Method and apparatus for backside via exposure processing - Patents.com
The integrated method using a central controller and machine learning to optimize BVR process tool interactions addresses via bending and breaking issues, enhancing yield and performance by adjusting profile parameters across multiple steps.
Patent Information
- Application Number
- JP2023506120
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2020-07-31
- Filing Date
- 2021-07-29
- Publication Date
- 2024-11-13
- Estimated Expiration
- 2041-07-29
AI Technical Summary
The existing backside via exposure (BVR) process in semiconductor manufacturing is prone to via bending or breaking during the final chemical mechanical polishing (CMP) step due to inadequate control over the complex interaction of multiple process steps, leading to substrate damage and reduced yield.
An integrated method and apparatus that utilizes a central controller to receive and analyze process tool operating parameters and sensor data, applying machine learning to adjust profile parameters across multiple process tools, including CMP, etch, and CVD chambers, to prevent via failure by optimizing the BVR process flow.
The integrated approach enhances control over the BVR process, reducing via bending and breaking, thereby increasing yield and improving substrate performance by compensating for process variations and ensuring precise substrate profiles.
Smart Images

Figure 0007583148000001 
Figure 0007583148000002 
Figure 0007583148000003
Abstract
Description
[Technical field]
[0001] FIELD OF THE PRESENT PRIOR ART Embodiments of the present principles relate generally to semiconductor processing of semiconductor substrates. [Background technology]
[0002] During the formation of integrated circuits, multi-packaging such as 3D packaging may be used. In such cases, vertical connections or vias may be used to connect integrated circuits found at different levels of the packaged circuit. A via is essentially a vertical conductor that penetrates a substrate. To access the via from the backside of the substrate to make further connections, a chemical mechanical polishing or CMP process may be used to expose the surface of the via. A backside via exposure (BVR) process may include backside grinding, polishing, and etching to expose or expose the via so that it extends above the backside surface. A dielectric film may then be deposited over the via, and another CMP process may be used to expose the connection surface of the via on the backside surface after deposition. The inventors have noticed that during the final CMP process, the via may bend or break off, causing damage to the via and the backside surface.
[0003] Accordingly, the present inventors have provided an improved method and apparatus for increasing control of the backside via exposure process. Summary of the Invention
[0004] Methods and apparatus are provided herein for increasing control of the backside via exposure process.
[0005] In some embodiments, a method for performing a backside via exposure process may include receiving, by a central controller, process tool operating parameters from process tools involved in the backside via exposure process; receiving, by the central controller, metrology sensor data from at least one or more of the process tools involved in the backside via exposure process; and modifying, by the central controller, the backside exposure process based at least in part on the process tool operating parameters and the metrology sensor data by adjusting profile parameters of two or more of the process tools involved in the backside via exposure process, wherein the profile parameters are configured to prevent backside via damage during a chemical mechanical polishing (CMP) process.
[0006] In some embodiments, the method also includes storing, by a central controller, process tool operating parameters and metrology sensor data in a database, determining, by the central controller, adjustments for two or more of the process tools by using a machine learning model based at least in part on the process tool operating parameters and the metrology sensor data, and transmitting and receiving, by the central controller, the metrology sensor data between two or more of the process tools involved in the backside via exposure process as the substrate is processed to automatically adjust the two or more of the process tools before a subsequent substrate is processed in the two or more of the process tools, wherein the backside via exposure process temporarily bonds the substrate to a backside of the substrate and backplates the substrate. the backside of the substrate includes a first chemical mechanical polishing (CMP) process, a second CMP process, and a third CMP process to expose the vias on the backside of the substrate, the backside of the substrate includes a first recess around the vias, a second recess around the vias, and a third CMP process to expose the vias on the backside of the substrate, the sensor metrology data for the bonding and backgrinding includes temporary bonding voids, amount of warpage, total thickness variation (TTV), and surface roughness data, the sensor metrology data for the CMP process includes total thickness variation (TTV), filled via profile, surface roughness, and via breakage data, the sensor metrology data for the etching includes recess depth and via height, via breakage, and surface roughness data, and / or the sensor metrology data for the deposition includes via nail height distribution and thickness uniformity, layer delamination, and seam data.
