Adaptive slurry dispensing system

Abstract

Provided herein is an advanced substrate polishing method that uses machine learning artificial intelligence (AI) algorithms, or software applications generated using AI, to control one or more aspects of the polishing process. The AI ​​algorithms are trained to simulate the polishing process using substrate processing data obtained from the polishing system and to make predictions about the polishing process and expected process results from the polishing process.

Description

Technical Field

[0001] The embodiments described herein generally relate to semiconductor device manufacturing, particularly chemical mechanical polishing (CMP) systems used in semiconductor device manufacturing, and related methods.

Background Art

[0002] Chemical mechanical polishing (CMP) is commonly used to fabricate high-density integrated circuits, to planarize a material layer on a substrate, to remove excess material from the surface of a lower material layer, or both. In a typical CMP process, a substrate is held within a carrier head that presses the backside of the substrate against a rotating polishing pad in the presence of a polishing fluid. The polishing pad is often formed of a polymeric material, and the asperity (roughness) of its surface facilitates the transfer of the polishing fluid to the interface between the material surface of the substrate and the underlying moving polishing pad. The polishing fluid typically comprises an aqueous solution of one or more chemical components and nanoscale polishing particles suspended in the aqueous solution, and is often referred to as a polishing slurry. The combination of chemical activity and mechanical movement provided by the polishing fluid, the relative movement between the substrate and the polishing pad, and the contact pressure therebetween causes material to be removed across the entire surface of the material layer of the substrate. Consumables such as the polishing pad and the polishing fluid are selected based on the desired CMP application.

[0003] General CMP applications include planarization of bulk films and removal of excess material in the damascene process. Planarization of bulk films, such as polishing of the interlayer dielectric (ILD), is typically used to smooth out unwanted recesses and protrusions on the surface of the material layer caused by underlying two - dimensional or three - dimensional features. Typical damascene CMP applications include shallow trench isolation (STI) and formation of interlayer metal wiring, where CMP is used to remove the material (overburden) that fills trenches, contacts, vias, or lines from the exposed surface (field) of one or more underlying layers where STI or metal wiring features are located.

[0004] Depending on the application, the results of the CMP process are typically characterized by a combination of interrelated metrics related to global polishing uniformity, local planarization performance, and surface defects induced by CMP. Such process results determine the performance, reliability, and / or operability of the resulting devices formed on the substrate. Process results that deviate from the limits of the process tolerances can lead to device failures and thus potentially reduce the yield of usable devices formed on the substrate. Typically, the tolerances of the process results decrease as the circuit density increases and the device feature sizes shrink.

[0005] To meet industry demands for reducing device form factors, advanced CMP systems have become dramatically more complex in order to control virtually all process variables (parameters) known to affect process results. Such advanced CMP systems include highly engineered and complex individual subsystems, each configured to control one or more process parameters to a desired set point. The controllable process parameters collectively define the substrate polish recipe. In many cases, the polish recipe for a single substrate CMP process includes multiple stages of a polish sequence, where one or more parameter set points are changed for each stage of the sequence.

[0006] Unfortunately, progress in CMP technology far outpaces the scientific understanding of the complex interactions between the surface, fluid, and abrasive grains at the polish interface. As a result, existing CMP models are generally not suitable for use during process development. Thus, CMP substrate processes are typically determined and / or refined based on conventional process development and improvement techniques, such as design of experiments (DOE) and trial and error. Typically, standard quality control measures prohibit experiments on production substrates containing devices intended for use or sale. As a result, DOE experiments are often performed using expensive test substrates, consuming valuable time on the CMP processing system. Thus, due to the associated time and costs, it is virtually impossible to thoroughly investigate the complex relationships between polish parameters, algorithms, consumables, device features, and process results for multiple individual polish processes used within a production facility.

[0007] Thus, conventional process improvement methods are insufficient to leverage the combined capabilities of the advanced CMP processing system's equipment and subsystems and cannot provide improved process results and a wider process window that might otherwise be achievable.

[0008] Correspondingly, in the art, there is a need for an advanced processing method that is not plagued by the drawbacks described above. SUMMARY OF THE INVENTION

[0009] Embodiments described in the specification generally relate to a chemical mechanical polishing (CMP) system used in the manufacture of electronic devices and an advanced substrate processing method used 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 liquid onto the 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 the substrate against the surface of the polishing pad according to the polishing recipe; (c) adjusting a first control parameter to maintain a first polishing parameter among the plurality of polishing parameters at or near its target value; (d) generating processing system data including time-series data of the polishing recipe and the first control parameter; and (e) simultaneously with (a)-(d), generating time-series in-situ result data 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 training data set including processing system data and in-situ result data for the polished substrate; receiving training data including the plurality of training data sets in an artificial intelligence (AI) training platform; training a machine learning AI algorithm using the training data; and using the trained machine learning AI algorithm to change one or more of the plurality of polishing parameters.

[0011] In one embodiment, a computer-readable medium includes instructions for performing a method of determining a polishing recipe. The method includes receiving, in an artificial intelligence (AI) training platform, training data including a plurality of training data sets, each of the training data sets including process system data and in-situ result data associated with a substrate polished by a polishing system. The process system data for each of the training data sets includes a polishing recipe including a plurality of polishing parameters and corresponding 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. 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 includes training a machine learning AI algorithm using the training data, and determining, using the trained machine learning AI algorithm, a functional relationship between the in-situ result data and the time series data of the first control parameter.

[0012] In one embodiment, a computer-implemented method for matching polishing performance among polishing systems is provided. The computer-implemented method includes receiving training data including a plurality of training data sets in an artificial intelligence (AI) training platform. Each of the training data sets includes process system data associated with an individual substrate among a first plurality of substrates polished using a first polishing system, and various substrates among the first plurality of substrates are polished using various combinations of a substrate carrier assembly from a plurality of substrate carrier assemblies of the first polishing system and a polishing station from a plurality of polishing stations of the first polishing system. The process system data for each of the training data sets is a polishing recipe including a plurality of polishing parameters and corresponding target values, wherein one or more of the plurality of polishing parameters are maintained at or near their target values using a corresponding closed-loop control system, and the polishing recipe includes 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 various substrate carrier assemblies of the first polishing system and / or between various polishing stations. The method further includes performing one or more corrective actions based on the identified differences.

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

[0014] Embodiments of the present disclosure also provide a computer-implemented method for polishing a substrate in one or more polishing systems. The computer-implemented method includes: (a) flowing a polishing liquid onto the surface of a polishing pad according to a polishing recipe, where the polishing recipe may include 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 among the plurality of polishing parameters at or near its target value by adjusting a first control parameter; (d) generating process system data that may include time-series data of the polishing recipe and the first control parameter; (e) simultaneously with (a)-(d), generating time-series in-situ result data using measurements obtained from an in-situ substrate monitoring system; repeating (a)-(e) for a plurality of substrates to obtain a corresponding plurality of training data sets, where each of the training data sets may include process system data and in-situ result data for the polished substrate; receiving training data that may include the plurality of training data sets in an artificial intelligence (AI) training platform, where at least a portion of the plurality of training data sets are received sequentially over time; and changing one or more of the plurality of polishing parameters based on an 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 comprising receiving, in an artificial intelligence (AI) training platform, training data including a plurality of training data sets, each of the training data sets including process system data associated with an individual substrate among a first plurality of substrates polished using a first polishing system, wherein various substrates among the first plurality of substrates are polished using various combinations of a substrate carrier assembly from a plurality of substrate carrier assemblies of the first polishing system and a polishing station from a plurality of polishing stations of the first polishing system, and the process system data for each of the training data sets is a polishing recipe including a plurality of polishing parameters and corresponding target values, wherein one or more of the plurality of polishing parameters are maintained at or near the target value using a corresponding closed-loop control system, the polishing recipe, and time-series data of control parameters of the closed-loop control system, the method further comprising training a machine learning AI algorithm using the training data, wherein the trained machine learning AI algorithm is configured to identify differences between various substrate carrier assemblies or various polishing stations of the first polishing system, and performing one or more corrective actions based on the identified differences.

[0016] To better understand the above features of the present disclosure, a more specific description of the present disclosure briefly summarized above can be provided by referring to the embodiments, some of which are shown in the accompanying drawings. However, it should be noted that since the present disclosure may admit other equally effective embodiments, the accompanying drawings show only typical embodiments of the present disclosure and should not be considered as limiting the scope of the present disclosure.

Brief Description of the Drawings

[0017]

Figure 1A

Figure 1B

Figure 1C

Figure 1D

Figure 2A

Figure 2B

Figure 2C

Figure 3

Figure 4A-C

Figure 5

BEST MODE FOR CARRYING OUT THE INVENTION

[0018] For ease of understanding, the same reference numerals are used to denote the same elements common to each figure whenever possible. It is assumed that the elements and features of one embodiment can be beneficially incorporated into other embodiments without further description.

