Mounting device, mounting method, and manufacturing method of semiconductor device

Abstract

To provide a technique capable of shortening a time required to determine the accuracy of bond positions before the start of production.SOLUTION: A mounting device includes a head table having a plurality of heads, a drive unit that moves the head table, a stage that holds a substrate to which a workpiece is supplied by the head, and a control device that controls the drive unit using a correction table. The control device uniquely associates a relationship between a position on the substrate to which a workpiece is supplied and the plurality of heads, and is configured to calculate a correction table for the second substrate to be used in a current production on the basis of the correction table for the first substrate used in a previous production, the layout of the first substrate, and the layout of the second substrate to be used in a current production.SELECTED DRAWING: Figure 18

Description

[Technical field]

[0001] The present disclosure relates to a mounting apparatus and is applicable to, for example, a flip chip bonder. [Background technology]

[0002] Mounting devices for electronic components such as semiconductor chips generally use a suction nozzle such as a collet provided on a bond head to bond electronic components onto a substrate (for example, JP 2022-46979 A). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2022-46979 Summary of the Invention [Problem to be solved by the invention]

[0004] An object of the present disclosure is to provide a technique capable of shortening the time required for determining the accuracy of bond positions before the start of production. Other objects and novel features will become apparent from the description of this specification and the accompanying drawings. [Means for solving the problem]

[0005] A brief summary of representative aspects of this disclosure is as follows. That is, the mounting device includes a head table having a plurality of heads, a drive unit that moves the head table, a stage that holds a substrate to which a workpiece is supplied by the head, and a control device that controls the drive unit using a correction table. The control device uniquely associates a relationship between a position on the substrate to which a workpiece is supplied and the plurality of heads, and is configured to calculate a correction table for the second substrate to be used in the current production based on the correction table for the first substrate used in the previous production, the layout of the first substrate, and the layout of the second substrate to be used in the current production. Effect of the Invention

[0006] According to the present disclosure, it is possible to reduce the time required to achieve precision in bond positions before production begins. [Brief description of the drawings]

[0007] [Figure 1] FIG. 1 is a top view showing an outline of a flip chip bonder according to an embodiment. [Diagram 2] FIG. 2 is a diagram for explaining the operation of the pickup flip head and the transfer head when viewed from the direction of the arrow A in FIG. [Diagram 3] FIG. 3 is a diagram for explaining the operation of the bond head when viewed from the direction of arrow B in FIG. [Figure 4] FIG. 4 is a block diagram showing a schematic configuration of a control system of the flip chip bonder shown in FIG. [Diagram 5] FIG. 5 is a flowchart showing a method for manufacturing a semiconductor device using the flip chip die bonder shown in FIG. [Figure 6] FIG. 6 is a flow chart for explaining how to achieve high accuracy in bond positions when starting up the apparatus in this embodiment. [Figure 7] FIG. 7 is a plan view showing an example of the jig plate. [Figure 8] FIG. 8 is a recognition image of the jig plate by the bond camera. [Figure 9] FIG. 9 is a diagram showing an example of the head offset. [Figure 10] FIG. 10 is a diagram showing an example of a bond offset. [Figure 11] FIG. 11 is a diagram showing an example of a bond layout of a substrate. [Figure 12] FIG. 12 is a diagram showing an example of the accuracy measurement result. [Figure 13] FIG. 13 is a diagram showing an example of the accuracy measurement results shown in FIG. 12 in the form of a table. [Figure 14]FIG. 14 is a diagram showing an example of the head offset of each bond head. [Figure 15] FIG. 15 is a diagram showing an example of the bond offset at each bond point. [Figure 16] FIG. 16 is a diagram showing an example of the progress of obtaining precision using the method shown in FIG. [Figure 17] FIG. 17 is a flow chart for explaining how to obtain precision in bond positions when switching product types in this embodiment. [Figure 18] FIG. 18 is a diagram showing an overview of a method for calculating the bond offset of a new recipe based on an adjusted recipe. [Figure 19] FIG. 19 is a diagram for explaining an example of calculation of the bond offset of B type based on the bond coordinates of A type, the bond offset of A type, and the bond coordinates of B type. [Figure 20] FIG. 20 is a board layout diagram for explaining one line bond. [Figure 21] FIG. 21 is a diagram showing an example of the amount of positional deviation for each head and for all the heads. [Figure 22] FIG. 22 is a diagram showing an example of changing the head offset. [Figure 23] FIG. 23 is a diagram showing an example of the progress of obtaining precision by the method shown in FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Hereinafter, the embodiments will be described with reference to the drawings. However, in the following description, the same components are given the same reference numerals and the repeated description may be omitted. Note that, in order to clarify the description, the drawings may show the width, thickness, shape, etc. of each part in a schematic manner compared to the actual embodiment, but these are merely examples and do not limit the interpretation of the present invention.

[0009] A flip chip bonder, which is one aspect of a mounting device, will be described with reference to Figures 1 to 3. Figure 1 is a top view showing an outline of a flip chip bonder in an embodiment. Figure 2 is a diagram explaining the operation of a pickup flip head and a transfer head when viewed from the direction of an arrow A in Figure 1. Figure 3 is a diagram explaining the operation of a bond head when viewed from the direction of an arrow B in Figure 1.

[0010] The flip chip bonder 1 is roughly comprised of a wafer supply unit 10, pickup units 20a, 20b, transfer stage units 30a, 30b, bonding units 40a, 40b, a transport unit 50, a substrate supply unit 60, a substrate unloading unit 70, and a control unit 80. The Y2-Y1 direction (Y direction) is the front-rear direction of the flip chip bonder 1, the X2-X1 direction (X direction) is the left-right direction, and the Z1-Z2 direction (Z direction) is the up-down direction. The wafer supply unit 10 is disposed on the front side of the flip chip bonder 1, and the bonding units 40a, 40b are disposed on the rear side.

