Die Bonding Apparatus and Method of Manufacturing Semiconductor Device

Abstract

To provide a technique capable of automating the registration work of a template matching model.SOLUTION: A die bonding device includes a control unit that controls an imaging device. The control unit images a wafer with the imaging device, detects a dicing groove, obtains the center coordinates of the die on the basis of the detected dicing groove, roughly positions the center position of the die, moves the center position of the die to the center of the optical axis of the imaging device on the basis of the center coordinates, images the die moved by the imaging device, detects the position of the pad held by the die on the basis of the image of the imaged die, detects a unique pad row with a unique pitch on the basis of the detected pad arrangement, and registers the detected unique pad row as a template model.SELECTED DRAWING: Figure 10

Description

Technical Field

[0001] The present disclosure relates to a die bonding apparatus and is applicable to a die bonder that positions by, for example, template matching.

Background Art

[0002] A die bonding apparatus such as a die bonder, which is one of semiconductor manufacturing apparatuses, is an apparatus that bonds (mounts and adheres) a semiconductor chip (hereinafter simply referred to as a die) onto a wiring board, a lead frame, etc. (hereinafter simply referred to as a substrate) or an already bonded die. In a die bonding apparatus, a die is vacuum-sucked by a pickup head or a bonding head, lifted at high speed, horizontally moved, lowered, and placed on an intermediate stage or a substrate.

[0003] When a die is vacuum-sucked by a pickup head or a bonding head, it is necessary to surely pick up the die. With the recent thinning of dies, the requirement is high. Therefore, the position of the die is recognized to detect the deviation of the die, and then the position of the die or the bonding head is corrected to pick up the die.

[0004] As a method for recognizing the position of a die, for example, there is a method of recognizing the position of a die by performing pattern matching (template matching) between imaging data at unique portions such as die alignment marks and pads for wire bonding and a template obtained in a mimicking operation in advance.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] However, in Patent Document 1, the unique part is selected by the operator of the die bonder using a GUI or the like. The determination of the registration area for the template matching model (unique part) is made by human judgment.

[0007] An object of the present disclosure is to provide a technology capable of automating the registration work of a template matching model.

Means for Solving the Problems

[0008] The outline of typical ones among the present disclosures will be briefly described as follows. That is, the die bonding apparatus includes a control unit that controls an imaging device. The control unit controls the imaging device to image a wafer, detects dicing grooves, obtains the center coordinates of the die based on the detected dicing grooves, roughly positions the center position of the die, moves the center of the die to the optical axis center of the imaging device based on the center coordinates, images the moved die by the imaging device, detects the positions of the pads of the die based on the image of the imaged die, detects a unique pad row having a unique pitch based on the detected arrangement of the pads, and is configured to register the detected unique pad row as a template model.

Effects of the Invention

[0009] According to the above die bonding apparatus, it is possible to automate the registration work of the template matching model.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Figure 24

Figure 25

Embodiments for Carrying Out the Invention

[0011] Hereinafter, examples and modification examples will be described with reference to the drawings. However, in the following description, the same reference numerals may be given to the same components and repeated descriptions may be omitted. Note that the drawings may schematically represent the width, thickness, shape, etc. of each part compared to the actual aspect for clearer explanation, but this is merely an example and does not limit the interpretation of the present disclosure.

Examples

[0012] FIG. 1 is a schematic top view showing the configuration of a die bonder in an example. FIG. 2 is a diagram for explaining the schematic configuration when viewed from the direction of arrow A in FIG. 1.

[0013] The die bonder 10 is roughly composed of a die supply unit 1 that supplies a die D to be mounted on a substrate S, a pickup unit 2, an intermediate stage unit 3, a bonding unit 4, a transport unit 5, a substrate supply unit 6, a substrate discharge unit 7, and a control unit 8 that monitors and controls the operations of each unit. The Y-axis direction is the front-rear direction of the die bonder 10, and the X-axis direction is the left-right direction. The die supply unit 1 is arranged on the front side of the die bonder 10, and the bonding unit 4 is arranged on the back side. Here, one or a plurality of product areas (hereinafter referred to as package area P), which will be the final one package, are printed on the substrate S.

[0014] First, the die supply unit 1 supplies the die D to be mounted on the package area P of the substrate S. The die supply unit 1 includes a wafer holding stage 12 that holds the wafer 11, and a pushing-up unit 13 indicated by a dotted line that pushes up the die D from the wafer 11. The die supply unit 1 is moved in the XY-axis directions by a driving means (not shown), and moves the die D to be picked up to the position of the pushing-up unit 13. Here, the wafer 11 is divided into a plurality of dies D, and the die D is held in the outer shape of the wafer by a dicing tape 16 described later.

[0015] The pickup unit 2 includes a pickup head 21 that picks up the die D, a Y driving unit 23 of the pickup head that moves the pickup head 21 in the Y-axis direction, and driving units (not shown) that move the collet 22 up and down, rotate it, and move it in the X-axis direction. The pickup head 21 has a collet 22 (see also FIG. 2) that sucks and holds the pushed-up die D at its tip, picks up the die D from the die supply unit 1, and places it on the intermediate stage 31. The pickup head 21 has driving units (not shown) that move the collet 22 up and down, rotate it, and move it in the X-axis direction.

[0016] The intermediate stage unit 3 includes an intermediate stage 31 that temporarily places the die D, and a stage recognition camera 32 for recognizing the die D on the intermediate stage 31.

[0017] The bonding unit 4 picks up the die D from the intermediate stage 31 and bonds it onto the package area P of the substrate S being conveyed, or bonds it in a form of laminating onto the die that has already been bonded onto the package area P of the substrate S. The bonding unit 4 includes a bonding head 41 having a collet 42 (see also FIG. 2) that adsorbs and holds the die D at its tip in the same manner as the pickup head 21, a Y drive unit 43 that moves the bonding head 41 in the Y-axis direction, a substrate recognition camera 44 that images a position recognition mark (not shown) of the package area P of the substrate S and recognizes the bonding position, and an XY drive unit 45 that drives the substrate recognition camera 44 in the X-axis direction and the Y-axis direction. With such a configuration, the bonding head 41 corrects the pickup position and posture based on the imaging data of the stage recognition camera 32, picks up the die D from the intermediate stage 31, and bonds the die D onto the substrate S based on the imaging data of the substrate recognition camera 44.

