Method including positioning dummy source die or target site to compensate for overlay error

Abstract

To provide a method for adjusting the position of a source die or a target site to compensate for an overlay error between a dummy source die and a different target site, which are previously joined to each other.SOLUTION: A method includes a step 624 of joining a dummy source die to a first target site of a target substrate, wherein the dummy source die has a measurement pattern, and the first target site also has a measurement pattern. The method further includes: a step 644 of collecting radiation data regarding the measurement pattern and at least part of the measurement pattern within a radiation area; a step 646 of analyzing the radiation data to determine an overlay error between the dummy source die and the first target site; and a step 648 of adjusting the position of a known good source die or the second target site of the target substrate to compensate for the overlay error between the dummy source die and the first target site.SELECTED DRAWING: Figure 6

Description

[Technical field]

[0001] Related Applications This application is related to U.S. patent application Ser. No. 18 / 491,556, entitled "Method Including Positioning a Source Die or a Destination Site to Compensate for Overlay Error," filed on the same day by Choi, Attorney Docket No. CNT C1159-US, which is assigned to the present assignee and is hereby incorporated by reference in its entirety.

[0002] The present disclosure relates to a method that involves adjusting the position of a source die or a target feature to compensate for overlay errors between a previously bonded dummy source die and a different target feature. [Background technology]

[0003] Advanced packaging techniques require high throughput and precise placement of dies. Hybrid bonding can be particularly challenging when misalignment tolerance is small. The apparatus may include a bridge and a base, where the bond head, alignment hardware, and alignment fiducials are coupled to the bridge, and the carriage is coupled to the base. The destination substrate chuck and alignment hardware may be coupled to the carriage. During the hybrid bonding sequence, the source die is held by the bond head and the destination substrate is held by the destination substrate chuck. The amount of misalignment may be measured relative to the source die, separately from the destination substrate chuck. Although some compensation may be made for the misalignment, the lack of endpoint detection may result in larger than desired overlay error. Better overlay error measurement is needed. Summary of the Invention

[0004] The method includes bonding a first dummy source die to a first target feature of a target substrate, the first dummy source die having a first dummy source metrology pattern and the first target feature having a first target metrology pattern, the method includes collecting first radiation data for the first dummy source metrology pattern and at least a portion of the first target metrology pattern in a first radiation area, where the collecting first radiation data is performed after bonding the first dummy source die to the first target feature, and may further include analyzing the first radiation data to determine a first overlay error between the first dummy source die and the first target feature, and adjusting a position of a known good source die or a second target feature of the target substrate to compensate for the first overlay error between the first dummy source die and the first target feature.

[0005] In another embodiment, the method further comprises applying radiation to the first radiation region, wherein the first dummy source metrology pattern in the first radiation region includes a portion that is relatively transparent to the radiation and a portion that is relatively non-transparent to the radiation, wherein the relatively transparent portion of the first dummy source metrology pattern includes a substrate, and wherein the relatively non-transparent portion of the first dummy source metrology pattern includes a substrate and a dummy source alignment mark, at least one of the dummy source alignment marks being disposed along a bond surface of the first dummy source die, wherein the dummy source alignment mark is only present within the relatively non-transparent portion, and wherein a dummy source alignment mark of the first dummy source metrology pattern is not present within the relatively transparent portion.

[0006] In certain implementations, the first target metrology pattern includes a target alignment mark, and during illumination, no more than 25% of the radiation is transmitted through the target alignment mark.

[0007] In a more particular implementation, during bonding, at least one of the dummy source alignment marks does not contact any of the target alignment marks.

[0008] In another embodiment, the method further includes applying radiation to the first radiation region, the first radiation region including at least a portion of each of the first dummy source measurement pattern and the first target measurement pattern, within the first radiation region, the first dummy source measurement pattern including a portion that is relatively transparent to the radiation and a portion that is relatively opaque to the radiation, a difference in transmittance between the relatively transparent and relatively opaque portions with respect to the radiation being at least 25%.

[0009] In yet another embodiment, the method further includes irradiating the first radiation region with radiation, within the first radiation region, the first dummy source measurement pattern includes portions that are relatively transparent to the radiation and portions that are relatively opaque to the radiation, and during irradiation, no more than 25% of the radiation transmits through the relatively opaque portions and at least 75% of the radiation transmits through all of the relatively transparent portions.

[0010] In yet another embodiment, the method further comprises irradiating the first radiation area with infrared light.

[0011] In a further embodiment, the step of collecting the first radiation data is performed using a radiation and a radiation detector, the radiation detector having a focal depth of up to 9 μm for the radiation.

[0012] In another implementation, the target substrate includes a known good target die and a known bad target die, and the first target feature is at least a portion of the bad target die.

[0013] In certain implementations, the method further includes bonding the known good source die to the second target feature of the target substrate after adjusting a position of the known good source die or the second target feature, the second target feature being at least a portion of the known good target die.

[0014] In another specific implementation, the method further includes bonding a second dummy source die to a third target feature of the target substrate, the second dummy source die having a second dummy source metrology pattern and the third target feature die having a second target metrology pattern. The method further includes collecting second radiation data of at least a portion of the second dummy source metrology pattern and the second dummy source metrology pattern in a second radiation area, and analyzing the second radiation data to determine a second overlay error between the second dummy source die and the third target feature. The bonding of the second dummy source die, collecting the second radiation data, and analyzing the second radiation data are performed after adjusting a position of a known good source die or the second target feature.

[0015] In another implementation, the first dummy source metrology pattern includes a first dummy source alignment mark.

[0016] In a particular implementation, the first target metrology pattern includes a first target alignment mark.

[0017] In another specific implementation, the first target metrology pattern includes a set of contact pads that are not part of an alignment mark.

[0018] In a further embodiment, the first dummy source metrology pattern and the first target metrology pattern have areas that include corresponding arrays of contact pads.

[0019] In a specific embodiment, the array of contact pads of the first dummy source measurement pattern is phase shifted by an angle in the range of 90° to 270° relative to the array of contact pads of the first target measurement pattern.

[0020] In another specific implementation, the array of contact pads of the first dummy source metrology pattern is phase shifted relative to the array of contact pads of the first target metrology pattern by an angle in the range of 150° to 210°.

[0021] In another embodiment, the analyzing step includes analyzing the first radiation data using spatial frequency analysis.

[0022] In a specific implementation, each of the first dummy source measurement pattern, the first target measurement pattern, or the first dummy source measurement pattern and the first target measurement pattern has a pitch and a feature size corresponding to the pitch. Each of the first dummy source measurement pattern, the first target measurement pattern, or the first dummy source measurement pattern and the first target measurement pattern has a pitch and a feature size corresponding to the pitch corresponds to a spatial frequency value. A quotient is the pitch divided by the feature size corresponding to the pitch, and the value is + / - 10% of the quotient and is in the range of 1.5 to 9.0.

[0023] In a further implementation, analyzing the first radiation data includes analyzing a pattern of a first image generated from the first radiation data in comparison to a pattern of a standard image. [Brief description of the drawings]

[0024] Embodiments are illustrated by way of example and not by way of limitation in the accompanying figures.

[0025] [Figure 1] FIG. 1 is a conceptual diagram of a system that may be used to transfer a die to a target substrate.

[0026] [Diagram 2] FIG. 2 is a bottom view of a bridge of the device in the system of FIG. 1.

[0027] [Diagram 3]3 is a cross-sectional view of the device of FIG. 1 taken along section line 3-3 of FIG. 2;

[0028] [Figure 4] 4 is a cross-sectional view of the die loading machine and apparatus of FIG. 1 taken along section line 4-4 of FIG. 2;

[0029] [Diagram 5] 11 is a flow chart illustrating a process for determining an overlay error compensation corresponding to an overlay error between a dummy source die and its corresponding target feature on a target substrate. [Figure 6] 11 is a flow chart illustrating a process for determining an overlay error compensation corresponding to an overlay error between a dummy source die and its corresponding target feature on a target substrate.

[0030] [Figure 7] 4 illustrates a cross-sectional view of a portion of the apparatus after loading a source substrate onto the source substrate chuck and after loading a target substrate onto the target substrate chuck.

[0031] [Figure 8] 2A is a top view of a target substrate chuck and a target substrate including known good and bad dies;

[0032] [Figure 9] 1 is a cross-sectional view of a portion of a dummy source die.

[0033] [Figure 10] 2 is a cross-sectional view of a portion of the system of FIG. 1 while a dummy source die is being held by a die loading machine.

[0034] [Figure 11] 11 is a cross-sectional view of the base of the system of FIG. 10 after a dummy source die has been transferred to the die carrier sheet.

[0035] [Figure 12]12 is a cross-sectional view of the system of FIG. 11 when positioning the carriage using an optical element under an alignment datum coupled to the bridge.

[0036] [Figure 13] 13 is a cross-sectional view of the system of FIG. 12 after a dummy source die has been placed under the bond head.

[0037] [Figure 14] 14 is a cross-sectional view of the system of FIG. 13 after a dummy source die has been transferred from the die carrier sheet to the bonding head.

[0038] [Figure 15] 15 is a cross-sectional view of the system of FIG. 14 when measuring alignment errors of a dummy source die.

[0039] [Figure 16] 16 is a cross-sectional view of the system of FIG. 15 after a dummy source die has been placed on the target substrate.

[0040] [Figure 17] FIG. 2 is a cross-sectional view of a portion of a target die including alignment marks.

[0041] [Figure 18] 17 is a cross-sectional view of the system of FIG. 16 after bonding a dummy source die to a target site on a target substrate.

[0042] [Figure 19] 19 is a top view of the target substrate chuck and target substrate of the system of FIG. 18 after bonding a dummy source die to the target feature.

[0043] [Figure 20] 13A is a cross-sectional view of a portion of a dummy source die and a target die in a metrology pattern for the die.

[0044] [Figure 21] 20 is a cross-sectional view of the system of FIG. 19 when collecting radiation data relating to the dummy source die and the target site.

[0045] [Figure 22] 1 illustrates a portion of a dummy source substrate in a radiation region and radiation data corresponding to a portion of a dummy source die.

[0046] [Figure 23] FIG. 13 is a diagram showing a measurement pattern of a dummy source die.

[0047] [Figure 24] FIG. 13 is a diagram showing a measurement pattern of a target die.

[0048] [Diagram 25] FIG. 13 shows radiation areas and measurement patterns of a dummy source die and target features.

[0049] [Figure 26] FIG. 26 illustrates a portion of the radiation field of FIG. 25 and a measurement pattern of a dummy source die and a target portion.

[0050] [Figure 27] 13 is a flow chart for transferring a source die to a target portion of a destination substrate. [Figure 28] 13 is a flow chart for transferring a source die to a target portion of a destination substrate.

[0051] [Figure 29] 4 illustrates a cross-sectional view of the system after moving the carriage under multiple source dies coupled to a source substrate chuck.

[0052] [Diagram 30] 30 is a cross-sectional view of the system of FIG. 29 after transferring a set of source dies to the array on the die carrier sheet.

[0053] [Diagram 31] 31 is a cross-sectional view of the system of FIG. 30 when positioning the carriage using an optical element under an alignment datum coupled to the bridge.

[0054] [Diagram 32] 32 is a cross-sectional view of the system of FIG. 31 after a set of source dies has been positioned under the array of bond heads.

[0055] [Diagram 33] 33 is a cross-sectional view of the system of FIG. 32 after the set of dies has been transferred from the array on the die carrier sheet to the array on the bond head.

[0056] [Diagram 34] 34 is a cross-sectional view of the system of FIG. 33 when measuring alignment error associated with a particular source die.

[0057] [Diagram 35] 35 is a cross-sectional view of the system of FIG. 34 after bonding a set of source dies to target sites on a target substrate.

[0058] [Diagram 36] 36 is a top view of the target substrate chuck and target substrate of the system of FIG. 35 after bonding the set of source dies to their corresponding target sites.

[0059] [Figure 37] FIG. 13 is a top view showing active contact between a source die and a target portion of a target substrate.

