Precise bonded head alignment using a general-purpose reference plate

JP7897985B2Active Publication Date: 2026-07-30ASMPT SINGAPORE PTE LTD
View PDF 9 Cites 0 Cited by

Patent Information

Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
ASMPT SINGAPORE PTE LTD
Filing Date
2025-06-05
Publication Date
2026-07-30

Smart Images

  • Figure 0007897985000001
    Figure 0007897985000001
  • Figure 0007897985000002
    Figure 0007897985000002
  • Figure 0007897985000003
    Figure 0007897985000003
Patent Text Reader

Abstract

To provide an alignment method and apparatus for accurately determining and correcting alignment offsets.SOLUTION: A semiconductor die is picked up with a collet, and the collet is then located between first and second camera systems. The first camera system views the collet and a reference plate, and the second camera system views the semiconductor die and the reference plate, in order to determine a position and orientation of the semiconductor die relative to the collet and the reference plate. Thereafter, the collet is moved to a position above a bonding position on a substrate, and the first camera system views the bonding position so that the position and orientation of the semiconductor die may be adjusted before the semiconductor die is bonded onto the bonding position.SELECTED DRAWING: Figure 1A
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to semiconductor die bonding, and more particularly to performing die bonding with improved accuracy and precision during the die bonding operation.

Background Art

[0002] During the die bonding operation, there are different types of offsets that need to be determined and corrected to ensure that the semiconductor die is accurately bonded onto the bonding position on the substrate. One offset is the offset between the semiconductor die and the bonding position, while another is the offset between the bonding position and the collet or bond head holding the semiconductor die.

[0003] The amount of such offsets is typically obtained by capturing images of the bond head and the semiconductor die using a first vision system, and then capturing images of the bond head and the substrate using a second vision system. However, after performing multiple bonding operations, the subject positions in the captured images shift due to factors such as module deformation and thermal drift. Therefore, when analyzing the aforementioned captured images after the bonding apparatus has been operating for some time, the bonding accuracy tends to be limited due to a decrease in camera calibration accuracy and optical aberration. With just these images, an accurate representation of the general offset is no longer possible.

[0004] In some conventional systems, a vision system with a small aperture number and a large depth of field is deployed in the vision system to be able to view the semiconductor die and the substrate within the same image. Alternatively, in order to guarantee a higher bonding accuracy, the bonding speed may have to be reduced.

[0005] One approach to achieving higher bonding accuracy is described in Patent Document 1, titled "High Precision Bond Head Positioning System and Apparatus." Nevertheless, this device may still suffer from the tilt of the optical system used in the bond head or vision system, which is of unknown cause and can negatively affect alignment accuracy. Furthermore, the position of the subject in the image may also shift after a period of use, and any inherent optical distortion and aberration in the optical system will also affect alignment accuracy.

[0006] It is beneficial to develop alignment methods and apparatus that avoid the aforementioned drawbacks of prior art. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] U.S. Patent No. 10,861,819 [Overview of the project] [Problems that the invention aims to solve]

[0008] Therefore, despite the typical problems faced by the use of prior art approaches, the object of the present invention is to provide alignment methods and apparatus that accurately determine and correct alignment offsets. [Means for solving the problem]

[0009] According to a first aspect of the present invention, a method is provided for joining a semiconductor die to a bonding position on a surface, the method comprising the steps of: picking up the semiconductor die with a collet; positioning the collet between a first camera system and a second camera system; observing the collet and a reference plate with the first camera system to determine the position and orientation of the collet relative to the reference plate when the collet is holding the semiconductor die; observing the semiconductor die and the reference plate with the second camera system to determine the position and orientation of the semiconductor die relative to the reference plate; moving the collet to a position above the bonding position; observing the bonding position with the first camera system to determine the position and orientation of the bonding position relative to the position and orientation of the collet and the position and orientation of the semiconductor die; and then adjusting the position and orientation of the semiconductor die with the collet to bond the semiconductor die to the bonding position.