[0018] The transfer unit 5 has a substrate transfer claw 51 that grips and transfers the substrate S, and a transfer lane 52 along which the substrate S moves. The substrate S moves by being driven by a ball screw (not shown) provided along the transfer lane 52 with a nut (not shown) of the substrate transfer claw 51 provided on the transfer lane 52. With such a configuration, the substrate S moves from the substrate supply unit 6 along the transfer lane 52 to the bonding position, and after bonding, moves to the substrate unloading unit 7 and delivers the substrate S to the substrate unloading unit 7.

[0019] Next, the configuration of the die supply unit 1 will be described with reference to FIG. 3. FIG. 3 is a schematic cross-sectional view showing the main part of the die supply unit shown in FIG. 1.

[0020] The die supply unit 1 includes a wafer holding stage 12 that moves in the horizontal direction (XY-axis direction), and a pushing-up unit 13 that moves in the vertical direction. The wafer holding stage 12 has an expand ring 15 that holds the wafer ring 14, and a support ring 17 that horizontally positions a dicing tape 16 to which a plurality of dies D are adhered and held by the wafer ring 14. The pushing-up unit 13 is disposed inside the support ring 17.

[0021] When the die supply unit 1 raises the die D, it lowers the expand ring 15 that holds the wafer ring 14. As a result, the dicing tape 16 held by the wafer ring 14 is stretched, the interval between the dies D widens, the push-up unit 13 pushes up the die D from below the die D, and the pick-up property of the die D is improved. A film-like adhesive material called a die attach film (DAF) 18 is attached between the wafer 11 and the dicing tape 16. In the wafer 11 having the die attach film 18, dicing is performed on the wafer 11 and the die attach film 18. Therefore, in the peeling process, the wafer 11 and the die attach film 18 are peeled from the dicing tape 16.

[0022] The die bonder 10 includes a wafer recognition camera 24 as an imaging device that recognizes the posture and position of the die D on the wafer 11, a stage recognition camera 32 as a second imaging device that recognizes the posture and position of the die D placed on the intermediate stage 31, and a substrate recognition camera 44 as a third imaging device that recognizes the mounting position on the bonding stage BS. It is necessary to correct the posture deviation between the recognition cameras for the stage recognition camera 32 involved in the pick-up by the bonding head 41 and the substrate recognition camera 44 involved in the bonding to the mounting position by the bonding head 41. In this embodiment, the positioning and surface inspection of the die D are performed using the coaxial illumination 26 (see FIG. 15) and the like together with the wafer recognition camera 24, the stage recognition camera 32, and the substrate recognition camera 44.

[0023] Next, the control unit 8 will be described with reference to FIG. 4. FIG. 4 is a block diagram showing the schematic configuration of the control system of the die bonder shown in FIG. 1.

[0024] The control system 80 includes a control unit 8, a drive unit 86, a signal unit 87, and an optical system 88. The control unit 8 mainly includes a control and arithmetic unit 81 mainly composed of a CPU (Central Processor Unit), a storage device 82, an input / output device 83, a bus line 84, and a power supply unit 85. The storage device 82 includes a main storage device 82a composed of a RAM that stores processing programs and the like, and an auxiliary storage device 82b composed of an HDD, an SSD, etc. that store control data, image data, etc. necessary for control. The input / output device 83 includes a monitor 83a that displays the device state, information, etc., a touch panel 83b for inputting the operator's instructions, a mouse 83c for operating the monitor, and an image capture device 83d for capturing image data from the optical system 88. Further, the input / output device 83 includes a motor control device 83e that controls the drive unit 86 such as the XY table (not shown) of the die supply unit 1, the ZY drive axis of the bonding head table, and the XY drive axis of the substrate recognition camera, and an I / O signal control device 83f that captures or controls signals from the signal unit 87 such as various sensor signals and switches such as lighting devices. The optical system 88 includes a wafer recognition camera 24, a stage recognition camera 32, and a substrate recognition camera 44. The control and arithmetic unit 81 fetches necessary data via the bus line 84, performs calculations, controls the pickup head 21, etc., and sends information to the monitor 83a, etc.

[0025] The control unit 8 stores the image data captured by the wafer recognition camera 24, the stage recognition camera 32, and the substrate recognition camera 44 via the image capture device 83d in the storage device 82. Based on the stored image data, using the control and arithmetic unit 81 with software programmed accordingly, the positioning of the package area P of the die D and the substrate S, as well as the surface inspection of the die D and the substrate S, are performed. Based on the positions of the package area P of the die D and the substrate S calculated by the control and arithmetic unit 81, the driving unit 86 is moved via the motor control device 83e by software. Through this process, the positioning of the die on the wafer is performed, and the die D is bonded onto the package area P of the substrate S by operating the driving units of the pickup unit 2 and the bonding unit 4. The wafer recognition camera 24, the stage recognition camera 32, and the substrate recognition camera 44 used are grayscale, color, etc., and the light intensity is digitized.

[0026] Next, the die bonding process (manufacturing method of semiconductor device) will be described with reference to FIG. 5. FIG. 5 is a flowchart for explaining the die bonding process in the die bonder shown in FIG. 1.

[0027] In the die bonding process of the embodiment, first, the wafer ring 14 holding the dicing tape 16 to which the die D divided from the wafer 11 is attached is stored in a wafer cassette (not shown) and carried into the die bonder 10. Also, the substrate S is prepared and carried into the die bonder 10.

[0028] (Wafer loading: Step P1) The control unit 8 takes out the wafer ring 14 holding the wafer 11 from the wafer cassette and places it on the wafer holding stage 12, and transports the wafer holding stage 12 to the reference position where the pickup of the die D is performed.

[0029] (Die transfer: Step P2) Next, the control unit 8 moves the wafer holding stage 12 on which the wafer 11 is placed in a pitch movement at a predetermined pitch and holds it horizontally, thereby placing the die D picked up first at the pick-up position. Note that the pick-up position of the die D is also the recognition position of the die D by the wafer recognition camera 24. The wafer 11 is inspected die by die in advance by an inspection apparatus such as a prober, map data indicating good or defective for each die is generated, and stored in the storage device 82 of the control unit 8. The determination as to whether the die D to be picked up is a good product or a defective product is made based on the map data or a defective mark on the die surface. When the die D is a defective product, the control unit 8 moves the wafer holding stage 12 on which the wafer 11 is placed in a pitch movement at a predetermined pitch, places the next die D to be picked up at the pick-up position, and skips the defective die D.

