with an imaging unit 22. The imaging unit 22 is an optical microscope, and captures an image including the division grooves 13b formed on the planned division lines 13 by capturing an image of the back surface 11b side of the chip 21.

[0049] The head portion 22a of the imaging unit 22 is disposed so as to face the holding surface 4a in the Z-axis direction. A condenser lens (not shown) whose optical axis is disposed along the Z-axis direction is provided inside the head portion 22a.

[0050] The imaging unit 22 has a light source section (not shown) such as an LED capable of emitting measurement light (e.g., white light), and the measurement light is irradiated from the head section 22a onto the chip 21 via a focusing lens so as to be focused near the back surface 11b.

[0051] The measurement light irradiated onto the chip 21 is reflected by the rear surface 11b or the like, and then guided to a line sensor (imaging element) (not shown) provided in the imaging unit 22 via a condenser lens.

[0052] A line sensor, also known as a linear image sensor, has multiple photodiodes, phototransistors, etc. arranged in a straight line, and constitutes a CCD (Charge-Coupled Device) image sensor or a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor.

[0053] The heads 18a and 22a are spaced apart enough so as not to affect the measurements. For example, the optical axes of the condenser lenses of the heads 18a and 22a are spaced apart by the radius of the wafer 11 or more.

[0054] In addition, a partition plate (not shown) that has been subjected to a predetermined surface treatment or has a light-absorbing film formed on its surface may be provided between the head portions 18a and 22a to prevent interference between the measurement light from the head portion 18a and the measurement light from the head portion 22a.

[0055] The rotation drive unit 8 and the Y-axis direction moving unit 12 function as a moving unit 24 that moves the thickness measurement unit 18 and the imaging unit 22 relative to the chuck table 4.

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

[0057] As a result, the thickness of each chip 21 is measured at a plurality of measurement points P A The chuck table 4, the angle measuring unit 10, the thickness measuring unit 18, the imaging unit 22, the moving unit 24, and the like are controlled by a control unit 26.

[0058] The control unit 26 of this embodiment is configured by a computer having a processor (processing device) represented by a CPU (Central Processing Unit), and storage devices including a main storage device and an auxiliary storage device.

[0059] The primary storage device includes a dynamic random access memory (DRAM), a static random access memory (SRAM), a read only memory (ROM), etc., and the secondary storage device includes a flash memory, a hard disk drive, a solid state drive, etc.

[0060] The auxiliary storage device stores software including a predetermined program. The functions of the control unit 26 are realized by operating the processing device and the like in accordance with this software. A part of the auxiliary storage device functions as a storage unit 28 that stores predetermined information obtained by measurements using the thickness measurement unit 18 and the imaging unit 22.

[0061] The storage unit 28 stores a plurality of measurement points P in a predetermined plane (for example, a plane corresponding to the back surface 11b of the wafer 11 before or after division) based on the notch 11c. A and the positions (i.e., coordinates) of the multiple measurement points P A The thickness data 28a of each chip 21 is stored, which is associated with the thickness at the chip 21.

[0062] At the start of thickness measurement, the positions of the moving block 14 and the chuck table 4 are adjusted so that the spot of the measurement light irradiated from the head portion 18a is positioned on the notch 11c. A The position of is automatically calculated by the control unit 26 based on the initial position of the spot, the rotation speed of the chuck table 4 and the movement speed of the moving block 14.

[0063] In this embodiment, the rotation speed of the chuck table 4 is 300 degrees / s, the movement speed of the moving block 14 is 9,308 μm / s, and the measurement points PA are arranged in a spiral shape on the rear surface 11b side (see FIG. 6).

[0064] 3, the head 22a of the imaging unit 22 is disposed a predetermined distance in the opposite direction (-Y direction) from the head 18a of the thickness measurement unit 18. Therefore, imaging of the back surface 11b starts after a predetermined time calculated by dividing the predetermined distance by the moving speed of the moving block 14.

[0065] The storage unit 28 also stores image data 28b captured by the imaging unit 22. The image data 28b is synthesized into a single image having an imaging range of a predetermined size (for example, approximately 335 mm square) by executing a first program stored in the auxiliary storage device on the processing device.

[0066] Furthermore, by performing image processing on the composite image using a second program stored in the auxiliary storage device, the position of the center line 13b1 in the width direction of the dividing groove 13b, which is perpendicular to the longitudinal direction of the dividing groove 13b, is identified for each of the multiple dividing grooves 13b.

[0067] For example, first, image processing such as edge detection is performed on the composite image to identify the coordinates of both edges of one dividing groove 13b in the width direction, and then the center position of the coordinates of both edges is calculated to identify the position of the center line 13b1 of one dividing groove 13b in the width direction.

