Bonding apparatus, bonding method, estimation method, and article manufacturing method

Abstract

The present invention provides a bonding apparatus for bonding a second member having a second bonding surface having a second pattern thereon to a first member having a first bonding surface having a first pattern thereon. The apparatus includes: a first imaging device configured to capture an image of the first bonding surface; a second imaging device configured to capture an image of the second bonding surface; and a controller configured to control the alignment of the first member and the second member, and the bonding process of bonding the second member to the first member, based on the position of the first pattern obtained from the image captured by the first imaging device and the position of the second pattern obtained from the image captured by the second imaging device.

Description

Technical Field

[0001] The present invention relates to a bonding device, a bonding method, an estimation method and an article manufacturing method. Prior Art

[0002] Japanese Patent Laid-Open No. 2021-190527 describes a method for determining the quality of the bond between a wafer and a substrate by measuring the amount of deformation of the wafer and substrate when ultrasonically bonded to the substrate. Japanese Patent Laid-Open No. 2011-169816 describes a method for inspecting the quality of the bond between a substrate and a semiconductor wafer by measuring the inclination of the semiconductor wafer relative to the substrate based on an image obtained by photographing the bond between the substrate and the semiconductor wafer from the side.

[0003] In bonding equipment used to bond first and second components, it is important to minimize the relative positional deviation between the pattern provided on the bonding surface of the first component and the pattern provided on the bonding surface of the second component after bonding. However, after bonding, the bonding surfaces of the first and second components cannot be visually inspected. Therefore, it is difficult to easily and accurately determine the relative position between the pattern on the first component and the pattern on the second component after bonding. The methods described in Japanese Patent Laid-Open Nos. 2021-190527 and 2011-169816 judge the bonding quality between the first component (substrate) and the second component (wafer) based solely on appearance, and are unable to determine the relative position between the pattern on the first component and the pattern on the second component after bonding. Summary of the Invention

[0004] The present invention provides a technique that facilitates obtaining the relative position between a pattern of a first member and a pattern of a second member, for example, after the first member and the second member are joined.

[0005] According to one aspect of the present invention, a bonding apparatus is provided for bonding a second member including a second bonding surface having a second pattern thereon to a first member including a first bonding surface having a first pattern thereon, the apparatus comprising: a first image capturing device configured to capture an image of the first bonding surface of the first member; a second image capturing device configured to capture an image of the second bonding surface of the second member; and a controller configured to capture an image of the second bonding surface based on a position of the first pattern obtained from the image captured by the first image capturing device and a position of the second pattern obtained from the image captured by the second image capturing device. The controller is configured to control a bonding process of aligning the first component and the second component and bonding the second component to the first component based on the position of the first component, wherein after the bonding process, the controller is configured to: obtain characteristic position information indicating the position of a characteristic portion of the second component relative to the first component based on the image obtained by photographing the second component bonded to the first component by the first image capturing device; and estimate the relative position between the first pattern and the second pattern after the bonding process based on the pre-acquired positional relationship information and the characteristic position information, the positional relationship information indicating the positional relationship between the characteristic portion and the second pattern in the second component.

[0006] Further features of the present invention will become apparent from the following description of exemplary embodiments with reference to the accompanying drawings. Simple diagram description

[0007] [Figure 1] is a schematic diagram showing a bonding apparatus according to a first embodiment;

[0008] [Figure 2] is a diagram of the stage viewed from the +Z direction;

[0009] [FIG. 3A] is a schematic diagram showing a state in which an image of a grain is captured;

[0010] [Figure 3B] is a schematic diagram showing an example of the configuration of the grains;

[0011] [Figure 4] is a flowchart showing the operating sequence of the bonding device in Example 1;

[0012] [FIG. 5] is a diagram schematically showing the bonding between a wafer and a die;

[0013] [Figure 6] is a flowchart showing the operation sequence of the bonding device in Example 1 (modification);

[0014] [Figure 7] is a schematic diagram showing a bonding apparatus according to a second embodiment;

[0015] FIG8 is a flow chart showing the operation sequence of the bonding apparatus in Example 3; and

[0016] [Figure 9] is a flow chart showing the operation sequence of the bonding device in Example 4. Implementation Method

[0017] Below, embodiments will be described in detail with reference to the accompanying drawings. Please note that the following embodiments are not intended to limit the scope of the claimed invention. While various features are described in the embodiments, this does not limit the invention to requiring all such features, and multiple such features may be combined as appropriate. Additionally, in the accompanying drawings, identical or similar components are assigned the same element numbers, and redundant descriptions are omitted.

[0018] In this specification and accompanying drawings, directions will be indicated on an XYZ coordinate system, where the plane parallel to the horizontal plane is defined as the XY plane. Directions parallel to the X, Y, and Z axes of the XYZ coordinate system are defined as the X-direction, Y-direction, and Z-direction, respectively. Rotation about the X-axis, rotation about the Y-axis, and rotation about the Z-axis are defined as θX, θY, and θZ, respectively. Control and drive (movement) regarding the X-axis, Y-axis, and Z-axis refers to control or drive (movement) regarding directions parallel to the X-axis, Y-axis, and Z-axis, respectively. Furthermore, control or drive regarding the θX-axis, θY-axis, and θZ-axis refers to control or drive regarding rotation about an axis parallel to the X-axis, rotation about an axis parallel to the Y-axis, and rotation about an axis parallel to the Z-axis, respectively.

[0019] In the embodiments described later, a wafer (substrate) having semiconductor devices formed thereon is used as the first member (the first bonding target), and dies (wafers) obtained by dividing the wafer having semiconductor devices formed thereon into pieces are used as the second member (the second bonding target). However, the first and second members are not limited to these bonding targets, and various changes and modifications can be made within the scope of the present invention.

[0020] As the first member, a substrate having a bonding surface (first bonding surface) on which a pattern (first pattern) is provided can be used. Examples of the substrate include, in addition to wafers having semiconductor devices formed thereon, silicon wafers, silicon wafers having wiring formed thereon, glass wafers, glass plates having wiring formed thereon, organic boards (PCBs) having wiring formed thereon, and metal plates. The substrate can be a wafer to which one or more dies have already been bonded.

[0021] As the second member, a die having a bonding surface (second bonding surface) provided with a pattern (second pattern) can be used. In addition to die obtained by dicing a wafer having semiconductor devices formed thereon, examples of die include die stacks, small material sheets, optical elements, MEMS, and structures.

[0022] In the embodiments described later, various temporary or permanent bonding methods can be used as the bonding method for the first and second components. Examples of bonding methods include bonding using an adhesive, temporary bonding using a temporary adhesive, bonding by hybrid bonding, atomic diffusion bonding, vacuum bonding, and bump bonding.

[0023] An industrial application example of the embodiment described later will be explained.

[0024] The first application example is the manufacture of stacked memory cells. When the bonding apparatus according to the embodiments described later is applied to the manufacture of stacked memory cells, a wafer (substrate) on which the memory cells are used as semiconductor devices serves as the first component, and a die (wafer) on which the memory cells are formed serves as the second component. For example, when manufacturing a stacked memory cell having eight memory cell layers, the second component (die) formed into eight memory cell layers is bonded to the first component (substrate) already having seven memory cell layers. Note that the last layer of the stacked memory cell may not be a memory cell layer, but rather a layer on which a driver for driving the memory cells is formed.

[0025] The second application example is heterogeneous integration of processors. Conventional processors are primarily system-on-chip (SoCs), in which logic circuits, static random access memory (SRAM), and other components are integrated into a single semiconductor device. In contrast, heterogeneous integration utilizes processes optimized for each component to form multiple types of components from separate wafers and then bond them together to produce a processor. This reduces processor costs and improves processor yield. When the bonding equipment described in the embodiments below is applied to heterogeneous integration, a wafer (substrate) on which logic devices, as semiconductor devices, are formed is used as the first component. Dies (wafers) separated after probing (such as SRAM, antennas, or drivers) serve as the second component. In heterogeneous integration, for example, different types of dies are bonded sequentially, increasing the number of components bonded to the first component. More specifically, when a die containing SRAM is bonded to a logic wafer, the logic wafer serves as the first component, and the die containing SRAM serves as the second component. When a die containing devices to be formed on an SRAM is bonded to a die containing the SRAM, the logic wafer and the SRAM die are bonded together as the first component, and the die containing the devices is the second component. When bonding multiple dies overlapping each other, it is preferable to start with the thinnest die in the bonding order to prevent the bonding head from interfering with the bonded dies.