[0030] (Die positioning: Step P3) Next, the control unit 8 photographs the main surface (upper surface) of the die D to be picked up by the wafer recognition camera 24 and acquires an image. Although details will be described later, the control unit 8 calculates the amount of displacement of the die D to be picked up from the acquired image and measures the position. Based on this amount of displacement, the control unit 8 moves the wafer holding stage 12 on which the wafer 11 is placed, and accurately positions the die D to be picked up at the pick-up position.

[0031] (Die surface inspection: Step P4) Next, the control unit 8 performs a surface inspection of the die D to be picked up using the image acquired by the wafer recognition camera 24. Here, when the control unit 8 determines that there is no problem on the surface of the die D, after the die crack inspection described later, it proceeds to the next step (Step P9 described later). However, when it determines that there is a problem, it performs a skip process or an error stop. The skip process skips Steps P9 and later for the die D and moves to Step P2.

[0032] (Substrate loading: Step P5, Substrate transfer: Step P6) The control unit 8 places the substrate S on the transfer lane 52 by the substrate supply unit 6. The control unit 8 moves the substrate transfer claws 51 that grip and transfer the substrate S to the bonding position.

[0033] (Substrate positioning: Step P7) Next, the control unit 8 moves the substrate recognition camera 44 to the imaging position (bonding tab imaging position) of the package area P to be bonded. The control unit 8 photographs the substrate S with the substrate recognition camera 44 and acquires an image. The control unit 8 calculates the amount of misalignment of the package area P of the substrate S from the acquired image and measures the position. Based on this amount of misalignment, the control unit 8 moves the substrate S and performs positioning or adjustment of the bonding position to accurately place the package area P to be bonded at the bonding position.

[0034] (Substrate surface inspection: Step P8) Next, the control unit 8 performs a surface inspection of the package area P of the substrate S using the image acquired by the substrate recognition camera 44. Here, the control unit 8 determines whether there is a problem in the surface inspection. If it is determined that there is no problem on the surface of the package area P of the substrate S, it proceeds to the next step (Step P9 described later). However, if it is determined that there is a problem, it visually checks the surface image or performs an inspection with changed high-sensitivity inspection or lighting conditions, etc. If there is a problem, it performs skip processing, and if there is no problem, it performs the processing of the next step. The skip processing skips the processing after Step P13 for the corresponding tab of the package area P of the substrate S and registers a defect in the substrate start information.

[0035] (Die handling: Step P9, Intermediate stage placement: Step P10) After the control unit 8 accurately places the die D to be picked up by the die supply unit 1 at the pick-up position, the die D is picked up from the dicing tape 16 by the pick-up head 21 including the collet 22 and placed on the intermediate stage 31.

[0036] (Die position inspection: Step P11) The control unit 8 detects the positional deviation (rotational deviation) of the die placed on the intermediate stage 31 by imaging with the stage recognition camera 32. The control unit 8 photographs the main surface (upper surface) of the die D with the stage recognition camera 32 and acquires an image. Although details will be described later, the control unit 8 calculates the amount of positional deviation of the die D from the acquired image and measures the position. When there is a positional deviation, the control unit 8 corrects the positional deviation by rotating the intermediate stage 31 about a plane parallel to the mounting surface having the mounting position by a turning drive device (not shown) provided on the intermediate stage 31.

[0037] (Die surface inspection: Process P12) The control unit 8 performs a surface inspection of the die D using the image acquired by the stage recognition camera 32. Here, when the control unit 8 determines that there is no problem on the surface of the die D, after the die crack inspection, it proceeds to the next process (process P13 described later). However, when it determines that there is a problem, it performs a skip process or an error stop. The skip process places the die on a defective product tray (not shown) or the like, skips the processes after process P13 for the die D, and moves to process P2.

[0038] (Die attach: Process P13) The control unit 8 picks up the die D from the intermediate stage 31 with the bonding head 41 including the collet 42 and performs die bonding on the package area P of the substrate S or on a die that has already been bonded to the package area P of the substrate S.

[0039] (Relative position inspection of die and substrate: Process P14) Next, the control unit 8 photographs the die D with the substrate recognition camera 44 and acquires an image. Although details will be described later, the control unit 8 calculates the amount of positional deviation of the die D from the acquired image and measures the position. After bonding the die D, the control unit 8 inspects whether the bonding position is accurate. At this time, similar to the die alignment, the center of the die and the center of the tab are obtained, and it is inspected whether the relative position is correct.

[0040] (Surface inspection of die and substrate: Process P15) The control unit 8 performs a surface inspection of the die D using the image acquired by the substrate recognition camera 44. Here, if the control unit 8 determines that there is no problem on the surface of the die D, it proceeds to the next process (process P2). However, if it determines that there is a problem, it performs a skip process or an error stop. In the skip process or error stop, a defect is registered in the substrate start information and the process moves to P2.

[0041] (Substrate transfer: process P16, substrate unloading: process P17) Thereafter, the die D is bonded one by one to the package area P of the substrate S according to the same procedure. When the bonding of one substrate is completed, the substrate S is moved to the substrate unloading section 7 by the substrate transfer claws 51 and the substrate S is handed over to the substrate unloading section 7.

[0042] (Wafer unloading: process P18) Thereafter, the die D is peeled off from the dicing tape 16 one by one according to the same procedure (process P9). When the pickup of all the dies D in the start category is completed, the dicing tape 16 and the wafer ring 14 etc. that held the dies D in the outer shape of the wafer 11 are unloaded into the wafer cassette.

[0043] As described above, the die D is mounted on the substrate S via the die attach film 18 and is carried out from the die bonder. Thereafter, it is electrically connected to the electrodes of the substrate S via wires such as Au in the wire bonding process. In the case of a stacked product, subsequently, the substrate S on which the die D is mounted is reloaded into the die bonder and the second die D is stacked on the die D mounted on the substrate S via the die attach film 18. After being carried out from the die bonder, it is electrically connected to the electrodes of the substrate S via an Au wire in the wire bonding process. The second die D is peeled off from the dicing tape 16 by the method described above and then transported to the pelletizing process and stacked on the die D. After the above process is repeated a predetermined number of times, the substrate S is transported to the molding process and the stacked package is completed by sealing the plurality of dies D and the Au wires with a molding resin (not shown).

[0044] The method of die positioning will be described with reference to FIGS. 6 to 9. FIG. 6 is a flowchart for explaining the imitation operation. FIG. 7 is a diagram showing an example of a unique part (selection area). FIG. 8 is a diagram showing examples of registered images and similar images. FIG. 9 is a flowchart for explaining the continuous bonding operation.

