Machine learning and integrated metrology for run-to-run optimization of chip-to-wafer alignment accuracy

Abstract

A method, apparatus, and system in an integrated bonding system for optimizing bond alignment between a die and a substrate includes bonding a first die to a first substrate using a preset alignment setting using a bonder of the integrated bonding system; transporting the bonded die-substrate combination to an on-board inspection tool of the integrated bonding system using a transport arm / robot of the integrated bonding system; inspecting alignment of bonds between the die and substrate of the bonded die-substrate combination at the on-board inspection tool to determine misalignment measurements indicative of misalignment of bonds between the die and substrate of the bonded die-substrate combination; determining correction measurements from the misalignment measurements using a machine learning process that are communicated to the bonder; and bonding a different die to a different substrate at the bonder using the determined machine learning based correction measurements.

Description

[Technical field]

[0001]

[0001] Embodiments of the present disclosure generally relate to methods, apparatus, and systems for processing substrates. In particular, embodiments of the present disclosure relate to methods, apparatus, and systems for improving chip-to-wafer bonding alignment accuracy. [Background technology]

[0002]

[0002] Accurate chip-to-wafer (C2W) alignment is crucial to ensure electrical connectivity. Currently, alignment optimization typically relies on post-bonding misalignment measurements in a standalone tool and manual input of compensation offsets into the bonder. This process is time-consuming, prone to human error, and has limited optimization rounds. Furthermore, checking the alignment performance of the bonder is typically performed only after part replacement, but not on a timely run-to-run (R2R) basis. Summary of the Invention

[0003]

[0003] A method in an integrated bonding system for optimizing bonding alignment between a die and a substrate includes bonding a first die to a first substrate using a preset alignment setting using a bonder of the integrated bonding system, transporting the bonded die-substrate combination to an on-board inspection tool of the integrated bonding system using a transport arm / robot of the integrated bonding system, inspecting alignment of the bonds between the die and substrate of the bonded die-substrate combination in the on-board inspection tool to determine misalignment measurements indicative of misalignment of the bonds between the die and substrate of the bonded die-substrate combination, determining correction measurements from the misalignment measurements using a machine learning process which are communicated to the bonder, and bonding a different die to a different substrate in the bonder using the determined machine learning based correction measurements.

[0004]

[0004] In some embodiments, the method may further include comparing the determined misalignment measurement to a threshold value to determine whether the determined misalignment of the bond between the die and the substrate of the bonded die-substrate combination is within an acceptable error range. If the determined misalignment is determined not to be within an acceptable error range, a correction measurement may be determined, if the determined misalignment is determined to be within an acceptable error range, and if the determined misalignment is determined to be worsening, a correction measurement may also be determined. Different dies may then be bonded to different substrates using the determined machine learning based correction measurement.

[0005]

[0005] An apparatus in an integrated bonding system for optimizing bonding alignment between a die and a substrate includes a processor and a memory coupled to the processor, the memory having at least one of a program or instructions stored therein, which, when executed by the processor, configures the apparatus to: cause a bonder of the integrated bonding system to bond a first die to a first substrate using a preset alignment setting; a transport arm / robot of the integrated bonding system to transport the bonded die-substrate combination to an on-board inspection tool of the integrated bonding system; the on-board inspection tool to inspect the alignment of the bond between the die and the substrate of the bonded die-substrate combination to determine misalignment measurements representative of misalignment of the bonds between the die and the substrate of the bonded die-substrate combination; determine, using a machine learning process, correction measurements from the misalignment measurements that are communicated to the bonder; and the bonder to bond a different die to a different substrate using the determined machine learning based correction measurements.

[0006]

[0006] In some embodiments, the apparatus further compares the determined misalignment measurement to a threshold value to determine whether the determined misalignment of the bond between the die and the substrate of the bonded die-substrate combination is within an acceptable error range. If the determined misalignment is determined not to be within an acceptable error range, a correction measurement is determined. If the determined misalignment is determined to be within an acceptable error range and the determined misalignment is determined to be worsening, a machine learning based correction measurement may be determined. The bonder may use the determined machine learning based correction measurement to bond different dies to different substrates.

[0007]

[0007] An integrated bonding system for optimizing bonding alignment between a die and a substrate includes a bonder that bonds a first die to a first substrate using preset alignment settings, a transport arm / robot that transports the bonded die-substrate combination to an on-board inspection tool of the integrated bonding system, and an on-board inspection tool that inspects the alignment of the bonds between the die and substrate of the bonded die-substrate combination to determine misalignment measurements representative of the misalignment of the bonds between the die and substrate of the bonded die-substrate combination and uses a machine learning process to determine correction measurements from the misalignment measurements which are communicated to the bonder, and the bonder bonds different die to different substrates using the determined machine learning based correction measurements.

[0008]

[0008] In some embodiments, the on-board detection tool further compares the determined misalignment measurement to a threshold value to determine whether the determined misalignment of the bond between the die and the substrate of the bonded die-substrate combination is within an acceptable error range. If the determined misalignment is determined not to be within an acceptable error range, a correction measurement is determined. If the determined misalignment is determined to be within an acceptable error range and the determined misalignment is determined to be worsening, a machine learning based correction measurement may be determined. The bonder may use the determined machine learning based correction measurement to bond different dies to different substrates.

[0009]

[0009] Other and further embodiments of the present disclosure are described below.

[0010]

[0010] The embodiments of the present disclosure, briefly summarized above and described in more detail below, can be understood by reference to the exemplary embodiments of the present disclosure illustrated in the accompanying drawings. However, since the present disclosure may admit of other equally effective embodiments, the accompanying drawings depict only typical embodiments of the present disclosure and therefore should not be considered as limiting the scope. [Brief description of the drawings]

[0011] [Figure 1]

[0011] A functional flow diagram of a typical chip-to-wafer (C2W) bonding process is shown. [Diagram 2]

[0012] 2 shows a high level block diagram of an integrated bonding system 200, in accordance with an embodiment of the present principles. [Diagram 3]

[0013] 1 shows a flow diagram of a method in an integrated bonding system of the present principles for die-to-substrate / wafer bonding including machine learning based alignment correction, in accordance with an embodiment of the present principles. [Figure 4]

[0014] FIG. 4 shows a functional diagram of one implementation of the method of FIG. 3 for die-to-substrate / wafer bonding including machine learning based alignment correction in a scenario of a single die type on a substrate where misalignment is detected, in accordance with an embodiment of the present principles. [Diagram 5]

[0015] 1 shows a flow diagram of an alternative method in an integrated bonding system of the present principles for bonding a die to a substrate / wafer including machine learning based alignment correction and including a threshold comparison procedure, in accordance with an embodiment of the present principles. [Figure 6A]

[0016] 1 shows a flow diagram of a first portion of an alternative method in an integrated bonding system of the present principles for a multi-die in a substrate bonding process including machine learning based alignment correction, in accordance with an embodiment of the present principles. [Figure 6B]

[0017] 1 shows a flow diagram of a second portion of an alternative method in an integrated bonding system of the present principles for a multi-die in a substrate bonding process including machine learning based alignment correction, in accordance with an embodiment of the present principles. [Figure 7]

[0018] FIG. 7 shows a functional diagram of one implementation of the method of FIG. 6 for die-to-substrate / wafer bonding including machine learning based alignment correction in a multi-die-type scenario on a substrate where misalignment is detected, according to an embodiment of the present principles. [Figure 8]

[0019] 1 depicts a high level block diagram of a controller / computing device suitable for use with multiple embodiments of an integrated bonding system in accordance with present principles. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012]

[0020] For ease of understanding, the same reference numbers have been used, where possible, to designate identical elements common to the figures. The figures are not to scale and may be simplified for clarity. Elements and features of one embodiment may be beneficially incorporated in other embodiments without further recitation.

