Bonding device and bonding method

Abstract

To provide a technique advantageous for highly accurately performing positioning when joining a second object to the predetermined spot of a first object in first and second directions.SOLUTION: A joining device for joining the second object to the first object includes a first holding part which holds the first object, a second holding part which holds the second object, a positioning mechanism which changes a relative position between the first holding part and the second holding part in the first and second directions, a first camera which images the first object, a second camera which images the second object, a support part which supports the second holding part and the first camera, and a control part which controls the positioning mechanism in the first and second directions, so as to position the second object to the joining spot of the first object on the basis of the output of the first camera and the output of the second camera.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a joining device and a joining method. [Background technology]

[0002] Patent Document 1 describes an apparatus for positioning a first object relative to a second object. The apparatus includes a moving body that moves linearly relative to the second object. The moving body is provided with a holding unit that holds the first object and a position identifying unit that identifies the position of the second object, attached at a predetermined interval along the moving direction of the moving body. The apparatus also includes a scale arranged along the moving direction of the moving body. The moving body is further provided with a first position detecting unit that detects the position of the holding unit based on the graduations on the scale, and a second position detecting unit that detects the scale position corresponding to the position of the second object, attached at a predetermined interval along the moving direction of the moving body. The apparatus also includes a control unit that moves the moving body to a position where the first position detecting unit detects the graduation position, thereby positioning the first object relative to the second object. This apparatus allows the first object to be positioned with high accuracy relative to the second object, even if the scale thermally expands. However, in this apparatus, the moving body moves in only one direction. Therefore, the device can only position the first object relative to the second object with high precision in only one direction. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6787612 Summary of the Invention [Problem to be solved by the invention]

[0004] An object of the present invention is to provide an advantageous technique for achieving high accuracy in positioning in a first direction and a second direction when joining a second object to a predetermined location of a first object. [Means for solving the problem]

[0005] One aspect of the present invention relates to a joining device that joins a first object to a plurality of second objects, the joining device including a first holding unit that holds the first object, a second holding unit that holds the second object, a first camera that images the first object, a second camera that images the second object, a support unit that supports the second holding unit and the first camera, a stage that supports the first holding unit and the second camera and is movable in a first direction and a second direction, and a drive mechanism that drives the stage. The system comprises a measuring instrument that measures the position of the stage and a control unit that controls the driving mechanism, wherein the stage has a plurality of marks formed on a plate, the plurality of marks including marks that are positioned differently from each other in the first direction and marks that are positioned differently from each other in the second direction, and the control unit calibrates the position of the stage based on an image of the plurality of marks captured by the first camera and the measurement results of the measuring instrument, and after the calibration, controls the driving mechanism to join the second object to the first object based on the output of the first camera, the output of the second camera, and the measurement results of the measuring instrument. [Effects of the Invention]

[0006] According to the present invention, there is provided an advantageous technique for realizing highly accurate positioning in a first direction and a second direction when joining a second object to a predetermined location of a first object. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 is a diagram schematically illustrating a configuration of a joining apparatus according to a first embodiment. [Figure 2] FIG. 2 is a diagram showing an example of the configuration of a wafer stage in the bonding apparatus of the first embodiment. [Figure 3] 4 is a flowchart showing a joining method in the joining device of the first embodiment. [Figure 4] 5 is a flowchart showing a method for calculating an offset of a die bonding position in the bonding apparatus of the first embodiment. [Figure 5] FIG. 10 is a diagram schematically illustrating the configuration of a joining device according to a second embodiment. [Figure 6] FIG. 10 is a diagram showing an example of the configuration of a wafer stage in a bonding apparatus according to a second embodiment. [Figure 7] FIG. 10 is a diagram schematically illustrating the configuration of a joining device according to a third embodiment. [Figure 8] FIG. 10 is a diagram showing an example of the configuration of a bonding stage in a bonding apparatus according to a third embodiment. [Figure 9] FIG. 10 is a diagram schematically illustrating the configuration of a joining device according to a fourth embodiment. [Figure 10] FIG. 10 is a diagram showing an example of the configuration of a bonding stage in a bonding apparatus according to a fourth embodiment. [Figure 11]FIG. 10 is a diagram schematically illustrating the configuration of a joining device according to a fifth embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0008] Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the scope of the invention claimed. Although multiple features are described in the embodiments, not all of these multiple features are necessarily essential to the invention, and multiple features may be combined arbitrarily. Furthermore, in the accompanying drawings, the same reference numerals are used to designate the same or similar components, and redundant explanations will be omitted.

[0009] In the following description, the first object is a wafer on which semiconductor devices are formed, and the second object is a singulated die containing semiconductor devices; however, the first object and the second object are not limited to these, and various modifications and variations are possible within the scope of the invention.

[0010] For example, the first object may be a silicon wafer, a silicon wafer on which wiring is formed, a glass wafer, a glass panel on which wiring is formed, an organic panel (PCB) on which wiring is formed, or a metal panel. Alternatively, the first object may be a substrate on which a die on which semiconductor devices are formed is bonded to a wafer on which semiconductor devices are formed.

[0011] For example, the second object may be a stack of several singulated dies, a small piece of material, an optical element, a MEMS, or other structure.

[0012] The bonding method is not limited to a specific one, and may be, for example, bonding using an adhesive, temporary bonding using a temporary adhesive, hybrid bonding, atomic diffusion bonding, vacuum bonding, bump bonding, or the like, and various temporary and permanent bonding methods can be used.

[0013] Here, industrial application examples will be described. A first application example is the manufacture of stacked memories. When applied to the manufacture of stacked memories, the first object may be a wafer on which memories are formed, and the second object may be the memories as individual dies. Typically, eight or more layers are stacked, so when bonding the eighth layer, the first object may be a substrate on which six layers of memories have already been bonded on the wafer. The final layer may be a driver that drives the memories.

[0014] A second application example is heterogeneous integration of processors. While conventional processors are mainly SoCs, in which logic circuits and SRAM are integrated into a single semiconductor chip, this technology manufactures processors by applying the optimal process for each element to a separate wafer and then bonding them together. This reduces processor costs and improves yield. When applied to heterogeneous integration, the first object can be a wafer containing a logic device, a semiconductor device, and the second object can be an SRAM, antenna, or driver die that has been singulated after probing. Since different dies are typically bonded sequentially, the number of bonded elements in the first object increases. For example, if bonding begins with an SRAM, when bonding the next SRAM element, the first object is the logic wafer to which the SRAM is bonded. When bonding multiple dies, it is desirable to bond the thinnest die first to prevent the bonding head from interfering with the already bonded dies.

[0015] A third application example is 2.5D bonding using a silicon interposer. A silicon interposer is a silicon wafer with wiring formed on it. 2.5D bonding involves bonding singulated dies to the silicon interposer to electrically bond the dies together. When applying die bonding to a silicon interposer, the first object may be a silicon wafer with wiring formed on it, and the second object may be a singulated die. Since multiple types of dies are typically bonded to a silicon interposer, the first object may include a silicon interposer with several dies already bonded to it. When bonding multiple dies, it is desirable to bond the thinner dies first to prevent the bonding head from interfering with the already bonded dies.

[0016] A fourth application example is 2.1D bonding using an organic interposer or glass interposer. An organic interposer is an organic panel (PCB substrate, CCL substrate) used as a package substrate with wiring formed on it, and a glass interposer is a glass panel with wiring formed on it. 2.1D bonding involves bonding singulated dies to an organic interposer or glass interposer and electrically connecting the dies using the wiring on the interposer. When applying die bonding to an organic interposer, the first object may be an organic panel with wiring formed on it. When applying die bonding to a glass interposer, the first object may be a glass panel with wiring formed on it, and the second object may be a singulated die. Since multiple types of dies are usually bonded to an organic interposer or glass interposer, the first object may include an organic interposer or glass interposer with several dies already bonded to it. When bonding multiple dies, it is desirable to bond the thinner dies first so that the bonding head does not interfere with the already bonded dies.

[0017] A fifth application example is temporary bonding in a fan-out package manufacturing process. Fan-out wafer-level packaging (WLP) is known, in which singulated dies are reassembled into wafers using a molding resin and packaged. Fan-out panel-level packaging (PLP) is also known, in which singulated dies are reassembled into panels and packaged. During packaging, rewiring from the dies to the bumps or rewiring connecting different dies is formed on a molded reconfigured substrate. If the die alignment accuracy is low, it becomes difficult to accurately align the rewiring pattern with the die when transferring the rewiring pattern using a step-and-repeat exposure tool. Therefore, accurate alignment of the dies is required. When applied to a fan-out package manufacturing process, the first object may be a substrate to be temporarily bonded, such as a metal panel, and the second object may be a singulated die. The singulated die can be temporarily bonded to a substrate, such as a metal panel, using a temporary bonding agent. The wafer or panel is then molded into a wafer or panel shape using a molding device, and after molding, the wafer or panel is peeled off from the substrate, such as a metal panel, to produce a reconstructed wafer or panel. When applying this bonding method, it is desirable to adjust the bonding position using a bonding device so as to correct for the array deformation caused by the molding process.

[0018] A sixth application example is heterogeneous substrate bonding. For example, in the field of infrared image sensors, InGaAs is known as a highly sensitive material. High-sensitivity, high-speed infrared image sensors can be manufactured by using InGaAs for the light-receiving sensor and silicon for the logic circuit that extracts data, which enables high-speed processing. However, InGaAs crystals are only mass-produced in small diameter wafers, such as 4 inches, which are smaller than the 300 mm silicon wafers that are the mainstream. Therefore, a method has been proposed in which an individualized InGaAs substrate is bonded to a 300 mm silicon wafer with a logic circuit formed on it. This bonding system can also be used for heterogeneous substrate bonding, which involves bonding substrates made of different materials and different sizes. When applying this technology to heterogeneous substrate bonding, the first object can be a large-diameter substrate such as a silicon wafer, and the second object can be a small piece of material such as InGaAs. The small piece of material is typically a sliced ​​crystal, preferably cut into a rectangular shape. First Embodiment FIG. 1 is a diagram schematically illustrating the configuration of a bonding apparatus BD according to a first embodiment. In FIG. 1, directions are indicated by an XYZ coordinate system. Typically, the XY plane is a plane parallel to a horizontal plane, and the Z axis is an axis parallel to the vertical direction. The X, Y, and Z axes are shown as examples of directions that are mutually orthogonal or intersect with each other. This is also true for the other drawings.