[0007] In some embodiments, a method for performing a backside via exposure process may include receiving, by a central controller, an initial profile of a substrate and a completed profile of the substrate for the backside via exposure process; receiving, by the central controller, process tool operating parameters from process tools involved in the backside via exposure process; receiving, by the central controller, sensor metrology data from at least one of the process tools involved in the backside via exposure process; determining, by the central controller, two or more individual process tool recipes based on the initial profile, the completed profile, the process tool operating parameters, and the sensor metrology data, where the individual process tool recipes are configured to distributely modify the initial profile to generate the completed profile; and sending, by the central controller, the two or more individual process tool recipes to associated process tools involved in the backside exposure process, resulting in profile modifications to the substrate.
[0008] In some embodiments, the method also includes storing, by the central controller, the process tool operating parameters and the metrology sensor data in a database; determining, by the central controller, an individual process tool recipe by using a machine learning model based at least in part on the process tool operating parameters; transmitting and receiving, by the central controller, the metrology sensor data between two or more of the process tools involved in the backside exposure process as the substrate is processed to automatically adjust the individual process tool recipe by effecting changes before a subsequent substrate is processed in the two or more of the process tools, and / or the backside exposure process includes temporarily bonding the substrate and backgrinding the backside of the substrate; and performing a first chemical-mechanical test on the backside of the substrate. The method may include performing a polishing (CMP) process, etching the backside of the substrate to create recesses around the vias, depositing a passivation layer on the backside of the substrate, and performing a second CMP process to expose the vias on the backside of the substrate; and / or the sensor metrology data for bonding and backgrinding includes temporary bonding voids, amount of warping, total thickness variation (TTV), and surface roughness data; the sensor metrology data for the CMP process includes total thickness variation (TTV), filled via profile, surface roughness, and via breakage data; the sensor metrology data for etching includes recess depth and via height, via breakage, and surface roughness data; and the sensor metrology data for deposition includes via nail height distribution and thickness uniformity, layer delamination, and seam data.
[0009] In some embodiments, a non-transitory computer-readable medium having instructions stored thereon that, when executed, cause the implementation of a method for performing a backside exposure process, the method may include receiving, by a central controller, process tool operating parameters from process tools involved in the backside via exposure process; receiving, by the central controller, sensor metrology data from at least one or more of the process tools involved in the backside via exposure process; and modifying, by the central controller, the backside exposure process based at least in part on the process tool operating parameters and the sensor metrology data by adjusting profile parameters of two or more of the process tools involved in the backside via exposure process, wherein the profile parameters are configured to prevent backside via damage during a chemical mechanical polishing (CMP) process.
[0010] In some embodiments, the non-transitory computer readable medium may include a method for determining, by a central controller, adjustments for two or more of the process tools by using a machine learning model based on the process tool operating parameters and the sensor metrology data, or for transmitting and receiving, by the central controller, sensor metrology data between two or more of the process tools involved in the backside via exposure process as a substrate is processed to automatically adjust the two or more of the process tools before a subsequent substrate is processed in the two or more of the process tools, temporarily bonding the substrate and backgrinding the backside of the substrate, and determining, by the central controller, adjustments for two or more of the process tools by using a machine learning model based on the process tool operating parameters and the sensor metrology data, and / or for transmitting and receiving sensor metrology data between two or more of the process tools involved in the backside via exposure process as a substrate is processed to automatically adjust the two or more of the process tools before a subsequent substrate is processed in the two or more of the process tools. to create a recess around the via, depositing a passivation layer on the back side of the substrate, and performing a CMP process to expose the via on the back side of the substrate; and / or the sensor metrology data for bonding and backgrinding includes temporary bonding voids, amount of warpage, total thickness variation (TTV), and surface roughness data; the sensor metrology data for the CMP process includes total thickness variation (TTV), filled via profile, surface roughness, and via breakage data; the sensor metrology data for etching includes recess depth and via height, via breakage, and surface roughness data; and the sensor metrology data for deposition includes via nail height distribution and thickness uniformity, layer delamination, and seam data.
[0011] Other and further embodiments are disclosed below.