[0010] According to a second aspect of the present invention, an apparatus is provided for bonding a semiconductor die to a bonding position on a surface, the apparatus comprising a collet for picking up a semiconductor die, a first camera system and a second camera system and a reference plate, wherein when the collet is holding the semiconductor die between the first and second camera systems, the first camera system is operable to view the collet and the reference plate to determine the position and orientation of the collet relative to the reference plate, and the second camera system is operable to view the semiconductor die and the reference plate to determine the position and orientation of the semiconductor die relative to the reference plate, and when the collet moves to a position above the bonding position, the first camera system is further operable to view the bonding position to determine the position and orientation of the collet and the position and orientation of the bonding position relative to the position and orientation of the semiconductor die before adjusting the position and orientation of the semiconductor die with the collet to bond the semiconductor die to the bonding position.

[0011] It would be convenient to describe the present invention in more detail thereafter by referring to the accompanying drawings illustrating specific preferred embodiments of the invention. Details of the drawings and related descriptions should not be understood to supersede the generality of the broad identification of the invention as defined by the claims.

[0012] A specific example of the alignment method and apparatus according to the present invention will be described below with reference to the attached drawings. [Brief explanation of the drawing]

[0013] [Figure 1A] This is a side view of a dial alignment system according to a preferred embodiment of the present invention, in which both the first downward-facing camera and the upward-facing camera are viewing the reference plate. [Figure 1B] This is a side view of the dial alignment system, with an upward-facing camera viewing the die and reference plate. [Figure 2A] This figure shows exemplary images captured by first downward and upward cameras near one corner of the die when performing a die pre-alignment process according to a preferred embodiment of the present invention. [Figure 2B] This figure shows exemplary images captured by first downward and upward cameras near one corner of the die when performing a die pre-alignment process according to a preferred embodiment of the present invention. [Figure 2C] This figure shows exemplary images captured by first downward and upward cameras near one corner of the die when performing a die pre-alignment process according to a preferred embodiment of the present invention. [Figure 2D] This figure shows exemplary images captured by a second downward and upward-facing camera near another, opposite corner of the die. [Figure 2E] This figure shows exemplary images captured by a second downward and upward-facing camera near another, opposite corner of the die. [Figure 2F]A diagram showing adjustment of a collet for aligning a die based on images captured by first and second downward cameras and an upward camera. [Figure 3A] An example of the field of view of the upward camera 24 with respect to the field of view of the downward camera. [Figure 3B] A diagram showing an image captured by the upward camera during the bonding process. [Figure 3C] A diagram showing an image captured by the downward camera during the bonding process. [Figure 4A] A diagram showing exemplary images captured by the first and second downward cameras at a position on a substrate to which a die is to be bonded. [Figure 4B] A diagram showing exemplary images captured by the first and second downward cameras at a position on a substrate to which a die is to be bonded. [Figure 4C] A diagram showing adjustment of a collet for aligning a die to a substrate based on the captured image of the substrate. [Figure 5] A diagram showing where a die is bonded to a substrate.

Best Mode for Carrying Out the Invention

[0014] FIG. 1A is a side view of a die alignment system according to a preferred embodiment of the present invention. The die alignment system is adapted to bond a semiconductor die onto a bonding position on a surface, such as the surface of a substrate.

[0015] The alignment system is generally configured to pick up and hold the semiconductor die 16 and then has a bond head 10 including a collet 12 for bonding the semiconductor die 16 onto a substrate. In use, the collet 12 is disposed between a first camera system (which can be in the form of a downward camera system) including a first downward camera 30 and a second downward camera 32, and a second camera system (which can be in the form of an upward camera system) including a single upward camera 24. Both the first downward camera 30 and the upward camera 24 are shown as looking at a reference plate 26 in FIG. 1A. When the collet 12 is disposed between the first and second camera systems, the first and second downward cameras 30, 32 are respectively disposed adjacent to opposite corners of the collet 12.