[0060] [Figure 38] 36 is a cross-sectional view of the system of FIG. 35 while another dummy source die is being held by the die loading machine.

[0061] [Figure 39] 39 is a top view of the target substrate chuck and target substrate of the system of FIG. 38 after bonding another dummy source die to its corresponding target site.

[0062] [Diagram 40] 40 is a cross-sectional view of the system of FIG. 39 when collecting radiation data relating to another dummy source die and its corresponding target site.

[0063] Those skilled in the art will appreciate that elements in the figures are illustrated for simplicity and clarity and have not necessarily been drawn to scale For example, the dimensions of some of the elements in the figures may be exaggerated relative to other elements to help understand implementations of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0064] The following description in combination with the drawings is provided to aid in understanding the teachings disclosed herein. The following discussion focuses on specific implementations of the present teachings. This focus is provided to help explain the teachings and should not be construed as a limitation on the scope or applicability of the teachings.

[0065] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention belongs. The materials, methods, and examples are illustrative only and are not intended to be limiting. Unless explained herein, many details regarding specific materials and processing operations are conventional and can be found in textbooks and other sources in the art.

[0066] A dummy source die may be bonded to a first target feature of a target substrate. The dummy source die and the first target feature may include a metrology pattern. Radiation data relating to at least a portion of the metrology pattern is collected and analyzed to determine an overlay error between the dummy source die and the first target feature. During subsequent die bonding, a position of a known good source die or a second target feature on the target substrate may be adjusted to compensate for the overlay error between the dummy source die and the first target feature.

[0067] The dummy source die does not have any electrical circuit elements that are electrically connected to electrical circuit elements in the target substrate and, therefore, does not have interconnects or other features that may interfere when collecting radiation data after the dummy source die is bonded to the first target portion.

[0068] There are many options for analyzing the radiation data. The analysis may or may not include generating an image. When the image is generated, further refinement may or may not be used to improve the image. The imaging may be performed at a focus range where alignment marks on the dummy source die and the target die can be focused to generate an image where the electrical circuit elements and interconnects on the target die are out of focus or blurred. Pattern recognition may be used to ensure that electrical circuit elements and interconnects under contact pads in the target die are not taken into account when determining the overlay error. Additionally, spatial frequency analysis may be used during imaging or when determining the overlay error.

[0069] If necessary or desired, the metrology patterns may be designed such that alignment marks in one of the metrology patterns do not overlap with alignment marks in the other metrology patterns. Ideally, the alignment marks in one of the metrology patterns are 180° out of phase from the alignment marks in the other metrology pattern. The overlay error may correspond to how far an alignment mark in one of the metrology patterns is from being 180° out of phase from an alignment mark in the other metrology pattern.

[0070] Multiple dummy source dies can be bonded to the target substrate to provide information regarding device drift, whether there is center-to-edge variation in overlay error or other positional relationships with respect to where the source dies are bonded on the target substrate, etc.

[0071] The method may be performed by a system that may take the form of an apparatus or include a combination of an apparatus and a die loading machine or other equipment used in hybrid bonding dies. The system and method of using the system will be better understood from the following description in conjunction with the corresponding figures.

[0072] 1-4 include a conceptual diagram of a system 100 that may include an apparatus 110 and a die loading machine 180. The system 100 may be used to transfer multiple dies to a target substrate. FIG. 1 includes a high level diagram of the system 100, and FIG. 2 includes a bottom view of the bridge 120. In FIG. 2, section line 3-3 corresponds to FIG. 3, and section line 4-4 corresponds to FIG. 4. FIG. 3 includes a cross-sectional view of the apparatus 110 to show components used in transferring multiple dies to a target substrate. FIG. 4 includes a cross-sectional view of the system 100 including a radiation tool 272 in the apparatus 110 and a die loading machine 180. FIG. 1-4 show the apparatus that make up the system 100, but do not show multiple dies or a target substrate. The dies and target substrate are described with respect to process flow charts described later in this specification.

[0073] 1-3, in one implementation, the apparatus 110 includes a bridge 120, a base 140, a controller 160, and a memory 162. The controller 160 may be coupled to the bridge 120, the base 140, one or more components coupled to the bridge 120 or the base 140, or a combination thereof. Each of the bridge 120 and the base 140 may be a support structure. The bridge 120 may be coupled to a source substrate chuck 122, an array of bonding heads 124, an alignment fiducial 128, and an optical element 130. The base 140 may be coupled to a carriage 146.

[0074] In Figures 3 and 4, the bridge 120, base 140, and components physically coupled to the bridge 120 or base 140, as well as the die loading machine 180, may be organized along an XZ direction, a YZ direction, a ZZ direction, or a combination thereof. With respect to Figures 3, 4, and other cross-sectional or side views therein, the X direction is between the left and right sides of the figure, the Z direction is between the top and bottom of the figure, and the Y direction is out of the plane of the figure. Unless otherwise noted, rotation occurs along the XY plane defined by the X and Y directions.

[0075] The source substrate chuck 122 may be a vacuum chuck, a pin-type chuck, a groove-type chuck, an electrostatic chuck, an electromagnetic chuck, etc. The source substrate chuck 122 may be coupled to the bridge 120 by being directly attached to the bridge 120 or may be coupled to the bridge 120 via a carriage (not shown). The source substrate chuck 122 has a source-holding surface facing the base 140 or a component coupled to the base 140.

[0076] 2 includes a bottom view of bridge 120 to show the general location of source substrate chuck 122, region 224 containing an array of bond heads, alignment datums, and optics, as well as radiation tool 272, which is addressed later in this specification.

[0077] 3 shows an array of bond heads 124. The array of bond heads 124 may be configured as a vector (rows or columns of bond heads), or as a matrix (at least two rows and at least two columns of bond heads), or as a staggered arrangement. The number of bond heads in the array of bond heads 124 may vary between rows, columns, or between rows and columns. Some array configurations are 3x1, 6x1, 2x2, 2x3, 2x4, 4x2, 10x10, or another rectangular shape, with a first number corresponding to the number of bond heads along a row or column and a second number corresponding to the number of bond heads along the other of the rows or columns.

[0078] Each of the bond heads in the array of bond heads 124 may include a die chuck and a body disposed between the die chuck and the bridge 120. Each of the die chucks may be a vacuum chuck, a pin-type chuck, a groove-type chuck, an electrostatic chuck, an electromagnetic chuck, a Bernoulli chuck, or the like. Alternatively, the die chuck may be a non-contact chuck, where the die chuck contacts a side of the die rather than a device side or a backside. The device side is the side of the die on which the electrical components are formed, the backside of the die is opposite the device side, and the side is disposed between the device and the backside of the die. When the die includes through-substrate vias (TSVs), the TSVs may be exposed along the backside of the die. The non-contact chuck may help reduce the likelihood that an activated surface for bonding will contact the bond head. In one implementation, the device side, the backside, or both the device side and the backside may have an activated surface. The side may not be activated for bonding.

[0079] The bond heads may be configured such that the die chucks have a limited range of motion relative to their corresponding bodies to provide better positioning when dies are transferred from the array of bond heads 124 to a target substrate (not shown in FIGS. 1-4) coupled to a target substrate chuck 148. Each bond head may include a positioning stage that moves each bond head independently in one or more of the X-direction, Y-direction, Z-direction, tip, tilt, and rotation.

[0080] 3, alignment fiducials 128 are coupled to bridge 120. Alignment fiducials 128 may include marks or other features that can aid in proper positioning of carriage 146 relative to bridge 120 or components coupled to bridge 120. Alignment fiducials 128 and optical elements 150 may be used during alignment operations. Further details regarding optical elements 150 are described below with respect to components coupled to base 140.

[0081] An optical element 130 is coupled to the bridge 120 and may be used to determine the pitch of the array of die carrier sheets 144. The optical element 130 may also be used to verify the presence or identity (e.g., the part number or type of die) of a die coupled to a die carrier sheet in the array of die carrier sheets 144 or a target substrate coupled to a target substrate chuck 148. Multiple optical elements 130 may be coupled to the bridge 120 as needed or desired. The optical element 130 may also be used to determine the position of a target site on a target substrate coupled to a target substrate chuck 148.

[0082] Carriage 146 is coupled to base 140. Carriage 146 may be a positioning stage and may provide translational motion along base 140 in the X, Y, or Z directions, or rotational motion about one or more axes, such as rotation about the Z axis along a plane aligned with the X and Y directions.

[0083] The array of die carrier sheets 144 is coupled to a carriage 146. The array of die carrier sheets 144 has a body and a die chuck. The body is coupled to the carriage 146. The bridge 120 or a component coupled to the bridge 120 is closer to the die chuck than the body of a die carrier sheet in the array of die carrier sheets 144. In one implementation, any one or more of the die carrier sheets can have a die chuck, which can be any type described with respect to the array of bond heads 124. The die carrier sheet and the bond head can be the same type or different types. In one embodiment, a die carrier sheet in the array of die carrier sheets 144 can be a pick-up head.

[0084] The array of die carrier sheets 144 can be configured as a vector (rows or columns of pick-up heads), or as a matrix (at least two rows and at least two columns of pick-up heads), or as a staggered arrangement. For a matrix, the number of bonding heads in the array of die carrier sheets 144 can vary from row to row, column to column, or row and column to row. Some array configurations are 3x1, 6x1, 2x2, 2x3, 2x4, 4x2, 10x10, or another rectangular shape, with a first number corresponding to the number of pick-up heads along a row or column and a second number corresponding to the number of pick-up heads along the other of the rows or columns.

[0085] Theoretically, the die from an entire source substrate could all be transferred at once. From a top view, in such a configuration, the array of die carrier sheets 144 would have fewer pick-up heads along rows closer to the top and bottom of the array compared to the row or pair of rows closest to the center of the array, and the array of die carrier sheets 144 would have fewer pick-up heads along columns closer to the left and right sides of the array compared to the column or pair of columns closest to the center of the array. After reading this specification, one skilled in the art would be able to determine an array configuration for an array of die carrier sheets 144 that meets the needs or desires of a particular application.

[0086] The array of die transport sheets 144 may be configured to have an adjustable pitch that can be reversibly changed between a source matching pitch and a bond head matching pitch. The array of die transport sheets 144 or carriages 146 may include motors, electrical components, etc. that can be actuated to move the die transport sheets to achieve the desired pitch. In one embodiment, the array of die transport sheets 144 may be at the source matching pitch when picking up a set of dies coupled to the source substrate chuck 122 and at the bond head matching pitch when transferring the set of dies to the array of bond heads 124. After the dies are transferred to the array of bond heads 124, the pitch of the array of die transport sheets 144 may be returned to the source matching pitch before picking up more dies.

[0087] The terms "transfer operation" and "transfer cycle" are used to aid in understanding the implementations described herein. A transfer operation begins with loading a first set of dies onto the array of the die transport sheet 144 and ends with bonding of a last set of dies bonded to a target portion of a target substrate on the target substrate chuck 148. A transfer cycle begins with loading a set of dies onto the array of the die transport sheet 144 and continues until the same particular set of dies is bonded to a target portion of a target substrate coupled to the target substrate chuck 148. A transfer operation may include one or more transfer cycles.

[0088] In one implementation, the die carrier sheets in the array of die carrier sheets 144 need not be pick-up heads. The die carrier sheets in the array of die carrier sheets 144 may or may not extend in the Z direction (towards the bridge 120 or a component coupled to the bridge 120). The dies may be loaded onto the die carrier sheets in the array of die carrier sheets 144 by a die loading machine 180. Alternatively, the dies may be loaded manually by a human operator.

[0089] 3 may be coupled to carriage 146 and used during alignment operations. Optical element 150 may be part of registration and alignment hardware used in aligning carriage 146 with alignment datum 128, identifying dies held by the array of bond heads 124, positioning the bond heads, and measuring alignment errors of dies held by the array of bond heads 124. Optical element 150 may include a lens optically coupled to a mirror, a prism, a grating, a light source, a fiber optic cable, an aperture, a tube, a camera, or combinations thereof.