[0012] The embodiments of the present principles, briefly summarized above and discussed in more detail below, can be understood by referring to the exemplary embodiments of the principles shown in the attached drawings. However, the attached drawings illustrate only typical embodiments of the principles and therefore should not be considered limiting in scope, since the principles may admit of other equally effective embodiments. [Brief description of the drawings]
[0013] [Figure 1] 1A-1C illustrate methods for performing a backside via exposure process in accordance with some embodiments of the present principles. [Diagram 2] 1A-1C are cross-sectional views of a substrate during a backside via exposure process, in accordance with some embodiments of the present principles. [Diagram 3] 11A-11C are cross-sectional views of a substrate after a second chemical mechanical process, according to some embodiments of the present principles. [Figure 4] FIG. 1 illustrates how a backside via exposure process can be performed to provide increased control of the BVR process, in accordance with some embodiments of the present principles. [Diagram 5] 1 is a cross-sectional view of a framework for integrating BVR process flows, in accordance with some embodiments of the present principles; [Figure 6] 1 is a cross-sectional view of a framework for performing a backside via exposure process, in accordance with some embodiments of the present principles. [Figure 7] 1A-1C illustrate a method for performing a backside via exposure process that matches the finished profile in accordance with some embodiments of the present principles. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0014] For ease of understanding, the same reference numbers have been used, where possible, to designate identical elements common to each of the figures. The figures are not to scale and may be simplified for clarity. Elements and features of one embodiment may be advantageously incorporated in other embodiments without further description.
[0015] The method and apparatus provide a solution to via bending or breaking that can occur during the final chemical mechanical polishing (CMP) step of a backside via exposure (BVR) process. Often, the BVR process relies solely on the final CMP to correct the substrate profile, and drastic correction inevitably leads to via damage. The BVR process is a complex interaction between several steps in the process, which may include, for example, a bonding process, a backgrinding process, an etching process, a deposition process, and two or more polishing processes. By integrating the BVR process and distributing the substrate profile changes across the BVR process flow, via bending or breaking can be prevented.
[0016] Metrology data can be leveraged to determine the profile of the incoming substrate and changes to the profile of the substrate along the BVR process flow and after the BVR process is completed. Sensors to receive metrology data can be placed on the process tools used in the BVR process and / or at the factory interface or load lock of the integrated process tool platform, etc. The metrology data can be acquired by a central controller and used to adjust profile settings at each process tool of the BVR process in real time and / or stored in a database for further processing. In some embodiments, machine learning based models can be used to determine the profile and / or operational changes for the process tools used in the BVR process. The framework of the present principles is not limited to only the BVR process and can also be incorporated into other process flows such as, but not limited to, via middle, via last, redistribution layer (RDL), copper pillar, and / or damascene process flows.
[0017] FIG. 1 is a method 100 for performing a BVR process, according to some embodiments. In some embodiments, block 104 is an optional process. FIG. 2 shows a cross-sectional view of a substrate during a BVR process, according to some embodiments, and will be discussed together with the method 100 of FIG. 1. In block 102, a substrate 218 is temporarily bonded to a carrier 202 via an adhesive layer 204 and undergoes a backgrinding process as shown in FIGS. 200A and 200B. The substrate 218 may have pre-formed vias 210 with a barrier / liner layer 206 interposed between the vias 210 and the bulk material 208 of the substrate 218. The carrier 202 provides rigidity to the substrate 218 during processing and is removed at a later stage of processing. The backside surface 216A of the substrate 218 may be uneven and may require a backgrinding procedure to uniformize the backside surface 216A to the backside surface 216B of the substrate 218. Backgrinding may also reduce or eliminate any warpage in the substrate 218. In some cases, the carrier 202 may not be properly centered with respect to the substrate 218 and / or the backgrinding process may not properly remove gouges or other defects in the backside surface 216B, creating other problems as the substrate 218 progresses through the BVR process flow. In some cases, high levels of warpage may not be completely removed during the backgrinding process, potentially resulting in a low center profile and a high edge profile on the substrate 218.
[0018] In block 104, an optional first CMP process is performed on the backside surface 216B. The optional first CMP process polishes the backside surface 216B of the substrate 218 so that subsequent processes in the BVR process flow can achieve the desired profile for each of the processes. The degree of polishing by the optional first CMP process can directly affect the success of the subsequent etching process. Too little polishing and etching will not remove enough of the dielectric layer to expose the via 210. Too much polishing and etching will expose too much of the via and may cause the via to bend or break during the subsequent CMP process. In some embodiments, the optional first CMP process of block 104 is skipped and the substrate proceeds directly from the backgrinding process of block 102 to the etching process of block 106. In block 106, the backside of the substrate 218 is etched to create a recess around the via 210 as shown in FIG. 200C. A portion of the bulk material 208 is etched away to expose the via 210. The barrier / liner layer 206 still covers the via 210. The bulk material 208 is etched away to allow the passivation layer 212 to be deposited over the via 210. If the etch rate is too slow, the via 210 will not be exposed and the passivation layer 212 will deposit over the bulk material 208 and not cover the via 210. If the etch rate is too aggressive, the subsequent passivation layer may not be thick enough to completely encompass the via 210, resulting in poor performance and even bending and breaking of the via 210 in the subsequent CMP process.