[0011] The pickup unit 20b, transfer stage unit 30b and bonding unit 40b are arranged in mirror image to the pickup unit 20a, transfer stage unit 30a and bonding unit 40a with respect to a line that passes through the die D to be picked up and extends in the Y direction, and are configured and operate in the same manner. The "b" in the reference numerals of the components of the pickup unit 20b, transfer stage unit 30b and bonding unit 40b corresponds to the "a" in the reference numerals of the components of the pickup unit 20a, transfer stage unit 30a and bonding unit 40a.

[0012] When there is no need to distinguish between the pickup units 20a and 20b, they will be referred to as pickup unit 20. When there is no need to distinguish between the transfer stage units 30a and 30b, they will be referred to as transfer stage unit 30. When there is no need to distinguish between the bonding units 40a and 40b, they will be referred to as bonding unit 40.

[0013] The wafer supply section 10 includes a wafer cassette lifter 11, a wafer holder 12, and a peeling unit 13.

[0014] A wafer cassette containing a wafer ring WR is loaded into a wafer cassette lifter 11 from outside the flip chip bonder 1. The wafer ring WR is taken out of the wafer cassette lifter 11 and supplied to a wafer holder 12, or taken out of the wafer holder 12 and carried out to the wafer cassette lifter 11. Here, the wafer ring WR is a jig to which a dicing tape DT with a wafer W affixed thereto is fixed and which can be attached to the wafer holder 12. The wafer W is divided into a plurality of dies D. The wafer W is, for example, a semiconductor wafer or a glass wafer, and the dies D as the workpiece are, for example, a semiconductor chip, a glass chip, or a MEMS (Micro Electro Mechanical Systems).

[0015] The wafer holder 12 moves in the X and Y directions by an XY table and a drive unit (not shown), and moves the die D to be picked up to the position of the peeling unit 13. The wafer holder 12 rotates the wafer ring WR in the XY plane by a drive unit (not shown). The peeling unit 13 moves in the Z1-Z2 directions by a drive unit (not shown). The peeling unit 13 peels the die D from the dicing tape DT.

[0016] The pickup unit 20a is located on the substrate supply unit 60 side (X2 direction side) with respect to the die D to be picked up. The pickup unit 20a includes a pickup flip head (PFH) 21a, a drive unit 23a, a transfer head (TRH) 25a, and a drive unit 27a. The pickup flip head 21a has a collet 22a that suction-holds the die D at its tip. The drive unit 23a raises and lowers, rotates, inverts, and moves the flip head 21a in the X direction. The transfer head 25a has a collet 26a that suction-holds the die D at its tip. The drive unit 27a raises and lowers and moves the transfer head 25a in the X direction. A wafer recognition camera 24 is provided directly above the die D to be picked up, and is shared by the pickup units 20a and 20b.

[0017] 2, the pickup flip head 21a picks up the die D based on the imaging data of the wafer recognition camera 24, rotates the pickup flip head 21a 180 degrees, inverts the die D so that it faces downward, and sets the die D in a position to be handed over to the transfer head 25a. The transfer head 25a receives the inverted die D from the pickup flip head 21a and places it on the transfer stage portion 30a.

[0018] The transfer stage unit 30a includes transfer stages (TRS) 31a1 and 31a2 on which a die D is temporarily placed, an undervision camera 34a, and an undervision correction mark 35a. The transfer stages 31a1 and 31a2 are movable in the Y direction by a drive unit (not shown).

[0019] The bonding section 40a is located on the substrate supply section 60 side with respect to the die D to be picked up. The bonding section 40a includes a bond head (BDH) 41a, a Y beam 43a of a gantry table, a bond camera 44a, and a bond head table 45a.

[0020] The bond head 41a includes heads L1, L2, L3, and L4 each having a collet 42a (C1, C2, C3, and C4) that suction-holds at its tip four dies D. The heads L1, L2, L3, and L4 are individually movable in the X direction, the Y direction, and a rotational direction within the XY plane.

[0021] The Y beam 43a extends in the Y direction so as to straddle the bond stage 46, and both ends of the Y beam 43a are supported by a pair of X beams (not shown) of a gantry table so as to be movable in the X direction. The Y beam 43a moves the bond head table 45a in the Y direction. The pair of X beams moves the Y beam 43a in the X direction. The bond head table 45a moves the bond head 41a in the Z direction.

[0022] The bond camera 44a is provided on the bond head table 45a. The bond camera 44a captures an image of a position recognition mark (not shown) on the substrate S, and recognizes the bond position. The substrate S is, for example, a wiring board, a panel, or the like.

[0023] With this configuration, the bond head 41a picks up the die D from the transfer stages 31a1, 31a2, and uses the undervision camera 34a and the bond camera 44a to capture an image of the position where the bond head 41a holds the die D. Based on this captured image data, a bond positioning correction position is calculated, and the bond head 41a is moved to bond the die D to the substrate S. When the die D is bonded to the substrate S, the substrate S is fixed by suction to the bond stage 46.

[0024] The transport section 50 includes transport rails 51, 52 along which the substrate S moves in the X direction. The transport rails 51, 52 are provided in parallel. With this configuration, the substrate S is carried out from the substrate supply section 60, moves along the transport rails 51, 52 to the bond stage 46 (bonding position), moves to the post-bonding substrate unloading section 70, and passes the substrate S to the substrate unloading section 70. While the die D is being bonded to the substrate S, the substrate supply section 60 carries out a new substrate S and waits on the transport rails 51, 52. The substrate S is carried into the substrate supply section 60 from outside the flip chip bonder 1, and the substrate S on which the die D is placed is carried out from the substrate unloading section 70 to outside the flip chip bonder 1.

[0025] Next, the control unit 80 will be described with reference to Fig. 4. Fig. 4 is a block diagram showing a schematic configuration of a control system of the flip chip bonder shown in Fig. 1.

[0026] The control system 8 includes a control unit (control device) 80, a drive unit 86, a signal unit 87, an optical system 88, etc. The control unit 80 mainly includes a control / arithmetic unit 81 composed of a CPU (Central Processing 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 and an auxiliary storage device 82b. The main storage device 82a is composed of a RAM (Random Access Memory) that stores processing programs and the like. The auxiliary storage device 82b is composed of a HDD (Hard Disk Drive) or SSD (Solid State Drive) that stores control data, image data, etc. required for control.