[0026] The third application example is 2.5D bonding using a silicon interposer. A silicon interposer is a silicon wafer with wiring formed thereon. 2.5D bonding involves bonding multiple types of dies to a silicon interposer, electrically connecting the multiple types of dies via the wiring on the interposer. When the bonding apparatus according to the embodiments described later is applied to 2.5D bonding, the silicon wafer with wiring formed thereon serves as the first component, and the separated dies serve as the second component. In 2.5D bonding, for example, multiple types of dies are bonded to a silicon interposer. Therefore, a structure with one or more dies already bonded to the silicon interposer is sometimes treated as the first component. It is important to note that when bonding multiple dies to a silicon interposer, the bonding order is preferably started with thinner dies to prevent the bonding head from interfering with the dies being bonded.

[0027] The fourth application example is 2.1D bonding using an organic interposer or a glass interposer. An organic interposer is an organic panel (PCB substrate or CCL substrate) used as a packaging substrate, on which wiring is formed. A glass interposer is a glass plate with wiring formed thereon. 2.1D bonding involves bonding multiple types of dies to an organic interposer or a glass interposer, electrically connecting the multiple types of dies via the wiring on the interposer. When the bonding apparatus according to the embodiments described later is applied to 2.1D bonding, in 2.1D bonding using an organic interposer, the organic panel with wiring formed thereon serves as the first component, and the separated dies serve as the second component. In contrast, in 2.1D bonding using a glass interposer, the glass plate with wiring formed thereon serves as the first component, and the separated dies serve as the second component. In 2.1D bonding, for example, multiple types of dies are bonded to an organic interposer or a glass interposer, so the structure of the organic interposer or glass interposer to which one or more dies are bonded is sometimes treated as the first component. Note that when bonding multiple dies to the interposer, as for the bonding order, it is better to start bonding from the thin dies so that the bonding head does not interfere with the bonded dies.

[0028] A fifth application example involves temporary bonding in the fan-out packaging manufacturing process. For example, fan-out packaging, an advanced packaging method used in semiconductor manufacturing processes, includes fan-out wafer-level packaging and fan-out panel-level packaging. Fan-out wafer-level packaging uses a molded resin to encapsulate and reconstruct separate dies into a wafer shape. Fan-out panel-level packaging uses a molded resin to encapsulate and reconstruct separate dies into a panel shape. In this type of fan-out packaging, wiring is formed from the die to the bumps, or wiring is formed on a molded reconstructed substrate to bond different types of dies. If the die array has low precision, it can be difficult to precisely align the wiring pattern with the die when transferring it using a step-and-repeat exposure system. Therefore, multiple dies must be precisely aligned in the fan-out package. When the bonding equipment according to the embodiments described later is applied to the fan-out packaging manufacturing process, a metal plate serves as the first component, and the separated dies serve as the second component. More specifically, a bonding machine temporarily serially bonds the separated dies to a metal plate using a temporary adhesive. The dies temporarily bonded to the metal plate are then molded into a wafer or plate shape using a molding machine and peeled from the metal plate after molding. This produces a reconstructed wafer or plate with the dies arranged thereon. Note that in the fan-out packaging manufacturing process, the array of the dies can change during the molding process. Therefore, when temporarily bonding the dies to the metal plate using the bonding machine, it is preferable to adjust the bonding position of each die on the metal plate to correct for variations in the array caused by the molding process.

[0029] The sixth application example is heterogeneous substrate bonding. For example, in infrared image sensors, InGaAs, a highly sensitive material, is used in the sensor cells configured to receive light, while silicon, capable of high-speed processing, is used in the logic circuits configured to extract data. This allows the manufacture of highly sensitive, high-speed infrared image sensors. However, currently, mass production of InGaAs crystals can only produce wafers as small as 4 inches in diameter, which is smaller than the mainstream 300 mm silicon wafers. Therefore, a method has been proposed for bonding the resulting die, obtained by slicing the InGaAs substrate into slices, to a 300 mm silicon wafer on which logic circuits are formed. The bonding equipment according to the embodiments described later can also be applied to heterogeneous substrate bonding, where substrates made of different materials and having different sizes are bonded. When the bonding equipment is applied to heterogeneous substrate bonding, a substrate with a large diameter, such as a silicon wafer, is used as the first component, and a die (chip) of a material such as InGaAs is used as the second component. It should be noted that the crystal grains (chips) of a material such as InGaAs may be crystal plates and are preferably cut into a rectangular shape.

[0030] <First embodiment> The first embodiment of the present invention will now be described. Figure 1 is a schematic diagram of a bonding apparatus 100 according to the first embodiment. In Figure 1, the direction perpendicular to the holding surface of a stage 43 (described later) holding a wafer 6 (substrate) is defined as the Z direction, and the directions perpendicular to each other on a plane parallel to the holding surface of the stage 43 are defined as the X and Y directions. The bonding apparatus 100 sequentially bonds a plurality of dies 51, which serve as the second bonding target, to predetermined positions on a wafer 6, which serves as the first member (the first bonding target). The plurality of dies 51 are arranged on a dicing tape adhered to a dicing frame 5.

[0031] In this embodiment, as shown in FIG1 , a bonding apparatus 100 includes a pick-up unit 3, a bonding unit 4, and a controller CNT. The pick-up unit 3 and the bonding unit 4 are mounted on a base 1 cushioned by a mounting member 2. In this embodiment, the pick-up unit 3 and the bonding unit 4 are mounted on the same base 1, but they can also be mounted on separate bases.

[0032] The pickup unit 3 includes a pickup head 31, a release head 32, and a frame holder 33. The pickup unit 3 picks up dies 51 one by one from the dicing tape adhered to the dicing frame 5. The frame holder 33 holds the dicing frame 5. The release head 32 pushes the target die 51 upward from the back side of the dicing tape adhered to the dicing frame 5, causing the target die 51 to protrude upward from the remaining dies. At this point, the target die 51 is partially peeled off from the dicing tape. The pickup head 31 holds (holds) the target die 51 pushed upward by the release head 32 using a vacuum force or other means, and peels (separates) the target die 51 from the dicing tape. The pickup head 31 is configured to be movable from the pickup unit 3 to the bonding unit 4. When moving from the pickup unit 3 to the bonding unit 4, the pickup head 31 rotates (flips the wafer) to flip the die 51 and transfer it to the bonding head 44, described later. At this time, the pickup head 31 contacts the bonding surface (second bonding surface) of the die 51. To cope with bonding methods such as hybrid bonding that utilize surface activation, the bonding surface of the die 51 is preferably treated with a highly stable surface treatment such as a diamond-like carbon (DLC) coating or a fluorine coating.

[0033] Bonding unit 4 includes stage 43, bonding head 44, die observation camera 45, wafer observation camera 46, and interferometer 47. Stage 43 is configured to move on stage base 41, and die observation camera 45 is mounted on stage 43. Bonding head 44, wafer observation camera 46, and interferometer 47 are mounted on upper base 42.

[0034] First, the configuration on the stage base 41 will be described. The stage 43 and the die observation camera 45 are installed on the stage base 41.

[0035] Stage 43 (first holder) includes a chuck 431 that holds wafer 6 (substrate) on a holding surface parallel to the X and Y directions using vacuum force, etc., and a drive mechanism 432 that drives chuck 431 (wafer 6). Drive mechanism 432 includes an actuator, such as a linear motor, and is configured to drive wafer 6 in the X, Y, and θZ directions. Drive mechanism 432 can also be configured to drive wafer 6 in the Z direction. Relative rotation of wafer 6 and die 51 in the θZ direction can be performed by rotating wafer 6 using stage 43, or alternatively, by rotating die 51 using bond head 44, described later. Stage 43 also includes a mirror 433 for measuring the position of stage 43 in the X and Y directions. Mirror 433 serves as a target for interferometer 47 (described later) that measures the position of stage 43 in the X and Y directions.

[0036] The die observation camera 45 (second image capture device) is a camera used to observe the bonding surface (second bonding surface) of the die 51. The die observation camera 45 can be configured to capture an image of the bonding surface of the die 51 while the die 51 is held by the bonding head 44 (second holder). In this embodiment, the die observation camera 45 is mounted on the stage 43 and can move in the X and Y directions as the stage 43 moves. The die observation camera 45 is used to acquire (measure) information indicating the position of a pattern (second pattern) provided on the bonding surface of the die 51, as well as information indicating the positional relationship between a characteristic portion of the die 51 and the pattern provided on the bonding surface of the die 51.

[0037] The characteristic portion of die 51 is an indicator that can be identified in both an image obtained by photographing the bonding surface of die 51 and an image obtained by photographing the back surface (the surface opposite the bonding surface) of die 51. In other words, the characteristic portion of die 51 is an indicator that can be used as a reference for the position of die 51 on both the bonding surface and the back surface. Examples of the characteristic portion of die 51 are the outer edge of die 51 and / or a through hole (through hole) extending from the bonding surface to the back surface of die 51.