[0019] The embodiments described in the specification generally relate to chemical mechanical polishing (CMP) systems used in the manufacture of electronic devices, and in particular, to advanced substrate processing methods used therewith.

[0020] Broadly, the advanced substrate polishing method of the present specification uses algorithms such as machine learning artificial intelligence (AI) algorithms, or software applications generated using AI algorithms, to control one or more aspects of the polishing process. Generally, an AI system utilizes a large dataset containing intelligent iterative processing algorithms to learn from the patterns and features of the data to be analyzed. By performing rounds of data processing, each time the AI system analyzes the data, it comprehensively tests and measures its own performance and develops additional expertise based on the analysis performed. Here, the AI algorithm is trained to simulate the polishing process using substrate processing data obtained from the polishing system and make predictions about the polishing process and the process results expected from the polishing process.

[0021] In some embodiments, the AI algorithm, or software application generated using the AI algorithm, is used to predict the planning time horizon of a desired polishing endpoint and adjust the composition of the polishing liquid with respect thereto, for example, by starting, stopping, or changing the flow rates of one or more polishing liquid components. As used herein, "polishing endpoint" refers to a point in a polishing process where it may be desirable to change one or more substrate polishing parameters such as the slurry composition, and does not necessarily indicate the end of the polishing process. For damascene applications, the ability to accurately predict the desired polishing endpoint and pre-adjust the polishing liquid composition (e.g., the slurry composition) based on that prediction promotes an improvement in local planarization performance compared to conventional reactive endpoint detection schemes. The improvement in local planarization performance results in desirable improvements in the performance, reliability, and yield of the resulting device. An example of poor local planarization that can be improved using the method provided herein is shown in FIG. 1A.

[0022] As further described below, a polishing liquid composition (e.g., a slurry composition) generally includes a mixture of one or more solid particles suspended in a liquid such as water. The solid particles are often referred to as abrasives and can include metal oxide fine powders such as CeO2, Fe2O3, Al2O3, SiO2 suspended in the liquid. The above liquid can include one or more of acids, bases, and various additives (e.g., corrosion inhibitors, pH adjusters) often disposed in water.

[0023] FIG. 1A is a schematic cross-sectional view showing poor local planarization, e.g., erosion up to a distance e and dishing up to a distance d, after a polishing process for removing overfill of a metal filling material from the field surface, i.e., the upper surface or the outer surface, of a substrate 1. Here, the substrate 1 includes a dielectric layer 2, a first metal interconnect feature 3a formed in the dielectric layer 2, and a plurality of second metal interconnect features 3b formed in the dielectric layer 2. The plurality of second metal interconnect features 3b are densely arranged to form a region 4 with a relatively high feature density. Typically, the metal interconnect features 3a, 3b are formed by depositing a metal filling material on the dielectric layer 2 and depositing the metal filling material into corresponding openings formed in the dielectric layer 2. Thereafter, the material surface of the substrate 1 is planarized using a CMP process to remove overfill of the filling material from the field surface 5 of the dielectric layer 2.

[0024] As shown, due to poor local planarization performance, the upper surface of the metal interconnect feature 3a is recessed by a distance d from the surrounding surface of the dielectric layer 2, which is known as dishing. Due to poor local planarization performance, unwanted depressions in the dielectric layer 2 also occur within the high feature density region 4, for example at a distance e, where the upper surface of the dielectric layer 2 within the region 4 is recessed from the plane of the field surface 5, which is known as erosion. The metal loss resulting from dishing and / or erosion causes unwanted variations in the effective resistance of the metal interconnect features 3a, 3b formed therefrom, and thus may affect the performance and reliability of the device.

[0025] In some embodiments, the AI algorithm is trained using data from one or more polishing systems operating within a production facility, i.e., a semiconductor device manufacturing facility. Training the AI algorithm using a production polishing system advantageously provides rich data that the AI algorithm can use and better understands the complex relationships between multiple variables for a particular polishing application. An exemplary manufacturing facility (Fab) 10 is schematically shown in FIG. 1B.

[0026] Here, Fab10 includes a plurality of polishing systems 20, one or more machine learning artificial intelligence (AI) algorithm (hereinafter, "AI") training platforms 30, a Fab production control system 40, one or more stand-alone substrate inspection and / or metrology stations 50, and other processing systems 60. The other processing systems 60 include substrate processing systems used in the manufacture of semiconductor devices that are seen both upstream and downstream of the polishing process in the substrate processing flow, and the substrate processing systems include, for example, an epitaxial system, a heat treatment system, a non-epitaxial deposition system, a lithography system, an etching system, an implantation system, and other polishing systems. In some embodiments, Fab10 further includes one or more electrical test systems 70, such as a parametric test and / or device yield test system that communicates with the Fab production control system 40.

[0027] Typically, 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 between the carrier assemblies 22, and a carrier transfer system 24 for moving the substrate carrier assemblies 22 between the carrier loading station 23 and the various polishing stations 21. Here, each of the polishing systems 20 further includes one or more substrate inspection systems 25, one or more measurement systems 26, and a cleaning system 27, which are integrated with the polishing system 20 to perform pre- and / or post-polishing (in-line) inspection, measurement, and cleaning of the substrates polished therein, respectively. Each of the polishing systems 20 includes a system controller 28 that directs and adjusts the operation of the various components and subsystems of the polishing system 20.

[0028] As shown, each of the AI training platforms 30 is communicatively connected to a corresponding system controller 28 using a communication link 29 such as an Ethernet or USB connection. In other embodiments, one or more of the AI training platforms 30 may be integrated with and form part of the system controller 28. 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, an individual AI training platform 30 may be used with more than one polishing system 20 to perform the methods described herein, and / or individual AI training platforms 30 may be communicatively connected to each other to share training data 111 (FIG. 1C). The training data 111 may be shared among the individual AI training platforms 30 at multiple different times. In one example, the training data 111 can be shared sequentially in a timely manner, which can include sharing at regular time intervals, or sharing during or after one or more sequentially executed processes are performed within the polishing system 20, and / or sharing during or between one or more asynchronous processes performed by multiple polishing systems 20. In other embodiments, one or more of the AI training platforms 30 are not physically located within Fab10, and the methods described herein are implemented using cloud computing techniques.

[0029] The Fab production control system 40 instructs the flow and processing of substrates passing through the production line, and collects and manages data related to both the substrates and the processing system. Typically, the system controller 28 communicates with the Fab production control system 40, and the Fab production control system 40 gives commands to the system controller 28 and receives information from the system controller 28. Here, the Fab production control system 40 further communicates with one or more stand-alone substrate inspection and / or measurement stations 50, other processing systems 60, and one or more electrical test systems 70. In some embodiments, the Fab production control system 40 transmits the information received from the stand-alone substrate inspection and / or measurement stations 50, other processing systems 60, and one or more electrical test systems 70 to the system controller 28, and the information is used as training data 111 (FIG. 1C) by the AI training platform 30. 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.

[0030] Figure 1C is a schematic diagram of a process improvement scheme 100 that can be used with the method described herein. The process improvement scheme 100 uses an AI training platform 30, which includes a processor and a memory block (PMB (processor and memory block) 104). This processor and memory block 104 is operable with a support circuit 32 to execute a machine learning AI algorithm, namely the AI algorithm 110 herein. The processor of PMB 104 (not shown separately) is one or a combination of computer processors suitable for executing the AI algorithm 110, for example, a programmable central processing unit (CPU), a graphics processing unit (GPU), a field programmable gate array (FPGA), an application-specific integrated circuit for machine learning (ASIC), or one or more of other suitable hardware implementations. The memory of PMB 104 (not shown separately) is operably connected to the processor, is non-transitory, and is any non-volatile type of memory of a size suitable for storing the AI algorithm 110, the learning data 111 used with the AI algorithm 110, and one or more machine learning AI models 112 generated using the AI algorithm 110. The support circuit 32 has conventionally been connected to a central processing unit and includes a cache, a clock circuit, an input / output subsystem, a power supply, etc., and combinations thereof.

[0031] Here, the AI algorithm 110 is trained using one or a combination of supervised learning models and unsupervised learning models with the learning data 111 stored in the memory of the PMB 104. In an example of a supervised learning model, the AI algorithm 110 can be trained to map input data such as time-series data of individual control parameters to output data such as individual processing results based on exemplary input-output pairs provided by the user. In an example of an unsupervised learning model, the AI algorithm 110 can be trained to find patterns and relationships in the training data 111 received over time with minimal user input. The training process can be carried out based on multiple data sets received from various in-situ sensors or ex-situ sensors over a long period of time.

[0032] In embodiments where the AI algorithm 110 includes a supervised model, a support vector machine (SVM) can be used, or any supervised learning model that can receive the learning data 111 and provide a continuous output indicating or predicting a processing result can be used. In embodiments where the AI algorithm 110 includes an unsupervised model, a neural network can be used, or any unsupervised learning model that can receive the training data 111 for training the AI algorithm 110 and provide a clustered and classified output indicating and / or predicting one or more processing results can be used. In some embodiments, such as embodiments where the training data includes images of various components of the polishing system 20 and / or images of substrates processed within the polishing system 20, the AI algorithm 110 can use a convolutional neural network.