[0068] The position of the center line 13b1 (see FIG. 7) of each dividing groove 13b is also stored in the storage unit 28. In this embodiment, the position of the center line 13b1 is also stored in the storage unit 28 at each measurement point P A Similarly to the position of the wafer 11, the position (ie, coordinates) is a position in a predetermined plane (for example, a plane corresponding to the back surface 11b of the wafer 11 after division) with the notch 11c as the reference.

[0069] For example, for one chip 21 shown in Figure 7, the position (X = c) of one center line 13b1 parallel to the Y-axis direction and the position (X = d) of another center line 13b1 parallel to the Y-axis direction are stored.

[0070] Furthermore, for that one chip 21, the position (Y=e) of one center line 13b1 parallel to the X-axis direction and the position (Y=f) of another center line 13b1 parallel to the X-axis direction are stored.

[0071] The auxiliary storage device stores a third program, which, when executed by the processing device, functions as a setting unit 30 that sets a thickness data non-reference region Q (see FIG. 10) on the outer periphery (surroundings) of each chip 21 based on the thickness data 28a of each chip 21.

[0072] The thickness data non-reference area Q is located at a predetermined distance R (see FIG. 9) or less from the center line 13b1 of one of the dividing grooves 13b, and is a measurement point P where the thickness data 28a is not referenced when calculating the thickness of the chip 21. B Includes.

[0073] The auxiliary storage device also stores a fourth program. The fourth program is executed by the processing device to measure the measurement points P A Measurement point P included in thickness data non-reference area Q B Multiple measurement points P excluding A The calculation unit 32 calculates the average thickness of the sample.

[0074] 4 to 10, a measurement method for measuring the thickness of a plurality of chips 21 using the measurement device 2 will be described. Fig. 4 is a flow chart of the measurement method.

[0075] First, as shown in Fig. 5, the front surfaces 11a of the plurality of chips 21 formed by dividing the wafer 11 are suction-held by the holding surface 4a via the tape 17 (holding step S10). Fig. 5 is a diagram showing the holding step S10.

[0076] In the holding step S10, the chip unit 23 is held by suction in a state in which an imaginary straight line W1 (see FIG. 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 positioned on the -Y direction side.

[0077] After the holding step S10, the thickness is measured by the thickness measuring unit 18, and the image of the dividing groove 13b from the rear surface 11b side is taken by the imaging unit 22 (measurement and imaging step S20).

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

[0079] In this embodiment, the rotation direction of the chuck table 4 is counterclockwise when viewed from above, and the movement direction of the moving block 14 is 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 step S20, the storage unit 28 stores the thickness data 28a (first storage step S30). A FIG.

[0081] Measurement point P in this embodiment A The number of measurement points P is approximately 800,000. In FIG. 6, for the sake of clarity, the number of measurement points P is shown as being significantly less than 800,000. A In the thickness measurement, the thickness is measured at each of measurement points P1, P2, P3, ... Pz (z is a natural number greater than or equal to 3, approximately 800,000 in this example) in order.

[0082] Therefore, if there are 968 rectangular areas on the wafer 11 where the devices 15 are formed, there are 968 measurement points P per chip 21. A The number of measurement points P per chip 21 is about 826 on average. A The average number is about 14.

[0083] As mentioned above, each measurement point P A The position of each center line 13b1 of each division groove 13b is defined in a predetermined plane (a plane corresponding to the back surface 11b of the wafer 11 before or after division) based on the notch 11c. After the measuring and imaging step S20, the memory unit 28 stores the position of the center line 13b1 of each division groove 13b obtained by processing the image data 28b (second storage step S40).

[0084] The first storage step S30 and the second storage step S40 may be performed in either order, or may be performed simultaneously. Fig. 7 is a diagram schematically showing the center lines 13b1 of the plurality of dividing grooves 13b to be stored in the second storage step S40 in one chip 21. For convenience of explanation, Fig. 7 shows the center lines 13b1 of the plurality of measuring points P A is omitted.

[0085] As described above, the position of each center line 13b1 is also defined within a predetermined plane (a plane corresponding to the back surface 11b of the wafer 11 before or after division) based on the notch 11c.

[0086] After the first storage step S30 and the second storage step S40, the setting unit 30 sets a thickness data non-reference region Q around each chip 21 based on the thickness data 28a of each chip 21 (setting step S50).

[0087] FIG. 8 shows a plurality of measurement points P A9 is a schematic diagram showing the center lines 13b1 (X=c), (X=d), (Y=e) and (Y=f) of the multiple dividing grooves 13b surrounding the chip 21, and FIG. 9 is a graph showing the thickness data non-reference area Q etc. set in the setting step S50.

