CN114284199A - Apparatus and method for enhanced microelectronic device handling - Google Patents

Abstract

The present application relates to apparatus and methods for enhanced microelectronic device handling. The apparatus includes: a bonding head; a bonding tip coupled to the bonding head and having a bonding tip surface configured to receive a microelectronic device thereon; a first 3D sensor carried by the bond head; a pick-up arm drive; a pick arm coupled to the pick arm drive and having a pick surface configured to receive the microelectronic device thereon; and a second 3D sensor drivingly carried by the pick arm. The apparatus further includes a controller configured to: receiving first image data from the first 3D sensor, the first image data including image data of the pickup surface of the pickup arm; and receiving second image data from the second 3D sensor, the second image data comprising image data of the engaging tip surface of the engaging tip.

Description

Apparatus and method for enhanced microelectronic device handling

Priority claim

The present application claims benefit of the filing date of U.S. provisional patent application serial No. 63/086,268 entitled "APPARATUS AND method FOR ENHANCED MICROELECTRONIC device handling (APPARATUS AND METHODS FOR improved MICROELECTRONIC device DEVICE HANDLING)" filed on month 1 of 2020.

Technical Field

Embodiments disclosed herein relate to apparatus and methods for enhanced handling of microelectronic devices. More particularly, embodiments disclosed herein relate to methods and apparatus for reducing the likelihood that microelectronic devices will be damaged during physical manipulation of such devices.

Background

As the performance of electronic devices and systems increases, there is an associated need to improve the performance of the microelectronic components of such systems while maintaining or even reducing the form factor (e.g., length, width, and height) of the microelectronic device or assembly. Such requirements are typically, but not exclusively, associated with mobile devices and high performance devices. In order to maintain or reduce the footprint and height of component assemblies, such as semiconductor dies, in the form of microelectronic devices, three-dimensional (3D) assemblies of stacked components equipped with so-called through-silicon vias (TSVs) for vertical electrical (e.g., signal, power, ground/bias) communication between the stacked components are becoming more common, incorporating component thickness reduction and employing pre-formed dielectric films in bond wires, such as the spaces between stacked components, to reduce bond wire thickness while increasing bond wire uniformity. Such dielectric films include, for example, so-called non-conductive films (NCFs) and Wafer Level Underfill (WLUF), such terms being often used interchangeably. While effective in reducing the height of 3D microelectronic device assemblies, the reduction of microelectronic device (e.g., semiconductor die) thickness to about 50 μm or less increases the fragility and susceptibility of the device to fracture under stresses, particularly compressive (e.g., impact) and bending stresses experienced during handling, such as during pick and place operations. Reducing bond wire thickness can also exacerbate the susceptibility of damage to such very thin microelectronic devices because the thin dielectric material (e.g., NCF) in the bond wires may no longer provide any cushioning effect or the ability to accommodate particulate contamination in the bond wires when, for example, the device is stacked on another device to form a 3D assembly. Non-limiting examples of microelectronic device assemblies including stacked microelectronic devices that may be subject to stress-induced cracking include semiconductor memory die assemblies, including so-called high bandwidth memory (HBMx), Hybrid Memory Cubes (HMC), and chip-to-wafer (C2W) assemblies, alone or in combination with other die functionality, such as logic.

Disclosure of Invention

Embodiments of the present disclosure include an apparatus for handling microelectronic devices. The apparatus includes: a bonding head; a bonding tip coupled to the bonding head and having a bonding tip surface configured to receive a microelectronic device thereon; a first 3D sensor carried by the bond head; a pick-up arm drive; a pick arm coupled to the pick arm drive and having a pick surface configured to receive the microelectronic device thereon; and a second 3D sensor drivingly carried by the pick arm. The apparatus further includes a controller configured to: receiving first image data from the first 3D sensor, the first image data including image data of the pickup surface of the pickup arm; and receiving second image data from the second 3D sensor, the second image data comprising image data of the engaging tip surface of the engaging tip.

Embodiments of the present disclosure include a method comprising: moving a pick arm to a first position relative to an engagement tip, wherein a pick surface of the pick arm is within a field of view of a first 3D sensor coupled to the engagement tip; capturing image data of the pickup surface via the first 3D sensor; and analyzing the image data of the pickup surface.

Embodiments of the present disclosure include an apparatus for handling microelectronic devices. The apparatus includes: a bonding head; a bonding tip coupled to the bonding head and having a bonding tip surface configured to receive a microelectronic device thereon; a pick-up arm drive; a pick arm coupled to the pick arm drive and having a pick surface configured to receive the microelectronic device thereon; and at least one 3D sensor configured to capture image data of the engagement tip surface and image data of the pickup surface.

Drawings

Fig. 1 shows a schematic diagram of an example pick and place apparatus incorporating a thermal compression bonding apparatus depicting microelectronic devices being removed from a dicing tape and transferred to a bonding tip of a bonding head for stacking on a substrate, in accordance with an embodiment of the present disclosure;

figure 2 is a photomicrograph showing a plurality of stacked semiconductor dies having cracks;

FIG. 3 is an enlarged micrograph showing cracks invaded by underfill material in the bond wire;

4A-4F schematically depict an example pick operation for removing semiconductor dies from a dicing tape using a pick arm and ejector of a pick and place apparatus;

fig. 5A and 5B schematically depict an example semiconductor die transfer operation from a pick arm to a bonding tip of a bonding head;

fig. 6A and 6B schematically depict an inspection system of a pick and transfer apparatus according to one or more embodiments of the present disclosure;

fig. 7 schematically depicts an inspection system of a pick and transfer apparatus in accordance with one or more embodiments of the present disclosure;

FIGS. 8A and 8B include a flow chart of a method of inspecting a pick surface of a pick arm and an engagement tip surface of an engagement tip in accordance with an embodiment of the present disclosure;

fig. 9 schematically depicts an inspection system of a pick and transfer apparatus in accordance with one or more embodiments of the present disclosure; and

fig. 10A and 10B include a flow chart of a method of inspecting an active surface of a semiconductor die according to an embodiment of the present disclosure.

Detailed Description

Embodiments of the present disclosure relate to methods and systems that enhance handling of microelectronic devices by reducing the magnitude and inconsistent application of stresses applied to such microelectronic devices during handling, such as during pick and place operations involving removal of microelectronic devices from a group of such devices by a pick arm and transfer of the removed devices to bonding tips of a bonding head for placement on a substrate or stacking with other devices.

As used herein, the terms "comprising," "including," "containing," "characterized by," and grammatical equivalents thereof are inclusive or open-ended terms that do not exclude additional, unrecited elements or method acts, and include the more limiting terms "consisting of …" and "consisting essentially of …," and grammatical equivalents thereof.

As used herein, the term "may" with respect to a material, structure, feature, or method act indicates that such is contemplated for implementing an embodiment of the present disclosure, and such term is used in preference to the more limiting term "is" to avoid or necessarily preclude any implication of other compatible materials, structures, features, and methods that may be used in combination therewith.

As used herein, the terms "longitudinal," "vertical," "lateral," and "horizontal" are with reference to a major plane of a substrate (e.g., base material, base structure, base construction, etc.) in or on which one or more structures and/or features are formed and are not necessarily defined by the earth's gravitational field. A "lateral" or "horizontal" direction is a direction substantially parallel to the major plane of the substrate, and a "longitudinal" or "vertical" direction is a direction substantially perpendicular to the major plane of the substrate. The major plane of the substrate is defined by the surface of the substrate having a relatively larger area than the other surfaces of the substrate.

As used herein, spatially relative terms (e.g., "below," "lower," "bottom," "above," "over," "upper," "top," "front," "back," "left," "right," and the like) may be used for ease of description to describe one element or feature's relationship to another element(s) or feature(s), as illustrated in the figures. Unless otherwise specified, spatially relative terms are intended to encompass different orientations of the material in addition to the orientation depicted in the figures. For example, if the materials in the figures are reversed, elements described as "below" or "above" or "upper" or "on top of" other elements or features would then be oriented "below" or "beneath" or "lower" or "on top of the other elements or features. Thus, those of ordinary skill in the art will appreciate that the term "above" can encompass both an orientation of above and below, depending on the context in which the term is used. The materials may be otherwise oriented (e.g., rotated 90 degrees, inverted, flipped) and the spatially relative descriptors used herein interpreted accordingly.

As used herein, the singular forms "a" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

As used herein, the terms "configured" and "configuration" refer to the size, shape, material composition, orientation, and arrangement of one or more of at least one structure and one or more of at least one apparatus that facilitates the operation of the structure and the apparatus in a predetermined manner.

As used herein, the term "substantially" with respect to a given parameter, property, or condition means and includes the degree to which the given parameter, property, or condition is satisfied to within a degree of variation (e.g., within acceptable manufacturing tolerances) as understood by one of ordinary skill in the art. By way of example, depending on the particular parameter, property, or condition being substantially met, the parameter, property, or condition may meet at least 90.0%, meet at least 95.0%, meet at least 99.0%, or even meet at least 99.9%.

As used herein, "about" or "approximately" with respect to a value of a particular parameter includes the value and the degree of change relative to the value within an acceptable tolerance for the particular parameter as understood by one of ordinary skill in the art. For example, "about" or "approximately" with respect to a numerical value may include additional numerical values in a range from 90.0% to 110.0% of the numerical value, such as in a range from 95.0% to 105.0% of the numerical value, in a range from 97.5% to 102.5% of the numerical value, in a range from 99.0% to 101.0% of the numerical value, in a range from 99.5% to 100.5% of the numerical value, or in a range from 99.9% to 100.1% of the numerical value.

As used herein, the terms "layer" and "film" mean and include a layer, sheet, or coating of a material that resides on a structure, which may be continuous or discontinuous between portions of the material, and which may be conformal or non-conformal, unless otherwise indicated.

As used herein, the term "substrate" means and includes a base material or construction on which additional materials are formed. The substrate may be a semiconductor substrate, a base semiconductor layer on a support structure, a metal electrode, a semiconductor substrate having one or more materials, layers, structures, or regions formed thereon. Materials on a semiconductor substrate may include, but are not limited to, semiconductive materials, insulative materials, conductive materials, and the like. The substrate may be a conventional silicon substrate or other bulk substrate comprising a layer of semiconductive material. As used herein, the term "bulk substrate" means and includes not only silicon wafers, but also silicon-on-insulator ("SOI") substrates, such as silicon-on-sapphire ("SOS") substrates and silicon-on-glass ("SOG") substrates, epitaxial layers of silicon on a base semiconductor foundation, and other semiconductor or optoelectronic materials such as silicon germanium, gallium arsenide, gallium nitride, and indium phosphide. The substrate may be doped or undoped. The term "substrate" also means and includes an organic substrate, such as a substrate having a plurality of metal layers in the form of traces, and interposed with a dielectric layer (e.g., a plexiglas woven polymer). For example, a conventional BGA package includes a plurality of dies and encapsulation (e.g., Epoxy Molding Compound (EMC)) on one side of an organic substrate and an array of solder balls on the other side.

As used herein, the term "microelectronic device" means and includes, by way of non-limiting example, semiconductor dies, dies exhibiting functionality through non-semiconducting activity, microelectromechanical systems (MEM) devices, substrates including a plurality of dies including conventional wafers, and other bulk substrates and partial wafers described above and substrate fragments including more than one die location.

