Processing system and processing method

Abstract

This processing system mounts a plurality of dies held on a carrier on a target substrate. The processing system comprises: a substrate holding unit that holds the target substrate and is configured to be movable in a horizontal direction; a pickup unit that picks up the dies from the carrier; a head unit that receives the dies from the pickup unit and joins the dies to the target substrate held by the substrate holding unit; and a control unit. On the basis of position information of the plurality of dies on one target substrate measured by a measurement unit for measuring the positions of the plurality of dies mounted on the target substrate, the control unit executes control for deriving a target-substrate-specific positional deviation component of the positions of the plurality of dies on another target substrate.

Description

Processing system and processing method

[0001] The present disclosure relates to a processing system and a processing method.

[0002] Patent Document 1 discloses a chip-on-wafer bonding method for mounting chips on a wafer, in which a surface activation process and a hydrophilization process are performed on the chips, a surface activation process and a hydrophilization process are performed on the substrate, and then multiple chips are bonded to the substrate.

[0003] Patent No. 6337400

[0004] The technology according to the present disclosure mounts multiple dies in appropriate positions on a target substrate.

[0005] One aspect of the present disclosure is a processing system that mounts multiple dies held on a carrier onto a target substrate, the processing system comprising: a substrate holding unit that holds the target substrate and is configured to be freely movable horizontally; a pickup unit that picks up the die from the carrier; a head unit that receives the die from the pickup unit and bonds the die to the target substrate held by the substrate holding unit; and a control unit, wherein the control unit performs control to derive positional misalignment components for each target substrate of the positions of the multiple dies on another target substrate based on positional information of the multiple dies on one of the target substrates measured by a measurement unit that measures the positions of the multiple dies mounted on the target substrate.

[0006] According to the present disclosure, multiple dies can be mounted in appropriate positions on a target substrate.

[0007] FIG. 1 is a plan view showing an outline of the configuration of a wafer on which a plurality of types of dies are mounted. FIG. 2 is a cross-sectional view showing an outline of the configuration of a first carrier that holds a first die. FIG. 3 is a cross-sectional view showing an outline of the configuration of a second carrier that holds a second die. FIG. 4 is a cross-sectional view showing an outline of the configuration of a second carrier that holds a second die. FIG. 5 is a plan view showing an outline of the configuration of a processing system. FIG. 6 is a flow diagram showing main steps of a die-on wafer manufacturing process. FIG. 7 is an explanatory diagram showing a schematic diagram of some steps of the die-on wafer manufacturing process. FIG. 8 is a perspective view showing an outline of the configuration of a first bonding apparatus. FIG. 9 is a cross-sectional view showing an outline of the configuration of an air supply unit. FIG. 10 is a side view and a plan view showing an outline of the configuration of a head unit. FIG. 11 is an explanatory diagram showing a schematic diagram of a bonding process of a first die to a wafer. FIG. 12 is an explanatory diagram showing a schematic diagram of some steps of the bonding process in the first bonding apparatus. FIG. 13 is an explanatory diagram showing a schematic diagram of some steps of the bonding process in the first bonding apparatus. FIG. 1 is an explanatory diagram schematically showing some steps of a bonding process in a first bonding apparatus; FIG. 2 is an explanatory diagram schematically showing some steps of a bonding process in a first bonding apparatus; FIG. 3 is a flow diagram showing main steps of a first die position correction method; FIG. 4 is an explanatory diagram schematically showing how the position of the first die is corrected; FIG. 5 is an explanatory diagram schematically showing positional misalignment components (shift component, scaling component, rotation component) on a die-by-die basis in local correction; and FIG. 6 is an explanatory diagram schematically showing positional misalignment components (shift component, scaling component, rotation component, orthogonal component) on a wafer-by-wafer basis in global correction.

[0008] In recent years, in the manufacturing process of semiconductor devices, in response to demands for even higher performance and higher density of devices, a chip-on-wafer (CoW) manufacturing process has been studied as one method of three-dimensional packaging technology. Chip-on-wafer manufacturing is carried out by a method disclosed in Patent Document 1, for example.

[0009] In a chip-on-wafer manufacturing process (hereinafter referred to as a "die-on-wafer manufacturing process"), as disclosed in Patent Document 1, a semiconductor chip (hereinafter referred to as a "die") is subjected to a surface activation treatment and a hydrophilization treatment, and a semiconductor substrate (hereinafter referred to as a "wafer") is also subjected to a surface activation treatment and a hydrophilization treatment, and then multiple dies are bonded to the wafer. Bonding of multiple dies to the wafer is performed, for example, using a bonding device. In the bonding device, the die, which is attached to the tip (lower end) of a head unit, is positioned opposite the wafer held by a wafer holder. The head unit is then lowered to place the die on the wafer and bond them.

[0010] To improve the positional accuracy of multiple dies on a wafer, the positions of multiple dies mounted on one wafer are measured, and the positions of multiple dies mounted on a subsequent wafer are corrected based on the measurement results. For example, the average positional deviation between the measured positions of all dies on one wafer and their target positions is calculated, and the horizontal positions of the multiple dies are corrected based on the average positional deviation. Alternatively, the average angular deviation between the measured angles calculated from the measured positions of all dies on one wafer using trigonometric functions and the target angles of all dies is calculated, and the rotational positions of the multiple dies are corrected based on the average angular deviation. In other words, in such a case, one wafer has one correction value for the horizontal position and one correction value for the rotational position.

[0011] However, when position correction for multiple dies is performed on one wafer using a single correction value as in the past, it is not possible to correct positional deviations caused by driving errors (e.g., scaling errors, rotation errors, and orthogonal errors) in the wafer holder or head unit. Therefore, there is room for improvement in the positional accuracy of multiple dies mounted on a wafer.

[0012] The technology disclosed herein mounts multiple dies at appropriate positions on a target substrate. Hereinafter, a processing system and a processing method according to the present embodiment will be described with reference to the drawings. Note that in this specification and the drawings, elements having substantially the same functional configurations are designated by the same reference numerals, and redundant description will be omitted.

[0013] In this embodiment, as shown in Fig. 1, multiple types of dies D1 and D2 are mounted on a wafer W as a target substrate. Specifically, multiple first dies D1 held on a first carrier C1 as shown in Figs. 2 and 3 and multiple second dies D2 held on a second carrier C2 as shown in Figs. 4 and 5 are bonded to the wafer W as a mounting target.

[0014] 1, the wafer W on which the first die D1 and the second die D2 are mounted is a semiconductor wafer such as a silicon substrate or a glass substrate used in the manufacturing process of semiconductor devices. For example, the diameter of the wafer W is 300 mm and the thickness is approximately 800 μm. The surface of the wafer W on which the first die D1 and the second die D2 are mounted is referred to as the front surface Wa, and the surface opposite to the front surface Wa is referred to as the back surface Wb.

[0015] The number and arrangement of the first die D1 and the second die D2 to be mounted on the front surface Wa of the wafer W are set to a desired pattern. A device layer (not shown) may be formed on the front surface Wa. This device layer is formed to correspond to a device layer E of the first die D1, which will be described later.

[0016] 2 and 3, the first carrier C1 holds a plurality of first dies D1. The first die D1 has a configuration in which, for example, a silicon layer S1 and a device layer E are stacked. A circuit is formed in the device layer E. As will be described later, when the first die D1 is bonded to the wafer W, high bonding accuracy is required for the first die D1, i.e., the first die D1 is a high-precision die. Note that the surface of the first die D1 on which the device layer E is formed is referred to as the front surface D1a, and the surface opposite to the front surface D1a is referred to as the back surface D1b.

[0017] 4 and 5, the second carrier C2 holds a plurality of second dies D2. Each second die D2 has a structure in which, for example, a silicon layer S2 and an oxide film F are stacked. No devices such as circuits are formed on the oxide film F. As will be described later, when the second die D2 is bonded to the wafer W, low bonding accuracy is required for the second die D2; that is, the second die D2 is a low-precision die. The surface of the second die D2 on which the oxide film F is formed is referred to as the front surface D2a, and the surface opposite to the front surface D2a is referred to as the back surface D2b.

[0018] As described above, the first die D1 and the second die D2 are different types of dies. The first die D1 has a circuit formed thereon, whereas the second die D2 does not. Therefore, the joining accuracy required for the first die D1 is relatively high, while the joining accuracy required for the second die D2 is relatively low. In addition, the shapes (size and thickness) of the first die D1 and the second die D2 are also different. Therefore, the number of first dies D1 held by the first carrier C1 and the number of second dies D1 held by the second carrier C2 are also different.

[0019] The first carrier C1 and the second carrier C2 have the same configuration. First, the configuration of the first carrier C1 will be described.

[0020] 2 and 3, the first carrier C1 has an upper surface that serves as an adsorption surface for holding a plurality of first dies D1 by electrostatic and vacuum adsorption. The first carrier C1 adsorbs and holds the silicon layer S1 on the back surface D1b side of the first die D1. The first carrier C1 has a structure in which a main body M1 and an insulating layer N1 are stacked.

[0021] The main body M1 has approximately the same diameter and thickness as the wafer W on which the first die D1 is mounted, e.g., a diameter of 300 mm and a thickness of approximately 800 μm. The main body M1 is made of any conductive material, such as silicon, aluminum, an aluminum alloy, stainless steel, alumina, zirconia, SiC, or titanium. In other words, a silicon substrate may be used as the main body M1. The main body M1 may have approximately the same diameter as the wafer W on which the first die D1 is mounted, but may have a different thickness (e.g., 500 μm to 1000 μm) from the wafer.

[0022] The main body M1 has a plurality of through holes H1 formed therethrough in the thickness direction. The through holes H1 can be formed at any position on the suction surface of the first carrier C1. For example, the through holes H1 may be formed one for each of the first dies D1 held by the first carrier C1, in other words, the same number as the number of first dies D1 held by the first carrier C1. Alternatively, for example, the through holes H1 may be formed in a number corresponding to each of the first dies D1 held by the first carrier C1, in other words, a number greater than the number of first dies D1 held by the first carrier C1.

[0023] Although the number, size, spacing, and arrangement of the through holes H1 are not particularly limited, it is desirable to determine the number, size, and spacing so as to ensure the strength (rigidity) of the first carrier C1 so that deformation such as bending does not occur. For example, the diameter of the through holes H1 is 0.5 mm to 5.0 mm, and the spacing (the center-to-center distance between adjacent through holes H1) is 0.5 mm to 5.0 mm.

[0024] The insulating layer N1 is formed on the surface of the main body M1 and constitutes the adsorption surface of the first die D1 on the first carrier C1. The insulating layer N1 has no through-holes and covers the surface of the main body M1. The insulating layer N1 has a thickness of, for example, several tens of μm, which is sufficient to hold the first die D1 on the first carrier C1 by electrostatic adsorption.

[0025] The insulating layer N1 is made of a flexible and insulating material, such as polyimide or EVA (ethylene-vinyl acetate copolymer). In this embodiment, "the insulating layer N1 is flexible" means that the insulating layer N1 on the main body M1 has an elastic modulus of 2 GPa or less, preferably 0.5 GPa or less. In this embodiment, "the insulating layer N1 has insulating properties" means that the insulating layer N1 on the main body M1 has a breakdown voltage of 30 kV or more, preferably 40 kV or more.