[0090] 3, the target substrate chuck 148 may be coupled to a carriage 146 that is coupled to the base 140. In one implementation, the target substrate chuck 148 is attached to the carriage 146. The target substrate chuck 148 may hold a target substrate including a target feature. The target substrate chuck 148 may be a vacuum chuck, a pin-type chuck, a groove-type chuck, an electrostatic chuck, an electromagnetic chuck, etc. The target substrate chuck 148 may be heated, cooled, or both heated and cooled. The target substrate chuck 148 may include a heater (not shown). In the same or a different embodiment, a fluid (not shown) may flow through the target substrate chuck 148 to increase or decrease the temperature of the target substrate chuck 148.

[0091] 1 and 4, a die loading machine 180 can be used to load the die carrier sheet of the apparatus 110. The die loading machine 180 can be used in place of or in conjunction with the source substrate chuck 122. For example, any, some, or all of the die to be bonded to the destination substrate can be coupled to a source substrate coupled to the source substrate chuck 122, and any, some, or all of the die to be bonded to the destination substrate can be placed on the die carrier sheet 144 by the die loading machine 180.

[0092] The die loading machine 180 may be a robot, a pick and place tool, etc. Referring to FIG. 4, the die loading machine 180 may include a base 482, a positioning shaft 484, a positioning arm 486, and a die transport head 488. The positioning shaft 484 may move in the X direction, the Y direction, or both the X and Y directions. The positioning shaft 484 may or may not be configured to extend or retract in the Z direction. The positioning arm 486 may be oriented along a plane defined by the X and Y directions and may rotate. The positioning arm 486 may or may not be configured to extend or retract in a direction from or towards the positioning shaft 484. The die transport head 488 may obtain a die or multiple dies from a source substrate (not shown in FIG. 4) or from another die location and transport the die or multiple dies to a plurality of die transport sheets 144 (shown in FIG. 3) in the apparatus 110.

[0093] 1-4, the system 100 may be operated using a controller 160 in communication with the bridge 120, any components coupled to the bridge 120, the base 140, any components coupled to the base 140, the die loading machine 180, or any combination thereof. The controller 160 may be operated using a computer readable program optionally stored in a memory 162. The controller 160 may include a processor (e.g., a central processing unit of a microprocessor or microcontroller), a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), or the like. The controller 160 may be within the device 110. In another implementation (not shown), the controller 160 may be at least a part of a computer external to the device 110, such computer being bidirectionally coupled to the system 100. The memory 162 may include a non-transitory computer readable medium including instructions for performing operations associated with or between transfer operations. The memory 162 may include data used or obtained when performing the method. In another implementation, bridge 120, components coupled to bridge 120, base 140, components coupled to base 140, or die loading machine 180 may include a local controller that provides some of the functionality provided by controller 160.

[0094] A method of using the system 100 will be described. Figures 5 and 6 include a process flow chart of the method described with respect to Figures 7-26, and Figures 27 and 28 include a process flow chart of the method described with respect to Figures 29-37. In practice, Figures 5 and 6 are used to determine an overlay error, and Figures 27 and 28 include a method of bonding a source die to a target feature, including adjusting the position of the source die, the target feature, or both, to compensate for the overlay error, as described with respect to Figures 5 and 6. Some parts of the system 100 shown in Figures 1-4, such as the controller 160 and the memory 162, are not shown in the figures. Figures 7-26 and Figures 29-38 are intended to simplify the understanding of the concepts described herein.

[0095] The method may include mounting the destination substrate on the destination substrate chuck at block 522 and mounting the source substrate on the source substrate chuck at block 524, as shown in FIG. 5. Referring to FIG. 7, the carriage 146 may be moved to allow easier access to the source substrate chuck 122 and the destination substrate chuck 148. The operations at blocks 522 and 524 may be performed in either order. Furthermore, mounting of the source substrate may be performed as late as possible after collecting radiation data at block 644 of FIG. 6. After reading this entire specification, one skilled in the art may determine when to mount the source substrate for the needs or desires of a particular application.

[0096] 7, a target substrate 748 is mounted on the target substrate chuck 148. The target substrate 748 may include a semiconductor wafer, a package substrate, a printed wiring board, a circuit board, an interposer, etc. A microelectronic device may be part of the target substrate 748, such as a semiconductor wafer. The package substrate, printed wiring board, circuit board, or interposer may or may not have a die mounted thereon. Some or all of the sides of the target substrate 748 may be activated for hybrid bonding. In one implementation, the activated surfaces for hybrid bonding are shown as dark bands along the exposed surface of the target substrate 748.

[0097] 8 includes a top view of the target substrate chuck 148 and the target substrate 748. The target substrate 748 includes a target die 820. Any or all of the target die 820 may include a microprocessor, a microcontroller, a graphics processing unit, a digital signal processor, a memory die (e.g., level 2 or level 3 cache, flash memory, etc.), a field programmable gate array (FPGA), a power transistor die, a power circuit die, a charge coupled device (CCD), an image sensor, a semiconductor circuit element, a die bond location on the target substrate, etc.

[0098] Each of the target die 820 can be a known good die (KGD) 824 or a bad die 834. A KGD passes one or more tests and is acceptable for bonding to another die, whether it is a target die, a source die, or another die. A bad die is one that has failed at least one test. The tests can be electrical opens, electrical shorts, other functional tests (e.g., processing operations not performed (missing implants, insulating layers not formed, missing etch operations, etc.), processing operations inadvertently performed one or more times, etc.), or the die is operating too slowly for a particular application (e.g., operating frequency is too low, read access time is too long, etc.).

[0099] 1-4 and 8, information regarding the KGD 824 and the bad die 834, and their locations relative to the target substrate 748, may be stored in memory 162 or another data storage unit (e.g., hard disk, database, etc.) external to the system 100 for later use when determining where the source die should be bonded to the target substrate 748. In one implementation, all of the target die 820, whether KGD 824 or the bad die 834, may have electrical circuit elements and electrical circuits including multiple electrical circuit elements. As used herein, an electrical circuit element may include a transistor, a capacitor, a resistor, or an inductor, and does not include a conductive structure per se (such as a contact pad, an interconnect, or a conductive plug between different levels of interconnects).

[0100] Referring to FIG. 7, a source substrate 722 can be mounted along the source substrate chuck 122. A plurality of dies 724 can be mounted on the source substrate 722. The plurality of dies 724 can include a plurality of production source dies. As used herein, a production source die is a die formed during a microelectronics manufacturing process and includes an electrical circuit element. A production source die can include a single electrical circuit element or can include an electrical circuit including multiple electrical circuit elements. A production source die can be a KGD or failed die. All or only a portion (but not all) of the plurality of dies 724 are transferred to a destination substrate 748. The source substrate 722 can be an adhesive tape that can be in the form of a tape frame or tape reel, a container having a grid that defines a matrix of areas that can hold the plurality of dies 724, or the like.

[0101] Each die in the plurality of dies 724 may include a plurality of contact pads (such as contact structures or conductive interconnects). The plurality of contact pads may be arranged on a rectangular, square, diagonal, or hexagonal grid, with each grid point within the area of ​​the die including a contact pad. The interface of the target substrate 748 may have an identical set of contact pads to match the contact pads on the plurality of dies 724. The arrangement of the contact pads on each die may have a spatial frequency characterized by its periodicity in two directions and the relative phase of the periodic arrangement in the two directions. The phase may be relative to a fixed point on the die, such as a corner of the die.

[0102] Any or all of the dies in the plurality of dies 724 may include a microprocessor, a microcontroller, a graphics processing unit, a digital signal processor, a memory die (e.g., level 2 or level 3 cache, flash memory, etc.), a field programmable gate array (FPGA), a power transistor die, a power circuit die, a charge coupled device (CCD), an image sensor, semiconductor circuit elements, etc. All dies in the plurality of dies 724 may include the same circuit element(s) and, if circuitry is present, the same circuit(s).

[0103] The die have a device side that has most or all of the electrical circuitry of the die, and a backside opposite the device side. In the embodiment shown in FIG. 7, the backside of the die in the plurality of dies 724 is disposed between the source substrate chuck 122 and the device side of the die. In another implementation, the device side of the die in the plurality of dies 724 is disposed between the source substrate chuck 122 and the backside of the die. The side of the die that faces the base 140 or components bonded to the base 140 is activated for hybrid bonding to the target substrate 748, and the dark bands along the bottom of the die in the plurality of dies 724 are used to indicate the activated surface. Any one or more dies may have TSVs or electrical components along the backside, and such dies may also include backside bonding sites that may act as future target substrate bonding sites that may be used later.

[0104] Before continuing with the method description, a portion of the dummy source die 924 will be described. Figure 9 includes a cross-sectional view of a portion of the dummy source die 924. The dummy source die 924 may be used in determining an overlay error between the dummy source die and a corresponding target feature of the target substrate 748. As used herein, a dummy die, such as the dummy source die 924, does not have electrical circuitry that is electrically connected or electrically connected to electrical circuitry in the KGD of the target substrate.

[0105] The dummy source die 924 includes a substrate 932, an insulating layer 942, and an alignment mark 944. In one implementation, the alignment mark 944 can be an array of dummy contact pads. The radiation can be used during subsequent radiation analysis operations. With respect to any particular layer, feature (e.g., alignment mark, interconnect, conductive plug, etc.), or material, the radiation can be transmitted through, absorbed by, or reflected by such layer, feature, or material. As used herein, "non-transparent" and variations thereof refer to a combination of being absorbed, reflected, or absorbed and reflected by such layer, feature, or material. The substrate 932 and insulating layer 942 are relatively transparent to radiation compared to the alignment mark 944, and the alignment mark 944 is relatively non-transparent to radiation compared to the substrate 932 and insulating layer 942. More details regarding metrology patterns are described in more detail later in this specification.

[0106] The substrate 932 can include a semiconductor substrate (e.g., Si, SiC, III-V semiconductor material, II-VI semiconductor material, etc.) with or without an insulating layer overlying the semiconductor substrate, a patterned polysilicon layer, etc. In one embodiment, the semiconductor material is Si, 4H-SiC, 6H-SiC, Al x Ga (1-x)N (0≦x≦1), or a combination of layers, each layer of the layer combination comprising one of the semiconductor materials. The insulating layer 942 may include one or more films of silicon oxide, silicon nitride, or silicon oxynitride. The alignment marks 944 may include a metal or metal alloy, such as Al, Al-1 wt% Cu, Cu, Ag, Au, Pt, Co, W, TiW, Ti, Ta, etc. More layers may be used for the dummy source die 924.

[0107] Zones 954 correspond to alignment marks 944, and zones 952 correspond to the regions between alignment marks 944. Zones 952 and 954 extend through, over, and under alignment marks 944. Any object or portion of dummy source die 924 that is within a particular zone 954 and outside of the corresponding alignment mark 944 is directly above or below the corresponding alignment mark 944, whether such object or portion is in contact with or spaced apart from the corresponding alignment mark 944. For example, with reference to FIG. 9 , a portion of substrate 932 in a particular zone 954 along the opposite side of alignment mark 944 is directly above the corresponding alignment mark 944 in the particular zone 954. Such portion does not directly overlie a portion of insulating layer 942 in an immediately adjacent zone 952 or a different alignment mark in a different zone 954.

[0108] 5, the method may further include moving the carriage to a dummy source die loading position, at block 526, and placing the dummy source die on the die carrier sheet, at block 528. Placing the dummy source die 924 may include one of the die carrier sheets removing the dummy source die 924 from the source substrate 722 on the source substrate chuck 122. In an alternative implementation, the die loading machine 180 loads the dummy source die 924 directly onto the carrier sheet 144, as will be described below with reference to FIG.