[0019] In block 108, as shown in FIG. 200D, a passivation layer 212 is deposited on the substrate 218. In some embodiments, the deposition process is performed in a CVD processing chamber or the like. As previously described, the success of the passivation layer deposition process is influenced by previous processes, such as an optional first CMP process and an etching process. In block 110, as shown in FIG. 200E, a second CMP process is performed to expose or uncover the top surface 214 of the via 210. The top surface 214 of the via 210 is used for subsequent electrical connection to other circuits and / or as an input / output to existing circuits on the substrate 218. As discussed above, the success of the second CMP process is highly dependent on the previous processes in the BVR process flow.
[0020] FIG. 3 illustrates a cross-sectional view 300 of a substrate after a second CMP process, according to some embodiments. When a previous process in the BVR process flow creates a situation that the second CMP process profile cannot fully compensate for (backgrinding cannot remove warpage, incorrect optional first CMP process profile, etching too deep / shallow, passivation layer too thin, etc.), bending and / or breaking of the via 210 may occur. In the cross-sectional view 300 of FIG. 3, a gorge 302 on the backside of the substrate 218 is created, compromising the substrate 218. Although the second CMP process may be able to compensate for small process tolerances in each of the previous process tools, the inventors have realized that a cumulative effect that can occur when two or more processes have profiles that produce a substrate surface that the second CMP process cannot compensate for can cause large differences, leading to damage to the substrate during the second CMP process.
[0021] The inventors have discovered that by using an integrated approach across the BVR process flow, bent or broken vias during the second CMP process can be eliminated, increasing yield and improving performance parameters of the substrates produced. FIG. 4 is a method 400 of implementing a BVR process to increase control of the BVR process, according to some embodiments. FIG. 5 is a cross-sectional view of a framework 500 for integrating a BVR process flow, according to some embodiments, and is used in discussing the method 400. In block 402, a central controller 502 receives process tool operating parameters from process tools involved in the BVR process flow. The process tool operating parameters may include, but are not limited to, minimum and / or maximum processing times, minimum and / or maximum etch rates, minimum and / or maximum deposition rates, minimum and / or maximum removal rates, minimum and / or maximum chamber pressures, minimum and / or maximum chamber temperatures, and / or slurry types, etc.
[0022] At block 404, in some embodiments, the central controller 502 receives sensor metrology data from at least one or more of the process tools involved in the BVR process flow. In some embodiments, the central controller 502 receives sensor metrology data from at least two or more of the process tools involved in the BVR process flow. Examples of different types of metrology data are discussed in more detail below. The central controller 502 is in communication with a first controller 508 of the first process tool 504 and a second controller 512 of the second process tool 506. In some embodiments, the first process tool 504 may have a first metrology sensor 510 that communicates (522) metrology data to the first controller 508. The first controller 508 communicates (526) first process tool operating parameters and / or metrology data, if available, to the central controller 502. Similarly, the second process tool 506 may have a second metrology sensor 514 that communicates (524) metrology data to the second controller 512. The second controller 512 communicates (528) the second process tool operating parameters and / or metrology data, if available, to the central controller 502. As discussed above, in some embodiments, the central controller 502 may also be in communication with metrology sensors that are not embedded in the process tooling, such as factory interfaces and / or load locks. In block 406, the central controller 502 may modify the BVR process based at least in part on the process tool operating parameters and / or sensor metrology data by adjusting profile parameters of two or more of the process tools involved in the BVR process flow. The profile parameters are configured to prevent backside via damage during a chemical mechanical polishing (CMP) process.