[0019] As shown in FIG. 1 , the bonding apparatus BD may include a pickup unit 3 and a bonding unit 4 disposed on a base 1 that is vibration-damped by a mount 2. While FIG. 1 illustrates an example in which the pickup unit 3 and the bonding unit 4 are mounted on a single base 1, the pickup unit 3 and the bonding unit 4 may be individually mounted on separate bases. The bonding apparatus BD may be configured to position and bond a die 51, which serves as a second object, to a bonding target location on a wafer 6, which serves as a first object. The die 51 may be provided in a state in which it is held by dicing tape attached to a dicing frame 5. The bonding apparatus BD may also include a control unit CNT that controls the pickup unit 3 and the bonding unit 4. The control unit CNT may be configured, for example, by a PLD (abbreviation for programmable logic device) such as an FPGA (abbreviation for field programmable gate array), an ASIC (abbreviation for application specific integrated circuit), a general-purpose or dedicated computer with an embedded program, or a combination of all or part of these.

[0020] The pickup unit 3 may include a pickup head 31 and a release head 32. The pickup unit 3 may use the release head 32 to peel the die 51 to be bonded to the wafer 6 from the dicing tape and then use the pickup head 31 to hold the die 51 by suction. The pickup head 31 may, for example, rotate the die 51 180 degrees and deliver it to the bonding head 423 of the bonding unit 4. The pickup head 31 may contact the bonding surface of the die 51. Therefore, in an application example of a bonding method that activates the surface to bond, such as hybrid bonding, it is desirable to reduce the contact area by using a highly stable surface such as a diamond-like carbon (DLC) coating or a fluorine coating on the surface that contacts the bonding surface, or by using a shape with a small contact area such as a high-density pin shape. Alternatively, a non-contact handling method such as a Bernoulli chuck or by holding the side or edge may be considered to avoid contact with the bonding surface.

[0021] The bonding unit 4 may include a stage base 41 and an upper base 42. A wafer stage 43 serving as a first holder may be mounted on the stage base 41. The wafer stage 43 may be driven in the X-axis direction (first direction) and the Y-axis direction (second direction) by a drive mechanism 436 such as a linear motor. The drive mechanism 436 may further be configured to drive the wafer stage 43 to rotate about an axis parallel to the Z-axis direction (third direction). Instead of the drive mechanism 436 driving the wafer stage 43 to rotate about an axis parallel to the Z-axis direction, the bonding head 423 may drive the die 51 to rotate about an axis parallel to the Z-axis direction. The drive mechanism 436 may constitute a positioning mechanism that changes the relative position of the wafer chuck 433 (or the wafer 6) serving as the first holder and the bonding head 423 (or the die 51) serving as the second holder.

[0022] A die observation camera 431 serving as a second camera may be mounted on the wafer stage 43. The die observation camera 431 is a second detector for detecting the position of a characteristic portion of a die 51 serving as a second object held by the bonding head 423. A bar mirror 432 may be provided on the wafer stage 43. The bar mirror 432 may be used as a target for the interferometer 422. A wafer chuck 433 serving as a first holder may be mounted on the wafer stage 43. The wafer chuck 433 holds a wafer 6 serving as a first object.

[0023] In the example of FIG. 1 , the wafer stage 43 functions as a support structure that supports a wafer chuck 433 as a first holder and a die observation camera 431 as a second camera. The wafer stage 43 as a support structure may have a first end face (left end face in FIG. 1 ) on the side of a path along which a die 51 as a second object is transferred to a bonding head 423 as a second holder, and a second end face (right end face in FIG. 1 ) opposite the first end face. The die observation camera 431 as a second camera may be disposed between an imaginary plane that passes through the center of the support structure and is parallel to the first end face and the first end face. Alternatively, from another perspective, the die observation camera 431 as a second camera may be disposed between a predetermined position on a path along which a die 51 as a second object is transferred to a bonding head 423 as a second holder and the wafer chuck 433 as a first holder. Such a configuration is advantageous in that it reduces the amount of movement of the wafer stage 43 in order to observe the die 51 held by the bonding head 423 with the die observation camera 431, thereby improving throughput.

[0024] A wafer observation camera 421 serving as a first camera may be mounted on the upper base 42. The wafer observation camera 421 is a first detector for detecting the positions of characteristic locations on a wafer 6 serving as a first object held by a wafer chuck 433. The control unit CNT may be configured to identify or calculate the positions of multiple bonding target locations on the wafer 6 based on the positions of the characteristic locations on the wafer 6 detected using the wafer observation camera 421. An interferometer 422 may also be mounted on the upper base 42 for measuring the position of the wafer stage 43 using a bar mirror 432. A bonding head 423 may also be mounted on the upper base 42. The bonding head 423 receives and holds a die 51 serving as a second object handed over from the pickup head 31 and bonds the die 51 to a bonding target location on the wafer 6. The bonding head 423 also functions as a second holder for holding the die 51 serving as the second object.

[0025] In the example of FIG. 1 , the upper base 42 is a support configured to support a bonding head 423 as a second holder and a wafer observation camera 421 as a first camera. The upper base 42 as a support may have a third end face (left end face in FIG. 1 ) on the side of a path for transporting a die 51 as a second object to the bonding head 423 as a second holder, and a fourth end face (right end face in FIG. 1 ) opposite the third end face. The wafer observation camera 421 as a first camera may be disposed between the third end face and a second imaginary plane that passes through the center of the support and is parallel to the third end face. Such a configuration is advantageous for reducing the amount of movement of the wafer stage 43 to observe the wafer 6 held by the wafer chuck 433 with the wafer observation camera 421, thereby improving throughput.

[0026] When bonding the die 51 as a second object to a bonding target portion of the wafer 6 as a first object, the bonding head 423 can drive the die 51 in the negative direction (downward) of the Z axis to bond the die 51 to the bonding target portion of the wafer 6. Alternatively, the drive mechanism 436 can drive the wafer stage 43 in the positive direction (upward) of the Z axis to bond the die 51 to the bonding target portion of the wafer 6. Alternatively, the drive mechanism (not shown) can drive the wafer chuck 433 in the positive direction (upward) of the Z axis to bond the die 51 to the bonding target portion of the wafer 6.

[0027] In the above description, the pickup head 31 rotates the die 51 by 180 degrees and passes it to the bonding head 423. However, a first die holding unit and a second die holding unit may be provided, and the die 51 may be passed from the first die holding unit to the second die holding unit midway, and then from the second die holding unit to the bonding head 423. Alternatively, a drive mechanism for driving the bonding head 423 may be provided, and the bonding head 423 may be driven to receive the die 51. Furthermore, in order to improve productivity, multiple pickup units, multiple pickup heads, multiple release heads, and multiple bonding heads may be provided.

[0028] FIG. 2 is a view of the wafer stage 43 as viewed from the positive direction of the Z axis. The wafer 6 is held by a wafer chuck 433. The wafer 6 or the wafer stage 43 can be positioned with respect to rotation about an axis parallel to the X-axis direction (first direction) and the Y-axis direction (second direction), which are orthogonal or intersecting each other, and the Z-axis direction (third direction), which is orthogonal to the X-axis direction and the Y-axis direction. Therefore, the wafer stage 43 can be provided with bar mirrors 432, more specifically, bar mirrors 432a and 432b. The bar mirror 432a can function as a target for the interferometers 422a and 422c. The control unit CNT can detect the position of the wafer stage 43 in the X-axis direction based on the output of the interferometer 422a, and can also detect the rotation of the wafer stage 43 around an axis parallel to the Z-axis direction based on the outputs of the interferometers 422a and 422c. The bar mirror 432b can function as a target for the interferometer 422b. The control unit CNT can detect the position of the wafer stage 43 in the Y-axis direction based on the output of the interferometer 422b. The control unit CNT can be configured to feedback control the wafer 6 or the wafer stage 43 with respect to rotations in the X-axis direction, the Y-axis direction, and around an axis parallel to the Z-axis direction orthogonal to the X-axis direction and the Y-axis direction, based on the outputs of the interferometers 422a, 422b, and 422c. The interferometer 422 and the control unit CNT may be understood as components of the positioning mechanism described above.

[0029] A reference plate 434 may be provided on the upper surface of the wafer stage 43. A plurality of marks 434a, 434b, and 434c may be arranged on the reference plate 434. The reference plate 434 is made of a material with a low thermal expansion coefficient, and the marks may be drawn with high positional accuracy. For example, the reference plate 434 may be formed by drawing marks on a quartz substrate using a drawing method used in semiconductor lithography processes. The reference plate 434 has a surface that is approximately flush with the surface of the wafer 6 and may be observed by the wafer observation camera 421. Alternatively, a separate camera may be provided for observing the reference plate 434. The wafer stage 43 may be configured by combining a coarse stage that can be driven over a large range and a fine stage that can be driven over a small range with high precision. In such a configuration, the die observation camera 431, bar mirrors 432a and 432b, wafer chuck 433, and reference plate 434 may be mounted on the fine stage to achieve high-precision positioning.

[0030] Here, a method for assuring the origin position, magnification, X-axis and Y-axis directions (rotation), and orthogonality of the wafer stage 43 using the reference plate 434 will be described. Mark 434a is observed with the wafer observation camera 421, and the output value of the interferometer when mark 434a is located at the center of the image output by the wafer observation camera 421 is set as the origin of the wafer stage 43. Next, mark 434b is observed with the wafer observation camera 421, and the Y-axis direction (rotation) and magnification in the Y-axis direction of the wafer stage 43 are determined based on the output value of the interferometer when mark 434b is located at the center of the image output by the wafer observation camera 421. Next, mark 434c is observed with the wafer observation camera 421, and the X-axis direction (rotation) and magnification in the X-axis direction of the wafer stage 43 are determined based on the output value of the interferometer when mark 434c is located at the center of the image output by the wafer observation camera 421.