[0016] The collet 12 includes a through - portion that enables light rays to move through the body of the collet 12. One approach is to create the through - portion from a transparent material incorporated into the body of the collet 12, for example, by attaching a glass plate 13 that is transparent in the through - portion. The glass plate 13 can extend through the entire thickness of the collet 12 to enable the first and second downward cameras 30, 32 to see through the collet 12, while it is also possible to form through - hole openings 18 along a portion of the thickness of the collet 12 to enable light rays to move through the body of the collet 12. If there are multiple downward cameras 30, 32 in the first camera system, the collet 12 should include a plurality of through - portions respectively corresponding to the locations of each camera included in the first camera system when the collet 12 is disposed between the first and second camera systems.

[0017] The transparent glass plate 13 preferably has a dichroic coating on its first surface, particularly its bottom surface, and a plating on its second surface, particularly its top surface, which includes a metallic fiducial mark. In one embodiment, the dichroic coating on the bottom surface of the glass plate 13 forms a reflective surface 14 for several wavelengths of light within a particular range, while a fiducial, such as metallic dots or other fiducial patterns, is plated on its top surface to form a collet reference mark 20. The dichroic coating on the reflective surface 14 allows at least one wavelength of light (e.g., green light) to be reflected to the top surface of the glass plate 13, while another wavelength of light (e.g., red or blue light) can pass through the reflective surface 14. The dots or other fiducial patterns covered on the top surface should be arranged at a predetermined density such that the dots are suitable for use as collet reference marks 20 with sufficient precision.

[0018] On the side of the bond head 10 opposite to the first and second downward-facing cameras 30, 32, the upward-facing camera 24 is positioned below a reference plate 26 made of a sheet of transparent material, which is placed between the bond head 10 and the upward-facing camera 24. In this arrangement, the upward-facing camera 24 can view the semiconductor die 16 through the reference plate 26. Advantageously, the reference plate 26 can be mounted on top of the upward-facing camera 24 so that it is movable with the upward-facing camera 24, allowing the semiconductor die 16 and the reference plate mark to be simultaneously visible to the upward-facing camera 24 through the reference plate 26.

[0019] The reference plate 26 includes a reference plate mark 28 formed on the surface of a sheet of transparent material, preferably the reference plate mark 28 being in the form of multiple dots or other patterns on the surface of the reference plate 26 that are visible by the upward-facing camera 24 and the downward-facing cameras 30, 32. The multiple dots can form a matrix or grid pattern with an optimal density of dots required for the specific intended application. Other suitable shapes or patterns may also be used.

[0020] The collet 12 and the downward-facing cameras 30 and 32 of the first camera system are preferably driven together using the first motion system 17 (see Figure 4A), while the second motion system 19 is preferably capable of driving only the collet 12, thereby moving it without moving the downward-facing cameras 30 and 32.

[0021] The collet reference mark 20 on the upper surface of the glass plate 13 is visible to the first and second downward-facing cameras 30, 32, and the semiconductor die 16 also includes a die reference mark 22 on its lower surface, which is visible to the upward-facing camera 24. The semiconductor die 16 can be positioned by the collet 12 such that both the die reference mark 22 and the reference plate mark 28 are within the depth of field of the upward-facing camera 24, enabling simultaneous inspection.

[0022] Figure 1A also shows that the collet reference mark 20 on the top surface of the transparent glass plate 13 is visible along the virtual image path 36 by the first camera system, in this case the downward-facing camera 30. Although the glass plate 13 and reference plate 26 of the collet 12 are at different heights, the reflected image of the collet reference mark 20 appearing along the virtual image path 36 and the image of the directly observed reference plate mark 28 appear together to the first and second downward-facing cameras 30, 32 as being at substantially the same height or level.