[0033] Here, the training data 111 includes the processing system data 114 generated by the polishing system 20 or its subsystems, and the corresponding processing result data 116 for one or more substrates processed on the polishing system 20. Here, the processing system data 114 used to train the AI algorithm 110 includes the polishing recipe parameter data 118, for example, individual polishing parameters and their corresponding target values, and the control parameter data 120 provided by one or more parameter control systems 201a - n as described in FIGS. 2A - 2C, and the process monitoring data 122 regarding the operation and processing performance of the subsystems and / or their consumables generated by, for example, additional sensors or measuring devices arranged within the polishing system 20. The processing system data 114 generated by the polishing system 20 or its subsystems can be represented by discrete values as provided in the polishing recipe, or can include time - series data, for example, a series of data points (or images) arranged in chronological order.

[0034] In some embodiments, the AI training platform 30 is communicably connected to one or more components of the polishing system 20, and at least a part of the processing system data 114 is received therefrom. In some embodiments, at least a part of the processing system data 114 is stored in the memory of the polishing system controller 28, and the AI training platform 30 receives the processing system data 114 therefrom.

[0035] The processing result data 116 is information regarding the planarization and / or removal of the material layer from the substrate during the polishing process obtained through measurement or inspection of the substrate, and includes information obtained from the measurement or inspection of the substrate. In some embodiments, the processing result data 116 includes an image of the surface of the substrate taken using, for example, a camera device.

[0036] Here, the processing result data 116 includes, for example, substrate measurement values (in-situ result data 124) acquired simultaneously with the polishing process and substrate measurement values (ex-situ result data 126) obtained subsequent to the polishing process, using an eddy current sensor or an optical sensor as described in FIG. 2A below. In some embodiments, the in-situ result data 124 includes time-series data. In some embodiments, the processing result data 116 includes the difference between measurement values acquired before the polishing process, such as the material removal rate or material removal uniformity, and measurement values acquired thereafter.

[0037] Here, the in-situ result data 124 includes time-series eddy current information and / or time-series optical signal information acquired using the in-situ substrate monitoring system 222 described in FIG. 2A. The in-situ result data 124 typically includes signal information and may include information derived from the signal information, such as the thickness of the material layer and the uniformity information of the material layer.

[0038] The ex-situ result data 126 can be generated using an appropriate measurement or inspection system typically found within a semiconductor device manufacturing facility. In some embodiments, at least a portion of the ex-situ result data 126 is generated using one or more in-line inspection systems 25 and / or measurement systems 26 of the polishing system 20, and a portion of the ex-situ result data 126 is received by the AI training platform 30 therefrom. In some embodiments, at least a portion of the ex-situ result data 126 is stored in the memory of the polishing system controller 28 communicatively connected to the in-line systems 25, 26, and the AI training platform 30 receives a portion of the ex-situ result data 126 from the processing system controller 28.

[0039] In some embodiments, at least a portion of the excitue result data 126 is generated using one or more stand-alone inspection and / or metrology stations 50 separate from the polishing system 20. Typically, in the same embodiment, the excitue result data 126 is collected and / or received from a Fab production control system 40 communicatively connected to each of the stand-alone inspection and / or metrology stations 50.

[0040] Examples of information that can form part of the excitue result data 126 include the material removal rate (MRR), the planarization (global flatness) of the material layer, the uniformity between substrates, i.e., wafer-to-wafer non-uniformity (WTWNU), the uniformity of the material removal rate across the substrate surface, and / or the uniformity of the thickness of the planarized material layer, collectively referred to as the within-wafer non-uniformity (WTWNU) indicator values, the planarization efficiency, the local flatness, e.g., within-die (WID) flatness, the undesired removal of the underlying material layer, e.g., oxide film loss, the erosion of the underlying material layer in regions of high feature density, the depression (dishing) of the material in trenches, contacts, vias, and / or line features, the defects induced by polishing on the substrate surface or on the exposed features formed on the substrate surface, and / or on the substrate surface or on the exposed features, and the defects induced by polishing on the substrate surface or on the exposed features formed on the substrate surface. Defects induced by CMP include mechanical related defects such as scratches and chemical related defects such as corrosion of metal features.

[0041] In some embodiments, the exit result data 126 includes images obtained from in-line and / or stand-alone measurement and / or inspection systems, such as an image of a substrate obtained using a camera device or other optical sensor. In some embodiments, the exit result data is an image generated by a measurement or inspection system and includes information obtained from the substrate, such as an image representing the thickness, flatness, defects, and / or stress map of the substrate and / or a material layer on the substrate surface.

[0042] In some embodiments, the training data 111 includes one or more of the substrate tracking data 128, the equipment system data 130, and the electrical test data 132. Here, the substrate tracking data 128 includes the identification information of the substrate, the information about the devices formed on the substrate, and the processing history of the substrate. Examples of device information include the size of the device, the device shape, the size of the feature, and the pattern density. The processing history typically includes the identification of the upstream processing system and the corresponding processing information (e.g., day / time information, the processing recipe used). The processing history may also include information obtained from upstream measurement and / or inspection systems.

[0043] The equipment system data 130 includes information regarding the equipment supply system coupled to the polishing system 20 and / or information related to environmental conditions surrounding the polishing system 20 such as, for example, temperature, particle count, and air flow. Examples of information related to the equipment supply system include information obtained from a deionized (DI) water supply system, a clean dry air (CDA) supply system, a chemical supply system, and a remote polishing fluid dispensing system. Typically, the remote polishing distribution system circulates the polishing fluid through the equipment line to supply a plurality of polishing systems 20 that are fluidly coupled to the equipment line at the point of use. Such a polishing fluid dispensing system is often configured for bulk mixing of the polishing fluid and may include one or more analyzers to facilitate the mixing process and / or continuously monitor the integrity of the polishing fluid. Monitoring the integrity of the polishing fluid involves using the analyzer to determine and monitor the chemical properties of the polishing fluid (e.g., pH, levels of oxidizing agents and additives, and their attenuation behavior) and the polishing properties of the polishing fluid including large particle counts (LPC), mean particle size distribution (PSD), density, weight percent solids, and viscosity. Information related to the equipment system including the integrity of the polishing fluid is communicated to the individual system controllers 28 of the plurality of polishing systems 20 and / or the fab production control system 40 and can be received therefrom by the AI training platform 30.

[0044] The electrical test data 132 can include, for example, parametric test information generated in a subsequent parametric test process and / or device test information generated in one or more subsequent device test processes using a dedicated test structure disposed within the die line between devices. In some embodiments, the electrical test data 132 includes images representing information obtained during the parametric test process and / or the device test process, including, for example, a device yield map representing the locations on the substrate of operational and failed devices.

[0045] Here, the training data 111 includes identification information such as substrate tracking information, system information, and timestamp information, and this identification information can be used to correlate the information received from each of the above-described data sources with a specific combination of a substrate, a polishing system, a polishing station, and a substrate carrier that forms a set of corresponding training data.

[0046] In some embodiments, the trained AI algorithm 110 is used to generate an AI model 112, such as a software algorithm, and the AI model 112 is communicated to the system controller 28 for use as instructions for instructing the operation of the polishing system 20.

[0047] FIG. 1D is a schematic diagram 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 used by the control system 150 to maintain the polishing parameters at or near a target value 156. As used herein, "target value" includes a desired setpoint, a value above a desired lower threshold, a value below a desired upper threshold, and a value between a desired lower threshold and upper threshold.

[0048] In FIG. 1D, the process control system 150 provides a closed feedback control loop for maintaining the polishing parameters at or near the target value 156. As shown, the process control system 150 includes a sensor 151, a controller 152, and a parameter control device 153 operably connected to the controller 152, such as an actuator. Here, the sensor 151, the controller 152, and the control device 153 are arranged such that information flows within a feedback loop 154, providing a closed-loop feedback control system.

[0049] During the polishing process, sensor 151 measures the actual value 155 of the polishing parameters (e.g., platen rotation speed, slurry flow rate, etc.), and controller 152 determines the error between the actual value 155 and the target value 156. To correct the error, controller 152 commands parameter control device 153 (e.g., an actuator (motor) connected to the platen, a slurry dispensing pump connected to the slurry supply system, etc.) to change control parameter 157 (e.g., motor current, pump pressure, pump speed, etc.), thereby causing a corresponding change in the polishing parameter output (e.g., platen rotation speed, slurry flow rate, etc.).

[0050] Parameter control system 150 is generally reactive. When the polishing parameters increase until the target value 156 is reached, the change in control parameter 157 by controller 152 indicates a response to the change in the polishing process. Similarly, changes in control parameter 157 between substrates for substantially similar polishing processes may indicate an undesirable process drift. Thus, in the embodiments of this specification, time-series control parameter data 120 is included in processing system data 114, and AI algorithm 110 can better understand the complex relationships among the subsystems, processing parameters, consumables, and substrates for a specific polishing process.