[0013]

[0021] Multiple embodiments of methods, apparatus, and systems for improving the accuracy of chip-to-wafer alignment are provided herein. For example, methods, apparatus, and systems for improving the accuracy of chip-to-wafer alignment may include a bonder having an on-board inspection tool to enable in-line measurement of any misalignment between the die and the substrate(s). Measurements from the on-board inspection tool may be communicated to a machine learning process / database. The machine learning process / database provides feedback to the bonder to correct any measured misalignment between the die(s) and the substrate(s) based on the feedback. In some embodiments, measurements from the on-board inspection tool may be made continuously. Alternatively, or in addition, measurements from the on-board inspection tool may be made at predetermined or dynamically determined intervals.

[0014]

[0022] FIG 1 shows a functional flow diagram 100 of a typical chip-to-wafer (C2W) bonding process. In FIG 1, the flow diagram 100 is divided into a die side 110 and a substrate side 150. The die side 110 of FIG 1 illustratively includes upstream processes 112, which may include patterning, chemical mechanical polishing / planarization (CMP), back grinding, dicing, etc. The die side 110 of FIG 1 further includes a die separation process 114, a wet cleaning process 116, a degassing process 118, a plasma process 120, a UV process 122, a die ejection and picking process 124, and a die flip process 126 for the bonding process.

[0015]

[0023] In the flow diagram 100 of FIG. 1, the substrate on the substrate side 150 may include a wafer with or without Si or a glass-supported Si wafer depending on the process flow and use case. In FIG. 1, the substrate side 150 illustratively includes an upstream process 152 that may include a CMP process. The substrate side 150 of FIG. 1 further includes a plasma process 154 and a bonding die process / tool ​​156 to generate a bonded wafer / die. In some embodiments of the present principles, the substrate side 150 may further include an optional wet cleaning process (not shown) and an optional degassing process (not shown). As shown in FIG. 1, after the bonding die process 156, the bonded wafer / die is transferred to another post-bonding misalignment measurement process on a standalone misalignment measurement tool 160. Any measured misalignment between the wafer and die measured by the standalone misalignment measurement tool 160 may be manually input as a compensation offset to the bonding die process / tool ​​156 to attempt to correct the measured misalignment. It should be noted that the exemplary processes illustrated in Figure 1 are not intended to be an exhaustive list or order of processes that may be included in a typical chip-to-wafer (C2W) bonding process, which may include more or fewer processes in the same or different order.

[0016]

[0024] FIG. 2 shows a high level block diagram of an integrated bonding system 200 according to an embodiment of the present principles. The integrated bonding system 100 of FIG. 2 includes a transport arm / robot (not shown) in a body 201, an integrated bonder 202, and an on-board misalignment measurement tool (illustratively an on-board metrology tool) 204. As shown in FIG. 2, at least one die (illustratively a plurality of separated dies) 206 and at least one substrate / wafer 208 (illustratively a single substrate / wafer with an adhesive organic tape) are loaded into the integrated bonding system 200 of FIG. 2. The die 206 is bonded to the substrate / wafer 208 in the bonder 202, and then the alignment between the die(s) 206 and the substrate / wafer 208 is measured in the on-board metrology tool 204 to determine any misalignment between the die(s) 206 and the substrate / wafer 208. In some embodiments of the present principles, the die(s) 206 may be bonded to the substrate / wafer 208 one at a time. The alignment may then be measured according to the present principles. In alternative embodiments of the present principles, the die(s) 206 may be bonded to the substrate / wafer 208 two or more at a time. The alignment may then be measured according to the present principles.

[0017]

[0025] As shown in Fig. 2, the integrated bonding system 200 of Fig. 2 may further include a number of other optional pre-treatment chambers, illustratively including a degassing chamber 218 for reducing moisture and other volatiles from, for example, adhesive tape, a plasma chamber 220 for cleaning residues, surface activation, and oxide reduction, a UV release chamber 222 for releasing die from dicing tape for die picking, a wet clean chamber 216 for cleaning the front and back sides of the substrate / wafer 208 and hydrating the substrate / wafer 208, and a factory interface and single or multiple atmospheric mainframe (FI and AMM) integration chambers 209 for superior Q-time control and environmental control, such as particle, light, and moisture control. The integrated bonding system 200 of Fig. 2 may further include a controller / computing device 800 for executing the methods and processes of the present principles (described in more detail below). In accordance with embodiments of the present principles, the controller / computing device 800 may be a local device or a remote server (such as a field service server) connected to the tool via means such as Ethernet. For example, in some embodiments of the present principles, the controller / computing device 800 may be associated with an on-board misalignment measurement tool of the present principles, such as on-board metrology tool 204 of FIG. 2. In particular, in some embodiments, the controller / computing device 800 may be in communication with at least one of a bonder and / or an on-board misalignment measurement tool of the present principles. In some embodiments, the controller / computing device 800 may be incorporated within at least one of a bonder and / or an on-board misalignment measurement tool of the present principles.

[0018]

[0026] Although a particular number and types of optional chambers are shown for integrated bonding system 200 of FIG. 2, other embodiments of the integrated bonding system of the present principles may include more or less different types of optional chambers in the same or different sequences necessary to complete the bonding process.

[0019]

[0027] As described above with respect to FIG. 2, an integrated bonding system of the present principles, such as integrated bonding system 200 of FIG. 2, may include a controller / computing device 800, taking the form of a computing device (described in more detail below), for performing the methods and computing functions of the present principles, such as, for example, machine learning processes. In various embodiments described in this disclosure, the controller / computing device 800 of the present principles is specifically described as a computing device. For example, in some embodiments, the on-board metrology tool 204 may implement the computing device 800 to perform the machine learning process and provide feedback information / data to the bonder 202 of the integrated bonding system 200 of FIG. 2 to correct any misalignment between the die and the substrate / wafer measured by the on-board metrology tool 204. In accordance with the present principles, data associated with the computing device (e.g., machine learning data, measurement data) may be stored in the computing device (described in more detail with respect to FIG. 8) or any other memory accessible to the computing device.