[0026] 4 and 5, the second carrier C2 has the same configuration as the first carrier C1, as described above, and is configured by stacking a main body M2 and an insulating layer N2. The configurations of the main body M2 and the insulating layer N2 are the same as those of the main body M1 and the insulating layer N1, respectively. The main body M2 has a plurality of through holes H2 formed therethrough in the thickness direction. Because the second die D2 held by the second carrier C2 is a different type from the first die D1, the number, size, spacing, and arrangement of the through holes H2 may be different from those of the through holes H1 of the first carrier C1.

[0027] The first carrier C1 in this embodiment is configured as described above, and attracts and holds the first die D1 on its attracting surface by generating an electrostatic (Coulomb) force between the first carrier C1 and the first die D1. Similarly, the second carrier C2 attracts and holds the second die D2 on its attracting surface by generating an electrostatic (Coulomb) force between the second carrier C2 and the second die D2.

[0028] In addition, in this embodiment, the first carrier C1 has an insulating layer N1 made of a flexible material with a small elastic modulus. After the first die D1 is placed on the first carrier C1, a gap is formed between the first die D1 and the insulating layer N1 before the first die D1 is electrostatically attracted and held. Due to the flexibility of the insulating layer N1, when the first die D1 is electrostatically attracted to the insulating layer N1, air escapes between the first die D1 and the insulating layer N1, creating a pseudo-vacuum between the first die D1 and the insulating layer N1. As a result, in addition to the electrostatic attraction between the first carrier C1 and the first die D1, a vacuum attraction force is generated due to the pseudo-vacuum state, creating a strong holding state using both the electrostatic attraction and the vacuum attraction. Similarly, the second carrier C2 attracts and holds the second die D2 using both the electrostatic attraction and the vacuum attraction.

[0029] The method of holding the dies D1 and D2 by the carriers C1 and C2 is not limited to this embodiment. For example, the dies D1 and D2 may be pressed against the carriers C1 and C2 to be crimped. In this case, pressing the dies D1 and D2 against the carriers C1 and C2 removes air from between the dies D1 and D2 and the insulating layers N1 and N2, generating a vacuum suction force between the dies D1 and D2 and the insulating layers N1 and N2, thereby adsorbing and holding the dies D1 and D2 to the carriers C1 and C2. In this case, the material of the main bodies M1 and M2 of the carriers C1 and C2 does not need to be conductive and can be any material. Furthermore, the material of the insulating layers N1 and N2 does not need to be insulating and can be any material.

[0030] Alternatively, for example, adhesive sheets may be used for the insulating layers N1 and N2 of the carriers C1 and C2. In this case, adhesive force is generated between the dies D1 and D2 and the insulating layers N1 and N2, thereby holding the dies D1 and D2 on the carriers C1 and C2. Alternatively, thermal release sheets may be used for the insulating layers N1 and N2. The thermal release sheet is a sheet that has adhesive force at room temperature but is released when heated. As will be described later, when the dies D1 and D2 are bonded to the wafer W, the dies D1 and D2 are released from the carriers C1 and C2. At this time, heating the insulating layers N1 and N2 can make it easier to release the dies D1 and D2.

[0031] Next, a processing system 1 according to this embodiment will be described. In the processing system 1, a plurality of first dies D1 held in a first carrier C1 and a plurality of second dies D2 held in a second carrier C2 are mounted on a wafer W. FIG. 6 is a plan view showing an outline of the configuration of the processing system 1.

[0032] 6, the processing system 1 has a configuration in which a load / unload station 10 and a processing station 20 are integrally connected. In the load / unload station 10, for example, FOUPs Fw, Fc1, and Fc2, each capable of accommodating a plurality of wafers W, a plurality of first carriers C1, and a plurality of second carriers C2, are loaded and unloaded between the load / unload station 10 and the outside. The processing station 20 is equipped with various processing devices for performing a series of processes described below.

[0033] The loading / unloading station 10 is provided with a FOUP mounting table 30. In the illustrated example, a plurality of FOUPs, for example, two FOUPs Fw, one FOUP Fc1, and one FOUP Fc2, are placed on the FOUP mounting table 30, lined up in the Y-axis direction. Note that the number and arrangement of the FOUPs Fw, Fc1, and Fc2 placed on the FOUP mounting table 30 are not limited to those in this embodiment and can be determined arbitrarily.

[0034] A transfer device 40 is provided adjacent to the FOUP mounting table 30 on the positive side of the X-axis. The transfer device 40 is configured to be movable on a transfer path 41 extending in the Y-axis direction. The transfer device 40 also has, for example, two transfer arms 42, 42 that hold and transfer wafers W, first carriers C1, and second carriers C2. Each transfer arm 42 is configured to be movable horizontally, vertically, around a horizontal axis, and around a vertical axis. Note that the configuration of the transfer arms 42 is not limited to this embodiment and may have any configuration. The transfer device 40 is configured to transfer wafers W, first carriers C1, and second carriers C2 to and from the FOUPs Fw, Fc1, and Fc2 on the FOUP mounting table 30, a transition device 60 (described later), and a buffer device 61 (described later).

[0035] The processing station 20 is provided with, for example, three processing blocks 21 to 23. The first processing block 21, the second processing block 22, and the third processing block 23 are arranged in this order from the negative side to the positive side of the X axis.

[0036] The first processing block 21 is provided with a transfer device 50, a transition device 60, a buffer device 61, an inspection device 70, a wafer surface modification device 71, a die surface modification device 72, a wafer surface cleaning device 73, a die surface cleaning device 74, a wafer surface hydrophilization device 75, and a die surface hydrophilization device 76. The number and arrangement of these various processing devices are not limited to those in this embodiment, and can be determined as desired.

[0037] The transfer device 50 is configured to be movable on a transfer path 51 extending in the X-axis direction. The transfer device 50 also has, for example, two transfer arms 52, 52 that hold and transfer the wafer W, the first carrier C1, and the second carrier C2. Each transfer arm 52 is configured to be movable horizontally, vertically, around a horizontal axis, and around a vertical axis, and is configured to be able to transfer the wafer W, the first carrier C1, and the second carrier C2 to each of the devices 60, 61, 70 to 76 in the first processing block 21, a transition device 90 described later, and a buffer device 91 described later.

[0038] The transition device 60 and the buffer device 61 are disposed on the negative side of the X-axis of the transport device 50. The transition device 60 and the buffer device 61 are stacked in this order vertically from the top. Note that a plurality of buffer devices 61 may be stacked.

[0039] The transition device 60 transfers the wafers W, the first carrier C1, and the second carrier C2 between the transfer device 40 and the transfer device 50. The buffer device 61 temporarily stores the wafers W, the first carrier C1, and the second carrier C2.

[0040] The inspection device 70, wafer surface modification device 71, and die surface modification device 72 are arranged on the positive side of the Y axis of the transfer device 50. The inspection device 70, wafer surface modification device 71, and die surface modification device 72 are stacked vertically in this order from the top. The wafer surface modification device 71 and die surface modification device 72 are arranged side by side in this order from the negative side to the positive side of the X axis.

[0041] The inspection device 70 inspects a wafer W on which a first die D1 and a second die D2 are mounted. The inspection device 70 captures an image of the wafer W using, for example, an IR camera, and inspects for the presence or absence of voids between the surface Wa of the wafer W and the dies D1 and D2. The inspection device 70 may also function as a measurement unit and inspect the positions of the dies D1 and D2 bonded to the surface Wa of the wafer W. That is, the inspection device 70 may inspect the positions (misalignment) of at least two alignment marks on the first die D1 relative to at least two alignment marks on the wafer W.

[0042] The wafer surface modification device 71 uses plasma to modify the surface Wa of the wafer W. In the wafer surface modification device 71, for example, under a reduced pressure atmosphere, a process gas such as oxygen gas, nitrogen gas, hydrogen gas, or a mixture of these gases is excited to form plasma and ionized. The oxygen ions, nitrogen ions, or hydrogen ions are irradiated onto the surface Wa of the wafer W, and the surface Wa is subjected to plasma processing and modified.

[0043] Like the wafer surface modification device 71, the die surface modification device 72 also uses plasma to modify the surfaces D1a and D2a of the dies D1 and D2. In the die surface modification device 72, for example, under a reduced pressure atmosphere, a process gas such as oxygen gas, nitrogen gas, hydrogen gas, or a mixture of these is excited to form plasma and ionized. The oxygen ions, nitrogen ions, or hydrogen ions are irradiated onto the surfaces D1a and D2a of the dies D1 and D2 held by the carriers C1 and C2, and the surfaces D1a and D2a are plasma-processed and modified.

[0044] The wafer surface cleaning device 73, die surface cleaning device 74, wafer surface hydrophilization device 75, and die surface hydrophilization device 76 are arranged on the negative side of the Y axis of the transfer device 50. The wafer surface cleaning device 73 and die surface cleaning device 74, and the wafer surface hydrophilization device 75 and die surface hydrophilization device 76 are stacked vertically from the top in this order. The wafer surface cleaning device 73 and die surface cleaning device 74 are arranged side by side in this order from the negative side to the positive side of the X axis. The wafer surface hydrophilization device 75 and die surface hydrophilization device 76 are arranged side by side in this order from the negative side to the positive side of the X axis.

[0045] The wafer surface cleaning device 73 cleans the front surface Wa of the wafer W. In the wafer surface cleaning device 73, while the wafer W held by, for example, a spin chuck is rotated, a cleaning liquid such as pure water, DI water, DHF, or IPA is supplied onto the wafer W. The supplied cleaning liquid then spreads over the front surface Wa of the wafer W, cleaning the front surface Wa.

[0046] The die surface cleaning device 74 also cleans the surfaces D1a and D2a of the dies D1 and D2, similarly to the wafer surface cleaning device 73. In the die surface cleaning device 74, a cleaning liquid is supplied onto the dies D1 and D2 held by the carriers C1 and C2 while the carriers C1 and C2 held by the chucks are rotated. The supplied cleaning liquid then spreads over the surfaces D1a and D2a of the dies D1 and D2, cleaning the surfaces D1a and D2a.

[0047] The wafer surface hydrophilization device 75 hydrophilizes and rinses the surface Wa of the wafer W. In the wafer surface hydrophilization device 75, pure water is supplied onto the wafer W while the wafer W held by, for example, a spin chuck is being rotated. The supplied pure water then spreads over the surface Wa of the wafer W, hydrophilizing the surface Wa. The surface Wa is also rinsed with the pure water.

[0048] Like the wafer surface hydrophilization device 75, the die surface hydrophilization device 76 hydrophilizes and rinses the surfaces D1a and D2a of the dies D1 and D2. In the die surface hydrophilization device 76, a cleaning liquid is supplied onto the dies D1 and D2 held by the carriers C1 and C2 while the carriers C1 and C2 held by, for example, spin chucks are rotated. The supplied cleaning liquid then spreads over the surfaces D1a and D2a of the dies D1 and D2, hydrophilizing the surfaces D1a and D2a. The surfaces D1a and D2a are also rinsed with pure water.

[0049] The second processing block 22 is provided with a transport device 80, a transition device 90, a buffer device 91, and a first bonding device 100. The number and arrangement of these various processing devices are not limited to those in this embodiment, and can be determined arbitrarily.