[0038] The die observation camera 45 is also used to measure the distances at multiple points along the height direction (Z direction) of the bonding surface of the die 51, that is, the height distribution of the bonding surface of the die 51. Specifically, the die observation camera 45 can be used to measure the height position of the die 51 held by the bonding head 44, the inclination of the die 51, and / or the flatness of the bonding surface. Note that in this embodiment, the pattern provided on the bonding surface of the die 51 can be defined as including a circuit pattern and a mark for measuring the position of the die 51.

[0039] Next, the mechanism mounted on the upper base 42 will be described. The bonding head 44, the wafer observation camera 46, and the interferometer 47 are mounted on the upper base 42.

[0040] Bonding head 44 (second holder) holds die 51 transferred from pickup head 31 using a vacuum force or the like, and drives die 51 in the -Z direction to bond die 51 to wafer 6. In this embodiment, bonding of die 51 to wafer 6 is performed by bonding head 44 driving die 51 in the -Z direction, but the present invention is not limited thereto. For example, bonding of die 51 to wafer 6 can be performed by stage 43 driving wafer 6 in the +Z direction, or by bonding head 44 and stage 43 driving die 51 and wafer 6 relative to each other.

[0041] The wafer observation camera 46 (first image capture device) is a camera used to observe the bonding surface (first bonding surface) of the wafer 6. The wafer observation camera 46 can be configured to capture an image of the wafer 6 while it is held by the stage 43 (first holder). The wafer observation camera 46 is used to obtain (measure) information indicating the position of a pattern (first pattern) provided on the bonding surface of the wafer 6, as well as information indicating the position of a characteristic portion of the die 51 relative to the wafer 6 after the wafer 6 and the die 51 are bonded. The wafer observation camera 46 is also used to measure the distances between multiple points in the height direction (Z direction) of the bonding surface of the wafer 6, that is, the height distribution of the bonding surface of the wafer 6. In other words, the wafer observation camera 46 can be used to measure the height position of the wafer 6 held by the stage 43, the tilt of the wafer 6, and / or the flatness of the bonding surface. The interferometer 47 illuminates a mirror 433 provided on the stage 43 with light and measures the position of the stage 43 based on the light reflected by the mirror 433.

[0042] The bonding apparatus 100 according to this embodiment is configured to rotate the (flip-chip) pickup head 31 and transfer the die 51 to the bonding head 44, but is not limited thereto. For example, two or more die holders may be provided between the pickup head 31 and the bonding head 44, and flip-chip bonding may be performed by transferring the die 51 to the two or more die holders, and then transferring the die 51 to the bonding head 44. A drive mechanism for driving the bonding head 44 may also be provided between the pickup unit 3 and the bonding unit 4, allowing the bonding head 44 to move to the bonding unit 4 and receive the die 51. To achieve higher productivity, the bonding apparatus 100 may include a plurality of pickup units 3, a plurality of pickup heads 31, a plurality of release heads 32, and a plurality of bonding heads 44.

[0043] The controller CNT is formed, for example, by a computer (information processing device) including a processor such as a central processing unit (CPU) and storage such as memory. The controller CNT controls the bonding process by controlling each unit of the bonding apparatus 100. The bonding process involves aligning the wafer 6 and the die 51 so that the pattern (first pattern) on the wafer 6 and the pattern (second pattern) on the die 51 overlap, and then bonding the die 51 to the wafer 6. More specifically, the controller CNT determines the position of the pattern (first pattern) provided on the bonding surface of the wafer 6 based on an image of the bonding surface (first bonding surface) of the wafer 6 captured by the wafer observation camera 46. Furthermore, the controller CNT determines the position of the pattern (second pattern) provided on the bonding surface of the die 51 based on an image of the bonding surface (second bonding surface) of the die 51 captured by the die observation camera 45. The controller CNT can control the bonding process based on the positions of the patterns on the wafer 6 and the die 51.

[0044] Figure 2 shows stage 43 as viewed from the +Z direction. Wafer 6 is held by chuck 431. To perform two-dimensional positioning, stage 43 includes a strip mirror 433x for position measurement in the X and θZ directions (rotational directions), and a strip mirror 433y for position measurement in the Y direction. Strip mirror 433x serves as a target for interferometers 47a and 47c, which measure position in the X direction. Interferometers 47a and 47c are spaced apart in the Y direction. The amount of rotation of stage 43 (in the θZ direction) can be determined from the difference between the measurement results of interferometers 47a and 47c. Strip mirror 433y serves as a target for interferometer 47b, which measures position in the Y direction. Interferometers 47a to 47c simultaneously measure the position of stage 43 in the X and Y directions, as well as the amount of rotation in the θZ direction. Controller CNT can perform real-time feedback control on the drive of stage 43 and accurately two-dimensionally position stage 43. In bonding apparatus 100 according to this embodiment, interferometers 47a to 47c accurately measure the position of stage 43, and based on the results of this position measurement, feedback control is performed on the drive mechanism of stage 43, thereby functioning as a positioning mechanism for stage 43.

[0045] A reference plate 434, including a plurality of marks 434a to 434c, is mounted on stage 43. Reference plate 434 is made of a material with a low thermal expansion coefficient and includes marks 434a to 434c formed (drawn) with high positional accuracy. For example, reference plate 434 can be formed by drawing marks on a quartz substrate using a drawing method used in semiconductor lithography processes. Reference plate 434 can be configured to have an upper surface that is substantially flush with the upper surface of wafer 6 held by stage 43 (chuck 431). In this embodiment, reference plate 434 can be observed by wafer observation camera 46, but the present invention is not limited to a separate reference plate observation camera. Stage 43 can be composed of a coarse movement stage capable of wide movement and a fine movement stage capable of precise movement within a small range on the coarse movement stage. In this case, the die observation camera 45, mirror 433, chuck 431, and reference plate 434 require precise positioning and are therefore preferably fixed to the fine movement stage.

[0046] A method of using the reference plate 434 to ensure the original position, magnification, rotation in the θZ direction, and orthogonality of the stage 43 will be described. When the wafer observation camera 46 is controlled to capture (observe) an image of the mark 434a, the controller CNT obtains the measurement values ​​of the interferometers 47a to 47c when the mark 434a is arranged at the center of the image obtained by the wafer observation camera 46, and sets the measurement values ​​as the origin of the stage 43. Then, when the wafer observation camera 46 is controlled to capture (observe) an image of the mark 434b, the controller CNT obtains the measurement values ​​of the interferometers 47a to 47c when the mark 434b is arranged at the center of the image obtained by the wafer observation camera 46. Based on the obtained measurement values, the controller CNT determines the Y-axis direction and Y magnification of the stage 43. Next, when the wafer observation camera 46 is controlled to capture (observe) an image of the mark 434c, the controller CNT obtains the measurement values ​​of the interferometers 47a to 47c when the mark 434c is arranged at the center of the image obtained by the wafer observation camera 46. Based on the obtained measurement values, the controller CNT determines the X-axis direction and X magnification. Specifically, the direction from mark 434b toward mark 434c on reference plate 434 is defined as the Y-axis of bonding apparatus 100, and the direction from mark 434a toward mark 434c is defined as the X-axis of bonding apparatus 100. Calibration is performed for the axis directions and orthogonality. Furthermore, the interval between mark 434b and mark 434c is defined as the scale reference for bonding apparatus 100 in the Y-direction, and the interval between mark 434a and mark 434c is defined as the scale reference for bonding apparatus 100 in the X-direction. Calibration is also performed. Due to changes in atmospheric pressure and temperature, the refractive index of the interferometer's optical path changes, causing changes in the measured value. Therefore, interferometers 47a to 47c are preferably calibrated at all times to ensure the origin position, magnification, rotation, and orthogonality of stage 43. To reduce variations in the measured values ​​of interferometers 47a to 47c, the space in which stage 43 moves can be covered with a temperature-controlled chamber to control the temperature.

[0047] In this embodiment, the reference plate 434 is attached to the stage 43 and an image of the reference plate 434 is captured (observed) by the wafer observation camera 46. However, the present invention is not limited to this. For example, the reference plate 434 can be attached to the upper base 42 so that the image of the reference plate 434 can be captured (observed) by the wafer observation camera 45. Even in this configuration, the origin position, magnification, rotation, and orthogonality of the stage 43 can be guaranteed. In this embodiment, the calibration is performed by capturing (observing) an image of the reference plate 434. However, the present invention is not limited to this. For example, calibration can be performed by contacting the reference surface. Alternatively, a position measurement mechanism using a white interferometer that ensures absolute values ​​can be used to accurately position the stage 43.