[0031] Specifically, the direction from mark 434b to mark 434a on the reference plate 434 serves as the Y-axis of the bonding device BD, and the direction from mark 434c to mark 434a serves as the X-axis of the bonding device BD, allowing for calibration of the axis direction and orthogonality. Calibration can also be performed using the distance between marks 434b and 434a as the Y-axis scale reference for the bonding device BD, and the distance between marks 434c and 434a as the X-axis scale reference for the bonding device BD. Because the refractive index of the optical path of an interferometer changes due to changes in atmospheric pressure and temperature, which can cause fluctuations in measurement values, it is desirable to perform calibration at any time to ensure the origin position, magnification, rotation, and orthogonality of the wafer stage 43. To reduce fluctuations in the interferometer measurement values, it is desirable to enclose the space in which the wafer stage 43 is located in a temperature-controlled chamber and control the temperature within the chamber.

[0032] In this embodiment, a form in which the reference plate on the wafer stage is observed with a wafer observation camera has been described, but even if the reference plate is attached to the upper base and observed with a die observation camera, the origin position, magnification, rotation, and orthogonality of the wafer stage can be guaranteed.

[0033] The above explanation is about an example of performing calibration by observing a reference plate. Alternatively, calibration may be performed by, for example, abutting against a reference surface, or high-precision positioning may be performed using a position measuring instrument whose absolute value is guaranteed, such as a white light interferometer.

[0034] The bonding method according to the first embodiment will be described below with reference to the flowchart in FIG. 3. This bonding method is controlled by the control unit CNT. In step 1001, a wafer 6 serving as a first object is loaded into the bonding device BD and held by the wafer chuck 433 (first holding step). Because foreign matter adhering to the bonding surface can cause bonding defects, the inside of the bonding device BD may be a space with a high level of cleanliness, such as Class 1. To maintain high cleanliness, the wafer 6 may be housed in a sealed container, such as a FOUP, and then loaded into the bonding device BD from that container. To further enhance cleanliness, the wafer 6 may be cleaned in the bonding device BD after being loaded. Pretreatment for bonding may also be performed. For example, in the case of adhesive bonding, an adhesive is applied to the wafer 6, and in the case of hybrid bonding, a treatment to activate the surface of the wafer 6 may be performed. The wafer 6 is roughly positioned by a pre-aligner (not shown) based on the notch or orientation flat and the wafer outer shape position, and then transported to a wafer chuck 433 serving as a first holding part on the wafer stage 43, and held by the wafer chuck 433.

[0035] In step 1002, the position of the characteristic portion (measurement target portion) of the wafer 6 is measured using the wafer observation camera 421, and the position of the bonding target portion is determined based on the measurement result. Here, the positional relationship (relative position) between the position of the characteristic portion (measurement target portion) of the wafer 6 and the bonding target portion is known. Focus adjustment for imaging the characteristic portion of the wafer 6 by the wafer observation camera 421 can be provided by providing a focus adjustment mechanism within the wafer observation camera 421. Alternatively, focus adjustment can be provided by providing a Z-axis drive mechanism on the wafer stage 43 and driving the wafer 6 about the Z axis using the Z-axis drive mechanism. While alignment marks for alignment are often formed on the wafer 6, if no alignment marks are formed, a characteristic portion whose position can be identified can be measured. The control unit CNT causes the wafer observation camera 421 to image the characteristic portion of the wafer 6 (first imaging step) and can detect the relative position of the image of the characteristic portion with respect to the center of the output image of the wafer observation camera 421 as the position of the characteristic portion (measurement target portion).

[0036] An offset amount may be calculated in advance to measure the relative position of the mark with respect to the reference point of the bonding device BD with high accuracy. This may include a process of driving the wafer stage 43 so that the mark on the reference plate 434 is within the field of view of the wafer observation camera 421 and measuring the position of the mark with the wafer observation camera 421. Based on the driving position of the wafer stage 43 at that time and the position of the mark measured with the wafer observation camera 421, an offset amount with respect to the position measured with the wafer observation camera 421 can be determined. Here, the reference point of the bonding device BD is generally set to the position of a specific mark on the reference plate 434, but it may be set to another location as long as it serves as a reference position.

[0037] Because the measurement range in the rotation direction using an interferometer or encoder is narrow, the amount of rotation that can be corrected by the wafer stage 43 is small. Therefore, if the amount of rotation of the wafer 6 is large, it is desirable to correct the rotation and re-hold the wafer 6. After re-holding, it is necessary to measure the mounting position of the wafer 6 again. During this operation, the surface position of the bonding surface of the wafer 6 may be measured during the autofocus operation when measuring the marks on the wafer, or using a first height measuring device (not shown). Because the thickness of the wafer 6 varies, it is advantageous to measure the surface position of the wafer 6 in order to accurately manage the gap between the wafer 6 and the die 51 during the bonding operation.

[0038] The wafer stage 43 uses the reference plate 434 to ensure the origin position, magnification, and X- and Y-axis directions (rotation) and orthogonality. Therefore, the positions of characteristic locations (measurement target locations) on the wafer 6 can be measured using the origin position of the wafer stage 43 and the X- and Y-axes as references. The wafer 6 may have bonding target locations (or semiconductor devices as bonding targets) at a regular interval. These bonding target locations (semiconductor devices) are manufactured by positioning multiple layers with high precision using semiconductor manufacturing equipment, so the bonding target locations (semiconductor devices) are generally arranged repeatedly with nano-level precision. Therefore, in the wafer alignment in step 1002, it is not necessary to measure the positions of each characteristic location corresponding to all bonding target locations (semiconductor devices). Therefore, the control unit CNT may be configured to perform a process (first measurement step) of determining the positions of multiple bonding target locations by measuring the positions of measurement target locations that are fewer than the number of bonding target locations and statistically processing the measurement results. This type of control is advantageous in improving throughput compared to the method of measuring the bonding points after each die bonding, as disclosed in Patent Document 1. Here, the multiple measurement target locations can be determined based on the arrangement information of the semiconductor device. To determine the positions of the multiple bonding target locations, the control unit CNT can calculate the origin position of the repeating arrangement of the multiple bonding target locations, the rotation amount and orthogonality in the X-axis and Y-axis directions, and the magnification error of the repeating period, based on the measurement results of the positions of the multiple measurement target locations.

[0039] Furthermore, it is desirable that wafer chuck 433 have a temperature control function for adjusting the temperature of wafer 6. This is because the thermal expansion coefficient of silicon wafers is 3 ppm / °C, and in the case of a 300 mm diameter wafer, a 1°C rise in temperature will cause the outermost periphery to move by 150 mm x 0.000003 = 0.00045 mm = 450 nm. If the bonding position moves after wafer alignment, bonding cannot be performed with high positional accuracy, so it is desirable to stabilize the wafer temperature with an accuracy of 0.1°C or less.

[0040] If the first object is an interposer on which wiring is formed, the multiple bonding locations are determined based on the arrangement of the repetitively formed wiring, not the arrangement of the semiconductor devices. Also, if the first object is a wafer or panel without a pattern, the wafer alignment in step 1002 is not performed.

[0041] The following describes the movement of the die as the second object, which is performed in parallel with or after the loading of the wafer as the first object and wafer alignment. In step 2001, a dicing frame 5 on which die 51, singulated by a dicer, are arranged on a dicing tape, is loaded into the bonding device BD. Because foreign matter adhering to the bonding surface can cause bonding defects, the dicing frame may be transported in a container that is tightly sealed and kept highly clean. To further enhance cleanliness, the die 51 on the dicing frame 5 may be washed inside the bonding device BD. The rotation direction and shift position of the dicing frame 5 may be roughly determined by a pre-aligner (not shown) based on the external shape of the dicing frame.

[0042] In step 2002, a die 51 as a second object is picked up by the pickup head 31. Specifically, the pickup head 31 and the release head 32 may be positioned at the position of the die 51 to be picked up. Then, while the die 51 to be picked up is sucked by the pickup head 31, the dicing tape is peeled off from the die 51 by the release head 32, and the die 51 may be held by the pickup head 31. The die 51 to be picked up may be determined based on, for example, known good die (KGD) information transmitted online to the bonding device BD. Normally, only good die are picked up, but when a bad die (KBD: known bad die) is to be bonded to a defective device location on the wafer 6, the bad die is picked up.

[0043] In step 2003, the die 51 as the second object picked up by the pickup head 31 is transferred to the bonding head 423 and held by the bonding head 423 (second holding step). When the die 51 is picked up by the pickup head 31, the semiconductor device surface faces the pickup head 31. On the other hand, the die 51 is transferred to the bonding head 423 so that the surface opposite the semiconductor device surface faces the bonding head 423. The transfer of the die 51 to the bonding head 423 may be performed directly from the pickup head 31 to the bonding head 423 or via multiple die holders. Furthermore, pretreatment for bonding may be performed while the die 51 is being transferred. Pretreatment may include, for example, a die cleaning process, application of an adhesive in the case of adhesive bonding, or a surface activation process in the case of hybrid bonding. Note that if the surface of the die 51 becomes inactive while being transferred to the bonding head 423, it is desirable to activate the bonding surface using an atmospheric pressure plasma activation device after the die 51 is mounted on the bonding head 423.

[0044] With the above steps, the wafer 6 as the first object and the die 51 as the second object are held by their respective holders. Next, the bonding flow will be described. In step 1003, the position of the die 51 as the second object held by the bonding head 423 can be measured (second measurement step). Specifically, the drive mechanism 436 can drive the wafer stage 43 so that a characteristic portion of the die 51 is within the field of view of the die observation camera 431. Focus adjustment can be provided by providing a focus adjustment mechanism within the die observation camera 431, or by providing a Z-axis drive mechanism in the bonding head 423 and driving the die 51 about the Z axis with the Z-axis drive mechanism. Focus adjustment can also be provided by providing a Z-axis drive mechanism in the wafer stage 43 on which the die observation camera 431 is mounted and driving the die observation camera 431 about the Z axis with the Z-axis drive mechanism.

[0045] Scribe lines on which alignment marks used for alignment in semiconductor manufacturing processes are formed can be removed by dicing. Therefore, the die 51 often does not have alignment marks for alignment. Therefore, the position of the die 51 can be measured using the end of an array of pads or bumps arranged on the die 51, a region with a non-periodic arrangement that allows its position to be identified, or the outer shape of the die as a characteristic point. The control unit CNT can cause the die observation camera 431 to capture an image of the die 51 (second imaging process) and determine the position of the characteristic point based on the relative position of the image of the characteristic point with respect to the center of the output image of the die observation camera 431. It is necessary to manage the offset amount when positioning the die 51 at the joining point based on the position of the die 51 measured using the die observation camera 431; this method will be described later.

