Three dimensional integrated circuits - Part 2: Alignment of stacked dies having fine pitch interconnect
1 Scope
This document provides specifications of initial alignment and alignment maintenance between multiple stacked integrated circuits during the die bonding process. These specifications define the alignment keys and operating procedures of the keys.
This document apply only if electrical coupling method of die-to-die alignment is used in the die stacking.
2 Normative references
The following documents are referred to in the text in such a way that some or all of their
content constitutes requirements of this document. For dated references, only the edition
cited applies. For undated references, the latest edition of the referenced document (including any amendments) applies.
IEC 63011-1 Integrated circuits − Three dimensional Integrated Circuits - Part 1: Terminology
Note: GB/T 43536.1-2023, Three dimensional integrated circuit – Part 1: Terminologies and definitions (IEC 63011-1:2018, IDT)
3 Terms and definitions
For the purposes of this document, the terms and definitions given in IEC 63011-1 apply.
3.1
die bonding
assembly step to adhere physically or chemically a die to another
3.2
bonder
apparatus performing die bonding
3.3
signal generator
apparatus generating electrical signals
3.4
alignment key
apparatus to monitor or adjust the alignment of the overlaid dies
3.5
aligner
apparatus to perform the alignment of the overlaid dies
4 Die alignment during three dimensional integration
4.1 Alignment during stacking
Once the upper die covers the bottom one, the patterns including align key on the bottom die are not seen any longer. Therefore, the image of the bottom die is stored in the memory. As the upper die is moved above the bottom die, the patterns on the upper die are compared with the stored patterns to be precisely aligned. The procedure is illustrated in Figure 1 . The cross-patters indicate alignment keys, and they are placed at the same location on every die. The alignment keys are also used as a positioning reference for all other patterns on the die. The position of patters on the upper die is compared with memorized images of the lower die because the pattern on the lower die is covered by the upper one and no longer seen by general alignment tools.
Figure 1 - Procedure of alignment of dies during die stacking
4.2 Alignment maintenance during die bonding
After the upper die is placed on top of the bottom one, the bonding process is proceeded to give the permanent physical contact on the bonder as shown in Figure 2. The bonding process is involved with thermal and mechanical agitation to provide the adhesive contact between the TSV and micro bump. Physical agitation destroys the alignment. The image of the bottom die is not observed by optical microscope using visible light. Although the infrared light penetrates the solid to the limited depth, the resolution deteriorates drastically as the thickness of the top die increases. In addition, the metallic piece of die holder blocks images in the infrared microscope. Another alignment sensor is desired to monitor the deviation from the perfect alignment using the electrical signal. The misalign information is, then, fed back to the aligner to compensate the misalignment. The aligner shall be capable of recovering the translational misalignment along the two principle axes parallel to the die surface, and rotational misalignment perpendicular to the die surface. The signal generator provides the source signal to be supplied to the die through the transmitter. The receiver collects the transmitted signal that is distorted by the amount of misalignment as depicted by a curve in Figure 2 below.
Figure 2 - Misalignment sensing and compensation by aligner
In order to convert the physical misalignment to the electrical signal during the die bonding step, the alignment keys shall sense the alignment when they are not in contact. The electrical or magnetic coupling is an efficient medium of alignment information. Figure 3 illustrates a possible example of alignment key deployment in the stacked two dies. The intensity of the received signal becomes strong when active alignment keys are placed on the facing surfaces of the two stacked dies, i.e. bottom of the upper die and top of the lower die. Both transmitter and receiver are located on the bottom die and the upper die does not have any active device so that the upper die does not need to have electricity. The power is provided through the fixed bottom die and the top floated upper die provides passive bridges. Then, the upper die is free to move for alignment recovery and bonding. The signal is emitted by the alignment key connected to the transmitter on bottom die, and it is coupled by left part of the bridge on bottom surface of the upper die. The signal travels to the right half of the bridge and couples back to the transponder that is connected to the receiver on the top surface of the bottom die. The attenuation of the received signal from the transmitted one is determined by the distance and misalignment of align keys. If the upper die shakes constantly, the amount of attenuation tells which direction is for the perfect alignment. The bridge on the upper die is exposed to the ambient, but the transponder on bottom die is covered with thin dielectric film to avoid direct contact between the alignment keys and to maximize the received signal as well. Clauses A.1 and A.2 show an example of the shape of a typical sensor element and the strength of coupling to misalignment.
Figure 3 - Adjustment for translational misalignment
4.3 Alignment measurement after die stacking
The alignment of the two dies may be disrupted by the mechanical or thermal agitation during the die bonding process. The typical cross-section view of the stacked dies with misalignment in the vertical interconnects is shown in Figure 4. After the upper die is completely bonded onto the lower die mechanically and electrically, the alignment is once again measured by an appropriate instrument. The quality of the final alignment of the three dimensional integrated circuits is delivered in the form of amount of misalignment. An example of a structure for detecting misalignment after bonding is shown in Annex A.3.
Figure 4 - Final alignment of vertical interconnects between the adjacent layers of dies
5 Alignment procedure
5.1 Initial die stacking
Place the bottom die and store the image of alignment key in the memory before the top die is brought above the bottom one. Then, compare the image of alignment key on the top die and that in the memory to make the initial alignment. The accuracy of the alignment shall be better than a half of spacing of alignment unit in the coupling alignment keys.
5.2 Final alignment
Turn on the electrical alignment key to track the alignment with electrical signal during the subsequent bonding process. Shake the upper die in X direction to get the best alignment along that direction using one of coupling alignment methods. And repeat the alignment check along Y direction. Rotate the upper die around the axis perpendicular to the die surface to compensate the rotational misalignment. Repeat translational and rotational alignment until the misalignment is small enough.
5.3 Assessment of alignment
Whenever a layer of die is stacked, the quality of final alignment shall be measured using an appropriate method, e.g. resistance measurement as described in IEC 63011-3.