[0048] In the bonding apparatus 100 having the above configuration, after bonding the die 51 to the wafer 6, it is important to minimize the relative positional deviation between the pattern provided on the bonding surface of the wafer 6 (first member) and the pattern provided on the bonding surface of the die 51. However, after bonding the die 51 to the wafer 6, the bonding surfaces of the wafer 6 and die 51 cannot be visually inspected. Therefore, it is difficult to easily and accurately determine the relative position between the pattern on the wafer 6 and the pattern on the die 51 after bonding. Here, the wafer observation camera 46 may be an infrared camera. In this method, infrared light is transmitted through the chip 51 and an image is captured, thereby measuring the relative position between the pattern on the wafer 6 and the pattern on the die 51 after bonding. However, such infrared cameras are expensive and may be detrimental to the cost of the bonding apparatus 100. Even with an infrared camera, the pattern on the wafer 6 and the pattern on the die 51 after bonding cannot be clearly captured, making it difficult to accurately measure the relative position between the pattern on the wafer 6 and the pattern on the die 51 after bonding.

[0049] In the bonding apparatus 100 according to this embodiment, after the bonding process between wafer 6 (the first component) and die 51 (the second component), controller CNT controls wafer observation camera 46 (the first image capture device) to capture an image of the appearance of die 51 bonded to wafer 6. Based on the image captured by wafer observation camera 46, controller CNT obtains characteristic position information indicating the position of a characteristic portion of die 51 relative to wafer 6. Next, controller CNT estimates the relative position of the pattern on wafer 6 and the pattern on die 51 after the bonding process based on the characteristic position information and positional relationship information indicating the previously acquired positional relationship between the characteristic portion of die 51 and the pattern on die 51. Consequently, the patterns of wafer 6 and die 51 after the bonding process can be easily and accurately acquired. An example of this embodiment will be described below.

[0050] [Example 1] The operation of the bonding apparatus 100 in Example 1 will be described with reference to Figures 3A, 3B, and 4. Figure 3A is a schematic diagram showing a state in which an image of a die 51 is captured. Figure 3B is a schematic diagram showing an example configuration of the die 51 used in Example 1. As shown in Figure 3B, the die 51 used in Example 1 has a bonding surface 51a and a back surface 51b opposite the bonding surface. The bonding surface 51a has a device pattern 501 (second pattern) and alignment marks 502. The die 51 has through-holes 503 extending from the bonding surface 51a to the back surface 51b. Although Figure 3B shows the through-holes 503 only on the back surface 51b, it is understood that at least some of the through-holes 503 may also be exposed on the bonding surface 51a. In this case, the through-holes 503 can function as the aforementioned feature portion. Figure 4 is a flow chart showing the operating sequence of the bonding apparatus 100 in Example 1. The controller CNT can execute the process shown in the flow chart of Figure 4.

[0051] In step S101, the controller CNT uses a wafer transport mechanism (not shown) to load the wafer 6, serving as the first component (the target for first bonding), onto the stage 43 (chuck 431) of the bonding apparatus 100. At this time, foreign matter adhering to the bonding surface of the wafer 6 can cause bonding failures, so it is desirable to maintain the space within the bonding apparatus 100 at a high cleanliness level of approximately Class 1. To maintain high cleanliness even for the wafer 6, it is desirable to store the wafer 6 in a highly airtight container that maintains a high cleanliness level and load the wafer 6 from this container onto the stage 43 of the bonding apparatus 100. The container is, for example, a front-opening unified unit box (FOUP).

[0052] To improve the cleanliness of wafer 6, a cleaning mechanism for cleaning wafer 6 can be provided in bonding apparatus 100. Bonding apparatus 100 can also include a mechanism for performing pre-processing on wafer 6 before the bonding process. For example, pre-processing involves applying adhesive to the bonding surface of wafer 6 in bonding using an adhesive, or activating the bonding surface of wafer 6 in hybrid bonding. After the position of wafer 6 in the θZ direction and the X and Y directions is measured by a pre-alignment unit (not shown), wafer 6 is roughly positioned based on the measurement results and transferred to chuck 431 of stage 43. The position of wafer 6 in the θZ direction can be measured by detecting the notch or orientation flat of wafer 6, and the position of wafer 6 in the X and Y directions can be measured by detecting the shape of wafer 6.

[0053] In step S102, the controller CNT uses the wafer observation camera 46 to perform wafer alignment. During wafer alignment, the wafer observation camera 46 captures an image of the bonding surface of the target region (bonding target) of the die 51 to be bonded on the wafer 6. Based on the captured image, the position of the pattern (first pattern) provided on the wafer 6 (target region) is determined. Note that the wafer 6 has a plurality of target regions.

[0054] Focus adjustment when imaging the bonding surface of wafer 6 can be performed using a focus adjustment mechanism provided on wafer observation camera 46 or by driving wafer 6 in the Z direction using the Z drive mechanism of stage 43. When alignment marks are provided on the bonding surface of wafer 6, the position of the pattern on wafer 6 can be determined using the alignment marks. Conversely, when alignment marks are not provided on the bonding surface of wafer 6, the position of the pattern on wafer 6 can be determined using a portion of the bonding surface that allows the position of the pattern to be specified (hereinafter sometimes referred to as a specifiable portion). For example, a portion of the pattern on wafer 6 can be used as the specifiable portion.

[0055] For example, the controller CNT can measure the position of the pattern on the wafer 6 by measuring the image position of an alignment mark or a specifiable portion relative to the center of the image captured by the wafer observation camera 46. Alignment marks or specifiable portions are sometimes referred to as alignment marks. For example, there is a method for accurately measuring the position of alignment marks relative to a reference point of the bonding apparatus 100. According to this method, the stage 43 is driven so that a mark formed on the reference plate 434 falls within the image capture field of view of the wafer observation camera 46, and the wafer observation camera 46 captures an image of the mark on the reference plate 434. Based on the position of the stage 43 at that time and the position of the mark in the image captured by the wafer observation camera 46, the reference point of the bonding apparatus 100 is determined. Based on the image captured by the wafer observation camera 46 of the alignment mark, the offset of the alignment mark relative to the reference point is determined. Therefore, the position of the alignment mark can be accurately measured based on the position of the reference point and the offset. In Example 1, the position of the mark on the reference plate 434 is used as the reference point of the bonding apparatus 100. However, the position of another location can be used as a reference.

[0056] Because interferometer 47 has a narrow measurement range in the θZ direction, the amount of rotation in the θZ direction that can be corrected by stage 43 is relatively small. If the amount of rotation of wafer 6 in the θZ direction is large, wafer 6 is preferably repositioned on stage 43 to correct the amount of rotation of wafer 6 in the θZ direction. When wafer 6 is repositioned on stage 43, the position of wafer 6 needs to be measured again. During step S102, the surface position of wafer 6 is preferably measured using a first height measurement mechanism (not shown) that measures the surface position of the bonding surface of wafer 6. This is because the thickness of wafer 6 varies, and the surface position of wafer 6 is important for accurately managing (controlling) the gap between wafer 6 and die 51 during the bonding process.

[0057] Because a reference plate is used to ensure the origin position, magnification, X and Y positions, θZ rotation, and orthogonality of stage 43, the position of wafer 6 mounted on stage 43 relative to the origin position of stage 43 can be measured. On wafer 6, target areas (bonding targets or target regions) for forming semiconductor devices are repeatedly arranged at a predetermined period. Specifically, wafer 6 includes a plurality of target areas, each of which is bonded to a die 51. Semiconductor devices are precisely positioned and manufactured in each target area using semiconductor manufacturing equipment. The plurality of target areas on wafer 6 are typically arranged in a repeating period with nanometer-level precision. For this reason, during wafer alignment in step S102, it is not necessary to measure the positions of all target areas on wafer 6; only the positions of the bonding targets of some target areas on wafer 6 are measured. More specifically, the positions of semiconductor devices (patterns or marks) in three or more of the plurality of target areas on wafer 6 are measured, and statistical processing is performed. Therefore, the array of the target area, the origin position of the array, the position in the X and Y directions, the rotation amount in the θZ direction, the orthogonality, and the magnification error of the repetition period can be calculated.

[0058] Chuck 431 may include a mechanism for controlling the temperature of wafer 6. This is because, given a silicon wafer's thermal expansion coefficient of 3 ppm / °C and a 300 mm wafer diameter, a 1°C temperature increase causes the outermost periphery to shift by 150 mm x 0.000003 mm = 0.00045 mm = 450 nm. If the bonding position (e.g., the target area) is shifted after wafer alignment, it may be difficult to precisely bond wafer 6 and die 51. Therefore, it is preferable to control the temperature of wafer 6 to keep the temperature variation within 0.1°C or less.