[0046] When measuring the position of the die 51, it is desirable to measure the positions of multiple characteristic locations within the die 51 and also measure the amount of rotation of the die 51. To measure the positions of multiple characteristic locations, the wafer stage 43 may be driven each time the position of each characteristic location is measured, or the field of view of the die observation camera 431 may be designed so that multiple characteristic locations can be observed at once. The die 51 can be rotated by rotating the wafer stage 43 during bonding. However, because the measurement range of the interferometer in the rotational direction is narrow, if the amount of rotation of the die 51 is large, it is desirable to correct the rotation and re-hold the die 51. After re-holding the die 51, it is necessary to measure the position of the die 51 again. During this operation, the surface position of the bonding surface of the die 51, which serves as the second object, may be measured during the autofocus operation during die position measurement or using a second height measuring device (not shown). Because the thickness of the die 51 varies, measuring the surface position of the die 51 is advantageous for accurately managing the gap between the wafer 6 and the die 51 during the bonding operation. Furthermore, the height of the die 51 at multiple positions may be measured, and the posture of the die 51 or the wafer 6 may be adjusted by a tilt mechanism (not shown) during bonding. Such a tilt mechanism may be incorporated into the wafer stage 43, the wafer chuck 433, or the bonding head 423.

[0047] In step 1004, the driving mechanism 436 drives the wafer stage 43 so that the die 51, which is the second object, is positioned at a bonding target location selected from a plurality of bonding target locations on the wafer 6, which is the first object. At this time, the control unit CNT can control the driving mechanism 436 so that the position of the wafer stage 43 is feedback-controlled based on the measurement results from the interferometer 422. Also, at this time, the control unit CNT can determine the target position of the wafer stage 43 based on the position and rotation amount of the wafer 6 and the position and rotation amount of the die 51 measured in steps 1002 and 1003, as well as the offset amount. Also, if a shift occurs due to the bonding operation as described below, the control unit CNT considers that amount as the offset amount.

[0048] In step 1005, a die 51 as a second object is bonded to a selected bonding target location of a wafer 6 as a first object (bonding process). The bonding operation may involve lifting and lowering the bonding head 423, or the wafer stage 43 or wafer chuck 433. To prevent a decrease in positioning accuracy during lifting and lowering, a highly reproducible lifting and lowering drive system may be employed, or the lifting and lowering may be performed under continuous feedback control. To perform lifting and lowering under continuous feedback control, when lifting and lowering the wafer stage 43, the width of the bar mirrors in the Z-axis direction may be designed so that the bar mirrors do not deviate from the optical path of the interferometer even during lifting and lowering. On the other hand, when lifting and lowering the bonding head 423 or wafer chuck 433, feedback control may be performed while monitoring the positional deviations of the bonding head 423 or wafer chuck 433 in the X-axis and Y-axis directions using an encoder or gap sensor. Furthermore, to precisely control the gap between the first object and the second object, a linear encoder may be provided to measure the Z-axis position of the lifting and lowering drive mechanism. Furthermore, when the first object and the second object come into contact, the wafer stage, which is feedback-controlled using an interferometer, is constrained, so the control method may be different before and after the contact, such as stopping the feedback control. Up to this point, we have described the process of bringing the die 51 into contact with the part of the wafer 6 to be bonded, but in the case of bump bonding, steps necessary for bonding, such as pressing the die 51 against the wafer 6 with a predetermined pressure, and a step of observing the bonded state after bonding, may be added.

[0049] When the bonding of one die 51 to the wafer 6 is completed, in step 1006, the control unit CNT determines whether the die 51 as the second object has been bonded to all of the multiple bonding target locations of the wafer 6 as the first object. Typically, tens to hundreds of semiconductor devices are arranged on one wafer 6, and a die 51 is bonded to each of them, so the bonding of the die 51 is repeated multiple times. If the bonding of the die 51 to all of the multiple bonding target locations of the wafer 6 has not been completed, the process returns to the die pick-up in step 2002. Note that in the example of FIG. 3, the above determination is made after the bonding operation in step 1005, and the die 51 is picked up in step 2002. However, the die pick-up in step 2002 may be performed in parallel between the die alignment in step 1003 and the bonding operation in step 1005. Furthermore, when multiple types of die are bonded to one semiconductor device, bonding of the next type of die may begin after bonding of one type of die to all semiconductor devices in one wafer 6 has been completed. In this case, the next type of die is picked up in the die pickup in step 2002. At this time, necessary steps such as the operation of carrying in the dicing frame on which the next type of die is mounted are carried out.

[0050] When bonding of the die 51 to all of the multiple bonding target locations on the wafer 6 is completed, the wafer 6 is unloaded from the bonding device BD in step 1007. The wafer 6 may be returned to the FOUP into which it was loaded, or may be returned to another container. However, in general, the wafers are returned to another container because the thickness of the wafers changes due to bonding and the gaps between the wafers need to be wider than before bonding.

[0051] The flow for bonding multiple second objects to one first object has been explained above, but this operation is repeated for the required number of first objects. Note that the number of dies on the dicing frame generally differs from the number of semiconductor devices on the wafer to which the dies are bonded, so the loading of the wafer and the loading of the dicing frame are not synchronized. If the dies on the dicing frame run out while dies are being bonded to one wafer, the next dicing frame is loaded. Also, if dies remain on the dicing frame after dies have been bonded to all the semiconductor devices on one wafer, those dies are used to bond to the next wafer.

[0052] Next, a method for managing the offset amount reflected in the bonding position drive in step 1004 with respect to the position of the die 51 measured using the die observation camera 431 will be described with reference to the flowchart in Fig. 4. The processing shown in the flowchart in Fig. 4 is controlled by the control unit CNT.

[0053] In step 3001, a wafer 6 as a first object is carried into the bonding apparatus SB and held by the wafer chuck 433. The wafer 6 has formed thereon alignment marks used for aligning the wafer 6 and marks for measuring bonding misalignment, which will be described later. The wafer 6 can also be prepared by, for example, applying a temporary adhesive to the areas to be bonded, so that the die 51 will not be misaligned after being mounted. The wafer 6 is roughly positioned by a pre-aligner (not shown) based on the notch or orientation flat and the wafer outer shape position, and then carried to the wafer chuck 433 as a first holding unit on the wafer stage 43 and held by the wafer chuck 433.

[0054] In step 3002, the position of the alignment mark on the wafer 6 is measured using the wafer observation camera 421, and the mounting position and rotation amount of the wafer 6 are calculated based on the measurement results. During this operation, the surface position of the bonding surface of the wafer 6 may be measured using a first height measuring device (not shown). Since the thickness of the wafer 6 varies, it is advantageous to measure the surface position of the wafer 6 in order to control the gap between the wafer 6 and the die 51 with high precision during the bonding operation.

[0055] In step 3003, a glass die with alignment marks is held by the bonding head 423. The reason for using a glass die is to check for bonding misalignment using the wafer observation camera 421 after bonding. Therefore, the die may be made of a material that transmits light of a wavelength detected by the wafer observation camera 421. For example, when observing using infrared light, a silicon die may be used. The die has alignment marks for measuring the position of the die and marks for measuring bonding misalignment formed thereon.

[0056] In step 3004, the position and rotation amount of the glass die with alignment marks held by the bonding head 423 are measured. During this operation, a second height measuring device (not shown) may be used to measure the surface position of the bonding surface of the glass die with alignment marks. Because the thickness of the glass die with alignment marks varies, measuring the surface position of the glass die with alignment marks is advantageous for accurately managing the gap between the wafer 6 and the die 51 during the bonding operation. Furthermore, the height of the glass die with alignment marks may be measured at multiple positions, and the attitude of the die 51 or the wafer 6 during bonding may be adjusted using a tilt mechanism (not shown). Such a tilt mechanism may be incorporated into the wafer stage 43, the wafer chuck 433, or the bonding head 423.

[0057] In step 3005, the driving mechanism 436 drives the wafer stage 43 so that the glass die with the alignment mark is positioned at a bonding target location selected from a plurality of bonding target locations on the wafer 6. At this time, the control unit CNT can control the driving mechanism 436 so that the position of the wafer stage 43 is fed back based on the measurement results from the interferometer 422. Also, at this time, the control unit CNT can determine the target position of the wafer stage 43 based on the position and rotation amount of the wafer 6 measured in steps 3002 and 3004, and the position, rotation amount, and offset amount of the glass die with the alignment mark.

[0058] In step 3006, similar to step 1005, a glass die with an alignment mark is bonded to the selected bonding target location on the wafer 6.

[0059] In step 3007, the bonding position is measured. Specifically, the drive mechanism 436 drives the wafer stage 43 so that the mark for measuring the bonding misalignment is within the field of view of the wafer observation camera 421, and the amount of bonding misalignment between the wafer 6 and the glass die is measured using the wafer observation camera 421. Examples of marks for measuring the bonding misalignment include a 30-μm-wide square frame on the wafer side and a 60-μm-wide square frame on the glass die side. The two frames are bonded so that they overlap, and the bonding misalignment can be calculated from the amount of misalignment between the two frames. The mark for measuring the bonding misalignment may be a circle instead of a square, or the mark on the wafer side may be an outer mark and the mark on the die side may be an inner mark. Alternatively, two different marks may be measured and the amount of misalignment may be detected from the distance between them. To determine the bonding misalignment, the amount of misalignment may be measured for multiple marks within the glass die. Measuring multiple marks within the glass die allows for measurement of bonding rotation errors and reduces measurement errors through statistical processing, enabling high-precision measurement of the bonding misalignment.

[0060] In step 3008, the control unit CNT calculates an offset amount based on the positional deviation measured using the die observation camera 431. The calculated offset amount may include, for example, shift amounts in the X-axis direction and the Y-axis direction, and a rotation amount around the Z-axis direction. Here, a glass die may be bonded to each of a plurality of bonding target locations on the wafer 6, and an offset amount may be calculated for each of the plurality of bonding target locations. Alternatively, a glass die may be bonded to each of a plurality of bonding target locations on the wafer 6, and a final offset amount may be calculated by averaging the offset amounts calculated for each of the plurality of bonding target locations.