[0023] To achieve the aforementioned effect, the reflective surface 14 is positioned at a distance from the collet reference mark 20 on the upper surface of the glass plate 13 such that the reflection 20' of the collet reference mark 20 appears to the first downward camera 30 at a height equal to the height of the reference plate mark 28, and both the reflection 20' of the reference plate mark 28 and the collet reference mark 20 appear to the first downward camera 30 within the limited depth of field of the first downward camera 30.

[0024] Therefore, the top surface on which the collet reference mark 20 is formed is positioned at a height such that the total distance traveled by light rays between the collet reference mark 20 on the top surface and the first camera system via the bottom surface (where the reflective surface 14 is located) is the same as the total distance traveled by light rays between the reference plate 26 and the first camera system, so that both the collet reference mark 20 and the reference plate 26 are within the depth of field of the first camera system. As a result, the images of the collet reference mark 20 and the reference plate mark 28 on the reference plate 26 appear to be at substantially the same height or level. Therefore, in a preferred embodiment of the present invention, it should be understood that the collet reference mark 20 should be illuminated by green light reflected by the reflective surface 14, while the reference plate mark 28 should be illuminated by red or blue light that does not reflect in the same way but can instead pass through the reflective surface 14.

[0025] At the position of the bond head 10 shown in Figure 1A, the collet 12 moves to the left in direction A using the second motion system 19 so that the semiconductor die 16 moves out of the field of view of the first downward camera 30, and the first downward camera 30 is then able to see the reference plate 26 through the aperture 18 and glass plate 13 without being obstructed by the presence of the semiconductor die 16. It will be understood that such a movement would not be necessary if the semiconductor die 16 were not so large as to obstruct the downward camera 30 from seeing the reference plate 26. At this position, the first downward camera 30 is configured to see the reference plate mark 28 on the reference plate 26 along the first optical path 34.

[0026] Meanwhile, on the opposite side of the reference plate 26, the upward-facing camera 24 is simultaneously viewing the reference plate mark 28 on the reference plate 26. Simultaneous viewing of the reference plate mark 28 by the first downward-facing camera 30 and the upward-facing camera 24 helps to align the position of the first downward-facing camera 30 with the position of the upward-facing camera 24.

[0027] Figure 1B is a side view of a dial alignment system in which an upward-facing camera 24 is viewing a transparent reference plate 26 and the semiconductor die 16 through the reference plate 26. In this position, the bond head 10 is moved to the right in direction B so that both the die reference mark 22 and the reference plate mark 28 are within the field of view of the upward-facing camera 24. At this time, the reference plate mark 28 may be obstructed from the field of view of the first downward-facing camera 30 by the semiconductor die 16, but the position and orientation of the semiconductor die 16 relative to the reference plate mark 28 can be determined from the image captured by the upward-facing camera 24. Meanwhile, the collet 12 is positioned so that the first downward-facing camera 30 can operate to view the collet reference mark 20 and determine the position and orientation of the collet 12 corresponding to the determined position and orientation of the semiconductor die 16. Therefore, based on the images captured in Figures 1A and 1B, the positions and / or orientations of the first downward-facing camera 30, the upward-facing camera 24, the collet 12, and the semiconductor die 16 relative to the fixed reference plate mark 28 can all be determined.

[0028] Figures 2A to 2C show exemplary images captured by an upward-facing camera 24 and a first downward-facing camera 30 near one corner of the semiconductor die 16 when performing a die pre-alignment process according to a preferred embodiment of the present invention. The first downward-facing camera 30 of the first camera system views the collet 12 and the reference or glass plate 13 to determine the position and orientation of the collet 12 by referring to the collet reference mark 20 on the glass plate 13. In Figure 2A, a top view of the bond head 10 provides a view of the glass plate 13 of the collet 12, and a portion of the unaligned semiconductor die 16 is visible through the glass plate 13. With respect to a bottom view of the bond head 10, it can be seen that the orientation of the semiconductor die 16 is out of alignment as required.