[0051] FIG. 2A is a schematic side cross-sectional view of a polishing station 21 and a carrier assembly 22 that can be used with the method described herein, according to one embodiment. Here, the polishing station 21 includes a plurality of subsystems each operable with one or a combination of parameter control systems 201a-n. Here, each of the parameter control systems 201a-n is configured to include a closed feedback control loop and may include any one or a combination of the elements of the process control system 150 described in FIG. 1D.

[0052] Typically, each of the control systems 201a - n includes one or more corresponding actuators 202a - n, process parameter sensors 203a - n, controllers 204a - n, and control parameter sensors 205a - n. The actuators 202a - n include any device or processing system operable to change a control parameter in response to a signal received from the controllers 204a - n, such as an electrical signal, a pneumatic signal, or a digital signal. Examples of common actuators 202a - n include, but are not limited to, electromechanical devices, electromagnetic devices, pneumatic devices, hydraulic devices, and combinations thereof, such as motors, servos, solenoids, valves, pumps, pistons, and regulators.

[0053] The process parameter sensors 203a - n include any device or combination of devices that can be used to measure the value of a process parameter or provide one or more measured values, where the actual value of the desired process parameter can be determined from the values and measured values of the process parameters described above. Examples of suitable process parameter sensors 203a - n include temperature sensors (e.g., IR sensors, pyrometers, thermocouples), pressure sensors, force sensors, position sensors, acceleration sensors, rotational 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.

[0054] Controllers 204a - n determine the difference between the actual value and the target value of the process parameter, i.e., the error, and are operable devices or systems that command the corresponding actuators 202a - n or the processing system to change its output, e.g., the control parameters described herein. Examples of suitable controllers 204a - n include proportional - integral (PI) controllers, proportional - integral - derivative (PID) controllers, and / or logic controllers, e.g., programmable logic controllers (PLCs) programmed to execute software including logic applications. In some embodiments, such as when the control parameter includes the output of the processing system, the system controller 28, or other computing device operable to execute a software algorithm, can be used as the controllers 204a - n. In some embodiments, one or more functions of individual controllers among the controllers 204a - n or combinations of the controllers 204a - n can be executed by the system controller 28.

[0055] Control parameter sensors 205a - n include any sensors suitable for measuring the output of the actuators 202a - n or the processing system, and the said output is used to maintain the process parameter at the target value. Examples of suitable sensors that can be used as the control parameter sensors 205a - n include any one or combination of the exemplary sensors previously described with respect to the process parameter sensors 203a - n. In some embodiments, such as in the case of a control system where it is not feasible to measure the control parameter, the control parameter or an approximation thereof can be determined using the signals and / or commands supplied by the controllers 204a - n to the corresponding actuators 202a - n or the processing system.

[0056] In other embodiments, any one or combination of the individual subsystems described below may operate using an open-loop control system, i.e., a non-feedback system.

[0057] Here, the plurality of subsystems includes 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 liquid supply system 216 and an in-situ substrate monitoring system 222. The operations of the polishing station 21 and the carrier assembly 22 are adjusted by a system controller 28.

[0058] The platen assembly 212 includes a platen 228 and a rotation speed control system 201a. The control system 201a is connected to the platen 228 and includes a platen actuator 202a, such as a motor, used to rotate the platen 228 around the platen axis A, a process parameter sensor 203a used to measure the rotation speed and / or rotation posture of the platen 228, a controller 204a, and a control parameter sensor 205a.

[0059] Here, the controller 204a, together with the sensor 203a, adjusts control parameters such as the motor current supplied to the platen actuator 202a to maintain the rotational speed of the platen 228 at or near a target value. The control parameter sensor 205a is used to measure the control parameters, and time-series control parameter data is generated therefrom. In some embodiments, a change in the control parameter of the motor current is caused as overfill of material is removed from the field surface of the substrate 242 (FIG. 2B) being pressed, due to a change in the friction between the surfaces at the polishing interface. Thus, in some embodiments, a change in the motor current can be used to detect a desired polishing endpoint of the polishing process. In other embodiments, the motor current is used to detect fluctuations in the amount of slurry supplied to the surfaces of the polishing pad and the substrate 242 at any instant during polishing. For example, higher friction sensed by the motor current can be caused by a decrease in the slurry flow rate or a change in the composition of the slurry composition.

[0060] The platen assembly 212 further includes a platen temperature control system 201b, which includes a fluid source 202b, such as a water source or a refrigerant source, a sensor 203b for measuring the temperature of the platen 228, and a controller 204b. The temperature of the platen can be used to detect fluctuations in the amount of slurry supplied to the polishing pad, fluctuations in the characteristics of the polishing pad (e.g., the amount of glazing), or fluctuations in the downward force applied to the substrate 242 at any given moment during polishing. The platen 228 is formed of a cylindrical metal body having one or more channels 234 formed therein. The one or more channels 234 are fluidly coupled to the fluid source 202b. The controller 204b, together with the sensor 203b, is used to maintain the temperature of the platen 228 at a target value by adjusting the flow rate of the coolant through the one or more channels 234 from the fluid source 202b. In some embodiments, the control parameter for controlling the temperature of the polishing platen 228 includes the flow rate of the coolant measured by a flow meter, such as a control parameter sensor 205b. For some polishing processes, it may be desirable to heat the platen 228. In such embodiments, the fluid source 202b can include a heated fluid, such as heated water and / or steam, and the target value can include a temperature above a lower threshold. In some embodiments, the platen 228 is heated using a heater (not shown), such as a resistive heating element disposed within and / or embedded within the cylindrical metal body.

[0061] The carrier assembly 22 includes a substrate carrier 238, a carrier shaft 239, and control systems 201c, 201d. The substrate carrier 238 will be described below with reference to FIG. 2B. 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 connected to the carrier shaft 239 and is used to rotate the carrier shaft 239, and thus the substrate carrier 238 and the substrate 242 disposed therein, about the carrier axis B. The controller 204c, together with the sensor 205c, is used to maintain the rotational speed of the substrate carrier 238 at or near a target value by adjusting control parameters such as the motor current supplied to the first actuator 202c. The control parameter sensor 205c is used to measure the control parameters provided to the first actuator 202c.

[0062] The control system 201d includes a second actuator 202d connected 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, together with the sensor 203d, is used to maintain the sweep speed of the substrate carrier 238 at or near a target value by adjusting control parameters such as the motor current supplied to the second actuator 202d. The control parameter sensor 205d is used to measure the control parameters provided to the second actuator 202d.

[0063] As shown in FIG. 2B, the substrate carrier 238 includes a housing 240, a base assembly 243, a substrate downward force control system 201f, and a carrier load control system 201g. The housing 240 is movably and sealably coupled to the base assembly 243 and, together with the base assembly 243, defines a loading chamber 244. The base assembly 243 includes a carrier base 246, an annular retaining ring 247 coupled to the carrier base 246, and a flexible membrane 248 coupled to the carrier base 246 and defining a plurality of plenums 249.

[0064] During polishing of the substrate, the plurality of plenums 249 are pressurized, whereby the flexible membrane 248 applies a force to the non-active surface (back surface) of the underlying substrate 242. The plurality of plenums 249 facilitate adjustment of the distribution of the force applied to the entire back surface of the substrate 242 by allowing differences in their internal pressures. The pressures within the various plenums 249 and the pressure differences between the various plenums 249 are maintained by the control system 201f, which includes a plurality of actuators 202f (e.g., back 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 the target pressure within each plenum 249, enabling fine control of the distribution of the force exerted by the flexible membrane 248 on the substrate 242.

[0065] One or more controllers 204f, together with the plurality of sensors 203f, maintain the pressure within the plenums 249 at their target values by adjusting the respective control parameters for the corresponding actuators 202f. The various control parameter values are measured by the corresponding control parameter sensors 205f.

[0066] During processing, the loading chamber 244 is also pressurized to apply a downward force on the carrier base 246 and thus on the retaining ring 247 surrounding the substrate 242. The downward force on the retaining ring 247 prevents the substrate 242 from slipping off the substrate carrier 238 when the polishing pad 231 (FIG. 2A) moves under the substrate 242. The contact pressure between the retaining ring 247 and the polishing pad 231 is adjusted by changing the target downward force on the retaining ring 247. The target downward force is maintained by a control system 201g, which includes an actuator 202g, such as a back pressure regulator, a sensor 203g for measuring the pressure in the load 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, together with the sensor 203g, adjusts the control parameters provided to the actuator 202g to maintain the pressure in the loading chamber 244 at or near its target value. Here, the various components of the control systems 201g, h collectively form an upper pneumatic assembly, i.e., here the UPA 241, and the assembly may further include regulators, valves, and pumps (not shown here) used to supply pressurized gas, such as clean dry air (CDA) and / or vacuum, to the plurality of plenums 249 and the loading chamber 245. In other embodiments, electromechanical devices may be used to apply a downward force on one or both of the substrate 242 and the retaining ring 247.