[0020]

[0028] Embodiments of the present principles provide a method in an integrated bonding system for optimizing bonding alignment between at least one die and at least one substrate, the method including at least: bonding a first die to a first substrate using a preset alignment setting using a bonder of the integrated bonding system, transporting the bonded die-substrate combination to an on-board inspection tool of the integrated bonding system using a transport arm / robot of the integrated bonding system, inspecting alignment of bonds between the die and substrate of the bonded die-substrate combination at the on-board inspection tool to determine misalignment measurements representative of misalignment of bonds between the die and substrate of the bonded die-substrate combination, determining correction measurements from the misalignment measurements using a machine learning process that are communicated to the bonder, and bonding a different die to a different substrate at the bonder using the determined machine learning based correction measurements.

[0021]

[0029] For example, in some embodiments of the present principles, a machine learning process is used to train a model that represents acceptable (e.g., within a tolerance) alignment between the die(s) and the substrate(s). That is, during training, the machine learning process may be exposed to examples of acceptably aligned die and substrates, and unacceptably aligned die and substrates. During training, the machine learning process may learn to identify bonded die and substrates that have acceptable alignment, in some embodiments using a loss function. The machine learning process of the present principles may further train, for example, on the amount of adjustments to be made to the bonder to have the bonder improve the alignment between the bonds between the die and substrate in a more accurate manner (i.e., without over-compensating or under-compensating). Subsequently, during bonding of the die and substrate, the neural network may be used in the machine learning process of the present principles to identify acceptable (i.e., within a tolerance) and unacceptable (i.e., outside of a tolerance) bonds between the die and substrate.

[0022]

[0030]

[0023] Figure 3 illustrates a flow diagram of a method 300 in an integrated bonding system of the present principles, such as integrated bonding system 200 of Figure 2, for die-to-substrate / wafer bonding including machine learning based alignment correction, in accordance with an embodiment of the present principles. Method 300 begins at 302, during which a bonder, such as bonder 202 of integrated bonding system 200 of Figure 2, bonds a first die to a first substrate with preset / historical alignment settings of bonder 202. Method 300 may proceed to 304.

[0023]

[0031] At 304, an integrated bonding system of the present principles, such as integrated bonding system 200 of Figure 2, transfers the bonded die / substrate, for example using a transfer arm / robot of integrated bonding system 200, to an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204 of integrated bonding system 200 of Figure 2. Method 300 may proceed to 306.

[0024]

[0032] At 306, an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204, inspects the alignment of the bond between the die and the substrate to determine whether there is a misalignment (or, in some embodiments described below, an unacceptable misalignment) between the die and the substrate to which the die is bonded. If there is a misalignment, the on-board metrology tool 204 measures the misalignment, the bonded first die / first substrate is discarded, and method 300 may proceed to 308. If there is no misalignment, the bonded first die / first substrate is saved, and method 300 may proceed to 312.

[0025]

[0033] At 308, a machine learning process of a computing device of an integrated inspection tool of the present principles, such as a machine learning process implemented by computing device 800 of on-board metrology tool 204, determines a correction signal / measurement from the misalignment measurement to be communicated to a bonder of the present principles, such as bonder 202. The correction signal / measurement allows, for example, the bonder to correct the misalignment between the first die and the first substrate. In some embodiments, the machine learning process of computing device 800 takes into account previous and current measurements to determine a correction signal / measurement to be communicated to bonder 202 to precisely adjust any measured misalignment between the first die and the first substrate, e.g., without over- or under-correcting for greater accuracy. In some embodiments, the collected and determined data and the machine learning data may be sent to an associated machine learning database to be stored for future use. That is, in accordance with the present principles, as described above, data associated with a computing device (e.g., machine learning data) may be stored in a memory of the computing device (described in more detail in connection with FIG. 8) or any other memory accessible to the computing device. The method 300 may proceed to 310.

[0026]

[0034] At 310, the determined correction signals / measurements are communicated to a bonder of the present principles, such as bonder 202 of integrated bonding system 200 of Figure 2. The method 300 may proceed to 312.

[0027]

[0035] At 312, an integrated bonding system of the present principles, such as integrated bonding system 200 of FIG. 2, determines whether there are any other die to be bonded to any substrate. For example, in some embodiments, bonder 202 may determine whether there are any other die to be bonded to any substrate. If it is determined that there are any additional die to be bonded to any substrate, the method may proceed to 314 for at least one of the remaining die. If it is determined that there are no die to be bonded to any substrate, method 300 may end.

[0028]

[0036] At 314, a bonder of the present principles, such as bonder 202 of integrated bonding system 200 of FIG. 2, compensates for the measured misalignment based on the machine learning determined correction signal / measurement when bonding a second die to a second substrate. That is, in accordance with the present principles, a different die is bonded to a different substrate using the new machine learning based alignment settings. Method 300 then returns to 304, where the bonded die / substrate is transferred to on-board metrology tool 204, e.g., using a transfer arm / robot, and method 300 may proceed as before. Method 300 of FIG. 3 may end when there are no more dies to be bonded to any additional substrates.

[0029]

[0037] FIG. 4 shows a functional diagram 400 of one implementation of the method 300 of FIG. 3 for die-to-substrate / wafer bonding including machine learning based alignment correction in a scenario of a single die type on a substrate where misalignment is detected, in accordance with an embodiment of the present principles.

[0030]

[0038] At 402, the bonder bonds a first die(s) of type A to a first substrate S1 with preset / history alignment settings.

[0031]

[0039] At 404, the bonded die / substrate S1-A is transferred to an on-board inspection tool of the present principles, such as the on-board metrology tool 204 of the integrated bonding system 200 of FIG. 2, for example using a transfer arm / robot of the body 201 of the integrated bonding system 200 of FIG. 2.

[0032]

[0040] At 406, the on-board inspection tool of the present principles measures the misalignment between the bonded die(s) A and the substrate S1. In some embodiments, using a machine learning process in the on-board metrology tool 204, correction signals / measurements are determined to correct the misalignment between the bonded die and the substrate. The collected and determined data may be sent to an associated machine learning database to be stored for future use.

[0033]

[0041] In one embodiment of FIG. 4, 450 shows an exemplary top view of misalignment between bonded die A and substrate S1, as exemplified by the respective copper bond pads in die A and substrate S1, and 460 shows an exemplary side view of the same misaligned bond between bonded die A and substrate S1.

[0034]

[0042] At 408, the corrected signals / measurements determined by the on-board metrology tool of the present principles are communicated to the bonder.

[0035]

[0043] In one embodiment of FIG. 4, 404, 406, and 408 are considered together as an on-board alignment check and machine learning cycle, which in FIG. 4 is on-board alignment check and machine learning cycle 1.