[0109] 1-4 and 10, the dummy source die 924 may be held by a die transport head 488 coupled to a positioning arm 486 of the die loading machine 180. The controller 160 or a local controller may send a signal for the carriage 146, the die loading machine 180, or both, to move the dummy source die 924 to overlap the optical element 150. The optical element 150 may be used to ensure that the dummy source die 924 is the correct die to be subsequently bonded to the target substrate 748. The optical element 150 may also be used to verify that the positioning of the dummy source die 924 is appropriate to allow transfer to the die transport sheet in the array of the die transport sheets 144. With reference to FIGS. 10 and 11, the die chuck of the die transport sheet in the array of the die transport sheets 144 may be extended toward the die transport head 488 to remove the dummy source die 924 from the die transport head 488, and the die chuck may be withdrawn from the die transport head 488.

[0110] In one implementation, the die chucks of the array of die carrier sheets 144 do not contact the active surfaces of the die, such as dummy source die 924. The dummy source die is depicted in FIG. 11 spaced apart from the die carrier sheets in the array of die carrier sheets 144 to indicate that the active surfaces of the dummy source die do not contact the die carrier sheets in the multiple die carrier sheets. The array of die carrier sheets 144 has a design that allows for picking up of the die along the side of the die, the side being between the device and backside of the die.

[0111] If the dummy source die 924 is too thin to be held by its sides, a backing plate (not shown) can be bonded to the die. For example, the die can have a thickness of less than 50 μm. The thickness of the backing plate, or the thickness of the backing plate and die combination, is sufficient to allow a pick-up head to pick up the backing plate or the backing plate and die combination without having the active surface of the die contact the pick-up head. The backing plate can have a thickness in the range of 100 μm to 500 μm.

[0112] The backing plate may be bonded to the dummy source die 924 using an adhesive compound. The backing plate may be later removed or may remain bonded to the die in the completed electrical device. The backing plate may be removed after the die is bonded to the target substrate 748. In one embodiment, the adhesive compound may be deactivated by exposure to actinic radiation. The actinic radiation may be in the range of 100 nm to 1000 nm. In such an embodiment, at least 70% of the actinic radiation is transmitted through the backing plate. In another embodiment, a solvent may be used to remove the adhesive compound from between the dummy source die and the backing plate.

[0113] In further implementations, the die chuck for the array of die transfer sheets 144 may have a design where the die chuck contacts the bottom surface of the dummy source die 924. After reading this entire specification, one of ordinary skill in the art can determine whether or not the array of die transfer sheets 144 should contact the dummy source die 924 for a particular application.

[0114] In another implementation, a dummy source die 924 and a production source die in the plurality of dies 724 can be attached to the source substrate 722. The dummy source die 924 can be removed from the source substrate 722. A method for picking up a die from the source substrate 722 is described in more detail with the process flow charts of Figures 27 and 28.

[0115] The method may further include, at block 542 of FIG. 5, transferring a dummy source die from the die carrier sheet to the bonding head. The carriage 146 is moved such that the alignment fiducials 128 are above the optical elements 150, as shown in FIG. 12. The information collected from the optical elements 150 may be received by the controller 160 or a local controller. The controller 160 or a local controller may use the information and send a signal for the carriage 146 to move to a desired position. In one implementation, as shown in FIG. 13, the carriage 146 is moved such that a bonding head in the array of bonding heads 124 is above the dummy source die 924. The movement may include moving the carriage in the X direction, the Y direction, rotating the carriage along the XY plane, or a combination thereof.

[0116] The die chucks for the die carrier sheets in the array of die carrier sheets 144 holding the dummy source die 924 can extend towards the bond heads in the array of bond heads 124, the die chucks for the bond heads can extend towards the die carrier sheet holding the dummy source die, or both. Figure 14 includes the dummy source die 924 after it has been transferred from the die carrier sheet to the bond head.

[0117] The method may include measuring the alignment of the dummy source die using an optical element at block 544 of FIG. 5. Referring to FIG. 15, an optical element 150 coupled to the carriage 146 may be used to position the dummy source die 924, measure an alignment error of the dummy source die 924, or both. Better positioning of the dummy source die 924 may be achieved by measuring an alignment error as a position of the dummy source die 924 on the bonding head relative to an ideal position of the dummy source die 924 on the bonding head. Information from the optical element 150 may be sent and received by the controller 160 or a local controller.

[0118] The controller 160 or the local controller can use the information to determine the alignment error and positioning amount of the dummy source die 924 so that the dummy source die 924 is more closely aligned with its corresponding target portion of the target substrate 748. The controller 160 or the local controller can send a signal so that the position of the dummy source die 924 is adjusted by moving the die chuck of the bond head using the limited range of movement of the die chuck relative to the corresponding body. Thus, in one implementation, moving the die chuck allows the position of the dummy source die 924 to be adjusted with respect to the target substrate 748 held by the target substrate chuck 148. The positioning and measuring alignment error can be performed iteratively until the alignment error is zero or an acceptably low value.

[0119] The method may include moving the carriage so that the bond head is over the target substrate in block 622 of Figure 6. In Figure 16, the carriage 146 is moved so that the dummy source die is over a corresponding target portion of the target substrate 748.

[0120] Before continuing with the method description, an exemplary structure for the target site will be described with reference to FIG. 17. FIG. 17 includes a cross-sectional view of a portion of one of the target dies 820. The KGD 824 and the failing die 834 can have a structure as shown in FIG. 17, including a portion of the metrology pattern including alignment marks 1744 that extend through the silicon oxide layer 1724 into the substrate 1700. The alignment marks 1744 can be an array of contact pads. The contact pads that are electrically connected to the electrical circuit elements are referred to herein as active contact pads. The array of contact pads can be dummy contact pads or active contact pads. The metrology pattern is addressed in more detail later in this specification with respect to FIG. 24. FIG. 17 includes only a portion of the substrate 1700, which can include a semiconductor substrate (not shown), an electronic component (not shown), an insulating layer 1702, interconnects 1704, and conductive plugs 1706.

[0121] Insulating layer 1702 may include multiple insulating layers including oxides, nitrides, or oxynitrides. Interconnects 1704 may be made of a metal or metal alloy such as Al, Al-1 wt% Cu, Cu, Ag, Au, Pt, Co, W, TiW, Ti, Ta, etc. Conductive plugs 1706 may include any of the materials described with respect to interconnects 1704. Silicon oxide layer 1724 may include silicon oxide, silicon nitride, or silicon oxynitride. Alignment marks 1744 may include any of the materials as described with respect to interconnects 1704. Alignment marks 1744 and interconnects 1704 may have the same or different materials.

[0122] 17, features other than the insulating layer may underlie at least a portion of the metrology pattern. Interconnects 1704 and conductive plugs 1706 underlie the metrology pattern. Portions of interconnects 1704 and conductive plugs 1706 are directly under alignment marks 1744 of the metrology pattern. The terms "directly under" and "directly under" were previously addressed with respect to FIG. 9. In other implementations, there may be more or fewer features other than insulating layer 1702, or no features at all, directly under the metrology pattern or alignment marks in the metrology pattern.

[0123] The method may further include bonding the dummy source die to a corresponding target portion of the target substrate at block 624 of FIG. 6. A die chuck for a bond head in the array of bond heads 124 may extend toward the target substrate 748, the target substrate chuck 148 may extend toward the array of bond heads 124, or both. Pressure is applied to bond the dummy source die 924 to its corresponding target portion of the target substrate 748 of FIG. 18. In one embodiment, the bond may be an oxide-oxide bond. The bonding pressure may be 0.5 N / cm. 2 ~20N / cm 2 The bonding can be performed at room temperature (e.g., at a temperature in the range of 20° C. to 25° C.) or above. The bonding is performed at a lower temperature than the subsequent anneal to expand the conductive metal in the die and at the target site. The temperature and pressure can be limited depending on the components in the membrane or device 110 present during bonding. For example, the temperature can be about 200° C. or less. After reading this specification, one of ordinary skill in the art can determine the pressure and temperature to be used for bonding.

[0124] 19 includes a top view of the target substrate 148 and the target substrate 748 after the dummy source die 924 has been bonded to a corresponding target site of the target substrate 748. To maintain a high yield of the target substrate 748, the dummy source die 924 may be bonded to a bad die 834 of the target substrate 748.

[0125] 20 includes a cross-sectional view of a dummy source die 924 and a portion of a bad die 834 in a metrology pattern for the dies. The alignment mark 944 for the dummy source die 924 may be at the bonding surface and contact the bad die 834, and the alignment mark 1744 for the bad die 834 may be at the bonding surface and contact the dummy source die 924. The alignment marks need not be at the bonding surface. For example, the alignment mark 944 in the dummy source die 924 may be spaced from the bonding surface by at least a portion of the insulating layer 942, such that the alignment mark 944 does not contact the bad die 834. Similarly, the alignment mark 1744 of the bad die 834 may be spaced from the bonding surface by at least a portion of the silicon oxide layer 1724, such that the alignment mark 1744 does not contact the dummy source die 924.

[0126] The method may include irradiating a metrology pattern in a radiation field at block 642 of Figure 6 and collecting radiation data corresponding to the metrology pattern in the radiation field at block 644. Figure 21 includes a cross-sectional view at this point in the method. The radiation tool 272 includes a radiation source 474, a reflector 476, which may be a formation of a frustum or another shape, and a radiation detector 478. The reflector 476 is optional and may or may not be present.

[0127] The radiation output by the radiation source 474 may be reflected by some components in the metrology pattern in the dummy source die 924 and the target site of the target substrate 748. The reflected radiation may be received by the radiation detector 478. In the same or a different implementation, the radiation source 474 may be absorbed and cause some components in the dummy source die 924 or the target site of the target substrate 748 to fluoresce, phosphoresce, or generate ionizing radiation, and the radiation detector 478 may detect the fluorescence, phosphorescence, or ionizing radiation emitted by such components. Thus, the radiation detector 478 may detect the same or a different form of radiation emitted by the radiation source 474. For purposes of this specification, components in the dummy source die 924 or the target site of the target substrate 748 that fluoresce, phosphoresce, or generate ionizing radiation in response to exposure to radiation are a type of non-transparent component.

[0128] In another implementation, radiation may be transmitted through the dummy source die 924 and the target substrate 748 may be detected by a radiation detector in the target substrate chuck 148, the carriage 146, the base 140, or another suitable part of the system 100. Like the radiation detector 478, the radiation detector for the transmitted radiation may detect the same type of radiation emitted by the radiation source 474 or another form of radiation, such as fluorescence, phosphorescence, or ionizing radiation.

[0129] The radiation can be selected such that the dummy source die 924 includes relatively non-transmissive and relatively transparent portions. As previously described, the alignment mark 944 can include a metal or metal alloy that is more transmissive, reflective, or absorbing and reflective of radiation than the insulating layer 942 and the substrate 932, which can include a semiconductor material, an insulating layer, or a combination thereof. The non-transmissive portions can be the alignment mark 944, and the relatively transparent portions can include the insulating layer 942 and the substrate 932. In one embodiment, the radiation is infrared (IR) and can have a wavelength in the range of 0.7 μm to 1000 μm, and in certain implementations 1 μm to 10 μm, or 1 μm to 2 μm.

[0130] The radiation detector 478, a radiation detector in the target substrate chuck 148, the carriage 146, the base 140, or another suitable portion of the system 100, or another component coupled to the radiation tool 272, may collect at least some radiation data. The radiation data may include measurements of the intensity of the detected radiation and the location where such intensity was detected. If the radiation detector 478 does not collect position information, such position information may be obtained from the carriage 146 or any other suitable component in the system 100.

[0131] FIG. 22 includes exemplary radiation data and position information that may be collected for the dummy source die 924. Referring to FIGS. 21 and 22, the detection signal from the radiation detector 478 is normalized using only the positions with the alignment marks 944 and 2644 and the positions between the alignment marks 944. The alignment mark 2644 is a specific alignment mark and is one of the alignment marks 944. Most of the radiation from the radiation source 474 is reflected by the alignment marks 944, having a value of 1 in FIG. 22, and significantly less radiation is reflected at the positions between the alignment marks, having a value of 0 in FIG. 22. In practice, not all of the radiation is reflected by the alignment marks 944, and at least some of the radiation may be reflected by the combination of the substrate 932 and the insulating layer 942.