[0067] The pad conditioner assembly 218 (FIG. 2A) is used to condition the polishing pad 231 by pressing the 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 201j - m for controlling various aspects of the pad conditioning process.

[0068] Typically, the conditioning disk 260 includes a conditioning surface with abrasive grains fixed, such as diamonds embedded in a metal alloy, and is used to polish and rejuvenate the surface of the polishing pad 231 and to remove polishing by - products or other debris from the surface. Since the abrasiveness of the conditioning disk 260 naturally dulls with use, the conditioning disk 260 is generally regarded as a consumable that requires periodic replacement.

[0069] The control systems 201j,k are used to maintain the rotational speed and sweep speed of the conditioning disk 260 at their respective target values while the conditioning disk 260 moves between the inner and outer radii of the polishing pad 231. The control system 201l is used to maintain the downward force applied to the conditioning disk 260 at a target value. In some embodiments, the 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 the polishing pad 231. In the same embodiment, 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 the system controller 28.

[0070] Here, the control system 201j includes a first actuator 202j coupled to one end of the conditioner arm 262 and used to rotate the conditioning disk 260 around the axis C, a sensor 203j for determining the rotational speed, and a controller 204j.

[0071] The control system 201k includes a second actuator 202k coupled to the distal end of the conditioner arm 262 remote from the first actuator 202j, one or more sensors 203k for determining the sweep speed and radial position of the conditioning disk 260 on the polishing pad, a controller 204k, and a control parameter sensor 205k. The control system 201g includes a third actuator 202l for applying a downward force to the conditioner arm 262, a sensor 203l for measuring the downward force, a controller 204l, and a control parameter sensor 205l. Here, the third actuator 202l is coupled to the distal end of the conditioner arm 262, in the vicinity of the second actuator 202k, and at a position distal from the conditioning disk 260. Each of the controllers 204j - l, together with the corresponding sensors 203j - l, maintains their respective processing parameters at or near their target values by adjusting the control parameters of the corresponding actuators 202j - l.

[0072] In some embodiments, the 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 downward force of the conditioning disk 260. Here, the control system 201m includes actuators 202j - l, a displacement sensor 203m connected to the conditioner 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 the radial direction. Here, the displacement sensor 203m is an inductive sensor that measures eddy currents to determine the distance between one 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 when the pad conditioning disk 260 is in contact with the platen 228 and the displacement when the pad conditioning disk 260 is in contact with the polishing pad 231 attached to the platen 228.

[0073] The system controller 28 compares the thickness profile of the polishing pad 231 determined using the displacement sensor 203m with the target thickness profile and determines the difference therebetween. Based on the difference, the system controller 28 generates a conditioning recipe, i.e., a set of conditioning parameters, that can be used to bring the actual thickness profile of the polishing pad 231 towards the target thickness profile. In some embodiments, the generated conditioning recipe changes the dwell time of the conditioning disk 260 and / or the downward force on the conditioning disk at one or more radial positions. The dwell time refers to the average time spent by the conditioning disk 260 at a certain radial position while the platen 228 rotates to move the polishing pad 231 under the conditioning disk 260 and the conditioning disk 260 is swept from the inner radius to the outer radius of the polishing pad 231.

[0074] A pad cooling assembly 220 (FIG. 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, heat is generated due to chemical activity and mechanical movement at the polishing interface, causing the temperatures of the substrate 242 and the polishing pad 231 to rise. Relatively high and / or unstable temperatures can consequently result in undesirable variations in the removal rate across the surface of the substrate 242 (intra-wafer non-uniformity), or undesirable variations in the removal rate between substrates (inter-wafer non-uniformity). In many damascene processes, relatively high temperatures degrade the quality of local planarization, resulting in a locally reduced flatness, erosion of the underlying layer, and / or dishing of features such as trenches, contacts, vias, and lines formed within the underlying layer. Thus, herein, the pad cooling assembly 220 is configured to cool the surface of the polishing pad 231 by supplying a non-reactive coolant, such as flakes of solid carbon dioxide (carbon dioxide snow), thereon. As the carbon dioxide snow sublimates (transitions from the solid phase to the gas phase without passing through an intermediate liquid phase), heat is removed from the surface of the polishing pad 231, and the temperature of the entire polishing process desirably decreases. Advantageously, sublimation of the carbon dioxide snow prevents undesirable dilution of the polishing fluid on the polishing pad. In other embodiments, the coolant includes a cryogenic fluid, i.e., a fluid having a boiling point below a threshold of 120 Kelvin, which is stored in liquid form and supplied to the surface of the polishing pad 231, such as, for example, liquid oxygen (LOX), liquid hydrogen, liquid nitrogen (LIN), liquid helium, liquid argon (LAR), liquid neon, liquid krypton, liquid xenon, liquid methane, or combinations thereof.

[0075] The pad cooling assembly 220 includes a coolant supply arm 275 disposed on the polishing pad 231, a plurality of nozzles 276 disposed on the coolant supply 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, such as an IR sensor or a pyrometer, are disposed to face the surface of the polishing pad 231 and are used to measure the temperature of the surface. In some embodiments, one or more sensors 203n include a thermal imaging system that generates a thermal image of the surface of the polishing pad 231.

[0076] The plurality of nozzles 276 are fluidly coupled to a coolant source 202n that provides vapor and solid carbon dioxide. The plurality of nozzles 276 generate carbon dioxide snow while the vaporous carbon dioxide expands therethrough and provide the carbon dioxide snow to the surface of the polishing pad 231. The controller 204n, together with the 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 for controlling the temperature of the surface of the polishing pad 231 includes a mass flow rate such that it is measured by the control parameter sensor 205n. In some embodiments, the supply of coolant to and / or the flow rate of coolant to individual nozzles 276 among the plurality of nozzles 276 are controlled individually. In the same embodiment, 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 a temperature distribution across the region.

[0077] Each control system 201a - n of the polishing system 20 described above uses a closed - loop feedback control method to maintain one or more polishing parameters at or near their respective target values by adjusting the associated control parameters. As described above, differences in control parameters between substrates (e.g., between wafers (WTW)), during polishing of individual substrates (e.g., within a wafer (WIW)), or both, are likely to indicate disturbances or changes 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 the target values using control systems 201a - l. 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. Thus, in some embodiments, an AI algorithm 110 using an unsupervised learning model can be used to identify and understand patterns in the control parameter data 120 and ultimately better understand the complex chemical and mechanical interactions between the surface, fluid, and abrasive at the polishing interface.

[0078] As described in the following method, 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, based thereon, adjust the composition of the polishing fluid at the polishing interface. Thus, the liquid supply system 216 herein is configured to stop flowing individual polishing fluid components to the surface of the polishing pad 231 and thus to the polishing interface, start flowing the individual polishing fluid components, and / or adjust the flow rate of the polishing fluid components based on instructions received from the system controller 28. In some embodiments, the instructions are in the form of a software algorithm, such as one or more machine - learning AI models 112 generated using the trained AI algorithm 110.

[0079] The liquid supply system 216 (FIG. 2C) is used to supply a polishing liquid containing individual liquid components to the surface of the polishing pad. The liquid supply system 216 includes a liquid distribution system 281, a liquid supply arm 282 including a plurality of nozzles 283, and an actuator 284 connected to the liquid supply arm 282. The liquid distribution system 281 is fluidly coupled to a plurality of polishing liquid sources 287a, 287b that supply the polishing liquid and / or liquid components. The actuator 284 is operable to swing the supply arm 282 over the polishing pad to position the plurality of nozzles 283 at desired radial dispensing positions over the polishing pad.

[0080] Here, the liquid distribution system 281 includes one or a combination of a plurality of valves 285a, pumps 285b, flow controllers 285c, and a polishing liquid mixing device 285d that can be used to control, measure, and supply the polishing liquid and / or individual polishing liquid components to the surface of the polishing pad 231. In some embodiments, the liquid distribution system 281 further includes one or more heaters (not shown) used to heat the individual polishing liquid and / or one or more individual polishing liquid components before and / or simultaneously with the supply of the liquid and / or the components to the surface of the polishing pad 231.

[0081] Here, the one or more polishing liquids and individual polishing components are supplied from the liquid distribution system 281 to the corresponding nozzles among the plurality of nozzles 283 using a plurality of supply lines 288 fluidly connected therebetween. In some embodiments, the liquid distribution system 281 is configured to supply one or more various polishing liquids and / or liquid components individually to various nozzles of the plurality of nozzles and / or to individually control the flow rates of the various polishing liquids or liquid components. Thus, the liquid distribution system 281 can be used to provide a desired dispersion of the polishing liquid and / or individual polishing liquid components dispensed on the surface of the polishing pad 231 and ultimately to provide a desired polishing liquid composition gradient across the surface of the pad 231.