[0023] In some embodiments, the central controller 502 may be in communication 534 with a data store 516 in which the central controller 502 may store metrology data and / or process tool operating parameters. In some embodiments, the central controller 502 may be in communication (538) with a model 518, which is in communication (536) with the data store 516. The model may be based on machine learning techniques, etc., and may utilize historical metrology data, real-time data, and / or process tool operating parameters, etc. In some embodiments, the central controller 502 may also be in communication (540, 542) with other process tools 520. The central controller 502 may use the model 518 and / or historical data stored in the data store 516 to facilitate determining how process profiles of the first process tool 504 and the second process tool 506 may be altered to prevent damage during the CMP process. Because the central controller 502 knows the process tool operating parameters and / or metrology data associated with each of the process outputs of the process tools, the central controller 502 can make a determination as to whether the combined process tool processes will produce a substrate profile that cannot be compensated for by the CMP process to prevent damage to the substrate as it progresses through the BVR process flow. After the determination, the central controller 502 communicates (530) any profile adjustments to the first process tool 504 and / or communicates (532) any profile adjustments to the second process tool 506. The profile adjustments can be made before the substrate run (e.g., based on historical metrology data and / or based on an initial profile - e.g., a substrate with significant bow that requires a profile change that differs from a profile for an untwisted or less twisted substrate) and / or in real time after the substrate run. In some embodiments, the central controller 502 can have an interface 556 that allows for input of a profile, e.g., an initial profile and / or a finished profile for the substrate.In some embodiments, the interface 556 may be a user interface for board profile submission and / or a digital and / or analog interface for use with other computing devices for board profile submission, or the like.
[0024] Herein, the first controller 508 and the second controller 512 control the operation of the first process tool 504 and the second process tool 506 or chamber, respectively, using direct and / or indirect control of the process tools. During operation, the first controller 508 and the second controller 512 enable data collection from the respective process tools and systems to optimize performance. The first controller 508 and the second controller 512 generally include a central processing unit (CPU) 550, a memory 552, and support circuits 554, respectively. The CPU 550 may be any form of general-purpose computer processor that may be used in an industrial environment. The support circuits 554 are coupled to the CPU 550 as usual and may include cache, clock circuits, input / output subsystems, power supplies, etc. Software routines, such as the methods described above or below, when stored in memory and executed by the CPU 550, may convert the CPU 550 into a special-purpose computer (controllers 508, 512).
[0025] The memory 552 is a form of computer-readable storage medium that includes instructions that, when executed by the CPU 550, facilitate the operation of semiconductor processes and equipment. The instructions in the memory 552 are in the form of a program product, such as a program that implements the method of the present principles. The program code may conform to any one of several different programming languages. As an example, the present disclosure may be implemented as a program product stored on a computer-readable storage medium for use with a computer system. The program of the program product defines the functions of the aspects (including the methods described herein). Exemplary computer-readable storage media include, but are not limited to, non-writeable storage media in which information is permanently stored (e.g., a read-only memory device in a computer, such as a CD-ROM disk readable by a CD-ROM drive, a flash memory, a ROM chip, or any type of solid-state non-volatile semiconductor memory), and writeable storage media in which changeable information is stored (e.g., a floppy disk in a diskette drive, or a hard disk drive, or any type of solid-state random access semiconductor memory). Such computer-readable storage media, when carrying computer-readable instructions that direct the functions of the methods described herein, are aspects of the present principles.
[0026] As used herein, the central controller 502 controls the operation of the process tools using direct control of the process tools and / or alternatively indirectly by controlling a computer (or controller) associated with the process tools. In operation, the central controller 502 enables data collection from each process tool to optimize the performance of the BVR process flow. The central controller 502 generally includes a central processing unit (CPU) 544, a memory 546, and support circuits 548. The CPU 544 may be any form of general-purpose computer processor that may be used in an industrial environment. The support circuits 548 are conventionally coupled to the CPU 544 and may include cache, clock circuits, input / output subsystems, power supplies, etc. Software routines, such as the methods described above, when stored in memory and executed by the CPU 544, may transform the CPU 544 into a special-purpose computer (the central controller 502).
[0027] The memory 546 of the central controller 502 is in the form of a computer-readable storage medium that includes instructions that, when executed by the CPU 544, facilitate the operation of semiconductor processes and equipment. The instructions in the memory 546 are in the form of a program product, such as a program that implements the methods of the present principles. The program code may conform to any one of several different programming languages. As an example, the present disclosure may be implemented as a program product stored on a computer-readable storage medium for use with a computer system. The program of the program product defines the functions of the aspects (including the methods described herein). Exemplary computer-readable storage media include, but are not limited to, non-writeable storage media in which information is permanently stored (e.g., a read-only memory device in a computer, such as a CD-ROM disk readable by a CD-ROM drive, a flash memory, a ROM chip, or any type of solid-state non-volatile semiconductor memory), and writeable storage media in which changeable information is stored (e.g., a floppy disk in a diskette drive, or a hard disk drive, or any type of solid-state random access semiconductor memory). Such computer-readable storage media, when carrying computer-readable instructions that direct the functions of the methods described herein, are aspects of the present principles.