[0036]

[0044] At 410, the bonder compensates for misalignment of the bonds between die A and substrate S1 based on the received correction signals / measurements.

[0037]

[0045] In 412, the bonder bonds a different die of the same type A to a different substrate S2 with new alignment settings based on the received correction signals / measurements.

[0038]

[0046] At 414, a second on-board alignment inspection and machine learning cycle is then performed. That is, in one embodiment of FIG. 4, after the bonder bonds a different die of the same type A to a different substrate S2 with a new alignment setting, the bonded die / substrate S2-A is transferred to the on-board inspection tool of the present principles, for example, using the transfer arm / robot of the main body 201 of the integrated bonding system 200 of FIG. 2, the on-board inspection tool of the present principles measures the misalignment between the bonded die(s) A and the substrate S2, and based on the latest misalignment data and historical data, a second correction signal / measurement is determined, and in some embodiments, a machine learning process is used to correct the misalignment between the bonded die and the substrate, and the correction signal / measurement determined by the on-board inspection tool of the present principles is communicated to the bonder.

[0039]

[0047] In one embodiment of FIG. 4, 470 illustrates a top view of misalignment between bonded die A and a second substrate S2, as exemplified by the respective copper bond pads in die A and substrate S2.

[0040]

[0048] In one embodiment of FIG. 4, at 416, the bonder compensates for the misalignment of the bond between die A and substrate S2 based on the received second correction signal / measurement.

[0041]

[0049] At 418, the bonder bonds a third die of the same type A to a different third substrate S3 with new alignment settings based on the received second correction signals / measurements.

[0042]

[0050] In some embodiments of the present principles, the misalignment measurements may be compared to at least one threshold measurement to determine whether the misalignment needs to be corrected in accordance with the present principles or whether the misalignment is within an acceptable margin of error to ensure perfect electrical contact between the bonded die and the substrate. For example, FIG. 5 illustrates a flow diagram of an alternative method 500 in an integrated bonding system of the present principles, such as integrated bonding system 200 of FIG. 2, for bonding a die to a substrate / wafer including machine learning based alignment correction and including a threshold comparison procedure, in accordance with an embodiment of the present principles. Method 500 begins at 502, during which a bonder, such as bonder 202 of integrated bonding system 200 of FIG. 2, bonds a first die to a first substrate with preset / historical alignment settings of bonder 202. Method 500 may proceed to 504.

[0043]

[0051] At 504, an integrated bonding system of the present principles, such as integrated bonding system 200 of Figure 2, transfers the bonded die / substrate, for example using a transfer arm / robot on the body 201 of integrated bonding system 200 of Figure 2, to an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204 of integrated bonding system 200 of Figure 2. Method 500 may proceed to 506.

[0044]

[0052] At 506, an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204, inspects the alignment of the bond between the first die and the first substrate to determine if there is a misalignment between the first die and the first substrate to which the first die is bonded. If there is a misalignment, the on-board metrology tool 204 measures the misalignment, the first bonded die / substrate is discarded, and method 500 proceeds to 508. If there is no misalignment, the bonded die / substrate S1-A is saved, and method 500 may skip to 514.

[0045]

[0053] At 508, the measured misalignment measurement may be compared to a threshold value representing a maximum allowable misalignment measurement. For example, in some embodiments of the present principles, the threshold value representing the maximum allowable misalignment measurement may be stored in a memory of a computing device of the present principles, such as, for example, computing device 800 of on-board metrology tool 204. That is, in accordance with the present principles, data associated with a computing device (e.g., threshold data) may be stored in a memory of the computing device (described in more detail in connection with FIG. 8) or any other memory accessible to computing device 800.

[0046]

[0054] If the misalignment measurement is greater than the threshold in 508, the method 500 may proceed to 510. If the misalignment measurement is less than the threshold, the method 500 may skip to 514. However, in some embodiments, when the misalignment measurement is less than the threshold but the amount of misalignment is trending up / worsening based on stored data of previous runs, the method 500 may take preventative measures based on machine learning / AI-enabled predictive analysis to provide updated correction offsets to the bonder so that the misalignment measurement can always be kept at a minimum. More specifically, in some embodiments, when the misalignment measurement is less than the threshold but the amount of misalignment is trending up / worsening, the method 500 may proceed to 510 instead of skipping to 514.

[0047]

[0055] At 510, a machine learning process of a computing device of an integrated inspection tool of the present principles, such as a machine learning process implemented by a computing device 800 of an on-board metrology tool 204, determines a correction signal / measurement to be communicated to the bonder 202 from the misalignment measurement (either outside a threshold tolerance or within a threshold tolerance but trending upward / worsening). The correction signal / measurement allows, for example, the bonder to correct the misalignment between the first die and the first substrate. In some embodiments, the machine learning process of the computing device 800 takes into account previous and current measurements to determine a correction signal / measurement to be communicated to the bonder 202 to precisely adjust any measured misalignment between the first die and the first substrate, for example, without over-correcting or under-correcting for greater accuracy. Method 500 may proceed to 512.

[0048]

[0056] At 512, the determined correction signals / measurements are communicated to a bonder of the present principles, such as bonder 202 of integrated bonding system 200 of FIG.

[0049]

[0057] At 514, the integrated bonding system of the present principles, such as integrated bonding system 200 of Figure 2, determines whether there are any other die to be bonded to any substrates. If it is determined that there are any additional die to be bonded to any substrates, the method may proceed to 516 for at least one of the remaining die. If it is determined that there are no die to be bonded to any substrates, method 500 may end.

[0050]

[0058] At 516, a bonder of the present principles, such as bonder 202 of integrated bonding system 200 of FIG. 2, compensates for the measured misalignment based on the machine learning determined correction signals / measurements when bonding the remaining selected die to the substrate. That is, in accordance with present principles, the remaining selected die are bonded to the substrate using the new machine learning based alignment settings. Method 500 can then return to 504, e.g., using a transfer arm / robot of body 201 of integrated bonding system 200 of FIG. 2, the bonded die / substrate is transferred to on-board metrology tool 204, and method 500 can proceed as before.

[0051]

[0059] In some embodiments of the present principles, multiple dies may be bonded to the same or different substrates in the same or different bonders of the present principles, such as bonder 202 of integrated bonding system 200 of Figure 2. For example, Figures 6A and 6B (collectively referred to herein as Figure 6) together show a flow diagram of an alternative method 600 in an integrated bonding system of the present principles, such as integrated bonding system 200 of Figure 2, for a multi-die in a substrate bonding process including machine learning based alignment correction, in accordance with an embodiment of the present principles. One embodiment of method 600 of Figure 6 is directed to the case where a first bonder is implemented to bond a first type of die to a substrate and a second bonder is implemented to bond a second type of die to the same or different substrate(s).

[0052]

[0060] 6 begins at 602 during which a first bonder bonds a first die of a first type to a first substrate with preset / history alignment settings of the first bonder. Method 600 may proceed to 604.