[0050] The transfer device 80 is configured to be movable on a transfer path 81 extending in the X-axis direction. The transfer device 80 also has, for example, two transfer arms 82, 82 that hold and transfer the wafer W, the first carrier C1, and the second carrier C2. Each transfer arm 82 is configured to be movable horizontally, vertically, around a horizontal axis, and around a vertical axis, and is configured to be able to transfer the wafer W, the first carrier C1, and the second carrier C2 to each of the devices 90, 91, 100 in the second processing block 22, a transition device 120 described later, and a buffer device 121 described later.

[0051] The transition device 90 and the buffer device 91 are disposed on the negative side of the X-axis of the transport device 80. The transition device 90 and the buffer device 91 are stacked in this order vertically from the top. Note that a plurality of buffer devices 91 may be stacked.

[0052] The transition device 90 transfers the wafers W, the first carrier C1, and the second carrier C2 between the transfer device 50 and the transfer device 80. The buffer device 91 temporarily stores the wafers W, the first carrier C1, and the second carrier C2.

[0053] For example, two first bonding apparatuses 100 are arranged in the positive direction of the Y axis of the transfer apparatus 80, and two are arranged in the negative direction of the Y axis of the transfer apparatus 80. The first bonding apparatus 100 bonds a first die D1 (high-precision die) held by a first carrier C1 to a wafer W. That is, high bonding precision is required of the first bonding apparatus 100 (hereinafter, such a bonding apparatus may be referred to as a "high-precision bonding apparatus"). The configuration of the first bonding apparatus 100 will be described in detail later.

[0054] The third processing block 23 is provided with a transport device 110, a transition device 120, a buffer device 121, and a second bonding device 130. The number and arrangement of these various processing devices are not limited to this embodiment and can be determined arbitrarily.

[0055] The transfer device 110 is configured to be movable on a transfer path 111 extending in the X-axis direction. The transfer device 110 also has, for example, two transfer arms 112, 112 that hold and transfer the wafer W and the second carrier C2. Each transfer arm 112 is configured to be movable horizontally, vertically, around a horizontal axis, and around a vertical axis, and is configured to be able to transfer the wafer W and the second carrier C2 to each of the devices 120, 121, and 130 in the third processing block 23.

[0056] The transition device 120 and the buffer device 121 are disposed on the negative side of the X-axis of the transport device 110. The transition device 120 and the buffer device 121 are stacked in this order vertically from the top. Note that a plurality of buffer devices 121 may be stacked.

[0057] The transition device 120 transfers the wafers W and the second carrier C2 between the transfer device 80 and the transfer device 110. The buffer device 121 temporarily stores the wafers W, the first carrier C1, and the second carrier C2.

[0058] For example, two second bonding devices 130 are arranged in the positive direction of the Y axis of the transfer device 110, and two are arranged in the negative direction of the Y axis of the transfer device 110. The second bonding device 130 bonds a second die D2 (low-precision die) held by a second carrier C2 to a wafer W. In other words, the bonding precision required of the second bonding device 130 is low (hereinafter, such a bonding device may be referred to as a "low-precision bonding device"). The configuration of the second bonding device 130 will be described in detail later.

[0059] The processing system 1 described above is provided with at least one control unit, i.e., a control device 140. The control device 140 processes computer-executable instructions that cause the processing system 1 to perform the various steps described in this disclosure. The control device 140 may be configured to control each element of the processing system 1 to perform the various steps described herein. In one embodiment, part or all of the control device 140 may be included in the processing system 1. The control device 140 may include a processing unit, a storage unit, and a communication interface. The control device 140 may be implemented, for example, by a computer. The processing unit may be configured to read from the storage unit a program that provides logic or routines that enable various control operations and execute the read program to perform various control operations. This program may be stored in the storage unit in advance or may be acquired via a medium when needed. The acquired program is stored in the storage unit and read from the storage unit by the processing unit for execution. The medium may be various computer-readable storage media or a communication line connected to the communication interface. The storage medium may be temporary or non-temporary. The processing unit may be a CPU (Central Processing Unit) or one or more circuits. The storage unit may include a RAM (Random Access Memory), a ROM (Read Only Memory), a HDD (Hard Disk Drive), an SSD (Solid State Drive), or a combination thereof. The communication interface may communicate with the processing system 1 via a communication line such as a LAN (Local Area Network).

[0060] Although the processing system 1 according to this embodiment is configured as described above, other processing devices may be further arranged in the processing system 1 depending on the purpose, and some processing devices may be arranged outside the processing system 1 depending on the purpose.

[0061] Next, a description will be given of a die-ion wafer manufacturing process performed in the processing system 1 configured as described above. Fig. 7 is a flow diagram showing the main steps of the die-ion wafer manufacturing process. Fig. 8 is an explanatory diagram schematically showing some steps of the die-ion wafer manufacturing process.

[0062] First, FOUPs Fw, Fc1, and Fc2, each containing a plurality of wafers W, a first carrier C1, and a second carrier C2, are placed on the FOUP mounting table 30 of the carry-in / out station 10. At this time, a plurality of first dies D1 are held in the first carrier C1 as shown in FIG. 8( a), and a plurality of second dies D2 are held in the second carrier C2 as shown in FIG. 8( b). The first carrier C1 contains the first dies D1 facing upward, and the second carrier C2 contains the second dies D2 facing upward. As described above, the first dies D1 are high-precision dies, and the second dies D2 are low-precision dies.

[0063] Next, the transfer device 40 removes the wafer W from the FOUP Fw and transfers it to the transition device 60. The wafer W transferred to the transition device 60 is transferred by the transfer device 50 to the wafer front surface cleaning device 73. In the wafer front surface cleaning device 73, the front surface Wa of the wafer W is cleaned with, for example, a cleaning liquid (St1 in FIG. 7 ).

[0064] Next, the wafer W is transferred by the transfer device 50 to the wafer surface modification device 71. In the wafer surface modification device 71, for example, plasma processing is performed under a reduced pressure atmosphere, and the surface Wa of the wafer W is modified (St2 in FIG. 7).

[0065] Next, the wafer W is transferred by the transfer device 50 to the wafer surface hydrophilization device 75. In the wafer surface hydrophilization device 75, for example, pure water is used to attach hydroxyl groups (silanol groups) to the surface Wa of the wafer W modified in St2, thereby hydrophilizing the surface Wa. The surface Wa is also rinsed with the pure water (St3 in FIG. 7).

[0066] Next, the wafer W is transferred by the transfer device 50 to the transition device 90, and further transferred by the transfer device 80 to the first bonding device 100. Note that if the bonding process has already been performed in the four first bonding devices 100, the wafer W is transferred to the buffer device 91 and temporarily stored in the buffer device 91.

[0067] While the wafer W is undergoing the above-described processes St1 to St3, the first die D1 held in the first carrier C1 is also processed. First, the transfer device 40 removes the first carrier C1 from the FOUP Fc1 and transfers it to the transition device 60. After being transferred to the transition device 60, the transfer device 50 transfers the first carrier C1 to the die surface cleaning device 74. In the die surface cleaning device 74, the surface D1a of the first die D1 is cleaned with, for example, a cleaning liquid (St4 in FIG. 7 ).

[0068] Next, the first carrier C1 is transported by the transport device 50 to the die surface modification device 72. In the die surface modification device 72, for example, plasma processing is performed under a reduced pressure atmosphere to modify the surface D1a of the first die D1 (St5 in FIG. 7).

[0069] Next, the first carrier C1 is transported by the transport device 50 to the die surface hydrophilization device 76. In the die surface hydrophilization device 76, hydroxyl groups (silanol groups) are attached to the surface D1a of the first die D1 modified in St5 using, for example, pure water, thereby hydrophilizing the surface D1a. The surface D1a is also rinsed with the pure water (St6 in FIG. 7).

[0070] Next, the first carrier C1 is transported by the transport device 50 to the transition device 90, and further transported by the transport device 80 to the first joining device 100. Note that if the joining process has already been performed in the four first joining devices 100, the first carrier C1 is transported to the buffer device 91 and temporarily stored therein.

[0071] While the wafer W is subjected to the processes St1 to St3 described above and the first die D1 is subjected to the processes St4 to St6 described above, the second die D2 held in the second carrier C2 is processed. First, the transfer device 40 removes the second carrier C2 from the FOUP Fc2 and transfers it to the transition device 60. The second carrier C2 transferred to the transition device 60 is then transferred by the transfer device 50 to the die surface cleaning device 74. In the die surface cleaning device 74, the surface D2a of the second die D2 is cleaned with, for example, a cleaning liquid (St7 in FIG. 7 ).

[0072] Next, the second carrier C2 is transported by the transport device 50 to the die surface modification device 72. In the die surface modification device 72, for example, plasma processing is performed under a reduced pressure atmosphere, and the surface D2a of the second die D2 is modified (St8 in FIG. 7).

[0073] Next, the second carrier C2 is transported by the transport device 50 to the die surface hydrophilization device 76. In the die surface hydrophilization device 76, hydroxyl groups (silanol groups) are attached to the surface D2a of the second die D2 modified in St8 using, for example, pure water, thereby hydrophilizing the surface D2a. The surface D2a is also rinsed with the pure water (St9 in FIG. 7).

[0074] Next, the second carrier C2 is transported by the transport device 50 to the transition device 90, then by the transport device 80 to the transition device 120, and then by the transport device 110 to the second joining device 130. Note that if the joining process has already been performed in the four second joining devices 130, the second carrier C2 is transported to the buffer device 121 and temporarily stored therein.

[0075] The wafer W that has been subjected to the processes of St1 to St3 described above is transferred to the first bonding apparatus 100, and the first carrier C1 that has been subjected to the processes of St4 to St6 described above is transferred to the first bonding apparatus 100. In the first bonding apparatus 100, the first die D1 held by the first carrier C1 is removed from the first carrier C1 and picked up. Furthermore, as shown in FIG. 8( c), the surface D1 a of the picked-up first die D1 is brought into contact with the surface Wa of the wafer W, and the first die D1 is pressed against the wafer W, thereby bonding the first die D1 to the wafer W (St10 in FIG. 7 ).

[0076] In St10, because the surface Wa of the wafer W and the surface D1a of the first die D1 have been modified in St2 and St5, respectively, van der Waals forces (intermolecular forces) are generated between the surfaces Wa and D1a, thereby bonding the surfaces Wa and D1a together. Furthermore, because the surface Wa of the wafer W and the surface D1a of the first die D1 have been hydrophilized in St3 and St6, respectively, the hydrophilic groups between the surfaces Wa and D1a form hydrogen bonds (intermolecular forces), thereby firmly bonding the surfaces Wa and D1a together.

[0077] As described above, the first bonding apparatus 100 is a high-precision bonding apparatus. The first die D1 includes a device layer E on which circuits and the like are formed, and high bonding precision is required. In St10, the first die D1 can be bonded to a desired position on the wafer W with high precision.