As used herein, the term "memory device" means and includes (by way of non-limiting example) semiconductors and other microelectronic devices that exhibit memory functionality but do not exclude other functionality, unless the context in which the term is used clearly indicates otherwise. In other words and by way of example only, the term "memory device" means and includes not only conventional memory in the form of DRAM, NAND and the like, but also, by way of example only, an Application Specific Integrated Circuit (ASIC), such as a system on a chip (SoC), a microelectronic device combining logic and memory, or a Graphics Processing Unit (GPU) incorporating memory.

As used herein, unless explicitly stated otherwise, the terms "metal" and "metallic material" mean and include elemental metals, metal alloys, and combinations (e.g., layers) of different and adjacent metals or metal alloys.

The description herein provides specific details such as size, shape, material composition, location, and orientation to provide a thorough description of embodiments of the disclosure. However, it will be understood and appreciated by those of ordinary skill in the art that embodiments of the present disclosure may not necessarily be practiced with these specific details, as embodiments of the present disclosure may be practiced in conjunction with conventional process actions and equipment used in the industry, with appropriate modification in light of the present disclosure. Additionally, the description provided below may not form a complete process flow. Only those process acts and structures necessary to understand the embodiments of the present disclosure are described in detail below.

The drawings presented herein are for illustration only and are not meant to be actual drawings of any particular material, component, structure, device, or system. Variations in the shapes of the illustrations as a result, for example, of manufacturing techniques and/or tolerances, are to be expected. Thus, the embodiments described herein should not be construed as limited to the particular shapes or regions illustrated, but are to include deviations in shapes that result, for example, from manufacturing. For example, a region illustrated or described as a frame may have rough and/or non-linear features, and a region illustrated or described as a circle may include some rough and/or linear features. Furthermore, acute angles between the illustrated surfaces may be rounded off, and vice versa. Thus, the regions illustrated in the figures are schematic and their shapes are not intended to illustrate the precise shape of a region and are not intended to limit the scope of the claims. The drawings are not necessarily to scale.

Embodiments may be described in terms of procedures described as flow charts (flow charts), structure diagrams, or block diagrams. Although a flowchart may describe the operational acts as a sequential process, many of these acts can be performed in another sequence, in parallel, or substantially concurrently. Additionally, the order of the actions may be rearranged. A process may correspond to a method, a thread, a function, a procedure, a subroutine, a subprogram, other structure, or a combination thereof. Further, the methods disclosed herein may be implemented in hardware, software, or both. If implemented in software, the functions may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. In the description and for convenience, the same or similar reference numerals may be used to identify common features and elements between the various drawings.

By way of background, microelectronic devices (e.g., semiconductor dies) may experience cracking when a pick arm of a pick and place apparatus contacts a die picked from a dielectric film (e.g., dicing tape) supported by a film frame because an ejector below the film carrying the die moves the die upward to meet the pick arm. In addition, when a picked die is transferred by the pick arm to a receiving assembly (e.g., a bonding tip of a bonding head used in a thermal compression bonding apparatus), contact on the die due to the movement of the pick arm toward the bonding tip can cause cracking. In either case, the impact may be caused by an incorrect over travel of the pick arm toward the die or the bonding tip or an angular misalignment (e.g., non-coplanar, non-parallel orientation) of the pick surface of the pick arm relative to the die surface or the bonding tip surface, such that the two surfaces are non-parallel as the pick arm approaches the die or the bonding tip. The magnitude of the pick surface contacting the die or the die contacting the bonding tip surface may initiate cracking if over travel occurs. If non-coplanarity occurs, point contact of the edge of the pick arm with the die surface or the die surface with the bonding tip surface causes an increase in force on the die by orders of magnitude, thereby initiating cracking. Additionally, if any of the pick-up surface, die surface, and/or bonding tip surface have protruding defects (e.g., microbumps, micro-ridges), contact of the defect with the die surface, pick-up surface, or bonding tip surface causes the force per unit area on the die at the defect to increase by orders of magnitude, causing cracking. Furthermore, contamination (e.g., particles) can inadvertently enter between the die surface and the pick surface or between the die surface and the bonding tip surface during the pick and place process, and can cause the force per unit area on the die at the contamination to increase by orders of magnitude and can induce cracking.

To better visualize the above-mentioned problems, fig. 1 schematically illustrates an example of a pick and place apparatus 100 in combination with an example of a thermal compression bonding apparatus 200. Additionally, fig. 2 is a photomicrograph showing a plurality of stacked semiconductor dies having cracks, and fig. 3 is an enlarged photomicrograph showing cracks invaded by underfill material in the bond wires. Referring concurrently to fig. 1-3, a microelectronic device (e.g., semiconductor die) S is depicted being removed from a singulated semiconductor wafer W on a dicing tape 102 with an active surface a of the semiconductor die facing upward and conductive elements (not shown) protruding from the active surface a in a non-conductive film (NCF), wherein cracks depicted in fig. 2 and 3 may occur. Next, the bonding tips 204 of the bonding heads 202 transferred to the thermal compression bonding apparatus 200 are implemented to be re-oriented and stacked on a substrate (e.g., a target semiconductor wafer including an array of die locations) in a "flip-chip" fashion, with the active surface a of the semiconductor die S facing downward toward the substrate. The pick and place apparatus 100 includes a pick arm 104 that is movable in X, Y and Z directions and rotatable about a transverse axis LA and about a longitudinal axis LO under the control of a programmed controller 122 including one or more microprocessors 124 by linear encoder equipped drive motors 120X, 120Y, 120Z and rotary encoder equipped drive motors 120LA and 120LO, the microprocessors 124 being in communication with a memory 126 storing an operating program and in closed loop communication with an optical sensor system 128 to align the pick arm 104 with a semiconductor die S to be removed from the dicing tape 102 by the pick arm 104 for transfer to the bonding tip surface 206. Suppliers of such equipment include, but are not limited to, ASM International of Almere, The Netherlands, Japan, Shinkawa, ltd.

As shown in fig. 1, semiconductor dies S singulated from a semiconductor wafer W are supported and adhered to a dicing tape 102. Typically, the dicing tape 102 (which may be a polymer film coated with UV-release glue and peripherally supported by a so-called film frame 103) supports the semiconductor wafer W during a so-called "singulation" operation, wherein individual semiconductor die locations on the wafer W are separated by a diamond coated dicing saw, after which the dicing tape 102 is stretched to separate singulated semiconductor dies S for removal from the dicing tape 102. At this point, the semiconductor die S is ready to be picked from the dicing tape 102, and the pick arm 104 of the pick and place apparatus 100 is suspended over the location of the semiconductor die S and optically aligned with the location of the semiconductor die S using the optical sensor system 128. At this point, the pick arm 104 has been rapidly moved to a position vertically above and in the plane of the lateral direction X, Y aligned with the semiconductor die S having the pick surface 106 parallel to the active surface a. Once aligned over the semiconductor die S, the pick arm 104 is quickly lowered vertically until a predetermined pre-programmed phase separation distance of, for example, about 100 μm and up to about 500 μm is reached between the pick surface 106 of the pick arm 104 and the active surface a of the semiconductor die S, after which the travel of the pick arm 104 is slowed significantly to implement a "soft contact" travel to contact the NCF over the active surface a. Between the time that the pick arm 104 slows and the time that the active surface a is in contact, the ejector 108 moves upward against the dicing tape 102 (as shown by vertical arrow E) in a synchronous upward movement with the pick arm 104 and presents the semiconductor die S to the pick surface 106. The pick arm 104 and pick surface 106 are equipped with a vacuum port 110 that is in selective communication with a vacuum source 110VS and actuated to pull the semiconductor die S up away from the dicing tape 102. Ideally, the pick arm 104 and the ejector 108 move in unison to minimize (e.g., substantially eliminate) the contact force between the pick surface 106 and the ejector 108 while substantially preventing any gap between the pick arm 104 and the NCF above the active surface a. For example, ideally, due to synchronization, the contact force of the pick arm 104 is minimized to no more than about 50 to 150 grams of contact force at most, and may be expected to be much smaller.

However, as discussed herein, due to the pick arm 104 and/or the ejector 108 being offset from the incorrect or mismatched encoder values of the calibration or control movement, the pick arm 104 may over-travel and the pick surface 106 may thus over-press the semiconductor die S, causing stress microcracks or even cracks, such as the cracks depicted in fig. 2 and 3, by at least one of the applied impact and the excessive force. Regardless of whether over travel occurs, the semiconductor die S is then removed for further handling by the pick arm 104 and transferred to the bonding tip surface 206 of the bonding tip 204 for arrangement on a substrate or another semiconductor die S, as discussed below.

Similarly, if the pickup surface 106 is not parallel to the active surface a or rotationally misaligned with the semiconductor die S, edge contact of the pickup surface 106 with the active surface a may occur. Non-coplanarity of the pick-up surface 106 relative to the active surface a as small as about 20 μm (referred to herein as "non-coplanarity") can result in damage to the semiconductor die S. In view of the foregoing, the term "non-coplanarity" may indicate that two object surfaces are not parallel to each other and/or that one surface is rotationally misaligned relative to the other.

Still referring to fig. 1, an example transfer operation from the pick arm 104 to the bonding tip 204 of the bonding head 202 of the thermal compression bonding apparatus 200 is illustrated in schematic form. In fig. 1, pick arm 104, which carries semiconductor die S through active surface a, has been lifted, moved in X, Y and Z directions, and rotated about lateral axis LA and longitudinal axis LO as needed to present a backside surface B of semiconductor die S to bonding tip surface 206 of bonding tip 204 of bonding head 202 of thermocompression bonding apparatus 200. As with the pick-up operation, the pick-up arm 104 is moved quickly until a pre-programmed separation distance of, for example, about 100 μm is reached between the backside surface B and the bonding tip surface 206, after which the pick-up arm 104 is moved relatively slowly toward the bonding tip 204, but not in contact with the bonding tip 204, the bonding tip 204 being at an elevated temperature provided by the resistive heater 208. The remaining standoff distance substantially isolates the heat of the semiconductor die S and the bonding tip 204, which would otherwise cause the NCF to become sticky and stick to the pick-up surface 106 or begin to cure prematurely before being stacked on the target substrate or another semiconductor die, thereby compromising bond wire integrity.

In addition, even if the apparatus is properly calibrated, variations in coplanarity of the pick-up surface 106 and the bonding tip surface 206 in terms of the vertical distance D can cause unintended spaced contact of the edge of the backside surface B of the semiconductor die S with the bonding tip surface 206. When the backside surface B is moved into relatively close proximity (e.g., less than about 150 μm to about 200 μm) to the bonding tip surface 206, the vacuum ports 210 within the bonding tips 204 are actuated and the vacuum ports 110 in the pick-up surface 106 are de-actuated to contactlessly transfer the semiconductor die S through its backside B to the bonding tip surface 206 in response to the air pressure differential. In some examples, the vacuum ports 110 in the pick surface 106 are inverted to generate a relatively small positive pressure in a cleaning sequence, and the semiconductor die S "expands" against the bonding tip surface 206 for contactless transfer. In an ideal case, the pick-up surface 106 has been properly calibrated to be parallel to the bonding tip surface 206 at a desired distance when manipulated close to the bonding head 202, so that transfer of the semiconductor die S is worst completed without any contact force distribution over the backside surface B of the semiconductor die S. However, during repeated use, the movement of the pick arm 104, the bonding tip 204, or both may deviate from calibration, and as a result, the pick surface 106, and thus the backside surface B of the semiconductor die S, is presented at an acute angle relative to the bonding tip surface 206, resulting in edge contact of the backside surface B with the bonding tip surface 206, resulting in edge cracking on the backside surface B of the semiconductor die S, such as the cracks depicted in fig. 2 and 3.