[0053]

[0061] At 604, an integrated bonding system of the present principles, such as integrated bonding system 200 of Figure 2, transfers the bonded first die / first substrate to an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204 of integrated bonding system 200 of Figure 2. Method 600 may proceed to 606.

[0054]

[0062] At 606, an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204, inspects the alignment of the bond between the first die and the first substrate to determine if there is a misalignment between the bonded first die / first substrate. If there is a misalignment, the on-board metrology tool 204 measures the misalignment, the bonded first die / first substrate is discarded, and method 600 proceeds to 608. If there is no misalignment, the bonded first die / first substrate is saved, and method 600 may proceed to 616 in parallel with 612.

[0055]

[0063] At 608, a machine learning process of a computing device of an integrated inspection tool of the present principles, such as a machine learning process implemented by computing device 800 of on-board metrology tool 204, determines a correction signal / measurement from the misalignment measurement to be communicated to the first bonder. The correction signal / measurement enables the bonder to correct the bonding misalignment in the bond between the first die and the first substrate. Method 600 may proceed to 610.

[0056]

[0064] At 610, the determined correction signal / measurement is communicated to the first bonder. The method 600 may proceed to 612.

[0057]

[0065] At 612, the integrated bonding system of the present principles, such as integrated bonding system 200 of FIG. 2, determines whether there are any other first type dies to be bonded to any substrate. For example, in some embodiments, the first bonder may determine whether there are any other first type dies to be bonded to any substrate. If it is determined that there are any additional first type dies to be bonded to any substrate, the method may proceed to 614. If it is determined that there are no dies to be bonded to any substrate, method 600 may end.

[0058]

[0066] At 614, the first bonder compensates for the measured bonding misalignment based on the machine learning determined correction signals / measurements when bonding the remaining selected first type die to the different selected substrate. For example, in some embodiments, the first bonder bonds the first type second die to the second substrate using the machine learning determined correction signals / measurements. Method 600 can then return to 604, where the bonded first type second die and second substrate are transferred to the on-board metrology tool 204, and method 600 can proceed as before.

[0059]

[0067] At 616, the second bonder bonds the second type die to the substrate on which the first type die was bonded with the preset / history alignment settings of the second bonder. That is, in some embodiments, detection of a misaligned first type die on the substrate bonded by the first bonder may act as a gate to bond the second type die onto the substrate in the second bonder. Method 600 may proceed to 618.

[0060]

[0068] At 618, an integrated bonding system of the present principles, such as integrated bonding system 200 of Figure 2, transfers the bonded first-second die type / substrate combination to an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204 of integrated bonding system 200 of Figure 2. Method 600 may proceed to 620.

[0061]

[0069] At 620, an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204, inspects the alignment of the bond between the second type die and the substrate on which the first type die is bonded to determine whether there is a misalignment between the bonded second type die and the substrate on which the first type die is bonded. If there is a misalignment, the misalignment is measured and the bonded first-second die type / substrate is discarded and method 600 proceeds to 622. If there is no misalignment, the bonded first-second die substrate is saved and method 600 may proceed to 626.

[0062]

[0070] At 622, a machine learning process of a computing device of an integrated inspection tool of the present principles, such as a machine learning process implemented by the computing device 800 of the on-board metrology tool 204, determines a correction signal / measurement from the misalignment measurement to be communicated to the second bonder. The correction signal / measurement enables the bonder to correct the bonding misalignment in the bond between the second die and the substrate. Method 600 may proceed to 624.

[0063]

[0071] At 624, the determined correction signal / measurement is communicated to the second bonder. The method 600 may proceed to 626.

[0064]

[0072] At 626, the integrated bonding system of the present principles, such as integrated bonding system 200 of FIG. 2, determines whether there are any other second type dies to be bonded to any substrate. For example, in some embodiments, a second bonder may determine whether there are any other second type dies to be bonded to any substrate. If it is determined that there are any additional second type dies to be bonded to any substrate, the method may proceed to 628. If it is determined that there are no dies to be bonded to any substrate, method 600 may end.

[0065]

[0073] At 628, the second bonder compensates for the measured bonding misalignment based on the machine learning determined correction signals / measurements when bonding the remaining selected second type die to the different selected substrate. For example, in some embodiments, the second bonder bonds another different second type die to the different substrate using the machine learning determined correction signals / measurements. Method 600 can then return to 618, and the bonded different second type die and the different substrate are transferred to the on-board metrology tool 204, and method 600 can proceed as before to 620, etc.

[0066]

[0074] Embodiments of the present principles for bonding multiple types of dies, such as method 600 of Fig. 6, may further include a threshold measurement procedure to determine whether misalignment needs to be corrected according to the present principles as shown in one embodiment of Fig. 5. For example, in an alternative embodiment of method 600 (herein referred to as method 600A) including a threshold comparison, the misalignment measured in step 606 may be compared to a threshold representing a maximum allowable misalignment measurement, for example, in step 607A (not shown). If the misalignment is greater than the threshold in step 607B (not shown), method 600A may proceed to 608 of method 600 described above. If the misalignment measurement is less than the threshold, method 600A may skip to 612 of method 600 described above. However, in some embodiments, when the misalignment measurement is below the threshold but the amount of misalignment is trending up / worsening based on stored data from previous runs, method 600A can take preventative action based on machine learning / AI-enabled predictive analysis to provide updated correction offsets to the bonder so that the misalignment measurement can always be kept at a minimum. More specifically, in some embodiments, when the misalignment measurement is below the threshold but the amount of misalignment is trending up / worsening, method 600A can proceed to 608 instead of skipping to 612. Method 600A can then proceed as described above in method 600.

[0067]

[0075] Additionally, while one embodiment of FIG. 6 illustrates bonding of two types of die, in some embodiments of the present principles, an integrated bonding system of the present principles, such as integrated bonding system 200 of FIG. 2, may be implemented to bond three or more types of die in accordance with the present principles.

[0068]

[0076] 7 shows a functional diagram 700 of one implementation of a method 600 for die to substrate / wafer bonding including machine learning based alignment correction in a scenario of multiple die types on a substrate where misalignment is detected, according to an embodiment of the present principles. In one example of FIG. 7, a first bonder, Bonder A, is implemented to bond a die of a first type, Type A, to a substrate S. Further, in one embodiment of FIG. 7, a second bonder, Bonder B, is implemented to bond a die of a second type, Type B, to a substrate S.

[0069]

[0077] In one embodiment of FIG. 7, in 702, a first bonder, Bonder A, bonds a first type A die to a first substrate S1 with preset / history alignment settings of the first bonder, Bonder A.

[0070]

[0078] At 704, an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204, inspects the alignment of the bond between die A and the first substrate S1 and determines that a misalignment exists between die A and the first substrate S1 to which the die is bonded. Thus, in one embodiment of Figure 7, the on-board metrology tool 204 measures the misalignment and the bonded die / substrate S1-A is discarded and there is no bonding of a second type B die at the second bonder B, preventing further loss of the die and substrate.