[0132] The metrology pattern is described with respect to FIG. 23 and FIG. 24. After reading this entire specification, one skilled in the art will understand that the metrology patterns shown in FIG. 23 and FIG. 24 are exemplary and are not limiting of the shape or size of features and spacing between features of other metrology patterns. FIG. 23 includes a top view of a metrology pattern 2340 of a dummy source die 924. The metrology pattern 2340 may include alignment marks used to help determine overlay error. In the embodiment shown in FIG. 23, the alignment marks are an array of dummy contact pads. As used herein, a dummy contact pad of a die is a contact pad that is not electrically connected to an electrical circuit element or electrical circuit within the die. In the same or another implementation, the metrology pattern 2340 may be along the bonding surface of the dummy source die 924, spaced apart from the bonding surface, or not along the bonding surface. The dummy source die 924 has substantially the same dimensions and mass (within 1%) as one of the multiple dies 724. The dummy source die 924 has dummy contact pads arranged in the same pattern as the dummy contact pads on the multiple dies 724, except that the phase is offset relative to the phase on one of the multiple dies 724. While each of the multiple dies 724 may include multiple layers of features patterned within the die, the dummy source die 924 may include only unpatterned layers or no interior layers but may include patterned contact pads.

[0133] FIG. 24 includes a top view of a metrology pattern 2440 of a target site. The metrology pattern 2440 can be a dedicated metrology structure that is not electrically connected to any electrical circuit elements or electrical circuits in the target site. In this implementation, the metrology pattern 2440 can include an alignment mark that is an array of dummy contact pads. In another implementation, the metrology pattern 2440 can correspond to a portion of a target site having closely packed contact pads that are electrically coupled to electrical circuit elements and are not designed to be part of a dedicated alignment mark. With reference to FIG. 8, such active contact pads can be part of a bad die 834 of a target substrate 748. The metrology pattern 2340 of a source 924 can have a pattern that corresponds to the closely packed active contact pads of a target die. In one implementation, the target site of the dummy source die 924 is identical to the target site of the multiple dies 724.

[0134] In any of the above implementations, the metrology pattern 2440 may be an array of contact pads (dummy or active) located along the bonding surface of the target die, or may be an array of metal or metal alloy features formed at a lower elevation within the target site. The target die may include a lowest level of interconnects, which may include gate electrodes and word lines, another level including bit lines and interconnects electrically connected to source regions, drain regions, gate electrodes, or combinations thereof, yet another level of interconnects, further levels of interconnects, and finally a level including contact pads. Due to electromagnetic radiation, parasitic coupling, etc., the alignment marks 1744 within the target site may be formed with the lower level of interconnects as opposed to the contact pad level. Thus, the metrology pattern 2440 may be spaced apart from and not located along the bonding surface.

[0135] 9 and 23, the portion of the dummy source die 924 within zone 952 is a relatively transmissive portion, and the portion of the dummy source die 924 within zone 954 is a relatively non-transmissive portion due to the presence of alignment mark 944. Within metrology pattern 2340, alignment mark 944 corresponds to a relatively non-transmissive portion, and from the top view of FIG. 23, the portion of metrology pattern 2340 outside the alignment mark is a relatively transmissive portion. Within metrology pattern 2440 of FIG. 24, alignment mark 1744 corresponds to a relatively non-transmissive portion, and from the top view of FIG. 24, the portion of metrology pattern 2440 outside the alignment mark is a relatively transmissive portion.

[0136] The difference in transmission between the relatively transparent and relatively non-transparent portions is at least 25%. In the same or different embodiments, (1) no more than 25% of the radiation is transmitted through the relatively non-transparent portions, (2) at least 75% of the radiation is transmitted through all of the relatively transparent portions, or (3) no more than 25% of the radiation is transmitted through the relatively non-transparent portions and at least 75% of the radiation is transmitted through all of the relatively transparent portions.

[0137] Figure 25 shows metrology patterns 2340 and 2440 after a dummy source die 924 has been bonded to a target portion of a target substrate 748. Radiation area 2540 corresponds to an area that may be used to determine overlay error and is shown in dashed lines. Figure 26 includes a portion of radiation area 2540 and is described in more detail below with respect to specific, non-limiting examples.

[0138] The metrology patterns 2340 and 2440 are designed such that when the alignment marks (such as an array of dummy contact pads for a particular example) in the metrology patterns 2340 and 2440 are phase shifted by an angle of 180°, the overlay error is zero. In one implementation, the phase shift can be in a range of 90° to 270°, in a range of 150° to 210°, or another suitable range that can be useful for determining the overlay error. The alignment marks 944 do not overlap any of the alignment marks 1744. The phase shift can be in one direction or in different directions. With reference to FIG. 25, the metrology patterns 2340 and 2440 are phase shifted in both the X and Y directions. The phase shift in more than one direction can help determine an X-direction correction for the overlay error, a Y-direction correction for the overlay error, a rotation correction (θ) in the plane defined by the X and Y directions, etc. The out-of-phase arrangement can provide a more accurate overlay error determination compared to overlaid alignment marks, at least in part due to the difficulty in detecting the edges of the alignment marks, especially when the alignment marks are made of the same material and are highly reflective. In one implementation, the metrology pattern 2340 is in the form of dummy contact pads of the dummy source die 924 that are identical to, but phase shifted from, the contact pads in the metrology pattern 2440 of one of the multiple dies 724. In one implementation, the metrology pattern 2440 is in the form of contact pads that are identical to the contact pads of the target portions of the target substrate 748.

[0139] The method may include analyzing the radiation data to determine an overlay error between the dummy source die and the corresponding target portion at block 646 of FIG. 6. The radiation data corresponding to the radiation area 2540 may be transmitted to the controller 160 or a local controller. The data may or may not be stored in memory 162. The data may be transmitted externally to a processor or memory external to the system 100. The data regarding the detected radiation and position may be further processed by the controller 160, a local controller, or a computer external to the system 100 to determine the overlay error. The following paragraphs describe various methods for analyzing the radiation data. Each of the different methods or a combination of such methods may be used for the analysis.

[0140] An image may be generated from the radiation data. Imaging may be performed using the radiation data to generate an image corresponding to radiation region 2540. The radiation region may include all of metrology pattern 2340, metrology pattern 2440, or both, or may include portions, but not all, of either or both of metrology patterns 2340 and 2440. In FIG. 25, radiation region 2540 includes portions of metrology patterns 2340 and 2440.

[0141] Imaging of radiation region 2540 may be aided by taking into account several considerations. Imaging may be performed using a radiation detector with a depth of focus capable of focusing on alignment marks in metrology patterns 2340 and 2440. The depth of focus may be 9 μm or less, 7 μm or less, or 5 μm or less. In certain implementations, alignment marks 944 and 1744 may be an array of dummy contact pads on the bonding surface of dummy source die 924 and corresponding target sites. The thickness of the dummy contact pads may range from 1 μm to 4 μm. Limiting the depth of focus may reduce the likelihood of focusing on other metal or metal alloy components at lower heights in the target sites, such as interconnects 1704 and conductive plugs 1706. Such components will be invisible or blurred. The depth of focus may be adjusted to the particular thickness used for alignment marks 944 and 1744. Thus, the depth of focus may be significantly less than 9 μm or greater than 9 μm if components adjacent to the metrology pattern 2440 do not interfere with imaging. In the same or different embodiments, other metal or metal alloy components at lower heights within the target site may be blurred.

[0142] After the image is generated, the controller 160, a local controller, or a computer external to the system 100 can determine the X-direction offset, the Y-direction offset, the rotational offset, or a combination thereof, between the measurement patterns 2340 and 2440 in the radiation field 2540. Alternatively, a human operator can instruct the controller 160, a local controller, or a computer external to the system 100 to use the image to measure the X-direction offset, the Y-direction offset, the rotational offset, or a combination thereof, between the measurement patterns 2340 and 2440 in the radiation field 2540.

[0143] Analysis of the radiation data can be performed without generating an image. The radiation data can be collected by scanning multiple rows in the X direction or multiple columns in the Y direction of the radiation area 2540. The alignment marks 944 and 1744 absorb, reflect, or have a combination of absorption and reflection of substantially more radiation than locations between the alignment marks 944 and 1744, allowing substantially more radiation to be transmitted between the alignment marks 944 and 1744 compared to radiation passing through the alignment marks 944 and 1744.

[0144] 22-26, when the radiation detector 478 receives the reflected radiation, the alignment marks 944 and 1744 reflect substantially more radiation than a location where the alignment marks 944 and 1744 are not present. In one implementation, the radiation detector 478 can output a relatively high signal when the radiation detector 478 is detecting radiation reflected by the alignment marks 944 and 2644 (see FIGS. 22 and 1744) and can output a relatively low signal when the radiation detector 478 is detecting radiation reflected between the alignment marks 944 and 1744. In the case of a radiation detector below the target substrate 748, the radiation detector can detect radiation that is transmitted through the dummy source die 924 and through all or a portion of the target portion of the target substrate 748. In this implementation, the radiation detector may output a relatively low signal corresponding to alignment marks 944 and 1744, and if present in metrology pattern 2440, interconnects 1204 and conductive plugs 1206 may output a relatively high signal when radiation detector 478 is detecting radiation reflected between alignment marks 944 and 1744. The examples described below are based on reflected radiation unless explicitly stated to the contrary.

[0145] The controller 160, a local controller, or a computer external to the system 100 can use the signal amplitude difference and position information to determine the size and location of the alignment marks 944 and 1744, as well as the pitch corresponding to the metrology patterns 2340 and 2440 within the radiation area 2540 of FIG. 25. The pitch can be the size of the feature and the space between the feature and an immediately adjacent feature in a particular direction, such as the X-direction or the Y-direction. The X-direction pitch can be the same as the Y-direction pitch or can be different. The features can be a pair of alignment marks 944 or alignment marks 1744. As a non-limiting example, the radiation data can include a scan along a row that can provide a signal similar to that shown in FIG. 22. The scan with the widest signal for a particular alignment mark 2644 can be analyzed. With reference to FIGS. 22 and 26, the alignment mark 944 closest to the right can be the particular alignment mark 2644. With respect to the scan, the center 2664 of the signal closest to the right is at a particular location, such as 2.5 μm from the right side of the radiation region 2540, when the alignment marks 944 including the particular alignment mark 2644 have an X-direction dimension, such as a width or diameter, of 1.0 μm, and an X-direction space between the alignment marks 944 including the particular alignment mark 2644 of 2.0 μm.

[0146] The Y dimension can be obtained by identifying which row the scan belongs to. A scan can refer to one or more lines of an image obtained by a CCD or one or more lines obtained by scanning using an optical element. As a further example, scan 0 may be along the upper boundary of the radiation area 2540 (closest to the top of FIG. 25), with each subsequent scan being displaced in the Y direction by 0.1 μm from the previous scan toward the bottom of FIG. 25. The scan data of FIG. 22 can correspond to scan 97, and thus the center of the alignment mark is 97 scans, 0.1 μm / scan, time 9.7 μm from the top edge of the radiation area 2540. Thus, the center of a particular alignment mark 2644 is 2.5 μm from the right side of the radiation area (X-direction coordinate) and 9.7 μm from the top of the radiation area 2540 (Y-direction coordinate). Other alignment marks within the radiation area 2540 can be analyzed in the same manner.

[0147] 23-26, the controller 160, a local controller, or a computer external to the system 100 may have some information regarding the radiation region 2540, such as the arrangement of alternating rows or alternating columns of alignment marks 944, including particular alignment marks 2644 and 1744. The controller 160, a local controller, or a computer external to the system 100 may use the X- and Y-directional coordinates of the alignment marks 944 and 1744 to determine the pitch corresponding to the alignment marks 944 and 1744 and the overlay error based on the radiation data collected for the radiation region 2540.

