Computer-implemented method for polishing substrates and matching polishing performance between polishing systems

Abstract

Advanced substrate polishing methods are provided herein that use a machine learning artificial intelligence (AI) algorithm or a software application generated using AI to control one or more aspects of the polishing process. The AI algorithm is trained using substrate processing data acquired from a polishing system to simulate the polishing process and make predictions about the polishing process and process the expected results based on the predictions.

Description

Technical Field

[0001] Embodiments described herein generally relate to semiconductor device fabrication, and more particularly to chemical mechanical polishing (CMP) systems and related methods used in semiconductor device fabrication. Background Art

[0002] Chemical mechanical polishing (CMP) is commonly used in the manufacture of high-density integrated circuits to planarize a layer of material on a substrate, remove excess material from the surface of an underlying material layer, or both. In a typical CMP process, a substrate is held in a carrier head that presses the back side of the substrate against a rotating polishing pad in the presence of a polishing fluid. The polishing pad is typically formed of a polymeric material having a surface roughness that facilitates the delivery of the polishing fluid to the interface between the material surface of the substrate and a moving polishing pad disposed thereunder. The polishing fluid generally comprises an aqueous solution of one or more chemical components and nano-scale abrasive particles suspended in the aqueous solution, commonly referred to as a polishing slurry. Material is removed across the surface of the material layer of the substrate by a combination of chemical and mechanical action provided by the polishing fluid, the relative motion of the substrate and the polishing pad, and the contact pressure therebetween. Consumables (e.g., polishing pad and polishing fluid) are selected based on the desired CMP application.

[0003] Common CMP applications include bulk film planarization and removal of excess material in damascene processing. Bulk film planarization (e.g., interlayer dielectric (ILD) polishing) is generally used to smooth out undesirable concavities and convexities in the surface of a material layer caused by underlying two-dimensional or three-dimensional features. Typical damascene CMP applications include shallow trench isolation (STI) and interlayer metal interconnect formation, where CMP is used to remove trench, contact, via, or line fill material (capping layer) from the exposed surface (field) of one or more underlying layers containing STI or metal interconnect features.

[0004] Depending on the application, CMP process results are typically characterized by a combination of interrelated metrics related to global polishing uniformity, local planarization performance, and CMP-induced surface defectivity. These process results can determine the performance, reliability, and / or operability of the resulting components formed on the substrate. Process results outside of process tolerance limits can lead to component failure, thereby inhibiting the yield of usable components formed on the substrate. Generally speaking, as circuit density increases and component feature sizes decrease, process result tolerances decrease.

[0005] To meet industry demands for shrinking component geometries, the complexity of advanced CMP systems has increased dramatically to provide control over virtually all process variables (parameters) known to influence process outcomes. Such advanced CMP systems include highly engineered and complex subsystems, each configured to control one or more process parameters to a desired set point. The controllable process parameters collectively define a substrate polishing recipe. Typically, a polishing recipe for a single substrate CMP process includes a multi-stage polishing sequence, where one or more parameter set points are varied for each stage of the sequence.

[0006] Unfortunately, to date, advances in CMP technology have far outpaced scientific understanding of the complex interactions of chemical and mechanical activity between surfaces, fluids, and abrasives at the polishing interface. As a result, existing CMP models are generally not suitable for use in process development. Consequently, CMP substrate processes are typically determined and / or improved using conventional process development and improvement techniques. Examples of such techniques include design of experiments (DOE) and trial and error. In general, standard quality control measures prohibit the performance of experiments on production substrates with components intended for use or sale. As a result, DOE experiments are typically performed using expensive test substrates while consuming valuable CMP process system time. Consequently, due to the time and cost associated therewith, it is practically impossible to thoroughly explore the complex relationships between polishing parameters, algorithms, consumables, component characteristics, and process results for the many individual polishing processes used in a production facility.

[0007] Thus, conventional process improvement approaches are not suitable for utilizing the combined capabilities of the devices and subsystems of advanced CMP processing systems and fail to provide the improved process results and wider process margins that would otherwise be achievable therewith.

[0008] Therefore, there is a need in the art for advanced processing methods that do not suffer from the above-mentioned disadvantages. Summary of the Invention

[0009] Embodiments of the present disclosure relate generally to chemical mechanical polishing (CMP) systems used in electronic device manufacturing, and more particularly to advanced substrate processing methods for use therewith.

[0010] In one embodiment, a computer-implemented method for generating a substrate polishing recipe is provided. The method includes polishing a substrate using a polishing system, comprising: (a) flowing a polishing fluid onto a surface of a polishing pad according to a polishing recipe, the polishing recipe comprising a plurality of polishing parameters and corresponding plurality of target values; (b) pressing a substrate against the surface of the polishing pad according to the polishing recipe; (c) maintaining a first polishing parameter of the plurality of polishing parameters at or near a target value for the first polishing parameter by adjusting a first control parameter; (d) generating processing system data, the processing system data comprising time-series data of the polishing recipe and the first control parameter; and (e) generating time-series in-situ result data concurrently with (a)-(d) using measurements obtained from an in-situ substrate monitoring system. The method further includes: repeating (a)-(e) for a plurality of substrates to obtain a corresponding plurality of training data sets, each of the training data sets comprising the processing system data and the in-situ result data for a polished substrate; receiving training data comprising the plurality of training data sets at an artificial intelligence (AI) training platform; training a machine learning AI algorithm using the training data; and changing one or more of the plurality of polishing parameters using the trained machine learning AI algorithm.

[0011] In one embodiment, a computer-readable medium includes instructions for executing a method for determining a polishing recipe. The method includes receiving training data comprising a plurality of training data sets at an artificial intelligence (AI) training platform, wherein each of the training data sets includes processing system data and in-situ result data associated with a substrate polished on a polishing system. The processing system data for each of the training data sets includes a polishing recipe comprising a plurality of polishing parameters and a corresponding plurality of target values; and time series data of a first control parameter used by a closed-loop control system to maintain a first polishing parameter of the plurality of polishing parameters at or near the target value, and the in-situ result data for each of the training data sets includes time series data generated using an in-situ substrate monitoring system. The method further includes training a machine learning AI algorithm using the training data; and determining a functional relationship between the in-situ result data and the time series data of the first control parameter using the trained machine learning AI algorithm.

[0012] In one embodiment, a computer-implemented method for matching polishing performance between polishing systems is provided. The computer-implemented method includes receiving training data comprising a plurality of training data sets at an artificial intelligence (AI) training platform. Each of the training data sets includes processing system data associated with a respective substrate of a first plurality of substrates polished using a first polishing system, wherein different substrates of the first plurality of substrates were polished using different combinations of substrate carrier assemblies from a plurality of substrate carrier assemblies and polishing stations from a plurality of polishing stations of the first polishing system. The processing system data for each of the training data sets includes a polishing recipe comprising a plurality of polishing parameters and a corresponding plurality of target values, wherein a corresponding closed-loop control system is used to maintain one or more of the plurality of polishing parameters at or near the target value of the one or more of the plurality of polishing parameters; and time series data of control parameters of the closed-loop control system. The method further includes training a machine learning AI algorithm using the training data. The trained machine learning AI algorithm is configured to identify differences between the different substrate carrier assemblies and / or the different polishing stations of the first polishing system. The method further includes implementing one or more corrective actions based on the identified differences.

[0013] Embodiments of the disclosure also provide a system of one or more computers that can be configured to perform specific operations or actions by installing software, firmware, hardware, or a combination thereof on the system, which, when operated, causes the system to perform the actions. One or more computer programs can be configured to perform specific operations or actions by including instructions that, when executed by a processor, cause a device to perform the actions. One general aspect includes a computer-implemented method for polishing a substrate within one or more polishing systems. The computer-implemented method may include: (a) flowing a polishing fluid onto a surface of a polishing pad according to a polishing recipe, the polishing recipe including a plurality of polishing parameters and corresponding plurality of target values; (b) pressing a substrate against the surface of the polishing pad according to the polishing recipe; (c) maintaining a first polishing parameter of the plurality of polishing parameters at or near the target value of the first polishing parameter by adjusting a first control parameter; (d) generating process system data, the process system data including time series data of the polishing recipe and the first control parameter; and (e) generating time series in situ result data using measurements obtained from an in situ substrate monitoring system concurrently with (a)-(d); repeating (a)-(e) for a plurality of substrates to obtain a corresponding plurality of training data sets, each of the training data sets including the process system data and the in situ result data for a polished substrate; receiving training data including the plurality of training data sets at an artificial intelligence (AI) training platform, wherein each of the plurality of training data sets is received in time sequence; and changing one or more of the plurality of polishing parameters based on analysis performed by a trained machine learning AI algorithm. Other embodiments of this aspect include corresponding computer systems, apparatus, and computer programs recorded on one or more computer storage components, each configured to perform the actions of the method.

[0014] Embodiments of the disclosure will also provide computer-implemented methods for polishing a substrate within one or more polishing systems. The computer-implemented method may include the following steps: (a) flowing a polishing fluid onto a surface of a polishing pad according to a polishing recipe, the polishing recipe including a plurality of polishing parameters and corresponding plurality of target values; (b) pressing a substrate against the surface of the polishing pad according to the polishing recipe; (c) maintaining a first polishing parameter of the plurality of polishing parameters at or near the target value of the first polishing parameter by adjusting a first control parameter; (d) generating process system data, the process system data including time series data of the polishing recipe and the first control parameter; and (e) generating time series in situ result data using measurements obtained from an in situ substrate monitoring system concurrently with (a)-(d); repeating (a)-(e) for a plurality of substrates to obtain a corresponding plurality of training data sets, each of the training data sets including the process system data and the in situ result data for a polished substrate; receiving training data including the plurality of training data sets at an artificial intelligence (AI) training platform, wherein at least a portion of the plurality of training data sets are received in time sequence; and changing one or more of the plurality of polishing parameters based on analysis performed by a machine learning AI algorithm.

[0015] Embodiments of the present disclosure also provide a computer-implemented method for matching polishing performance between polishing systems. The method includes receiving training data comprising a plurality of training data sets at an artificial intelligence (AI) training platform, wherein each of the training data sets comprises processing system data associated with respective substrates of a first plurality of substrates polished using a first polishing system, different substrates of the first plurality of substrates being polished using different combinations of substrate carrier assemblies from a plurality of substrate carrier assemblies and polishing stations from a plurality of polishing stations of the first polishing system, and the processing system data for each of the training data sets comprises a polishing recipe comprising a plurality of polishing parameters and corresponding plurality of target values, wherein a corresponding closed-loop control system is used to maintain one or more of the plurality of polishing parameters at or near the target value of the one or more of the plurality of polishing parameters, and time series data of control parameters of the closed-loop control system; and training a machine learning AI algorithm using the training data, wherein the trained machine learning AI algorithm is configured to identify differences between the different substrate carrier assemblies and / or the different polishing stations of the first polishing system, and implement one or more corrective actions based on the identified differences. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] A more detailed description of the disclosure, briefly summarized above, and the manner in which the above-described features of the disclosure may be understood in detail may be obtained by reference to embodiments, some of which are illustrated in the accompanying drawings. It is to be noted, however, that the drawings illustrate only typical embodiments of this disclosure and are therefore not to be considered limiting of its scope, for the disclosure may admit to other equally effective embodiments.