[0045] The die positioning algorithm mainly uses pattern matching using two or more template models (search algorithm using template matching), and performs calculations using the generally known normalized correlation formula. The result is taken as the matching rate. Template matching has imitation operations and continuous bonding operations of reference learning.

[0046] First, the imitation operation will be described with reference to FIG. 6. The imitation operation is an operation performed in advance prior to continuous bonding in the die bonding process shown in FIG. 5.

[0047] (Step S1) The control unit 8 transports the reference wafer to the pickup position in the same manner as in process P2.

[0048] (Step S2) The control unit 8 images the reference wafer by the wafer recognition camera 24 to acquire an image.

[0049] (Step S3) Although details will be described later, the control unit 8 performs rough positioning of the die using the dicing groove of the reference wafer, that is, the contour of the die, based on the image acquired in step S2. Then, the detected position of the die is moved to the pickup point, that is, the center of the field of view (optical axis center) of the wafer recognition camera 24.

[0050] (Step S4) The control unit 8 images the reference wafer again by the wafer recognition camera 24 to acquire the image PCr shown in FIG. 7.

[0051] (Step S5) The control unit 8 selects, from within the image PCr, a unique portion UA as shown in FIG. 7 as a template model by the method described later.

[0052] (Step S6) The control unit 8 stores the positional relationship (coordinates) between the selected unique portion (selection area) UA and the reference wafer in the storage device 82.

[0053] (Step S7) The control unit 8 stores the image of the selection area (template image) PT, the reference work image, and their coordinates in the storage device 82.

[0054] Next, the continuous bonding operation in the die bonding process shown in FIG. 5 will be described with reference to FIG. 9.

[0055] (Step S11) As described in the process P2 shown in FIG. 5, the control unit 8 conveys the product wafer as a member to the pickup position for continuous bonding. In the die position inspection at the intermediate stage 31, as described in the process P10 shown in FIG. 5, the die D as a member for continuous bonding is placed on the intermediate stage 31. In the relative position inspection between the die and the substrate, as described in the process P13 shown in FIG. 5, the die D as a member for continuous bonding is die-bonded to the substrate S.

[0056] The die positioning in the process P3 shown in FIG. 5, the die position inspection in the process P11, and the relative position inspection between the die and the substrate in the process P14 are performed by the following steps S12 to S14.

[0057] (Step S12) In process P3, the control unit 8 captures an image of the product wafer using the wafer recognition camera 24 and obtains the image PCn shown in FIG. 8. In process P11, the control unit 8 captures an image of the die D on the intermediate stage 31 using the stage recognition camera 32 and obtains the image PCn shown in FIG. 8. In process P14, the control unit 8 captures an image of the die D on the substrate S using the substrate recognition camera 44 and obtains the image PCn shown in FIG. 8.

[0058] (Step S13) As shown in FIG. 8, the control unit 8 compares the template image PT saved in the mimicking operation with the acquired image PCn of the product die, and calculates the coordinates of the image PTn of the most similar part.

[0059] (Step S14) The control unit 8 compares the coordinates of the image PTn with the coordinates measured using the reference wafer, and calculates the position of the product die (the offset between the image PTn and the template image PT).

[0060] Note that between step S11 and step S12, rough positioning similar to steps S2 and S3 in the mimicking operation may be performed.

[0061] Next, the automatic registration sequence of the template model shown in steps S2 to S5 in FIG. 6 will be described with reference to FIGS. 10 and 11. FIG. 10 is a flowchart for explaining the automatic registration sequence of the template model. FIG. 11 is a diagram showing a rough positioning method using a dicing groove (dicing line).

[0062] This embodiment is premised on the fact that the accuracy of the dicing groove (the accuracy of the die position obtained from the accuracy of the dicing groove) is worse than the pad pitch. The basic idea of this embodiment is: (A) Detect the dicing groove to roughly position the die (step S21), (B) Align the die with the optical axis based on the roughly positioned position (step S22), (C) Automatically detect the pad position based on the roughly positioned position (step S23), (D) Automatically detect a unique pad arrangement based on the detected pad position, and register the unique pad arrangement area as a unique part (selection area) (step S24).

[0063] (Step S21: Rough positioning of the die) The control unit 8 images the wafer by the wafer recognition camera 24 (step S2), and detects the dicing groove of the wafer, that is, the contour of the die. Using the detected dicing groove, rough positioning of the die is performed (step S3).

[0064] As shown in FIG. 11, the control unit 8 detects the dicing grooves DLx, DLy surrounding the target die Dm from the imaging data obtained from the wafer recognition camera 24, and roughly positions the center position Dc of the die Dm. Here, let the coordinates of the center position Dc be (xd, yd).

[0065] (Step S22: Alignment of the die with the optical axis) The control unit 8 moves the die Dm to the pickup point, that is, the center of the field of view (center of the optical axis), based on the position of the die detected in step S21 (step S3). For example, as shown in FIG. 11, when there is a deviation between the center position Dc of the die Dm and the center OAC of the optical axis indicated by a cross, the control unit 8 moves the die Dm in the direction of the arrow so that the center position Dc of the die Dm coincides with the center OAC of the optical axis.

[0066] (Step S23: Automatic detection of the pad position) The control unit 8 images the die using the wafer recognition camera 24 and detects the pad positions provided on the die from the captured image. The pad positions are obtained, for example, from the area of the blob, aspect ratio, shape, position relative to the edge of the die, arrangement between pads, etc.

[0067] The area of the template model must have a unique pattern compared to other areas in order to prevent misdetection. Since the pad portion formed on the surface of the die D can be the pattern with the best positional accuracy, in this embodiment, the pads are automatically detected and registered as a model. Note that imaging the entire die is essential for die surface inspection and the like, and it is difficult to simply determine whether the pads in the image at that magnification are square or not. Also, the same dies are arranged on the wafer, and even if a portion with a unique pattern is selected for a single die, there may be a similar pattern on the adjacent die, resulting in a risk of misdetection.

[0068] A method for automatically detecting the pad positions will be described with reference to FIG. 12. FIG. 12(a) is a diagram showing randomly arranged pads, FIG. 12(b) is a diagram showing aligned pads, FIG. 12(c) is a diagram showing pads arranged on four sides, and FIG. 12(d) is a diagram showing an image of the pads and the wiring connected to the pads.

[0069] (C-1) Template matching Create a model image of the average shape of the pads, perform template matching based on it, and judge by score and coordinates.