[0148] A spatial frequency analysis may be performed on the radiation data collected from radiation area 2540, for metrology patterns 2340 and 2440. The spatial frequency analysis may be performed in the spatial frequency domain, with or without an image of radiation area 2540. The spatial frequency analysis may be performed using a value based on the detected pitch and feature size. The spatial frequency may be set to exclude features such as word lines, bit lines, or other interconnects, and isolated features. Considering the above, the pitch divided by the feature size is the quotient, and the value of the spatial frequency analysis may be the quotient + / - 10% to account for manufacturing or other variances. In an alternative implementation, optical techniques may be used to blur features that are in planes of no interest. The value may range from 1.5 to 9.0.

[0149] The wordline or bitline may have no pitch in one of the directions, or if present, the segmented wordline or segmented bitline may have a relatively large length in a direction and a relatively small space between the segments in that same direction, so for a segmented bitline or segmented wordline, the quotient (pitch divided by feature size) will be less than 1.1.

[0150] The isolated features may have no pitch or may have a relatively small feature size and a relatively large space between the isolated features. If the isolated features have a pitch, the pitch divided by the feature size may be greater than 10, 50, or more. Alternatively, the isolated features may vary widely in size. The X-dimension, the Y-dimension, or both may be as small as 1.0 μm, or as large as 10 μm, 50 μm, or more. The spatial frequency analysis may include lower and / or upper limits for the X-dimension, the Y-dimension, or both. In a non-limiting example, each alignment mark 944 and 1744 may be designed to have an X-dimension of 1.5 μm and a Y-dimension of 1.5 μm. In practice, the X-dimension and the Y-dimension may vary by up to 0.2 μm due to manufacturing or other variations. Thus, the features have X-dimension and Y-dimension in the range of 1.3 μm to 1.7 μm.

[0151] The spatial analysis may be performed on only those features that meet one or more of: (1) a value associated with the quotient (pitch divided by feature size); (2) the feature size (X-dimension, Y-dimension, or both), another arbitrary criterion, or a combination thereof.

[0152] The analysis of the radiation data may include comparing the image corresponding to radiation region 2540 to a standard image corresponding to radiation region 2540. The standard image may be an image from different bonded dummy source dies bonded to different target sites of different target substrates, where the bonded different dies and different target sites have no overlay error or relatively low overlay error. In the same or another implementation, the standard image may be a computer-generated image corresponding to zero overlay error. The computer-generated image may be based on theoretical data and does not include actual radiation data from the bonded dies.

[0153] Pattern recognition can be useful in both reflected and transmitted radiation modes, especially when radiation passes through the target substrate 748. The metrology pattern 2440 may have other features close to or underlying the metrology pattern 2440 in the target portion, such as the interconnects 1204 and conductive plugs 1206. Pattern recognition can determine that the other features are not part of the metrology pattern 2440, and such other features may be excluded from the analysis when determining the overlay error.

[0154] As mentioned above, different techniques can be performed when analyzing the radiation data or a derivative thereof (e.g., an image generated from the radiation data). The analysis can be performed using any particular technique or a combination of techniques can be used. The analysis can further continue to determine the overlay error between the dummy source die 924 and its corresponding target feature.

[0155] As shown in Fig. 25, alignment mark 944 is offset in both the X and Y directions with respect to alignment mark 1744. Radiation data can be used to determine if and by how much, or if at least a portion of alignment mark 944 is from alignment mark 1744. The following are non-limiting examples intended to illustrate, but not limit the scope of the invention as defined in the claims. For overlay errors, positive values ​​in the X direction are closer to the right side of Figs. 25 and 26, and positive values ​​in the Y direction are closer to the top of Figs. 25 and 26.

[0156] 25 and 26, each of the alignment marks 944, including particular alignment marks 2644 and 1744, may have a diameter of 1.0 μm. In the X direction (along a row of alignment marks), the spacing between alignment marks 944 is 2.0 μm and the spacing between alignment marks 1744 is 2.0 μm. The X direction pitch of alignment marks 944 is 3.0 μm and the X direction pitch of alignment marks 1744 is 3.0 μm. In the Y direction (along a row of alignment marks), the spacing between alignment marks 944 is 3.0 μm and the spacing between alignment marks 1744 is 3.0 μm. The Y direction pitch of alignment marks 944 is 4.0 μm and the Y direction pitch of alignment marks 1744 is 4.0 μm.

[0157] Metrology patterns 2340 and 2440 are designed such that, in the absence of overlay error, alignment marks 944 and 1744 are 180° out of phase with each other in the X and Y directions. Thus, when alignment marks 944 and 1744 are 180° out of phase with each other, along a row of alignment marks 944 (X direction), the center of alignment mark 944 is 1.5 μm away from each line that defines the center of alignment mark 1744 in the column immediately adjacent to alignment mark 1744. Along a column of alignment marks 944 (Y direction), the center of alignment mark 944 is 2.0 μm away from each line that defines the center of alignment mark 1744 in the row immediately adjacent to alignment mark 1744.

[0158] Figure 26 illustrates how overlay error can be determined based on metrology patterns 2340 and 2440, and based on the pitch and X- and Y-direction offsets discussed above. In Figure 26, as discussed above, alignment mark 2644 is a particular one of alignment marks 944 and is used to improve understanding of how to determine overlay error.

[0159] In the specific example, alignment mark 2644 is 1.2 μm away from the line defined by the centers of alignment marks 1744 in the column immediately to the left of alignment mark 2644, and 1.8 μm away from the line defined by the centers of alignment marks 1744 in the column immediately to the right of alignment mark 2644. With zero overlay error, both distances would be 1.5 μm. Alignment mark 2644 is 0.3 μm closer to the left in FIG. 26. The overlay error is therefore −0.3 μm in the X direction.

[0160] Continuing with the same example, alignment mark 2644 is 2.6 μm away from the line defined by the centers of alignment marks 1744 in the row immediately above alignment mark 2644, and 1.4 μm away from the line defined by the centers of alignment marks 1744 in the row immediately below the particular alignment mark 944. With zero overlay error, both distances would be 2.0 μm. Alignment mark 2644 is 0.6 μm closer to the bottom of FIG. 26 than it should be. Thus, the overlay error is −0.6 μm in the Y direction. The overlay error between the dummy source die 924 and the target feature is −0.3 μm in the X direction and −0.6 μm in the Y direction.

[0161] The overlay error may or may not have a rotational component. For example, a line defined by the centers of a row or column of alignment marks 944 should be parallel to a line defined by the centers of immediately adjacent rows or columns of alignment marks 1744. If a rotational overlay error is present, the line corresponding to alignment mark 944 will intersect with the line corresponding to alignment mark 1744. The intersection of the lines may occur within radiation region 2540 or outside radiation region 2540. In this particular example, there is no rotational overlay error.

[0162] The method may include, at block 648 of FIG. 6, determining an overlay error compensation based at least in part on the overlay error. Continuing with the example, when the source die is subsequently positioned, the compensation for the overlay error is the same value as the overlay error, but in the opposite direction to the overlay error. Thus, for the dummy source die, the dummy source die 924 should have been 0.3 μm closer to the right side of FIG. 26 (X-direction overlay error compensation of +0.3 μm) and 0.6 μm closer to the top of FIG. 26 (Y-direction overlay error compensation of +0.6 μm). Or, for the target substrate 748, the target substrate 748 should have been 0.3 μm closer to the left side of FIG. 26 (X-direction overlay error compensation of −0.3 μm) and 0.6 μm closer to the bottom of FIG. 26 (Y-direction overlay error compensation of −0.6 μm).

[0163] When positioning the next source die or its corresponding target feature on the target substrate 748, the next source die should be 0.3 μm closer to the right of FIG. 26 to compensate for the X-direction overlay error and 0.6 μm closer to the top of FIG. 26 to compensate for the Y-direction overlay error. The previous example included many specific details. The overlay error may be determined for multiple alignment marks 944, not just one alignment mark.

[0164] The method can continue with other die being bonded to other target sites on target substrate 148. This portion of the method corresponds to the process flow charts of Figures 27 and 28, and the illustrations of Figures 29-37.

[0165] The method includes performing alignment and metrology on the plurality of dies on the source substrate and the plurality of die carrier sheets at block 2722 of FIG. 27. Referring to FIG. 29, the alignment and metrology operations can be performed on the array of the plurality of dies 724 and the die carrier sheet 144. The optical element 150 can be used to gather information about the plurality of dies 724. The information from the optical element 150 can be sent to the controller 160 or a local controller and used to determine a source pitch of the plurality of dies 724 coupled to the source substrate chuck 122. The source pitch can include an X-direction source pitch, a Y-direction source pitch, or both an X-direction source pitch and a Y-direction source pitch. If necessary or desired, the information can be used to identify or confirm that the dies in the plurality of dies 724 are in the correct relative position of the target site of the target substrate 748. The plurality of dies 724 can be a production source KGD that is coupled to a KGD 824 of the target substrate 748.

[0166] The method may further include, at block 2724 of FIG. 27, changing the pitch of the array of the die carrier sheet to a source matching pitch. The source matching pitch may be the same as or within a tolerance of the source pitch. Similar to the source pitch, the source matching pitch may include an X-direction source matching pitch, a Y-direction source matching pitch, or both an X-direction and a Y-direction source matching pitch. The tolerance may account for slight differences that may be due to repeatability of the equipment or manufacturing process. As used herein, the tolerance may be within 2.0%, 1.0%, or 0.5% of the desired value. For example, the source matching pitch may be within 2.0%, 1.0%, or 0.5% of the source pitch.

[0167] The method may include loading the set of dies onto the array of die carrier sheets at block 2726 of FIG. 27. Referring to FIG. 30, the die chuck of the array of die carrier sheets 144 may be extended toward the source substrate 722 to pick up the set of dies 3024 and pull the die chuck back from the source substrate 722. The set of dies 3024 may be a set of production source KGD. Picking up dies is a specific type of loading. Another type of loading for the system 100 may be to load any or all of the die carrier sheets using the die loading machine 180 used to load the dummy source die 924. With respect to the source substrate 722, the dies that are picked up may be the dies that are closest to each other, or one or more other dies may be between the dies that are picked up, as shown in FIG. 30. Within the plurality of dies 724, the dies that are not picked up remain coupled to the source substrate chuck 122, as shown in FIG. 30.

[0168] In one embodiment, the die chuck of the array of die carrier sheet 144 does not contact the activated surfaces of the set of dies 3024. The die chuck of the array of die carrier sheet 144 may be a Bernoulli chuck. Although the set of dies 3024 is held by the die chuck, the set of dies 3024 is depicted in FIG. 30 as being spaced apart from the array of die carrier sheet 144 to show that the activated surfaces of the dies do not contact the die carrier sheet. The array of die carrier sheet 144 may have a design that allows for picking up the dies along a side of the die, the side being between the device and backside of the die.

[0169] If the die is too thin to be held by its sides, a backing plate can be bonded to the die. For example, the die can have a thickness of less than 50 μm. The thickness of the backing plate, or the thickness of the backing plate and die combination, is sufficient to allow a pick-up head to pick up the backing plate or the backing plate and die combination without having the activated surface of the die contact the pick-up head. The backing plate can have a thickness in the range of 100 μm to 500 μm.

[0170] The backing plate may be bonded to the die using an adhesive compound. The backing plate may be later removed or may remain bonded to the die in the completed electrical device. After the die is bonded to the target substrate 748, the backing plate may be removed. In one embodiment, the adhesive compound may be inactivated by exposure to actinic radiation. The actinic radiation may be in the range of 100 nm to 1000 nm. In such an embodiment, at least 70% of the actinic radiation is transmitted through the backing plate. In another embodiment, a solvent may be used to remove the adhesive compound from between the die and the backing plate.

[0171] In another implementation, a die may not have an activated surface, but may have relatively weak components along the surface that is bonded to the target substrate 748, and such surface should not contact a die carrier sheet in the array of die carrier sheets 144. A die carrier sheet as described for dies with activated surfaces may be used for dies that have weak components along the surface that faces the die carrier sheet.

[0172] In further implementations, the die chuck for the array of die carrier sheets 144 may have a design where the die chuck contacts the bottom side of the set of dies 3024. After reading this entire specification, one of ordinary skill in the art can determine whether the array of die carrier sheets 144 should contact the device side or backside of the dies and determine a design that meets the needs or desires of a particular application.