[0017] Figure 1A is a schematic cross-sectional view of a portion of a substrate illustrating undesirably poor local planarization performance.

[0018] Figure 1B is a schematic representation of a semiconductor component fabrication facility (Fab).

[0019] Figure 1C is a schematic representation of a machine learning artificial intelligence (AI) training system that may be used with the methods described herein, according to one embodiment.

[0020] Figure 1D is a schematic representation of an exemplary closed-loop feedback control system that may be used with the polishing systems described herein.

[0021] Figure 2A is a schematic side cross-sectional view of an exemplary polishing system that can be used to perform the methods described herein, according to one embodiment.

[0022] Figure 2B is a schematic side cross-sectional view of an exemplary substrate carrier.

[0023] Figure 2C Shown from different perspectives Figure 2A Schematic side cross-sectional view of a polishing system.

[0024] Figure 3 is a diagram illustrating a method of polishing a substrate according to one embodiment.

[0025] Figures 4A-4C are schematic cross-sectional views of a substrate illustrating different stages of a polishing process performed according to the methods set forth herein.

[0026] Figure 5 is a diagram illustrating a method for matching performance between different polishing systems according to one embodiment.

[0027] To facilitate understanding, identical reference numerals have been used, where possible, to designate identical components that are common to the figures. It is contemplated that components and features of one embodiment may be beneficially incorporated in other embodiments without further recitation. DETAILED DESCRIPTION

[0028] Embodiments of the present disclosure generally relate to chemical mechanical polishing (CMP) systems used in electronic device manufacturing, and more particularly to advanced substrate processing methods for use therewith.

[0029] Generally speaking, the advanced substrate processing methods herein use algorithms (e.g., machine learning artificial intelligence (AI) algorithms) or software applications generated using AI algorithms to control one or more aspects of the polishing process. Generally speaking, AI systems utilize large data sets with intelligent, iterative processing algorithms to learn from patterns and features in the data they analyze. Each time the AI system analyzes data by performing a round of data processing, it will generally test and measure its own performance and develop additional expertise based on the analysis performed. In this article, substrate processing data acquired from a polishing system is used to train an AI algorithm to simulate the polishing process, make predictions about the polishing process, and process the results expected based on the predictions.

[0030] In some embodiments, an AI algorithm or a software application generated using an AI algorithm is used to predict the time range of the desired polishing endpoint and based on this, adjust the composition of the polishing fluid through, for example, by starting, stopping or changing the flow rate of one or more polishing fluid components. As used herein, "polishing endpoint" refers to a point in the polishing process at which one or more substrate polishing parameters (e.g., slurry composition) may need to be changed, and does not necessarily represent the end of the polishing process. For inlay applications, the ability to accurately predict the desired polishing endpoint and preemptively adjust the polishing fluid composition (e.g., slurry composition) based on the prediction is beneficial for improving local planarization performance when compared to conventional reactive endpoint detection schemes. Improvements in local planarization performance result in desired improvements in the performance, reliability, and yield of the resulting components. Examples of poor local planarization that can be improved using the methods provided herein are shown in Figure 1A middle.

[0031] As discussed further below, a polishing fluid composition (e.g., a slurry composition) generally comprises a mixture of one or more types of solid particles suspended in a liquid (e.g., water). The solid particles are often referred to as abrasives and may include finely divided metal oxides, such as CeO2, Fe2O3, Al2O, and SiO2, suspended in a liquid. The liquid may include one or more of an acid, a base, and various additives (e.g., corrosion inhibitors, pH adjusters), typically disposed in water.

[0032] Figure 1Ais a schematic cross-sectional view illustrating poor localized planarization (e.g., erosion to a distance e and dishing to a distance d) after a polishing process used to remove an overlying layer of metal fill material from a field surface (i.e., an upper or outer surface) of a substrate 1. Here, substrate 1 features a dielectric layer 2, a first metal interconnect feature 3a formed in dielectric layer 2, and a plurality of second metal interconnect features 3b formed in dielectric layer 2. The plurality of second metal interconnect features 3b are closely spaced to form a region 4 with a relatively high feature density. Generally, metal interconnect features 3a, 3b are formed by depositing metal fill material onto dielectric layer 2 and into corresponding openings formed therein. A CMP process is then used to planarize the material surface of substrate 1 to remove the overlying layer of fill material from the field surface 5 of dielectric layer 2.

[0033] As shown, poor local planarization performance causes the upper surface of metal interconnect feature 3a to be recessed by a distance d relative to the surrounding surface of dielectric layer 2, also known as dishing. Poor local planarization performance also causes undesirable dishing (e.g., distance e) of dielectric layer 2 in high feature density region 4, where the upper surface of dielectric layer 2 in region 4 is recessed relative to the plane of field surface 5, also known as erosion. Metal loss caused by dishing and / or erosion may result in undesirable changes in the effective resistance of metal interconnect features 3a, 3b formed thereby, thereby affecting device performance and reliability.

[0034] In some embodiments, the AI algorithm is trained using data from one or more polishing systems operating in a production capacity (i.e., operating in a semiconductor device manufacturing facility). Using a production polishing system to train the AI algorithm advantageously provides a large amount of data that the AI algorithm can use to better understand the complex relationships between the many variables of a particular polishing application. An exemplary fabrication facility (Fab) 10 is schematically shown in FIG. Figure 1B middle.

[0035] Here, the Fab 10 includes multiple polishing systems 20, one or more machine learning artificial intelligence (AI) algorithm (hereinafter referred to as "AI") training platforms 30, a Fab production control system 40, one or more independent substrate inspection and / or metrology stations 50, and other processing systems 60. Other processing systems 60 include substrate processing systems used in the manufacture of semiconductor components, both upstream and downstream of the polishing process in the substrate processing flow, such as epitaxial systems, thermal processing systems, non-epitaxial deposition systems, lithography systems, etching systems, implant systems, and other polishing systems. In some embodiments, the Fab 10 further includes one or more electrical test systems 70, such as parameter test and / or component yield test systems, which communicate with the Fab production control system 40.

[0036] In general, each of the polishing systems 20 includes a plurality of polishing stations 21, a plurality of substrate carrier assemblies 22, a carrier loading station 23 for transferring substrates to and from the carrier assemblies 22, and a carrier transport system 24 for moving the substrate carrier assemblies 22 between the carrier loading station 23 and the different polishing stations 21. Here, each of the polishing systems 20 further includes one or more substrate inspection systems 25, one or more metrology systems 26, and a cleaning system 27, which are integrated with the polishing system 20 to perform pre-polishing and / or post-polishing (in-line) inspection, measurement, and cleaning, respectively, on the substrates being polished therein. Each of the polishing systems 20 includes a system controller 28, which directs and coordinates the operation of the various components and subsystems of the polishing system 20.

[0037] As shown, each of the AI training platforms 30 is communicatively coupled to a corresponding system controller 28 using a communication link 29 (e.g., an Ethernet or USB connection). In other embodiments, one or more AI training platforms 30 may be integrated with the system controller 28 to form a portion thereof. In some embodiments, the AI training platform 30 communicates directly with one or more components or subsystems of the polishing system 20. In some embodiments, each AI training platform 30 may be used with more than one polishing system 20 to perform the methods described herein, and / or each AI training platform 30 may be communicatively coupled to each other to share training data 111 therebetween. Figure 1C ). The training data 111 can be shared between the various AI training platforms 30 at multiple different times. In one example, the training data 111 can be shared sequentially in time, which can include sharing at regular time intervals or during or after one or more sequentially executed processes running within the polishing system 20 and / or one or more asynchronous processes running in multiple polishing systems 20. In other embodiments, one or more of the AI training platforms 30 is not physically located in the Fab 10, and cloud computing technology is used to implement the methods described herein.

[0038] The Fab production control system 40 directs the flow and processing of substrates as they travel through the production line and collects and manages data related to both the substrates and the processing systems. Generally speaking, the system controller 28 communicates with the Fab production control system 40, which provides instructions to the system controller 28 and receives information from it. Here, the Fab production control system 40 further communicates with one or more independent substrate inspection and / or metrology stations 50, other processing systems 60, and one or more electrical test systems 70. In some embodiments, the Fab production control system 40 passes information received from the independent substrate inspection and / or metrology stations 50, other processing systems 60, and one or more electrical test systems 70 to the system controller 28 for use as training data 111 ( Figure 1C In some embodiments, the Fab production control system 40 communicates directly with the AI training platform 30 via a corresponding communication link 29. The communication link 29 may include a conventional wired or wireless type of communication link.

[0039] Figure 1C is a schematic representation of a processing improvement scheme 100 that can be used with the methods described herein. The processing improvement scheme 100 uses an AI training platform 30 including a processor and memory module (PMB 104), which is operable with support circuitry 32 to execute a machine learning AI algorithm (referred to herein as AI algorithm 110). The processor of PMB 104 (not shown separately) is one or a combination of computer processors (e.g., one or more of a programmable central processing unit (CPU), a graphics processing unit (GPU), a field programmable gate array (FPGA), a machine learning application specific integrated circuit (ASIC), or other suitable hardware implementations) suitable for executing AI algorithm 110. The memory of PMB 104 (not shown separately) is operably coupled to the processor and is non-transitory and represents any non-volatile type of memory having a size suitable for storing AI algorithm 110, training data 111 to be used with AI algorithm 110, and one or more machine learning AI models 112 generated using AI algorithm 110. Support circuits 32 are conventionally coupled to the processing unit and include cache, clock circuits, input / output subsystems, power supplies, etc., as well as combinations of the foregoing.

[0040] Here, the AI algorithm 110 is trained using one or a combination of supervised and unsupervised learning models using training data 111 stored in the memory of the PMB 104. In one example of a supervised learning model, the AI algorithm 110 can be trained to map input data (e.g., time series data of various control parameters) to output data (e.g., various processing results) based on example input-output pairs provided by a user. In an example unsupervised learning model, the AI algorithm 110 can be trained to find patterns and relationships in the training data 111, which is received over time with minimal user input. The training process can be based on multiple data sets received from various in-situ or ex-situ sensors over an extended period of time.

[0041] In embodiments where the AI algorithm 110 comprises a supervised model, a support vector machine (SVM), a regression model, or any supervised learning model capable of receiving training data 111 and providing a continuous output indicative of or predicting a process outcome may be used. In embodiments where the AI algorithm 110 comprises an unsupervised model, a neural network or any unsupervised learning model capable of receiving training data 111 to train the AI algorithm 110 to provide a clustered and categorized output indicative of and / or predicting one or more process outcomes may be used. In some embodiments, such as embodiments where the training data comprises images of various components of the polishing system 20 and / or substrates processed therein, the AI algorithm 110 may utilize a convolutional neural network.