[0028] 6 is a cross-sectional view 600 of a framework for performing a BVR process, according to some embodiments. In the illustrated framework, a central controller 502 is in communication 610 with a bonding and backgrinding process tool 602 that bonds a substrate to a carrier and backgrinds the substrate in preparation for a subsequent BVR process. In some embodiments, the carrier is formed of a glass material. The central controller 502 may receive metrology data from the bonding and backgrinding process tool 602, for example, regarding temporary bonding voids, bow, total thickness variation (TTV), and / or surface roughness of the substrate. The central controller 502 may incorporate metrology data from the bonding and backgrinding process tool 602 for temporary bonding quality assurance and zonal substrate CMP control.
[0029] The central controller 502 is also in communication 612 with a CMP process tool 604 that provides an initial polishing process to smooth the backside surface of the substrate and a final polishing process to expose vias on the backside surface of the substrate. The central controller 502 may receive metrology data from the CMP process tool 604 for the initial polishing process, such as for TTV, filled via profile, and / or surface roughness. In some embodiments, for example, the CMP process tool 604 may incorporate a near-infrared sensor that may map thickness across the substrate and provide TTV metrology data, etc. The central controller 502 may incorporate metrology data from the CMP process tool 604 for incoming through-silicon via (TSV) etch uniformity and / or substrate recess etch, etc. The central controller 502 may also use the metrology data to facilitate adjusting the BVR process flow and process tool profile, such as for center-to-edge TTV control. The central controller 502 may receive metrology data from the CMP process tool 604 for the final polishing process, such as for via bending and / or breakage. The central controller 502 may incorporate metrology data from the CMP process tool 604, such as for substrate recess etching by an etch process tool 606 and / or passivation layer deposition by a CVD process tool 608. The central controller 502 may also use the metrology data to facilitate adjusting BVR process flows and process tool profiles, such as for center-edge down force control.
[0030] The central controller 502 is also in communication (614) with an etch process tool 606 for etching the substrate around the via. The central controller 502 may receive metrology data from the etch process tool 606 regarding recess depth / via height, via bending and / or fracture, and / or surface roughness, etc. The central controller 502 may incorporate metrology data from the etch process tool 606 for passivation layer deposition and / or initial CMP polishing of the substrate, etc. The central controller 502 may also use the metrology data to facilitate adjusting the BVR process flow and process tool profile for M or W shaped plasma profile for center-edge deposition thickness compensation.
[0031] The central controller 502 is also in communication (616) with a CVD process tool 608 for depositing a passivation layer on the substrate. The central controller 502 may receive metrology data from the CVD process tool 608 regarding TSV nail height distribution / thickness uniformity, layer delamination, seams, etc. The central controller 502 may incorporate the metrology data from the CVD process tool 608 for final CMP polishing of the substrate and / or substrate recess etching, etc. The central controller 502 may also use the metrology data to facilitate adjusting the BVR process flow and process tool profiles for film thickness control.
[0032] By having access to metrology data, process tool recipes, and / or other process tool operating parameters, the central controller 502 can distribute the processing of substrates across the BVR process flow without requiring a single process tool, such as a CMP tool, to compensate at the end of the flow for all necessary corrections to the substrate. The central controller 502 can also use and / or distribute metrology data and profile information between two or more of the process tools to facilitate recipe adjustments for a given profile of the process tool. The central controller 502 can also use the metrology data to calibrate and fine-tune the BVR process flow before, during, and / or after each processed substrate. For example, it may be found that a slurry mixture used in an optional first CMP process is too aggressive, causing more material to be removed from the substrate than expected. The central controller 502 can receive metrology data from the CMP process tool, determine correction profile adjustments to distribute across two or more subsequent process tools, and communicate the distributed changes in real time to the two or more subsequent process tools to accommodate unexpected results during the optional first CMP process. In another example, the etch process may be found to be overly aggressive. The central controller 502 may receive each depth / via height metrology data from the etch process tool and incorporate the changes into a post-correction profile adjustment, which may be distributed, for example, as an increased deposition rate in the deposition process, a decreased polishing rate in the second CMP process, and so on. Similarly, the central controller 502 may make adjustments with a higher edge etch rate (e.g., with an etch process tool) and / or a higher center deposition rate (e.g., with a CVD deposition process tool) for substrates with higher edge thicknesses.