[0029] In Figure 2B, the bond head 10 is moving to the left in direction A using the second motion system 19, and the aperture 18 is shown to be at the upper right corner of the collet 12. From the top view, the reference plate 26 is shown to be roughly below the collet 12, but the reference plate mark 28 can still be seen by the first downward camera 30 through the aperture 18 along the first optical path 34. At the same time, if the collet reference mark 20 is within the field of view of the first downward camera, an image of the collet reference mark 20 can be optionally captured by the first downward camera 30 along the virtual optical path 36. From the bottom, the upward camera 24 correspondingly sees the reference plate mark 28 on the reference plate 26. With these top and bottom images of the reference plate mark 28, a relationship can be established between the upward camera 24 and the downward cameras 30, 32 along the first optical path 34.

[0030] The upward-facing camera 24 of the second camera system should also view the semiconductor die 16 and the reference plate 26 to determine the position and orientation of the semiconductor die 16 relative to the reference plate 26. In Figure 2C, the bond head 10 is moved to the right in direction B using the second motion system 19 so that the die reference mark 22 of the semiconductor die 16 moves into the field of view of the upward-facing camera 24. At this position, the first downward-facing camera 30 captures an image of the collet reference mark 20 from above along a second optical path 42 illuminated by green light reflected by the reflective surface 14. On the other side, the upward-facing camera 24 captures an image of the die reference mark 22 relative to the reference plate mark 28 along a third optical path 44 from the bottom illuminated by red or blue light (with wavelengths that the reflective surface 14 does not reflect).

[0031] Subsequently, the bond head 10 moves so that the second downward-facing camera 32 is positioned above the reference plate 26, and the upward-facing camera 24 is looking at the opposite corner of the collet 12, as shown in Figure 2D. Initially, the bond head 10 moves to the right in the drawing, and the aperture 18 is shown at the lower left corner of the bond head 10. From the top view, the reference plate 26 is shown to be roughly below the collet 12, but the reference plate mark 28 can be seen by the second downward-facing camera 32 through the aperture 18 along the first optical path 34 at this opposite corner of the collet 12. At the same time, if the collet reference mark 20 is within the field of view of the second downward-facing camera 32, an image of the collet reference mark 20 can be optionally captured by the second downward-facing camera 32 along the virtual optical path 36. From the bottom view, the upward-facing camera 24 correspondingly looks at the reference plate mark 28 on the reference plate 26 below the aperture 18.

[0032] In Figure 2E, the bond head 10 moves to the left so that the die reference mark 22 on the semiconductor die 16 moves into the field of view of the upward-facing camera 24. At this position, the second downward-facing camera 32 captures an image of the collet reference mark 20 from above, while the upward-facing camera 24 captures an image of the die reference mark 22 relative to the reference plate mark 28 from below.

[0033] In Figure 2F, based on the orientation of the semiconductor die 16 as determined from the images of its opposing corners captured as described with reference to Figures 2C and 2E, the bond head 10 moves linearly along the horizontal plane in the XY direction and (if necessary) rotates so that the semiconductor die 16 is pre-aligned to the desired orientation with an appropriate level of precision.

[0034] It will be understood that the field of view of the first or second downward-facing cameras 30, 32 and the upward-facing camera 24 often differ. If so, there may be difficulties in relating the die reference marks 22 on the semiconductor die 16 to the collet reference marks 20 when only some of the collet reference marks 20 within the field of view of the downward-facing cameras 30, 32 are within the field of view of the upward-facing camera 24. Also, when the semiconductor die 16 is held by the collet 12 to determine the offset of the semiconductor die 16, the die reference marks 22 on the semiconductor die 16 cannot be seen directly by the first and second downward-facing cameras 30, 32.

[0035] Figure 3A shows an example of the field of view 11' of the upward-facing camera 24, which differs from the field of view 11 of the first or second downward-facing cameras 30, 32. In particular, the rear field of view 11' is smaller than the front field of view 11.