[0071]

[0079] In one embodiment of FIG. 7, 750 illustrates an exemplary side view of misaligned bonding between a bonded die A and a first substrate S1, as illustrated by respective copper bond pads in die A and substrate S1.

[0072]

[0080] At 706, bonder A compensates for the measured bonding misalignment based on the machine learning determined correction signals / measurements determined from the measured misalignment and bonds another type A die to a second substrate S2 using the new machine learning based alignment settings.

[0073]

[0081] In one embodiment of Figure 7, an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204, inspects 708 the alignment of the bonds between die A and the second substrate S2 and determines that there is only a slight misalignment, within an acceptable margin of error, between die A and the first substrate S1 to which the die is bonded. Thus, in one embodiment of Figure 7, on-board metrology tool 204 measures the misalignment and the bonded die / substrate S2-A is saved.

[0074]

[0082] In one embodiment of FIG. 7, 760 illustrates an exemplary side view of bonding within an acceptable tolerance between a first type A die and a substrate S2, as illustrated by respective copper bonding pads in the die A and the substrate S2.

[0075]

[0083] In one example of FIG. 7, at 710, bonder A bonds another type A die to a third substrate S3 using a new machine learning based alignment setting to attempt to offset the measured bonding misalignment and optimize the alignment of the bond between the first type die A and the substrate S3 based on the machine learning determined correction signals / measurements determined from the measured misalignment of the die A on the substrate S2. For example, in one example of FIG. 7, 770 shows an exemplary side view of the optimized bonding alignment of the first type A die to the substrate S3 as exemplified by the die A and respective copper bonding pads in the substrate S3. In one example of FIG. 7, the bonding of the first type A to the substrate S3 is optimal as shown at 770.

[0076]

[0084] At 712, the process continues with an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204, inspecting the alignment of the bonds between the first type A and the third substrate S3 to determine whether any misalignment exists between the first type A and the third substrate S3.

[0077]

[0085] In one embodiment of FIG. 7, in parallel with 712, since the alignment of the bond between die A and the second substrate S2 was within the acceptable tolerance, at 714 a second bonder B bonds a second type B die to the bonded substrate S2-A with the preset / history alignment settings.

[0078]

[0086] At 716, an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204, inspects the bond misalignment between the second type B die and the bonded substrate S2-A and determines that there is only a slight misalignment, within an acceptable tolerance, between the die B and the bonded substrate S2-A to which the second type B die is bonded. Thus, in one embodiment of Figure 7, the on-board metrology tool 204 measures the misalignment and the bonded die / substrate S2-A-B is saved.

[0079]

[0087] In one embodiment of FIG. 7, 780 illustrates an exemplary side view of bonding within an acceptable tolerance between a second type B die and substrate S2-A, as illustrated by respective copper bonding pads in die A and substrate S2-A.

[0080]

[0088] In one example of FIG. 7, at 718, a second bonder B bonds another type B die to the bonded substrate S3-A using a new machine learning based alignment setting to attempt to offset the measured bonding misalignment and optimize the alignment of the bond between the second type die B and the bonded substrate S3-A based on the machine learning determined correction signal / measurement determined from the measured misalignment of the die B on the bonded substrate S2-A. For example, in one example of FIG. 7, 790 illustrates an exemplary side view of an optimized bond between the second type B die and the substrate S3-A, as illustrated by the respective copper bond pads in the die B and the substrate S3-A. In one example of FIG. 7, the bond between the second type B die and the bonded substrate S3-A is optimal as shown at 790.

[0081]

[0089] At 720, the process continues with an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204, inspecting the alignment of the bonds between the die B and the bonded substrate S3-A to determine whether any misalignment exists between the second type B die and the bonded substrate S3-A.

[0082]

[0090] In the above described embodiments of the present principles, although embodiments are described as all die bonds on all substrates being inspected for misalignment in accordance with the present principles, the precision of how often or how many bonds are inspected for misalignment in accordance with the present principles can be varied. For example, in some embodiments, the precision of how often or how many bonds are inspected for misalignment can be based on the amount of time available to perform the bonding process and / or the amount of processing / memory capabilities of the integrated bonding system of the present principles. In some embodiments, the precision of how often or how many bonds are inspected for misalignment can be based on predictions of a machine learning / artificial intelligence process of the present principles. This machine learning / artificial intelligence process can predict when a changing misalignment measurement may be out of tolerance based on trends in previous and current misalignment measurements.

[0083]

[0091] In some embodiments, the inspection accuracy may be dynamic. For example, in some embodiments, the sampling rate may depend on the maturity and stability of the bonding process. That is, the initial sampling rate may be very high, up to 100% (all dies are inspected for all bonded substrates). Thereafter, during ramp-up, for example, the sampling rate may be 20-80%. Then, during mass production of a stable, high-yield process, the sampling rate may be less than 10%. The above-mentioned inspection accuracy is merely exemplary, and the frequency of bonding inspection may include virtually any frequency in accordance with the present principles.

[0084]

[0092] According to the present principles, the die / substrate may have multiple measurement points. For example, in some embodiments, for thin dies (thickness less than 200um), measurements may be taken at 4 points, such as at the corners of the die, or even 5 points (corners + center) to capture potential misalignment caused by die warpage. For other embodiments and for thicker dies (thickness greater than 200um), measurements may be taken at 4 points, but if the Cu bonding pads are large (size greater than 10um) and the specifications / tolerances are not tight, it may be possible to reduce the measurements to 2 points at the corners of the die for quick check during mass production.

[0085]

[0093] In some embodiments of the present principles, as described above, some alignment measurements can include five measurements sampled per die (four corners and one in the center), and there can be as many as 500 dies per wafer. Thus, the scale of the control system can be as large as 2500 control loops (5 x 500). Since there is no interaction between each control loop, they can be treated as 2500 single-input, single-output control loops, which we call MSISO (Multiple Single-Input and Single-Output). We propose an MSISO control system designed on top of an exponentially weighted moving average (EWMA) model. The MSISO control system of the present principles allows multiple SISO parameters to be configured in the same context. That is, the MSISO control model design of the present principles simplifies the configuration of multiple SISO parameters by reducing the strategy logic of multiple parameters to a single step. The MSISO control system of the present principles supports a control system with a large number of inputs and outputs, such as 1000 inputs and 1000 outputs. The advantage of the MSISO design of the present principles is that all 2500 control loops can be calculated in one go, instead of looping through each control loop one by one (2500 loops). The MSISO method of the present principles significantly improves the calculation time (or reduces the calculation load) compared to currently available methods that treat the control loops individually.