[0061] An example of positioning when bonding a die to a wafer will be described below using the measurement results of the wafer and die positions and a predetermined offset amount. Note that although the signs are reversed depending on how the coordinate direction is taken, the following example follows the coordinate system shown in the figure. The position of the wafer 6 measured in step 1002 (the position relative to the reference point of the bonding device BD) is (Wx, Wy), and the rotation amount is Wθ. Also, the position of the die 51 relative to the center of the image captured in step 1003 is (Dx, Dy), and the rotation amount is Dθ. Also, the shift amount generated during bonding is (Px, Py), and the rotation amount is Pθ. Also, the offset amount calculated in step 3008 is (X0, Y0), θ0.

[0062] If the offset amount in step 3008 is correctly determined, Wx = Wy = Wθ = Dx = Dy = Dθ = 0. When the same process as in step 3008 is used, bonding can be performed with high precision by sending the wafer stage 43 to (X0, Y0) and θ0 for bonding.

[0063] If the position of the wafer 6 deviates from the reference position of the wafer stage 43, for example, if it deviates in the positive direction, it can be corrected by moving the wafer stage 43 in the negative direction by that amount. Therefore, during bonding, it is sufficient to drive the wafer stage 43 to (X0-Wx, Y0-Wy), θ0-Wθ.

[0064] On the other hand, if the position of the die 51 is deviated from the reference of the bonding head 423, for example, if it is deviated in the positive direction, it can be corrected by moving the wafer stage 43 in the positive direction by that amount. Therefore, in order to adjust the bonding position, the wafer stage 43 should be driven to (X0-Wx+Dx, Y0-Wy+Dy), θ0-Wθ+Dθ during bonding.

[0065] Furthermore, since the amount of shift that occurs during bonding is shifted by that amount to reach the bonding position, if there is a shift in the positive direction, the wafer stage 43 should be moved by the same amount to perform bonding. Therefore, during bonding, the wafer stage 43 should be driven to (X0-Wx+Dx+Px, Y0-Wy+Dy+Py), θ0-Wθ+Dθ+Pθ. Second Embodiment The second embodiment will be described below, but matters not mentioned in the second embodiment may follow the first embodiment. Figure 5 is a diagram schematically showing the configuration of a bonding apparatus BD of the second embodiment. In the bonding apparatus BD of the second embodiment, the position of the wafer stage 43 is measured using an encoder.

[0066] Specifically, instead of the interferometer 422 and bar mirror 432 in the bonding apparatus BD of the first embodiment, the bonding apparatus BD of the second embodiment employs an encoder scale 424 and an encoder head 435. The encoder head 435 is a two-dimensional encoder head mounted on the wafer stage 43. The encoder scale 424 is a two-dimensional encoder scale mounted on the upper base 42. The encoder scale 424 has a two-dimensional scale so that the position of the wafer stage 43 can be measured within the movable range of the wafer stage 43. The encoder head 435 measures the position of the wafer stage 43 in the X-axis direction and the Y-axis direction.

[0067] The encoder scale 424 is made of a material with a low thermal expansion coefficient, allowing the scale to be drawn with high positional accuracy. In one example, the encoder scale 424 may be formed by drawing a scale on a quartz substrate using a drawing method used in semiconductor lithography processes. The wafer stage 43 may have a configuration in which a fine movement stage, which is driven with high precision over a small range, is mounted on a coarse movement stage, which is driven over a large range. In such a configuration, the encoder head 435 may be provided on the fine movement stage for high-precision positioning. The control unit CNT may be configured to feedback-control the wafer 6 or the wafer stage 43 with respect to rotation around an axis parallel to the X-axis, Y-axis, and Z-axis, which is perpendicular to the X-axis, Y-axis, and Z-axis, based on the output of the encoder head 435. The drive mechanism 436 may constitute a positioning mechanism that changes the relative position of the wafer stage 43 (or the wafer 6) as a first holder and the bonding head 423 (or the die 51) as a second holder. The encoder head 435 and the control unit CNT may also be understood as components of the positioning mechanism.

[0068] FIG. 6 is a view of the wafer stage 43 as viewed from the positive direction of the Z axis. A method for using the reference plate 434 to ensure the origin position, magnification, and X- and Y-axis directions (rotation) and orthogonality of the wafer stage 43 will be described with reference to FIG. 6. Mark 434a is observed by the wafer observation camera 421, and the output value of the encoder head 435 when mark 434a is positioned at the center of the image output by the wafer observation camera 421 is set as the origin of the wafer stage 43. Next, mark 434b is observed by the wafer observation camera 421, and the Y-axis direction (rotation) and magnification in the Y-axis direction of the wafer stage 43 are determined based on the output value of the encoder head 435 when mark 434b is positioned at the center of the image output by the wafer observation camera 421. Next, mark 434c is observed by wafer observation camera 421, and the X-axis direction (rotation) and magnification in the X-axis direction of wafer stage 43 are determined based on the output value of encoder head 435 when mark 434c is positioned at the center of the output image of wafer observation camera 421. In other words, the axis direction and orthogonality can be calibrated by defining the direction from mark 434b to mark 434a on reference plate 434 as the Y-axis of bonding device BD and the direction from mark 434c to mark 434a as the X-axis of bonding device BD. Calibration can also be performed using the distance between mark 434b and mark 434a as the Y-axis scale reference for bonding device BD, and the distance between mark 434c and mark 434a as the X-axis scale reference for bonding device BD. Because the encoder scale 424 expands due to heat, which causes fluctuations in the measurement values ​​obtained by encoder head 435, it is desirable to perform calibration at any time to ensure the origin position, magnification, rotation, and orthogonality of the wafer stage 43. Instead of using a two-dimensional encoder, a linear encoder may be used for each of the X and Y axes.

[0069] Instead of the above configuration, multiple encoder heads may be arranged and switched between depending on the positions of the parts to be joined, which is advantageous for reducing the footprint. Alternatively, a pair of encoder heads may be arranged symmetrically with respect to the parts to be joined, which is advantageous for improving the position measurement accuracy.

[0070] The above explanation is about an example of performing calibration by observing a reference plate. Alternatively, calibration may be performed by, for example, hitting a reference surface, or a calibration mechanism may be provided inside the encoder to create a position measuring instrument with guaranteed absolute values. <Third embodiment> The third embodiment will be described below, but matters not mentioned in the third embodiment may follow the first embodiment. Figure 7 is a diagram schematically showing the configuration of a bonding apparatus BD of the third embodiment. In the bonding apparatus BD of the third embodiment, the relative position of the wafer 6 as the first object and the die 51 as the second object is changed or adjusted by positioning the bonding head 453.

[0071] The bonding unit 4 may include an upper base 42 and a lower base 44. The upper base 42 may support a bonding stage 45. The bonding stage 45 may be driven in the X-axis direction (first direction) and the Y-axis direction (second direction) by a driving mechanism 437 such as a linear motor. The driving mechanism 437 may be configured to further drive the bonding stage 45 to rotate about an axis parallel to the Z-axis direction (third direction). Instead of driving the bonding stage 45 to rotate about an axis parallel to the Z-axis direction, the driving mechanism 437 may drive the wafer chuck 443 to rotate about an axis parallel to the Z-axis direction. The driving mechanism 437 may constitute a positioning mechanism that changes the relative position of the wafer chuck 433 (or the wafer 6) as a first holder and the bonding head 453 (or the die 51) as a second holder.

[0072] A wafer observation camera 451 serving as a first camera may be mounted on the bonding stage 45. The wafer observation camera 451 is a first detector for detecting the position of a characteristic portion of a wafer 6 serving as a first object held by the wafer chuck 433. The bonding stage 45 may further be mounted with a bonding head 453 serving as a second holder that receives and holds a die 51 serving as a second object handed over from the pickup head 31 and bonds it to a bonding target portion of the wafer 6. In the example of FIG. 7 , the bonding stage 45 may constitute a support that supports the bonding head 453 serving as the second holder and the wafer observation camera 451 serving as the first camera. A bar mirror 452 may be provided on the bonding stage 45. The bar mirror 452 may be used as a target for the interferometer 442.

[0073] A die observation camera 441 serving as a second camera may be mounted on the lower base 44. The die observation camera 441 is a second detector for detecting the position of a characteristic portion of a die 51 serving as a second object held by the bonding head 453. A wafer chuck 433 serving as a first holder may be mounted on the lower base 44. The wafer chuck 433 holds a wafer 6 serving as the first object. An interferometer 442 may also be mounted on the lower base 44 for measuring the position of the bonding stage 45 using a bar mirror 452. In the example of FIG. 7 , the lower base 44 functions as a support structure for supporting the wafer chuck 433 serving as the first holder and the die observation camera 441 serving as the second camera.

[0074] When bonding the die 51 as a second object to a bonding target portion of the wafer 6 as a first object, the bonding head 453 can drive the die 51 in the negative direction (downward) of the Z axis to bond the die 51 to the bonding target portion of the wafer 6. Alternatively, the drive mechanism 437 can drive the bonding stage 45 in the negative direction (downward) of the Z axis to bond the die 51 to the bonding target portion of the wafer 6. Alternatively, the drive mechanism (not shown) can drive the wafer chuck 433 in the positive direction (upward) of the Z axis to bond the die 51 to the bonding target portion of the wafer 6.