[0036] Calibration is first performed by capturing images of one or more of all collet reference marks 20 within the field of view of the downward-facing cameras 30 and 32 with the downward-facing cameras. In this example, two or more central collet reference marks 20A are within the field of view 11' of the upward-facing camera 24. In addition, two or more peripheral collet reference marks 20B located outside the field of view 11' of the upward-facing camera 24 are included in the images.

[0037] The position of the peripheral collet reference mark 20B outside the field of view 11' is determined relative to the central collet reference mark 20A inside the field of view 11'. The positions of these central and peripheral collet reference marks 20A and 20B are recorded and averaged to create translation vectors corresponding to their relative positions.

[0038] Figure 3B shows an image captured within its field of view 11' by an upward-facing camera 24, which includes a die reference mark 22 and a central collet reference mark 20A when the semiconductor die 16 is held by the collet 12. A translation vector 15 is calculated between the die reference mark 22 and at least one of the central collet reference marks 20A. Since these are generated based on calculations referencing a rigid body, the use of the translation vector 15 ensures consistent measurements despite the multiple bonding cycles that have been performed.

[0039] Figure 3C shows images captured within the field of view 11 by downward-facing cameras 30, 32, compared to the field of view 11' of upward-facing camera 24, when the semiconductor die 16 is present. Reference marks visible to downward-facing cameras 30, 32 include the reference plate mark 28 and the central and peripheral collet reference marks 20A, 20B. However, the die reference mark 22 is on the opposite side of the semiconductor die 16 facing upward-facing camera 24 and is not visible to downward-facing cameras 30, 32.

[0040] Based on the field of view 11' of the upward-facing camera 24, the position of the die reference mark 22 relative to the central collet reference mark 20A is determined. Through calibration from the image in Figure 3A, the position of the die reference mark 22 relative to the central collet reference mark 20A (and therefore the peripheral collet reference mark 20B) can be determined. In this way, the projected position of the die reference mark 22, which is identical to the position of the die reference mark 22 obtained by directly viewing it with the upward-facing camera 24, can be compared, and any differences can be explained.

[0041] After calculating the position of the die reference mark 22 from both upward (Figure 3B) and downward (Figure 3C) images, its actual distance from one or more central collet reference marks 20A is recorded. Using this information, the position and orientation of the entire semiconductor die 16 can be determined after checking the positions of at least two opposing corners of the semiconductor die 16. The precise position of the semiconductor die 16 relative to the collet 12 can then be determined. This approach overcomes any constraints from the limited field of view 11' of the upward camera 24. After calculating the position of the semiconductor die 16, as well as the offset of its orientation relative to the collet 12, the die position can be precisely adjusted.

[0042] In general, the above pre-alignment process should be able to achieve an accuracy level of less than 1 micron. To ensure that this level of accuracy is achieved, the process described in Figures 2B to 2F can be repeated multiple times until the required level of accuracy of less than 1 micron is achieved.

[0043] After a level of pre-alignment accuracy is achieved with respect to the semiconductor die 16 alone, the semiconductor die 16 is then to be aligned with respect to the bonding position on the substrate 52 to which it is to be bonded. Figures 4A to 4B show exemplary images captured by the first and second downward-facing cameras 30, 32 at the position of the semiconductor die 16 above the bonding position on the substrate 52 to which it is to be bonded.

[0044] In Figure 4A, the bond head 10, collet 12, and downward-facing cameras 30 and 32 are positioned above the bonding location on the substrate 52 by the first motion system 17. The first downward-facing camera 30 of the first camera system can then observe the bonding location on the substrate 52 to determine the position and orientation of the bonding location, and can determine the position and orientation of the bonding location relative to the position and orientation of the collet 12 and semiconductor die 16.