[0042] Herein, the training data 111 includes processing system data 114 generated by the polishing system 20 or its subsystems, and corresponding processing result data 116 of one or more substrates processed on the polishing system 20. Here, the processing system data 114 used to train the AI algorithm 110 includes: polishing recipe parameter data 118, such as various polishing parameters and their corresponding target values; Figure 2A-2C 1a-201n described in detail; and process monitoring data 122 generated, for example, by additional sensors or measurement components provided in the polishing system 20, relating to the operation and process performance of the subsystems and / or their consumables. The process system data 114 generated by the polishing system 20 or its subsystems can be represented by discrete values (such as those provided in a polishing recipe) or can include time series data, such as a series of data points (or images) arranged in time sequence.

[0043] In some embodiments, the AI training platform 30 is communicatively coupled to one or more components of the polishing system 20 and receives at least a portion of the processing system data 114 from the one or more components. In some embodiments, at least a portion of the processing system data 114 is stored in a memory of the polishing system controller 28, and the AI training platform 30 receives the processing system data 114 from the memory.

[0044] Process result data 116 includes information related to planarization and / or removal of a material layer of the substrate during the polishing process, the information being obtained by measuring or inspecting the substrate, including information derived therefrom. In some embodiments, process result data 116 includes an image of the surface of the substrate, for example, captured using a camera assembly.

[0045] Here, the processing result data 116 includes, for example, the following Figure 2A 1. In some embodiments, the process result data 116 includes substrate measurements obtained concurrently with the polishing process using an eddy current sensor or an optical sensor (in-situ result data 124), and substrate measurements taken after the polishing process (ex-situ result data 126). In some embodiments, the in-situ result data 124 includes time series data. In some embodiments, the process result data 116 includes the difference between measurements taken before the polishing process and measurements taken after the polishing process, such as material removal rate or material removal uniformity.

[0046] Here, the in-situ result data 124 includes the use Figure 2A The in-situ substrate monitoring system 222 described in the foregoing can be used to obtain time-series eddy current information and / or time-series optical signal information. The in-situ result data 124 generally includes signal information and can include information derived from the signal information, such as material layer thickness and material layer uniformity information.

[0047] The ex-situ result data 126 can be generated using any suitable metrology or inspection system typically found in semiconductor device fabrication facilities. In some embodiments, at least a portion of the ex-situ result data 126 is generated using one or more sequentially connected inspection systems 25 and / or metrology systems 26 of the polishing system 20, and the portion of the ex-situ result data 126 is received at the AI training platform 30 from the one or more sequentially connected inspection systems 25 and / or metrology systems 26. In some embodiments, at least a portion of the ex-situ result data 126 is stored in a memory of a polishing system controller 28, which is communicatively coupled to the sequentially connected systems 25, 26, and the AI training platform 30 receives the portion of the ex-situ result data 126 from the processing system controller 28.

[0048] In some embodiments, at least a portion of the ex situ result data 126 is generated using one or more independent inspection stations and / or metrology stations 50 that are separate from the polishing system 20. Generally speaking, in those embodiments, the ex situ result data 126 is collected and / or received from a Fab production control system 40 that is communicatively coupled to each of the independent inspection stations and / or metrology stations 50.

[0049] Examples of information that may form part of the ex-situ results data 126 include: material removal rate (MRR); material layer planarization (global planarity); substrate-to-substrate uniformity, i.e., wafer-to-wafer non-uniformity (WTWNU); uniformity of material removal rate across the surface of the substrate and / or uniformity of the thickness of the planarized material layer, collectively referred to as the wafer-within-wafer non-uniformity (WIWNU) metric; planarization efficiency; local planarity, such as intra-grain (WID) planarity; undesirable removal of underlying material layers, such as oxide loss; erosion of underlying material layers in high feature density areas; depression of material in trenches, contacts, vias, and / or line features (dish-ing); and polishing-induced defects at or in the substrate surface and / or in exposed features formed in the substrate surface. CMP-induced defects include mechanically related defects (e.g., scratches) and chemically related defects (e.g., corrosion of metal features).

[0050] In some embodiments, the ex-situ result data 126 includes images obtained from a sequentially connected and / or independent metrology and / or inspection system, such as images of a substrate acquired using a camera assembly or other optical sensor. In some embodiments, the ex-situ result data includes images generated by a metrology or inspection system that represent information obtained from the substrate, such as thickness, planarity, defectivity, and / or stress maps of a material layer on the substrate and / or a surface of the substrate.

[0051] In some embodiments, the training data 111 includes one or more of substrate tracking data 128, facility system data 130, and electrical test data 132. Here, the substrate tracking data 128 includes identification information of the substrate, information related to components formed on the substrate, and the substrate's processing history. Examples of component information include component dimensions, component geometry, feature size, and pattern density. The processing history generally includes the identification of the upstream processing system and corresponding processing information, such as date / time information and the processing recipe used with the date / time. The processing history may also include information obtained from upstream metrology systems and / or inspection systems.

[0052] Facility system data 130 includes information related to the facility supply system coupled to the polishing system 20 and / or the environmental conditions surrounding the polishing system 20, such as temperature, particle counts, and airflow. Examples of information related to the facility supply system include information obtained from a deionized (DI) water supply system, a clean dry air (CDA) supply system, a chemical delivery system, and a remote polishing fluid distribution system. Typically, a remote polishing distribution system circulates polishing fluid through a facility pipeline for delivery to multiple polishing systems 20, which are fluidically coupled to the facility pipeline at the point of use. Such polishing fluid distribution systems are typically configured for bulk mixing of polishing fluid and may include one or more analyzers to facilitate the mixing process and / or continuously monitor polishing fluid health. Monitoring polishing fluid health includes using analyzers to determine and monitor the chemical properties of the polishing fluid (e.g., pH, oxidizer and additive levels and their decay behavior) as well as the abrasive properties of the polishing fluid, including large particle count (LPC), mean particle size distribution (PSD), density, weight percent solids, and viscosity. Information related to the facility systems, including polishing fluid operating conditions, may be transmitted to the individual system controllers 28 of the plurality of polishing systems 20 and / or to the Fab production control system 40 and received therefrom by the AI training platform 30 .

[0053] Electrical test data 132 may include parametric test information generated at subsequent parametric test operations, for example, using specialized test structures disposed in tangents between components, and / or component test information generated at one or more subsequent component test operations. In some embodiments, electrical test data 132 includes images representing information obtained during the parametric test operations and / or component test operations, such as component yield maps representing locations on the substrate of operable and failed components.

[0054] Here, the training data 111 includes identification information, such as substrate tracking information, system information, and timestamp information, which can be used to associate the information received from each of the above data sources with a specific substrate, polishing system, polishing station, and substrate carrier combination to form a corresponding training data set.

[0055] In some embodiments, the trained AI algorithm 110 is used to generate an AI model 112 , such as a software algorithm, which is passed to the system controller 28 for use as instructions to direct the operation of the polishing system 20 .

[0056] Figure 1Dis a schematic representation of a control system 150 that can be used to generate control parameter data 120. The control parameter data 120 includes time series data of one or more control parameters 157 that the control system 150 uses to maintain a polishing parameter at or near a target value 156. As used herein, a "target value" includes a desired set point, a value greater than a desired lower threshold, a value less than a desired upper threshold, and values between the desired lower and upper thresholds.

[0057] exist Figure 1D , a process control system 150 provides a feedback control closed loop to maintain a polishing parameter at or near a target value 156. As shown, the process control system 150 includes a sensor 151, a controller 152, and a parameter control component 153 (e.g., an actuator) operably coupled to the controller 152. Here, the sensor 151, the controller 152, and the control component 153 are arranged so that information flows in a feedback loop 154 to provide a closed-loop feedback control system.

[0058] During the polishing process, sensor 151 measures an actual value 155 of a polishing parameter (e.g., platen rotational speed, polishing fluid flow rate, etc.), and controller 152 determines an error between actual value 155 and a target value 156. To correct the error, controller 152 instructs parameter control component 153 (e.g., an actuator (motor) coupled to the platen, a slurry dispensing pump connected to the slurry delivery system, etc.) to change a control parameter 157 (e.g., motor current, pump pressure, pump speed, etc.), which results in a corresponding change in the polishing parameter output (e.g., platen rotational speed, slurry flow rate, etc.).

[0059] The parameter control system 150 is generally reactive, such that once a polishing parameter rises to a target value 156, changes in the control parameter 157 by the controller 152 indicate a response to the change in the polishing process. Similarly, for substantially similar polishing processes, changes in the control parameter 157 from substrate to substrate may indicate undesirable process drift. Therefore, in embodiments herein, the time-series control parameter data 120 is included in the process system data 114 to enable the AI algorithm 110 to better understand the complex relationships between subsystems, process parameters, consumables, and substrates for a particular polishing process.

[0060] Figure 2Ais a schematic side cross-sectional view of a polishing station 21 and a carrier assembly 22 according to one embodiment and which can be used with the methods described herein. Here, the polishing station 21 includes a plurality of subsystems, each of which can operate with one or a combination of parameter control systems 201a-201n. Here, each of the parameter control systems 201a-201n is configured to include a feedback control closed loop and can include Figure 1D Any one or combination of the elements of the process control system 150 described in .

[0061] In general, each of the control systems 201a-201n includes one or more corresponding actuators 202a-202n, process parameter sensors 203a-203n, controllers 204a-204n, and control parameter sensors 205a-205n. The actuators 202a-202n include any component or processing system operable to change a control parameter in response to a signal (e.g., an electrical, pneumatic, or digital signal) received from the controllers 204a-204n. Examples of common actuators 202a-202n include, but are not limited to, motor assemblies, electromagnetic assemblies, pneumatic assemblies, hydraulic assemblies, and combinations thereof, such as electric motors, servos, solenoids, valves, pumps, pistons, and regulators.

[0062] The process parameter sensors 203a-203n include any component or combination of components that can be used to measure the value of a process parameter or can be used to provide one or more measurement values, wherein the actual value of the desired process parameter can be determined based on the one or more measurement values. Examples of suitable process parameter sensors 203a-203n include temperature sensors (such as IR sensors, pyrometers, and thermocouples), pressure sensors, force sensors, position sensors, acceleration sensors, speed sensors, rotary encoders, electrical signal detection sensors, electrochemical sensors, pH sensors, concentration sensors, optical sensors, inductive sensors, flow sensors (mass and / or volume), and combinations thereof.

[0063] The controllers 204a-204n include components or systems that are operable to determine the difference (i.e., error) between the actual value of a process parameter and the target value of the process parameter and instruct the corresponding actuators 202a-202n or processing systems to change their outputs (e.g., control parameters described herein). Examples of suitable controllers 204a-204n include proportional-integral (PI) controllers, proportional-integral-derivative (PID) controllers, and / or logic controllers, such as programmable logic controllers (PLCs) that have been programmed to execute software including logic applications. In some embodiments, for example, when the control parameters include the output of the processing system, the system controller 28 or another computing component that is operable to execute a software algorithm can be used as the controller 204a-204n. In some embodiments, one or more of the functions of each or a combination of the controllers 204a-204n can be performed by the system controller 28.