[0088] The horizontal axis of FIG. 9 is the distance from one center line 13b1 (X=c) adjacent to the chip 21 in the −X direction shown in FIG. 8 to each measurement point P A The zero point on the horizontal axis corresponds to the position of the center line 13b1 (X=c). The vertical axis in FIG. 9 represents the distance (μm) from the measurement point P A is the thickness (μm) of the chip 21 at

[0089] In FIG. 9, in each of a plurality of unit areas divided in 10 μm increments from one center line 13 b 1, a measurement point P A The average thickness and standard deviation are calculated for each unit area. The line graph shown in Fig. 9 is drawn by connecting the average values ​​for each unit area.

[0090] The bars arranged parallel to the vertical axis in Fig. 9 represent the measurement points P A The standard deviation of the thickness is shown schematically, with (average value + σ) indicated at the top of the bar and (average value - σ) indicated at the bottom of the bar.

[0091] When creating the graph of FIG. 9, first, a plurality of measurement points P located in the range surrounded by four center lines 13b1 (X=c), (X=d), (Y=e), and (Y=f) are A The thickness data 28a is extracted.

[0092] Then, from the center line 13b1 (X=c), each measurement point P A The distance to each measurement point P A In data including thickness at and , outliers in thickness are removed using known processing such as M-estimation, which allows for robust data analysis.

[0093] As shown in FIG. 9, in the vicinity of the distance 0 μm (i.e., the center line 13 b 1), the measurement point PA The thickness variation in the vicinity of the center line 13b1 is considered to be a result of the thickness at the dividing line 13 where no insulating film or the like is formed.

[0094] In contrast, the mean value becomes substantially constant and the standard deviation (i.e., the length of the bars) becomes smaller as the distance from the center line 13b1 increases. Such changes in the mean value and standard deviation suggest that the thickness of the chip 21 becomes constant as the distance from the center of the back surface 11b of the chip 21 increases.

[0095] Therefore, in this embodiment, first, for each unit area in order from closest to the center line 13b1, the standard deviation and the moving average of the average thickness (i.e., the difference between the average values ​​in adjacent unit areas) are determined according to predetermined determination conditions.

[0096] The predetermined criteria in this embodiment are that the standard deviation is 0.5 μm or less and the moving average of the average thickness is 0.1 μm or less. The setting unit 30 determines that the distance R is the distance from the center line 13b1 to the unit area that first satisfies these criteria.

[0097] Then, the area farther than the distance R from the center line 13b1 as the starting point is designated as the measurement point P used when calculating the thickness of the chip 21. A In other words, the setting unit 30 sets the area from the center line 13b1 to the distance R as the measurement points P that are not used to calculate the average thickness. B is set to the thickness data non-reference area Q where

[0098] 9, the area from one center line 13b1 to a position 500 μm away does not satisfy the above-mentioned criteria, and is therefore a thickness data non-reference area Q. In contrast, the area in the +X direction beyond the 500 μm position is a valid area T.

[0099] 8 and 9, the thickness data non-reference area Q and the effective area T are described with reference to one center line 13b1 (X=c) adjacent to one chip 21 in the −X direction.

[0100] However, in the setting step S50, in accordance with the above-mentioned judgment conditions, the distance R1 is further calculated based on the center line 13b1 (X=d), the distance R2 is calculated based on the center line 13b1 (Y=e), and the distance R3 is calculated based on the center line 13b1 (Y=f).

[0101] In this way, a thickness data non-reference area Q is set in a rectangular annular area including an area from the center line 13b1 (X=c) to a distance R in the +X direction, an area from the center line 13b1 (X=d) to a distance R1 in the -X direction, an area from the center line 13b1 (Y=e) to a distance R2 in the -Y direction, and an area from the center line 13b1 (Y=f) to a distance R3 in the +Y direction.

[0102] Then, the area surrounded by the thickness data non-reference area Q is set as the valid area T. In this way, the setting unit 30 extracts the corresponding thickness data 28a, removes abnormal thickness values, calculates the distance R etc. according to the judgment conditions, and sets the thickness data non-reference area Q.

[0103] The setting unit 30 similarly sets a thickness data non-reference area Q (that is, a valid area T) for each of the plurality of chips 21 on the wafer 11.

[0104] FIG. 10 shows a thickness data non-reference area Q in the setting step S50 and a plurality of measurement points P located in the valid area T and to be calculated in the calculation step S60. A FIG.

[0105] After the setting step S50, the calculation unit 32 calculates the measurement points P A Measurement point P included in thickness data non-reference area Q B Multiple measurement points P excluding A The average thickness is calculated (calculation step S60).

[0106] In the calculation step S60, for one chip 21, the measurement point P included in the thickness data non-reference area Q is calculated. BMultiple measurement points P excluding A Thereafter, in processes such as picking up the chips 21, each average value is treated as the thickness of the corresponding chip 21.

[0107] In this way, the calculation unit 32 calculates the distance between the plurality of measurement points P A The calculation unit 32 calculates the average thickness for each of the plurality of chips 21 on the wafer 11 in the same manner.