[0045] Along a fifth optical path 54 extending between the first downward-facing camera 30 and the substrate 52, the bond head 10 first moves diagonally towards the upper left of the drawing using the second motion system 19 so that the opposing corners of the substrate 52 become visible to the first downward-facing camera 30 through the aperture 18 at one corner of the bond head 10 and to the second downward-facing camera 32 through the aperture 18 at the opposite corner of the bond head 10. It should be noted that if the area of ​​the semiconductor die 16 does not obstruct the first and second downward-facing cameras 30, 32 from viewing the opposing corners of the substrate 52, then there may be no need at all to move the bond head 10 with the second motion system 19 for the first and second downward-facing cameras 30, 32 to view the substrate 52.

[0046] In Figure 4B, the bond head 10 is moved diagonally towards the lower right of the drawing using the second motion system 19 so that the collet reference marks 20 and 21 on the collet 12 can be seen from above by the first and second downward-facing cameras 30 and 32, respectively.

[0047] The positions and orientations of the collet reference marks 20 and 21 relative to the die reference mark 22 on the semiconductor die 16 can be seen from Figures 2B to 2E, and the positions and orientations of the collet reference marks 20 and 21 relative to the bonding position on the substrate 52 can be seen, making it possible to accurately align the semiconductor die 16 to the bonding position. Even after such determination of any relative misalignment, Figure 4C shows the adjustment of the bond head 10 to adjust the position and orientation of the semiconductor die 16 at the collet 12 to align the semiconductor die 16 with the substrate 52 based on the captured image of the bonding position on the substrate 52.

[0048] The alignment process performed between the semiconductor die 16 and the junction position can achieve an accuracy level of 50 nanometers or less. To ensure that this level of accuracy is achieved, the process described in Figures 4A to 4C can be repeated multiple times until the required level of accuracy of less than 50 nanometers is achieved.

[0049] Once the required level of precision is achieved, the semiconductor die 16 is bonded to the bonding position on the substrate 52 with high precision and accuracy, as shown in Figure 5, and the collet 12 is operated so that the orientation of the semiconductor die 16 is precisely aligned with the bonding position on the substrate 52.

[0050] It should be understood that the required depth of field for the alignment apparatus and method according to the embodiments described in the present invention is relatively small, and therefore a larger numerical aperture can be achieved to obtain high alignment accuracy without incurring high component costs. Further accuracy can be achieved by associating the collet reference mark 20 and the die reference mark 22 with the same reference plate mark 28.

[0051] Pre-alignment of the semiconductor die 16 with respect to a reference plate 26 mounted on top of the upward-facing camera 24 allows for pre-alignment with the reference plate mark 28, minimizing errors caused by factors such as distortion, aberration, module deformation, or thermal drift that may occur during use. Furthermore, the reference plate mark 28 reduces the sensitivity of the upward-facing and downward-facing cameras 24, 30, and 32 to inclination, simplifying the design of the upward-facing camera 24 without the need to adapt it to view reference marks positioned at different subject levels.

[0052] Furthermore, this alignment system guarantees long-term bonding accuracy and significantly improves accuracy compared to prior art systems without negatively impacting bonding speed.

[0053] The inventions described herein are subject to variations, alterations, and / or additions other than those specifically described, and it should be understood that the inventions include all such variations, alterations, and / or additions that fall within the spirit and scope of the above description. [Explanation of Symbols]

[0054] 10 Bondhead 11 Field of view of the first or second downward-facing cameras 30, 32 11' Field of view of upward-facing camera 24 12 Colettes 13 Glass Plate 14 Reflective surface 15 Translation Vectors 16 Semiconductor Dies 17. The First Motor System 18 Aperture 19. The second motor system 20 Collet Reference Marks 20' reflection 20A Center collet reference mark 20B Peripheral collet reference mark 21 Collet reference mark 22 Die standard mark 24 Upward-facing camera 26. Standard plate 28 Standard Plate Mark 30. First downward-facing camera 32. Second downward-facing camera 34. The First Optical Path 36 Virtual Image Light Path 42. Second optical path 44 The third optical path 52 circuit boards 54 The fifth optical path