[0064] The control parameter sensors 205a-205n include any sensor suitable for measuring an output of an actuator 202a-202n or a processing system that is used to maintain a processing parameter at a target value. Examples of suitable sensors that can be used as the control parameter sensors 205a-205n include any one or a combination of the example sensors described above with respect to the processing parameter sensors 203a-203n. In some embodiments, such as for control systems where measuring a control parameter is not feasible, the control parameter or its approximate value can be determined using signals and / or instructions provided by the controller 204a-204n to the corresponding actuator 202a-202n or processing system.

[0065] In other embodiments, any one or combination of the various subsystems described below may be operated using an open-loop control system (ie, a non-feedback system).

[0066] Here, the plurality of subsystems include a platen assembly 212, a carrier assembly 22, a pad conditioner assembly 218, and a pad cooling assembly 220. The polishing station 21 further includes a fluid delivery system 216 and an in-situ substrate monitoring system 222. Operation of the polishing station 21 and carrier assembly 22 is coordinated by a system controller 28.

[0067] The pressure plate assembly 212 includes a pressure plate 228 and a speed control system 201a. The control system 201a includes a pressure plate actuator 202a (e.g., an electric motor), a process parameter sensor 203a, a controller 204a, and a control parameter sensor 205a. The pressure plate actuator 202a is coupled to the pressure plate 228 and is used to rotate the pressure plate 228 about the pressure plate axis A. The process parameter sensor 203a is used to measure the speed and / or rotation direction of the pressure plate 228.

[0068] Here, the controller 204a in combination with the sensor 203a maintains the rotational speed of the platen 228 at or near a target value by adjusting a control parameter (e.g., motor current) provided to the platen actuator 202a. The control parameter sensor 205a is used to measure the control parameter, and time series control parameter data is generated based on the control parameter. In some embodiments, the change in the control parameter of the motor current is caused by the substrate 242 ( Figure 2B ) is caused by a change in the friction between the surfaces at the polishing interface when the coating of material is removed from the polishing pad. Thus, in some embodiments, a change in motor current can be used to detect the desired polishing endpoint of the polishing process. In other embodiments, the motor current can be used to detect a change in the amount of slurry delivered to the polishing pad and the surface of the substrate 242 at any moment during polishing. For example, a higher friction force sensed by the motor current may be caused by a decrease in slurry flow or a change in the composition of the slurry composition.

[0069] The platen assembly 212 further includes a platen temperature control system 201b, which includes a fluid source 202b (e.g., a water or coolant source), a sensor 203b, and a controller 204b. The platen temperature control system 201b includes a fluid source 202b (e.g., a water or coolant source). The sensor 203b is used to measure the temperature of the platen 228. The platen temperature can be used to detect changes in the amount of slurry delivered to the polishing pad, changes in polishing pad properties (e.g., the amount of polishing), or changes in the downforce applied to the substrate 242 at any moment during polishing. The platen 228 is formed from a cylindrical metal body having one or more channels 234 formed therein. The one or more channels 234 are fluidically coupled to the fluid source 202b. The controller 204b, in combination with the sensor 203b, is used to maintain the temperature of the platen 228 at a target value by adjusting the flow rate of coolant from the fluid source 202b through the one or more channels 234. In some embodiments, the control parameter(s) used to control the temperature of the polishing platen 228 include a coolant flow rate measured by a flow meter (e.g., control parameter sensor 205b). For some polishing processes, it may be desirable to heat the platen 228. In those embodiments, the fluid source 202b may include a heated fluid (e.g., heated water and / or steam), and the target value may include a temperature greater than a lower threshold. In some embodiments, the platen 228 is heated using a heater (not shown) (e.g., a resistive heating element disposed and / or embedded in the cylindrical metal body).

[0070] The carrier assembly 22 includes a substrate carrier 238, a carrier shaft 239, and control systems 201c, 201d. Figure 2B1 and 2. The substrate carrier 238 is depicted in FIG. The control system 201c includes a first actuator 202c, a controller 204c, a rotational speed sensor 203c, and a control parameter sensor 205c. The first actuator 202c is coupled to a carrier shaft 239 and is configured to rotate the carrier shaft 239, thereby rotating the substrate carrier 238 and the substrate 242 disposed therein about a carrier axis B. The controller 204c, in combination with the sensor 205c, is configured to maintain the rotational speed of the substrate carrier 238 at or near a target value by adjusting a control parameter (e.g., motor current) provided to the first actuator 202c. The control parameter sensor 205c is configured to measure the control parameter provided to the first actuator 202c.

[0071] The control system 201d includes a second actuator 202d coupled to the carrier shaft 239 and / or the first actuator 202c, a controller 204d, a sweep speed sensor 203d, and a control parameter sensor 205d. The controller 204d, in combination with the sensor 203d, is used to maintain the sweep speed of the substrate carrier 238 at or near a target value by adjusting a control parameter (e.g., motor current) provided to the second actuator 202d. The control parameter sensor 205d is used to measure the control parameter provided to the second actuator 202d.

[0072] like Figure 2B As shown in FIG, the substrate carrier 238 includes a housing 240, a base assembly 243, a substrate downforce control system 201f, and a carrier load control system 201g. The housing 240 is movably and sealingly coupled to the base assembly 243 to define a loading chamber 244 therewith. The base assembly 243 includes a carrier base 246, an annular retaining ring 247 coupled to the carrier base 246, and a flexible diaphragm 248 coupled to the carrier base 246 to define a plurality of plenums 249 therewith.

[0073] During substrate polishing, the plurality of air chambers 249 are pressurized, causing the flexible diaphragm 248 to apply a force to the inactive (backside) surface of the substrate 242 beneath it. The plurality of air chambers 249 facilitate adjustment of the distribution of the force applied across the backside surface of the substrate 242 by allowing pressure differentials therein. The pressures in the various air chambers 249 and the pressure differentials therebetween are maintained by a control system 201f, which includes a plurality of actuators 202f (e.g., backside pressure regulators, valves, etc.), a plurality of sensors 203f, one or more controllers 204f, and one or more control parameter sensors 205f. The control system 201f is used to maintain a target pressure in each of the air chambers 249, thereby allowing precise control of the distribution of the force applied by the flexible diaphragm 248 to the substrate 242.

[0074] One or more controllers 204f in combination with a plurality of sensors 203f maintain the pressures in the air chamber 249 at their target values by adjusting the respective control parameters of their corresponding actuators 202f. The different control parameter values are measured by their corresponding control parameter sensors 205f.

[0075] During processing, the loading chamber 244 is also pressurized to apply a downward force to the carrier base 246, and therefore to the retaining ring 247 surrounding the substrate 242. The downward force on the retaining ring 247 prevents the polishing pad 231 ( Figure 2A ) moves under the substrate 242 as the substrate 242 slides from the substrate carrier 238. The contact pressure between the retaining ring 247 and the polishing pad 231 is adjusted by varying the target downforce on the retaining ring 247. The target downforce is maintained by a control system 201g, which includes an actuator 202g (e.g., a backside pressure regulator), a sensor 203g for measuring the pressure in the loading chamber 244 and / or the contact load between the retaining ring 247 and the polishing pad 231, a controller 204g for maintaining the target pressure in the loading chamber 244, and a control parameter sensor 205g. The controller 204g, in combination with the sensor 203g, maintains the pressure in the loading chamber 244 at or near its target value by adjusting the control parameters provided to the actuator 202g. Here, the various components of the control systems 201g and 201h collectively form an upper pneumatic assembly (here, UPA 241), which may further include regulators, valves, and pumps (not shown) for providing pressurized gas (e.g., clean dry air (CDA) and / or vacuum) to the plurality of air chambers 249 and the load chamber 245. In other embodiments, a motor assembly may be used to apply downward pressure to one or both of the base plate 242 and the retaining ring 247.

[0076] Pad adjuster assembly 218 ( Figure 2A ) is used to condition the polishing pad 231 by pressing a conditioning disk 260 against the surface of the polishing pad 231 before, after, or during polishing of the substrate 242. Here, the pad conditioner assembly 218 includes the conditioning disk 260, a conditioner arm 262 for sweeping the rotating conditioning disk 260 between the inner and outer radii of the polishing pad 231, and a plurality of control systems 201 j - 201 m for controlling various aspects of the pad conditioning process.

[0077] Generally speaking, the conditioning disk 260 includes a fixed abrasive conditioning surface (e.g., diamonds embedded in a metal alloy) and is used to grind and recondition the surface of the polishing pad 231 and remove polishing byproducts and other debris therefrom. The conditioning disk 260 is generally considered a processing consumable that requires periodic replacement because the abrasiveness of the conditioning disk 260 naturally dulls with use.

[0078] Control systems 201j, 201k are used to maintain the rotational speed and sweep speed of conditioning disk 260 at respective target values as conditioning disk 260 oscillates between the inner and outer radii of polishing pad 231. Control system 2011 is used to maintain the downward force applied to conditioning disk 260 at a target value. In some embodiments, pad conditioner assembly 218 further includes a control system 201m that can be used to provide and / or maintain a desired polishing pad thickness profile across the surface of polishing pad 231. In those embodiments, the desired polishing pad thickness profile is maintained by adjusting one or a combination of the rotational speed, sweep speed, and downward force according to instructions provided by a software algorithm executed by system controller 28.

[0079] Here, the control system 201j includes a first actuator 202j coupled to the end of the adjuster arm 262 where it is used to rotate the adjuster disk 260 about the axis C, a sensor 203j for determining the rotational speed, and a controller 204j.

[0080] Control system 201k includes a second actuator 202k coupled to the end of adjuster arm 262 distal from first actuator 202j, one or more sensors 203k for determining the sweep velocity and / or radial position of adjustment disk 260 on the polishing pad, a controller 204k, and a control parameter sensor 205k. Control system 201g includes a third actuator 202l for applying a downforce on adjuster arm 262, a sensor 203l for measuring the downforce, a controller 204l, and a control parameter sensor 205l. Here, third actuator 202l is coupled to the end of adjuster arm 262 adjacent to second actuator 202l and distal from adjustment disk 260. Each of the controllers 204j-204l, in combination with the respective sensors 203j-203l, maintains the respective process parameters at or near their target values by adjusting the control parameters of the respective actuators 202j-202l.

[0081] In some embodiments, a control system 201m is used to maintain a desired polishing pad thickness profile by adjusting one or a combination of the rotational speed, sweep speed, and downforce of the conditioning disk 260. Here, the control system 201m includes actuators 202j-202l, a displacement sensor 203m coupled to an adjuster arm 262, and a system controller 28. The displacement sensor 203m is used to determine the thickness of the polishing pad 231 and the profile of the pad thickness in a radial direction across the polishing pad 231. Here, the displacement sensor 203m is an inductive sensor that measures eddy currents to determine the distance between the end of the sensor 203m and the surface of the metal platen 228 disposed thereunder. The thickness of the polishing pad 231 is determined using the difference between the known displacement of the pad conditioning disk 260 when it contacts the platen 228 and the displacement of the pad conditioning disk 260 when it contacts the polishing pad 231 mounted on the platen 228.