[0027] The input / output device 83 includes a monitor 83a for displaying the device status and information, a touch panel 83b for inputting an operator's instruction, a mouse 83c for operating the monitor 83a, and an image capture device 83d for capturing image data from the optical system 88. The input / output device 83 further includes a motor control device 83e and an I / O signal control device 83f. The motor control device 83e controls the drive unit 86. The drive unit 86 includes the drive units 23a, 23b, 27a, and 27b of the pickup units 20a and 20b, the Y beams 43a and 43b, and the bond head tables 45a and 45b. The I / O signal control device 83f captures signals from and controls the signal unit 87. The signal unit 87 includes switches and volumes for controlling the brightness of various sensors and lighting devices. The optical system 88 includes the wafer recognition camera 24, the undervision recognition cameras 34a and 34b, and the bond cameras 44a and 44b. The control / arithmetic unit 81 takes in necessary data via a bus line 84, performs calculations, controls the bond heads 41a, 41b, etc., and sends information to the monitor 83a, etc.

[0028] A part of the manufacturing process of a semiconductor device using flip chip bonder 1 (a manufacturing method and a mounting method of a semiconductor device) will be described with reference to Fig. 5. Fig. 5 is a flowchart showing a manufacturing method of a semiconductor device using the flip chip bonder shown in Fig. 1. In the following description, the operation of each part constituting flip chip bonder 1 is controlled by a control unit 80.

[0029] The following description focuses on the pickup unit 20a, transfer stage unit 30a, and bonding unit 40a, but the same is true for the pickup unit 20b, transfer stage unit 30b, and bonding unit 40b. Note that the pickup unit 20b, transfer stage unit 30b, and bonding unit 40b operate in parallel to the pickup unit 20a, transfer stage unit 30a, and bonding unit 40a to the extent that they do not interfere with each other.

[0030] [Process S1: Wafer loading process] A wafer cassette (not shown) containing a wafer ring WR is loaded into a wafer cassette lifter 11. A dicing tape DT to which a die D separated from a wafer W is attached is attached to the wafer ring WR. The wafer supply unit 10 removes the wafer ring WR from the wafer cassette filled with the wafer ring WR and carries it into the wafer holder 12. The wafer W is, for example, a semiconductor wafer, and the die D as a workpiece is a semiconductor chip.

[0031] [Process S2: Board loading process] The transport jig storing the substrate S is inserted into the substrate supply section 60. The substrate supply section 60 removes the substrate S from the transport jig. The removed substrate S is carried into the bonding section 40 via the transport section 50.

[0032] [Process S3: Pick-up process] After step S1, wafer holder 12 moves so that the desired die D can be picked up from dicing tape DT. Wafer recognition camera 24 photographs die D. Positioning and surface inspection of die D are performed based on image data acquired by photographing. The image data is processed to calculate the amount of deviation (X, Y, and θ directions) of die D on wafer holder 12 from the die position reference point of the flip chip bonder, and positioning is performed. Note that the die position reference point is previously held at a predetermined position of wafer holder 12 as the initial setting of the device. The image data is processed to perform surface inspection of die D.

[0033] The peeling unit 13 moves upward so that the upper surface of the peeling unit 13 comes into contact with the back surface of the dicing tape DT. Then, the peeling unit 13 adsorbs the dicing tape DT. The pickup flip head 21a descends while drawing a vacuum on the collet 22a, lands on the die D to be peeled, and adsorbs the die D. The pickup flip head 21a rises and peels the die D from the dicing tape DT. As a result, the die D is picked up by the pickup flip head 21a.

[0034] The pickup flip head 21a moves upward and toward the transfer head 25a. The pickup flip head 21a rotates 180 degrees, and faces the front surface of the die D downward, in other words, the back surface of the die D upward, and assumes a position for handing the die D over to the transfer head 25a. The collet 26a of the transfer head 25a picks up the die D from the collet 22a of the pickup flip head 21a. This completes the handover of the die D. The pickup flip head 21a is then turned over, and the suction surface of the collet 22a faces downward.

[0035] The transfer head 25a moves to the transfer stage 31a1. The transfer head 25a places the die D held by the collet 26a on the transfer stage 31a1. At this time, the surface of the die D faces downward. The transfer head 25a moves to a delivery position for the die D. The transfer stage 31a1 moves to a delivery position for the bond head 41a.

[0036] [Process S4: Bonding process] The transport unit 50 transports the substrate S to the bond stage 46 .

[0037] The bond camera 44a captures an image of the die D held on the transfer stage 31a1, and the collet 42a of the bond head 41a picks up a plurality of dies D (four in this embodiment) from the transfer stage 31a1. In this way, the die D is handed over.

[0038] The bond head 41a moves from the transfer stage 31a1 onto the substrate S. At this time, the bond camera 44a takes an image of the undervision correction mark 35a while moving, and the undervision camera 34a takes an image of the four dies D that are moving. The Y beam 43a moves in the X direction, and the bond head 41a moves in the Y direction. Positioning and surface inspection of the die D are performed based on the image data acquired by shooting. The image data is processed to calculate the amount of deviation (X, Y, and θ directions) of the die D on the transfer stage 31a1 from the die position reference point of the flip chip bonder, and positioning is performed. Note that the die position reference point is previously held at a predetermined position of the transfer stage 31a1 as the initial setting of the device. The image data is processed to perform surface inspection of the die D.

[0039] The bond camera 44a photographs the substrate S placed on the bond stage 46. Positioning and surface inspection of the substrate S are performed based on image data acquired by photographing. The image data is processed to calculate the amount of deviation (X, Y, and θ directions) of the substrate S from the substrate position reference point of the flip chip bonder 1. Note that the substrate position reference point is previously held at a predetermined position of the bonding unit 40 as the initial setting of the device. The image data is processed to perform surface inspection of the substrate S.