[0059] Note that in Example 1, wafer 6 was used as the first member (the target of the first bonding). If an interposer with wiring formed thereon were used as the first member, the array of repeatedly formed wiring would be measured, not the array of semiconductor devices. If a wafer or plate without a pattern is used as the first member, wafer alignment in step S102 is not necessary.

[0060] The above steps S101 and S102 are processes for the wafer 6 serving as the first member (the first bonding target). In parallel with steps S101 and S102, processes (steps S201 to S203) for the die 51 serving as the second member (the second bonding target) are performed.

[0061] In step S201, the controller CNT uses a conveyor mechanism (not shown) to load the dicing frame 5 onto the pickup unit 3 (on the frame holder 33). The dicing frame 5 is a frame with an opening in the center, and a dicing tape is adhered to the dicing frame 5 to cover the opening. A plurality of dies 51, separated by a cutter such as a dicer, are arranged on the dicing tape. Conventionally, the dicing frame 5 is conveyed from an unsealed magazine. However, if foreign matter adheres to the bonding surface 51a of the die 51, it can cause bonding failure. Therefore, the dicing frame 5 must be conveyed within a container that is airtight and maintains a high degree of cleanliness. To improve the cleanliness of the die 51, a cleaning mechanism for cleaning the die 51 on the dicing frame 5 (dicing tape) can be provided within the bonding apparatus 100. After a pre-alignment unit (not shown) roughly determines the dicing frame 5's rotation in the θZ direction and its displacement position (position in the X and Y directions) based on its external shape, the dicing frame 5 can be conveyed to the frame holder 33.

[0062] In step S202, the controller CNT controls the pickup head 31 and the release head 32 to pick up a die 51 from the dicing frame 5 (dicing tape). More specifically, the controller CNT moves the pickup head 31 and the release head 32 to the position of the die 51 to be picked up (hereinafter referred to as the target die 51). The controller CNT drives the release head 32 in the +Z direction to push the target die 51 upward from the back side of the dicing tape. In this state, the controller CNT drives the pickup head 31 in the -Z direction, bringing the pickup head 31 and the target die 51 into contact with each other. The target die 51 is then held (clamped) by the pickup head 31 using a vacuum force, etc., and can be peeled from the dicing tape by driving the pickup head 31 in the +Z direction. The target die 51 to be picked up is determined based on information about non-defective dies (known good dies: KGD) transmitted online to the bonding equipment 100. Typically, only non-defective dies are picked up as target dies 51. However, for a target area on the wafer 6 having defective devices, a defective die (known as a bad die: KBD) may be picked up as a target die 51 .

[0063] In step S203, the controller CNT delivers (transfers) the target die 51 picked up by the pickup head 31 to the bonding head 44 of the bonding unit 4. More specifically, as shown in FIG1 , the controller CNT drives the pickup head 31 in the X direction to pick up the target die 51 and positions the pickup head 31 below the bonding head 44. The controller CNT then drives the pickup head 31 in the +Z direction to transfer the target die 51 from the pickup head 31 to the bonding head 44. When the target die 51 is picked up by the pickup head 31, the bonding surface 51a of the target die 51 is oriented in the +Z direction and in contact with the pickup head 31. Conversely, when transferring the target die 51 from the pickup head 31 to the bonding head 44, the bonding surface 51a of the target die 51 needs to be oriented in the -Z direction. To achieve this, the pickup head 31 can be driven to rotate, causing the target die 51 to be flipped (flip-chipped) while being transferred to the bonding head 44.

[0064] While Example 1 describes an example in which the pickup head 31 directly transports the target die 51 to the bonding head 44, the present invention is not limited thereto. For example, if one or more transport mechanisms are provided along the transport path from the target die 51 to the bonding head 44, the target die 51 can be transported to the bonding head 44 by passing the target die 51 to the one or more transport mechanisms. A mechanism for performing pre-processing on the target die 51 during the bonding process can be provided within the bonding apparatus 100. Pre-processing can include, for example, applying adhesive to the bonding surface 51a of the target die 51 in adhesive bonding or activating the bonding surface 51a of the target die 51 in hybrid bonding. The target die 51 can also be cleaned as pre-processing. Pre-processing can also be performed while the target die 51 is being transported to the bonding head 44.

[0065] Through the above process, the wafer 6 is held by the stage 43 , and the target die 51 is held by the bonding head 44 .

[0066] Subsequently, in step S103, the controller CNT performs die alignment using the die observation camera 45. During die alignment, as shown in FIG3A , the die observation camera 45 is positioned below the target die 51 held by the bonding head 44 by driving the stage 43 on which the die observation camera 45 is mounted. The die observation camera 45 captures an image of the bonding surface 51a of the target die 51 and, based on the captured image, determines the position of the pattern 501 (second pattern) disposed on the bonding surface 51a of the die 51.

[0067] Focus adjustment when imaging the bonding surface 51a of the target die 51 can be performed using a focus adjustment mechanism provided on the die observation camera 45 or by driving the die observation camera 45 in the Z direction using the Z drive mechanism of the stage 43. If the Z drive mechanism is provided on the bonding head 44, focus adjustment can be performed by driving the target die 51 in the Z direction using the Z drive mechanism of the bonding head 44. In Example 1, alignment marks 502 are provided on the bonding surface 51a of the target die 51. Therefore, the position of the pattern 501 of the target die 51 can be determined using alignment marks 502. For typical dies, alignment marks are often provided on the scribe line and removed along with the scribe line. In this case, the position of the die pattern can be determined using a portion of the bonding surface that allows the pattern's position to be specified (hereinafter sometimes referred to as a specifyable portion). Examples of specifyable portions include the ends of an array of pads or bumps arranged on the bonding surface, a region having a non-periodic array, or the outer edge (outline) of the die.

[0068] For example, the controller CNT can measure the position of the pattern 501 of the target die 51 by measuring the image position of the projected alignment mark 502 or the specifiable portion relative to the center of the image captured by the die observation camera 45. Measuring the position of the target die 51 can include measuring the rotation amount of the target die 51 (rotation along the θZ direction). The rotation amount of the target die 51 can be measured, for example, by acquiring the position of each specifiable portion on the bonding surface 51a of the target die 51 based on the image captured by the die observation camera 45. The position of each specifiable portion can be acquired based on a plurality of images acquired by individually capturing each specifiable portion while the die observation camera 45 is driven by the stage 43. Alternatively, when the entire target die 51 falls within the imaging field of view of the die observation camera 45, the position of each specifiable portion can be acquired by capturing an image of the entire bonding surface 51a of the target die 51 by the die observation camera 45. During the bonding process, the rotation of the target die 51 can be corrected by rotating the wafer 6 using the stage 43. However, the interferometer 47 has a narrow measurement range in the θZ direction. Therefore, if the rotation of the target die 51 is large, it is desirable to reposition the target die 51 on the bonding head 44 to correct the rotation of the target die 51. When the target die 51 is repositioned on the bonding head 44, the position of the target die 51 must be measured again.

[0069] During step S103, the surface position of the bonding surface 51a of the target die 51 is preferably measured using a second height measurement mechanism (not shown) that measures the surface position of the bonding surface 51a of the target die 51. Because the thickness of the target die 51 varies, the surface position of the target die 51 is important for accurately managing (controlling) the gap between the wafer 6 and the target die 51 during the bonding process. Furthermore, the heights of multiple locations on the bonding surface 51a of the target die 51 (i.e., the height distribution of the bonding surface 51a) can be measured to adjust the relative posture of the wafer 6 and the target die 51 based on the measurement results during the bonding process. This relative posture can be adjusted using a tilt mechanism mounted on the stage 43 and / or the bonding head 44.

[0070] In step S103, the controller CNT uses the die observation camera 45 to obtain positional relationship information. As described above, the positional relationship information is information indicating the positional relationship between the characteristic portion of the target die 51 and the pattern 501 of the target die 51. As described above, the characteristic portion of the die 51 is an indicator that can be identified in both an image obtained by photographing the bonding surface 51a of the target die 51 and an image obtained by photographing the back surface 51b of the target die 51. The characteristic portion of the target die 51 is the outer edge (external shape) of the target die 51 and / or the through hole 503. For example, the controller CNT can control the die observation camera 45 to photograph the bonding surface 51a of the target die 51 and obtain the positional relationship between the characteristic portion of the target die 51 and the pattern 501 of the target die 51 in the obtained image as the positional relationship information. When the outer edge of the target die 51 is used as the characteristic portion, the die observation camera 45 is preferably configured so that the entire target die 51 falls within the field of view of the image. In Example 1, the example of obtaining positional relationship information using the die observation camera 45 in step S103 has been described. However, the positional relationship information only needs to be obtained before the bonding process (step S105) described later, and the positional relationship information can be obtained in advance using, for example, an external device.