[0082] The system controller 28 compares the thickness profile of the polishing pad 231 determined using the displacement sensor 203m with the target thickness profile to determine the difference therebetween. Based on the difference, the system controller 28 generates an adjustment recipe (i.e., a set of adjustment parameters) that can be used to drive the actual thickness profile of the polishing pad 231 toward the target thickness profile. In some embodiments, the generated adjustment recipe changes the dwell time of the adjustment disk 260 and / or the downward force on the adjustment disk at one or more radial locations. The dwell time refers to the average duration that the adjustment disk 260 spends at a radial location as the adjustment disk 260 sweeps from the inner radius to the outer radius of the polishing pad 231 as the pressure plate 228 rotates to move the polishing pad 231 under the adjustment disk 260.

[0083] Pad cooling assembly 220 ( Figure 2C) is used to maintain the polishing surface of the polishing pad 231 within a desired temperature range or at a desired temperature set point. In a typical polishing process, chemical and mechanical activity at the polishing interface generates heat, which in turn increases the temperature of the substrate 242 and the polishing pad 231. Relatively high and / or unstable temperatures can lead to undesirable removal rate variations across the surface of the substrate 242 (intra-wafer non-uniformity) or from substrate to substrate (wafer-to-wafer non-uniformity). For many damascene processes, relatively high temperatures can degrade local planarization, resulting in poor local planarity, erosion of underlying layers, and / or dishing of trenches, contacts, vias, or line features formed in the underlying layers. Therefore, herein, the pad cooling assembly 220 is configured to cool the surface of the polishing pad 231 by delivering a non-reactive coolant, such as a thin sheet of solid carbon dioxide (CO2 snow), to the surface. As the CO2 snow sublimes (transitions from a solid phase to a gas phase without passing through an intermediate liquid phase), heat is removed from the surface of the polishing pad 231, thereby desirably reducing the overall temperature of the polishing process. Beneficially, the sublimation of the carbon dioxide snow prevents undesirable dilution of the polishing fluid on the polishing pad. In other embodiments, the coolant comprises a cryogenic fluid, i.e., a fluid having a boiling point equal to or less than a threshold of 120 Kelvin, which is stored and delivered to the surface of the polishing pad 231 in liquid form, such as liquid oxygen (LOX), liquid hydrogen, liquid nitrogen (LIN), liquid helium, liquid argon (LAR), liquid neon, liquid krypton, liquid xenon, liquid methane, or a combination thereof.

[0084] The pad cooling assembly 220 includes a coolant delivery arm 275 positioned above the polishing pad 231, a plurality of nozzles 276 disposed on the coolant delivery arm 275, and a control system 201n. Here, the control system 201n includes a coolant source 202n, one or more sensors 203n, a controller 204n, and a control parameter sensor 205n. One or more sensors 203n (e.g., IR sensors or pyrometers) are positioned facing the surface of the polishing pad 231 and are used to measure its temperature. In some embodiments, one or more of the sensors 203n include a thermal imaging system that generates a thermal image of the surface of the polishing pad 231.

[0085] A plurality of nozzles 276 are fluidically coupled to a coolant source 202n, which provides vapor and solid carbon dioxide to the plurality of nozzles 276. The plurality of nozzles 276 generate carbon dioxide snow as the vapor carbon dioxide expands therethrough and delivers the carbon dioxide snow to the surface of the polishing pad 231. A controller 204n, in combination with a sensor 203n, maintains the temperature of the polishing pad 231 at a target value by adjusting the mass flow rate of carbon dioxide provided from the coolant source 202n to the nozzles 276. Here, the control parameter(s) for controlling the temperature of the surface of the polishing pad 231 include the mass flow rate measured by the control parameter sensor 205n. In some embodiments, the delivery and / or flow rate of coolant to each of the plurality of nozzles 276 is independently controlled. In those embodiments, the pad cooling assembly 220 can be used to adjust the temperature of a region of the surface of the polishing pad 231 to maintain a desired uniformity of temperature or temperature distribution across the surface.

[0086] Each of the control systems 201a-201n of the polishing system 20 described above utilizes a closed-loop feedback control method to maintain one or more polishing parameters at or near a respective target value by adjusting respective control parameters associated with the one or more polishing parameters. As discussed above, differences in control parameters between substrates (e.g., wafer-to-wafer (WTW)), during polishing of individual substrates (e.g., within-wafer (WIW)), or both, may indicate a disturbance or change in the polishing process. Such disturbances or changes in the polishing process are unlikely to be caused by changes in the polishing parameters maintained at or near a target value using the control systems 201a-201n. Instead, such disturbances or process changes are likely to occur at the polishing interface and include changes in the surface of the substrate 242, changes in the surface of the polishing pad 231, changes in the composition, properties, and / or volume of the polishing fluid, and combinations thereof. Therefore, in some embodiments, an AI algorithm 110 utilizing an unsupervised learning model may be used to identify and understand patterns in the control parameter data 120 to better understand the complex chemical and mechanical interactions between the surface, fluid, and abrasive at the polishing interface.

[0087] As discussed in the method below, in some embodiments, the AI algorithm 110 is trained to determine a functional relationship between one or more control parameters and in-situ substrate measurement data, and to adjust the polishing fluid composition at the polishing interface based on the functional relationship. Thus, herein, the fluid delivery system 216 is configured to stop the flow of each polishing fluid component to the surface of the polishing pad 231, start the flow of each polishing fluid component to the surface of the polishing pad 231, and / or adjust the flow rate of each polishing fluid component to the surface of the polishing pad 231 based on instructions received from the system controller 28, thereby stopping the flow of each polishing fluid component to the polishing interface, starting the flow of each polishing fluid component to the polishing interface, and / or adjusting the flow rate of each polishing fluid component to the polishing interface. In some embodiments, the instructions are in the form of software algorithms, such as one or more machine learning AI models 112 generated using the trained AI algorithm 110.

[0088] Fluid delivery system 216 ( Figure 2C ) is used to deliver polishing fluid (including various fluid components) to the surface of the polishing pad. The fluid delivery system 216 includes a fluid distribution system 281, a delivery arm 282 including a plurality of nozzles 283, and an actuator 284 coupled to the fluid delivery arm 282. The fluid distribution system 281 is fluidically coupled to a plurality of polishing fluid sources 287a, 287b, which deliver polishing fluid and / or fluid components to the fluid distribution system 281. The actuator 284 is operable to swing the delivery arm 282 over the polishing pad to position the plurality of nozzles 283 at desired radial distribution locations on the polishing pad.

[0089] Here, the fluid distribution system 281 includes one or a combination of a plurality of valves 285a, a pump 285b, a flow controller 285c, and a polishing fluid mixing device 285d. The flow controller 285c can be used to control, measure, and deliver the polishing fluid and / or individual polishing fluid components to the surface of the polishing pad 231. In some embodiments, the fluid distribution system 281 further includes one or more heaters (not shown) that are used to heat the individual polishing fluids and / or one or more individual polishing fluid components before and / or while delivering the fluids and / or components to the surface of the polishing pad 231.

[0090] Here, one or more polishing fluids and individual polishing components are delivered from the fluid distribution system 281 to corresponding ones of the plurality of nozzles 283 using a plurality of delivery lines 288 fluidly coupled between the fluid distribution system 281 and the plurality of nozzles 283. In some embodiments, the fluid distribution system 281 is configured to independently deliver one or more different polishing fluids and / or fluid components to different ones of the plurality of nozzles 283 and / or independently control the flow rates of the different polishing fluids or fluid components to the different nozzles. Thus, the fluid distribution system 281 can be used to provide a desired distribution of the polishing fluid and / or individual polishing fluid components distributed onto the surface of the polishing pad 231 to provide a desired polishing fluid composition gradient across the surface of the polishing pad 231.

[0091] In some embodiments, the fluid distribution system 281 further includes a mixing device 285d that can be used to adjust the composition of the polishing fluid by adding one or more polishing fluid components to the polishing fluid before delivering the resulting mixture to the surface of the polishing pad 231. In some embodiments (not shown), the mixing station is disposed on the fluid delivery arm 282.

[0092] Examples of individual polishing fluid components that can be independently delivered to the surface of the polishing pad 231, delivered to desired locations on the surface of the polishing pad, and / or added to the polishing fluid using the mixing device 285d include: an abrasive solution in which nano-sized silicon oxide or metal oxide particles are suspended; a complexing agent; a corrosion inhibitor; an oxidizing agent; a pH adjuster and / or buffer, a polymeric additive, a passivating agent, an accelerator, a surfactant, or a combination thereof.

[0093] In some embodiments, the fluid delivery system 216 further includes an optical sensor (e.g., camera 299) positioned above and facing the polishing pad 231. In some embodiments, the camera 299 is a digital camera (e.g., a CCD camera) configured to generate a digital image or stream of digital images of the object it is positioned to view. The optical sensor can be used to determine the distribution of the polishing fluid and / or polishing fluid components across the surface of the polishing pad 231. In some embodiments, one or more of the polishing fluids and / or polishing fluid components include an optical marker, such as a conventional water-soluble dye or fluorophore. In those embodiments, the images captured using the optical sensor can be analyzed to determine the distribution of the polishing fluid across the surface of the polishing pad 231 and / or to determine the composition gradient of the polishing fluid components across the surface of the polishing pad 231.

[0094] In some embodiments, the polishing fluid distribution and / or composition at the surface of the polishing pad 231 is adjusted based on the analysis of the image by starting, stopping, or changing the flow rate of one or more individual polishing fluid components to one or more of the individual nozzles 283. In some embodiments, a closed-loop feedback control system 280 is used to continuously adjust the polishing fluid distribution and / or composition at the surface of the polishing pad 231 to a target distribution and / or composition. For example, the control system 280 herein includes the system controller 28, an optical sensor (e.g., camera 299) for determining the polishing fluid distribution and / or composition at the surface of the polishing pad 231, and the fluid distribution system 281. In another example, the control system 280 herein includes the system controller 28, an electrochemical sensor (not shown), or a pH sensor (not shown) for determining the polishing fluid composition at the surface of the polishing pad 231 and / or within the fluid distribution system 281. Based on the analysis of the image acquired from the optical sensor, the system controller 28 directs the fluid distribution system 281 to change one or more control parameters associated with delivering the polishing fluid and / or polishing fluid components to the surface of the polishing pad 231. For example, the control parameters may include starting, stopping, or changing the flow rate of individual polishing fluids and / or polishing fluid components provided to the plurality of nozzles 283 as a whole or to individual nozzles within the plurality of nozzles.

[0095] In some embodiments, one or more of the images captured using the optical sensor (e.g., a time series of multiple captured images) includes processed monitoring in-situ measurement data 122, which can be used as training data 111 for the AI algorithm 110 training method provided herein.