[0040] Based on the data of the undervision correction mark 35a photographed by the bond camera 44a, the image data of the die D photographed by the undervision camera 34a, and the image data of the substrate S photographed by the bond camera 44a, the bond head 41a sequentially places the four dies D held by the collets C1, C2, C3, and C4 on the substrate S. At this time, the surfaces of the dies D face downward.

[0041] The bond camera 44a photographs the die D bonded to the substrate S. Based on the image data acquired by photographing, an inspection is performed to check whether the die D is bonded at a desired position (inspection of the relative position between the die D and the substrate S), etc. In addition, a surface inspection is performed based on the image data acquired by photographing.

[0042] The bond head 41a moves to a position for delivery to and from the transfer stage 31a2. The transfer stage 31a1 moves to a position for delivery to and from the transfer head 25a.

[0043] In the above description, the transfer stage 31a1 is used as an example, but the same applies to the case where the transfer stage 31a2 is used.

[0044] [Process S5: Board unloading process] The substrate S on which the die D is bonded (mounted) is transported to the substrate unloading section 70. The substrate unloading section 70 stores the substrate S in a transport jig. The transport jig storing the substrate S is unloaded from the flip chip bonder 1.

[0045] The precision of the bond positions before the start of production is explained using the bonding section 40a as an example. The precision of the bond positions before the start of production is adjusted when the device is started up and when the product type is changed.

[0046] First, the process of achieving precision in bond positions when starting up the flip chip bonder 1 will be described with reference to Fig. 6. Fig. 6 is a flow chart for explaining the process of achieving precision in bond positions when starting up the device in this embodiment.

[0047] [Step S21: Correction of machine coordinate system errors] The error originating from the machine coordinate system is measured using a jig plate with a lattice mark to create a correction table. The error originating from the machine coordinate system will be explained with reference to Fig. 7 and Fig. 8. Fig. 7 is a plan view showing an example of a jig plate. Fig. 8 is a recognition image of the jig plate by a bond camera.

[0048] In the flip chip bonder 1, the bond head 41a picks up the die D from the transfer stage 31a1 and moves it to the substrate S to be mounted, where it mounts (mounts and bonds). The flip chip bonder 1 is equipped with a bond camera 44a, which detects the position of the substrate S by capturing an image of the substrate S, and alignment (positioning) of the die D during mounting is performed based on the position detection result. That is, alignment during mounting is performed based on the optical coordinate system of the bond camera 44a.

[0049] However, the position of the optical system coordinates of the bond camera 44a is not necessarily at the position indicated by the control data, and positional deviation occurs due to various factors. This positional deviation is caused by the mechanical error (error originating from the machine coordinate system) of the gantry table that moves the bond camera 44a. This mechanical error is not necessarily uniform, and each position in the mounting area (bond stage 46) that positions the substrate S shows a unique positional deviation amount. For this reason, when starting up the flip chip bonder 1, it is necessary to measure these positional deviation amounts for each position in the mounting area and create a correction table.

[0050] Therefore, a correction table for correcting these positional deviations is created by detecting the amount of positional deviation between the optical coordinate system of the bond camera 44a in the control data and the optical coordinate system of the actual bond camera 44a. A measurement jig plate P is used to create this correction table. The jig plate P is manufactured so that its shape and dimensions are in accordance with reference values ​​so as to prevent positioning errors, and reference marks M are provided on the surface in a grid pattern at specified positions with specified accuracy, as shown in Figure 7.

[0051] First, in the flip chip bonder 1, the jig plate P is positioned and set on the bond stage 46 in the same manner as in the case of the substrate S on which the die D is to be mounted.

[0052] Next, by recognizing reference marks etc. at the corners of the jig plate P, the amount of mounting deviation of the jig plate P itself on the bond stage 46 is calculated, and the position of each reference mark M is defined by the XY coordinates on the device.

[0053] Next, the gantry table is driven to move the bond camera 44a onto the jig plate P positioned at a predetermined position, and the reference marks M are photographed one after another to recognize the positions of the reference marks M. At this time, the movement of the bond camera 44a is controlled in the control data so that the origin of the optical coordinate system coincides with the center of the reference mark M. However, in the position recognition result by the bond camera 44a, the origin 0 does not necessarily coincide with the center of the reference mark M due to mechanical errors of the gantry table. FIG. 8 shows an image diagram when position recognition is performed on the reference mark Mn of the jig plate P. Here, position recognition detects a positional deviation of Δxn, Δyn with respect to the origin 0 of the optical coordinate system for the reference mark Mn (center coordinates (Xn, Yn) in the mechanical coordinate system).

[0054] That is, when the die D is mounted near the coordinates (Xn, Yn) in the machine coordinate system, it is necessary to further add corrections of Δxn and Δyn to the position of the substrate S recognized by the bond camera 44a. Therefore, by determining the positional deviation amounts (Δxn, Δyn) for all the reference marks M at the start-up of the flip chip bonder 1 after the assembly is completed, it is possible to obtain positional correction amount data for correcting the positional deviation inherent to the coordinate system XY of the gantry table.

[0055] Next, positional deviations due to causes other than mechanical errors will be described with reference to Figures 9 and 10. Figure 9 is a diagram showing an example of head offset. Figure 10 is a diagram showing an example of bond offset.

[0056] The method using the jig plate P detects the amount of positional deviation of the bond camera 44a in the bond stage 46. In addition to this positional deviation, there is also a positional deviation of the bond head 41a (deviation from the positional relationship between the bond head 41a and the bond camera 44a). In this specification, this positional deviation is referred to as head positional deviation, and a correction value obtained by inverting the sign of the head positional deviation amount is referred to as head offset (HO). For example, FIG. 9 shows an example in which the head L1 is shifted by Δx in the X1 direction and Δy in the Y1 direction. The bond head 41a also has a rotational deviation in the XY plane.