[0070] (C-1-1) As an example, for each wafer type, the wafer recognition camera 24 captures images to collect pad images. The images of these pads are averaged and synthesized to obtain an average pad image. This is effective when the variation among wafers of the same type is small. Synthesizing the actual images is to consider the influence of probe marks. Here, a probe mark is a mark formed when a probe needle contacts a pad in a wafer test to confirm that the device function can be correctly realized on the wafer at the end of the pre-process of the device manufacturing process. The average pad image is registered in the template model to detect the pad. The determination of whether it is a pad is made by setting a threshold for the score. If a location above the threshold is found, that location is taken as the candidate position of the pad, and its coordinates are obtained.

[0071] (Determination after candidate position detection) Die D is square-shaped in plan view. Except for products with a center pad where the pad is arranged near the center of the die, pad Dp is arranged around die D. Also, pad Dp is arranged in the vicinity of at least one side of the four sides forming the square.

[0072] As shown in Fig. 12(a), pad Dp does not arrange randomly. As shown in Fig. 12(b), in most cases, even if the pitches are different, pad Dp arranges in one or two rows along the contour of the die. Here, in Fig. 12(b), pad Dp is arranged on the upper side in the drawing of the contour of die D extending along the X-axis direction.

[0073] Therefore, based on the acquired coordinates, select a group of candidate positions of the pads that have no deviation in the vertical (Y-axis direction) or horizontal (X-axis direction) in the group of candidate positions of the pads. Also, select candidate positions of the pads that are within a predetermined range from the edge of the die D. For example, when the die size is small, such as 1 to 5 mm, one row of candidate positions within 10 to 20% of the die size from the edge of the die D is set as the pad position. When the die size is larger than 5 mm, one or two rows of candidate positions within a predetermined distance (L) from the edge of the die D are set as the pad position. Here, L is, for example, 100 to 400 μm. In this embodiment, as shown in FIG. 12(c), the pads Dp are arranged along the four sides of the die D, assuming up to two rows on each side.

[0074] (C-1-2) As another example, for each type of wafer, image is captured by the wafer recognition camera 24 and a plurality of representative pads are pre-registered as models in multiple types. Search is performed for each model, and the result with a large number of candidates is used. This is effective when there is a large variation among wafers of the same type. Representing the actual image is to consider the influence of the probe marks. Set the candidate positions of the pads with a large number of candidates and acquire their coordinates. The same processing as (C-1-1) is used for the determination after the candidate positions are detected.

[0075] (C-2) Binarization Perform binarization processing to convert a grayscale image to 0 and 1 based on an arbitrary threshold value, and determine from the area, aspect ratio, and coordinates of the floating blobs by blob analysis. Binarization is performed using a discriminant analysis method or the like within the range of the pad candidate position conditions (within the die edge region) described in (Candidate position detection after determination) of (C-1-1). Here, the discriminant analysis method is a method of determining a threshold value so that pixels belonging to a pixel class close to black (a class with a density value close to 0) and pixels belonging to a pixel class close to white (a class with a density value close to the maximum value) are separated and distributed on both sides of a certain threshold value, and the separation degree of the distribution is increased. When the candidate position is not detected, the threshold value is variably changed by about 10% of the average brightness of the measurement region and repeated until the candidate position is found. The pad size is set to 20 to 150 μm, and those that are inclined obliquely are excluded. The aspect ratio is from a square to 1:2. Here, the direction along the direction in which the ends of adjacent dies extend is 1, and the direction along the direction intersecting the ends of the die is 2.

[0076] As shown in FIG. 12(d), since the wiring WL connected to the pad Dp may be reflected, perform a process of erasing the wiring WL by methods such as morphological opening (performing dilation, which is a process of expanding by 1 pixel, N times after performing contraction, which is a process of shrinking by 1 pixel, N times), closing, dilation, erosion, etc., and then incorporate the detection process into a retry or the like. It is determined whether or not the detected candidate position is regarded as a pad by confirming whether it satisfies the conditions shown in the determination after candidate position detection in (C1-1-1).

[0077] (C-3) Machine learning Judge by machine learning. Learn the image of the pad and the image of the non-pad area, and detect the position of the pad.

[0078] (Step S24: Selection of unique pad array area) Since the unique pads are the same size, they do not have a unique shape. Therefore, in the embodiment, a unique pad array is used by utilizing the arrangement of a plurality of pads. The control unit 8 selects a unique array area from the arrangement between the pads and automatically registers it as a template model.

[0079] A method for finding a unique pad array will be described with reference to FIGS. 13 and 14. FIG. 13(a) shows an example of a pad arrangement with a unique pad pitch, and FIG. 13(b) is a histogram of the pad pitches in the pad arrangement shown in FIG. 13(a). FIG. 14(a) shows an example of a pad arrangement without a unique pad pitch, and FIG. 14(b) is a histogram of the pad pitches in the pad arrangement shown in FIG. 14(a).

[0080] Use a list of candidate positions for pads arranged in a straight line. For example, if the pads are arranged in a straight line in the X direction, list the pad pitches in the X direction and create a histogram thereof.

[0081] In the pad array shown in FIG. 13(a), as shown in FIG. 13(b), the frequency of the pad pitch K1 with the minimum pad pitch is 9, the frequency of the pad pitch K2 with the next shortest pad pitch is 2, the frequency of the pad pitch K4 with the next shortest pad pitch after the pad pitch K2 is 1, and the frequency of the pad pitch K6 with the next shortest pad pitch after the pad pitch K4 is 2.

[0082] In the histogram, the areas between pads with a frequency of 1 can be regarded as basically unique regions. Therefore, the region including two pads with the pad pitch K4 of the pad pitch is the candidate position MAP of the model area.

[0083] In the pad array shown in FIG. 14(a), as shown in FIG. 14(b), the frequency of the pad pitch K1 with the minimum pad pitch is 10, the frequency of the pad pitch K2 with the next shortest pad pitch is 2, and the frequency of the pad pitch K6 with the next shortest pad pitch after the pad pitch K2 is 2.

[0084] In the pad array shown in Fig. 14(a), there is no region with a degree of 1. When there is no region with a degree of 1, candidate positions are sequentially determined from the regions with a lower degree. The region including adjacent pads or the region with the minimum pad pitch is used as the candidate position, and it is always investigated whether there is a similar pattern within the search area. If not, that region is taken as the model region. If there are two or more types with the minimum degree, the one with the shorter pad - to - pad distance is prioritized. Thus, the region including two pads with a pad pitch K2 of the pad - to - pad distance, adjacent pads, and the region with the minimum pitch K1 becomes the candidate position MAP1 of the model area. Also, the region including two pads with a pad pitch K2 of the pad - to - pad distance and the region of adjacent pads becomes the model area MAP2.