[0078] In St10, when all the first dies D1 held in the first carrier C1 are bonded to the wafers W, the first carrier C1 is transferred to the transition device 90 by the transfer device 80, transferred to the transition device 60 by the transfer device 50, and further transferred to the FOUP Fc1 by the transfer device 40. On the other hand, if the first die D1 remains in the first carrier C1 after St10, the first carrier C1 may remain in the first bonding device 100, and the first die D1 may be bonded to the subsequent wafer W. Alternatively, if the first die D1 remains in the first carrier C1 after St10, the first carrier C1 may be transferred to the buffer device 91 by the transfer device 80 and temporarily stored in the buffer device 91.

[0079] On the other hand, in St10, when the first die D1 is bonded to all of the desired positions on the wafer W, the wafer W is transferred by the transfer device 80 to the transition device 120, and then transferred by the transfer device 110 to the second bonding device 130. Note that if the bonding process has already been performed in the four second bonding devices 130, the wafer W is transferred to the buffer device 121 and temporarily stored in the buffer device 121.

[0080] The second carrier C2, which has been subjected to the processes of St7 to St9 described above, is transferred to the second bonding device 130. In the second bonding device 130, the second die D2 held by the second carrier C2 is removed from the second carrier C2 and picked up. Furthermore, as shown in FIG. 8D, the surface D2a of the picked-up second die D2 is brought into contact with the surface Wa of the wafer W, and the second die D2 is pressed against the wafer W, thereby bonding the second die D2 to the wafer W (St11 in FIG. 7).

[0081] In St11, because the surface Wa of the wafer W and the surface D2a of the second die D2 have been modified in St2 and St8, respectively, van der Waals forces (intermolecular forces) are generated between the surfaces Wa and D2a, thereby bonding the surfaces Wa and D2a together. Furthermore, because the surface Wa of the wafer W and the surface D2a of the second die D2 have been hydrophilized in St3 and St9, respectively, the hydrophilic groups between the surfaces Wa and D2a form hydrogen bonds (intermolecular forces), thereby firmly bonding the surfaces Wa and D2a together.

[0082] As described above, the second bonding apparatus 130 is a low-precision bonding apparatus. Since the second die D2 does not have a circuit or the like formed thereon and therefore requires low bonding precision, the second die D2 is bonded to a desired position on the wafer W with low precision in St11. In this case, since low bonding precision is sufficient, the time required for St11 can be shortened.

[0083] In St11, when all the second dies D2 held in the second carrier C2 are bonded to the wafers W, the second carrier C2 is transferred to the transition device 120 by the transfer device 110, transferred to the transition device 90 by the transfer device 80, transferred to the transition device 60 by the transfer device 50, and further transferred to the FOUP Fc2 by the transfer device 40. On the other hand, if the second die D2 remains in the second carrier C2 after St11, the second carrier C2 may remain in the second bonding device 130, and the second die D2 may be bonded to the subsequent wafer W. Alternatively, if the second die D2 remains in the second carrier C2 after St11, the second carrier C2 may be transferred to the buffer device 121 by the transfer device 110 and temporarily stored in the buffer device 121.

[0084] 1 , the wafer W is transferred to the transition device 120 by the transfer device 110, transferred to the transition device 90 by the transfer device 80, and transferred to the inspection device 70 by the transfer device 50. The inspection device 70, for example, takes an image of the wafer W and inspects for the presence or absence of voids between the front surface Wa of the wafer W and the dies D1 and D2 (St12 in FIG. 7 ). Note that the inspection device 70 may inspect the positions of the dies D1 and D2 bonded to the front surface Wa of the wafer W.

[0085] Next, the wafer W is transferred to the transition device 60 by the transfer device 50, and further transferred to the FOUP Fw by the transfer device 40. In this way, a series of die-on wafer manufacturing processes is completed.

[0086] The FOUPs into which the wafers W, the first carrier C1, and the second carrier C2 are recovered do not necessarily have to be the same FOUPs that housed the wafers W, the first carrier C1, and the second carrier C2, respectively, when they were loaded in. That is, for example, the FOUPs that housed the wafers W, the first carrier C1, and the second carrier C2, respectively, may each carry out different members, or new FOUPs or the like may be loaded into the processing system 1 for carrying out the wafers W, the first carrier C1, and the second carrier C2.

[0087] According to the above embodiment, the processing system 1 includes the first bonding apparatus 100, which is a high-precision bonding apparatus, and the second bonding apparatus 130, which is a low-precision bonding apparatus. The first die D1 includes a device layer E on which circuits and the like are formed, and therefore requires high bonding accuracy. The first bonding apparatus 100 can bond the first die D1 to a desired position on the wafer W with high accuracy. The second die D2 does not include circuits and therefore requires low bonding accuracy. The second bonding apparatus 130 bonds the second die D2 to a desired position on the wafer W with low accuracy. Since low bonding accuracy is sufficient, the time required for bonding the second die D2 to the wafer W can be shortened. Therefore, the throughput of the die-on-wafer manufacturing process can be improved while appropriately mounting multiple types of first dies D1 and second dies D2 on the wafer W with the required bonding accuracy.

[0088] In this embodiment, the second bonding device 130 is a low-precision bonding device, and as described below, the device configuration can be reduced to lower precision specifications compared to a high-precision bonding device. As a result, the device cost of the second bonding device 130 can be reduced, and the device cost of the processing system 1 can be reduced.

[0089] Next, the configurations of the above-described first welding apparatus 100 and second welding apparatus 130 will be described. The required welding precision differs between the first welding apparatus 100 and the second welding apparatus 130, so first the configuration of the first welding apparatus 100 will be described, and then the configuration of the second welding apparatus 130 will be described, focusing on the differences from the first welding apparatus 100. Figure 9 is a perspective view showing an outline of the configuration of the first welding apparatus 100.

[0090] As shown in Fig. 9, the first bonding apparatus 100 has a stage 200. The stage 200 is divided into a de-bonding area 201 and a bonding area 202. The de-bonding area 201 and the bonding area 202 are arranged side by side in the horizontal direction (X-axis direction). In the de-bonding area 201, the first die D1 held by the first carrier C1 is de-bonded from the first carrier C1 and picked up. In the bonding area 202, the first die D1 picked up in the de-bonding area 201 is bonded to a wafer W.

[0091] A carrier holding unit 210 is provided in the separation area 201. The carrier holding unit 210 has a holding surface of the first carrier C1 on its upper surface, and holds the first carrier C1 holding the first die D1 in a state where the surface D1a of the first die D1 faces upward.

[0092] A moving mechanism 211 is provided below the carrier holding unit 210. The moving mechanism 211 is configured to allow the carrier holding unit 210 to move in horizontal directions (X-axis direction and Y-axis direction). The moving mechanism 211 is also configured to allow the carrier holding unit 210 to rotate around a vertical axis. The driving unit of the moving mechanism 211 is not particularly limited, but a linear motor, for example, is used.

[0093] 10 , the carrier holding part 210 is provided with an air supply part 220. The air supply part 220 has an air cap 221 and an air supply source 222. The air cap 221 is provided inside the carrier holding part 210 and supplies air to the through-hole H1 of the first carrier C1 held by the carrier holding part 210. The air supply source 222 stores air therein and supplies air to the air cap 221.

[0094] The air supply unit 220 supplies air from the air cap 221 to the back surface D1b of the first die D1 on the first carrier C1 through the through hole H1. This applies air pressure to the first die D1, thereby reducing the adhesion between the first die D1 and the insulating layer N1 on the suction surface of the first carrier C1. This allows the first die D1 to be detached from the suction surface of the first carrier C1.

[0095] 9 , a pickup unit 230 is provided above the carrier holding unit 210. The pickup unit 230 picks up and transports the first die D1 from the first carrier C1. Specifically, the pickup unit 230 holds the first die D1 that has been lifted up (lifted up) due to a decrease in adhesion between the first die D1 and the insulating layer N1 caused by the supply of air from the air supply unit 220, removes the first die D1 from the first carrier C1, and then transports the held first die D1 from the removal area 201 to the bonding area 202.

[0096] The pickup unit 230 may hold the first die D1 in any manner. However, since the surface D1a of the first die D1 held by the pickup unit 230 is the device surface of the device layer E, the pickup unit 230 must hold the first die D1 without damaging the device surface. For example, the pickup unit 230 may be a collet chuck, or a non-contact chuck that can hold the first die D1 from above without contact using the Bernoulli effect or ultrasonic squeeze effect. Alternatively, instead of holding the surface D1a of the first die D1, the pickup unit 230 may be configured to hold the first die D1 by clamping the side surface of the first die D1 that has been raised by, for example, supplying air.

[0097] The pickup unit 230 is supported by a moving mechanism 231. The moving mechanism 231 supports and moves the pickup unit 230. The moving mechanism 231 has an arm 232, a rail 233, a support member 234, and a drive unit 235. The tip of the arm 232 supports the pickup unit 230, and the base end is attached to the rail 233. The rail 233 extends in the X-axis direction and is supported by the support member 234. The drive unit 235 moves the pickup unit 230 in the X-axis direction (a direction along the rail 233), the Y-axis direction, and the vertical direction. The drive unit 235 also rotates the pickup unit 230 around a horizontal axis (around the X-axis), allowing the pickup unit 230 to flip the front and back surfaces of the held first die D1. The drive unit 235 also rotates the pickup unit 230 around a vertical axis, allowing the position of the held first die D1 in the rotational direction (θ-axis direction). The drive source of the drive unit 235 is not particularly limited, but may be, for example, a linear motor.

[0098] A first imaging unit 240 is provided above the carrier holding unit 210. For example, a camera is used as the first imaging unit 240. The first imaging unit 240 captures images of at least two alignment marks of the first die D1 on the first carrier C1 held by the carrier holding unit 210 from above. The first imaging unit 240 is fixed, and the carrier holding unit 210 is moved to capture images of the at least two alignment marks of the first die D1. The captured images are output to the control device 140, which measures the position of the first die D1 on the first carrier C1. The control device 140 controls the moving mechanism 231 based on the measured position of the first die D1, and adjusts the position of the pickup unit 230 so that the pickup unit 230 properly picks up the first die D1.

[0099] A second imaging unit 241 is provided on the positive X-axis side of the carrier holding unit 210, below the pickup unit 230. The second imaging unit 241 may be, for example, a camera. The second imaging unit 241 captures images of at least two alignment marks on the first die D1 held by the pickup unit 230 from below. The second imaging unit 241 is fixed, and the pickup unit 230 is moved to capture images of the at least two alignment marks on the first die D1. The captured images are output to the control device 140, which measures the position of the first die D1 held by the pickup unit 230. The control device 140 controls the moving mechanism 231 based on the measured position of the first die D1, and adjusts the position of the pickup unit 230 so that the pickup unit 230 can transfer the first die D1 to an appropriate position of a bond head 261, which will be described later.

[0100] A wafer holding part 250 serving as a substrate holding part is provided in the bonding region 202. The wafer holding part 250 has a holding surface for the wafer W on its upper surface, and holds the wafer W with the front surface Wa, which is the mounting surface of the first die D1, facing upward. The wafer holding part 250 is supported by a support plate 251.

[0101] A moving mechanism 252 is provided below the support plate 251. The moving mechanism 252 is configured to move the wafer holding part 250 and the support plate 251 in horizontal directions (X-axis direction and Y-axis direction). The moving mechanism 252 is also configured to rotate the wafer holding part 250 and the support plate 251 around a vertical axis. The drive part of the moving mechanism 252 is not particularly limited, but a linear motor, for example, is used.