While die micro-cracking and cracking have generally been a problem, as noted above, the problem is exacerbated by the constant reduction in die and bond wire thickness. Significant examples of die cracking with respect to yield reduction due to handling issues become evident when the die thickness reaches about 60 μm to about 65 μm, the number and severity further increases when the die thickness reaches about 50 μm, and is expected to be further exacerbated when the die thickness reaches about 30 μm or less in response to industry requirements for highly stacking more and more microelectronic devices in a given form factor.

Referring still generally to fig. 1-3, to further aid the reader in understanding embodiments of the present disclosure, in practice, a dielectric film in the form of an NCF is adhered to and onto the active surface of a bulk semiconductor substrate in the form of generally a wafer (e.g., a silicon wafer on which integrated circuit systems have been fabricated). The wafer is then singulated, such as by a diamond coated dicing blade, along so-called "street" between adjacent semiconductor die locations on the wafer while supported on dicing tape supported on a film frame to provide individual semiconductor dies each having an NCF on their active surfaces. Even though the NCF may be laminated to the protective film during shipping and handling, once the NCF is laminated to the active surface of the wafer, the protective film peels off before singulation of the wafer into individual semiconductor dies occurs, exposing the upper, now uncovered, exposed surface of the NCF to contamination during singulation and during subsequent die handling from adjacent dies from which residual contaminants on the singulated NCF may fall.

The particles generated during the singulation process may be inorganic (e.g., silicon chips) or organic (e.g., NCF residues, dicing tape residues or particulates from other sources within the clean room environment). Silicon particles can, for example, cause the die to crack when the particle size exceeds the bond wire thickness, while organic particles can, when located on an Under Bump Metallization (UBM) conductive element, such as a solder-covered conductive pillar or solder bump, cause the solder not to wet, compromising electrical communication between stacked dies.

It has been determined that particulate contamination on the exposed NCF surface causes significant reduction in die yield, particularly when a blade cutting (e.g., singulation) process is employed, resulting in a large amount of particulate debris. Significantly, die yield due to NCF contamination gradually worsens during processing from the beginning of wafer detachment from the carrier wafer, after wafer (e.g., post) notching, after wafer lamination to dicing tape supported on a film frame, and during post dicing. If a 60 μm contaminant particle size is used as a baseline for determining damaged dies (e.g., die failure rate), the yield gradually decreases from almost 100% after detachment to about 90% after dicing, with about half of the defective dies of the wafer being damaged by silicon (e.g., solid) particles and about half being damaged by organic (e.g., transparent) particles. However, if a 20 μm contaminant particle size is used as a baseline, the yield drops abruptly from over 95% after detachment to less than 75% after dicing, again with about half of the defective die of the wafer being damaged by silicon (e.g., solid) particles and about half being damaged by organic (e.g., transparent) particles. Since about 15 μm thick NCFs are common and about 10 μm thick NCFs are contemplated, it is readily appreciated that even minute contaminant particles of about 15 μm or less in size can significantly increase die failure rates. In addition, as the industry moves toward so-called "zero bond wire thicknesses" of less than about 5 μm, the use of plasma treated silicon oxide or organic materials to join adjacent microelectronic devices electrically connected by stacked very thin (e.g., about 30 μm) aligned Cu-to-Cu conductive elements of adjacent microelectronic devices can significantly reduce the yield of these fragile devices, even if minute particles of contaminants are present on the surface of the microelectronic devices (e.g., the active surface of the semiconductor die).

The importance of contaminants to yield loss during pick and place operations can be characterized as facilitating stress concentration on the surface of the semiconductor die by substantially limiting contact of, for example, the surface of the pick arm with one or more contaminant particles of a size (e.g., diameter) greater than the thickness of the NCF and the height of the conductive elements (e.g., copper pillars protruding from and within the NCF on the active surface of the die). Thus, rather than spreading the pick-up arm contact force across the entire NCF and conductive elements and reducing the force per unit area on the die active surface, the entire force can be concentrated only on a few discrete points on the active surface on which contaminant particles reside and which protrude above the NCF.

Referring now to fig. 4A-4F, a conventional pick operation that can result in a fracture, such as the fractures depicted in fig. 2 and 3, is depicted in schematic form. For example, fig. 4A-4F depict a conventional pick operation using a pick arm and ejector of the pick and place apparatus 100 to remove a semiconductor die S configured with an active surface a having conductive elements (e.g., metal posts) from a dicing tape for stacking and thermal compression bonding.

Referring to fig. 4A, a microelectronic device in the form of a semiconductor die S is supported on and adhered to a dicing tape 102 of the pick and place apparatus 100. Typically, the dicing tape 102 supports singulated semiconductor dies S to be removed from the dicing tape 102. The pick arm 104 of the pick and place apparatus 100 is suspended over and optically aligned with the location of the semiconductor die S. As previously noted, the pick arm 104 is movable in the X, Y and Z directions, and is also rotatable about the lateral axis LA and the longitudinal axis LO under the control of a programmed controller by an encoder-equipped drive motor.

Once aligned over the semiconductor die S, the pick arm 104 is quickly lowered vertically as shown in fig. 4B until a predetermined preprogrammed separation distance (e.g., 100 μm as depicted) is reached between the pick surface 106 of the pick arm 104 and the active surface a of the semiconductor die S, after which the travel of the pick arm 104 is slowed significantly to implement a "soft contact" travel to contact the NCF over the active surface a. Between the time that the pick arm 104 slows and the time that the active surface a is in contact, the ejector 108 moves upward against the dicing tape 102 in synchronization with the pick arm 104 (as shown by vertical arrow E) and presents the semiconductor die S to the pick surface 106, the pick surface 106 being equipped with vacuum ports 110 that are actuated to pull the semiconductor die S up away from the dicing tape 102. Ideally, due to synchronization, the contact force of the pick arm 104 is minimized to no more than about 50 grams to 150 grams of contact force at most, and may be expected to be much smaller. However, due to the pick arm 104 and/or the ejector 108 being offset from the incorrect or mismatched encoder values of the calibration or control movement, as shown in fig. 4C, the pick arm 104 may overtravel and the pick surface 106 may thus over-press the semiconductor die S, causing stress microcracks and cracks from at least one of the applied impact and the excessive force, the damaged semiconductor die S then being removed from the dicing tape 102 by the pick arm 104 for further handling, as shown in fig. 4F.

Similarly, if the pickup surface 106 is not parallel to the active surface a or rotationally misaligned with the semiconductor die S, edge contact of the pickup surface 106 with the active surface a may occur, as shown in fig. 4D. A non-coplanarity of the pick-up surface 106 as small as about 20 μm relative to the active surface a may result in damage to the semiconductor die S.

In addition, the presence of contaminants in the form of inorganic or organic particles P from the singulation process on the active surface a (such as any of the above-described contaminants) or NCF residue from previous devices picking up on the pick-up surface 106 may cause damage forces to be concentrated on the active surface a of the semiconductor die S, as shown in fig. 4E. Likewise, the presence of defects D in the semiconductor die S exhibited at the active surface a and/or the presence of defects in the surface of the pick surface may result in damage forces being concentrated on the active surface a of the semiconductor die S, as shown in fig. 4E. Unfortunately, cracks and microcracks C are not readily detectable during semiconductor die handling, and their presence may often not be apparent prior to assembly with other semiconductor dies, which assembly and subsequent application of normal forces by the bond head during thermal compression bonding of the die stack can also exacerbate the microcracks from becoming cracked dies, as shown in fig. 2 and 3.

Referring now to fig. 5A and 5B, conventional transfer operations that can result in a rupture, such as the rupture depicted in fig. 2 and 3, are depicted in schematic form. For example, fig. 5A and 5B schematically depict conventional pick-up and transfer operations from the pick-up arm 104 to the bonding tip 204 of the bonding head 202 of the thermal compression bonding apparatus 200. As shown in fig. 5A, the pick arm 104 carrying the semiconductor die S through the active surface a has been moved in the X, Y and Z directions and rotated about the lateral and longitudinal axes as needed to present a backside surface B to the bonding tip surface 206 of the bonding tip 204 of the bonding head 202 of the thermocompression bonding apparatus 200. The pick arm 104 moves quickly until a pre-programmed separation distance after which the pick arm 104 moves more slowly toward the engagement tip 204 but not in contact with the engagement tip 204. When the backside surface B is in close proximity to the bonding tip surface 206, the vacuum port 210 of the bonding tip 204 is actuated and the vacuum port 110 in the pick arm 104 is de-actuated to contactlessly transfer the semiconductor die S to the bonding tip 204 in response to the pressure differential. Thus, the semiconductor die S is thermally isolated from the bonding tip 204 for as long as possible.

As previously noted, in some examples, the vacuum to the vacuum ports 110 in the bonding tip 204 and at the pick surface 106 may be reversed to provide a small positive pressure, and the semiconductor die S "expands" against the bonding tip surface 206 for contactless transfer. As shown in fig. 5A, in an ideal case, the pick-up surface 106 has been properly calibrated to be parallel to the bonding tip surface 206 when manipulated into proximity to the bonding head 202, so that transfer of the semiconductor die S is worst completed without any contact force distribution over the backside surface B of the semiconductor die S. However, as depicted in fig. 5B and mentioned above, during repeated use, the movement of the pick arm 104 can deviate from calibration, and as a result, the pick surface 106, and thus the back side surface of the semiconductor die S, is presented at an acute angle (e.g., "tilt") relative to the bonding tip surface 206, causing the back side surface B to edge contact EC with the bonding tip surface 206, causing edge cracking on the back side surface B. The edge contacts EC may be along a line when the pick surface 106 is rotationally oriented to match the engagement tip surface 206, or they may be point contacts that include corners of the backside surface B of the semiconductor die S when rotationally misaligned and tilted in a plane perpendicular to the longitudinal axis LO. It is found that a slight angular displacement D of about 80 μm of the edge of the backside surface B from parallel between the backside surface B and the bonding tip surface 206 can cause micro-cracking or fractures in the semiconductor die S from this edge contact. Similarly, as discussed above, contaminants in the form of inorganic or organic particles from the singulation process on the active surface a, defects in the semiconductor die S and/or defects in the surface of the bonding tip surface 206 exhibited at the active surface a, or the presence of NCF residue from previous devices picking up on the pick-up surface 106, may cause damage forces to be concentrated on the active surface a and edge contact on the backside surface B due to the non-parallel orientation of the backside surface B.

Embodiments of the present disclosure address the above-mentioned problems. For example, fig. 6A and 6B depict illustrations of an inspection system 603 including a bond head assembly 600 and a pick assembly 601, according to an embodiment of the present disclosure. Bond head assembly 600 may include a bond head 602, a bond tip 604 defining a bond tip surface 606, and a first three-dimensional (3D) sensor 608. The pick assembly 601 may include a pick arm 612 defining a pick surface 613, a pick arm drive 614, and a second 3D sensor 616.