[0057] Furthermore, there is a positional deviation of the mounting position of the die D by the collet 42a (a positional deviation between the collet 42a and the bond head 41a). In this specification, this positional deviation is called a bond positional deviation, and a correction value obtained by inverting the sign of the bond positional deviation amount is called a bond offset (BO). For example, Figure 10 shows an example in which the collet C1 of the head L1 is misaligned in the Y2 direction.

[0058] As shown in Fig. 10, a bond head 41a is equipped with a plurality of heads L1 to L4, each of which has a collet C1 to C4 attached to its shaft tip, so that the heads can move up and down independently. For example, the shaft centers of the heads L1 to L4 may be eccentric with respect to the optical axis of the bond camera 44a. Therefore, the greater the difference between the height of the bond camera 44a (bond head) and the height of the top surface of the substrate S, the more different the coordinates (without eccentricity) of the tip positions of the collets C1 to C4 moved to the target position based on the bond camera 44a and the coordinates (with eccentricity) of the tip positions of the collets C1 to C4 when the heads L1 to L4 are lowered after being moved to the same position based on the heads L1 to L4.

[0059] As a result, accurate positional correction may not be possible when mounting the die D simply by performing correction based on the result of recognizing the reference mark M using only the bond camera 44a described above.

[0060] Next, the creation of the correction table will be described with reference to Figures 11 to 15. Figure 11 is a diagram showing an example of a bond layout of a substrate.

[0061] [Step S22: Accuracy measurement with all-point bonding] (Step S22a: Linking the bond positions on the board to the head) The bonding points are uniquely linked to the heads L1 to L4 used for bonding. In other words, the bonding points are operated so that they are bonded only by a specific head. This improves the positioning and repeatability of the machine precision.

[0062] For example, as shown in Fig. 11, the board S has four mounting areas BA1 to BA4. The mounting area BA1 is X=1 to 62 and Y=1 to 51. The mounting area BA2 is X=1 to 62 and Y=52 to 102. The mounting area BA3 is X=63 to 124 and Y=1 to 51. The mounting area BA4 is X=63 to 124 and Y=52 to 102.

[0063] The bonding points of the two mounting areas BA1 and BA2 on the X1 side are bonded by heads R1 to R4 of the bond head 41b. The bonding points of the two mounting areas BA3 and BA4 on the X2 side are bonded by heads L1 to L4. Each mounting area BA1 to BA4 is divided into eight areas in the Y direction, and heads L1 to L4 are assigned to bond to those areas.

[0064] For example, the eight divided areas of mounting area BA3 are assigned, in order from the Y1 side, to an area bonded by head L1, an area bonded by head L2, an area bonded by head L3, an area bonded by head L4, an area bonded by head L1, an area bonded by head L2, an area bonded by head L3, and an area bonded by head L4. Mounting area BA3 is assigned areas bonded by heads L1 to L4, just like mounting area BA4. Mounting areas BA1 and BA2 are assigned areas bonded by heads R1 to R4, just like mounting areas BA3 and BA4.

[0065] (Step S22b: Positioning of production board) A production substrate S is carried into the flip chip bonder 1, and the position and dimensions of the carried-in substrate S are measured and corrected by the bond camera 44a.

[0066] First, the carried-in substrate S is positioned and set on the bond stage 46 .

[0067] Next, by recognizing the reference marks MS and the like at the corners of the substrate S, the amount of mounting misalignment of the substrate S itself on the bond stage 46 is calculated, and the position of the substrate S is specified by the XY coordinates on the device. Also, by recognizing the positions of the edges, etc. of the substrate S, the dimensions of the substrate S are measured. In Fig. 11, five reference marks MS (at the four corners and the center) are provided in each of the four mounting areas.

[0068] (Step S22c: Bond all points) The die D is bonded to each bond point on the substrate S with the head offset and bond offset set to "0".

[0069] (Step S22d: Bond position accuracy measurement) The bond camera 44a photographs the die D bonded to the substrate S. Based on the image data acquired by photographing, the bond position accuracy is measured using a reference (e.g., a reference mark MS) on the substrate S, and the error (amount of positional deviation) is examined and stored in a storage device.

[0070] An example of error in measuring bond position accuracy will be described with reference to Figures 12 and 13. Figure 12 shows an example of accuracy measurement results, plotting all errors at each bond point. The center of the cross in Figure 12 indicates the average amount of positional deviation at all bond points. Figure 13 shows the example of accuracy measurement results shown in Figure 12 in table format.

[0071] In FIG. 13, "Total" is the error at all bond points bonded by the bond heads 41a and 41b. "L" is the error at all bond points bonded by the bond head 41a (for example, half of the bond points on the X2 side shown in FIG. 11), including heads L1 to L4. "R" is the error at all bond points bonded by the bond head 41b (for example, half of the bond points on the X1 side shown in FIG. 11), including heads R1 to R4. "Xdiff" is the amount of misalignment in the X direction, "Ydiff" is the amount of misalignment in the Y direction, and "Tdiff" is the amount of misalignment in the rotation direction. "average" is the average value, "max" is the maximum value, and "min" is the minimum value. "R" is "max" - "min". "USL" is the upper limit of the allowable range, and "LSL" is the lower limit of the allowable range. "Cp" and "CpK" are process capabilities. "3sigma" is three times the standard deviation (3σ). "stdev" is the standard deviation.

[0072] For example, "Total", "Xdiff", and "average" are the average values ​​of the positional deviations in the X direction of heads L1 to L4 and R1 to R4, and the value is "-7.2". "Total", "Ydiff", and "average" are the average values ​​of the positional deviations in the Y direction of heads L1 to L4 and R1 to R4, and the value is "-0.5". These are the values ​​at the centers of the crosses shown in FIG. 12.

[0073] [Step S23: Accuracy determination] If the bond accuracy is OK (error is within the allowable range), the process ends. Two or three full-point bonds are required until accuracy is achieved. If the bond accuracy is NG (error exceeds the allowable range), the process proceeds to step S24.