[0085] <Modification Example> Hereinafter, several representative modification examples of the embodiments will be exemplified. In the description of the following modification examples, for parts having the same configuration and function as those described in the above - mentioned embodiments, the same reference numerals as those in the above - mentioned embodiments may be used. And for the description of such parts, within a technically non - contradictory range, the description in the above - mentioned embodiments may be appropriately incorporated. Also, a part of the above - mentioned embodiments and all or part of a plurality of modification examples may be appropriately and combinatorially applied within a technically non - contradictory range.

[0086] (First Modification Example) The dicing tape 16 is pressed against the support ring 17 so as not to sag during the pickup process, thereby obtaining tension and being maintained in a flat plane. These processes are called expansion processes. When the expanded wafer 11 has a thickness of less than 200 to 300 μm, warpage occurs in the die D due to the expansion tension. When the thickness of the wafer 11 around 2011 was about 100 μm, for example, by adjusting the illumination intensity, the detection of the dicing groove was possible. However, in recent years, for ultra-thin wafers of several tens of μm, for example, 20 μm, the die warps upward even without the expansion process. As shown in FIG. 15, when the expansion process is performed, the warping of the die becomes more intense. Therefore, even if the illumination intensity is adjusted in the conventional illumination system, the brightness difference between the dicing groove and the periphery of the die cannot be obtained, and the dicing groove, that is, the contour of the die may not be clearly defined. As a result, it becomes difficult to detect the dicing groove, and it becomes difficult to roughly position the die.

[0087] A method for detecting the dicing groove of a warped die will be described with reference to FIGS. 15 to 19. FIG. 15 is a diagram showing a wafer recognition camera, coaxial illumination, and a wafer in a first modified example. FIG. 16 is a diagram showing an imaging image of the wafer. FIG. 17 is a diagram showing the irradiation light of the coaxial illumination and the reflected light by the wafer. FIG. 18 is a diagram showing an imaging image in which the contour of the die is clear. FIG. 19 is a diagram showing an imaging image in which the die is rotated in the θ direction.

[0088] First, the lighting device will be described with reference to FIG. 15. Coaxial lighting is used as the lighting device. A lens 25 (objective lens) is attached to the wafer recognition camera 24, and an image of the main surface of the die D is captured through the lens 25. Between the lens 25 and the die D, a coaxial lighting 26 having a light emitting portion 27 and a half mirror (semi-transmissive mirror) 28 inside is arranged. The irradiation light from the surface light source 27a of the light emitting portion 27 is reflected by the half mirror 28 on the same optical axis OA as the wafer recognition camera 24 and irradiated onto the wafer 11. The scattered light irradiated onto the wafer 11 on the same optical axis as the wafer recognition camera 24 is reflected by the wafer 11, and the specularly reflected light among them passes through the half mirror 28 and reaches the wafer recognition camera 24, forming an image of the wafer 11.

[0089] In the die D arranged in a grid pattern within the wafer 11, when the optical axis OA of the coaxial lighting 26, the wafer recognition camera 24, and the lens 25 are approximately aligned, as shown in FIG. 16, the die Dm directly below the optical axis OA becomes brighter than the surrounding dies. This is because the warped die has properties similar to a concave mirror, and the reflection direction of the irradiated illumination light is significantly different between the die Dm directly below the optical axis OA and the surrounding dies. For dies other than Dm, the reflected light from the warped portion does not reach the wafer recognition camera 24.

[0090] Thus, in order to make the contour shape of the die stand out, it is necessary to adjust the irradiation range of the illumination light. As shown in FIG. 15, the wafer recognition camera 24 is located directly above the die Dm as the subject, and the surface light source 27a of the coaxial lighting 26 is located away from the optical axis OA in the horizontal direction. However, as shown in FIG. 17, the apparent surface light source 27v is directly above as seen from the subject by the half mirror 28.

[0091] When the die has no warp, the maximum light emitting surface size (ESmax) of the illumination is set to about twice the maximum die size (DSmax). The light emitting surface size is adjusted according to the die size, and the adjusted light emitting surface size (ESA) is set to about twice the die size.

[0092] On the other hand, when there is warping of the die, the light-emitting surface size is adjusted according to the degree and die size, and the adjusted light-emitting surface size (ESA) is set to about 1 to 3 times the die size. Therefore, the maximum light-emitting surface size (ESmax) of the illumination is also increased accordingly.

[0093] As shown in FIG. 18, if the contour shape of the die can be lifted, the coordinates of each side Do1, Do2, Do3, Do4 of the outer peripheral portion of the die Dm can be obtained by using various edge detection processes typified by second derivative processing. Here, in image processing, an edge refers to a boundary point between an object and a background, and the contour of the object is formed by the continuous connection of these boundary points. The contour of the object has a continuously changing density value from light to dark or from dark to light. The second derivative processing is a method of defining the position where the second derivative of the density change crosses zero as an edge.

[0094] For example, for side Do1, a first detection line extending along the Y-axis direction is set, and it is detected where an edge is present on the first detection line. The Y coordinate of the edge obtained from the first detection line is set as the position of side Do1. For side Do2, a second detection line extending along the X-axis direction is set, and it is detected where an edge is present on the second detection line. The X coordinate of the edge obtained from the second detection line is set as the position of side Do2. For side Do3, it is detected where an edge is present on the first detection line. The Y coordinate of the edge obtained from the first detection line is set as the position of side Do3. For side Do4, it is detected where an edge is present on the second detection line. The X coordinate of the edge obtained from the second detection line is set as the position of side Do4.

[0095] As shown in FIG. 19, by detecting the inspection points of each side Do1 to Do4 of the outer peripheral portion not at just one point but at a plurality of points, and using the least squares method or averaging method, etc., even if the die Dm is rotated in the θ direction, the positions and directions of each side Do1 to Do4 can be obtained. Once the four sides Do1 to Do4 of the die Dm are obtained, for example, the intersection of the center lines of each pair of opposite sides can be set as the die center, and the center coordinates of the die Dm can be obtained.