[0033] The inventors have also discovered that the integrated approach of the present principles allows for individual process tool recipes or substrate profiles to be determined based on initial and final substrate profiles. As used herein, "profile" or "substrate profile" refers to the physical parameters of a substrate, and "recipe" or "process tool recipe" refers to tool operating parameter settings that allow a process tool to modify the substrate's profile. Each individual process tool in the BVR process flow may utilize the incoming substrate's incoming profile or physical parameters and the outgoing substrate's outgoing profile or physical parameters. When the incoming substrate or outgoing substrate profile changes, the process tool recipe is modified to compensate for the profile change. If one of the process tools in the BVR process flow unexpectedly modifies the outgoing substrate profile (e.g., an incorrect recipe, or possibly an incoming substrate profile that is out of range or not as expected (e.g., a twisted substrate)), the modified substrate profile from that process tool will have a cumulative effect as the substrate progresses through the BVR process flow.
[0034] 7 is a method 700 for performing a BVR process that matches an initial substrate profile to a completed substrate profile, according to some embodiments. In block 702, a central controller receives an initial profile of the substrate for the BVR process and a completed profile of the substrate. The initial profile may be from actual metrology measurements of the substrate and / or from an expected profile (with some margin of error, etc.). In block 704, the central controller receives process tool operating parameters from the process tools involved in the BVR process flow. As previously mentioned, the process tool operating parameters may include, but are not limited to, minimum and / or maximum processing times, minimum and / or maximum etch rates, minimum and / or maximum deposition rates, minimum and / or maximum removal rates, minimum and / or maximum chamber pressures, minimum and / or maximum chamber temperatures, and / or slurry types, etc.
[0035] In block 706, the central controller receives sensor metrology data from at least one of the process tools involved in the BVR process flow. The sensor metrology data may include historical metrology data from a previous substrate run. In block 708, the central controller determines two or more individual process tool recipes based on the initial profile, the completed profile, the process tool operating parameters, and the sensor metrology data, where the individual process tool recipes are configured to modify the substrate in a distributed manner to generate the completed profile. Even if the initial profile of the substrate is inaccurate (not based on actual metrology data), the BVR process flow can be adjusted to compensate for differences detected by metrology sensors from the process tools. The metrology data can then be used to facilitate updating recipes used by the process tools, resulting in profile changes that help compensate for the unexpected initial profile metrology. In some embodiments, machine learning based models can be used to assist in determining the individual tool recipes and profiles. In block 710, the central controller sends the two or more individual process tool recipes to the associated process tools involved in the BVR process flow, resulting in profile changes to the substrate being processed.
[0036] Embodiments according to the present principles may be implemented in hardware, firmware, software, or any combination thereof. Embodiments may also be implemented as instructions stored using one or more computer-readable media, which may be read and executed by one or more processors. A computer-readable medium may include any mechanism for storing or transmitting information in a form readable by a machine (e.g., a computing platform, or a "virtual machine" running on one or more computing platforms). For example, a computer-readable medium may include any suitable form of volatile or non-volatile memory. In some embodiments, a computer-readable medium may include a non-transitory computer-readable medium.
[0037] While the forgoing is directed to embodiments of the present principles, other and further embodiments of the principles may be devised without departing from the basic scope thereof.
Claims
1. A method for performing a backside via exposure process comprising a plurality of steps, comprising: The steps include: temporarily bonding the substrate with the vias formed therein and backgrinding the backside of the substrate; Etching the backside of the substrate to create a recess around the via; depositing a passivation layer on the back side of the substrate; performing a chemical mechanical polishing (CMP) process to remove a portion of the passivation layer and expose the via on the backside of the substrate; Including, The method comprises: receiving, by a central controller, process tool operating parameters from a process tool associated with the backside via exposure process, the process tool operating parameters including at least one of a minimum processing time, a maximum processing time, a minimum etch rate, a maximum etch rate, a minimum deposition rate, a maximum deposition rate, a minimum removal rate, a maximum removal rate, a minimum chamber pressure, a maximum chamber pressure, a minimum chamber temperature, a maximum chamber temperature, or a slurry type; receiving, by the central controller, sensor metrology data from at least one or more of the process tools involved in the backside via exposure process; modifying, by the central controller, at least two steps of the plurality of steps of the backside via exposure process to prevent backside via damage during the chemical mechanical polishing (CMP) process based on the process tool operating parameters and the sensor metrology data; The method includes:
2. 10. The method of claim 1, wherein the process tools involved in the backside via exposure process include a CMP chamber, an etch chamber, and a chemical vapor deposition (CVD) chamber.
3. The method of claim 1 , further comprising storing, by the central controller, the process tool operating parameters and the sensor metrology data in a database.