[0071] In step S104, controller CNT drives stage 43 to align wafer 6 and target die 51 so that the pattern on wafer 6 and pattern 501 on target die 51 overlap. More specifically, controller CNT drives stage 43 so that the target area of ​​target die 51 to be bonded on wafer 6 is positioned below target die 51 held by bonding head 44. Then, controller CNT aligns wafer 6 and target die 51 based on the position of the pattern on wafer 6 obtained in step S102 and the position of pattern 501 on target die 51 obtained in step S103. At this point, wafer 6 and target die 51 are preferably aligned to minimize relative rotational deviation and / or posture deviation between them. If the relative position between wafer 6 and target die 51 in the X and Y directions changes (shifts) during the bonding process described later, this change in relative position can be used as an offset to align wafer 6 and target die 51. This offset can be determined in advance through experiments, simulations, etc.

[0072] In step S105, controller CNT bonds target die 51 to wafer 6 by reducing the gap between wafer 6 and target die 51 (bonding process). The bonding process can be performed by driving target die 51 in the Z direction by bonding head 44 or by driving wafer 6 in the Z direction using stage 43. Alternatively, the bonding process can be performed by driving target die 51 and wafer 6 relative to each other in the Z direction by bonding head 44 and stage 43. To precisely control the gap (gap) between wafer 6 and target die 51, a detector (e.g., an encoder) can be provided to detect the position of bonding head 44 and / or stage 43 in the Z direction. During the bonding process, ultrasonic waves can be applied to bonding head 44 and / or stage 43 while wafer 6 and target die 51 are in contact with each other (i.e., ultrasonic bonding can be employed). After bonding wafer 6 and target die 51, controller CNT releases bonding head 44 from holding target die 51 and widens the gap between wafer 6 and target die 51. Note that the bonding process can be understood as including the alignment in step S104 described above.

[0073] To improve the alignment accuracy between wafer 6 and target die 51 even during the bonding process, the relative position between wafer 6 and target die 51 in the X and Y directions can be controlled. For example, when the relative position between wafer 6 and target die 51 in the X and Y directions is controlled by driving stage 43, the width of mirror 433 in the Z direction is preferably set so that light from interferometer 47 can illuminate mirror 433 even when stage 43 is driven in the Z direction. Alternatively, a detector (e.g., an encoder or gap sensor) can be provided to detect the relative position between bonding head 44 and stage 43 in the X and Y directions. In this case, during the bonding process, the relative position between bonding head 44 and stage 43 in the X and Y directions can be detected (monitored) by the detector while feedback control of the relative position is performed. If wafer 6 and target die 51 come into contact, the position of stage 43, which is feedback-controlled based on the measurement results of interferometer 47, is limited. Therefore, the method for controlling the relative position between the wafer 6 and the target die 51 in the X and Y directions is preferably switched before and after contact by, for example, stopping the feedback process when contact between the wafer 6 and the target die 51 begins. In bump bonding, in step S105, the process required for bump bonding can be performed by, for example, pressing the target die 51 against the wafer 6 with a predetermined pressure (pressing pressure).

[0074] In step S106, controller CNT determines whether die 51 has been bonded to all target regions on wafer 6. Typically, dozens to hundreds of semiconductor devices are formed on a single wafer 6 as multiple target regions, and die 51 can be bonded to each target region. If a target region (next target region) to which die 51 is to be bonded exists on wafer 6, the process returns to step S202. If no next target region exists on wafer 6, meaning die 51 has been bonded to all target regions on wafer 6, the process proceeds to step S107.

[0075] In Example 1, we explained an example in which the presence of the next target region is determined after the bonding process, and the process returns to step S202. However, the determination of the presence of the next target region can be performed before the bonding process ends. In this case, step S202 can be performed in parallel with the bonding process. That is, in parallel with the bonding process, the die 51 to be bonded to the next target region is picked up from the dicing frame 5 (dicing tape). When bonding multiple types of die 51 to each target region (semiconductor device) on the wafer 6, one type of die is bonded to all target regions on the wafer 6, and then the bonding of the next type of die begins. When bonding the next type of die begins, the dicing frame 5 with the next type of die arranged thereon is loaded (step S201), and then the die picking in step S202 can be performed.

[0076] In step S107, controller CNT executes an estimation process to estimate the relative position between the pattern on wafer 6 and the pattern 501 on target die 51 after the bonding process. The estimation process may include steps S107a through S107c. In Example 1, the estimation process may be performed for each target region on wafer 6, i.e., each die 51 bonded to wafer 6.

[0077] In step S107a, as shown in FIG3A , controller CNT controls wafer observation camera 46 to capture an image of the exterior of die 51 bonded to wafer 6. The image obtained includes the back surface 51b of die 51 bonded to wafer 6 and the portion of wafer 6 surrounding die 51. In step S107b, controller CNT obtains feature position information based on the image obtained in step S107a. As described above, feature position information is information indicating the position of a feature portion of die 51 relative to wafer 6. The feature portion of die 51 is the outer edge (external shape) of die 51 and / or through-hole 503. In step S107b, the position of the feature portion on back surface 51b of die 51 is obtained.

[0078] In step S107c, controller CNT estimates (calculates) the relative position between the pattern on wafer 6 and the pattern 501 on die 51 after the bonding process based on the characteristic position information obtained in step S107b and the positional relationship information obtained in step S103. More specifically, controller CNT can estimate the relative position by converting the position of the characteristic portion of die 51 in the characteristic position information into the position of pattern 501 on die 51 based on the positional relationship information.

[0079] In step S107d, the controller CNT outputs the estimated result from step S107c. The estimated result can be output by displaying information on a user interface (e.g., a display) of the bonding apparatus 100 or by transmitting the estimated result information to an external computer. The controller CNT can output information indicating the relative position between the pattern on the wafer 6 and the pattern 501 on the die 51 after the bonding process (e.g., numerical information indicating a deviation in the relative position) as the estimated result, or output information indicating the estimated result of the relative position. The evaluation result can be, for example, an evaluation of whether the estimated deviation in the relative position falls within an allowable range.

[0080] FIG5 schematically illustrates the bonding between wafer 6 and die 51. Reference numeral 5A in FIG5 shows the bonding surface of target area 6a on wafer 6, with pattern 601 represented by nine circles. Reference numeral 5B in FIG5 shows the bonding surface of die 51, with pattern 501 represented by lines. Reference numeral 5C in FIG5 shows the relative position between pattern 601 on wafer 6 and pattern 501 on die 51, as identified by controller CNT during alignment in step S104. During alignment in step S104, controller CNT identifies the relative position based on the position of pattern 601 on wafer 6, previously determined in step S102, and the position of pattern 501 on die 51, previously determined in step S103. That is, during alignment in step S104, controller CNT does not actually observe pattern 601 on wafer 6 or pattern 501 on die 51. Reference numeral 5D in FIG5 shows the relative position between the pattern 601 of the wafer 6 and the pattern 501 of the die 51 after the bonding process. Note that the relative position between the pattern 601 of the wafer 6 and the pattern 501 of the die 51 will be simply referred to as the "relative position" in some cases.

[0081] (Case 1) in Figure 5 represents an ideal state in which the relative position identified by the controller CNT during alignment and the relative position after the bonding process are consistent. However, in reality, there may be errors between the relative position identified by the controller CNT during alignment and the relative position after the bonding process, as shown in (Case 2) to (Case 4) in Figure 5. According to Example 1, the estimation process in step S107 allows for easy and highly accurate estimation of the relative position after the bonding process. The estimated relative position after the bonding process can be used, for example, to understand (identify) the bonding state between wafer 6 and die 51 or to provide feedback to a later bonding process.

[0082] In step S108, controller CNT uses a wafer transport mechanism (not shown) to unload wafer 6, to which die 51 has been bonded, from stage 43 (chuck 431). Wafer 6 can be returned to the FOUP used to load wafer 6, or to a container other than the FOUP. However, since the thickness of the entire wafer 6 to which die 51 is bonded varies, it is preferably returned to another container. The operating sequence of bonding apparatus 100 for bonding die 51 to a single wafer 6 has been described above. The flowchart of FIG. 4 is repeated as die 51 is bonded to each wafer 6.