[0075] FIG. 8 is a view of the bonding stage 45 as viewed from the negative direction of the Z axis. The bonding head 453 holds a die 51. The die 51 can be positioned with respect to rotation about an axis parallel to the X-axis direction (first direction) and the Y-axis direction (second direction), which are orthogonal or intersecting each other, and the Z-axis direction (third direction), which is orthogonal to the X-axis direction and the Y-axis direction. Therefore, the bonding stage 45 can be provided with bar mirrors 452, more specifically, bar mirrors 452a and 452b. The bar mirror 452a can function as a target for the interferometers 422a and 422c. The control unit CNT can detect the position of the bonding stage 45 in the X-axis direction based on the output of the interferometer 422a, and can detect the rotation of the bonding stage 45 about an axis parallel to the Z-axis direction based on the outputs of the interferometers 422a and 422c. The bar mirror 452b can function as a target for the interferometer 422b. The control unit CNT can detect the position of the bonding stage 45 in the Y-axis direction based on the output of the interferometer 422b. The control unit CNT can be configured to feedback-control the die 51 or the bonding stage 45 with respect to rotations in the X-axis direction, the Y-axis direction, and around an axis parallel to the Z-axis direction orthogonal to the X-axis direction and the Y-axis direction, based on the outputs of the interferometers 422a, 422b, and 422c. The interferometer 422 and the control unit CNT may be understood as components of the positioning mechanism described above.

[0076] A reference plate 454 is provided on the underside of the bonding stage 45. A plurality of marks 454a, 454b, and 454c are arranged on the reference plate 454. The reference plate 454 is made of a material with a low thermal expansion coefficient, allowing the marks to be drawn with high positional accuracy. For example, the reference plate 454 may be formed by drawing marks on a quartz substrate using a drawing method used in semiconductor lithography processes. The reference plate 454 has a surface that is approximately flush with the surface of the die 51 and can be observed by the die observation camera 441. Alternatively, a separate camera may be provided for observing the die observation camera 441. The bonding stage 45 may be configured by combining a coarse-motion stage that can be driven over a large range and a fine-motion stage that can be driven over a small range with high precision. In such a configuration, the wafer observation camera 451, bar mirrors 452a and 452b, bonding head 453, and reference plate 454 may be mounted on the fine-motion stage to achieve high-precision positioning.

[0077] Here, a method for ensuring the origin position, magnification, X-axis and Y-axis directions (rotation), and orthogonality of the bonding stage 45 using the reference plate 454 will be described. Mark 454a is observed with the die observation camera 441, and the output value of the interferometer when mark 454a is located at the center of the output image of the die observation camera 441 is set as the origin of the bonding stage 45. Next, mark 454b is observed with the die observation camera 441, and the Y-axis direction (rotation) and Y-axis magnification of the bonding stage 45 are determined based on the output value of the interferometer when mark 454b is located at the center of the output image of the die observation camera 441. Next, mark 454c is observed with the die observation camera 441, and the X-axis direction (rotation) and X-axis magnification of the bonding stage 45 are determined based on the output value of the interferometer when mark 454c is located at the center of the output image of the die observation camera 441.

[0078] Specifically, the direction from mark 454b to mark 454a on the reference plate 454 serves as the Y-axis of the bonding device BD, and the direction from mark 454c to mark 454a serves as the X-axis of the bonding device BD, allowing for calibration of the axis direction and orthogonality. Calibration can also be performed using the distance between mark 454b and mark 454a as the Y-axis scale reference for the bonding device BD, and the distance between mark 454c and mark 454a as the X-axis scale reference for the bonding device BD. Because the refractive index of the optical path of an interferometer changes due to fluctuations in air pressure and temperature, which can cause fluctuations in measurement values, it is desirable to perform calibration at any time to ensure the origin position, magnification, rotation, and orthogonality of the bonding stage 45. To reduce fluctuations in the interferometer measurement values, it is desirable to enclose the space in which the bonding stage 45 is located in a temperature-controlled chamber and control the temperature within the chamber.

[0079] In this embodiment, the reference plate on the bonding stage is observed by the die observation camera. Alternatively, the reference plate may be attached to the lower base and observed by the wafer observation camera, which can ensure the origin position, magnification, rotation, and orthogonality of the bonding stage.

[0080] The above explanation is about an example of performing calibration by observing a reference plate. Alternatively, calibration may be performed by, for example, abutting against a reference surface, or high-precision positioning may be performed using a position measuring instrument whose absolute value is guaranteed, such as a white light interferometer.

[0081] In the third embodiment, since the bonding position and the location measured by the interferometer are separated, it is desirable to correct the Abbe error. Also, the error may be reduced by measuring on both sides of the bonding stage.

[0082] The bonding flow of the third embodiment will be described below with reference to the flowchart in Fig. 3. This bonding flow is controlled by the control unit CNT. In step 1001, a wafer 6 as a first object is carried into the bonding device BD and held by the wafer chuck 433. The wafer 6 is roughly positioned by a pre-aligner (not shown) based on the notch or orientation flat and the wafer outer shape position, and then carried to the wafer chuck 443 as a first holder on the lower base 44 and held by the wafer chuck 433.

[0083] In step 1002, the mounting position of the wafer 6 is measured using the wafer observation camera 451. Focus adjustment may be provided by providing a focus adjustment mechanism within the wafer observation camera 451, or by providing a Z-axis drive mechanism in the wafer chuck 443 and driving the wafer 6 about the Z axis using the Z-axis drive mechanism. Alternatively, focus adjustment may be provided by providing a Z-axis drive mechanism in the bonding stage 45 and driving the wafer observation camera 451 about the Z axis using the Z-axis drive mechanism. While alignment marks for alignment are often formed on the wafer 6, if no alignment marks are formed, a feature whose position can be identified can be measured. The control unit CNT can detect the relative position of the image of the feature with respect to the center of the output image of the wafer observation camera 451 as the position of the feature.

[0084] An offset amount may be calculated in advance to measure the relative position of the mark with respect to the reference point of the bonding device BD with high accuracy. This may include a process of driving the bonding stage 45 so that the mark on the reference plate 454 is within the field of view of the wafer observation camera 451, and measuring the position of the mark with the wafer observation camera 451. The driving position of the bonding stage 45 at that time and the offset amount relative to the position measured using the wafer observation camera 451 can be determined. Here, the reference point of the bonding device BD is generally set to the position of a specific mark on the reference plate 454, but it may be set to another location as long as it serves as a reference position.

[0085] Because the measurement range of the interferometer in the rotational direction is narrow, the amount of rotation that can be corrected by the bonding stage 45 is small. Therefore, if the amount of rotation of the wafer 6 is large, it is desirable to correct the rotation and re-hold the wafer 6. After re-holding, it is necessary to measure the mounting position of the wafer 6 again. During this operation, the surface position of the bonding surface of the wafer 6 may be measured using a first height measuring device (not shown). Since the thickness of the wafer 6 varies, it is advantageous to measure the surface position of the wafer 6 in order to accurately manage the gap between the wafer 6 and the die 51 during the bonding operation.

[0086] The origin position, magnification, direction (rotation) of the X-axis and Y-axis, and orthogonality of the bonding stage 45 are guaranteed using the reference plate 454, so the position of the mounted wafer 6 is measured based on the origin position of the bonding stage 45 and the X-axis and Y-axis.

[0087] The following describes the movement of the die as the second object, which is performed in parallel with or after the loading of the wafer as the first object and wafer alignment. In step 2001, a dicing frame 5 on which die 51, which have been singulated by a dicer, are arranged on a dicing tape, is loaded into the bonding device BD. In step 2002, the die 51 as the second object is picked up by the pickup head 31.

[0088] In step 2003, the die 51 as the second object picked up by the pickup head 31 is transferred to the bonding head 453. When the die 51 is picked up by the pickup head 31, the semiconductor device surface faces the pickup head 31. On the other hand, the die 51 is transferred to the bonding head 453 so that the surface opposite to the semiconductor device surface faces the bonding head 453. The die 51 may be transferred to the bonding head 453 directly from the pickup head 31 to the bonding head 453, or may be transferred via multiple die holders. In addition, pretreatment for bonding may be performed while the die 51 is being transferred. The pretreatment may include, for example, a die cleaning process, application of an adhesive in the case of adhesive bonding, or a surface activation process in the case of hybrid bonding. Note that if the surface of the die 51 becomes inactive while being transferred to the bonding head 453, it is desirable to perform a process of activating the bonding surface using an atmospheric pressure plasma activation device after the die 51 is mounted on the bonding head 453. With the above steps, the wafer 6 as the first object and the die 51 as the second object are held by their respective holders. Next, the bonding flow will be described. In step 1003, the position of the die 51 as the second object held by the bonding head 453 can be measured. Specifically, the driving mechanism 437 can drive the bonding stage 45 so that a characteristic portion of the die 51 is within the field of view of the die observation camera 441. Focus adjustment can be provided by providing a focus adjustment mechanism within the die observation camera 441, or by providing a Z-axis driving mechanism in the bonding head 453 and driving the die 51 about the Z axis using the Z-axis driving mechanism.

[0089] Scribe lines on which alignment marks used for alignment in semiconductor manufacturing processes are formed can be removed by dicing. Therefore, the die 51 often does not have alignment marks for alignment. Therefore, the position of the die 51 can be measured using the end of an array of pads or bumps arranged on the die 51, a region with a non-periodic arrangement that allows its position to be identified, or the outer shape of the die as a characteristic feature. The control unit CNT can determine the position of the characteristic feature based on the relative position of the image of the characteristic feature with respect to the center of the output image of the die observation camera 441. It is necessary to manage the offset amount when positioning the die 51 at the joining location based on the position of the die 51 measured using the die observation camera 441; this method will be described later.

[0090] When measuring the position of the die 51, it is desirable to measure the positions of multiple characteristic locations within the die 51 and also measure the amount of rotation of the die 51. To measure the positions of multiple characteristic locations, the bonding stage 45 may be driven each time the position of each characteristic location is measured, or the field of view of the die observation camera 441 may be designed so that multiple characteristic locations can be observed at once. The die 51 can be rotated by rotating the bonding stage 45 during bonding. However, because the measurement range of the interferometer in the rotational direction is narrow, if the amount of rotation of the die 51 is large, it is desirable to correct the rotation and re-hold the die 51. After re-holding the die 51, it is necessary to measure the position of the die 51 again. During this operation, a second height measuring device (not shown) may be used to measure the surface position of the bonding surface of the die 51, which serves as the second object. Because the thickness of the die 51 varies, it is advantageous to measure the surface position of the die 51 in order to accurately manage the gap between the wafer 6 and the die 51 during the bonding operation. Furthermore, the height of the die 51 at multiple positions may be measured, and the posture of the die 51 or the wafer 6 may be adjusted by a tilt mechanism (not shown) during bonding. Such a tilt mechanism may be incorporated into the wafer chuck 433 or the bonding head 453.