[0096] In-situ substrate monitoring system 222 ( Figure 2A ) is used to monitor the thickness of a layer of material on the substrate surface and / or detect changes in the substrate surface when material is removed from the substrate surface. Information collected using the in-situ substrate monitoring system 222 can be used as in-situ result data 124. Here, the in-situ substrate monitoring system 222 includes a controller 290 for one or both of the optical system 291 and the eddy current monitoring system 292. The optical system 291 includes a light source (not shown) and an optical sensor 289, which are respectively positioned to direct light toward the substrate 242 through a window (not shown) formed in the polishing pad 231 and receive light reflected from the substrate 242. The controller 290 analyzes the reflected light to determine one or more properties of the substrate surface based on the reflected light. For example, the optical system 291 can be used to detect changes in the reflectivity of the substrate surface, such as to determine the removal of a metal layer from the substrate surface, detect scattering of light reflected from the substrate surface, such as to determine changes in the flatness of the substrate surface, and / or determine the thickness of a transparent film (e.g., a dielectric layer) disposed on the substrate surface using interferometry techniques.

[0097] The eddy current monitoring system 292 includes an eddy current assembly 294, which includes an eddy current generator and a sensor disposed in the surface of the platen 228. The eddy current monitoring system 292 uses the eddy current assembly 294 to sense and measure eddy currents in a layer of conductive material (e.g., a metal layer) on the substrate, and the current monitoring system determines the thickness of the conductive material layer based on the eddy current. In some embodiments, the eddy current monitoring system 292 is used to determine a thickness profile across a radius of the substrate 242 as the substrate is swept over the eddy current monitoring system 292.

[0098] In some embodiments, one or both of the optical system 291 and the eddy current monitoring system 292 are used in combination with an endpoint algorithm executed on a controller of the polishing system (e.g., on the system controller 28) to trigger a change in polishing conditions based on the thickness of the material layer and / or based on the removal of overburden material from underlying field surfaces.

[0099] The system controller 28 directs the operation of the polishing system 20 and its various components and subsystems. In some embodiments, one or more, or all, of the functions of each of the controllers 204a-204n can be performed by the system controller 28. Herein, the system controller 28 can operate in conjunction with the AI training platform 30 to implement the methods described herein. The system controller 28 includes a programmable central processing unit (CPU 295) that operates in conjunction with memory 296 (e.g., non-volatile memory) and support circuits 297. For example, in some embodiments, the CPU 295 is one of any form of general-purpose computer processor, such as a programmable logic controller (PLC), used in industrial environments to control various polishing system components and sub-processors. The memory 296 coupled to the CPU 295 is non-transitory and is typically one or more readily accessible memories, such as random access memory (RAM), read-only memory (ROM), a floppy disk drive, a hard disk, or any other form of digital storage (local or remote). Support circuits 297 are conventionally coupled to the CPU 295 and include cache memory, clock circuits, input / output subsystems, power supplies, etc., and combinations thereof, coupled to various components of the polishing system 20 to facilitate control of the substrate polishing process.

[0100] Herein, memory 296 is in the form of a computer-readable storage medium (e.g., non-volatile memory) containing instructions that, when executed by CPU 295, facilitate the operation of polishing system 200. Illustrative computer-readable storage media include, but are not limited to: (i) non-writable storage media (e.g., read-only storage elements within a computer, such as CD-ROM disks readable by a CD-ROM drive, flash memory, ROM chips, or any type of solid-state non-volatile semiconductor memory), on which information can be permanently stored; and (ii) writable storage media (e.g., floppy disks within a disk drive, or hard disk drives, or any type of solid-state random-access semiconductor memory), on which information can be altered. The instructions in memory 296 are in the form of a program product (e.g., a middleware application, a device software application, etc.), such as a program that implements the methods of the present disclosure. In some embodiments, the present disclosure can be implemented as a program product stored on a non-transitory computer-readable storage medium for use with a computer system. Thus, the program(s) of the program product define the functionality of the embodiments (including the methods described herein).

[0101] Figure 3 Is shown using Figure 1C 100 for processing a substrate using the process improvement 100 described in

[0026] . It is contemplated that at least a portion of the method 300 may be performed on the polishing system 20 and may incorporate any of its features and functions, including various control systems used therewith. Applications of the method 300 include, but are not limited to, bulk material planarization applications (e.g., interlayer dielectric (ILD) applications) and damascene polishing applications (e.g., shallow trench isolation (STI) and metal interconnect polishing applications).

[0102] At activity 302, method 300 includes polishing a substrate using a polishing system, such as the polishing system 20 described above. Activity 302 will include a number of activities including activities 304-312.

[0103] At activity 304, method 300 includes flowing a polishing fluid composition (e.g., slurry) according to a polishing recipe onto a surface of a polishing pad in polishing system 20. The flow rate and / or amount of the polishing fluid composition provided to a defined radial location on the surface of polishing pad 231 can be controlled using commands sent from system controller 28 to actuator 284 and / or fluid distribution system 281.

[0104] At activity 306, method 300 includes pressing a substrate against a surface of a polishing pad in the presence of a polishing fluid according to a polishing recipe. Here, the polishing recipe is defined by a plurality of polishing parameters (including substrate carrier rotation speed, substrate carrier translation speed, platen rotation speed, substrate downforce, retaining ring downforce, polishing composition flow rate(s), rinse solution flow rate(s), and pad conditioning parameters) and their corresponding target values. The target values include a desired set point, a value greater than a desired lower threshold, a value less than a desired upper threshold, and a value between the desired lower and upper thresholds. Activity 306 includes pressurizing one or more of the plurality of plenums 249 so that the flexible diaphragm 248 in the substrate carrier applies a force to the inactive (backside) surface of the substrate 242 to force the frontside surface against the polishing pad 231.

[0105] The target values may include a combination of fixed values (e.g., predetermined set points or thresholds) and values determined by one or more software algorithms that are executed on the controller of the polishing system before, after, and / or during the polishing process. For example, in some embodiments, the duration of a phase of a polishing sequence is determined using an endpoint algorithm executed on the controller of the polishing system. In some embodiments, one or more of the target values is determined by a trained AI algorithm 110, for example, as part of an iterative continuous improvement process. In some embodiments, one or more of the target values is determined using a machine learning AI model 112 generated by the trained AI algorithm 110. In those embodiments, the machine learning AI model 112 may include a software algorithm executed by the system controller 28 of the polishing system 20.

[0106] In a typical polishing process, a polishing recipe for a single substrate includes a multi-stage polishing sequence in which one or more polishing parameter target values are changed at each stage of the sequence. In some embodiments, one or more stages of the multi-stage polishing sequence are performed at a first polishing station before the substrate is moved to a second polishing station and, in some cases, to a third polishing station for the remainder of the polishing sequence.

[0107] Examples of polishing parameters that can be used to define a polishing recipe include, but are not limited to: platen rotation speed; platen temperature; substrate carrier rotation speed; substrate carrier sweep speed; substrate carrier sweep start and stop positions (inner radial position and outer radial position on the polishing pad); substrate downforce (downward pressure applied to the back side of the substrate); distribution of downforce across the substrate; retaining ring downforce (downward pressure applied to the retaining ring); the difference between substrate downforce and retaining ring downforce; polishing pad surface temperature; polishing pad surface temperature uniformity and / or distribution; flow rate of the polishing fluid and / or individual polishing fluids, including starting and stopping the flow of the polishing fluid or components; temperature of the polishing fluid and / or individual polishing fluid components; polishing fluid composition prior to delivery to the polishing pad (e.g., as output from a polishing fluid mixing system) or on the surface of the polishing pad (e.g., as a result of dispensing individual polishing fluid components); and polishing fluid distribution and / or composition gradient across the surface of the polishing pad, and duration (time).

[0108] Generally speaking, the polishing recipe further includes process parameters related to the conditioning of the polishing pad before, after, and / or during the polishing process, referred to herein as pad conditioning parameters. Examples of pad conditioning parameters include: the rotational speed of the conditioning disk, the downforce applied to the conditioning disk against the polishing pad, the dwell time of the conditioning disk over one or more portions of the polishing pad, and the sweep speed of the conditioning disk across the surface of the polishing pad. As briefly discussed above, one or more of the pad conditioning parameters can be used in conjunction with a position sensor of the conditioner assembly to determine the conditioning disk dwell time. In some embodiments, the pad conditioning parameters can also include the polishing pad thickness and / or a profile of the polishing pad thickness measured from a location adjacent to the center of the polishing pad to a location radially outward thereof.

[0109] At activity 308, method 300 includes maintaining one or more polishing parameters at or near their target values by adjusting respective control parameters corresponding to the one or more polishing parameters. Here, a closed-loop control system is used to maintain the one or more polishing parameters at or near their target values. Thus, in some embodiments, maintaining the polishing parameters at or near their target values includes: (1) determining a difference between an actual value of the polishing parameter and its target value; (2) changing a control parameter of the control system corresponding to the polishing parameter based on the determined difference; and (3) continuously repeating (1) and (2) to provide closed-loop control of the polishing parameter.

[0110] As used herein, a control parameter includes an output from an actuator and / or system that results in a corresponding change in the actual value of a polishing parameter. The control parameters of a particular control system are distinct from the polishing parameters of that system. However, as will be appreciated from the description of at least some of the control systems herein, at least some of the parameters described above as exemplary polishing parameters can also be used as control parameters in a different control system. For example, in an embodiment where the polishing pad thickness profile is used as a polishing parameter in a closed-loop system, one or more of the various parameters of regulator downforce, rotational speed, and dwell time can be used as control parameters and adjusted to provide a desired pad thickness profile.

[0111] In some embodiments, at least one of the processing parameters of activity 308 includes pad surface temperature, and the corresponding control parameter includes the mass flow rate of a coolant (e.g., carbon dioxide snow) delivered to the surface of the polishing pad. In some embodiments, a controller 204b, in combination with a sensor 203b, is used to control the temperature of the platen 228 at a target value by adjusting the flow rate of coolant from a fluid source 202b through one or more channels 234 in the polishing platen 228. In some embodiments, the control parameter(s) used to control the temperature of the polishing platen 228 include the coolant flow rate measured by a flow meter (e.g., control parameter sensor 205b).

[0112] At activity 310, the method 300 includes generating the processing system data 114. Here, the processing system data 114 includes time series data of a polishing recipe and a first control parameter.

[0113] At activity 312 , the method 300 includes, concurrently with activities 304 through 310 , generating time-series in-situ result data using measurements obtained from an in-situ substrate monitoring system, such as the in-situ substrate monitoring system 222 described herein.

[0114] In some embodiments, at activity 312, a camera 299 ( Figure 2A) is configured to provide a signal (e.g., a video signal stream) that is monitored and analyzed by one or more software algorithms running within the camera or system controller 28 to detect changes or variations in the optical properties of the polishing pad surface and / or the polishing fluid composition disposed thereon. In one example, the camera is an IR camera that is configured to detect a temperature gradient across the polishing pad surface and / or temperature changes over time. The software algorithm can be used to detect the temperature and / or temperature changes on the polishing pad surface and / or the polishing fluid composition disposed thereon in real time. The camera 299 and / or system on which the algorithm is running is then adapted to provide a signal (including time-series in-situ result data) to the system controller 28 and / or to provide a signal including training data to the artificial intelligence (AI) training platform 30. In addition, flow rate sensing components and / or polishing fluid composition detection components (e.g., pH sensors, abrasive particle concentration sensors) coupled to components within the fluid distribution system 281 can also be configured to deliver signals regarding the amount and / or composition of one or more polishing fluid compositions dispensed on the polishing pad surface while the camera monitors the polishing pad surface. During subsequent activities, the artificial intelligence (AI) training platform 30A analyzes the time series in situ result data provided in the signals provided by the camera 299 and the flow rate sensing component and / or (multiple) polishing fluid composition detection components to detect interactions between these different types of data, and then in subsequent activities, uses components in the pad cooling component 220 to cause changes in the temperature of the polishing pad and / or changes in the composition of the polishing fluid composition based on the data received over time.