[0083] Since the number of dies 51 on the dicing frame 5 and the number of target areas on the wafer 6 are generally different, the loading of the wafer 6 and the loading of the dicing frame 5 are often asynchronous. If the dies 51 on the dicing frame 5 are exhausted during the bonding of the dies 51 to a wafer 6, the next dicing frame 5 can be loaded into the bonding apparatus 100. If the dies 51 remain on the dicing frame 5 even after the bonding of the dies 51 to a wafer 6 is completed, the remaining dies 51 can be used for the next wafer 6.

[0084] In the flowchart of Figure 4 , after the determination in step S106 , the estimation process in step S107 is performed on each target area on wafer 6 (i.e., each die 51 bonded to wafer 6 ). However, as shown in the flowchart of Figure 6 , the estimation process in step S107 can be performed immediately after the bonding process in step S105 on the die 51 bonded to wafer 6 in the bonding process. This is to reduce idle waiting time in the bonding unit 4 , such as when it takes time to pick up the die 51 in step S202 . This allows the bonding process in step S105 and the estimation process in step S107 to be flexibly switched or interchanged depending on the timing of picking up the die 51 in step S202 . Some bonding methods require a suitable amount of time after the bonding process in step S105 until the bonded state of the die 51 on wafer 6 stabilizes. This suitable time is, for example, the curing time of the adhesive in adhesive bonding or the time from the start to the end of covalent bonding on an activated bonding surface in room temperature bonding using a surface activation method. If the estimation process in step S107 is performed before the bonding state stabilizes, there may be a discrepancy between the final bonding state and the estimated result. To address this, it is preferable to set an arbitrary delay time between the bonding process in step S105 in FIG6 and the estimation process in step S107. Note that steps S101 to S108 in FIG6 are similar to steps S101 to S108 in FIG4 , and their detailed description will be omitted.

[0085] [Example 2] In Example 2, we will explain how the estimated results obtained in step S107 described in Example 1 are processed. Specifically, we will explain an example of outputting the estimated results obtained in step S107d. Note that Example 2 essentially inherits from Example 1 described above and conforms to Example 1 with the following exceptions.

[0086] The simplest processing of the estimated results is to display the estimated results in step S107 on the user interface (e.g., a display) of the bonding equipment 100 as information indicating the bonding accuracy between the pattern on wafer 6 and the pattern on die 51. The estimated results can also be output as supplementary information for the wafer 6 unloaded in step S108. The estimated results can be displayed on the user interface as information indicating the bonding accuracy of all or some of the dies 51 bonded to wafer 6. In many semiconductor manufacturing sites, in addition to the bonding equipment 100, manufacturing equipment and processing equipment used to perform other processes are also connected via a network, and information is exchanged online between these devices. Even the bonding equipment 100 can publish information indicating the bonding accuracy of all or some of the dies bonded to the wafer online, or output information indicating bonding accuracy online at any time in response to an online information request. All information obtained during the process described in the flowcharts of Figures 4 and 6 described in Example 1 can also be output online.

[0087] As described above, after the bonding process between wafer 6 (first component) and die 51 (second component), bonding apparatus 100 according to this embodiment controls wafer observation camera 46 (first image capture device) to capture the appearance of die 51 bonded to wafer 6. Bonding apparatus 100 estimates the relative position between the pattern on wafer 6 and the pattern on die 51 after the bonding process based on feature position information obtained from the image captured by wafer observation camera 46 and previously acquired positional relationship information. Consequently, the patterns of wafer 6 and die 51 after the bonding process can be easily and accurately acquired.

[0088] <Second embodiment> A second embodiment according to the present invention will be described. The second embodiment is substantially a successor to the first embodiment and can follow the first embodiment except for the matters mentioned below.

[0089] FIG7 is a schematic diagram of a bonding apparatus 100' according to a second embodiment. The bonding apparatus 100' according to this embodiment includes a die recovery unit 61 (die recovery container) that recovers die 51 that have undergone poor bonding in the bonding unit 4. The bonding apparatus 100' according to the second embodiment has a similar configuration to the bonding apparatus 100 according to the first embodiment, except that the bonding apparatus 100' includes the die recovery unit 61. Description of components other than the die recovery unit 61 will be omitted. FIG7 does not show the pickup unit 3 and the controller CNT.

[0090] The die recovery unit 61 has a structure capable of receiving the die 51 from the pickup head 31 and can be configured to hold (store) a plurality of die 51 received from the pickup head 31. The die recovery unit 61 can be mounted on the base 1 on which the bonding unit 4 is mounted, or can be mounted on another base or a mechanism independent of the other components. The die recovery unit 61 can be configured to be easily separated from the bonding apparatus 100'. For example, the die recovery unit 61 can be configured to be removed from the bonding apparatus 100' while storing the recovered die 51. An example of this embodiment will be described below.

[0091] [Example 3] FIG8 is a flowchart showing the operation sequence of the bonding apparatus 100′ in Example 3. The controller CNT can execute the process in the flowchart of FIG8 . Note that steps S101 to S108 and S201 to S203 in the flowchart of FIG8 are the same as those described in the first embodiment using the flowcharts of FIG4 and FIG6 , and their detailed description will be omitted.

[0092] In step S301, controller CNT determines whether the relative position deviation between the pattern on wafer 6 and the pattern 501 on target die 51 after the bonding process falls within an allowable range based on the estimated process results from step S107. The allowable range can be pre-set based on the line width, size, etc. of the pattern on wafer 6 and / or the pattern on target die 51. The allowable range can also be set based on the electrical characteristics of the objects to be bonded between wafer 6 and target die 51. If the relative position deviation falls within the allowable range, the process proceeds to step S106. If the relative position deviation does not fall within the allowable range, the process proceeds to step S302. Note that a die 51 whose relative position deviation does not fall within the allowable range can be considered a die 51 with a bonding defect. A die 51 whose relative position deviation does not fall within the allowable range is sometimes referred to as a "die 51 with a bonding defect."

[0093] In step S302, the controller CNT separates the poorly bonded die 51 from the wafer 6 (separation process). More specifically, the controller CNT drives the stage 43 so that the poorly bonded die 51 on the wafer 6 is positioned below the bonding head 44. The controller CNT then narrows the gap between the bonding head 44 and the wafer 6 (the poorly bonded die 51). When the poorly bonded die 51 contacts the bonding head 44, the controller CNT controls the bonding head 44 to hold (pick up) the die 51. While the bonding head 44 holds the poorly bonded die 51, the controller CNT widens the gap between the bonding head 44 and the wafer 6. As a result, the poorly bonded die 51 can be separated from the wafer 6.

[0094] This separation process addresses the fact that die 51 can be separated from wafer 6 before the bonding between wafer 6 and die 51 is complete. That is, the separation process can be performed before the bonding between wafer 6 and die 51 is complete. For example, in room-temperature bonding using a surface activation method, covalent bonding between molecules does not begin until the gap between wafer 6 and die 51 becomes on the order of 0.1 nm or less. Wafer 6 and die 51 remain in a state of incomplete bonding, and die 51 can be separated from wafer 6. Typically, in room-temperature bonding using a surface activation method, the gap between wafer 6 and die 51 does not become on the order of 0.1 nm or less unless die 51 is pressed onto wafer 6. Therefore, in room-temperature bonding using a surface activation method, the estimation process in step S107 is performed before pressing die 51 onto wafer 6, and die 51 can be pressed onto wafer 6 based on the estimation results. If the estimated result is satisfactory (i.e., the deviation in the amount of rotation falls within the allowable range), die 51 is pressed onto wafer 6. If the estimated result is unsatisfactory (i.e., the deviation in the amount of rotation falls outside the allowable range), a separation process is performed. In adhesive bonding, before the adhesive hardens, wafer 6 and die 51 are in a state prior to being fully bonded, and die 51 can be separated from wafer 6. The separation process in step S302 is performed at a time when the bonded state can be released, according to the bonding method.

[0095] In step S303, the controller CNT delivers (transfers) the die 51 held by the bonding head 44 to the pickup head 31. More specifically, the controller CNT drives the pickup head 31 in the X direction to position the pickup head 31 below the bonding head 44. Then, the controller CNT drives the pickup head 31 in the +Z direction to transfer the die 51 from the bonding head 44 to the pickup head 31.

[0096] In step S304, the controller CNT moves the pickup head 31 to the die recovery unit 61 and delivers the die 51 held by the pickup head 31 to the die recovery unit 61. Thus, the separation and recovery of the die 51 with poor bonding is completed. After step S304 ends, the process proceeds to step S106. Note that the bonding process for a new die 51 can be performed again on the target area of ​​the wafer 6 that has been separated from the temporarily bonded die 51.