[0082] In some embodiments, the liquid dispensing system 281 further includes a mixing device 285d that can be used to adjust the composition of the polishing liquid by adding one or more polishing liquid components thereto, and then the resulting mixture is supplied to the surface of the polishing pad 231. In some embodiments (not shown), the mixing station is disposed on the liquid supply arm 282.

[0083] Examples of individual polishing liquid components that can be individually supplied to the surface of the polishing pad 231 to a desired position on the polishing pad surface, and / or individual polishing liquid components that can be added to the polishing liquid using the mixing device 285d, include polishing liquids in which nanoscale silica particles or metal oxide particles are suspended, complexing agents, corrosion inhibitors, oxidizing agents, pH adjusters and / or buffers, polymer additives, passivating agents, accelerators, surfactants, or combinations thereof.

[0084] In some embodiments, the liquid supply system 216 further includes an optical sensor, such as a camera 299, disposed 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 a stream of digital images of an object at which it is positioned to look. The optical sensor can be used to determine the distribution of the polishing liquid and / or polishing liquid components across the surface of the polishing pad 231. In some embodiments, one or more of the individual polishing liquids and / or individual polishing liquid components include an optical marker, such as a conventional water-soluble dye or fluorescent substance. In the same embodiments, the images captured using the optical sensor can be analyzed to determine the distribution of the polishing liquid across the surface of the polishing pad 231 and / or to determine the compositional gradient of the individual polishing liquid components across the surface of the polishing pad 231.

[0085] In some embodiments, the distribution and / or composition of the polishing liquid on the surface of the polishing pad 231 is adjusted by starting, stopping, or changing the flow rate of one or more individual polishing liquid components to one or more of the individual nozzles 283 based on the analysis of an image. In some embodiments, the distribution and / or composition of the polishing liquid on the surface of the polishing pad 231 is continuously adjusted relative to a target distribution and / or composition using a closed-loop feedback control system 280. For example, here, the control system 280 includes a system controller 28, an optical sensor (e.g., camera 299) used to determine the distribution and / or composition of the polishing liquid on the surface of the polishing pad 231, and a liquid distribution system 281. In other embodiments, here, the control system 280 includes a system controller 28 and an electrochemical sensor (not shown) or a pH sensor (not shown) used to determine the composition of the polishing liquid on the surface of the polishing pad 231 and / or within the liquid distribution system 281. Based on the analysis of the image obtained from the optical sensor, the system controller 28 instructs the liquid distribution system 281 to change one or more control parameters regarding the supply of the polishing liquid and / or polishing liquid components to the surface of the polishing pad 231. For example, the control parameters may include starting, stopping, and / or changing the flow rate of the individual polishing liquid and / or polishing liquid components supplied to a collective plurality of nozzles 283 or to individual nozzles among the plurality of nozzles.

[0086] In some embodiments, one or more images captured using an optical sensor, such as a plurality of captured images in a time series, include process monitoring in-situ measurement data 122, and the process monitoring in-situ measurement data 122 can be used as training data 111 for the training method of the AI algorithm 110 provided herein.

[0087] The in-situ substrate monitoring system 222 (FIG. 2A) is used to monitor the thickness of a material layer on the substrate surface and / or to detect changes in the substrate surface as 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, and the light source and the optical sensor 289 are respectively arranged to direct light through a window (not shown) formed in the polishing pad 231 towards the substrate 242 and receive the light reflected therefrom. The controller 290 analyzes the reflected light and then determines one or more characteristics of the substrate surface. For example, the optical system 291 can be used to detect changes in the reflectivity of the substrate surface to determine, for example, the removal of a metal layer from the substrate surface, to detect the scattering of light reflected from the substrate surface to determine, for example, changes in the flatness of the substrate surface, and / or to use the technique of interference spectroscopy to determine the thickness of a transparent film, such as a dielectric layer, disposed on the substrate surface.

[0088] The eddy current monitoring system 292 includes an eddy current assembly 294, and the eddy current assembly 294 includes an eddy current generator and a sensor disposed on the surface of the platen 228. The eddy current monitoring system 292 uses the eddy current assembly 294 to induce and measure eddy currents in a conductive material layer (such as a metal layer) on the substrate, and further the eddy current monitoring system then determines the thickness of the conductive material layer. In some embodiments, the eddy current monitoring system 292 is used to determine a thickness profile across the radius of the substrate 242 as the substrate is swept thereover.

[0089] 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, such as the system controller 28, to trigger a change in the polishing conditions based on the thickness of the material layer and / or the removal of overfill from the underlying field surface.

[0090] The system controller 28 is used to direct the operation of the polishing system 20 and various components and subsystems of the polishing system 20. In some embodiments, individual controllers among the controllers 204a - n or one or more or all of the functions of the controllers 204a - n can be executed by the system controller 28. Here, the system controller 28, in combination with the AI training platform 30, is operable to execute the methods described herein. The system controller 28 includes a programmable central processing unit (CPU 295), and the central processing unit (CPU 295) is operable with a memory 296 (e.g., non - volatile memory) and support circuitry 297. For example, in some embodiments, the CPU 295 is one of any form of general - purpose computer processor used in industrial settings, such as a programmable logic controller (PLC) for controlling various polishing system components and sub - processors. The memory 296 connected to the CPU 295 is non - transitory and is typically one or more readily available memories, e.g., random access memory (RAM), read - only memory (ROM), floppy disk drive, hard disk, or other form of local or remote digital storage. The support circuitry 297 is conventionally connected to the CPU 295 and includes caches, clock circuits, input / output subsystems, power supplies, etc. connected to various components of the polishing system 20 to facilitate control of the substrate polishing process, and combinations thereof.

[0091] Here, the memory 296 is in the form of a computer-readable storage medium (e.g., non-volatile memory) containing instructions, which, when executed by the CPU 295, facilitate the operation of the polishing system 200. Exemplary computer-readable storage media include, but are not limited to, (i) non-writable storage media where information is permanently stored (e.g., read-only memory devices in a computer such as a CD-ROM disk readable by a CD-ROM drive, flash memory, ROM chips, or any type of solid-state non-volatile semiconductor memory), and (ii) writable storage media where modifiable information is stored (e.g., floppy disks in a disk drive or a hard disk drive, or any type of solid-state random access semiconductor memory). The instructions in the memory 296 are in the form of a program product (e.g., middleware application, device software application, etc.) such as a program that executes the method of the present disclosure. In some embodiments, the present disclosure can be realized as a program product stored on a non-transitory computer-readable storage medium for use in a computer system. Accordingly, the program of the program product defines the functions of the embodiments (including the methods described herein).

[0092] FIG. 3 is a diagram showing a method 300 for processing a substrate using the process improvement scheme 100 described in FIG. 1C. At least a part of the method 300 can be executed on the polishing system 20, and it is contemplated that any of the features and functions of the polishing system 20 can be incorporated, including the individual control systems used with the polishing system 20. The applications of the method 300 include, but are not limited to, planarization applications of bulk materials such as interlayer dielectric (ILD) applications, and damascene polishing applications such as shallow trench isolation (STI) applications and metal interconnect polishing applications.

[0093] In activity 302, method 300 includes polishing a substrate using a polishing system such as the polishing system 20 described above. Activity 302 includes a plurality of activities including activities 304 to 312.

[0094] In activity 304, method 300 includes flowing a polishing liquid composition (e.g., slurry) onto the surface of the polishing pad within the polishing system 20 according to a polishing recipe. The flow rate and / or amount of the polishing liquid composition provided at a predetermined radial position on the surface of the polishing pad 231 can be controlled using commands transmitted from the system controller 28 to the actuator 284 and / or the liquid distribution system 281.

[0095] In activity 306, method 300 includes pressing the substrate against the surface of the polishing pad in the presence of the polishing liquid. Here, the polishing recipe is determined by a plurality of polishing parameters including the substrate carrier rotation speed, the substrate carrier movement speed, the platen rotation speed, the downward force on the substrate, the downward force on the retaining ring, the flow rate of the polishing composition, the flow rate of the rinse liquid, and the pad adjustment parameters, and the target values corresponding thereto. The target values include a desired set value, a value exceeding a desired lower threshold value, a value below a desired upper threshold value, and a value between the desired lower and upper threshold values. Activity 306 includes pressurizing one or more of the plurality of plenums 249 to apply a force to the flexible film 248 within the substrate carrier against the non-active surface (back surface) of the substrate 242, and pressing the front surface against the polishing pad 231.

[0096] The target value can include a combination of a fixed value, e.g., a predetermined setpoint or threshold value, and a value determined by one or more software algorithms executed on a controller of the polishing system before, after, and / or simultaneously with the polishing process. For example, in some embodiments, the duration of a stage of a polishing sequence is determined using an endpoint algorithm executed on a controller of the polishing system. In some embodiments, one or more target values are determined by a trained AI algorithm 110, e.g., as part of an iterative continuous improvement process. In some embodiments, one or more target values are determined using a machine learning AI model 112 generated by a trained AI algorithm 110. In the same embodiment, the machine learning AI model 112 can include a software algorithm executed by a system controller 28 of the polishing system 20.