[0173] The method may further include modifying the pitch of the array of the die carrier sheet to a bond head matching pitch at block 2742 of FIG. 27. The bond head matching pitch may be the same as the bond head pitch of the array of bond heads 124 or may be within an acceptable tolerance range. The bond head matching pitch and the bond head pitch may include an X-direction bond head matching pitch, a Y-direction bond head matching pitch, or both an X-direction and a Y-direction bond head matching pitch. The tolerance may account for slight differences that may be due to repeatability of the equipment or manufacturing process. As used herein, the tolerance may be within 2.0%, 1%, or 0.5% of the desired value. For example, the bond head matching pitch may be within 2.0%, 1%, or 0.5% of the bond head pitch.

[0174] The method may include moving the carriage to an alignment position at block 2744 of Figure 27. In Figure 31, the carriage 146 is moved so that the alignment fiducial 128 is above the optical element 150. The information collected from the optical element 150 may be received by the controller 160 or a local controller. The controller 160 or local controller may use the information for subsequent movement of the carriage 146 to a desired location.

[0175] The method may further include moving the carriage 146 so that the set of dies 3024 may be transferred from the array of die carrier sheet 144 to the array of bond heads 124. In the embodiment shown in Figure 32, the carriage 146 is moved so that the array of bond heads 124 is above the set of dies 3024. The movement may include moving the carriage 146 in the X direction, the Y direction, rotating the carriage along the XY plane, or a combination thereof.

[0176] The method may further include transferring the set of dies from the array of die carrier sheets to the array of bond heads at block 2822 of Figure 28. The die chucks for the array of die carrier sheets 144 may extend toward the array of bond heads 124, the die chucks for the bond heads in the array of bond heads 124 may extend toward the array of die carrier sheets 144, or both. Figure 33 includes a set of dies 3024 after transferring the set from the array of die carrier sheets 144 to the array of bond heads 124.

[0177] The method may include measuring the alignment of the set of dies using an optical element at block 2824 of FIG. 28. An optical element 150 coupled to the carriage 146 may be used in positioning any one or more of the dies in the set of dies 3024, measuring the alignment error of the set of dies 3024, or both, as shown in FIG. 34. Better positioning of the dies may be achieved by measuring the alignment error as the position of each die on each bond head relative to the ideal position of the die on the bond head. Information from the optical element 150 may be transmitted to and received by the controller 160 or a local controller. The controller 160 or a local controller may use the information to determine the alignment error and the amount of positioning of the die such that the die is more closely aligned with its corresponding target portion of the target substrate 748. The controller 160 or a local controller may send a signal such that the position of each die is adjusted by moving the die chuck of the bond head using the limited range of movement of the die chuck relative to the corresponding body. Thus, in one implementation, moving the die chuck allows the position of the die to be adjusted relative to the target substrate 748 held by the target substrate chuck 148. Positioning and measuring alignment errors may be performed iteratively until the alignment error is zero or an acceptably low value.

[0178] The method may include adjusting the positions of the set of dies or the target features to compensate for the overlay error at block 2842 of FIG. 28. Adjusting the positions to compensate for the overlay error may or may not be performed in addition to adjusting the positions to achieve proper alignment. The adjustments may compensate for the overlay error between the dummy source die 924 and its corresponding target die, allowing for better electrical and physical connection between the active contact pads of the set of dies 3024 and the active contact pads of the target features of the target substrate 748 when the set of dies 3024 is subsequently bonded to the target substrate 748. Ideally, the overlay error compensation will fully compensate for the overlay error, but due to manufacturing variations or differences between the dummy source die and the production source die, the overlay error compensation may undercompensate or overcompensate the overlay error.

[0179] With respect to die positioning, the overlay error compensation is in the opposite direction to the overlay error. For the particular example described with respect to FIG. 26, the X-direction overlay error is −0.3 μm and the Y-direction overlay error is −0.6 μm. When positioning the pair of dies 3024, the X-direction overlay error compensation is +0.3 μm and the Y-direction overlay error compensation is +0.6 μm. Before bonding, the position of the pair of dies 3024 can be adjusted by moving the pair of dies 3024 in the X-direction +0.3 μm and in the Y-direction +0.6 μm. Referring briefly to FIG. 36, before bonding the pair of dies 3024, the pair is moved 0.3 μm to the right and 0.6 μm up.

[0180] In an alternative implementation, instead of positioning the set of dies 3024, the target substrate 748 can be positioned. For positioning the target substrate 748, the overlay error compensation is in the same direction as the overlay error. When positioning the target substrate 748, the X-direction overlay error compensation is −0.3 μm and the Y-direction overlay error compensation is −0.6 μm. Before bonding, the position of the target substrate 748 can be adjusted by moving the set of dies 3024 −0.3 μm in the X-direction and −0.6 μm in the Y-direction. Referring briefly to FIG. 36, before bonding the set of dies 3024, the target substrate 748 can be moved −0.3 μm in the X-direction (left side) and −0.6 μm in the Y-direction (downward).

[0181] The overlay error may include a rotational component. In this example, the overlay does not include a rotational component. However, in another implementation, the overlay error may include a rotational component having an angle θ in a counterclockwise direction (−θ). When positioning the set of dies 3024, the overlay error compensation may rotate the set of dies 3024 by the angle θ in a clockwise direction (+θ). When positioning the target substrate 748, the overlay error compensation may rotate the target substrate 748 by the angle θ in a counterclockwise direction (−θ).

[0182] The method can be used to compensate for X-direction overlay error, Y-direction overlay error, rotational overlay error, or a combination thereof. Compensating for overlay error allows for more accurate bonding of active contact pads between source dies and target sites on a destination substrate. Compensating for overlay error can be performed before or after adjusting the position of the set of dies 3024 with respect to misalignment (block 2824 of FIG. 28 ) or moving the carriage 146, as described below with respect to block 2844.

[0183] The method may include moving the carriage so that the array of bond heads is over the target substrate at block 2844 of Figure 28. The carriage 146 is moved so that the set of dies 3024 is over the target portion of the target substrate 748.

[0184] The method may further include bonding the set of dies to corresponding target sites of the target substrate at block 2846 of Figure 28. A die chuck for the array of bond heads 124 may be extended toward the target substrate 748, the target substrate chuck 148 may be extended toward the array of bond heads 124, or both. Pressure is applied to bond the set of dies 3024 to corresponding target sites of the target substrate 748 of Figure 35. In one embodiment, the bond may be an oxide-oxide bond. The bonding pressure may be 0.5 N / cm 2 ~20N / cm 2 The bonding can be performed at room temperature (e.g., at a temperature in the range of 20° C. to 25° C.) or above. The bonding is performed at a lower temperature than the subsequent anneal to expand the conductive metal in the die and target site. The temperature and pressure can be limited depending on the components in the film or device 110 present during bonding. For example, the temperature may be about 200° C. or less. After reading this specification, one of ordinary skill in the art can determine the pressure and temperature to be used for bonding.

[0185] FIG. 36 includes a top view of the target substrate chuck 148 and the target substrate 748 after the set of dies 3024 are bonded to corresponding target sites of the target substrate 748. The target sites are part of the KGD 824 of the target substrate 748. FIG. 37 includes a top view of bonded contact pads 3724 at the target sites. Interconnects, electrical circuit elements, electrical circuitry including electrical circuit elements, or combinations thereof are present in the source die and the KGD 824, but are not shown in FIG. 37 to simplify understanding of how the active contact pads between the source die and the KGD 824 overlap each other. Each of the bonded contact pads 3724 may include an active contact pad 3726 of a source die in the set of dies 3024 and an active contact pad of a corresponding target site of the KGD 824. The overlap between the active contact pads is very good, but not exact. Through adjustment, the overlay error was nearly compensated for, but not completely compensated for. Thus, in a particular example, adjusting the position using overlay error compensation may have insufficiently compensated for the overlay error between the source die and its corresponding target feature. In another example having a different source die, adjustment using overlay error compensation may fully compensate or overcompensate for the overlay error between the dummy source die 924 and its corresponding target feature on the target substrate 748.

[0186] In decision step 2848 of Figure 28, a determination is made as to whether another set of die should be transferred to the destination substrate. If no more die are to be transferred ("no" branch), the transfer operation is completed and the method of transferring die ends. If more die are to be transferred ("YES" branch from decision step 2848 of Figure 28), the method continues by transitioning to block 2724 of Figure 27.

[0187] If necessary or desired, another dummy source die can be bonded to the target substate 748 and used to determine another overlay error. This method can be any of the methods previously described with respect to Figures 5-26. The method for the other dummy source die can be the same or different from the method used for the dummy source die 924.

[0188] FIG. 38 includes a cross-sectional view of the system in which a dummy source die 3824 is held by the die transport head 488. The method for the dummy source die 3824 may or may not include adjusting the position of the dummy source die 3824 to compensate for the overlay error relative to the dummy source die 924. FIG. 39 includes a top view of the target substrate chuck 148 and the target substrate 748 after the dummy source die 3824 has been bonded to the target feature of the bad die 834. FIG. 40 includes a cross-sectional view during collection of radiation data corresponding to the metrology pattern of the other dummy source die 3824 and its corresponding target feature of the target substrate 748. The dummy source die 924 and the set of dies 3024 are present above the target substrate 748 but are not shown in FIG. 40 because FIG. 40 is along a different cross-sectional plane.

[0189] The overlay error of the dummy source die 924 and 3824 may be different. The information provided by the overlay error for the dummy source die 3824 may depend on when the dummy source die 3824 was bonded relative to the bonded set of the source KGD.

[0190] Bonding of the dummy source die 3824 may be performed prior to bonding the set of dies 3024 to the target features of the KGD 824. Radiation data for the dummy source die 3824 and its corresponding target features collected prior to bonding any source KGD may be useful to verify that the overlay error compensation for the dummy source die 924 was sufficient or to verify whether the overlay error compensation should be modified to reduce the amount of under-compensation or over-compensation associated with the overlay error compensation for the dummy source die 924.

[0191] Bonding of the dummy source die 3824 may be performed (1) after bonding the set of dies 3024 to the target site KGD 824 and before bonding another set of source dies to other target sites on the target substrate 748 (hereinafter, “intermediate bonding”), or (2) after all target sites on the KGD 824 have been bonded with source KGDs from the multiple dies 724 (hereinafter, “final bonding”). Radiation data regarding the dummy source die 3824 and its corresponding target sites collected for intermediate bonding or final bonding can help determine if the overlay error changes over time from the time of bonding the dummy source die 924. The change may be due to vibration or other movement of any one or combination of components in the apparatus 110, a temperature change in the apparatus 110, or another change in environmental conditions in the apparatus 110.

[0192] The difference in overlay error between the dummy source die 924 and 3824 can provide other information in addition to or instead of the time information. Referring to FIG. 39, the center of the target substrate 748 is closer to the dummy source die 3824 than the dummy source die 924. Based on the overlay error of the dummy source die 924 and 3824, a variation in the overlay error from center to edge can be seen. As an example, the source KGD should be bonded to the target portion of the target die 3924. The target die 3924 is a specific KGD 824. Due to different overlay errors, the overlay error compensation can be predicted based on different positions of the target substrate 748.

[0193] Continuing with the particular example from above, dummy source die 924 may have an X-direction overlay error of -0.3 μm and a Y-direction overlay error of -0.6 μm, where positive values ​​in the X-direction are closer to the right side of FIG. 36 and positive values ​​in the Y-direction are closer to the top of FIG. 36. Dummy source die 3824 may have an X-direction overlay error of -0.1 μm and a Y-direction overlay error of -0.2 μm. Referring briefly to FIG. 37, active contact pads 3726 and 3728 may have the same overlay error as dummy source die 3824.

[0194] The distance from the center of the target feature of the target die 3924 to the center of the target substrate 748 can have a value half the value of the distance between the centers of the target features of the bad die 834 bonded to the dummy source die 924 and 3824. Therefore, the overlay error compensation for the source KGD for each direction is as follows:

[0195] OEC KGD =-((OE 924 + OE 3824 ) / 2), where

[0196] OEC KGD is the overlay error correction to the source KGD,

[0197] O.E. 924 is the overlay error of the dummy source die 924,

[0198] O.E. 3424 is the overlay error of the dummy source die 3824.