[0108] In this embodiment, the thickness of each chip 21 can be calculated without including the thickness near the planned division line 13, so the influence of the thickness of the outer periphery of each chip 21 can be reduced compared to when the thickness at the planned division line 13 is considered as the thickness of the chip 21, and the thickness of each chip 21 can be measured more accurately.

[0109] In addition, the structures, methods, etc. according to the above-described embodiments can be modified as appropriate without departing from the scope of the present invention. For example, if an orientation flat is formed instead of the notch 11c, the center position of the length of the straight line that constitutes the orientation flat will perform the same function as the notch 11c.

[0110] The average thickness (i.e., thickness) of each of the plurality of chips 21 may be displayed to the worker via a display device (not shown) along with the overall view of the wafer 11 shown in FIG. 6. Also, chips 21 whose average thickness is outside a predetermined tolerance may be displayed to the worker in a conspicuous manner. This allows the worker to easily distinguish between good and bad chips 21. [Explanation of symbols]

[0111] 2: measuring device, 4: chuck table (holding table), 4a: holding surface, 4b: frame 4c: recess, 4d: porous plate, 4e: groove, 4f: hole, 6: column, 6a: rotation axis 8: Rotation drive unit, 10: Angle measurement unit 12: Y-axis direction moving unit, 14: moving block 11: wafer, 11a: front surface, 11b: back surface, 11c: notch W0: Center position, W1: Straight line 13: planned division line, 13a: width, 13b: division groove, 13b1: center line 15: Device, 17: Tape (protective material), 21: Chip, 23: Chip unit 16: first arm, 18: thickness measurement unit, 18a: head section 20: second arm, 22: imaging unit, 22a: head section 24: Mobile unit, 26: Control unit 28: Storage unit, 28a: Thickness data, 28b: Image data 30: Setting unit, 32: Calculation unit P1,P2,P3,P A ,P B ,P Z : Measurement point Q: Thickness data non-reference area, R, R1, R2, R3: distance, T: valid area S10: Retention step, S20: Measurement and imaging step, S30: First storage step S40: Second storage step, S50: Setting step, S60: Calculation step

Claims

1. A method for measuring thicknesses of a plurality of chips, comprising: a wafer having devices formed in each of areas partitioned in a grid pattern by a plurality of planned dividing lines set on a surface, each having a predetermined width, the wafer being divided into a plurality of chips by forming dividing grooves within each planned dividing line; The measurement method is a holding step of holding the front surfaces of the plurality of chips on a holding surface of a holding table via a protective member; a first storage step of measuring the thickness of each of the plurality of chips held on the holding surface at a plurality of measurement points and storing thickness data for each chip in which the positions of the plurality of measurement points in a predetermined plane corresponding to the back surface of the wafer are associated with the thickness at the plurality of measurement points; a second storage step of capturing an image of the backside of the plurality of chips held by the holding surface and storing the position of the center line of each division groove in the predetermined plane; a setting step of setting thickness data non-reference areas around each chip based on the thickness data of each chip, the thickness data non-reference areas including measurement points that are located at a predetermined distance or less from the center line of each dividing groove and whose thickness data is not referenced when calculating the thickness of each chip; a calculation step of calculating an average thickness value at each of the plurality of measurement points, excluding measurement points included in the thickness data non-reference area from the plurality of measurement points, for each of the plurality of chips; A measuring method comprising:

2. A measuring device for measuring thicknesses of a plurality of chips, comprising: a wafer having devices formed in each of areas partitioned in a grid pattern by a plurality of planned dividing lines set on a surface, each having a predetermined width, the wafer being divided into a plurality of chips by forming dividing grooves within each planned dividing line; The measuring device comprises: a holding table having a holding surface capable of holding the front surfaces of the plurality of chips via a protective member; a thickness measurement unit that measures the thickness of each of the plurality of chips held on the holding surface at a plurality of measurement points; an imaging unit that images the backside of the plurality of chips held on the holding surface; a moving unit that moves the thickness measuring unit and the imaging unit relative to the holding table; a control unit for controlling at least the holding table, the thickness measuring unit, the imaging unit, and the moving unit; The control unit a storage unit that stores thickness data for each chip, in which the positions of the plurality of measurement points in a predetermined plane corresponding to the back surface of the wafer are associated with the thicknesses at the plurality of measurement points, and the positions of the center lines of the division grooves imaged by the imaging unit in the predetermined plane; a setting unit that sets thickness data non-reference areas around each chip based on the thickness data, the thickness data non-reference areas including measurement points that are located at a predetermined distance or less from the center line of the dividing groove and whose thickness data is not referenced when calculating the thickness of each chip; a calculation unit that calculates an average thickness value at each of the plurality of measurement points, excluding measurement points included in the thickness data non-reference area from the plurality of measurement points, for each of the plurality of chips; A measuring device comprising:

Images