[0086]

[0094] In some embodiments of the present principles, alignment measurements taken by an on-board integrated inspection tool of the present principles, such as on-board metrology tool 204 of integrated bonding system 200 of FIG. 2, may be weighted such that less reliable measurements (i.e., measurements taken during less favorable conditions) are weighted less highly than more reliable measurements (i.e., measurements taken during better conditions). For example, in some embodiments, the current last data point that is deemed unstable may be weighted 30% and the more reliable measurements from the past may be weighted 70%. The weighting may be a tuning parameter that can be adjusted by the user.

[0087]

[0095] Embodiments of the present principles may further implement a moving horizon technology that may be applied to each control loop. For example, a tuning control loop to determine the alignment of a bonded die to a substrate may be based on the last 10 data points (or a moving window of 10). In one example shown, the horizon length (e.g., 10) is a tuning parameter that may be adjusted by the user.

[0088]

[0096] Figure 8 illustrates a high level block diagram of a controller / computing device 800 suitable for use with embodiments of an integrated bonding system in accordance with the present principles, such as integrated bonding system 200 of Figure 2. In some embodiments, computing device 800 may be configured to implement methods of the present principles as processor-executable program instructions 822 (e.g., program instructions executable by processor(s) 810) in various embodiments.

[0089]

[0097] In one embodiment of FIG. 8, computing device 800 includes one or more processors 810a-810n coupled to system memory 820 via an input / output (I / O) interface 830. Computing device 800 further includes a network interface 840 coupled to input / output (I / O) interface 830, and one or more input / output devices 850, such as a cursor control device 860, a keyboard 870, and one or more display(s) 880. In various embodiments, a user interface may be generated and displayed on display 880. In some cases, multiple embodiments are implemented using a single instance of computing device 800, while in other embodiments, multiple such systems, or multiple nodes comprising computing device 800, may be configured to host various portions or instances of various embodiments. For example, in one embodiment, some elements may be implemented via one or more nodes of computing device 800, which are distinct from the nodes implementing other elements. In another example, multiple nodes may implement computing device 800 in a distributed manner.

[0090]

[0098] In various embodiments, computing device 800 may be any of a variety of types of devices, including, but not limited to, a personal computer system, a desktop computer, a laptop, notebook, tablet or netbook computer, a mainframe computer system, a handheld computer, a workstation, a network computer, a camera, a set-top box, a mobile device, a consumer device, a video game console, a handheld video game device, an application server, a storage device, a peripheral device such as a switch, modem, router, or generally any type of computing or electronic device.

[0091]

[0099] In various embodiments, computing device 800 may be a uniprocessor system including one processor 810, or a multiprocessor system including several processors 810 (e.g., two, four, eight, or another suitable number). Processor 810 may be any suitable processor capable of executing instructions. For example, in various embodiments, processor 810 may be a general-purpose or embedded processor implementing any of a variety of instruction set architectures (ISAs). In a multiprocessor system, each of processors 810 may typically implement the same ISA, but this is not necessarily the case.

[0092]

[0100] The system memory 820 may be configured to store program instructions 822 and / or data 832 accessible by the processor 810. In various embodiments, the system memory 820 may be implemented using any suitable memory technology, such as static random access memory (SRAM), synchronous dynamic RAM (SDRAM), non-volatile / flash memory, or any other type of memory. In one embodiment shown, program instructions and data implementing any of the elements of the embodiments described above may be stored within the system memory 820. In other embodiments, the program instructions and / or data may be received, transmitted, or stored on various types of computer-accessible media or similar media separate from the system memory 820 or the computing device 800.

[0093]

[0101] In one embodiment, I / O interface 830 may be configured to coordinate I / O traffic between processor 810, system memory 820, and any peripheral devices within the device, including other peripheral interfaces such as network interface 840 or input / output devices 850. In some embodiments, I / O interface 830 may perform any necessary protocol, timing, or other data conversions to convert data signals from one component (e.g., system memory 820) into a format suitable for use by another component (e.g., processor 810). In some embodiments, I / O interface 830 may include support for devices connected via various types of peripheral buses, such as, for example, variations of the Peripheral Component Interconnect (PCI) bus standard and the Universal Serial Bus (USB) standard. In some embodiments, the functionality of I / O interface 830 may be split into two or more separate components, such as, for example, a northbridge and a southbridge. Also, in some embodiments, some or all of the functionality of I / O interface 830, such as the interface to system memory 820, may be incorporated directly into processor 810.

[0094]

[0102] Network interface 840 may be configured to allow data to be exchanged between computing device 800 and other devices connected to a network (e.g., network 890), such as one or more external systems, or between nodes of computing device 800. In various embodiments, network 890 may include one or more networks, including, without limitation, a local area network (LAN) (e.g., an Ethernet or enterprise network), a wide area network (WAN) (e.g., the Internet), a wireless data network, some other electronic data network, or some combination thereof. In various embodiments, network interface 840 may support communications over any suitable type of wired or wireless general-purpose data network, such as an Ethernet network, over a digital fiber communications network, over a storage area network such as a Fibre Channel SAN, or over any other suitable type of network and / or protocol.

[0095]

[0103] The input / output devices 850, in some embodiments, may include one or more display terminals, keyboards, keypads, touchpads, scanning devices, voice or optical recognition devices, or any other devices suitable for inputting or accessing data by one or more computer systems. Multiple input / output devices 850 may be present within a computer system or distributed among various nodes of the computing device 800. In some embodiments, similar input / output devices may be separate from the computing device 800 and interact with one or more nodes of the computing device 800 via wired or wireless connections, such as via the network interface 840.

[0096]

[0104] Those skilled in the art will appreciate that computing device 800 is merely exemplary and is not intended to limit the scope of the embodiments. In particular, computer systems and devices may include any combination of hardware or software capable of performing the indicated functions of the various embodiments, including computers, network devices, Internet appliances, PDAs, wireless telephones, pagers, etc. Computing device 800 may also be connected to other devices not shown, or may alternatively operate as a stand-alone system. In addition, functionality provided by the illustrated components may be combined into fewer components or distributed among additional components in some embodiments. Similarly, in some embodiments, some functionality of the illustrated components may not be provided and / or other additional functionality may be available.

[0097]

[0105] Computing device 800 can communicate with other computing devices based on a variety of computer communication protocols, such as Wi-Fi, Bluetooth® (and / or other standards for exchanging data over short distances include protocols using short wavelength radio transmissions), USB, Ethernet, cellular, ultrasonic local area communication protocols, etc. Computing device 800 may further include a web browser.

[0098]

[0106] Although computing device 800 is illustrated as a general purpose computer, computing device 800 may be programmed to perform various specialized control functions and configured to operate as a specialized specific computer in accordance with the present principles, and embodiments may be implemented in hardware, for example as an application specific integrated circuit (ASIC). Accordingly, the process steps described herein are intended to be broadly construed as being equally performed by software, hardware, or a combination thereof.

[0099]

[0107] While the forgoing is directed to several embodiments of the present disclosure, other and further embodiments of the disclosure may be devised without departing from the basic scope thereof.