[0199] When the source KGD for the target die 3924 is moved, the overlay error compensation is in the opposite direction compared to the overlay error, and therefore the overlay error for the dummy source dies 924 and 3824 has a minus in front of it.

[0200] Using the above values, the overlay error compensation when positioning the source KGD relative to the target feature of the target die 3924 is:

[0201] For the X direction,

[0202] OEC KGD =-((-0.3μm-0.1μm) / 2), or +0.1μm, and

[0203] For the Y direction,

[0204] OECKGD =-((-0.6μm-0.2μm) / 2), or +0.2μm.

[0205] At least part of the positioning may include moving the source KGD 0.1 μm to the right of FIG. 39 (to compensate for overlay error in the X direction) and 0.2 μm to the top of FIG. 39 (to compensate for overlay error in the Y direction).

[0206] When the target substrate 748 is moved for source KGD, the movement is in the same direction compared to the overlay error. The overlay error compensation when moving the target substrate 748 is −0.1 μm in the X direction (in a top view of the target substrate 748, the target substrate 748 is moved 0.1 μm to the left) and −0.2 μm in the Y direction (in a top view of the destination substrate 748, the target substrate 748 is moved 0.2 μm downward).

[0207] Although the above examples use specific numerical values ​​to illustrate how overlay error can be used to determine overlay error compensation, they are provided for purposes of understanding and are not intended to limit the scope of the invention as defined by the claims.

[0208] After all the desired source dies (dummy source dies and source KGD) are bonded to the target substrate 748, the hybrid bonding process can continue. The hybrid bonding process can include three steps including a bonding operation, a first anneal to expand and contact the metals with each other in the die and at the target site, and a second anneal to cause the metal atoms to cross the metal-metal interface and reduce the contact resistance. The above-mentioned methods can correspond to the bonding operation of the hybrid bonding process. The target substrate 748 can be removed from the apparatus 110 or moved to a different part of the apparatus 110 or a different tool to perform the anneal operation.

[0209] The methods described with respect to the process flow charts of Figures 5, 6, 27, 28 need not be performed in the order shown in those figures. For example, loading the destination substrate in block 522 may be performed before or after mounting the source substrate in block 524. Furthermore, loading the destination substrate in block 522 and loading the source substrate in block 524 may be performed simultaneously with at least some of the loading operations. In another implementation, loading the source substrate in block 524 may be performed after bonding the dummy source die 924 or both the dummy source die 924 and 3824 to the destination substrate 748. Similar to blocks 522 and 524 of Figure 5, the operations described with respect to blocks 2742 (changing pitch for the die carrier sheet) and 2744 (moving the carriage) of Figure 27 may be performed simultaneously with at least some of the operations for blocks 2742 and 2744, or may be performed in either order. After reading this specification, one of ordinary skill in the art can determine a process flow that meets the needs or desires of a particular application.

[0210] Many of the implementations described herein can help determine overlay error for a die and a target feature of a target substrate. A dummy source die can be bonded to a target feature of a target substrate. The dummy source die does not have any electrical circuit elements that are electrically connected to electrical circuit elements in the target substrate. Thus, the dummy source die may not have interconnects or other features that interfere when collecting radiation data after the dummy source die is bonded to the target feature. In one implementation, an alignment mark in a metrology pattern of a dummy source die may reflect substantially more radiation compared to other portions of the metrology pattern or transmit substantially less radiation compared to other portions of the metrology pattern.

[0211] The radiation can be collected and analyzed. The analysis may or may not include generating an image. When the image is generated, further refinement may or may not be used to improve the image. The imaging can be focused on the alignment marks in the dummy source die and the target die, and can be performed with a focus range that can generate an image in which the electrical circuit elements and interconnects in the target die are out of focus or blurred. Pattern recognition can be used so that the electrical circuit elements and interconnects under the contact pads in the target die are not taken into account when determining the overlay error. Additionally, spatial frequency analysis can be used when imaging or when determining the overlay error.

[0212] The metrology patterns of the dummy source die and its corresponding target features may be significantly out of phase to produce a more accurate determination of the overlay error. With reference to FIG. 26, the alignment marks 944 and 1744 do not overlap and are approximately 180° out of phase in the same or different implementations. Compare the metrology patterns with the bonded contact pads 3724 in FIG. 37. If the metrology patterns were designed to maximize the overlap between the alignment marks between the metrology patterns, as with the bonded active contact pads shown in FIG. 37, it would be substantially more difficult to determine the overlay error. In many cases, the contact pads are made of the same material, such as Cu or Al-1 wt% Cu. The periphery of the contact pad 3726 that is over the contact pad 3728 can be very difficult to detect. Thus, an out-of-phase arrangement, as shown in FIG. 26, allows the overlay error to be determined more accurately.

[0213] The resulting overlay error for the dummy source die and its corresponding target feature may include an X-direction overlay error, a Y-direction overlay error, a rotational overlay error, or any combination thereof. Overlay error compensation is used for a subsequent source die, such as a source KGD, that is bonded to a target substrate, and the overlay error compensation helps to reduce the overlay error of the subsequent die. Overlay error compensation can increase the overlap between the active contact pads of the source KGD to the active contact pads at the target feature of the target KGD on the target substrate. The increased overlap between the active contact pads helps to reduce the contact resistance.

[0214] Two or more dummy source dies can be bonded to a target substrate to provide information about drift of the device 110 and to determine whether the overlay error changes depending on the position at which the source die is bonded to the target portion of the target substrate (e.g., center-to-edge variation relative to the target substrate).

[0215] It should be noted that not all of the operations described in the general description and examples above are required, some of the specific operations may not be required, and one or more additional operations may be performed in addition to the operations described. Furthermore, the order in which the operations are listed is not necessarily the order in which they are performed.

[0216] Benefits, other advantages, and solutions to problems have been described above with respect to particular embodiments. However, the benefits, advantages, solutions to problems, and any features that may give rise to or make more pronounced any benefit, advantage, or solution should not be construed as required, necessary, or essential features of any or all of the claims.

[0217] The specification and illustrations of the embodiments described herein are intended to give a general understanding of the structure of the various implementations. The specification and illustrations are not intended to exhaustively and comprehensively describe all of the elements and features of the apparatus and systems that use the structures or methods described herein. Separate implementations may also be provided in combination in a single implementation, and conversely, for brevity, various features described in the context of a single implementation may also be provided separately or in any subcombination. Furthermore, references to values ​​described in ranges include each and every value within that range. Many other implementations may become apparent to those skilled in the art only after reading this specification. Other implementations may be used and derived from this disclosure, such that structural substitutions, logical substitutions, or other changes may be made without departing from the scope of this disclosure. Thus, this disclosure should be considered illustrative and not limiting.

Claims

1. bonding a first dummy source die to a first target portion of a target substrate, where the first dummy source die has a first dummy source measurement pattern; the first target portion has a first target measurement pattern; acquiring first radiation data relating to the first dummy source measurement pattern and at least a portion of the first target measurement pattern in a first radiation region, where the acquiring of the first radiation data is performed after bonding the first dummy source die to the first target feature; analyzing the first radiation data to determine a first overlay error between the first dummy source die and the first target feature; adjusting a position of a known good source die or a second target feature of the target substrate to compensate for a first overlay error between the first dummy source die and the first target feature; The method according to claim 1, further comprising:

2. irradiating the first radiation region with radiation; in the first radiation region, the first dummy source measurement pattern includes a portion that is relatively transparent to the radiation and a portion that is relatively non-transparent to the radiation; the relatively transmissive portion of a first dummy source measurement pattern includes a substrate, the relatively non-transmissive portion of the first dummy source measurement pattern includes a substrate and dummy source alignment marks, at least one of the dummy source alignment marks being disposed along a bonding surface of the first dummy source die; the dummy source alignment marks are only within the relatively non-transparent portions, and dummy source alignment marks of the first dummy source measurement pattern are not within the relatively transparent portions.

2. The method of claim 1 .

3. the first target measurement pattern includes a target alignment mark; 3. The method of claim 2, wherein during irradiation, no more than 25% of the radiation is transmitted through the target alignment mark.

4. 4. The method of claim 3, wherein during bonding, at least one of the dummy source alignment marks does not contact any of the target alignment marks.

5. irradiating the first radiation region with radiation; the first radiation region includes at least a portion of each of the first dummy source measurement pattern and the first target measurement pattern, and within the first radiation region, the first dummy source measurement pattern includes a portion that is relatively transparent to the radiation and a portion that is relatively non-transparent to the radiation; a difference in transmittance between the relatively transparent portion and the relatively non-transparent portion with respect to the radiation is at least 25%; 2. The method of claim 1 .

6. The method further includes irradiating the first radiation region with radiation; Within the first radiation region, the first dummy source measurement pattern includes a portion that is relatively transparent to the radiation and a portion that is relatively non-transparent to the radiation; during irradiation, no more than 25% of the radiation is transmitted through the relatively non-transparent portions, and at least 75% of the radiation is transmitted through all of the relatively transparent portions; 2. The method of claim 1 .

7. 2. The method of claim 1 further comprising irradiating the first radiation area with infrared radiation.

8. 2. The method of claim 1, wherein the step of acquiring the first radiation data is performed using a radiation and a radiation detector, the radiation detector having a focal depth of up to 9 μm for the radiation.

9. the target substrate includes known good target die and known bad target die; the first target portion is at least a portion of the defective target die; 2. The method of claim 1 .

10. 10. The method of claim 9, further comprising bonding the known good source die to the second target feature of the target substrate after adjusting a position of the known good source die or the second target feature, the second target feature being at least a portion of the known good target die.

11. bonding a second dummy source die to a third target feature of the target substrate; the second dummy source die has a second dummy source measurement pattern; the third target feature die has a second target metrology pattern; acquiring second radiation data of at least a portion of the second dummy source measurement pattern and the second dummy source measurement pattern in a second radiation region; analyzing the second radiation data to determine a second overlay error between the second dummy source die and the third target feature; having 10. The method of claim 9, wherein the steps of bonding a second dummy source die, collecting second radiation data, and analyzing the second radiation data are performed after adjusting a position of a known good source die or the second target feature.

12. 2. The method of claim 1, wherein the first dummy source metrology pattern includes a first dummy source alignment mark.

13. The method of claim 12 , wherein the first target metrology pattern includes a first target alignment mark.

14. 13. The method of claim 12, wherein the first target metrology pattern includes a set of contact pads that are not part of an alignment mark.

15. 2. The method of claim 1 , wherein the first dummy source metrology pattern and the first target metrology pattern have areas that include arrays of corresponding contact pads.

16. 16. The method of claim 15, wherein the array of contact pads of the first dummy source metrology pattern is phase shifted relative to the array of contact pads of the first target metrology pattern by an angle in the range of 90° to 270°.

17. 16. The method of claim 15, wherein the array of contact pads of the first dummy source metrology pattern is phase shifted relative to the array of contact pads of the first target metrology pattern by an angle in the range of 150° to 210°.

18. The method of claim 1 , wherein the analyzing step comprises analyzing the first radiation data using spatial frequency analysis.

19. each of the first dummy source measurement pattern, the first target measurement pattern, or the first dummy source measurement pattern and the first target measurement pattern has a pitch and a feature size corresponding to the pitch; the first dummy source measurement pattern, the first target measurement pattern, or each of the first dummy source measurement pattern and the first target measurement pattern has a pitch and a feature size corresponding to the pitch, which corresponds to a spatial frequency value; 20. The method of claim 18, wherein a quotient is the pitch divided by the size of the feature corresponding to the pitch, the quotient being + / - 10% of the quotient and within the range of 1.5 to 9.

0.

20. 2. The method of claim 1, wherein analyzing the first radiation data includes analyzing a pattern in a first image generated from the first radiation data in comparison with a pattern in a standard image.

Images