[0096] In order to obtain the outline of the die located at the center of the optical axis OA by edge detection, as shown in Fig. 17, the size of the irradiation area IA is set according to the die size by the shielding area SA. The irradiation area IA is set so that the light reflected at the outer warp of the outermost die of the wafer 11 shown in Fig. 15 does not enter. For example, in Fig. 16, the outside of the white-line rectangle becomes the shielding area SA. In a general-purpose die bonder, since the die size is often not of one type, there is a mechanism that can adjust the irradiation area (light-emitting area LA or shielding area SA) in a concentric circle (concentric square) shape according to the optical axis OA.

[0097] The adjustment mechanism of the irradiation area in the coaxial illumination shown in Fig. 15 will be described with reference to Fig. 20. Fig. 20(a) is a diagram showing the case where the irradiation area is adjusted by a shielding plate, Fig. 20(b) is a diagram showing the case where the irradiation area is adjusted by a surface-emitting light source, and Fig. 20(c) is a diagram showing the case where the irradiation area is adjusted by a liquid crystal panel.

[0098] As shown in Fig. 20(a), a shielding plate 27b of an appropriate size for each product type is placed in front of the surface-emitting light source 27a (on the half mirror 28 side) so as to form the irradiation area IA adjusted to the warp. Here, since the amount of warping varies depending on the wafer thickness and the type of DAF, the appropriate size for each product type is investigated first.

[0099] As shown in Fig. 20(b), the surface-emitting light source 27a is composed of a plurality of LEDs (Light Emitting Diodes) so that the size of the light-emitting area LA can be variable. The LEDs are arranged in concentric circles or concentric squares, and the lighting is controlled by the distance from the center.

[0100] As shown in Fig. 20(c), a liquid crystal panel 27c is placed in front of the surface-emitting light source 27a to make the shielding area SA flexible. The liquid crystal panel 27c has a control unit for area control (not shown).

[0101] Also, the above description has been made on the premise that the die Dm is below the optical axis OA. However, the case where this is not so will be described with reference to FIGS. 21 and 22. FIG. 21 is a diagram showing the toothless state of the wafer. FIG. 22 is an imaging image showing a state where the die center and the optical axis center are displaced.

[0102] When the position directly below the optical axis OA is located at the boundary between dies or, as shown in FIG. 21, is located at the toothless part of the toothless wafer where some dies do not exist on the dicing tape 16, fine adjustment shall be automatically or manually performed so that the die Dm is located directly below the optical axis OA.

[0103] Also, as shown in FIG. 22, when the deviation between the center of the die Dm and the optical axis center OAC indicated by a cross is large, the contour of the die Dm may become unclear. In that case, it can be automated by performing a retry process of moving the clearer side of the contour (the left side in the figure) in the direction of the arrow so as to gradually approach the optical axis center OAC and performing contour detection again. Here, the initial alignment of the die Dm and the optical axis OA may be manual or automatic.

[0104] (Second Modification Example) When the die position can be pre-positioned more accurately than the pad pitch by rough positioning such as by detecting the dicing groove, as in the embodiment, a model including a single pad or a small number of pads may be used without worrying about the arrangement of the pads (pad array). If the rough positioning of the die is performed at runtime (during continuous bonding), the template model does not have to be unique within the range of the rough positioning accuracy.

[0105] The basic idea of the second modification example is obtained based on the fact that the accuracy of the dicing groove (the accuracy of the die position obtained by the accuracy of the dicing groove) is generally better than the pad pitch (K). Also, the basic idea of the second modification example is to detect the dicing groove, roughly position the die, set a search area based on the rough positioning position, and finely position the die by matching processing with a predetermined template.

[0106] The second modification example targets die D without unique parts such as a unique pad array or alignment marks as shown in FIG. 7, and detects the position of die D using a pad array Pr arranged at a non-unique fixed pitch as shown in FIG. 23. Hereinafter, the method for detecting the position of die D in the second modification example will be described with reference to FIGS. 11 and 24.

[0107] FIG. 24 is an explanatory diagram of pattern matching in the second modification example. FIG. 24 shows a fine positioning method having a predetermined accuracy by performing pattern matching in a search area obtained based on rough positioning at a position cut out within the region R1 shown in FIG. 23. FIG. 24(a) is a diagram showing the size of the search area, FIG. 24(b) is a diagram showing the size of the template, and FIG. 24(c) is a diagram showing the positional deviation.

[0108] First, as shown in FIG. 11, similar to step S2 of the imitation operation in the embodiment, dicing grooves DLx and DLy surrounding the target die Dm are detected from the imaging data obtained from the wafer recognition camera 24, and rough positioning of the center position Dc of die Dm is performed. Here, let the coordinates of the center position Dc be (xd, yd).

[0109] Next, as shown in Fig. 24(a), the size (Xk × Yk) of the search area Kg for pattern matching is set to be at most K (pad pitch) wider than the size (Xt × Yt) of the template Tp including the pad Dp in the X and Y directions, respectively, as shown in Fig. 24(b). This ensures that the template fits uniquely within the search area Kg. Here, let the length of the search area Kg in the X direction be Xk and in the Y direction be Yk, and the length of the template Tp in the X direction be Xt and in the Y direction be Yt. By ensuring that the template fits uniquely within the search area Kg, the center position Tc of the template in the imaging data obtained from the wafer recognition camera 24 by template matching can be detected. Here, let the coordinates of the center position Tc be (xt, yt). Note that in the above description, the search area Kg is set to be at most K (pad pitch) wider, but it can be set to be wider within the range of less than or equal to K (pad pitch) and greater than or equal to the accuracy of the dicing groove (the accuracy of the die position obtained from the accuracy of the dicing groove). In other words, the search area Kg is an area obtained by adding a length that is smaller than the pitch of the regular pattern array in the pad array Pr and larger than the accuracy of the dicing groove (the die position obtained from the accuracy of the dicing groove) to each of the X and Y directions of the size of the template Tp.

[0110] As a result, as shown in Fig. 24(c), the center position Kc of the search area Kg, that is, the position deviation (ΔX, ΔY), which is the correction value from the center position Tc of the template when there is no position deviation, can be detected. Here, let the coordinates of the center position Kc be (xk, yk).

[0111] As shown in Fig. 23, this process can also be performed at another location, and if possible, at a position in the region R2 at the diagonal position of the region R1 to obtain the rotational deviation Δθ.

[0112] The condition for the template such that the template Tp fits uniquely is that no identical or similar pattern (pattern) occurs within the search area Kg. That is, the size of the template Tp is larger than the size of the pad Dp and is formed across two or more pads Dp. In Fig. 24, the number of pads Dp involved across is 3.