[0091] In step 1004, the driving mechanism 437 drives the bonding stage 45 so that the die 51, which is the second object, is positioned at a bonding target location selected from a plurality of bonding target locations on the wafer 6, which is the first object. At this time, the control unit CNT can control the driving mechanism 437 so that the bonding stage 45 is feedback-controlled based on the measurement results from the interferometer 442. Also, at this time, the control unit CNT can determine the target position of the bonding stage 45 based on the position and rotation amount of the wafer 6 and the position and rotation amount of the die 51 measured in steps 1002 and 1003, as well as the offset amount. Also, if a shift occurs due to the bonding operation as described below, the control unit CNT considers that amount as the offset amount.

[0092] In step 1005, a die 51 as a second object is bonded to a selected bonding target location of a wafer 6 as a first object. The bonding operation may involve raising and lowering the bonding stage 45 or the bonding head 453, or the wafer chuck 443. To prevent a decrease in positioning accuracy during the raising and lowering operation, a highly reproducible lifting and lowering drive system may be employed, or the lifting and lowering operation may be performed under continuous feedback control. To perform the lifting and lowering operation under continuous feedback control, when raising and lowering the bonding stage 45, the width of the bar mirrors in the Z-axis direction may be designed so that the bar mirrors do not deviate from the optical path of the interferometer even during the raising and lowering operation. On the other hand, when raising and lowering the bonding head 453 or the wafer chuck 443, feedback control may be performed while monitoring the positional deviations of the bonding head 453 or the wafer chuck 443 in the X-axis and Y-axis directions using an encoder or gap sensor. Furthermore, to precisely control the gap between the first object and the second object, a linear encoder may be provided to measure the Z-axis position of the lifting and lowering drive mechanism. Furthermore, when the first object and the second object come into contact, the bonding stage 45, which is feedback-controlled using an interferometer, is constrained, so the control method may be different before and after the contact, such as stopping the feedback control. Up to this point, we have described the process of bringing the die 51 into contact with the bonding target portion of the wafer 6, but in the case of bump bonding, steps required for bonding, such as pressing the die 51 against the wafer 6 with a predetermined pressure, and a step of observing the bonding state after bonding, may be added.

[0093] Step 1006 and subsequent steps are the same as in the first embodiment, and therefore the explanation will be omitted.

[0094] Next, a method for managing the offset amount reflected in the bonding position drive in step 1004 with respect to the position of the die 51 measured using the die observation camera 441 will be described with reference to the flowchart in Fig. 4. The processing shown in the flowchart in Fig. 4 is controlled by the control unit CNT.

[0095] In step 3001, a wafer 6 as a first object is carried into the bonding apparatus SB and held by the wafer chuck 433. The wafer 6 has marks formed thereon that are used for aligning the wafer 6 and marks for measuring bonding misalignment. The wafer 6 can also be prepared by, for example, applying a temporary adhesive to the areas to be bonded to prevent misalignment of the die 51 after mounting the die 51. The wafer 6 is roughly positioned by a pre-aligner (not shown) based on the notch or orientation flat and the outer diameter position of the wafer, and then carried to the wafer chuck 443 as a first holder on the lower base 44 and held by the wafer chuck 433.

[0096] In step 3002, the position of the alignment mark on the wafer 6 is measured using the wafer observation camera 451, and the mounting position and rotation amount of the wafer 6 are calculated based on the measurement results. In this operation, the surface position of the bonding surface of the wafer 6 may be measured using a first height measuring device (not shown). Since the thickness of the wafer 6 varies, it is advantageous to measure the surface position of the wafer 6 in order to control the gap between the wafer 6 and the die 51 with high precision during the bonding operation.

[0097] In step 3003, a glass die with alignment marks is held by a bonding head 453. The reason for using a glass die is to check for bonding misalignment using a wafer observation camera 451 after bonding. Therefore, the die may be made of a material that transmits light of a wavelength detected by the wafer observation camera 451. For example, when observing using infrared light, a silicon die may be used. The die has alignment marks for measuring the position of the die and marks for measuring bonding misalignment formed thereon.

[0098] In step 3004, the position and rotation amount of the glass die with alignment marks held by the bonding head 453 are measured. During this operation, a second height measuring device (not shown) may be used to measure the surface position of the bonding surface of the glass die with alignment marks. Because the thickness of the glass die with alignment marks varies, measuring the surface position of the glass die with alignment marks is advantageous for accurately managing the gap between the wafer 6 and the die 51 during the bonding operation. Furthermore, the height of the glass die with alignment marks may be measured at multiple positions, and the attitude of the die 51 or the wafer 6 during bonding may be adjusted using a tilt mechanism (not shown). Such a tilt mechanism may be incorporated into the wafer chuck 433 or the bonding head 453.

[0099] In step 3005, the driving mechanism 437 drives the bonding stage 45 so that the glass die with the alignment mark is positioned at a bonding target location selected from a plurality of bonding target locations on the wafer 6. At this time, the control unit CNT can control the driving mechanism 437 so that the position of the bonding stage 45 is fed back based on the measurement results from the interferometer 442. Also, at this time, the control unit CNT can determine the target position of the bonding stage 45 based on the position and rotation amount of the wafer 6 measured in steps 3002 and 3004, and the position, rotation amount, and offset amount of the glass die with the alignment mark.

[0100] In step 3006, similar to step 1005, a glass die with an alignment mark is bonded to the selected bonding target location on the wafer 6.

[0101] In step 3007, the bonding position is measured. Specifically, the driving mechanism 437 drives the bonding stage 45 so that the mark for measuring the bonding deviation comes into the field of view of the wafer observation camera 451, and the amount of bonding deviation between the wafer 6 and the glass die is measured using the wafer observation camera 451. In step 3008, the control unit CNT calculates an offset amount based on the positional deviation measured using the wafer observation camera 451. The calculated offset amount may include, for example, shift amounts in the X-axis direction and the Y-axis direction, and a rotation amount around the Z-axis direction. Here, a glass die may be bonded to each of a plurality of bonding target locations on the wafer 6, and an offset amount may be calculated for each of the plurality of bonding target locations. Alternatively, a glass die may be bonded to each of a plurality of bonding target locations on the wafer 6, and a final offset amount may be calculated by averaging the offset amounts calculated for each of the plurality of bonding target locations.

[0102] Below, an example of positioning when bonding a die to a wafer will be described using the measurement results of the wafer and die positions and a predetermined offset amount. Note that although the signs are reversed depending on how the coordinate direction is taken, the following example follows the coordinate system shown in the figure. The position of the wafer 6 measured in step 1002 (the position relative to the reference point of the bonding device BD) is (Wx, Wy), and the rotation amount is Wθ. Also, the position of the die 51 relative to the center of the image captured in step 1003 is (Dx, Dy), and the rotation amount is Dθ. Also, the shift amount generated during bonding is (Px, Py), and the rotation amount is Pθ. Also, the offset amount calculated in step 3008 is (X0, Y0), θ0.

[0103] If the offset amount in step 3008 is correctly calculated, Wx = Wy = Wθ = Dx = Dy = Dθ = 0. When the same process as in step 3008 is used, bonding can be performed with high precision by sending the bonding stage 45 to (X0, Y0) and θ0 for bonding.

[0104] If the position of the wafer 6 deviates from the reference position of the wafer stage 43, for example, if it deviates in the positive direction, it can be corrected by moving the bonding stage 45 in the negative direction by that amount. Therefore, during bonding, it is sufficient to drive the bonding stage 45 to (X0+Wx, Y0+Wy), θ0+Wθ.

[0105] On the other hand, if the position of the die 51 deviates from the reference of the bonding head 423, for example, if it deviates in the positive direction, it can be corrected by moving the bonding stage 45 in the negative direction by that amount. Therefore, in order to adjust the bonding position, the bonding stage 45 should be driven to (X0+Wx-Dx, Y0+Wy-Dy), θ0+Wθ-Dθ during bonding.

[0106] Furthermore, since the amount of shift that occurs during bonding is offset by that amount to reach the bonding position, if there is a shift in the forward direction, bonding can be performed by moving the bonding stage 45 in the reverse direction by the same amount. Therefore, during bonding, the bonding stage 45 can be driven to (X0+Wx-Dx-Px, Y0+Wy-Dy-Py), θ0+Wθ-Dθ-Pθ.

[0107] <Fourth embodiment> The fourth embodiment will be described below, but matters not mentioned in the fourth embodiment may follow the third embodiment, or the first embodiment via the third embodiment. Figure 9 is a diagram schematically showing the configuration of a bonding apparatus BD of the second embodiment. In the bonding apparatus BD of the fourth embodiment, the position of the bonding stage 45 is measured using an encoder.

[0108] Specifically, instead of the interferometer 442 and bar mirror 452 in the bonding apparatus BD of the third embodiment, the bonding apparatus BD of the fourth embodiment employs an encoder scale 444 and an encoder head 455. The encoder head 455 is a two-dimensional encoder head mounted on the bonding stage 45. The encoder scale 444 is a two-dimensional encoder mounted on the lower base 44. The encoder scale 444 has a two-dimensional scale so that the position of the bonding stage 45 can be measured within the range of motion of the bonding stage 45. The encoder head 455 measures the position of the bonding stage 45 in the X-axis direction and the Y-axis direction.

[0109] The encoder scale 444 is made of a material with a low thermal expansion coefficient, and the scale can be written with high positional accuracy. In one example, the encoder scale 444 can be formed by writing a scale on a quartz substrate using a writing method used in semiconductor lithography processes. The bonding stage 45 can be configured by combining a coarse movement stage that moves over a large range and a fine movement stage that moves over a small range with high precision. In such a configuration, the encoder head 455 can be fixed on the fine movement stage for high-precision positioning. The drive mechanism 437 can constitute a positioning mechanism that changes the relative position of the wafer chuck 443 (or wafer 6) as a first holder and the bonding head 453 (or die 51) as a second holder. The encoder head 455 and the control unit CNT can also be understood as components of the positioning mechanism.