4. 10. The method of claim 1, comprising determining, by the central controller, changes to the at least two processes by using machine learning models based on the process tool operating parameters and the sensor metrology data.
5. 10. The method of claim 1, further comprising: transmitting and receiving, by the central controller, sensor metrology data between two or more of the process tools involved in the backside via exposure process as a substrate is processed, to automatically adjust the two or more of the process tools before a subsequent substrate is processed in the two or more of the process tools.
6. 2. The method of claim 1, wherein the steps of the backside via exposure process further include performing a chemical mechanical polishing (CMP) process on the backside of the substrate after the step of backgrinding the backside of the substrate and before the step of creating the recess around the via.
7. The method of claim 1 , wherein the sensor metrology data for the bonding and backgrinding includes temporary bonding voids, amount of warpage, total thickness variation (TTV), and surface roughness data.
8. 2. The method of claim 1, wherein the sensor metrology data for the chemical mechanical polishing (CMP) process includes total thickness variation (TTV), filled via profile, surface roughness, and via breakage data.
9. The method of claim 1 , wherein the sensor metrology data for etching includes recess depth and via height, via breakage, and surface roughness data.
10. The method of claim 1 , wherein the sensor metrology data for a deposition includes via nail height distribution and thickness uniformity, layer delamination, and seam data.
11. 1. A non-transitory computer readable medium having stored thereon instructions that, when executed, cause a method for performing a backside via exposure process including a plurality of steps, comprising: The steps include: temporarily bonding the substrate with the vias formed therein and backgrinding the backside of the substrate; Etching the backside of the substrate to create a recess around the via; depositing a passivation layer on the back side of the substrate; performing a chemical mechanical polishing (CMP) process to remove a portion of the passivation layer and expose the via on the backside of the substrate; Including, The method comprises: receiving, by a central controller, process tool operating parameters from a process tool associated with the backside via exposure process, the process tool operating parameters including at least one of a minimum processing time, a maximum processing time, a minimum etch rate, a maximum etch rate, a minimum deposition rate, a maximum deposition rate, a minimum removal rate, a maximum removal rate, a minimum chamber pressure, a maximum chamber pressure, a minimum chamber temperature, a maximum chamber temperature, or a slurry type; receiving, by the central controller, sensor metrology data from at least one or more of the process tools involved in the backside via exposure process; modifying, by the central controller, at least two steps of the plurality of steps of the backside via exposure process to prevent backside via damage during the chemical mechanical polishing (CMP) process based on the process tool operating parameters and the sensor metrology data; A non-transitory computer readable medium comprising:
12. A non-transitory computer-readable medium as described in claim 11, wherein the method further includes storing, by the central controller, the process tool operating parameters and the sensor measurement data in a database.
13. A non-transitory computer-readable medium as described in claim 11, wherein the method includes determining, by the central controller, changes to the at least two processes by using a machine learning model based on the process tool operating parameters and the sensor measurement data.
14. The non-transitory computer-readable medium of claim 11, the method further comprising: the central controller transmitting and receiving sensor measurement data between two or more of the process tools involved in the backside via exposure process as a substrate is processed, and automatically adjusting the two or more of the process tools before a subsequent substrate is processed in the two or more of the process tools.
15. 12. The non-transitory computer-readable medium of claim 11, wherein the steps of the backside via exposure process further comprise performing a chemical mechanical polishing (CMP) process on the backside of the substrate after the step of backgrinding the backside of the substrate and before the step of creating the recess around the via.
16. 12. The non-transitory computer readable medium of claim 11, wherein the sensor metrology data for the bonding and backgrinding includes temporary bonding voids, amount of warpage, total thickness variation (TTV), and surface roughness data.
17. The non-transitory computer-readable medium of claim 11, wherein the sensor measurement data for the chemical mechanical polishing (CMP) process includes total thickness variation (TTV), buried via profile, surface roughness, and via breakage data.
18. The non-transitory computer-readable medium of claim 11, wherein the sensor measurement data for etching includes recess depth and via height, via breakage, and surface roughness data.
19. The non-transitory computer-readable medium of claim 11, wherein the sensor measurement data for deposition includes via nail height distribution and thickness uniformity, layer delamination, and seam data.
Citation Information
Patent Citations
Apparatus and method for acoustic monitoring and control of TSV exposure process
JP2016517185A
In-Situ Acoustic Monitoring of Chemical Mechanical Polishing
US20160013085A1