[0113] Next, the die detailed positioning process flow in the second modification example described above will be described with reference to FIG. 24 for the case shown in FIG. 25.

[0114] First, two search areas Kg for pattern matching are set. For example, the search area Kg shown in FIG. 24(a) is set in the areas R1 and R2 shown in FIG. 23 (step S31). As shown in FIG. 11, the control unit 8 images the dicing grooves DLx and DLy surrounding the die Dm to be picked up by the wafer recognition camera 24 to obtain an image (imaging data) (step S32). Then, the control unit 8 detects the dicing grooves DLx and DLy from the imaging data obtained in step S32, and roughly positions the center position Dc (xd, Yd) of the die (step S33). The control unit 8 moves the die Dm to the pickup point, that is, the center of the field of view (optical axis center), based on the position of the die Dm detected in step S23 (step S34).

[0115] Next, the control unit 8 images the target die Dm by the wafer recognition camera 24 to obtain an image (step S35). Thereafter, based on the rough positioning of the center position Dc of the die obtained in step S23, the control unit 8 cuts out the two search areas determined in step S31, performs matching processing with the template Tp shown in FIG. 24(b) determined in advance, and performs fine positioning in the two search areas Kg (step S36). In this case, the position with the highest score is used as the fine coordinates of the fine positioning position.

[0116] To reconfirm the fine positioning result, it is confirmed that the relative relationship of the fine coordinates of the two search areas is the same as that at the time of setting in step S31 within a predetermined range (step S37). Thereafter, the center position (X, Y, θ) of the die Dm including the target rotation is obtained from the fine coordinates of the two fine positioning positions. Note that the number of times of reconfirming the search area may be three or more if higher accuracy is required (step S38).

[0117] According to the embodiments and modifications, one or more of the following effects can be obtained.

[0118] (1) By detecting the dicing groove, the center coordinates of the die can be calculated without performing imitation.

[0119] (2) By calculating the center coordinates of the die, the die center can be moved to the camera center (optical axis center) or the pickup point.

[0120] (3) By detecting the dicing groove, the size of the die can be obtained.

[0121] (4) By obtaining the center coordinates and size (shape) of the die, the determination conditions for the pad candidates can be set.

[0122] (5) By template matching or binarization, pad candidates can be detected based on the shape, area of the blob, aspect ratio, and position (coordinates).

[0123] (6) Assuming that the pads are arranged linearly, by using the determination conditions for the pad candidates in (4) above, it is possible to identify whether the pad candidates in (5) above are pads.

[0124] (7) By obtaining the pad pitch as the arrangement of the pads arranged linearly, a unique pad array with a unique pitch can be identified and detected.

[0125] (8) Since the detected pad array is unique, it can be automatically registered as a template model.

[0126] (9) By automatically registering the template model, the imitation operation can be automated.

[0127] (10) Since variations caused by human operations can be eliminated through automation, the quality can be stabilized by eliminating the human judgment process. As the quality of the device specifications, there are the stability of positioning, the stable operation of the device, etc., and as the quality of the products produced by the device, there is accuracy, etc.

[0128] (11) By setting the size of the irradiation area according to the die size in coaxial illumination, it is possible to detect the dicing grooves of a thin and warped die.

[0129] (12) Since the unique pad arrangement for each product type can be registered in advance, by detecting the pad arrangement with an imaging device, the product type data can be selected. Therefore, the product type data can be automatically selected when the wafer is loaded.

[0130] As described above, the disclosure made by the present inventors has been specifically described based on examples and modification examples. However, it goes without saying that the present disclosure is not limited to the above examples and modification examples and can be variously changed.

[0131] For example, in the example, die appearance inspection recognition is performed after die position recognition, but die position recognition may be performed after die appearance inspection recognition.

[0132] Also, in the example, DAF is attached to the back surface of the wafer, but DAF may not be provided.

[0133] Also, in the example, one pickup head and one bonding head are provided respectively, but two or more of each may be provided. Also, in the example, an intermediate stage is provided, but the intermediate stage may not be provided. In this case, the pickup head and the bonding head may be used in common.

[0134] In the embodiment, bonding is performed with the surface of the die facing upward. However, after picking up the die, the front and back of the die may be inverted so that the back surface of the die faces upward for bonding. In this case, the intermediate stage may not be provided. This apparatus is called a flip chip bonder.

Explanation of Signs

[0135] 10 ··· Die bonder (die bonding apparatus) 24 ··· Wafer recognition camera (imaging apparatus) 8 ··· Control unit 11 ··· Wafer D ··· Die

Claims

1. An imaging device that images a wafer having a dicing groove surrounding a die, a light emitting unit having a surface emitting light source and a semi-transmissive mirror that irradiates illumination light emitted from the light emitting unit onto the upper surface of the die, and an illumination device that irradiates the die with the illumination light along the optical axis of the imaging device, a control unit that controls the imaging device and the illumination device, comprising: the illumination device is configured to limit an irradiation area of the light emitting unit in order to detect the dicing groove, the control unit: irradiates illumination light with the illumination device, images the wafer with the imaging device, and detects the dicing groove, obtains center coordinates of the die based on the detected dicing groove and performs rough positioning of the center position of the die, moves the center of the die to the center of the optical axis of the imaging device based on the center coordinates, re-images the moved die with the imaging device, cuts out a search area of the re-imaged die based on the rough positioning, and performs a matching process between the search area and a predetermined template to determine a fine positioning position of the center position, a die bonding device.

2. In the die bonding device according to Claim 1, the control unit is configured to correct a pick-up position of the die based on the fine positioning position, a die bonding device.

3. In the die bonding device according to Claim 1, the light emitting unit of the illumination device includes a shielding plate or a liquid crystal panel that sets an irradiation area in front of the surface emitting light source, a die bonding device.

4. In the die bonding device according to Claim 1, the surface emitting light source of the light emitting unit of the illumination device is composed of a plurality of LEDs, the control unit is configured to set an irradiation area by turning on and off the plurality of LEDs, a die bonding device.

5. (a) A step of loading a substrate into the die bonding device according to Claim 1, (b) A step of loading a wafer ring holder that holds a dicing tape to which a product wafer having a dicing groove surrounding a die is attached, (c) A step of picking up the die of the product wafer, (d) A step of bonding the picked-up die onto the substrate or a die already bonded to the substrate, a method for manufacturing a semiconductor device including.

Images