[0110] 10, a method for ensuring the origin position, magnification, X-axis and Y-axis directions (rotation), and orthogonality of the bonding stage 45 using the reference plate 454 will be described. The output value of the interferometer when mark 454a is positioned at the center of the output image of the die observation camera 441 is set as the origin of the bonding stage 45. Next, mark 454b is observed with the die observation camera 441, and the Y-axis direction (rotation) and Y-axis magnification of the bonding stage 45 are determined based on the output value of the encoder head 455 when mark 454b is positioned at the center of the output image of the die observation camera 441. Next, mark 454c is observed with the die observation camera 441, and the X-axis direction (rotation) and X-axis magnification of the bonding stage 45 are determined based on the output value of the encoder head 455 when mark 454c is positioned at the center of the output image of the die observation camera 441.

[0111] In other words, the direction from mark 454b to mark 454a on the reference plate 454 is defined as the Y-axis of the bonding device BD, and the direction from mark 454c to mark 454a is defined as the X-axis of the bonding device BD, and calibration of the axis direction and orthogonality can be performed. Calibration can also be performed using the distance between mark 454b and mark 454a as the scale reference for the Y-axis of the bonding device BD, and the distance between mark 454c and mark 454a as the scale reference for the X-axis of the bonding device BD. Because the encoder head 455 expands due to heat, which causes fluctuations in the measurement values ​​obtained by the encoder head 455, it is desirable to perform calibration at any time to ensure the origin position, magnification, rotation, and orthogonality of the bonding stage 45. Instead of using a two-dimensional encoder, linear encoders may be used for each of the X-axis and Y-axis.

[0112] Instead of the above configuration, multiple encoder heads may be arranged and switched between depending on the positions of the parts to be joined, which is advantageous for reducing the footprint. Alternatively, a pair of encoder heads may be arranged symmetrically with respect to the parts to be joined, which is advantageous for improving the position measurement accuracy.

[0113] The above explanation is about an example of performing calibration by observing a reference plate. Alternatively, calibration may be performed by, for example, hitting a reference surface, or a calibration mechanism may be provided inside the encoder to create a position measuring instrument with guaranteed absolute values. Fifth Embodiment The fifth embodiment will be described below, but matters not mentioned in the fifth embodiment may follow the first embodiment. FIG. 11 is a diagram schematically showing the configuration of a bonding apparatus BD of the fourth embodiment. In the bonding apparatus BD of the first embodiment, a die observation camera 431 is mounted on the wafer stage 43, but in the bonding apparatus BD of the fifth embodiment, a die observation camera 411 is fixed to a position directly below the bonding head 423. The die observation camera 411 may be fixed to, for example, the upper base 42 or the stage surface plate 41. In other words, the wafer chuck 433 as the first holding unit and the die observation camera 411 as the second camera may be supported by different supports.

[0114] If the die observation camera 411 can be displaced relative to the bonding head 423, the amount of displacement may be measured and corrected. For example, by placing a predetermined mark on the bonding head 423 and observing it with the die observation camera 411, the amount of displacement of the die observation camera 411 relative to the bonding head 423 can be detected. Sixth Embodiment Next, a method for manufacturing an article (such as a semiconductor IC element, a liquid crystal display element, or a MEMS) using the aforementioned bonding apparatus BD will be described. The article is manufactured by preparing a first object, preparing a second object, bonding the first object and the second object together using the aforementioned bonding apparatus to produce a bonded product, and then processing the bonded product through other well-known processes. These other well-known processes include probing, dicing, bonding, packaging, and the like. This method for manufacturing an article makes it possible to manufacture higher-quality articles than conventional methods.

[0115] The invention is not limited to the above-described embodiments, and various changes and modifications can be made without departing from the spirit and scope of the invention. Accordingly, the following claims are appended to apprise the public of the scope of the invention. [Explanation of symbols]

[0116] 6: wafer (first object), 42: upper base (supporting part), 43: wafer stage, 51: die (second object), 421: wafer observation camera (first camera), 423: bonding head (second holding part), 431: die observation camera (second camera), 433: wafer chuck (first holding part), 436: driving mechanism, CNT: control part

Claims

1. A joining device for joining a plurality of second objects to a first object, a first holding unit that holds the first object; a second holding unit that holds the second object; a first camera that captures an image of the first object; a second camera that captures an image of the second object; a support portion that supports the second holding portion and the first camera; a stage that supports the first holding unit and the second camera and is movable in a first direction and a second direction; a driving mechanism that drives the stage; a measuring instrument for measuring the position of the stage; a control unit that controls the drive mechanism, the stage has a plurality of marks formed on a plate; the plurality of marks include marks that are positioned differently from one another in the first direction and marks that are positioned differently from one another in the second direction, The control unit calibrating the position of the stage based on images of the marks captured by the first camera and the measurement results of the measuring instrument; After the calibration, the driving mechanism is controlled to join the second object to the first object based on the output of the first camera, the output of the second camera, and the measurement result of the measuring instrument. A joining device characterized by:

2. the control unit determines a target position of the stage based on the output of the first camera and the output of the second camera, and controls the drive mechanism so that the stage moves to the target position based on the measurement result of the measuring instrument.

2. The joining device according to claim 1.

3. the plurality of marks includes three or more marks; 2. The joining device according to claim 1.

4. the plurality of marks include a first mark, a second mark that is different in position from the first mark in the first direction and is the same in position as the first mark in the second direction, and a third mark that is different in position from the first mark in the second direction and is the same in position as the first mark in the first direction; The control unit controlling the drive mechanism so that the first mark is positioned within a field of view of the first camera, and determining an origin of the stage based on an image of the first mark captured by the first camera and a measurement result of the measuring instrument; controlling the drive mechanism so that the second mark is positioned within a field of view of the first camera, and determining a reference in the first direction of the stage based on an image of the second mark captured by the first camera, an image of the first mark captured by the first camera, and a measurement result of the measuring instrument; controlling the drive mechanism so that the third mark is positioned within a field of view of the first camera, and determining a reference in the second direction of the stage based on an image of the third mark taken by the first camera, an image of the first mark taken by the first camera, and a measurement result of the measuring instrument; 4. The joining device according to claim 3.

5. a pickup unit disposed at a position distant from the stage; the second holding unit holds the second object picked up by the pickup unit.

2. The joining device according to claim 1.

6. the second camera captures an image of the second object when the stage moves so that the first object held by the first holding unit and the second object held by the second holding unit face each other.

2. The joining device according to claim 1.

7. the stage has a first end face on a position side where the second holding unit acquires the second object, and a second end face opposite to the first end face, and the second camera is disposed between the first end face and a plane that passes through a center of the stage and is parallel to the first end face.

2. The joining device according to claim 1.

8. the support unit has a third end surface on a position side where the second holding unit acquires the second object and a fourth end surface opposite to the third end surface, and the first camera is disposed between the third end surface and a plane that passes through a center of the support unit and is parallel to the third end surface. The joining device according to claim 7 .

9. the control unit controls the driving mechanism so as to perform feedback control of the stage based on the measurement result of the measuring instrument.

2. The joining device according to claim 1.

10. The measuring instrument is supported by the support portion.

2. The joining device according to claim 1.

11. the pickup unit includes a release head that peels the second object from the dicing tape, and a pickup head that holds the second object peeled off by the release head.

6. The joining device according to claim 5.

12. the pickup head delivers the second object to the second holding unit; The joining device according to claim 11 .

13. the pickup head rotates after picking up the second object so that the second object faces the second holding unit, and delivers the second object to the second holding unit. The joining device according to claim 12 .

14. the first direction and the second direction are directions along a surface of the stage facing the second holding unit; 2. The joining device according to claim 1.

15. the measuring instrument measures the position of the stage in the first direction and the second direction; The joining device according to claim 14 .

16. the stage is rotated by the driving mechanism around an axis parallel to a third direction perpendicular to the first direction and the second direction; The joining device according to claim 14 .

17. the measuring instrument measures a rotation angle of the stage around an axis parallel to the third direction; 17. The joining device according to claim 16.

18. the stage is disposed between the measuring instrument and the pickup unit; 6. The joining device according to claim 5.

19. the number of the plurality of measurement target locations of the first object imaged by the first camera is smaller than the number of the plurality of second objects joined to the first object; 2. The joining device according to claim 1.

20. The plurality of second objects are each the same type of second object.

2. The joining device according to claim 1.

21. the plurality of second objects include a plurality of types of second objects; 2. The joining device according to claim 1.

22. The second object to be picked up by the pickup unit is determined based on information about the quality of the second object.

6. The joining device according to claim 5.

23. the stage includes a coarse movement stage that is movable within a first range, and a fine movement stage that is movable within a range smaller than the first range; 2. The joining device according to claim 1.

24. the control unit determines an offset amount used to control the drive mechanism based on an image of the mark captured by the first camera.

2. The joining device according to claim 1.

25. The calibration of the stage position adjusts at least one of the origin position, magnification, rotation, and orthogonality of the stage.

2. The joining device according to claim 1.

26. the control unit controls the distance between the first object and the second object based on the measurement results of the surface positions of the plurality of second objects.

2. The joining device according to claim 1.

27. the control unit measures a surface position of each of the plurality of second objects at a plurality of mutually different positions, and controls the tilt of at least one of the first object and the second object based on the results of the surface position measurements.

2. The joining device according to claim 1.

28. A joining method for joining a first object to a plurality of second objects using a stage that is movable in a first direction and a second direction and has a plurality of marks formed on a plate, the method comprising: a first holding step of holding the first object by a first holding part supported by the stage; a second holding step of holding the second object by a second holding unit; a calibration step of calibrating the position of the stage based on images of the marks and measurement results of a measuring instrument; a first imaging step of imaging the first object held by the first holding unit with a first camera after the calibration step; a second imaging step of imaging the second object held by the second holding unit with a second camera supported on the stage after the calibration step; a measuring step of measuring the position of the stage by the measuring instrument; a joining step of moving the stage based on the image captured in the first imaging step, the image captured in the second imaging step, and a measurement result in the measurement step, and joining the second object to the first object, the plurality of marks include marks whose positions are different from each other in the first direction and marks whose positions are different from each other in the second direction; A joining method characterized by:

29. Providing a first object; Providing a second object; forming a joined body by joining the second object to the first object by the joining method according to claim 28; treating the bonded material to obtain an article; A method for manufacturing an article, comprising:

Images