[0115] In another example, at activity 312, the camera 299 ( Figure 2A ) is configured to detect the condition of the polishing pad surface, such as whether the polishing pad surface has a desired amount of "pad conditioning". In this case, camera 299 is positioned and configured to detect the roughness and / or amount of roughness found on the polishing surface of the polishing pad to determine the condition of the polishing pad surface. In some embodiments, camera 299 is replaced with a profilometer or other component configured to detect and measure the degree of surface roughness. Surface roughness can be determined by R a 、R rms 、R Sk or R pThe surface roughness detected by the camera or similar component can include irregularities in the pad material on the polishing surface of the polishing pad up to about 10-50 microns in size. In addition, the flow rate sensing component and / or the polishing fluid composition detection component (e.g., a pH sensor, an abrasive particle concentration sensor) can also be configured to deliver a signal regarding the amount and / or composition of the polishing fluid composition dispensed on the surface of the polishing pad while the camera is monitoring the state of the polishing pad surface. The time-series in-situ result data provided by the signals provided by the camera 299 or similar component and the flow rate sensing component and / or the polishing fluid composition detection component can be used by the artificial intelligence (AI) training platform 30A and the system controller 28 to cause adjustment processing to occur based on the interaction of different types of detected data, use the pad cooling component 220 to cause changes in the temperature of the polishing pad, and / or cause changes in the composition of the polishing fluid composition. Signals from these components can be provided to the system controller 28 and / or signals including training data can be delivered to the artificial intelligence (AI) training platform 30.

[0116] In another example, at activity 312, the camera 299 ( Figure 2A) is configured to detect the coverage and / or flow of polishing fluid across one or more areas of the polishing pad surface as the polishing fluid is dispensed onto the polishing pad. In this case, the camera 299 is positioned and configured to detect the amount of spread of the polishing fluid across the polishing surface of the polishing pad to determine the status of one or more of the components in the fluid delivery distribution system 281, such as detecting a blockage in one or more of the nozzles 283, detecting a change in the output of the fluid pump, and / or detecting a change in the position of the fluid delivery arm 282 relative to a desired position above the polishing pad surface and / or relative to the position of the substrate carrier 238 above the polishing pad. The amount of spread of the polishing fluid across the polishing surface of the polishing pad can be measured or determined by coverage of the horizontal area of the polishing pad or as a percentage of the field of view (FOV) of the camera 299. In some cases, the camera is also configured to detect a temperature gradient across the polishing pad surface and / or a change in temperature over time. In addition, the flow rate sensing assembly and / or polishing fluid composition detection assembly (e.g., pH sensor, abrasive particle concentration sensor) can also be configured to deliver a signal regarding the amount and / or composition of the polishing fluid composition dispensed onto the surface of the polishing pad while the camera monitors the coverage and / or flow of the polishing fluid across one or more regions of the polishing pad surface. The time-series in-situ result data provided from the signals provided by the camera 299 and the flow rate sensing assembly and / or polishing fluid composition detection assembly(ies) can be used by the artificial intelligence (AI) training platform 30A and the system controller 28 to, in subsequent activities, cause adjustments to the position of the fluid delivery arm 282 to adjust the position at which the polishing fluid is delivered to the surface of the polishing pad, cause an increase in the flow rate of the polishing fluid from one or more of the nozzles 283, cause a change in the temperature of the polishing pad using the pad cooling assembly 220, and / or cause a change in the composition of the polishing fluid composition based on detected interactions of different types of data during the subsequent activities.

[0117] At activity 314, method 300 includes repeating activities 304 through 312 for a plurality of substrates to obtain a corresponding plurality of training data sets. Here, each of the training data sets includes processing system data and in-situ result data that may be associated with a corresponding polished substrate.

[0118] At activity 316, the method 300 includes receiving, at an artificial intelligence (AI) training platform 30, training data 111 including a plurality of training data sets. In some embodiments, the plurality of training data sets includes data related to an amount of a slurry composition dispensed during a polishing process, a concentration of the dispensed slurry composition during a polishing process, a temperature of a polishing pad after the slurry composition is dispensed during a polishing process, polishing pad characteristics during a portion of a polishing process, and a time between pad conditioning processes, received from one or more polishing systems 20 over time to detect interactions between different data sets.

[0119] In one example, the plurality of training data sets collected and subsequently analyzed by the artificial intelligence (AI) training platform 30 include detection of trends in polishing process result data (e.g., dishing, wafer-to-wafer non-uniformity (WTWNU), planarization efficiency, and local flatness) based on detected interactions between data found in the training data sets including: detection of one or more polishing fluid compositions, detection of differences between different polishing fluid compositions (e.g., use of different abrasives or different amounts of a type of abrasive), detection of a certain type of substrate (e.g., oxide polishing process or metal polishing process), detection of polishing fluid flow rate, and / or detected trends in polishing pad temperature during multiple polishing processes performed in the one or more polishing systems 20.

[0120] In another example, at activity 316, the plurality of training data sets collected and subsequently analyzed by the artificial intelligence (AI) training platform 30 include detection of trends in optical properties of the surface of the polishing pad and / or a polishing fluid composition disposed thereon, and trends in variations in one or more polishing fluid compositions, or differences between different polishing fluid compositions (e.g., use of different abrasives or different amounts of a type of abrasive) on a certain type of substrate (e.g., oxide polishing or metal polishing).

[0121] In another example, at activity 316, the plurality of training data sets collected and subsequently analyzed by the artificial intelligence (AI) training platform 30 include detected coverage and / or flow of polishing fluid across one or more regions of the polishing pad surface, detected polishing fluid flow rate, and / or detected trends in the temperature of the polishing pad during a plurality of polishing processes performed in the one or more polishing systems 20.

[0122] At activity 318, method 300 includes training machine learning AI algorithm 110 using training data 111 to produce machine learning AI model 112. During activity 318, artificial intelligence (AI) training platform 30 can use machine learning AI model 112 to perform analysis on data currently received from various sources.

[0123] In one example, at activity 318, the artificial intelligence (AI) training platform 30 can determine that a detected increasing trend in polishing pad surface temperature may be caused by an increase in the concentration of abrasive particles in the polishing fluid composition or a decrease in dispensed polishing fluid based on receipt of data generated by the camera 299 and one or more polishing fluid composition detection components and use of the machine learning AI model 112. Based on previous and current analysis performed by the artificial intelligence (AI) training platform, the artificial intelligence (AI) training platform can determine that the detected increasing trend in polishing pad surface temperature is caused by improper mixing of a batch of polishing fluid composition or drift in a dosing mechanism responsible for controlling the composition of the process solution based on similar previously detected deviations occurring in one or more of the polishing systems 20.

[0124] In another example, the artificial intelligence (AI) training platform 30 can determine, based on receipt of data generated by the camera 299 and one or more polishing fluid composition detection components and use of the machine learning AI model 112, that a detected drift in the optical property of the surface of the polishing pad may be caused by a reduced effectiveness of a pad conditioning disk (e.g., the disk is wearing) based on similar previously detected trends in one or more of the polishing systems 20.

[0125] As discussed above, in another example, the artificial intelligence (AI) training platform 30 can, based on receipt of data generated by the camera 299 and other related sensors and use of the machine learning AI model 112, determine that a detected change in fluid coverage on one or more areas of the surface of the polishing pad may be caused by a blockage in one or more of the nozzles 283, a change in the output of the fluid pump, and / or a change in the position of the fluid delivery arm 282 relative to a desired position above the polishing pad surface based on similar previously detected trends in one or more of the polishing systems 20.

[0126] At activity 320, method 300 includes changing one or more of the plurality of polishing parameters in the process recipe based on the analysis performed using the machine learning AI model 112 during activity 318. In one example, the one or more polishing parameters changed based on the analysis performed by the AI algorithm may include adjusting an amount of a slurry composition dispensed during a current polishing process or a future polishing process, adjusting a concentration of a slurry composition dispensed during a current polishing process or a future polishing process, adjusting a temperature of a polishing pad after a slurry composition is dispensed during a current polishing process or a future polishing process, and / or causing a pad conditioning process to start or stop. Based on the analysis performed by the AI algorithm using the system controller 28 or the Fab production control system 40, respectively, the one or more of the plurality of polishing parameters changed may also be implemented on one or more polishing systems 20.

[0127] In one example, upon detecting a trend of increasing polishing pad surface temperature resulting from improper mixing of a batch of polishing fluid composition or drift in a polishing fluid component metering mechanism responsible for controlling the composition of the process solution, the artificial intelligence (AI) training platform 30 can instruct the system controller 28, or a user through the use of a graphical user interface (GUI) connected to the system controller 28, to replace the polishing fluid composition or metering mechanism and / or adjust one or more process variables in a polishing process recipe being run on current or future substrates being processed in the polishing system 20.

[0128] In another example, where a detected drift in the optical property of the surface of the polishing pad is caused by a reduced effectiveness of a pad conditioning disk, the artificial intelligence (AI) training platform 30 can instruct the system controller 28, or a user by using a GUI connected to the system controller 28, to replace the pad conditioning disk, adjust the dwell time of the conditioning disk on certain portions of the polishing pad, and / or adjust one or more process variables in a polishing process recipe being run on current or future substrates being processed in the polishing system 20.

[0129] As discussed above, in another example, upon detecting a drift in coverage and / or flow of the polishing fluid across one or more regions of the polishing pad surface, the artificial intelligence (AI) training platform 30 can instruct the system controller 28 to adjust the position of the fluid delivery arm 282 to adjust the location at which the polishing fluid is delivered to the surface of the polishing pad, cause an increase in the flow rate of the polishing fluid from one or more of the nozzles 283, use the pad cooling assembly 220 to cause a change in the temperature of the polishing pad, cause a change in the composition of the polishing fluid delivered from one or more of the nozzles 283, and / or adjust one or more process variables in a polishing process recipe being run on a current or future substrate being processed in the polishing system 20.

[0130] In some embodiments, method 300 includes removing a covering layer of material from a surface of a substrate, such as Figures 4A-4C Schematically shown in FIG. Figure 4A A substrate 400 is shown prior to a polishing process. The substrate 400 includes one or more material layers 401, 402, such as an epitaxial (Si) layer and a silicon nitride (SiN) layer disposed on the substrate 400. A plurality of openings are formed in the one or more material layers 401, 402 to form a patterned surface. A fill material layer 403 (e.g., an oxide layer (SiO2)) is deposited onto the patterned surface to fill the plurality of openings. The fill material disposed in the openings forms a plurality of features 403a (e.g., shallow trench isolation features), and a capping layer 403b of the fill material layer 403 still needs to be removed by polishing.