[0097] In a typical polishing process, a polishing recipe for one substrate includes a multi-stage polishing sequence, where one or more polishing parameter target values are changed for each stage of the sequence. In some embodiments, one or more stages of the multi-stage polishing sequence are executed at a first polishing station, after which the substrate is moved to a second polishing station and, optionally, moved again to a third polishing station to execute the remaining polishing sequence.

[0098] Examples of polishing parameters that can be used to determine a polishing recipe include platen rotation speed, platen temperature, substrate carrier rotation speed, substrate carrier sweep speed, substrate carrier sweep start position and substrate carrier sweep stop position (radially inner and outer positions on the polishing pad), downward force on the substrate (downward pressure applied to the back surface of the substrate), dispersion of the downward force across the substrate, downward force on the retaining ring (downward pressure applied to the retaining ring), difference between the downward force on the substrate and the downward force on the retaining ring, polishing pad surface temperature, uniformity and / or distribution of the polishing pad surface temperature, flow rate of the polishing liquid and / or individual polishing liquid components including start and stop of flowing the polishing liquid or polishing liquid components, temperature of the polishing liquid and / or individual polishing liquid components, polishing liquid composition before being supplied to the polishing pad (e.g., as an output from a polishing liquid mixing system), or polishing liquid composition before being supplied onto the surface of the polishing pad (e.g., as a result of dispensing of individual polishing liquid components), dispersion and / or composition gradient of the polishing liquid across the surface of the polishing pad, and duration (time), including but not limited to these.

[0099] Typically, a polishing recipe further includes processing parameters related to conditioning of the polishing pad before, after, and / or during the polishing process, herein referred to as pad conditioning parameters. Examples of pad conditioning parameters include rotation speed of the conditioning disk, downward force applied on the conditioning disk in contact with the polishing pad, dwell time of the conditioning disk over one or more portions of the polishing pad, and sweep speed of the conditioning disk across the surface of the polishing pad. As briefly described above, one or more pad conditioning parameters can be used with a position sensor of the conditioner assembly to determine the dwell time of the conditioning disk. In some embodiments, the pad adjustment parameters can also include the thickness of the polishing pad and / or the profile of the thickness of the polishing pad measured from a position near the center of the polishing pad to a radially outer position from the stop.

[0100] In activity 308, method 300 includes maintaining one or more polishing parameters at or near their target values by adjusting their respective control parameters. Here, the one or more polishing parameters are maintained at or near their target values using a closed-loop control system. Thus, in some embodiments, maintaining the polishing parameters at or near their target values includes the following. That is, (1) determining the difference between the actual value and the target value of the polishing parameter, (2) changing the 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 for the polishing parameter.

[0101] As used herein, a control parameter includes the output from an actuator and / or system that causes a corresponding change in the actual value of a polishing parameter. The control parameters for a particular control system are different from the polishing parameters of that system. However, as can be seen 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 function as control parameters in different control systems. For example, in an embodiment where a polishing pad thickness profile is used as a polishing parameter in a closed-loop system, one or more of the individual parameters of the conditioner's downward force, rotational speed, and dwell time can be used as control parameters and adjusted to provide the desired pad thickness profile.

[0102] In some embodiments, at least one of the processing parameters of activity 308 includes the pad surface temperature, and the corresponding control parameter includes the mass flow rate of a coolant, such as carbon dioxide snow, supplied to the surface of the polishing pad. In some embodiments, controller 204b, together with sensor 203b, is used to control the temperature of platen 228 to a target value by adjusting the flow rate of the coolant from fluid source 202b through one or more channels 234 within polishing platen 228. In some embodiments, the control parameter for controlling the temperature of polishing platen 228 includes the flow rate of the coolant measured by a flow meter, such as control parameter sensor 205b.

[0103] In activity 310, method 300 includes generating process system data 114. Here, process system data 114 includes a polishing recipe and time-series data of the first control parameter.

[0104] In activity 312, method 300 includes generating time-series in-situ result data using measurements obtained from an in-situ substrate monitoring system, such as in-situ substrate monitoring system 222 described herein, simultaneously with activities 304 - 310.

[0105] In some embodiments, in activity 312, a camera 299 (FIG. 2A) positioned to view the polishing surface (e.g., the upper surface) of the polishing pad 231 is monitored and analyzed by one or more software algorithms executed within the camera or within the system controller 28 to detect changes or variations in the optical properties of the polishing pad surface and / or the polishing liquid composition disposed on the surface, and provides a signal (e.g., a video signal stream) configured to do so. In one example, the camera is an IR camera configured to detect temperature gradients across the polishing pad surface and / or variations in temperature over time. The software algorithm can be used to detect in real time the temperature and / or temperature variations of the surface of the polishing pad and / or the polishing liquid composition disposed on the polishing pad. In this case, the camera 299 and / or the system in which the algorithm is executed is adapted to provide a signal containing time-series in-situ result data to the system controller 28 and / or is adapted to provide a signal containing training data to the artificial intelligence (AI) learning platform 30. Additionally, a flow rate sensing device and / or a polishing liquid composition sensing device (e.g., a pH sensor, a polishing particle concentration sensor, etc.) connected to components within the liquid dispensing system 281 can also be configured to transmit signals regarding the amount and / or composition of one or more polishing liquid compositions dispensed onto the surface of the polishing pad while the camera is monitoring the surface of the polishing pad. The time-series in-situ result data provided by the signals provided by the camera 299 and the flow rate sensing device and / or the polishing liquid composition detection device is analyzed by the artificial intelligence (AI) training platform 30A during subsequent activities to detect interactions between the various types of data described above, and then in subsequent activities, based on the data received over time, using the components found in the pad cooling assembly 220 and / or the composition of the polishing liquid composition, a change in the temperature of the polishing pad is caused.

[0106] In other examples, in activity 312, camera 299 (FIG. 2A) is configured to detect the condition of the polishing pad surface, for example, configured to detect whether the polishing pad surface has a desired amount of "pad conditioning". In this case, camera 299 is positioned and configured to detect the amount of roughness and / or asperity seen on the polishing surface of the polishing pad to determine the condition of the polishing pad surface. In some embodiments, camera 299 is replaced by a profilometer (surface profiler) or other device configured to detect and measure the degree of surface roughness. The surface roughness can be characterized by any value of Ra, Rrms, RSk, Rp. The surface roughness detected by the camera or similar device is the unevenness of the pad material on the polishing surface of the polishing pad, which may include unevenness with a size of up to about 10 to 50 microns. Additionally, a flow rate sensing device and / or a polishing liquid composition sensing device (e.g., a pH sensor, a polishing particle concentration sensor, etc.) can also be configured to supply signals regarding the amount and / or composition of the polishing liquid composition dispensed onto the surface of the polishing pad while the camera is monitoring the condition of the polishing pad surface. The in-situ result data in time series provided by the signals supplied by camera 299 or similar devices, and the flow rate sensing device and / or the polishing liquid composition detection device can be used by the artificial intelligence (AI) training platform 30A and the system controller 28 to cause a change in the temperature of the polishing pad and / or to cause a change in the composition of the polishing liquid composition based on the detected interactions between various types of data. Signals from these devices can be provided to the system controller 28, and / or signals containing training data can be provided to the artificial intelligence (AI) training platform 30.

[0107] In other embodiments, in activity 312, camera 299 (FIG. 2A) is configured to detect the coverage rate of the polishing liquid and / or the flow of the polishing liquid over one or more regions of the polishing pad surface while the polishing liquid is being dispensed onto the polishing pad. In this case, camera 299 is positioned and configured to detect the spread amount of the polishing liquid over the polishing surface of the polishing pad and to determine the state of one or more components within liquid distribution system 281, for example, detecting obstacles within one or more nozzles 283, detecting fluctuations in the output of the fluid pump, and / or detecting fluctuations in the position of liquid supply arm 282 relative to a desired position on the polishing pad surface and / or relative to the position of substrate carrier 238 on the polishing pad. The spread amount of the polishing liquid over the polishing surface of the polishing pad can be measured or determined by the coverage rate of the horizontal region of the polishing pad or the percentage of the field-of-view (FOV) of camera 299. In some cases, the camera is also configured to detect the temperature gradient across the polishing pad surface and / or the temperature change over time. Additionally, a flow rate sensing device and / or a polishing liquid composition sensing device (e.g., a pH sensor, a polishing particle concentration sensor) can also be configured to supply signals regarding the amount and / or composition of the polishing liquid composition being dispensed onto the polishing pad surface while camera 299 monitors the coverage rate of the polishing liquid and / or the flow of the polishing liquid over one or more regions of the polishing pad surface. The time-series in-situ result data provided by the signals from camera 299 and the flow rate sensing device and / or the polishing liquid composition detection device is used by artificial intelligence (AI) training platform 30A and system controller 28 to adjust the position of liquid supply arm 282 to adjust the position where the polishing liquid is supplied to the surface of the polishing pad in the next activity, to cause an increase in the flow of the polishing liquid exiting from one or more nozzles 283, to cause a change in the temperature of the polishing pad using pad cooling assembly 220, and / or to cause a change in the composition of the polishing liquid composition based on the detected interactions of various types of data during subsequent activities.