Bond head assembly 600 may include any of the bond heads and bond tips described above. In some embodiments, the first 3D sensor 608 may be mounted to the bond head 602 or at least oriented proximate (e.g., adjacent) to the bond head 602. Additionally, the first 3D sensor 608 is operatively coupled to the controller 610. The pick assembly 601 may include any of the pick arms described above. Further, a second 3D sensor 616 may be mounted to the pick arm drive 614 or oriented at least proximate (e.g., adjacent) to the pick arm drive 614, the second 3D sensor 616 operably coupled to the controller 610.

In some embodiments, when the inspection system 603 is within a first position, as depicted in fig. 6A, the first 3D sensor 608 can be oriented such that the field of view 620 of the first 3D sensor 608 will include the pick surface 613 of the pick arm 612 and/or the back side of the semiconductor die during a transfer process (e.g., any of the transfer processes described above). Further, when the inspection system 603 is within the second position, as depicted in fig. 6B, the second 3D sensor 616 may be oriented such that the field of view 622 of the second 3D sensor 616 will include the engagement tip surface 606. In one or more embodiments, as described in more detail below, the controller 610 may receive image data from the first and second 3D sensors 608, 616 to identify defects, particles, and/or contamination on the surface and to identify non-coplanarity of the pick-up arm 612 relative to the bonding tip 604. Thus, the controller 610 may control the orientation of the pick arm 612 and the semiconductor die (e.g., semiconductor die S) and/or the bonding tip 604 using the image data to ensure that the semiconductor die is properly oriented during the transfer process (e.g., to avoid the aforementioned misalignment and edge contact).

In one or more embodiments, the first 3D sensor 608 may be offset from the engagement tip surface 606 by a distance D1 in a range of about 0.00mm to about 5.0mm in a direction orthogonal to the engagement tip surface 606. For example, the distance D1 may be about 0.5 mm. In some embodiments, D1 may be selected to ensure that the field of view 620 of the first 3D sensor 608 maintains a view of the pick surface 613 of the pick arm 612 and/or the back side of the semiconductor die during the transfer process. In view of the foregoing, the position of the engagement tip surface 606 relative to the first 3D sensor 608 may be known, and vice versa. The second 3D sensor 616 may be offset from the pickup surface 613 by a distance D2 in a range of about 0.00mm to about 5.0mm in a direction normal to the pickup surface 613. For example, the distance D2 may be about 0.5 mm. In some embodiments, D2 may be selected to ensure that the field of view 622 of the second 3D sensor 616 maintains a view of the engagement tip surface 606 of the engagement tip 604. Additionally, the distance D1 and the distance D2 may be selected to ensure that the first and second 3D sensors 608, 616 do not interfere within the pick and transfer process (e.g., do not contact and risk damaging the semiconductor die). In view of the foregoing, the position of the pickup surface 613 relative to the second 3D sensor 616 may be known, and vice versa. In one or more embodiments, the distance D1 and the distance D2 may be adjusted depending on the process performed by the bond head assembly 600 and the pick assembly 601 and/or the thickness of the semiconductor die as the bond head assembly 600 and the pick assembly 601 are manipulated.

In some embodiments, one or more of first 3D sensor 608 and second 3D sensor 616 may be configured to detect objects and surface topology in three dimensions. For example, one or more of the first 3D sensor 608 and the second 3D sensor 316 may include a laser system. In one or more embodiments, the laser system may include a high speed, high resolution laser sensor. For example, the laser system may be scanned, measured and controlled at a speed of about 5kHz, 7kHz or 10 kHz. Additionally, the laser system may include a three-dimensional profile sensor. In addition, the laser system may provide a resolution in the range of about 0.008 to about 0.054 mm. As a non-limiting example, the laser system may include one manufactured by LMI technology

2500 series lasers. As additional non-limiting examples, the laser system may include a Keyence laser profiler manufactured by Keyence Corporation of America, USA, and/or a CheckBox laser scanner manufactured by Automated Precision, Inc. In some embodiments, both first 3D sensor 608 and second 3D sensor 616 may include laser systems. In one or more embodiments, one or more of first 3D sensor 608 and second 3D sensor 616 may include a camera, an infrared sensor, a stereo camera, or other 3D sensor. In some embodiments, the first 3D sensor 608 may be different from the second 3D sensor 616.

As noted above, in some embodiments, the first 3D sensor 608 may be mounted to the bond head 602 and the second 3D sensor 616 may be mounted to the pick arm drive 614. Thus, movement of the bonding head 602 and bonding tip 604 may cause the first 3D sensor 608 to move, and movement of the pick arm drive 614 and pick arm 612 may cause the second 3D sensor 616 to move. Accordingly, movement of the second 3D sensor 616 may track (e.g., match) movement of the pick arm 612. Likewise, movement of the first 3D sensor 608 may track (e.g., match) any movement of the bond head 602.

In some embodiments, controller 610 is operably coupled to a pick and place apparatus, such as pick and place apparatus 100 (fig. 1), that includes pick assembly 601, and a thermal compression bonding apparatus, such as thermal compression bonding apparatus 200 (fig. 2), that includes bonding head assembly 600. Thus, the controller 610 controls the operation of the pick-up assembly 601 and the bond head assembly 600 in addition to the operation of the first and second 3D sensors 608, 616.

The controller 610 may include a processor 624 coupled to a memory 626 and input/output components 628. Processor 624 may include a microprocessor, a field programmable gate array, and/or other suitable logic devices. The memory 626 may include volatile and/or nonvolatile media (e.g., ROM, RAM, magnetic disk storage media, optical storage media, flash memory devices, and/or other suitable storage media) and/or other types of computer-readable storage media configured to store data. The memory 626 may store algorithms for operating the pick arm and the first and second 3D sensors 608, 616, edge detection, image processing, image segmentation, feature extraction for execution by the processor 624. In some embodiments, the processor 624 may be configured to send data to a computing device, such as a server or personal computer, operably coupled (e.g., via the internet) to the controller 610. Input/output components 628 may include a display, touch screen, keyboard, mouse, and/or other suitable type of input/output device configured to accept input from an operator and provide output to the operator.

Figure 7 depicts a diagram of a pick assembly 601 along with semiconductor dies 702 and ejectors 706 on a dicing tape 704, according to an embodiment of the present disclosure. The dicing tape 704 and the ejector 706 may include any of the dicing tapes and ejectors described above. As noted above, the pick assembly 601 may include a pick arm 612 defining a pick surface 613, a pick arm drive 614, and a second 3D sensor 616, as described above with respect to fig. 6.

When the pick assembly 601 is in the orientation depicted in fig. 7 (e.g., immediately prior to a pick process (e.g., a pick process similar to any of the above), the field of view 622 of the second 3D sensor 616 may include the active surface 708 of the semiconductor die 702. Accordingly, as described in more detail below, the controller 610 may receive image data from the second 3D sensor 616 regarding the active surface 708 of the semiconductor die 702 and may identify defects and/or contamination on the active surface 708 and may identify the tilt of the pick arm 612 relative to the semiconductor die 702 and/or the active surface 708 of the semiconductor die 702.

To facilitate an understanding of embodiments of the present disclosure, an example inspection process is described herein. Fig. 8A and 8B show a flow chart of a method 800 of verifying a pick-up surface (e.g., pick-up surface 613) of a pick-up arm (e.g., pick-up arm 612) and an engagement tip surface (e.g., engagement tip surface 606) of an engagement tip (e.g., engagement tip 604).

The method 800 may include orienting the pick arm to a position where the pick arm is oriented within a field of view of a first 3D sensor (e.g., first 3D sensor 608), as shown in act 802 of fig. 8A. For example, orienting the pick arm (act 802) may include optically aligning the pick arm with the first 3D sensor such that the pick surface is within a field of view of the first 3D sensor. In some embodiments, the controller may cause the pick arm to be oriented within the field of view of the first 3D sensor by moving the pick arm in the X, Y and Z directions and/or rotating the pick arm about a lateral axis (e.g., lateral axis LA) and about a longitudinal axis (e.g., longitudinal axis LO). In one or more embodiments, orienting the pick arm (act 802) may include orienting the pick arm at least substantially directly underneath the first 3D sensor (e.g., the position depicted in fig. 6A). In other embodiments, orienting the pick arm (act 802) may include orienting the pick arm to a typical position for implementing semiconductor die transfer. For example, orienting the pick arm (act 802) may include orienting the pick arm to any of the positions described above with respect to fig. 5A.

In response to orienting the pick arm within the field of view of the first 3D sensor, the method 800 may include capturing and receiving image data from the first 3D sensor, as shown in act 804 of fig. 8A. For example, a controller (e.g., controller 610) operably coupled to the pick arm and the first 3D sensor may cause the first 3D sensor to capture image data and may receive image data from the first 3D sensor, as shown in act 804 of fig. 8A. The image data may include image data (e.g., images, light data, laser data, or other image data) of a pickup surface of the pickup arm. As noted above, in some embodiments, the first 3D sensor may include a laser system and the image data may include three-dimensional contour data. In some embodiments, the image data may include one or more of still image data (e.g., one or more discrete image data packets (e.g., one or more still images)) or video data (e.g., at least substantially continuous image data).

After receiving the image data, method 800 includes analyzing the image data, as shown in act 806 of fig. 8A. For example, the controller may analyze the image data. In some embodiments, analyzing the image data may include identifying an object represented in the image data, as shown in act 808 of fig. 8A. In some embodiments, analyzing the image data may include applying one or more image segmentation techniques to the image data to identify objects and surface topology within the image data. For example, analyzing the image data may include applying one or more image segmentation techniques to the image data to identify one or more of the pick arm, a pick surface of the pick arm, and any other objects and surface topology (e.g., particles, contamination, defects) on or at the pick surface of the pick arm. In one or more embodiments, segmenting the image data may include positioning objects and boundaries (e.g., lines, curves) within the image represented in the image data and correlating the objects with the pick-up arm, the pick-up surface of the pick-up arm, and any other objects and surface topology (e.g., particles, contamination, defects) on or at the pick-up surface of the pick-up arm.

The results of the image segmentation may include a set of segmentations that collectively cover the entire image within the image data and/or a set of contours extracted from the image (e.g., edge detection). Segmenting the image data may include applying any conventional image segmentation process to the image data. In further embodiments, analyzing the image data may also or alternatively include utilizing one or more feature extraction techniques or grayscale analysis techniques to detect and isolate various desired portions or shapes of the image data, such as features (e.g., a pick-up arm, a pick-up surface of a pick-up arm, and any objects (e.g., particles, contamination, defects) on or at the pick-up surface of the pick-up arm)). In some embodiments, identifying the object represented within the image data may include determining a size of the object (e.g., particle, contamination, defect).

In addition, analyzing the image data may optionally further include determining a position and orientation of a pick surface of the pick arm relative to an engagement tip surface of the engagement tip, as shown in act 810 of fig. 8A. For example, via the analysis described above with respect to acts 806 and 808 above, the controller may identify the pick surface of the pick arm, and based on the known position of the first 3D sensor relative to the engagement surface of the engagement tip, the controller may determine the orientation of the pick surface relative to the engagement surface. For example, the controller can determine whether the angle defined between the pick-up surface and the bonding tip surface is within an acceptable range (e.g., less than 80 μm) to avoid damage to the semiconductor die during a subsequent transfer process. In other words, the controller may identify potential non-coplanarity and the degree (e.g., severity) of non-coplanarity between the pick surface and the engagement tip surface.