[0102] A head unit 260 is provided above the wafer holding unit 250. The head unit 260 receives and holds the first die D1 from the pickup unit 230, transports the held first die D1 to the wafer W, and bonds the first die D1 to the wafer W. As shown in FIG. 11 , the head unit 260 has a bond head 261, a glass plate 262, and a base 263.

[0103] The bond head 261 receives from above and holds the first die D1 held by the pickup unit 230. The holding surface (lower surface) of the bond head 261 for the first die D1 has a convex shape with the center protruding downward.

[0104] The bond head 261 may hold the first die D1 in any manner. Because the back surface D1b of the first die D1 held by the bond head 261 is the silicon layer S1, the bond head 261 does not necessarily need to be configured with a non-contact chuck or the like, unlike the pickup unit 230. For example, a vacuum chuck may be used for the bond head 261, and the bond head 261 may suck and hold the first die D1 by vacuuming it using a vacuum mechanism (not shown).

[0105] The glass plate 262 is larger than the bond head 261 and holds the entire upper surface of the bond head 261. The method of holding the bond head 261 by the glass plate 262 is arbitrary. For example, the glass plate 262 may hold the bond head 261 by suction using a vacuum mechanism (not shown). In such a case, the bond head 261 is configured to be detachable from the glass plate 262 and to be replaceable.

[0106] The base 263 supports the outer periphery of the upper surface of the glass plate 262. The base 263 has an annular shape, and an opening 263a is formed in the center of the base 263. The opening 263a is formed to be at least larger than the bond head 261, that is, the bond head 261 held by the glass plate 262 is disposed inside the opening 263a in a plan view.

[0107] A plurality of, for example, four alignment marks 264 are provided on the lower surface of the glass plate 262. The four alignment marks 264 are arranged outside the bond head 261 held by the glass plate 262 and inside the opening 263 a in plan view.

[0108] As shown in Figure 9, a third imaging unit 280 (described later) is provided below the head unit 260, and a fourth imaging unit 281 (described later) and a fifth imaging unit 282 (described later) are provided above the head unit 260. These imaging units 280, 281, and 282 move slightly over time, so they need to be aligned at the desired timing. This alignment is performed by the imaging units 280, 281, and 282 capturing images of the four alignment marks 264. For this reason, the glass plate 262 is made transparent.

[0109] Furthermore, a fourth imaging unit 281 and a fifth imaging unit 282, which will be described later, each capture an image of the wafer W held by the wafer holding unit 250 below the head unit 260. The fourth imaging unit 281 and the fifth imaging unit 282 each capture an image of the wafer W through the glass plate 262, outside the bond head 261 and inside the opening 263 a in a plan view. For this reason, the glass plate 262 is configured to be transparent.

[0110] The head unit 260 is supported by a moving mechanism 270. The moving mechanism 270 supports and moves the head unit 260. The moving mechanism 270 has an arm 271, a rail 272, a support member 273, and a drive unit 274. The tip of the arm 271 supports the base 263 of the head unit 260, and the base end is attached to the rail 272. The rail 272 extends in the X-axis direction and is supported by the support member 273. The drive unit 274 moves the head unit 260 in the X-axis direction (the direction along the rail 272), the Y-axis direction, and the vertical direction. The drive unit 274 can also rotate the head unit 260 around the vertical axis to adjust the position of the held first die D1 in the rotational direction (θ-axis direction). The drive source of the drive unit 274 is not particularly limited, but a linear motor, for example, can be used.

[0111] A third imaging unit 280 is provided below the head unit 260. The third imaging unit 280 may be, for example, a camera. With the wafer holder 250 retracted to a position where it does not overlap the third imaging unit 280 in a plan view, the third imaging unit 280 images at least two, preferably three or more, alignment marks of the first die D1 held by the bond head 261 from below. The third imaging unit 280 is fixed, and the bond head 261 is moved to capture images of the at least two alignment marks of the first die D1. The captured images are output to the control device 140, which measures the position of the first die D1 held by the bond head 261. The control device 140 calculates the position of the first die D1, for example, by calculating the midpoint of the positions of the at least two alignment marks, and then calculates the center of gravity of the first die D1. In this case, the center of gravity position of the first die D1 may be calculated based on the positions of three or more alignment marks using a method similar to the local correction model using the least squares method represented by equations (1) and (2) described below.

[0112] A fourth imaging unit 281 is provided above the head unit 260, specifically above the opening 263a of the base 263. The fourth imaging unit 281 may be, for example, a camera. The fourth imaging unit 281 captures images of at least two, preferably three or more, alignment marks on the wafer W held by the wafer holder 250 from above. At least one of the fourth imaging unit 281 and the wafer holder 250 is moved to capture images of the at least two alignment marks on the wafer W. The captured images are output to the control device 140, which measures the position of the wafer W held by the wafer holder 250. The control device 140 calculates the center of gravity of the wafer W by calculating, for example, the midpoint between the positions of the at least two alignment marks as the position of the wafer W. In this case, the center of gravity of the wafer W may be calculated based on the positions of the three or more alignment marks using a method similar to a local correction model using the least squares method represented by equations (1) and (2) described below.

[0113] The control device 140 adjusts the position of the first die D1 relative to the wafer W based on the position of the first die D1 measured using the third imaging unit 280 and the position of the wafer W measured using the fourth imaging unit 281. That is, the control device 140 adjusts the position of the first die D1 relative to the wafer W based on the positions of the centers of gravity of the first die D1 and the wafer W. For example, when using the least squares method as described above, the control device 140 adjusts the position of the first die D1 relative to the wafer W so as to minimize the amount of misalignment of the measured alignment marks. Specifically, the control device 140 controls at least one of the moving mechanism 252 and the moving mechanism 270 to move at least one of the first die D1 and the wafer W, thereby adjusting the position of the first die D1 relative to the wafer W so that the first die D1 is bonded to an appropriate position relative to the wafer W.

[0114] A fifth imaging unit 282 serving as a measurement unit is provided above the head unit 260, specifically above the opening 263a of the base 263. The fifth imaging unit 282 is, for example, an IR camera. The fifth imaging unit 282 captures an image of the first die D1 mounted on the wafer W from above. The captured image is output to the control device 140, which measures the position of the first die D1 relative to the wafer W. That is, the control device 140 measures the positions (misalignments) of at least two, preferably three or more, alignment marks on the first die D1 relative to at least two, preferably three or more, alignment marks on the wafer W. The control device 140 then inspects the bonding accuracy of the first die D1 relative to the wafer W.

[0115] The fourth imaging unit 281 and the fifth imaging unit 282 may each be supported by an arm 271 of the moving mechanism 270 via a support member (not shown), and may be configured to be movable in the horizontal direction (X-axis direction and Y-axis direction) by the moving mechanism 270. Alternatively, the fourth imaging unit 281 and the fifth imaging unit 282 may each be supported by a moving mechanism (not shown) separate from the moving mechanism 270, and may be configured to be movable in the horizontal direction (X-axis direction and Y-axis direction) by the moving mechanism.

[0116] Next, a method for bonding the first die D1 to the wafer W using the first bonding apparatus 100 configured as described above will be described. Specifically, as shown in FIG. 12 , the first die D1 is bonded to a planned bonding position Wd on the wafer W. At this time, the first die D1 is bonded to the wafer W using a plurality of alignment marks Wm (e.g., at least two) formed at the planned bonding position Wd on the wafer W and a plurality of alignment marks D1m (e.g., at least two) formed on the first die D1 so that the amount of positional misalignment between the alignment marks Wm and D1m is minimized. Note that four alignment marks Wm and D1m are illustrated in FIG. 12 .

[0117] 13 , a wafer W and a first carrier C1 holding a plurality of first dies D1 are transported to the first bonding apparatus 100. The wafer W is held by the wafer holding part 250 with its front surface Wa facing upward. At this time, the wafer holding part 250 is disposed in a position that does not overlap with the third imaging part 280 in a plan view, that is, on the positive X-axis side of the third imaging part 280 in the illustrated example. The first carrier C1 is held by the carrier holding part 210 with the front surface D1a of the first die D1 facing upward.

[0118] Next, the first imaging unit 240 captures an image of at least two alignment marks D1m of one of the first dies D1 to be joined, among the first dies D1 on the first carrier C1 held by the carrier holding unit 210. The captured image is output to the control device 140, which measures the position of the first die D1 on the first carrier C1. The control device 140 controls the moving mechanism 231 based on the measured position of the first die D1, moves the carrier holding unit 210, and adjusts the position of the pickup unit 230 and the first die D1 to be picked up by the pickup unit 230.

[0119] Next, the air supply unit 220 selectively supplies air to the through hole H1 corresponding to one of the first dies D1 to be joined, causing the insulating layer N1 to expand, thereby pushing up the one of the first dies D1 from below and lifting it up.

[0120] 14, the pickup unit 230 is lowered to hold the surface D1a of one of the first dies D1 that has been lifted up. The pickup unit 230 may hold the first die D1 before lifting the first die D1 with air, or may hold the first die D1 at the same time as lifting the first die D1 with air. Subsequently, the pickup unit 230 is raised to pick up the first die D1 from the first carrier C1.

[0121] 15 , the pickup unit 230 is moved in the positive direction of the X-axis to above the second imaging unit 241. The second imaging unit 241 captures images of at least two alignment marks D1m of the first die D1 held by the pickup unit 230 from below. The captured images are output to the control device 140, which measures the position of the first die D1 held by the pickup unit 230. The control device 140 controls the moving mechanism 231 based on the measured position of the first die D1, and adjusts the position of the pickup unit 230 so that the first die D1 is delivered to the center of the bond head 261.

[0122] Next, the pickup unit 230 is rotated around the horizontal axis (around the X axis), thereby inverting the front and back surfaces of the first die D1, so that the back surface D1b of the first die D1 held by the pickup unit 230 faces upward.

[0123] Next, the pickup unit 230 is further moved in the positive direction of the X-axis until it is below the bond head 261. Subsequently, the pickup unit 230 transfers the first die D1 to the bond head 261. The first die D1 is placed in the center of the bond head 261.

[0124] Next, as shown in FIG. 16 , the third imaging unit 280 captures images of at least two, preferably three or more, alignment marks D1m on the first die D1 held by the bond head 261 from below. The captured images are output to the control device 140, which measures the position of the first die D1 held by the bond head 261. As described above, the control device 140 calculates the position of the center of gravity of the first die D1 by calculating, for example, the midpoint of the positions of at least two alignment marks D1m as the position of the first die D1. In this case, the center of gravity of the first die D1 may be calculated based on the positions of the three or more alignment marks D1m using a method similar to the local correction model using the least squares method represented by equations (1) and (2) described below.

[0125] Next, as shown in FIG. 17 , the wafer holding unit 250 is moved below the bond head 261 (fourth imaging unit 281). Subsequently, the fourth imaging unit 281 captures images of at least two, preferably three or more, alignment marks Wm on the wafer W held by the wafer holding unit 250 from above. The captured images are output to the control device 140, which measures the position of the wafer W held on the wafer W. As described above, the control device 140 calculates the center of gravity of the wafer W by calculating, for example, the midpoint between the positions of at least two alignment marks Wm as the position of the wafer W. In this case, the center of gravity of the wafer W may be calculated based on the positions of the three or more alignment marks Wm using a method similar to a local correction model using the least squares method represented by equations (1) and (2) described below.