Claims

1. A method for joining a semiconductor die to a bonding position on a surface, The steps include: lifting the semiconductor die with a collet, The steps include: positioning the collet between the first camera system and the second camera system; The first camera system observes the collet and the reference plate, and when the collet is holding the semiconductor die, the position and orientation of the collet relative to the reference plate is determined. The steps include: observing the semiconductor die and the reference plate with the second camera system and determining the position and orientation of the semiconductor die relative to the reference plate; A step of moving the collet to a position above the aforementioned joining position, The first camera system observes the bonding position and determines the position and orientation of the bonding position relative to the position and orientation of the collet and the position and orientation of the semiconductor die. The steps include then adjusting the position and orientation of the semiconductor die with the collet to bond the semiconductor die to the bonding position, Methods that include...

2. The method according to claim 1, wherein the first camera system includes a downward-facing camera system, and the second camera system includes an upward-facing camera system.

3. The method according to claim 1, wherein the collet includes a see-through portion that allows a ray of light to travel through the body of the collet.

4. The method according to claim 3, wherein the see-through portion includes a transparent material.

5. The method according to claim 4, wherein the transparent material includes a transparent plate having a dichroic coating on its first surface to form a reflective surface for only some wavelengths of light within a specific range.

6. The method according to claim 5, wherein the transparent plate includes a second surface opposite to the first surface, and the second surface includes a collet reference mark.

7. The method according to claim 6, wherein the collet reference mark is formed by plating a metal pattern onto the second surface.

8. The method according to claim 6, wherein the second surface is positioned at a height such that the total distance traveled by light rays from the second surface through the first surface to the first camera system is equal to the total distance traveled by light rays between the reference plate and the first camera system, and the collet reference mark and the reference plate are both within the depth of field of the first camera system.

9. The method according to claim 8, wherein the image of the collet reference mark and the image of the reference plate appear to be at substantially the same height or level.

10. The method according to claim 1, wherein the reference plate includes a sheet of transparent material and a reference plate mark formed on the surface of the sheet of transparent material.

11. The method according to claim 10, wherein the reference plate mark includes a plurality of dots or other fiducial patterns.

12. The method according to claim 10, wherein the sheet of transparent material is mounted on the second camera system such that the semiconductor die and the reference plate mark can be simultaneously viewed by the second camera system through the reference plate.

13. The method according to claim 1, wherein the first camera system is movable together with the collet while being driven by a first motion system, and the collet is driveable by a second motion system to move without moving the first camera system.

14. The method according to claim 13, further comprising the step of driving the collet with the second motion system to move the semiconductor die away from the field of view of the first camera system before viewing the reference plate with the first camera system.

15. The method according to claim 14, wherein the collet includes a plurality of see-through portions, each corresponding to the location of each camera included in the first camera system when the collet is positioned between the first and second camera systems.

16. The method according to claim 1, wherein the first camera system includes first and second cameras, respectively, positioned adjacent to opposing corners of the collet when the collet is positioned between the first and second camera systems.

17. An apparatus for bonding a semiconductor die to a bonding position on a surface, A collet for picking up the aforementioned semiconductor die, A first camera system and a second camera system, Reference plate and Includes, When the collet is holding the semiconductor die between the first and second camera systems, the first camera system is operable to view the collet and the reference plate in order to determine the position and orientation of the collet relative to the reference plate, and the second camera system is operable to view the semiconductor die and the reference plate in order to determine the position and orientation of the semiconductor die relative to the reference plate. As the collet moves to a position above the bonding position, the first camera system is further operable to view the bonding position in order to determine the position and orientation of the collet and the position and orientation of the bonding position relative to the position and orientation of the semiconductor die, before adjusting the position and orientation of the semiconductor die with the collet to bond the semiconductor die onto the bonding position. Device.