Based on the objects and/or unacceptable surface topology identified via the analysis described above with respect to acts 806 and 808 of fig. 8A, method 800 may optionally include initiating one or more remedial actions, as shown in act 812 of fig. 8A. For example, in response to identifying (e.g., detecting) one or more objects and surface topology (e.g., defects, contamination, particles, damage), the controller may implement remedial action.

In some embodiments, initiating the one or more remedial actions may include initiating a cleaning process to clean the pick surface of the pick arm, as shown in act 814 of fig. 8A. For example, in response to identifying particles or contamination on the active surface of the semiconductor die, the controller may initiate a cleaning process to clean the pick surface of the pick arm. In some embodiments, the cleaning process may include cleaning the pick arm with a cleaning liquid. For example, the cleaning process may include dipping the pick arm in a cleaning solvent (e.g., alcohol) at the cleaning station. In additional embodiments, the cleaning process may include cleaning the pickup surface with air. For example, the cleaning process may include blowing air (e.g., blow) across the pick surface of the pick arm. In further embodiments, the cleaning process may include scrubbing the pick-up surface with a brush. In still further embodiments, the cleaning process may include sweeping the pickup surface with a vacuum.

In some embodiments, initiating the one or more remedial actions may include replacing the pick arm, as shown in act 816 of fig. 8A. For example, in response to identifying a defect on the pick surface of the pick arm, the controller may initiate a replacement process to replace the pick arm. For example, the controller may trigger an alarm indicating that the pick arm is defective and needs to be replaced.

In some embodiments, initiating the one or more remedial actions may include triggering an alarm, as shown in act 818 of fig. 8A. For example, in response to identifying particles, contamination, defects, and/or damage, the controller may trigger an alarm requesting operator attention and inspection. In some embodiments, which remedial actions are initiated by the inspection system may be based at least in part on the determined size of the contamination, defect, and/or particle. For example, some size ranges of contamination, defects, and/or particles may only require warning alerts, while others may require stopping operation of the pick and place equipment and operator attention.

As noted above, the method 800 may include determining a position and orientation of a pick surface of a pick arm relative to an engagement tip surface of an engagement tip. In response to identifying non-coplanarity between the pickup surface and the engagement tip surface (e.g., an indication that an angle defined between the pickup surface and the engagement tip surface is outside an acceptable range or that the pickup arm has deviated from calibration), the method 800 may include recalibrating the pickup arm and/or the engagement tip, as shown in act 820 of fig. 8A. For example, in response to identifying non-coplanarity between the pick surface and the engagement tip surface, the controller may initiate recalibration of the pick arm and/or the engagement tip. In some embodiments, the pick-up arm and/or the engagement tip may be recalibrated by any conventional method. In one or more embodiments, based on the determined position of the pick-up surface of the pick-up arm relative to the engagement tip surface of the engagement tip, the controller may cause the pick-up arm and/or the engagement tip to move in X, Y and the Z-direction. In addition, the controller can cause the pick-up arm to rotate about a transverse axis (e.g., transverse axis LA) and about a longitudinal axis (e.g., longitudinal axis LO) to at least substantially achieve coplanarity between the pick-up surface and the engagement tip surface.

The method 800 may further include orienting the pick arm to a position in which the engagement tip surface is within a field of view of a second 3D sensor (e.g., second 3D sensor 616) coupled to the pick arm, as shown in act 822 of fig. 8A. For example, orienting the pick arm (act 822) may include optically aligning the engagement tip surface with the second 3D sensor such that the engagement tip surface is within a field of view of the second 3D sensor. In some embodiments, the controller may orient the pick arm, and thus the second 3D sensor, by moving the pick arm in the X, Y and Z directions and/or rotating the pick arm about the lateral axis and about the longitudinal axis. In one or more embodiments, orienting the pick arm (act 822) may include orienting the pick arm such that the second 3D sensor is at least substantially directly below the engagement tip surface (e.g., the position depicted in fig. 6B). In other embodiments, orienting the pick arm (act 822) may include orienting the pick arm to a typical position for performing semiconductor die transfer. For example, orienting the pick arm (act 822) may include orienting the pick arm to any of the positions described above with respect to fig. 5A.

In response to orienting the pick arm to a position in which the engagement tip surface is within the field of view of the second 3D sensor, the method 800 may include causing the second 3D sensor to capture image data from the second 3D sensor and receive the image data, as shown in act 824 of fig. 8A. For example, a controller operably coupled to the pick arm and the second 3D sensor may cause the second 3D sensor to capture image data and may receive image data from the second 3D sensor. The image data may include image data (e.g., images, light data, laser data, or other image data) of the engaging tip surface of the engaging tip. As noted above, in some embodiments, the second 3D sensor may include a laser system and the image data may include three-dimensional profile data. In some embodiments, the image data may include one or more of still image data (e.g., one or more discrete image data packets (e.g., one or more still images)) or video data (e.g., at least substantially continuous image data).

After receiving the image data, method 800 includes analyzing the image data, as shown in act 826 of fig. 8A. For example, the controller may analyze the image data. In some embodiments, analyzing the image data may include identifying an object represented in the image data, as shown in act 828 of fig. 8B. For example, the controller may analyze the image data to identify one or more of the bonding tip, the bonding tip surface of the bonding tip, and any objects and surface topology (e.g., particles, contamination, defects) on or at the bonding tip surface of the bonding tip via any of the manners described above with respect to act 806 of fig. 8A. In some embodiments, identifying the object represented within the image data may include determining a size of the object and a surface topology (e.g., particles, contamination, defects).

In addition, analyzing the image data may optionally further include determining a position and orientation of a pickup surface of the pickup arm relative to an engagement tip surface of the engagement tip, as shown in act 830 of fig. 8B. For example, via any of the analyses described above with respect to acts 806, 808, 810, 826, and 828 described above, the controller may identify the engagement tip surface of the engagement tip, and based on the known position of the second 3D sensor relative to the pickup surface of the pickup arm, the controller may determine the orientation of the pickup surface relative to the engagement surface. For example, the controller can determine whether the angle defined between the pick-up surface and the bonding tip surface is within an acceptable range (e.g., less than 80 μm) to avoid damage to the semiconductor die during subsequent transfer processes. In other words, the controller may identify non-coplanarity and the degree (e.g., severity) of non-coplanarity between the pick surface and the engagement tip surface.

Referring concurrently to acts 810 and 830 of fig. 8A and 8B, in some embodiments, the controller may utilize a combination of image data received from the first 3D sensor and image data received from the second 3D sensor to determine a position and orientation of the pick surface of the pick arm relative to the engagement tip surface of the engagement tip.

Based on the objects and/or unacceptable surface topology identified via the analysis described above with respect to acts 826 and 828 of fig. 8B, method 800 may optionally include initiating one or more remedial actions, as shown in act 832 of fig. 8B. For example, in response to identifying (e.g., detecting) one or more objects (e.g., defects, contamination, particles, damage), the controller may implement remedial action.

In some embodiments, initiating the one or more remedial actions may include initiating a cleaning process to clean the bond tip surface of the bond head, as shown in act 834 of fig. 8B. For example, in response to identifying particles or contamination on the engagement tip surface of the engagement tip, the controller may initiate a cleaning process to clean the engagement tip surface of the engagement tip. In additional embodiments, the cleaning process may include cleaning the engagement tip surface with air. For example, the cleaning process may include blowing gas (e.g., a blow spray) across the engagement tip surface of the engagement tip. In further embodiments, the cleaning process may include scrubbing the engaging tip surface with a brush. In yet further embodiments, the cleaning process may include sweeping the engagement tip surface with a vacuum.

In some embodiments, initiating the one or more remedial actions may include replacing the engagement tip, as shown in act 836 of fig. 8B. For example, in response to identifying a defect on the engagement tip surface of the engagement tip, the controller may initiate a replacement process to replace the engagement tip. For example, the controller may trigger an alarm indicating that the engagement tip is defective and needs to be replaced.

In some embodiments, initiating the one or more remedial actions may include triggering an alarm, as shown in act 838 of fig. 8B. For example, in response to identifying particles, contamination, defects, and/or damage, the controller may trigger an alarm requesting operator attention and inspection. As noted above, in some embodiments, which remedial actions are initiated by the inspection system may be based at least in part on the determined size of the contamination, defects, and/or particles. For example, some size ranges of contamination, defects, and/or particles may only require warning alerts, while others may require stopping operation of the pick and place equipment and operator attention.

As noted above, the method 800 may include determining a position and orientation of a pick surface of a pick arm relative to an engagement tip surface of an engagement tip (e.g., acts 810 and 830). In response to identifying non-coplanarity between the pick-up surface and the engagement tip surface of the pick-up arm (e.g., an indication that an angle defined between the pick-up surface and the engagement tip surface is outside an acceptable range or that the pick-up arm has deviated from calibration), the method 800 may include recalibrating the pick-up arm and/or the engagement tip, as shown in act 840 of fig. 8B. For example, in response to identifying non-coplanarity between the pick-up surface of the pick-up arm and the engagement tip surface, the controller may initiate recalibration of the pick-up arm and/or the engagement tip. In some embodiments, the pick-up arm and/or the engaging tip may be recalibrated by any conventional method. In one or more embodiments, based on the determined position of the pick surface of the pick arm relative to the engagement tip surface of the engagement tip, the controller may cause one or more of the pick arm and the engagement tip to move in X, Y and the Z direction. In addition, the controller may cause the pick arm to rotate about a transverse axis (e.g., transverse axis LA) and about a longitudinal axis (e.g., longitudinal axis LO) to at least substantially achieve coplanarity between the pick surface and the engagement tip surface.

Still referring to fig. 8A and 8B, in some embodiments, method 800 may include only acts within acts 802-820. In other embodiments, method 800 may include only acts within acts 822-840. For example, acts 802-820 may include methods separate from acts 822-840.

Referring also to acts 802-840, in some embodiments, method 800 or portions thereof (e.g., acts 802-820 or acts 822-840) may be performed intermittently throughout a pick-and-transfer process (e.g., the pick-and-transfer process described above with respect to fig. 1). In additional embodiments, the method 800, or portions thereof, may be performed between each individual pick-up and transfer process of each individual semiconductor die of a wafer. In further embodiments, method 800, or portions thereof, may be performed at the beginning of a pick and transfer process for a given wafer.

Still referring to fig. 8A and 8B, the inspection systems (e.g., inspection system 603) and methods (e.g., method 800) described herein may provide advantages over conventional pick and transfer apparatus and methods. For example, by detecting defects and/or particles on critical surfaces during pick-up and transfer, the inspection systems and methods described herein may avoid utilizing surfaces (e.g., pick-up surfaces, bonding tip surfaces) that may cause semiconductor die cracking. Furthermore, by detecting defects and/or particles on critical surfaces during pick and transfer processes, the inspection systems and methods described herein may implement a remedial process to prepare the surfaces for use in the pick and transfer process, which reduces the risk of damaging the semiconductor die.