[0074] [Step S24: Calculation of HO and BO] The head offset (HO) will be described with reference to Fig. 14. Fig. 14 is a diagram showing an example of the head offset of each bond head, and shows the average deviation amount for each head.

[0075] The measured errors (positional deviations) at each bond point are averaged for each head L1-L4 and R1-R4 to calculate the average deviation (average error). A table (head offset table) is created to correct the average deviations for each head L1-L4 and R1-R4 shown in FIG. 14 for each head L1-L4 and R1-R4, and stored in a storage device. The head offsets are the values ​​shown in FIG. 14 with their signs inverted. FIG. 14 shows the average position deviations for each head L1-L4 and R1-R4. "X" is the position deviation in the X direction, "Y" is the position deviation in the Y direction, and "T" is the position deviation in the rotation direction. For example, the offset in the X direction of head L1 is "7.3", the offset in the Y direction is "1.0", and the offset in the rotation direction is "-0.005".

[0076] The bond offset (BO) will be described with reference to Fig. 15. Fig. 15 is a diagram showing an example of the bond offset at each bond point.

[0077] The error at each bond point is calculated by subtracting the average error for each head L1-L4, R1-R4 from the measured error at each bond point. The sign of the calculated error is inverted to calculate the bond offset, and a correction table as shown in FIG. 15 is created. The correction table is stored in a storage device. In FIG. 15, "#X" is the X-direction position of the bond point on the substrate S, "Y" is the Y-direction position of the bond point on the substrate S, "OffsetX" is the X-direction bond offset at the bond point on the substrate S, "OffsetY" is the Y-direction bond offset at the bond point on the substrate S, and "OffsetT" is the rotational bond offset at the bond point on the substrate S. For example, in the first row, the X-direction bond offset at the bond point (X,Y)=(3,1) is "-2.6", the Y-direction bond offset is "-2", and the rotational bond offset is "0".

[0078] [Step S25: Accuracy measurement with all-point bonding] (S25a: Positioning of production boards) As in step S22b, the substrate is positioned.

[0079] (S25b: All points bonded) In step S24, the die D is bonded to each bond point on the substrate S under the conditions (head offset and bond offset) reflected therein.

[0080] (S25c: Bond position accuracy measurement) As in step S22d, the bond position accuracy is measured.

[0081] An example of precision adjustment using the method shown in Fig. 6 will be described with reference to Fig. 16. Fig. 16 is a diagram showing an example of the progress of precision adjustment using the method shown in Fig. 6.

[0082] A (A-1, A-2, A-3) and B (B-1, B-2, B-3) are different devices, but show the error in precision adjustment for the same layout board. The bond point is 11632. A-1 and B-1 are plots of all the errors at each bond point when all points are bonded with head offset and bond offset set to "0" (first all points bond in the flow shown in FIG. 6). A-2 and B-2 are plots of all the errors at each bond point in the second all points bond in the flow shown in FIG. 6. A-3 and B-3 are plots of all the errors at each bond point in the third all points bond in the flow shown in FIG. 6. In the example shown in FIG. 16, the bond precision is within the specified range by three all points bonds. In B-1, the head offsets of heads L1 to L4 and R1 to R4 vary greatly.

[0083] Next, the process of achieving precision in bond positions when switching product types will be described with reference to Fig. 17. Fig. 17 is a flow chart for explaining the process of achieving precision in bond positions when switching product types in this embodiment.

[0084] When the product type changes, the number of dies mounted on the board and the pitch (bond layout) will differ, so it is necessary to create and register the head offset and bond offset for each. In other words, when creating and registering a recipe for a new product type, it is necessary to carry out the steps from step S22 onwards in the flow shown in Figure 6.

[0085] However, it is necessary to perform all-point bonding multiple times, which takes time from achieving the required precision to starting production. In addition, multiple all-point bonding processes consume a large number of dies.

[0086] For example, the size of the board S (referred to as board S of type A (T_A)) shown in FIG. 11 is about 600 mm×600 mm. The board S of type A has four mounting areas with 62 bond points in the X direction and 51 bond points in the Y direction (number of dies mounted: 62×51×4=12648). For example, the board S of type B (T_B) has four areas with 58 bond points in the X direction and 54 bond points in the Y direction (number of dies mounted: 58×54×4=12528). The board S of type C (T_B) has four areas with 56 bond points in the X direction and 55 bond points in the Y direction (number of dies mounted: 56×55×4=12320). The board S of type B and the board S of type C have the same board size as the board S of type A, but the number of dies mounted and the pitch are different from those of the board S of type A.

[0087] Therefore, in the embodiment, a correction table at the time of changing the production type (for example, a correction table for product type B) is created based on a correction table obtained during previous production (for example, a correction table for product type A).

[0088] [Step S31: Calculation of BO] FIG. 18 is a diagram showing an overview of a method for calculating the bond offset of a new recipe based on an adjusted recipe.

[0089] The bond offset value of the recipe for the new variety to be used in the current production is calculated from the correction value whose accuracy was confirmed in another recipe during the previous production (for example, the bond offset calculated by the method shown in Fig. 6). In other words, the bond offset of the new recipe is calculated from the created and adjusted recipe (hereinafter referred to as the adjusted recipe) by referring to the bond layout of the newly created recipe (hereinafter referred to as the new recipe).

[0090] For example, as shown in FIG. 18, the bond offset of a B-type substrate is calculated based on the bond layout of a A-type (T_A) substrate S, the bond offset of the A-type substrate S, and the bond layout of a B-type (T_B) substrate S. Here, the other recipe is the bond layout and bond offset of the A-type substrate S. The bond offset of the A-type substrate S is the bond offset calculated by the method shown in FIG. 6 (for example, the bond offset shown in FIG. 15), and the accuracy has been confirmed. In addition, the new recipe is the bond layout and bond offset of the B-type substrate S. In addition, the bond layout is the coordinates (bond coordinates) of each bond point.

[0091] More specifically, based on the bond coordinate data and the correction table of the substrate of the adjusted recipe and the bond coordinate data of the substrate of the new recipe, the correction value (bond offset) of the new recipe is calculated to create a correction table.