[0131] Figure 4B The capping layer 403b is shown partially removed using a polishing process, and Figure 4C The capping layer 403b is shown completely removed and the desired planar features 403a are shown remaining in the patterned surface.

[0132] Generally speaking, changes in the surface of the substrate 400 as the blanket of filler material 403b is removed (cleared) from the substrate 400 can be detected in time-series data generated using the in-situ substrate monitoring system 222. In some embodiments, such changes are detected using an endpoint algorithm executing on a controller of the polishing system. The endpoint algorithm triggers a change in the polishing process when the blanket of material is cleared from the field surface of the substrate during an STI or damascene process. Unfortunately, such reactive endpoint detection schemes can result in over-polishing of the substrate surface, leading to undesirable dishing and erosion of features in its surface.

[0133] In some embodiments, the AI algorithm 110 is trained to identify a functional relationship between the time-series in-situ result data 124 and the processing system data 114 (e.g., time-series data of individual or combined control parameters). The functional relationship can be used by the trained AI algorithm 110 and / or the generated machine learning AI model 112 to predict a time range for the polishing endpoint before, rather than simultaneously with, the start of the removal of the material overburden from the substrate surface. Based on the predicted time range, the polishing fluid composition at the surface of the polishing pad can be altered to provide better local planarization performance.

[0134] In some embodiments, changing one or more of the plurality of polishing parameters based on the machine learning AI model 112 at activity 318 includes changing a composition of a polishing fluid disposed on the surface of the polishing pad based on the functional relationship. In some embodiments, changing the composition of the polishing fluid includes starting, stopping, or changing a flow rate of each polishing fluid component delivered to the surface of the polishing pad.

[0135] In some embodiments, the training data 111 used to train the machine learning AI algorithm 110 further includes Figure 1B and Figure 1C Any portion or combination of the substrate tracking data 128, facility system data 130, and electrical test data 132 described in.

[0136] Figure 5 is a diagram illustrating a method 500 of matching polishing performance between polishing systems.

[0137] At activity 502, method 500 includes receiving, at an artificial intelligence (AI) training platform 30, training data comprising a plurality of training data sets. Here, different ones of the plurality of training data sets correspond to substrates polished using different combinations of polishing stations and substrate carrier assemblies of a polishing system. Each of the training data sets includes processing system data associated with each of the substrates polished using the polishing system.

[0138] Here, each training data set includes processing system data 114, which includes polishing recipe data 118 and control parameter data 120. The polishing recipe data 118 includes a plurality of polishing parameters and a plurality of target values corresponding thereto. The control parameter data 120 includes time series data of control parameters of one or more closed-loop control systems. The one or more closed-loop control systems are used to maintain the corresponding polishing parameters at or near their target values.

[0139] At activity 504 , method 500 includes training a machine learning AI algorithm using the training data. Here, the trained machine learning AI algorithm is configured to recognize differences between different substrate carrier assemblies and / or different polishing stations of the polishing system.

[0140] At activity 506 , method 500 includes implementing one or more corrective actions based on the identified differences.

[0141] In some embodiments, the method 500 is used to identify differences between different substrate carrier assemblies and / or different polishing stations across multiple polishing systems and implement one or more corrective actions based thereon.

[0142] Advantageously, the machine learning AI system and AI algorithm training methods described herein can be used to better understand and utilize the combined capabilities of devices and subsystems of advanced CMP processing systems, thereby improving polishing results, making process margins desirably wider, and improving process consistency of the polishing system.

[0143] While the foregoing is directed to embodiments of the present disclosure, other and further embodiments of the disclosure may be devised without departing from the basic scope thereof, which is determined by the claims that follow.

Claims

1. A computer-implemented method for polishing a substrate, the method comprising: Polish the substrate using a polishing system including: (a) flowing a polishing fluid onto a surface of a polishing pad according to a polishing recipe, the polishing recipe comprising a plurality of polishing parameters and corresponding plurality of target values; (b) placing a substrate against the surface of the polishing pad according to the polishing recipe; (c) maintaining a first polishing parameter among the plurality of polishing parameters at or near a target value of the first polishing parameter by adjusting a first control parameter; (d) generating processing system data, the processing system data including time series data of the polishing recipe and the first control parameter; and (e) using measurements obtained from an in-situ substrate monitoring system concurrently with (a)-(d) to generate time series in-situ result data; Repeating (a)-(e) for a plurality of substrates to obtain a corresponding plurality of training data sets, each of the training data sets comprising the processing system data and the in-situ result data for a polished substrate; Receiving, at an artificial intelligence (AI) training platform, training data comprising the plurality of training data sets, wherein at least a portion of the plurality of training data sets are received in a time sequence; and One or more of the plurality of polishing parameters are changed based on analysis of the received training data performed by a machine learning AI algorithm.

2. The method of claim 1 , wherein the target value comprises a desired set point for each of the polishing parameters, a value greater than a desired lower threshold, a value less than a desired upper threshold, and / or a value between the desired lower threshold and the desired upper threshold.

3. The method of claim 1, wherein: The in-situ results data includes data derived from a signal provided from a camera positioned to view at least a portion of the surface of the polishing pad and configured to detect a change in temperature of at least a portion of the surface of the polishing pad.

4. The method of claim 3, wherein: The first polishing parameter includes a temperature of the surface of the polishing pad, and The first control parameter includes a flow rate of a coolant delivered to the surface of the polishing pad or a flow rate of a polishing fluid delivered to the surface of the polishing pad.

5. The method of claim 1 , wherein the in-situ result data comprises: data derived from a signal provided by a camera positioned to detect where the polishing fluid is dispensed on the surface of the polishing pad, or Data is derived from a signal provided from a camera positioned to detect an amount of coverage of the polishing fluid dispensed from a polishing fluid delivery nozzle onto the surface of the polishing pad.

6. The method of claim 5, wherein the first control parameter comprises: the flow rate of the polishing fluid delivered to the surface of the polishing pad, or The polishing fluid delivery nozzle is positioned relative to the surface of the polishing pad.

7. The method of claim 1 , wherein the in-situ result data comprises: data derived from a signal provided from a camera positioned to detect a temperature of at least a portion of the surface of the polishing pad, and Data is derived from a signal provided from a sensor configured to detect a constituent of the polishing fluid.

8. The method of claim 7, wherein: The first polishing parameter includes a temperature of the surface of the polishing pad, and The first control parameter includes a flow rate of a coolant delivered to the surface of the polishing pad or a flow rate of a polishing fluid delivered to the surface of the polishing pad.

9. The method of claim 1, wherein: The in-situ result data comprises data derived from a signal provided from a camera positioned to detect the roughness of the surface of the polishing pad or positioned to detect an optical property of the surface of the polishing pad, The first polishing parameters include pad conditioning parameters of the surface of the polishing pad, and The first control parameter includes a rotational speed of the conditioning disk, a downforce applied to the conditioning disk against the polishing pad, a dwell time of the conditioning disk on one or more portions of the surface of the polishing pad, or a sweep speed of the conditioning disk across the surface of the polishing pad.

10. The method of claim 1, wherein maintaining the first polishing parameter at or near a target value of the first polishing parameter comprises: i. determining the difference between the actual value of the first polishing parameter and the target value of the first polishing parameter; ii. based on the determined difference, changing the first control parameter of the first control system; as well as iii. Continuously repeating i. and ii. to provide closed-loop control of the first polishing parameter.

11. The method of claim 10, wherein the first polishing parameter comprises a temperature of the surface of the polishing pad.

12. The method of claim 11, wherein: The polishing fluid includes a slurry composition, and The first control parameter includes a flow rate or amount of the slurry composition delivered to the surface of the polishing pad.

13. The method of claim 12, wherein the first control parameter comprises a flow rate of a coolant delivered to the surface of the polishing pad.

14. The method of claim 10, wherein changing one or more of the plurality of polishing parameters based on analysis of the received training data performed by the machine learning AI algorithm further comprises: Using the training data to train a machine learning AI algorithm, and wherein The trained machine learning AI algorithm identifies the functional relationship between the time series in-situ result data and the time series data of the first control parameter, and Changing one or more of the plurality of polishing parameters includes changing a composition of the polishing fluid disposed on the surface of the polishing pad based on the functional relationship.

15. The method of claim 14, wherein changing the composition of the polishing fluid comprises: The flow rate of each polishing fluid component delivered to the surface of the polishing pad is started, stopped, or changed.

16. The method of claim 1, wherein the training data used to train the machine learning AI algorithm further comprises one or a combination of the following: substrate tracking data including a processing history of one or more of the plurality of substrates and / or information related to components formed on one or more of the plurality of substrates; facility system data, including information generated using one or more facility supply systems, including analytical information of polishing fluid delivered to the polishing system from a remote polishing fluid distribution system; and Electrical test data includes electrical test information generated from one or more of the plurality of substrates during a post-polish electrical test measurement operation.

17. A computer-implemented method for matching polishing performance between polishing systems, the method comprising: Training data including multiple training data sets is received at an artificial intelligence (AI) training platform, wherein Each of the training data sets includes processing system data associated with a respective substrate of a first plurality of substrates polished using a first polishing system, Different ones of the first plurality of substrates are polished using different combinations of substrate carrier assemblies from the plurality of substrate carrier assemblies and polishing stations from the plurality of polishing stations of the first polishing system, and The processing system data for each of the training data sets includes: a polishing recipe comprising a plurality of polishing parameters and a corresponding plurality of target values, wherein a corresponding closed-loop control system is used to maintain one or more of the plurality of polishing parameters at or near the target value of the one or more of the plurality of polishing parameters; and Time series data of control parameters of the closed-loop control system; and training a machine learning AI algorithm using the training data, wherein the trained machine learning AI algorithm is configured to identify differences between the different combinations of substrate carrier assemblies or different polishing stations of the first polishing system; and One or more corrective actions are implemented based on the identified differences.

18. The method of claim 17, wherein The plurality of training data sets further include processing system data associated with respective substrates of a second plurality of substrates polished using a second polishing system, different ones of the second plurality of substrates are polished using different combinations of substrate carrier assemblies from the plurality of substrate carrier assemblies and polishing stations from the plurality of polishing stations of the second polishing system, The trained machine learning AI algorithm is configured to identify differences in the different combinations of substrate carrier assemblies and / or the different polishing stations of the first polishing system and the second polishing system; and One or more corrective actions are implemented based on the identified differences.

19. The method of claim 18, wherein each of the plurality of training data sets further comprises time-series in-situ result data obtained from in-situ substrate monitoring systems corresponding to the plurality of polishing stations of the first polishing system and the second polishing system.

20. The method of claim 19, wherein the in-situ substrate monitoring system comprises a camera positioned to view and configured to detect a change in temperature of at least a portion of a surface of a polishing pad disposed within the first polishing system.

Images