[0126] Based on the position of the first die D1 measured using the third imaging unit 280 and the position of the wafer W measured using the fourth imaging unit 281, the control device 140 adjusts the position of the first die D1 relative to the wafer W so that the first die D1 is bonded to an appropriate position relative to the wafer W. Specifically, when the least squares method is used as described above, the control device 140 adjusts the position of the first die D1 relative to the wafer W so that the amount of misalignment between the measured alignment marks Wm and D1m is minimized.

[0127] 18, the bond head 261 is lowered, the surface D1a of the first die D1 held by the bond head 261 is brought into contact with the surface Wa of the wafer W, and the first die D1 is pressed to bond the first die D1 to the wafer W. Subsequently, the bond head 261 is raised.

[0128] As described above, the holding surface (lower surface) of the bond head 261 for holding the first die D1 has a convex shape with a central portion protruding downward. For example, if the first die D1 is held off-center on the bond head 261, pressing the first die D1 against the wafer W will result in poor surface pressure accuracy of the first die D1 relative to the wafer W, resulting in voids between the wafer W and the first die D1. In this regard, in this embodiment, the position of the pickup unit 230 is adjusted based on the image of the first die D1 captured by the second imaging unit 241, so that the first die D1 is appropriately held at the center of the bond head 261. Therefore, when bonding the first die D1 to the wafer W, the surface pressure accuracy of the first die D1 relative to the wafer W can be improved, and voids can be suppressed.

[0129] In this manner, the bonding operation of the first die D1 to the wafer W is performed. The bonding operation of the first die D1 is performed independently and continuously for each of the plurality of first dies D1 held by suction on the first carrier C1.

[0130] After the first die D1 has been bonded to all of the desired positions on the wafer W, the fifth imaging unit 282 then captures an image of the first die D1 mounted on the wafer W from above. The captured image is output to the control device 140, which measures the position of the first die D1 relative to the wafer W. That is, the control device 140 measures the positions (misalignments) of at least two, preferably three or more, alignment marks D1m on the first die D1 relative to at least two, preferably three or more, alignment marks Wm on the wafer W. Then, the bonding accuracy of the first die D1 relative to the wafer W is inspected. In this way, a series of bonding processes in the first bonding apparatus 100 is completed.

[0131] According to the above embodiment, various position adjustments (alignments) are performed in the first bonding apparatus 100. That is, the first imaging unit 240 is used to adjust the positions of the pickup unit 230 and the first die D1 picked up by the pickup unit 230. The second imaging unit 241 is used to adjust the position of the pickup unit 230 so that the first die D1 is delivered to the center of the bond head 261. The third imaging unit 280 and the fourth imaging unit 281 are used to adjust the position of the first die D1 relative to the wafer W. This allows the first die D1 to be bonded to an appropriate position on the wafer W. Furthermore, it is also possible to suppress voids when bonding the first die D1 to the wafer W. Therefore, the first bonding apparatus 100 can bond the first die D1 to the wafer W with high bonding accuracy.

[0132] Furthermore, in the first bonding apparatus 100, by increasing the rigidity of the moving mechanism 211, the arm 232 of the moving mechanism 231, the moving mechanism 252, and the arm 271 of the moving mechanism 270, it is possible to improve the accuracy of the movement of the carrier holding unit 210, the movement of the pickup unit 230, the movement of the wafer holding unit 250, and the movement of the head unit 260. In this case, it is possible to improve the accuracy of bonding the first die D1 to the wafer W in the first bonding apparatus 100.

[0133] Furthermore, in the first bonding apparatus 100, by improving the performance of each of the drive unit of the moving mechanism 211, the drive unit 235 of the moving mechanism 231, the drive unit of the moving mechanism 252, and the drive unit 274 of the moving mechanism 231, it is possible to speed up the movement of the carrier holding unit 210, the movement of the pickup unit 230, the movement of the wafer holding unit 250, and the movement of the head unit 260. As a result, it is possible to shorten the time required for the bonding process in the first bonding apparatus 100.

[0134] Next, the configuration of the second joining device 130 will be described, focusing on the differences from the first joining device 100. The second joining device 130 is a low-precision joining device, and compared to the first joining device 100, the device configuration can be kept to lower precision specifications.

[0135] For example, the second bonding apparatus 130 can omit at least one of the first imaging unit 240, the third imaging unit 280, and the fourth imaging unit 281 because low positional accuracy of the second die D2 relative to the wafer W is sufficient. Also, the second imaging unit 241 can be omitted if it is not necessary to suppress voids when bonding the second die D2 to the wafer W. Furthermore, the fifth imaging unit 282 can be omitted if low bonding accuracy of the second die D2 relative to the wafer W is sufficient and inspection is not necessary.

[0136] Furthermore, in the second bonding apparatus 130, by improving the performance of each of the drive units of the moving mechanism 211, the drive unit 235 of the moving mechanism 231, the drive unit of the moving mechanism 252, and the drive unit 274 of the moving mechanism 270, it is possible to speed up the movement of the carrier holding unit 210, the movement of the pickup unit 230, the movement of the wafer holding unit 250, and the movement of the head unit 260. On the other hand, since improving the performance of each drive unit in this way results in the drive unit becoming larger, it is preferable to reduce the rigidity of, for example, the moving mechanism 211, the rigidity of the arm 232 of the moving mechanism 231, the rigidity of the moving mechanism 252, and the rigidity of the arm 271 of the moving mechanism 270. In this regard, since the bonding precision required for the second bonding apparatus 130 is low, it is possible to reduce these rigidities.

[0137] As described above, the second bonding apparatus 130 can have a lower-precision device configuration and can also have lower-precision bonding conditions than the first bonding apparatus 100. This reduces the time required for bonding processing in the second bonding apparatus 130. Note that the second bonding apparatus 130 can have the same device configuration as the first bonding apparatus 100, and only the bonding conditions can be set to lower-precision specifications.

[0138] The bonding conditions are conditions for bonding the first die D1 and the second die D2 to the wafer W, and are equipment parameters. The equipment parameters include, for example, the acceleration / deceleration and speed of the movement of the carrier holding unit 210, the pickup unit 230, the wafer holding unit 250, and the head unit 260 by the movement mechanisms 211, 231, 252, and 270, and the waiting time until the drive of the drive units of these movement mechanisms 211, 231, 252, and 270 stabilizes. The equipment parameters also include whether or not to capture images by the image capturing units 240, 241, 280, 281, and 282, and the image capturing time.

[0139] In the first welding device 100 and the second welding device 130 of the above embodiment, the release region 201 and the welding region 202 are arranged side by side in the horizontal direction (X-axis direction), but they may be arranged stacked in the vertical direction, for example. In this case, the footprints (occupied areas) of the first welding device 100 and the second welding device 130 can be kept small.

[0140] In the above embodiment, the first carrier C1 holds a plurality of first dies D1, and the second carrier C2 holds a plurality of second dies D2, but one carrier may hold a plurality of first dies D1 and a plurality of second dies D2. Alternatively, one carrier may hold three or more types of dies, and the pattern of dies held by one carrier is arbitrary.

[0141] In the above embodiment, the case where the first die D1 and the second die D2 are mounted on the wafer W has been described. However, three or more types of dies may be mounted on the wafer W. For example, a third die D3 may be mounted on the wafer W in addition to the first die D1 and the second die D2. In such a case, the processing system 1 may be provided with a dedicated bonding device (not shown) for bonding the third die D3 to the wafer W. Alternatively, for example, if the third die D3 is a low-precision die, the second die D2 and the third die D3 may be bonded to the wafer W by replacing the bond head in the second bonding device 130.

[0142] In the above embodiment, the first bonding apparatus 100, which is a high-precision bonding apparatus as described above, bonds the plurality of first dies D1 to the wafer W with high bonding accuracy. In this regard, in the present embodiment, in order to further improve the positional accuracy of the plurality of first dies D1 on the wafer W, the positions of the plurality of first dies D1 on the subsequent second wafer W2 are corrected based on the measurement results of the positions of the plurality of first dies D1 on the first wafer W1. Note that the second wafer W2 may be a wafer processed next to the first wafer W1, or may be a wafer processed second or subsequent to the first wafer W1.

[0143] Here, in the conventional position correction, for example, the average value of the positional misalignment between the positions of the alignment marks D1m of all the first dies D1 on the first wafer W1 and the alignment marks Wm on the first wafer W1 is calculated, and then the horizontal positions (positions in the positive X-axis direction and positions in the Y-axis direction) of the multiple first dies D1 relative to the second wafer W2 are corrected based on the average positional misalignment.

[0144] Furthermore, for example, the average value of the angular misalignment between the measured angle calculated using a trigonometric function from the positions of the alignment marks D1m of all the first dies D1 on the first wafer W1 and the target angle based on the alignment marks Wm on the first wafer W1 is calculated, and the rotational direction positions (θ direction positions) of the multiple first dies D1 are corrected based on the average value of the angular misalignment.

[0145] That is, in the conventional position correction, one correction value for the horizontal position and one correction value for the rotational position are provided for one second wafer W2. In contrast, because the positional misalignment of the multiple first dies D1 on the second wafer W2 is not constant, it is not possible to perform appropriate correction for each of the multiple first dies D1.

[0146] Here, the wafer holder 250 is moved horizontally by the moving mechanism 252 and rotated about the vertical axis. The head unit 260 is moved horizontally and vertically by the moving mechanism 270 and rotated about the vertical axis. Therefore, the wafer holder 250 and the head unit 260 contain drive axis error components (drive errors). For example, when the wafer holder 250 moves horizontally, if the actual movement amount differs from the target movement amount, the wafer holder 250 will be positioned offset from the target position, resulting in a scaling error (scaling). Furthermore, for example, when the wafer holder 250 moves in the X-axis direction, if the wafer holder 250 rotates, a rotation error (rotation) will occur. Furthermore, for example, when the wafer holder 250 moves in the X-axis direction, if the wafer holder 250 is positioned offset in the Y-axis direction, an orthogonality error (orthogonality) will occur.

[0147] Due to the influence of the driving error, the positional misalignment of the first dies D1 on the second wafer W2 is not constant. Furthermore, when the positions of the first dies D1 on one second wafer W2 are corrected using a single correction value, as in the conventional method, it is not possible to correct the positional misalignment of the first dies D1 caused by the driving error of the wafer holder 250 or the head unit 260.

[0148] Therefore, in this embodiment, the positional misalignment of the first die D1 caused by driving errors of the wafer holder 250 and the head unit 260 is corrected. Specifically, the positions of the multiple first dies D1 mounted on the second wafer W2 are corrected on a die-by-die basis, and then the positions of the multiple first dies D1 are further corrected on a wafer-by-wafer basis. In the following description, die-by-die correction of the positions of the multiple first dies D1 is referred to as local correction, and wafer-by-wafer correction of the positions of the multiple first dies D1 is referred to as global correction. In this embodiment, the positional misalignment of the first die D1 that remains even after the local correction (hereinafter referred to as "positional misalignment residue") is corrected by global correction. FIG. 19 is a flow chart showing the main steps of the positional correction method for the first die D1. FIG. 20 is an explanatory diagram schematically showing the process of correcting the position of the first die D1.