[0092] A method for calculating the bond offset of a new recipe will be described with reference to Fig. 19. Fig. 19 is a diagram for explaining an example of calculating the bond offset of B type based on the bond coordinates of A type, the bond offset of A type, and the bond coordinates of B type.

[0093] The position of the lattice consisting of the bond coordinates of the adjusted recipe where the bond coordinates of the new recipe are included is calculated, and the correction values ​​of the bond coordinates of the new recipe are calculated by interpolation from the correction values ​​of the four points of the lattice.

[0094] In FIG. 19, the bond layouts of the A-type substrate S and the B-type substrate S are shown overlapping each other. The numbers on the inside are the lattice numbers of the A-type substrate S, and the numbers on the outside are the lattice numbers of the B-type substrate S. As can be seen from the positions of the lattice numbers (1 to 10) in the X direction, the pitch of the bond points of the A-type substrate S in the X direction is smaller than the pitch of the bond points of the B-type substrate S. As can be seen from the positions of the lattice numbers (1 to 6) in the Y direction, the pitch of the bond points of the A-type substrate S in the Y direction is smaller than the pitch of the bond points of the B-type substrate S. The lattice point Gp(X7,Y5) of the B-type substrate S is located among the four lattice points G1(X8,Y5), G2(X8,Y4), G3(X7,Y5), and G4(X7,Y4) of the A-type substrate S.

[0095] The bond coordinates of lattice points G1(X8,Y5), G2(X8,Y4), G3(X7,Y5), and G4(X7,Y4) of the A-type board are (x1,y1), (x2,y2), (x3,y3), and (x4,y4), and the bond offsets in the X direction at lattice points G1(X8,Y5), G2(X8,Y4), G3(X7,Y5), and G4(X7,Y4) are u1 to u4, and the bond offsets in the Y direction at lattice points G1(X8,Y5), G2(X8,Y4), G3(X7,Y5), and G4(X7,Y4) are v1 to v4.

[0096] If the bond coordinates and bond offsets at lattice points G1 to G4 are G1(x1,y1,u1,v1), G2(x2,y2,u2,v2), G3(x3,y3,u3,v3) and G4(x4,y4,u4,v4), G1(x1,y1,u1,v1) =(256978.6,265958.1,-0.6,1.8) G2(x2,y2,u2,v2) =(256978.6,271477.1,0.0,1.5) G3(x3,y3,u3,v3) =(261527.7,271477.1,-1.2,-2.8) G4(x4,y4,u4,v4) =(261527.7,265958.1,-1.1,-2.4) It is.

[0097] Here, the bond coordinates of the lattice point Gp (X7, Y5) of the substrate S of type B are (xp, yp), the bond offset in the X direction at the lattice point Gp is ​​up, and the bond offset in the Y direction is vp. If the bond coordinates and bond offset at the lattice point Gp are Gp (xp, yp, up, vp), and up = Xo, vp = Yo, then Gp(xp,yp,up,vp) =(259856.8,267504,5,Xo,Yo) It is.

[0098] By interpolating (Xo,Yo), (Xo,Yo)=(-0.80,-0.72) It becomes.

[0099] The bond offsets are calculated by performing the above calculations for all bond points of type B. Then, a correction table is created and stored in the storage device.

[0100] [Step S32: Accuracy measurement using 1 line bond] The one-line bond will be described with reference to Fig. 20. Fig. 20 is a board layout diagram for explaining the one-line bond.

[0101] (Step S32a: Positioning of production board) First, the production board is positioned in the same manner as in step S22b.

[0102] (Step S32b: 1 line bond) Next, using the bond offset calculated in step 31 and the adjusted head offset, as shown in FIG. 20, heads L1 to L4 bond die D to each bond point in a row LL in the Y direction on the X2 side of the substrate S, and heads R1 to R4 bond die D to each bond point in a row RL in the Y direction on the X1 side of the substrate S.

[0103] (Step S32c: Bond position accuracy measurement) As in step S22d, the bond position accuracy is measured, and the error (amount of positional deviation) is checked and stored in the storage device.

[0104] [Step S33: Accuracy determination] If the bond accuracy is OK (error is within the allowable range), the process ends. One or two single-line bonds are required until accuracy is achieved. If the bond accuracy is NG (error exceeds the allowable range), the process proceeds to step S34.

[0105] [Step S34: Calculation of HO] The calculation and reflection of head offset will be described with reference to Figs. 21 and 22. Fig. 21 is a diagram showing an example of the positional deviation amount for each head and for all the heads. "L1" to "L4" and "R1" to "R4" in Fig. 21 indicate the positional deviation amounts of heads L1 to L4 and R1 to R4, and "TOTAL" indicates the positional deviation amount of all the heads. Fig. 22 is a diagram showing an example of changing the head offset. "BC" in Fig. 22 is the head offset before the change, and "AC" is the head offset after the change.

[0106] As shown in Fig. 21, the errors (positional deviations) at the measured bond points are averaged for heads L1 to L4 and R1 to R4 to calculate the average deviation (average error). Here, the number in the center of the cross in Fig. 21 is the average deviation.

[0107] Next, head offsets that can be individually corrected for the heads L1 to L4 and R1 to R4 are used to correct the average deviation amounts of the heads L1 to L4 and R1 to R4, and adjust the overall average value.

[0108] For example, in the case of the results shown in FIG. 21, the head offset after the change is calculated by subtracting the average deviation amount (average value) of each of the heads L1 to L4 and R1 to R4 from the head offset before the change of each of the heads L1 to L4 and R1 to R4. For example, as shown in FIG. 22, the head offset in the X direction of head L1 before the change is "6.5", and the head offset in the Y direction is "8.4". Also, as shown in FIG. 21, the average deviation amount in the X direction of head L1 is "0.2", and the average deviation amount in the Y direction is "1.1". Therefore, the head offset in the X direction after the change is "6.5-0.2=6.3", and the head offset in the Y direction is "8.4-1.1=7.3".