[0097] [Example 4] FIG9 is a flowchart showing the operation sequence of the bonding apparatus 100′ in Example 4. The controller CNT can execute the process in the flowchart of FIG9 . Note that steps S101 to S108 and S201 to S203 in the flowchart of FIG9 are the same as those described in the first embodiment using the flowcharts of FIG4 and FIG6 , and their detailed description will be omitted.

[0098] In step S301, controller CNT determines whether the relative positional deviation between the pattern on wafer 6 and the pattern 501 on target die 51 after the bonding process falls within an allowable range based on the estimated results of the estimation process in step S107. In step S302, controller CNT separates the die 51 with poor bonding from wafer 6 (separation process). Steps S301 and S302 are the same as those described in Example 3, and their detailed description will be omitted.

[0099] In step S401, the controller CNT re-bonds the die 51, separated from the wafer 6 in step S302, to the wafer 6. More specifically, the controller CNT re-aligns the wafer 6 and the die 51 to correct the relative position deviation based on the relative position deviation estimated in step S107. The controller CNT then re-bonds the die 51 to the wafer 6 by reducing the gap between the wafer 6 and the die 51. This process allows the die 51 to be re-bonded to the wafer 6, correcting the relative position deviation between the pattern on the wafer 6 and the pattern on the die 51 that occurred during the first bonding. After the die 51 is re-bonded to the wafer 6, the process proceeds to step S107.

[0100] As described above, according to this embodiment, based on the estimation results of the estimation process in step S107, it is determined whether the relative position deviation between the pattern of wafer 6 and the pattern of die 51 after the bonding process falls within the allowable range. If the relative position deviation is not within the allowable range, die 51 is separated from wafer 6. Therefore, a new die 51 can be bonded to the target area of ​​wafer 6 that has been separated from die 51, or the separated die 51 can be bonded again.

[0101] <Example of the method for manufacturing an article> A method for manufacturing an article (semiconductor IC components, liquid crystal devices, MEMS, etc.) using the above-described bonding apparatus will be described. The article manufacturing method according to an embodiment of the present invention is suitable for manufacturing articles such as microdevices (e.g., semiconductor devices) or components having microstructures. The article manufacturing method according to this embodiment includes: using the above-described bonding apparatus to bond a second component to a first component; processing the first component to which the second component is bonded; and manufacturing an article from the processed first component. In the above-described bonding apparatus, by estimating the relative position between the pattern of the first component and the pattern of the second component after the bonding process, the bonding state between the first and second components can be easily and accurately determined. Furthermore, based on the estimated relative position, dies can be rebonded in the bonding apparatus. Information on the estimated relative position can also be reflected in subsequent steps. The subsequent steps are known steps and include probing, dicing, bonding, and packaging. The article manufacturing method according to this embodiment surpasses conventional methods in at least one of the following: article performance, quality, productivity, and production cost.

[0102] Other embodiments The embodiments of the present invention may also be implemented by reading and executing computer-executable instructions (e.g., one or more programs) recorded on a storage medium (which may also be more fully referred to as a "non-transitory computer-readable storage medium") to perform one or more functions of the above embodiments and / or by a computer of a system or device that may include one or more circuits (e.g., application-specific integrated circuits (ASICs)) for performing one or more functions of the above embodiments, and by a method performed by a computer of the system or device (e.g., by reading and executing computer-executable instructions from a storage medium to perform the functions of one or more embodiments and / or controlling one or more circuits to perform the above embodiments). The functions of one or more embodiments are implemented. A computer may include one or more processors (e.g., a central processing unit (CPU), a microprocessor unit (MPU)) and may include a network of separate computers or separate processors to read and execute computer-executable instructions. The computer-executable instructions may be provided to the computer, for example, from a network or a storage medium. The storage medium may include, for example, one or more of a hard drive, random access memory (RAM), read-only memory (ROM), storage in a distributed computing system, an optical disc (such as a compact disc (CD), a digital versatile disc (DVD), or a Blu-ray Disc (BD)™), a flash memory device, a memory card, and the like.

[0103] Although the present invention has been described with reference to exemplary embodiments, it should be understood that the invention is not limited to the disclosed exemplary embodiments. The scope of the following claims should be accorded the broadest interpretation so as to encompass all such modifications and equivalent structures and functions.

[0104] 1: Base 2: Mounting parts 3: Pickup unit 4:Joint unit 5: Cutting the frame 6: Wafer 31: Pickup head 32: Release head 33: Frame retainer 41: Carrier base 42: Upper base 43: Carrier 44:Joint head 45: Grain observation camera 46: Wafer observation camera 47,47a,47b,47c:Interferometer 51: Grain 51a: Joining surface 51b: back surface 61: Grain recovery unit 100,100':Joining equipment 431: Chuck 432: driving mechanism 433: Mirror 433x,433y: Strip mirror 434: Reference Board 434a, 434b, 434c: Marking 501: Component pattern 502: Alignment mark 503:Through hole 601: Pattern 5A,5B,5C,5D:Reference number 6a: Target Area CNT:Controller S101, S102, S103, S104, S105, S106, S107, S107a, S107b, S107c, S107d, S108, S201, S202, S203, S301, S302, S303, S304, S401: Steps

Claims

1. A bonding apparatus for bonding a second component including a second pattern to a first component including a first pattern, comprising: a first image capturing device; a second image capturing device; and a controller configured to control a bonding process for bonding the second component to the first component, wherein the controller is configured to determine the relative position between the first pattern and the second pattern after the bonding process based on feature position information and positional relationship information, wherein the feature position information indicates the position of a feature portion of the second component relative to the first component and is obtained based on an image of the bonding process of bonding the second component to the first component, the image being captured by the first image capturing device, wherein the positional relationship information indicates the positional relationship between the feature portion and the second pattern, and wherein the positional relationship information is obtained based on an image of a pattern of a second bonding surface including the second pattern before the bonding process, the image being captured by the second image capturing device.

2. The device of claim 1 further includes the second image capturing device configured to capture an image of the second pattern disposed on the second member; wherein the first image capturing device is configured to capture an image of the first pattern disposed on the first member; and wherein the controller is configured to: obtain information about the relative position between the first member and the second member based on the position of the first pattern in the image captured by the first image capturing device and the position of the second pattern in the image captured by the second image capturing device; and control the alignment of the first member and the second member, and the bonding process of bonding the second member to the first member based on the information obtained about the relative position between the first member and the second member.

3. The device of claim 1, wherein the characteristic portion of the second member is an indicator that can be identified by capturing an image obtained by photographing an image of a second mating surface that is the mating surface of the second member and by photographing an image obtained by photographing a surface opposite to the second mating surface.

4. The device of claim 1, wherein the characteristic portion of the second member is the outer edge of the second member.

5. The device of claim 1, wherein the characteristic portion of the second member is a through hole extending from a second mating surface, which is a mating surface of the second member, to a surface opposite to the second mating surface.

6. The device of claim 2, wherein the controller is configured to obtain the feature location information based on an image obtained by the second imaging device prior to the bonding process.

7. The device of claim 1, wherein the controller is configured to determine the relative position by converting the position of the feature portion of the second member in the feature position information into the position of the second pattern based on the position relationship information.

8. The device of claim 1, wherein the controller is configured to output information representing the relative position determined based on the positional relationship information and the characteristic position information.

9. The device of claim 1, wherein the controller is configured to output an evaluation result of the relative position determined based on the positional relationship information and the characteristic position information.

10. The device of claim 1, wherein the controller is configured to perform a separation process to separate the second component from the first component based on an evaluation result of the relative position determined based on the positional relationship information and the characteristic position information.

11. The device of claim 1 further comprises: a first retainer for retaining the first member; and a second retainer for retaining the second member, wherein the controller is configured to control the engagement process by relatively driving the first retainer and the second retainer.

12. The apparatus of claim 11, wherein the first image capturing device is configured to capture an image of a first mating surface, which is a mating surface of the first member, in a state in which the first member is held by the first holder, and the second image capturing device is configured to capture an image of a second mating surface, which is a mating surface of the second member, in a state in which the second member is held by the second holder.

13. A joining method for joining a second component including a second pattern to a first component including a first pattern, comprising: performing a joining process of joining the second component to the first component; determining the relative position between the first pattern and the second pattern based on feature position information and positional relationship information, wherein the feature position information indicates the position of a feature portion of the second component relative to the first component and is obtained based on an image of joining the second component to the first component by the joining process, and wherein the positional relationship information indicates the positional relationship between the feature portion and the second pattern.

14. A method of manufacturing an article, comprising: joining a second member to the first member using a joining method as defined in claim 13; processing the first member that has been joined to the second member; and manufacturing an article from the processed first member.

Images