Likewise, by detecting non-coplanarity between the pick surface and the engagement tip surface, the inspection systems and methods described herein may avoid pick and transfer actions when there is no coplanarity between the pick surface and the engagement tip surface, which reduces the risk of cracking due to edge contact and non-uniform force distribution. Furthermore, by detecting non-coplanarity between the pick surface and the bonding tip surface, the inspection systems and methods described herein may correct the orientation of the pick arm and/or the bonding tip surface, which reduces the risk of damaging the semiconductor die.

By reducing the risk of damaging semiconductor dies, the inspection systems and methods described herein reduce the risk of yield loss and improve the reliability of pick and transfer equipment and processes. Furthermore, reducing the risk of damaging the semiconductor die due to contamination, particle impact, tool and/or product defects, and non-coplanarity, the inspection systems and methods described herein may be capable of handling and processing semiconductor dies having reduced thicknesses than conventional pick and transfer equipment and processes. Additionally, the inspection systems and methods described herein may further enable near-zero bond wire packages (e.g., systems and processes that may further enable near-zero bond wire packages).

Fig. 9 depicts a diagram of an inspection system 903 including a bond head assembly 900 and a pick assembly 901, in accordance with an embodiment of the present disclosure. The bond head assembly 900 may include a bond head 902 and a bond tip 904 defining a bond tip surface 906. The pick assembly 901 may include a pick arm 912 and a pick arm drive 914 that define a pick surface 913. The bond head assembly 900 may include any of the bond heads and bond tips described above. The pick assembly 901 may include any of the pick arms described above.

The inspection system 903 may further include a single 3D sensor 909. In one or more embodiments, a single 3D sensor 909 may be attached (e.g., mounted) to a separate portion of the pick and place apparatus. In some embodiments, a single 3D sensor 909 may be oriented such that both the pickup surface 913 and the engagement tip surface 906 are within a field of view 911 of the single 3D sensor 909. In other embodiments, the single 3D sensor 909 may be movable (e.g., rotatable in a direction 915) between a first position and a second position, wherein the pickup surface 913 is within a field of view 911 of the single 3D sensor 909 when the single 3D sensor 909 is oriented in the first position, and the engagement tip surface 906 is within the field of view 911 of the single 3D sensor 909 when the single 3D sensor 909 is oriented in the second position. In a further embodiment, the inspection system 903 may also include a mirror element M that the inspection system 903 may utilize to change the field of view 911 of the single 3D sensor 909 to include the pickup surface 913 when in the first position and the engagement tip surface 906 when in the second position.

In addition, a single 3D sensor 909 is operably coupled to the controller, and the inspection system 903 can perform any analysis and implement any of the remedial measures described above with respect to fig. 6A, 6B, and 8.

Embodiments of the present disclosure further include utilizing the inspection system described herein during the pick and transfer process. Accordingly, to facilitate an understanding of the embodiments described herein, example pick-up and transfer processes that utilize first and second 3D sensors (e.g., first and second 3D sensors 608, 616) to improve the pick-up and transfer processes are described below. Fig. 10A and 10B show a flow chart of a method 1000 of picking up and transferring semiconductor dies according to an embodiment of the present disclosure.

In some embodiments, the method 1000 may include orienting a pick arm (e.g., pick arm 104) and a second 3D sensor (e.g., second 3D sensor 616) to a location to begin a pick and transfer process, as shown in act 1002 of fig. 10A. For example, the method may include orienting the pick arm such that the pick arm and the second 3D sensor are suspended over and optically aligned with the location of the semiconductor die, as shown in act 1002 of fig. 10A. For example, the pick arm and the second 3D sensor may be oriented via any of the manners described above with respect to fig. 1 and 4A.

In response to orienting the pick arm and the second 3D sensor, method 1000 may include capturing image data from the second 3D sensor and receiving the image data, as shown in act 1004 of fig. 10A. For example, a controller (e.g., controller 610) operably coupled to the pick arm and the second 3D sensor may cause the second 3D sensor to capture image data and may receive image data from the second 3D sensor. The image data may include image data (e.g., image, light data, laser data) of the active surface of the semiconductor die. The image data may further include image data of additional semiconductor dies and/or dicing tape (e.g., dicing tape 102) of the wafer. As noted above, in some embodiments, the second 3D sensor may include a laser system, and the image data may include three-dimensional contour data. In some embodiments, the image data may include one or more of still image data, such as one or more discrete image data packets (e.g., one or more still images), and video data, such as at least substantially continuous image data.

After receiving the image data, method 1000 includes analyzing the image data, as shown in act 1006 of fig. 10A. For example, the controller may analyze the image data. In some embodiments, analyzing the image data may include identifying an object represented in the image data, as shown in act 1008 of fig. 10A. The controller may analyze the image data to identify the semiconductor die, the active surface of the semiconductor die, and any objects and surface topology (e.g., particles, contamination, defects) on or at the active surface of the semiconductor die via any of the manners described above with respect to acts 806 and 808 of fig. 8A. In some embodiments, identifying the object represented within the image data may include determining a size of the object (e.g., particle, contamination, defect). In addition, the image data can be analyzed to determine the pitch, size, and height of conductive elements protruding from the active surface of such a die, such as when the semiconductor die is configured with such structures.

In addition, analyzing the image data may optionally further include determining a position and orientation of a pick surface of the pick arm relative to an active surface of the semiconductor die, as shown in act 1010 of fig. 10A. For example, via the analysis described above with respect to acts 1006 and 1008 above, the controller may identify an active surface of the semiconductor die, and based on the known position of the second 3D sensor relative to the pick surface of the pick arm, the controller may determine an orientation of the pick surface relative to the active surface. For example, the controller may determine whether the angle defined between the active surface and the pick surface is within an acceptable range (e.g., less than between about 30 μm to about 50 μm) to avoid damage to the semiconductor die during a subsequent pick process. In other words, the controller can identify non-coplanarity and the degree (e.g., severity) of non-coplanarity between the pickup surface and the active surface.

Based on the objects and unacceptable surface topology identified via the analysis described above with respect to acts 1006 and 1008 of fig. 10A, method 1000 may optionally include initiating one or more remedial actions, as shown in act 1012 of fig. 10A. For example, in response to identifying (e.g., detecting) one or more objects and surface topology (e.g., defects, contamination, particles, damage), the controller may implement remedial action.

In some embodiments, initiating the one or more remedial actions may include discarding the semiconductor die, as shown in act 1014 of fig. 10A. For example, in response to identifying a defect and/or damage to the semiconductor die, the controller may cause the semiconductor die to be discarded (e.g., not removed from the dicing tape or otherwise discarded).

In some embodiments, initiating the one or more remedial actions may include initiating a cleaning process to clean the active surface of the semiconductor die, as shown in act 1016 of fig. 10A. For example, in response to identifying particles or contamination on the active surface of the semiconductor die, the controller may initiate a cleaning process to clean the active surface of the semiconductor die. In one or more embodiments, the cleaning process can include cleaning the active surface of the semiconductor die with air. For example, the cleaning process may include blowing gas (e.g., a blow) across the active surface of the semiconductor die. In yet further embodiments, the cleaning process may include sweeping the active surface of the semiconductor die with a vacuum.

In some embodiments, initiating the one or more remedial actions may include triggering an alarm, as shown in act 1018 of fig. 10A. For example, in response to identifying particles, contamination, defects, and/or damage, the controller may trigger an alarm requesting operator attention and inspection. In some embodiments, the alarm may be triggered in response to a selected number (e.g., one, two, three, four, or more) of particles, contamination, defects, and/or damage. In addition, in response to identifying particles, contamination, defects, and/or damage, the controller may cause the semiconductor die to rework. Further, in response to identifying particles, contamination, defects, and/or damage, the controller may initiate another inspection (e.g., a review) of the semiconductor die. As noted above, in some embodiments, which remedial actions are initiated by the inspection system may be based at least in part on the determined size of the contamination, defects, and/or particles. For example, some size ranges of contamination, defects, and/or particles may only require warning alerts, while others may require stopping operation of the pick and place equipment and require operator attention.

As noted above, the method 1000 may include determining a position and orientation of a pick surface of a pick arm relative to an active surface of a semiconductor die. In response to identifying non-coplanarity between the pick-up surface of the pick-up arm and the active surface of the semiconductor die (e.g., an indication that an angle defined between the pick-up surface and the active surface is outside an acceptable range or that the pick-up arm has deviated from calibration), the method 1000 may include recalibrating the pick-up arm, as shown in act 1020 of fig. 10A. For example, the controller may recalibrate the pick-up arm via any of the manners described above with respect to acts 820 and 840 of fig. 8A and 8B.

After determining that the active surface of the semiconductor die is free of defects, contamination, and particles and determining that coplanarity exists between the pick surface of the pick arm and the active surface of the semiconductor die, the method 1000 may include picking up the semiconductor from the dicing tape, as shown in act 1022 of fig. 10A. For example, the method 1000 may include picking up semiconductor dies from the dicing tape via any of the manners described above with respect to fig. 1, 4A-4C, and 4F.

After picking up the semiconductor die from the dicing tape, the method 1000 includes lifting the pick arm and the second 3D sensor, moving the pick arm and the second 3D sensor in the X, Y and Z directions, and/or rotating the pick arm and the second 3D sensor about the lateral axis LA and the longitudinal axis LO to present the semiconductor die to the bonding tip surface and for performing a subsequent transfer process, as shown in act 1024 of fig. 10A. For example, the pick arm and second 3D sensor may be moved and oriented to present the semiconductor to the bonding tip surface and used to perform the subsequent transfer process via any of the manners described above with respect to fig. 1 and 5A.

In response to the orientation pick arm and the second 3D sensor being used to perform the transfer process, method 1000 may include capturing and receiving image data via and from the first 3D sensor (e.g., first 3D sensor 608) and/or the second 3D sensor, as shown in act 1026 of fig. 10A. For example, a controller operably coupled to the pick arm, the engagement tip, the first 3D sensor, and the second 3D sensor may cause the first 3D sensor and/or the second 3D sensor to capture image data and may receive image data from the first 3D sensor and/or the second 3D sensor. The image data may include image data (e.g., images, light data, laser data) of the bonding tip surface and/or the back side of the semiconductor die. As noted above, in some embodiments, the first 3D sensor and the second 3D sensor may include a laser system, and the image data may include three-dimensional contour data. In some embodiments, the image data may include one or more of still image data, such as one or more discrete image data packets (e.g., one or more still images), and video data, such as at least substantially continuous image data.

After receiving the image data, method 1000 includes analyzing the image data, as shown in act 1028 of fig. 10B. For example, the controller may analyze the image data. In some embodiments, analyzing the image data may identify an object represented in the image data, as shown in act 1030 of FIG. 10B. The controller may analyze the image data to identify the semiconductor die, the active surface of the semiconductor die, the bonding tip surface, and any other objects (e.g., particles, contamination, defects) on or at the active surface of the semiconductor die and/or on or at the bonding tip surface of the bonding tip, via any of the manners described above with respect to acts 806 and 808 of fig. 8A and 8B. In some embodiments, identifying the object represented within the image data may include determining a size of the object (e.g., particle, contamination, defect).