[0109] The above calculation is performed for each of the heads L1 to L4 and R1 to R4, the head offsets are calculated, and the updated head offsets are stored in the storage device.

[0110] [Step S35: Accuracy measurement with all-point bonding] As in step S25, the substrate is positioned, all points are bonded, and the bond position accuracy is measured.

[0111] In the flow shown in FIG. 17, the measurement result is fed back only to the head offset, but if necessary, the measurement result may be fed back to the bond offset.

[0112] FIG. 23 is a diagram showing an example of the progress of precision adjustment by the method shown in FIG. 17. C (C-1, C-2, C-3) is the same equipment as A (A-1, A-2, A-3) in FIG. 16, but the type of board is different. D (D-1, D-2, D-3) is the same equipment as B (B-1, B-2, B-3) in FIG. 16, but the type of board is different. C (C-1, C-2, C-3) and D (D-1, D-2, D-3) are precision adjustment for boards with the same layout. Also, C-1, C-2 and D-1, D-2 have 202 bond points on the left and right lines (2 lines). C-3 and D-3 are all bonded, and the bond points are 11800.

[0113] C-1 and D-1 are diagrams plotting all the errors at each bond point when one line bond is performed on the left and right sides using the bond offset calculated in step S31 (the first one line bond in the flow shown in Figure 17). C-2 and D-2 are diagrams plotting all the errors at each bond point of the second one line bond in the flow shown in Figure 17. C-3 and D-3 are diagrams plotting all the errors at each bond point of all bonds in the flow shown in Figure 17. In the example shown in Figure 17, the bond accuracy falls within the specified range with the second one line bond.

[0114] In the embodiment, a correction table with high accuracy (with the same level of accuracy as the method of repeating all-point bonding) can be created by consuming a small number of sample dies even for a new substrate.

[0115] The invention made by the present inventors has been specifically described above based on embodiments and examples. However, it goes without saying that the present invention is not limited to the above embodiments and examples, and various modifications are possible.

[0116] For example, in the embodiments, an example has been described in which an inversion mechanism is provided in the pickup flip head, the transfer head receives the die from the pickup flip head and places it on the transfer stage, and the die is inverted / non-inverted on the transfer stage, but this is not limited to this, and it is also possible not to provide an inversion mechanism in the pickup head and to invert / non-invert the die on the transfer stage.

[0117] Although an example in which there are four bond heads has been described in the embodiment, the present invention is not limited to this, and there may be one or more bond heads.

[0118] In the embodiment, an example has been described in which the bond camera is provided on the same bond table as the bond head, but the bond camera may be provided on a table different from the bond table, or the bond camera may be fixed and the substrate may be moved.

[0119] In the embodiment, an example has been described in which a bond camera is used to measure the bond position accuracy, but a measuring device other than a camera may also be used.

[0120] In the embodiment, an example in which there are two Y beams has been described, but there may be only one.

[0121] In the embodiment, a bond head has been described as an example, but the present invention can also be applied to a multi-head device, for example, a multi-head coating device. [Explanation of symbols]

[0122] 45a···Bond head table (head table) 43a Y beam (drive unit) L1~L4...head 46 Bond Stage 80... Control unit (control device)

Claims

1. a head table having a plurality of heads; A drive unit for moving the head table; a stage for holding a substrate to which a workpiece is supplied by the head; A control device that controls the driving unit using a correction table; Equipped with The control device includes: A relationship between a position on the substrate to which a workpiece is supplied and the plurality of heads is uniquely linked, a mounting device configured to calculate a correction table for a second board to be used in a current production, based on a correction table for a first board used in a previous production, a layout of the first board, and a layout of a second board to be used in a current production.

2. In the mounting device of claim 1, The correction table is a head offset for each head and all bond offsets of positions (bond points) to which a workpiece should be supplied, the head offset is a correction value for correcting a positional deviation of the head, The bond offset is a correction value that corrects the positional deviation between the position where the workpiece is supplied and the bond point.

3. In the mounting device of claim 2, The control device is a mounting device configured to calculate a correction table for the second substrate based on a correction table for the first substrate, coordinate (bond coordinate) data of the bond point on the first substrate, and coordinate (bond coordinate) data of the bond point on the second substrate.

4. In the mounting device of claim 3, The control device includes: Calculating at which position the bond coordinates of the second substrate are included in a lattice formed by the bond coordinates of the first substrate; A correction value at the bond coordinates of the second substrate is calculated by interpolation from the correction values ​​of the four points of the lattice; A mounting apparatus configured to create a correction table for the second substrate based on the calculated correction value.

5. In the mounting device of claim 2, The control device includes: Based on the correction table, a workpiece is bonded to each bond point in one row on a substrate having the same layout as the second substrate; A mounting apparatus configured to measure bond position accuracy at each of the bond points.

6. In the mounting apparatus of claim 5, The control device includes: The positional deviation amount at each bond point measured is averaged for each head to calculate an average deviation amount; A mounting device configured to correct the head offset for each of the heads based on the average deviation amount.

7. In the mounting device of claim 2, The control device includes: Setting the head offset and the bond offset to "0" and bonding the workpiece to all of the bond points on a substrate having the same layout as the first substrate; A mounting apparatus configured to measure bond position accuracy for each of the bond points.

8. The mounting device according to claim 7, The control device includes: Calculating an average error by averaging the measured errors at each of the bond points for each of the heads; calculating the head offset for each of the heads based on the average error; calculating an error at each of the bond points by subtracting the average error for each of the heads from the error at each of the bond points measured; A mounting apparatus configured to calculate the bond offset for each of the bond points based on the calculated error.

9. A step of carrying a third substrate having the same layout as the second substrate into the mounting apparatus of claim 1; bonding a workpiece to the third substrate; Implementation methods including:

10. A step of carrying a third substrate having the same layout as the second substrate into the mounting apparatus of claim 1; bonding a die to the third substrate; A method for manufacturing a semiconductor device comprising the steps of:

Images