[0149] First, the positions of the multiple first dies D1 mounted on the first wafer W1 are measured (St21 in FIG. 19 ). The position measurement in St21 may be performed using the fifth imaging unit 282 of the first bonding apparatus 100 or the inspection apparatus 70. In St21, the X-axis position (X coordinate) and Y-axis position (Y coordinate) of the alignment mark D1m in the first die D1 are measured, and the X-axis position (X coordinate) and Y-axis position (Y coordinate) of the first die D1 in the second wafer W2 are also measured. Alternatively, a separately provided mark for measuring misalignment may be used. Note that a large number of measurement points (three or more) is desirable for accurate measurement of misalignment. In St21, the amount of misalignment of the first die D1 relative to the first wafer W1, i.e., the amount of misalignment of the alignment mark D1m of the first die D1 relative to the alignment mark Wm of the wafer W, is also measured.

[0150] 20A shows an example of measurement results of the positions of the first dies D1 on the first wafer W1, i.e., a schematic diagram of an example of the positions of the first dies D1 on the second wafer W2 before the position correction of this embodiment. In FIG. 20A, arrows indicate positional misalignment components of the first dies D1.

[0151] Next, based on the measurement results of St21, a local correction model (first alignment analysis model) is used as a die-by-die model to derive die-by-die positional misalignment components of the multiple first dies D1 on the second wafer W2 (St22 in Figure 19).

[0152] The local correction model is a model using the least squares method of four parameters, and in the local correction model, the X-axis component is expressed by the following formula (1), and the Y-axis component is expressed by the following formula (2): dx ｉ = Tx + S × X ｉ -Rot x Y ｉ + Rsi(x ｉ ) ... (1) dy ｉ = Ty + S × Y ｉ +RotxX ｉ + Rsi(y ｉ) (2) where i: measurement position on the first die D1 (i = 1, 2, ..., n) dx: positional deviation amount of the first die D1 in the X-axis direction dy: positional deviation amount of the first die D1 in the Y-axis direction X: coordinate in the X-axis direction within the first die D1 Y: coordinate in the Y-axis direction within the first die D1 Tx: shift component of the first die D1 in the X-axis direction Ty: shift component of the first die D1 in the Y-axis direction S: scaling component of the first die D1 Rot: rotation component of the first die D1 Rsi: uncorrectable residue

[0153] In the above formulas (1) and (2), i is the measurement position on the first die D1, that is, the number (serial number) of a plurality of, for example, at least two, alignment marks D1m, and dx, dy, X, and Y are the measurement results measured in St21, respectively.

[0154] Tx and Ty are shift components (translation) of the first die D1 in die units. The arrows in FIG. 21A schematically show an example of the shift components of the first die D1. That is, the shift components are translation components that shift the first die D1 in die units in the horizontal direction relative to the second wafer W2.

[0155] S is a scaling component (scaling) of the first die D1 in die units. The arrows in Fig. 21(b) schematically show an example of the scaling component of the first die D1. That is, the scaling component is a component by which the first die D1 is enlarged or reduced in die units, and is a ratio to the original size.

[0156] Rot is a rotation component (Rotation) of the first die D1 in die units. The arrows in Fig. 21C schematically show an example of the rotation component of the first die D1. That is, the rotation component is a component of the rotation of the first die D1 in die units.

[0157] In St22, four coefficients, Tx, Ty (shift component), S (scaling component), and Rot (rotation component), in the local correction model are derived. These four coefficients are the die-by-die positional misalignment components of the first die D1. Specifically, Tx, Ty, S, and Rot are derived using the following equation (3).

[0158] Next, based on the die-by-die positional misalignment components derived in St22, the positions of the multiple first dies D1 on the second wafer W2 are corrected on a die-by-die basis, that is, local correction is performed (St23 in FIG. 19).

[0159] In St23, the correction value X in the X-axis direction of the first die D1 is offset , the correction value Y in the Y-axis direction of the first die D1 offset , the correction value S in the enlargement / reduction direction of the first die D1 offset , the correction value Rot in the rotation direction (θ-axis direction) of the first die D1 offset are calculated from the following formulas (4) to (7), respectively. where j is the number of the first die D1 (j=1, 2, . . . , n) X offset Y: Correction value of the first die D1 in the X-axis direction offset : Correction value S in the Y-axis direction of the first die D1 offset : Correction value in the enlargement / reduction direction of the first die D1 Rot offset : Correction value for the rotation direction of the first die D1

[0160] 20(b) schematically illustrates an example of the positions of the multiple first dies D1 on the second wafer W2 after local correction. In FIG. 20(b), the arrows indicate misalignment components of the first die D1. Comparing FIG. 20(a) before local correction with FIG. 20(b) after local correction, it can be seen that the misalignment of the first die D1 is reduced by local correction. However, as described above, local correction cannot correct the misalignment of the first die D1 caused by driving errors of the wafer holder 250 or the head unit 260, and therefore misalignment residue remains.

[0161] Next, the head unit 260 is feedback-controlled based on the correction value calculated in St23 (St24 in FIG. 19). offset , correction value Y offset , correction value Rot offset Based on each of the above, the position of the head unit 260 in the X-axis direction, the Y-axis direction, and the rotation direction is adjusted, and the position of the first die D1 held by the head unit 260 is corrected. offsetBased on this, the pressure when the head unit 260 presses the first die D1 against the second wafer W2 is adjusted, and the enlargement / reduction of the first die D1 is corrected. Alternatively, as described above, the bond head 261 has a convex shape with a central portion protruding downward, and the enlargement / reduction of the first die D1 may be corrected by replacing the bond head 261 so as to change the curvature of this convex shape.

[0162] In St24, the pickup unit 230 may be feedback-controlled. In this case, the correction value X offset , correction value Y offset , correction value Rot offset Based on each of the above, the position of the pickup unit 230 in the X-axis direction, Y-axis direction, and rotation direction is adjusted, and the position of the first die D1 held by the pickup unit 230 is corrected.

[0163] Next, in parallel with St24, the misalignment residue of the first die D1 that remains after the local correction in St23 is estimated and derived (St25 in FIG. 19). In St25, the misalignment residue Model(x) of the first die D1 in the X-axis direction and the misalignment residue Model(y) of the first die D1 in the X-axis direction are calculated using the following formulas (8) and (9), respectively. Model(x)=dx ｉ -(X offset -Rot offset ×Y ｉ )...(8) Model(y)=dy ｉ -(Y offset +Rot offset ×X ｉ ) (9) where, i: measurement position on the first die D1 (i=1, 2, . . . , n) Model(x): misalignment residue in the X-axis direction of the first die D1 Model(y): misalignment residue in the Y-axis direction of the first die D1

[0164] Next, based on the misalignment residue derived in St25, a global correction model (second alignment analysis model) is used as a target substrate unit model to derive wafer-unit misalignment components for multiple first dies D1 on the second wafer W2 (St26 in Figure 19).

[0165] The global correction model is a model using the least squares method of six parameters, and in the global correction model, the X-axis component is expressed by the following formula (10), and the Y-axis component is expressed by the following formula (11). Model(x ｉ ) = Tx + Sx × X ｊ -(Rot+Ort)×Y ｊ + Rsi(x ｊ )...(10) Model(y ｉ ) = Ty + Sy x Y ｊ +RotxX ｊ + Rsi(y ｊ ) (11) where j: number of the first die D1 (j=1, 2, . . . , n) Model(x): estimated positional misalignment amount in the X-axis direction of the first die D1 after local correction Model(y): estimated positional misalignment amount in the Y-axis direction of the first die D1 after local correction X: coordinate of the first die D1 in the X-axis direction within the second wafer W2 Y: coordinate of the first die D1 in the Y-axis direction within the second wafer W2 Tx: shift component of the first die D1 in the X-axis direction Ty: shift component of the first die D1 in the Y-axis direction Sx: scaling component of the first die D1 in the X-axis direction Sy: scaling component of the first die D1 in the Y-axis direction Rot: rotation component of the first die D1 Ort: orthogonal component of the first die D1 Rsi: uncorrectable residue

[0166] In the above formulas (10) and (11), j is the number (serial number) of the multiple first dies D1. Model(x) and Model(y) are the misalignment residues of the first die D1 derived in St25. X and Y are the measurement results measured in St21.

[0167] Tx and Ty are shift components (translation) of the first dies D1 on the second wafer W2 in a wafer-by-wafer basis. The arrows in FIG. 22A schematically show an example of the shift components of the first dies D1. That is, the shift components are translation components that shift the first dies D1 on the second wafer W2 in a horizontal direction in a wafer-by-wafer basis.

[0168] Sx and Sy are scaling components (scaling) of the plurality of first dies D1 on the second wafer W2 on a wafer-by-wafer basis. The arrows in FIG. 22B schematically show an example of the scaling components of the plurality of first dies D1. That is, the scaling components are components by which the plurality of first dies D1 on the second wafer W2 are enlarged or reduced on a wafer-by-wafer basis.

[0169] Rot is a rotation component (Rotation) of the plurality of first dies D1 on the second wafer W2 per wafer. The arrows in FIG. 22C schematically show an example of the rotation component of the plurality of first dies D1. That is, the rotation component is a component of the rotation of the plurality of first dies D1 on the second wafer W2 per wafer.

[0170] Ort is the orthogonality of the first dies D1 on the second wafer W2 on a wafer-by-wafer basis. The arrows in FIG. 22D schematically show an example of the orthogonality of the first dies D1. That is, the orthogonality is a component that causes the orthogonality of the first dies D1 on the second wafer W2 to deviate from that of the first dies D1 on a wafer-by-wafer basis.

[0171] In St26, six coefficients of the global correction model are derived: Tx, Ty (shift components), Sx, Sy (scaling components), Rot (rotation component), and Ort (orthogonal component). These six coefficients are misalignment components of the first die D1 per wafer. Specifically, Tx, Ty, Sx, Sy, Rot, and Ort are derived using the following equation (12) so as to minimize the misalignment residue of the first die D1.

[0172] Next, based on the misalignment components of each wafer derived in St26, the positions of the first dies D1 on the second wafer W2 are corrected for each wafer, that is, global correction is performed (St27 in FIG. 19).

[0173] In St27, the correction value X in the X-axis direction of the first die D1 is offset , the correction value Y in the Y-axis direction of the first die D1 offset , the correction value Rot in the rotation direction (θ-axis direction) of the first die D1 offsetare calculated from the following formulas (13) to (15), respectively. offset = Tx + Sx × X ｊ -OrtxY ｊ ... (13) Y offset = Ty + Sy × Y ｊ ...(14) Rot offset =Rot (15) where j is the number of the first die D1 (j=1, 2, . . . , n) offset Y: Correction value of the first die D1 in the X-axis direction offset : Correction value of the first die D1 in the Y-axis direction Rot offset : Correction value for the rotation direction of the first die D1

[0174] 20(c) schematically illustrates an example of the positions of the multiple first dies D1 on the second wafer W2 after the global correction. In FIG. 20(c), the arrows indicate misalignment components of the first die D1. Comparing FIG. 20(b) before the global correction with FIG. 20(c) after the global correction, it can be seen that the misalignment of the first die D1 is reduced by the global correction. Therefore, the residual misalignment of the first die D1 can be minimized and reduced by the global correction.