[0100]

[0108] In the above description, numerous specific details, examples, and scenarios are described to provide a more complete understanding of the present principles. However, it will be understood that embodiments of the present principles can be practiced without such specific details. Moreover, such examples and scenarios are provided for illustrative purposes and are not intended to limit the present teachings in any manner. Using the included description, one skilled in the art should be able to implement the appropriate functionality without undue experimentation.

[0101]

[0109] References herein to "one embodiment" or the like indicate that one described embodiment may include a particular feature, structure, or characteristic, but not all embodiments necessarily include the particular feature, structure, or characteristic. Such phrases do not necessarily refer to the same embodiment. Moreover, when a particular feature, structure, or characteristic is described with respect to one embodiment, it is believed to be within the knowledge of one of ordinary skill in the art to apply such feature, structure, or characteristic to multiple other embodiments, whether or not expressly stated.

Claims

1. 1. A method in an integrated bonding system for optimizing bonding alignment between a die and a substrate, comprising: bonding a first die to a first substrate using a bonder of the integrated bonding system using a preset alignment setting; transferring the bonded die-substrate combination to an on-board inspection tool of the integrated bonding system using a transfer robot of the integrated bonding system; inspecting, at the on-board inspection tool, an alignment of the bond between the die and the substrate of the bonded die-substrate combination to determine a misalignment measurement representative of a misalignment of the bond between the die and the substrate of the bonded die-substrate combination; determining, from the misalignment measurements, correction measurements that are communicated to the bonder using a machine learning process; and bonding, at the bonder, different die to different substrates using the determined machine learning based corrected measurements.

2. The method of claim 1 , further comprising determining whether there are any other die to be bonded to any substrate.

3. 10. The method of claim 1, further comprising: comparing the determined misalignment measurement to a threshold value to determine whether the determined misalignment of the bond between the die and the substrate of the bonded die-substrate combination is within an acceptable error range.

4. The method of claim 3 , wherein the threshold value represents a maximum allowable misalignment of the bond between the die and the substrate of the bonded die-substrate combination.

5. 4. The method of claim 3, wherein if the determined misalignment is determined to be not within the acceptable error range, the correction measurements are determined, and if the determined misalignment is determined to be within the acceptable error range, another die may be bonded to the first substrate.

6. The method of claim 3 , wherein the determined misalignment is determined to be within the acceptable error range and the correction measurement is determined if the determined misalignment is determined to be worsening.

7. The method of claim 5 , wherein the another die comprises at least one of a die of a different type and a die bonded to the first substrate using a different bonder of the integrated bonding system.

8. 10. The method of claim 1 , after bonding the different die to the different substrate using the determined machine learning based corrected measurements, the method of claim 1 further comprising returning to transport the bonded die-substrate combination to the on-board inspection tool of the integrated bonding system.

9. 1. An apparatus in an integrated bonding system for optimizing bonding alignment between a die and a substrate, comprising: A processor, a memory coupled to the processor, the memory storing therein at least one of a program or an instruction, the program or the instruction being executable by the processor, and when executed by the processor, causing the device to a bonder of the integrated bonding system bonding a first die to a first substrate using a preset alignment setting; a transfer robot of the integrated bonding system transferring the bonded die-substrate combination to an on-board inspection tool of the integrated bonding system; the on-board inspection tool inspecting alignment of the bonds between the die and the substrate of the bonded die-substrate combination to determine misalignment measurements representative of misalignment of bonds between the die and the substrate of the bonded die-substrate combination; determining, from the misalignment measurements, correction measurements that are communicated to the bonder using a machine learning process; and The apparatus is configured to cause the bonder to bond different dies to different substrates using the determined machine learning based corrected measurements.

10. The apparatus comprises:

10. The apparatus of claim 9, further configured to determine whether there are any other die to be bonded to any substrate.

11. The apparatus comprises:

10. The apparatus of claim 9, further configured to compare the determined misalignment measurement to a threshold value to determine whether the determined misalignment of the bond between the die and the substrate of the bonded die-substrate combination is within an acceptable error range.

12. The apparatus of claim 11 , wherein the threshold value represents a maximum allowable misalignment of the bond between the die and the substrate of the bonded die-substrate combination.

13. 12. The apparatus of claim 11, wherein if the determined misalignment is determined to not be within the acceptable error range, the correction measurement is determined, and if the determined misalignment is determined to be within the acceptable error range, another die may be bonded to the first substrate.

14. 14. The apparatus of claim 13, wherein the another die comprises at least one of a die of a different type and a die bonded to the first substrate using a different bonder of the integrated bonding system.

15. The apparatus of claim 11 , wherein the determined misalignment is determined to be within the acceptable error range and the correction measurement is determined if the determined misalignment is determined to be worsening.

16. The apparatus comprises:

10. The apparatus of claim 9, further configured to: after bonding the different die to the different substrate using the determined machine learning based correction measurements, again transport the bonded die-substrate combination to the on-board inspection tool of the integrated bonding system; again the on-board inspection tool inspecting alignment of the bonds between the die and the substrate of the bonded die-substrate combination to determine misalignment measurements representative of misalignment of the bonds between the die and the substrate of the bonded die-substrate combination; again determine correction measurements from the misalignment measurements using a machine learning process that are communicated to the bonder; and again the bonder bonding the different die to the different substrate using the determined machine learning based correction measurements.

17. 1. An integrated bonding system for optimizing bonding alignment between a die and a substrate, comprising: a bonder for bonding a first die to a first substrate using a preset alignment setting; a transfer robot that transfers the bonded die-substrate combination to an on-board inspection tool of the integrated bonding system; and an on-board inspection tool that inspects alignment of the bonds between the die and the substrate of the bonded die-substrate combination to determine misalignment measurements representative of misalignment of bonds between the die and the substrate of the bonded die-substrate combination and uses a machine learning process to determine correction measurements from the misalignment measurements that are communicated to the bonder; The bonder uses the determined machine learning based corrected measurements to bond different dies to different substrates.

18. 20. The system of claim 17, further comprising at least one of a plasma chamber for pre-treating at least one of the die and the substrate, a UV release chamber, a wet clean chamber, an integration chamber, and a degassing chamber.

19. 20. The system of claim 17, further comprising a different bonder for bonding at least one die of a different type than the first die to at least one substrate including the first substrate.

20. 20. The system of claim 17, wherein the on-board inspection tool further compares the determined misalignment measurement to a threshold value to determine if a determined misalignment of the bond between the die and the substrate of the bonded die-substrate combination is within an acceptable error range, and if the determined misalignment is determined to not be within the acceptable error range, a correction measurement is determined, if the determined misalignment is determined to be within the acceptable error range and if the determined misalignment is determined to have worsened, a correction measurement is determined, and the bonder bonds a different die to a different substrate using the determined machine learning based correction measurement.

Images