In addition, analyzing the image data may optionally further include determining a position and orientation of a bonding tip surface of the bonding tip relative to the back side of the semiconductor die, as shown in act 1032 of fig. 10B. For example, via the analysis described above with respect to acts 1028 and 1030 above, the controller may identify the bonding tip surface of the bonding tip and the back side of the semiconductor die, and based on the known position of the first 3D sensor relative to the bonding tip surface, the controller may determine an orientation of the bonding tip surface relative to the back side of the semiconductor die. For example, the controller may determine whether an angle defined between the bonding tip surface and the back side of the semiconductor die is within an acceptable range (e.g., less than 30 μm) to avoid damage to the semiconductor die during a subsequent transfer process. In other words, the controller may identify the non-coplanarity and the degree (e.g., severity) of non-coplanarity between the bonding tip surface and the back side of the semiconductor die.

Based on the objects identified via the analysis described above with respect to acts 1028 and 1030 of fig. 10B, method 1000 may optionally include initiating one or more remedial actions, as shown in act 1033 of fig. 10B. For example, in response to identifying (e.g., detecting) one or more objects (e.g., defects, contamination, particles, damage), the controller may implement remedial action.

In some embodiments, initiating the one or more remedial actions may include discarding the semiconductor die, as shown in act 1034 of fig. 10B. For example, in response to identifying a defect and/or damage to the semiconductor die, the controller may cause the semiconductor die to be discarded (e.g., not removed from the dicing tape or otherwise discarded). Additionally, in some embodiments, the controller may cause the semiconductor die to rework in response to identifying particles, contamination, defects, and/or damage. Further, in response to identifying particles, contamination, defects, and/or damage, the controller may initiate another inspection (e.g., a review) of the semiconductor die. As noted above, in some embodiments, which remedial actions are initiated by the inspection system may be based at least in part on the determined size of the contamination, defects, and/or particles. For example, some size ranges of contamination, defects, and/or particles may only require warning alerts, while others may require stopping operation of the pick and place equipment and operator attention.

In some embodiments, initiating the one or more remedial actions may include initiating a cleaning process to clean the back side of the semiconductor die and/or the bond tip surface of the bond head, as shown in act 1036 of fig. 10B. For example, in response to identifying particles or contamination on the back side of the semiconductor die, the controller may initiate a cleaning process to clean the back side of the semiconductor die. In one or more embodiments, the cleaning process can include cleaning the back side of the semiconductor die with air. For example, the cleaning process may include blowing gas (e.g., a blow) across the active surface of the semiconductor die. Additionally, in response to identifying particles or contamination on the engagement tip surface of the engagement tip, the controller may initiate a cleaning process to clean the engagement tip surface of the engagement tip. In some embodiments, the cleaning process may include cleaning the engagement tip with a cleaning liquid. For example, the cleaning process may include dipping the bonding tips in a cleaning solvent at a cleaning station. In additional embodiments, the cleaning process may include cleaning the engagement tip surface with air. For example, the cleaning process may include blowing (e.g., blowing) air across the engagement tip surface of the engagement tip. In further embodiments, the cleaning process may include scrubbing the engagement tip surface with a brush. In yet further embodiments, the cleaning process may include sweeping the bonding tip surface of the semiconductor die with a vacuum.

In some embodiments, initiating the one or more remedial actions may include replacing the engagement tip, as shown in act 1037 of fig. 10B. For example, in response to identifying a defect on the engagement tip surface of the engagement tip, the controller may initiate a replacement process to replace the engagement tip. For example, the controller may trigger an alarm indicating that the engagement tip is defective and needs to be replaced.

In some embodiments, initiating the one or more remedial actions may include triggering an alarm, as shown in act 1038 of fig. 10B. For example, in response to identifying particles, contamination, defects, and/or damage, the controller may trigger an alarm requesting operator attention and inspection. In some embodiments, the alarm may be triggered in response to a selected number (e.g., one, two, three, four, or more) of particles, contamination, defects, and/or damage. In addition, in response to identifying particles, contamination, defects, and/or damage, the controller may cause the semiconductor die to rework. Further, in response to identifying particles, contamination, defects, and/or damage, the controller may initiate another inspection (e.g., a review) of the semiconductor die.

As noted above, the method 1000 may include determining a position and an orientation of a bonding tip surface of a bonding tip relative to a back side of a semiconductor die. In response to identifying non-coplanarity between the engagement tip surface and the back surface (e.g., an indication that an angle defined between the engagement tip surface and the back surface is outside an acceptable range or that the pick arm has deviated from calibration), the method 1000 may include recalibrating the pick arm, as shown in act 1040 of fig. 10B. For example, the controller may recalibrate the pick arm via any of the manners described above with respect to acts 820 and 840 of fig. 8A and 8B.

After determining that the back side of the semiconductor die and the bonding tip surface of the bonding tip are free of defects, contamination, and particles and that the bonding tip surface is parallel to the back side, the method 1000 may include transferring the semiconductor from the pick-up arm to the bonding tip, as shown in act 1042 of fig. 10B. For example, the method 1000 may include transferring the semiconductor die from the pick arm via any of the manners described above with respect to fig. 1 and 5A.

Still referring to fig. 10A and 10B and method 1000, any of the acts of method 800 of fig. 8A and 8B may be interposed between, before, and/or after the acts of fig. 10A and 10B. Additionally, the method 1000 may provide any of the advantages described above with respect to fig. 8A and 8B.

Referring also to fig. 4A-10, in some embodiments, the inspection systems described herein may utilize one or more machine learning models to analyze image data and detect objects represented in the image data. In one or more embodiments, applying the one or more machine learning models may include analyzing the image data by applying machine learning and/or deep learning techniques, including providing a training corpus to a matching learning algorithm or neural network to train a machine to recognize objects within the image data and correlate events of the objects with process actions previously performed on the semiconductor die. In some embodiments, the inspection system may analyze the image data using one or more of a regression model (e.g., a set of statistical processes used to estimate the relationships between variables), a classification model, and/or a phenomenological model. Additionally, the machine learning model may include quadratic regression analysis, logistic regression analysis, support vector machines, gaussian process regression, integration models, or any other regression analysis. Further, in yet further embodiments, the machine learning model may include decision tree learning, regression trees, boosting trees, gradient boosting trees, multi-level perceptrons, one-to-many, naive bayes, k-nearest neighbors, association rule learning, neural networks, deep learning, pattern recognition, or any other type of machine learning.

For example, the inspection system may apply one or more of the above-described machine learning techniques to the results of image data and image analysis (e.g., detected objects). In addition, by applying one or more machine learning techniques to the data, the inspection system can identify evidence of contamination, defects, particles, and/or non-coplanarity. In some embodiments, the operational flow and/or logic of the inspection system may include an action flow for use in different cases. For example, the first operational flow may include a decision tree utilized in detecting evidence of contamination, defects, particles, and/or non-coplanarity. Another operational flow may include a decision tree utilized when evidence of contamination, defects, particles, and/or non-coplanarity is not detected. In view of the foregoing, the operational flow may be related to feedback data detected in the image data and/or data acquired from subsequent steps of processing the semiconductor die.

As non-limiting examples, the inspection system may utilize the feedback loop of the inspection system by identifying evidence of contamination, defects, particles, and/or non-coplanarity, determining previous process steps (e.g., singulation, formation) performed on the semiconductor die, previous use of equipment, pick-up arms, and/or bonding tips. In other words, via machine learning model techniques, the verification system can learn correlations between: (1) evidence of contamination, defects, particles, contamination rate, defect rate and/or particle rate and/or non-coplanarity and previous process steps and (2) previous use of pick-up arms and/or bonding tips and equipment previously used during the packaging process. In other words, the verification system may learn the relationship between: (1) evidence of contamination, defects, particles, contamination rates, defect rates and/or particle rates and/or non-coplanarity and (2) the operational and/or logic flows of inspection systems and other equipment that pick and place devices and handle and manufacture semiconductor dies. For example, machine learning models are trained via supervised and/or unsupervised learning, as is known in the art. After a sufficient number of iterations, the machine learning model becomes a trained machine learning model. In some embodiments, the machine learning model may also be trained based on historical data from previous uses of the inspection system and the pick and place equipment (e.g., image data, data reflecting previous uses of the pick and place equipment, evidence of contamination, defects, particles, and/or non-coplanarity data), and/or expert input data and/or related literature. In one or more embodiments, via machine learning analysis, the inspection system can learn how to adjust the orientation of the pick-up arm, ejector, and/or engagement tip to achieve coplanarity. In additional embodiments, via machine learning analysis, the inspection system may be aware of contamination, defects, particles, and/or contamination rates, defect rates, and/or particle rates indicative of wear within the pick and place equipment and/or other equipment that handles and/or manufactures semiconductor dies. In additional embodiments, via machine learning analysis, the inspection system may be aware of contamination, defects, particles and/or contamination rates, defect rates and/or particle rates indicating that specific maintenance is required or that adjustments or changes to earlier processes are required.

Referring also to fig. 4A-10, the inspection systems and methods described herein are equally applicable to other semiconductor manufacturing processes, such as, for example, tape and reel systems and picking wafers from carriers.

Embodiments of the present disclosure include an apparatus for handling microelectronic devices. The apparatus includes: a bonding head; a bonding tip coupled to the bonding head and having a bonding tip surface configured to receive a microelectronic device thereon; a first 3D sensor carried by the bond head; a pick-up arm drive; a pick arm coupled to the pick arm drive and having a pick surface configured to receive the microelectronic device thereon; and a second 3D sensor drivingly carried by the pick arm. The apparatus further includes a controller configured to: receiving first image data from the first 3D sensor, the first image data including image data of the pickup surface of the pickup arm; and receiving second image data from the second 3D sensor, the second image data comprising image data of the engaging tip surface of the engaging tip.

Embodiments of the present disclosure include a method comprising: moving a pick arm to a first position relative to an engagement tip, wherein a pick surface of the pick arm is within a field of view of a first 3D sensor coupled to the engagement tip; capturing image data of the pickup surface via the first 3D sensor; and analyzing the image data of the pickup surface.

Embodiments of the present disclosure include an apparatus for handling microelectronic devices. The apparatus includes: a bonding head; a bonding tip coupled to the bonding head and having a bonding tip surface configured to receive a microelectronic device thereon; a pick-up arm drive; a pick arm coupled to the pick arm drive and having a pick surface configured to receive the microelectronic device thereon; and at least one 3D sensor configured to capture image data of the engagement tip surface and image data of the pickup surface.

Embodiments of the present disclosure further include:

embodiment 1. an apparatus for handling microelectronic devices, comprising: a bonding head; a bonding tip coupled to the bonding head and having a bonding tip surface configured to receive a microelectronic device thereon; a first 3D sensor carried by the bond head; a pick-up arm drive; a pick arm coupled to the pick arm drive and having a pick surface configured to receive the microelectronic device thereon; and a second 3D sensor drivingly carried by the pick arm.

Embodiment 2. the apparatus of embodiment 1, wherein at least one of the first 3D sensor or the second 3D sensor comprises a laser system.

Embodiment 3 the apparatus of any of embodiments 1 or 2, further comprising a controller operably coupled to the first 3D sensor and the second 3D sensor, the controller comprising: at least one processor; and at least one non-transitory computer-readable storage medium storing instructions thereon that, when executed by the at least one processor, cause the controller to: receiving first image data from the first 3D sensor, the first image data comprising image data of the pickup surface of the pickup arm; and receiving second image data from the second 3D sensor, the second image data comprising image data of the engaging tip surface of the engaging tip.