[0175] Next, the wafer holding part 250 is feedback-controlled based on the correction value calculated in St27 (St28 in FIG. 19). offset , correction value Y offset , correction value Rot offset Based on these, the position of the wafer holding part 250 in the X-axis direction, the Y-axis direction, and the rotation direction is adjusted, and the position of the first die D1 relative to the second wafer W2 held by the wafer holding part 250 is corrected. Note that if the wafer holding part 250 is not configured to be movable in the rotation direction, the correction value Rot offset The X-axis and Y-axis directions of the wafer holder 250 are adjusted by separating the X-axis and Y-axis directions.

[0176] In addition, in St28, the head unit 260 may be feedback-controlled. In such a case, the correction value X offset , correction value Y offset , correction value Rot offsetBased on each of the above, the position of the head unit 260 in the X-axis direction, Y-axis direction, and rotation direction is adjusted, and the position of the first die D1 held by the head unit 260 is corrected.

[0177] According to the above embodiment, the residual positional misalignment of the first die D1 that remains after the local corrections in St22 and St23 can be reduced by performing the global corrections in St26 and St27. That is, the positional misalignment of the first die D1 caused by the driving errors of the wafer holder 250 and the head unit 260 can be corrected by the global correction. Therefore, the positional precision (accuracy) of the multiple first dies D1 mounted on the wafer W can be improved.

[0178] Furthermore, since the global correction model used in the global corrections of St26 and St27 uses the least squares method, it is possible to derive the wafer-by-wafer misalignment component (correction coefficient) of the first die D1 from the measurement results of multiple points so that the sum of squares of the misalignment residue (error) of the first die D1 is minimized. In such a case, approximation is performed from the measurement results of multiple points, so the misalignment residue (error) after correction can be reduced. Furthermore, since the distribution and amount of the misalignment component can be derived on a wafer-by-wafer basis, it is possible to appropriately correct the positions of multiple first dies D1. Furthermore, since the misalignment component is calculated for each first die D1, it is possible to appropriately correct the position of each first die D1.

[0179] As described above, the conventional position correction method corrects the horizontal position of the first die D1 using the average value of the positional misalignment amounts of all the first dies D1 on the first wafer W1, and corrects the rotational position of the first die D1 using the average value of the angular misalignment amounts calculated using a trigonometric function. Such correction values ​​cannot classify the positional misalignment components (shift component, scaling component, and rotation component) of the first die D1, and therefore the position correction of the first die D1 cannot be optimally performed. In particular, the conventional position correction method cannot calculate the scaling component of the positional misalignment components. Furthermore, the conventional position correction method calculates the angular misalignment amount using a trigonometric function, and therefore the rotational component of the positional misalignment components of the first die D1 can only be calculated from data on two points, resulting in a large error.

[0180] In this regard, the local correction model of St22 and St23 in this embodiment uses the least squares method, so that Tx, Ty (shift component), S (scaling component), and Rot (rotation component) can be derived as the positional deviation components of the first die D1. Therefore, according to this embodiment, the accuracy of the local correction itself can be improved compared to conventional positional correction methods, and the local correction of the first die D1 can be performed appropriately.

[0181] Furthermore, in St21, the position of the first die D1 on the first wafer W1 is measured using either the fifth imaging unit 282 of the first bonding apparatus 100 or the inspection apparatus 70. In St24, at least one of the head unit 260 and the pickup unit 230 is automatically feedback-controlled for the second wafer W2 based on the local correction result of St23. In St28, at least one of the wafer holder 250 and the head unit 260 is automatically feedback-controlled for the second wafer W2 based on the global correction result of St27. In this manner, in the present embodiment, in one processing system, the position measurement of the first die D1 on the first wafer W1 and the position correction (local correction and global correction) of the first die D1 on the subsequent second wafer W2 can be automatically performed inline.

[0182] In the above embodiment, the local corrections in St22 and St23 are performed, and then the global corrections in St26 and St27 are performed. However, the local corrections in St22 and St23 may be omitted. In such a case, the global corrections in St26 and St27 are performed directly based on the position measurement results of the first die D1 in St21. Specifically, in the global correction model, Model(x ｊ ) and Model(y ｊ ) are the positional misalignment amounts dx and dy of the first die D1 measured in St21. Even in this case, the positional precision (accuracy) of the multiple first dies D1 mounted on the wafer W can be improved.

[0183] In the above embodiment, a local correction model is used in the local corrections of St22 and St23. However, as in the conventional position correction method, an average value of the positional misalignment amounts and an angular misalignment amount calculated using a trigonometric function may be used. In such a case, after performing the local correction using the conventional position correction method, global corrections of St26 and St27 are performed. As a result, the positional accuracy (precision) of the multiple first dies D1 mounted on the wafer W can be improved.

[0184] Furthermore, in the above embodiments, the local correction models of St22 and St23 and the global correction models of St26 and St27 were each models using the least squares method, but the method is not limited to the least squares method as long as it is a model that can optimize the positional misalignment component (correction coefficient) of the first die D1.

[0185] In addition, in the above embodiment, the position measurement of the first die D1 on the first wafer W1 is performed inside the processing system 1 using either the fifth imaging unit 282 of the first bonding apparatus 100 or the inspection apparatus 70, but it may also be performed outside the processing system 1.

[0186] In the above embodiment, the position of the first die D1 is corrected. However, the position of the second die D2 may be corrected using a similar position correction method. In this case, local correction and global correction are performed individually depending on the types of the dies D1 and D2. For example, the control device 140 may store recipes for local correction and global correction depending on the types of the dies D1 and D2.

[0187] The embodiments disclosed herein should be considered to be illustrative in all respects and not restrictive. The above-described embodiments may be omitted, substituted, or modified in various ways without departing from the scope and spirit of the appended claims. For example, the components of the above-described embodiments may be arbitrarily combined. Such an arbitrary combination naturally provides the functions and effects of each of the components involved in the combination, and also provides other functions and effects that are apparent to those skilled in the art from the description of this specification.

[0188] Furthermore, the effects described herein are merely descriptive or exemplary and are not limiting. In other words, the technology according to the present disclosure may achieve other effects that would be apparent to a person skilled in the art from the description of this specification, in addition to or in place of the above-described effects.

[0189] REFERENCE SIGNS LIST 1 Processing system 70 Inspection device 100 First bonding device 130 Second bonding device 140 Control device 230 Pickup unit 250 Wafer holding unit 260 Head unit 282 Fifth imaging unit C1 First carrier C2 Second carrier D1 First die D2 Second die W Wafer

Claims

1. A processing system for mounting a plurality of dies held on a carrier onto a target substrate, comprising: a substrate holding unit that holds the target substrate and is configured to be movable horizontally; a pickup unit that picks up the dies from the carrier; a head unit that receives the dies from the pickup unit and bonds the dies to the target substrate held by the substrate holding unit; and a control unit. The control unit executes control to derive a position deviation component in units of the target substrate of the positions of the plurality of dies on another target substrate based on the position information of the plurality of dies on one target substrate measured by a measurement unit that measures the positions of the plurality of dies mounted on the target substrate.

2. The processing system according to claim 1, wherein the control unit executes control to adjust at least one of the substrate holding unit and the head unit so as to correct the positions of the plurality of dies on the other target substrate based on the position deviation component in units of the target substrate.

3. The position deviation component in units of the target substrate includes a shift component of the plurality of dies on the other target substrate, a magnification / reduction component of the plurality of dies on the other target substrate, a rotation component of the plurality of dies on the other target substrate, and an orthogonal component of the plurality of dies on the other target substrate. The processing system according to claim 1.

4. The processing system according to claim 1, wherein the control unit executes control to derive the position deviation component in units of the target substrate using a target substrate unit model using the least squares method.

5. The processing system according to claim 1, wherein the control unit executes control to derive a position deviation component in units of the die of the positions of the plurality of dies on another target substrate based on the position information of the plurality of dies on one target substrate measured by the measurement unit.

6. The processing system according to claim 5, wherein the control unit executes control to adjust at least one of the head unit and the pickup unit so as to correct the positions of the plurality of dies on the other target substrate based on the position deviation component in units of the die.

7. The position deviation component in units of the die includes a shift component of the die on the other target substrate, a magnification / reduction component of the die on the other target substrate, and a rotation component of the die on the other target substrate. The processing system according to claim 5.

8. The processing system according to claim 5, wherein the control unit executes control to derive a positional deviation component of each die using a die unit model using the least squares method.

9. The control unit derives a positional deviation component of each die, and after performing control to correct the positions of the plurality of dies on the other target substrate, controls to derive a positional deviation residue of the plurality of dies on the other target substrate, and based on the positional deviation residue, controls to derive a positional deviation component of each target substrate unit and correct the positions of the plurality of dies on the other target substrate. The processing system according to claim 5.

10. A processing method for mounting a plurality of dies held by a carrier on a target substrate, including holding the target substrate with a substrate holding unit configured to be movable horizontally, picking up the die from the carrier using a pickup unit, receiving the die from the pickup unit with a head unit and bonding the die to the target substrate held by the substrate holding unit using the head unit, measuring the positions of the plurality of dies mounted on the target substrate using a measuring unit, and deriving a positional deviation component of each target substrate unit of the positions of the plurality of dies on another target substrate based on the position information of the plurality of dies on one target substrate measured by the measuring unit.

11. The processing method according to claim 10, including adjusting at least one of the substrate holding unit and the head unit so as to correct the positions of the plurality of dies on the other target substrate based on the positional deviation component of each target substrate unit.

12. The positional deviation component of each target substrate unit includes a shift component of the plurality of dies on the other target substrate, a magnification / reduction component of the plurality of dies on the other target substrate, a rotation component of the plurality of dies on the other target substrate, and an orthogonal component of the plurality of dies on the other target substrate. The processing method according to claim 10.

13. The processing method according to claim 10, including deriving the positional deviation component of each target substrate unit using a target substrate unit model using the least squares method.

14. The processing method according to claim 10, comprising deriving a die unit position deviation component of the positions of the plurality of dies in the other target substrate based on the position information of the plurality of dies in the one target substrate measured by the measurement unit.

15. The processing method according to claim 14, comprising adjusting at least one of the head unit and the pickup unit so as to correct the positions of the plurality of dies in the other target substrate based on the die unit position deviation component.

16. The processing method according to claim 14, wherein the die unit position deviation component includes a shift component of the die in the other target substrate, a magnification / reduction component of the die in the other target substrate, and a rotation component of the die in the other target substrate.

17. The processing method according to claim 14, comprising deriving the die unit position deviation component using a die unit model using the least squares method.

18. The processing method according to claim 14, comprising deriving a position deviation residue of the plurality of dies in the other target substrate after deriving the die unit position deviation component and correcting the positions of the plurality of dies in the other target substrate, and deriving a position deviation component of the target substrate unit based on the position deviation residue and correcting the positions of the plurality of dies in the other target substrate.

Images