The apparatus of embodiment 3, further comprising instructions that, when executed by the at least one processor, cause the controller to: analyzing the first image data; analyzing the second image data; detecting one or more of a defect, a particle, or a contamination on the pickup surface of the pickup arm based at least in part on the analysis of the first image data; and detecting one or more of defects, particles, or contamination on the bonding tip surface of the bonding tip based at least in part on the analysis of the second image data.

The apparatus of embodiment 5, the apparatus of embodiment 4, further comprising instructions that, when executed by the at least one processor, cause the controller to: determining a position of the pick surface of the pick arm relative to the engagement tip surface of the engagement tip based at least in part on the analysis of at least one of the first image data and the second image data; and determining whether the pick surface is parallel to the engagement tip surface based at least in part on the determined position of the pick surface of the pick arm relative to the engagement tip surface of the engagement tip.

Embodiment 6 the apparatus of any of embodiments 4 or 5, further comprising instructions that, when executed by the at least one processor, cause the controller to: initiating a remedial action in response to detecting a defect, particle, or contamination on one or more of the pick surface of the pick arm or the engagement tip surface of the engagement tip.

Embodiment 7 the apparatus of any of embodiments 5 or 6, further comprising instructions that, when executed by the at least one processor, cause the controller to: initiating a remedial action in response to determining that the pickup surface is not parallel to the engagement tip surface.

Embodiment 8 the apparatus of embodiment 6, wherein initiating a remedial action comprises at least one of: cleaning one or more of the pick surface or the engagement tip surface; replacing one or more of the pick arm or the engagement tip; or trigger an alarm.

Embodiment 9 the apparatus of embodiment 7, wherein initiating remedial action comprises recalibrating one or more of the pick arm or the engagement tip.

The embodiment 10 the apparatus of any of embodiments 4-9, further comprising instructions that, when executed by the at least one processor, cause the controller to: causing the pick arm to move to a first position in which the engagement tip surface is within a field of view of the second 3D sensor; capturing image data of the engagement tip surface with the second 3D sensor; causing the pick arm to move to a second position in which the pick surface is within a field of view of the first 3D sensor; and capturing image data of the pickup surface with the first 3D sensor.

Embodiment 11. a method, comprising: moving a pick arm to a first position relative to an engagement tip, wherein a pick surface of the pick arm is within a field of view of a first 3D sensor coupled to the engagement tip; capturing first image data of the pickup surface via the first 3D sensor; and analyzing the first image data of the pickup surface.

Embodiment 12 the method of embodiment 11, further comprising: moving the pick arm to a second position relative to an engagement tip, wherein an engagement tip surface of the engagement tip is within a field of view of a second 3D sensor coupled to the pick arm; capturing second image data of the engagement tip surface via the second 3D sensor; and analyzing the second image data of the bonding tip surface.

Embodiment 13. the method of embodiment 12, further comprising: determining an orientation of the pickup surface of the pickup arm relative to the engagement tip surface of the engagement tip based at least in part on the analysis of at least one of the first image data of the pickup surface and the second image data of the engagement tip surface; and determining whether an angle defined between the pickup surface and the engagement tip surface is within an acceptable range based at least in part on the determined orientation of the pickup surface of the pickup arm relative to the engagement tip surface of the engagement tip.

Embodiment 14. the method of embodiment 13, further comprising: in response to determining that the angle defined between the pickup surface and the engagement tip surface is outside of an acceptable range, initiating a remedial action.

Embodiment 15 the method of embodiment 14, wherein initiating remedial action comprises recalibrating one or more of the pick arm or the engagement tip.

Embodiment 16. the method of any of embodiments 13-15, further comprising: detecting one or more of a defect, a particle, or a contamination on the pickup surface of the pickup arm based at least in part on the analysis of the first image data; and detecting one or more of a defect, particle, or contamination on the bonding tip surface of the bonding tip based at least in part on the analysis of the second image data.

Embodiment 17. the method of any of embodiments 12-16, further comprising: detecting one or more of a defect, a particle, or a contamination on the pickup surface of the pickup arm based at least in part on the analysis of the first image data; and detecting one or more of a defect, particle, or contamination on the bonding tip surface of the bonding tip based at least in part on the analysis of the second image data.

Embodiment 18 the method of embodiment 17, further comprising: initiating a remedial action in response to detecting a defect, particle, or contamination on one or more of the pick surface of the pick arm or the engaging tip surface of the engaging tip.

The embodiment 19. the method of any of embodiments 12-16, wherein analyzing one or more of the first image data or the second image data comprises applying one or more machine learning techniques to one or more of the first image data or the second image data.

Embodiment 20 an apparatus for handling microelectronic devices, comprising: a bonding head; a bonding tip coupled to the bonding head and having a bonding tip surface configured to receive a microelectronic device thereon; a pick-up arm drive; a pick arm coupled to the pick arm drive and having a pick surface configured to receive the microelectronic device thereon; and at least one 3D sensor configured to capture image data of the engagement tip surface and image data of the pickup surface.

While certain illustrative embodiments have been described in connection with the figures, those of ordinary skill in the art will recognize and appreciate that embodiments encompassed by the present disclosure are not limited to those explicitly shown and described herein. Rather, many additions, deletions, and modifications to the embodiments described herein may be made without departing from the scope of the embodiments encompassed by the present disclosure (e.g., the scope of the appended claims, including legal equivalents). In addition, features from one disclosed embodiment may be combined with features of one or more other disclosed embodiments while still being encompassed within the scope of the disclosure.

Claims (20)

1. An apparatus for handling microelectronic devices, comprising:

a bonding head;

a bonding tip coupled to the bonding head and having a bonding tip surface configured to receive a microelectronic device thereon;

a first 3D sensor carried by the bond head;

a pick-up arm drive;

a pick arm coupled to the pick arm drive and having a pick surface configured to receive the microelectronic device thereon; and

a second 3D sensor carried by the pick arm drive.

2. The apparatus of claim 1, wherein at least one of the first 3D sensor or the second 3D sensor comprises a laser system.

3. The apparatus of any of claims 1 or 2, further comprising a controller operably coupled to the first 3D sensor and the second 3D sensor, the controller comprising:

at least one processor; and

at least one non-transitory computer-readable storage medium storing instructions thereon, which when executed by the at least one processor, cause the controller to:

receiving first image data from the first 3D sensor, the first image data including image data of the pickup surface of the pickup arm; and

receiving second image data from the second 3D sensor, the second image data comprising image data of the engaging tip surface of the engaging tip.

4. The apparatus of claim 3, further comprising instructions that, when executed by the at least one processor, cause the controller to:

analyzing the first image data;

analyzing the second image data;

detecting one or more of a defect, a particle, or a contamination on the pickup surface of the pickup arm based at least in part on the analysis of the first image data; and

detecting one or more of defects, particles, or contamination on the bonding tip surface of the bonding tip based at least in part on the analysis of the second image data.

5. The apparatus of claim 4, further comprising instructions that, when executed by the at least one processor, cause the controller to:

determining a position of the pick surface of the pick arm relative to the engagement tip surface of the engagement tip based at least in part on the analysis of at least one of the first image data and the second image data; and

determining whether the pick surface is parallel to the engagement tip surface based at least in part on the determined position of the pick surface of the pick arm relative to the engagement tip surface of the engagement tip.

6. The apparatus of claim 4, further comprising instructions that, when executed by the at least one processor, cause the controller to: initiating a remedial action in response to detecting a defect, particle, or contamination on one or more of the pick surface of the pick arm or the engaging tip surface of the engaging tip.

7. The apparatus of claim 5, further comprising instructions that, when executed by the at least one processor, cause the controller to: initiating a remedial action in response to determining that the pickup surface is not parallel to the engagement tip surface.

8. The apparatus of claim 6, wherein initiating a remedial action comprises at least one of: cleaning one or more of the pick surface or the engagement tip surface; replacing one or more of the pick arm or the engagement tip; or trigger an alarm.

9. The apparatus of claim 7, wherein initiating remedial action comprises recalibrating one or more of the pick arm or the engagement tip.

10. The apparatus of claim 4, further comprising instructions that, when executed by the at least one processor, cause the controller to:

causing the pick arm to move to a first position in which the engagement tip surface is within a field of view of the second 3D sensor;

capturing image data of the engagement tip surface with the second 3D sensor;

causing the pick arm to move to a second position in which the pick surface is within a field of view of the first 3D sensor; and

capturing image data of the pickup surface with the first 3D sensor.

11. A method, comprising:

moving a pick arm to a first position relative to an engagement tip, wherein a pick surface of the pick arm is within a field of view of a first 3D sensor coupled to the engagement tip;

capturing first image data of the pickup surface via the first 3D sensor; and

analyzing the first image data of the pickup surface.

12. The method of claim 11, further comprising:

moving the pick arm to a second position relative to an engagement tip, wherein an engagement tip surface of the engagement tip is within a field of view of a second 3D sensor coupled to the pick arm;

capturing second image data of the engagement tip surface via the second 3D sensor; and

analyzing the second image data of the engaging tip surface.

13. The method of claim 12, further comprising:

determining an orientation of the pickup surface of the pickup arm relative to the engagement tip surface of the engagement tip based at least in part on the analysis of at least one of the first image data of the pickup surface and the second image data of the engagement tip surface; and

determining whether an angle defined between the pickup surface and the engagement tip surface is within an acceptable range based at least in part on the determined orientation of the pickup surface of the pickup arm relative to the engagement tip surface of the engagement tip.

14. The method of claim 13, further comprising: in response to determining that the angle defined between the pickup surface and the engagement tip surface is outside of an acceptable range, initiating a remedial action.

15. The method of claim 14, wherein initiating remedial action comprises recalibrating one or more of the pick arm or the engagement tip.

16. The method of claim 13, further comprising:

detecting one or more of a defect, a particle, or a contamination on the pickup surface of the pickup arm based at least in part on the analysis of the first image data; and

detecting one or more of defects, particles, or contamination on the bonding tip surface of the bonding tip based at least in part on the analysis of the second image data.

17. The method of any one of claims 12-16, further comprising:

detecting one or more of a defect, a particle, or a contamination on the pickup surface of the pickup arm based at least in part on the analysis of the first image data; and

detecting one or more of defects, particles, or contamination on the bonding tip surface of the bonding tip based at least in part on the analysis of the second image data.

18. The method of claim 17, further comprising: initiating a remedial action in response to detecting a defect, particle, or contamination on one or more of the pick surface of the pick arm or the engaging tip surface of the engaging tip.

19. The method of any of claims 12-16, wherein analyzing one or more of the first image data or the second image data comprises applying one or more machine learning techniques to one or more of the first image data or the second image data.

20. An apparatus for processing a microelectronic device, comprising:

a bonding head;

a bonding tip coupled to the bonding head and having a bonding tip surface configured to receive a microelectronic device thereon;

a pick-up arm drive;

a pick arm coupled to the pick arm drive and having a pick surface configured to receive the microelectronic device thereon; and

at least one 3D sensor configured to capture image data of the engagement tip surface and image data of the pickup surface.

Images