[0108] In activity 314, method 300 includes repeatedly performing activities 304-312 on a plurality of substrates to obtain a corresponding plurality of training data sets. Here, each training data set includes process system data and in-situ result data that can be correlated with the corresponding polished substrate.

[0109] In activity 316, method 300 includes receiving training data 111 that includes a plurality of training data sets in an artificial intelligence (AI) training platform 30. In some embodiments, the plurality of training data sets includes data regarding the dispense amount of the slurry composition during the polishing process, the concentration of the slurry composition dispensed during the polishing process, the temperature of the polishing pad after the slurry composition is dispensed during the polishing process, the characteristics of the polishing pad during a portion of the polishing process, and the time between pad adjustment processes received over time from one or more polishing systems 20 for detecting interactions between various data sets.

[0110] In one example, the plurality of training data sets collected and then analyzed by the artificial intelligence (AI) training platform 30 include the detection of one or more polishing liquid compositions, the detection of differences between the compositions of various polishing liquid compositions (e.g., the use of different abrasives or the amount of one type of abrasive), the detection of a specific type of substrate (e.g., an oxide polishing process or a metal polishing process), the detection of the polishing liquid flow rate, and / or the detected tendency of the temperature of the polishing pad between a plurality of polishing processes performed within one or more polishing systems 20. Based on the detected interactions between the data found in the training data sets, the tendency in the polishing process result data, such as the detection of dishing, wafer-to-wafer non-uniformity (WTWNU), planarization efficiency, and local flatness, is included.

[0111] In other embodiments, in activity 316, a plurality of training data sets collected by the artificial intelligence (AI) training platform 30 and then analyzed include trends in the optical properties of the surface of the polishing pad and / or the polishing liquid composition disposed on the polishing pad, and trends in the variation of the composition of one or more polishing liquid compositions, or differences between the compositions of various polishing liquid compositions on a particular type of substrate (e.g., an oxide polishing process or a metal polishing process) (e.g., the use of various abrasives or the amount of one type of abrasive).

[0112] In other embodiments, in activity 316, a plurality of training data sets collected by the artificial intelligence (AI) training platform 30 and then analyzed include the detected coverage rate of the polishing liquid across one or more regions of the polishing pad surface and / or the flow of the polishing liquid, the detected flow rate of the polishing liquid, and / or the detected trend in the temperature of the polishing pad during a plurality of polishing processes executed within one or more polishing systems 20.

[0113] In activity 318, method 300 includes generating a machine learning AI model 112 by training a machine learning AI algorithm 110 using training data 111. During activity 318, the artificial intelligence (AI) training platform 30 can use the machine learning AI model 112 to analyze data currently received from various sources.

[0114] In one example, in activity 318, the artificial intelligence (AI) training platform 30 can determine, based on the receipt of data generated by camera 299 and one or more polishing liquid composition detection devices and the use of machine learning AI model 112, that a detected upward trend in the surface temperature of the pad polishing pad can be caused by an increase in the concentration of abrasive particles in the polishing liquid composition or a decrease in the dispensed polishing liquid. Based on previous and current analyzes performed by the artificial intelligence (AI) training platform, the artificial intelligence (AI) training platform can, based on previously detected similar deviations occurring in one or more polishing systems 20, determine that a detected upward trend in the surface temperature of the polishing pad can be caused by an inappropriate mixing of a batch of polishing liquid composition or a drift of a dosing mechanism tasked with the control of the composition of the treatment solution.

[0115] In other embodiments, the artificial intelligence (AI) training platform 30 can determine, based on the receipt of data generated by camera 299 and one or more polishing liquid composition detection devices and the use of machine learning AI model 112, that a detected drift in the optical properties of the surface of the polishing pad can be caused by a decrease in the effectiveness of the pad adjustment disk (e.g., the disk is worn), based on a previously detected similar trend within one or more polishing systems 20.

[0116] As previously mentioned, in other examples, based on the receipt of data generated by camera 299 and other related sensors and the use of machine learning AI model 112, the artificial intelligence (AI) training platform 30 can, based on a previously detected similar trend within one or more polishing systems 20, determine that a detected change in the coverage rate of fluid across one or more regions of the surface of the polishing pad can be caused by an obstacle within one or more nozzles 283, a variation in the output of the fluid pump, and / or a variation in the position of the liquid supply arm 282 relative to a desired position on the surface of the polishing pad.

[0117] In activity 320, method 300 includes changing one or more of the plurality of polishing parameters of the processing recipe based on the analysis performed using the machine learning AI model 112 during activity 318. In one example, one or more of the polishing parameters changed based on the analysis performed by the AI algorithm may include adjusting the dispensing amount of the slurry composition during the current polishing process or a future polishing process, adjusting the concentration of the slurry composition dispensed during the current polishing process or a future polishing process, adjusting the temperature of the polishing pad after the slurry composition has been dispensed during the current polishing process or a future polishing process, and / or starting or stopping a pad conditioning process. One or more of the plurality of polishing parameters to be changed may also be implemented using the system controller 28 or the Fab production control system 40 respectively in one polishing system 20 or a plurality of polishing systems 20 based on the analysis performed by the AI algorithm.

[0118] In one example, when it is detected that the upward trend of the surface temperature of the polishing pad is caused by improper mixing of a batch of polishing liquid composition or drift of the dispensing mechanism of the polishing liquid components tasked with the task of controlling the composition of the processing solution, the artificial intelligence (AI) training platform 30 can use the system controller 28 or the user using the graphical user interface (GUI) connected to the system controller 28 to replace the polishing liquid composition or the dispensing mechanism, and / or adjust one or more processing variables of the polishing process recipe being executed on the current or future substrates processed within the polishing system 20.

[0119] In other instances, if it is detected that the detected drift in the optical properties of the surface of the polishing pad is caused by a decrease in the effectiveness of the pad conditioning disk, the artificial intelligence (AI) training platform 30 can instruct the system controller 28, or a user using a GUI connected to the system controller 28, to replace the pad conditioning disk, adjust the conditioning disk dwell time on a particular portion of the polishing pad, and / or adjust one or more process variables of the polishing process recipe being executed on the current or future substrates being processed within the polishing system 20.

[0120] As described above, in other instances, if a drift in the coverage rate and / or flow of the polishing liquid across one or more regions of the polishing pad surface is detected, the artificial intelligence (AI) training platform 30 can instruct the system controller 28 to adjust the position of the liquid supply arm 282 to adjust the position where the polishing liquid is supplied to the polishing pad surface, increase the flow of the polishing liquid exiting from one or more nozzles 283, cause a change in the temperature of the polishing pad using the pad cooling assembly 220, cause a change in the composition of the polishing liquid composition supplied from one or more nozzles 283, and / or adjust one or more process variables of the polishing process recipe being executed on the current or future substrates being processed within the polishing system 20.

[0121] In some embodiments, method 300 includes removing overfill of material from the surface of a substrate, as schematically shown in FIGS. 4A-4C. FIG. 4A shows a substrate 400 before a polishing process, on which one or more material layers 401, 402, such as an epitaxial (Si) layer and a silicon nitride (SiN) layer, are disposed. 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, such as an oxide layer (SiO2), is deposited on the patterned surface to fill the plurality of openings. The fill material disposed within the openings forms a plurality of features 403a, such as shallow trench isolation features, and an overfill layer 403b of the fill material layer 403, which is to be removed in a polishing process, remains.

[0122] FIG. 4B shows a partial removal of the overfill layer 403b using a polishing process, and FIG. 4C shows a complete removal of the overfill layer 403b and the desired planar features 403a remaining on the patterned surface.

[0123] Typically, changes in the surface of the substrate 400 as the overfill layer 403b of the fill material is removed can be detected in the time-series data generated using the in-situ substrate monitoring system 222. In some embodiments, such changes are detected using an endpoint algorithm running on a controller of the polishing system. When the overfill of material is removed from the field surface of the substrate in a STI or metal damascene process, the endpoint algorithm triggers a change in the polishing process. Unfortunately, such a reactive endpoint detection scheme can lead to over-polishing of the substrate surface, causing undesirable dishing and erosion of the features on the substrate surface.

[0124] In some embodiments, the AI algorithm 110 is trained to detect a functional relationship between the in-situ result data 124 and the processing system data 114, such as individual or combined time-series data for one or more control parameters. The functional relationship described above can be used by the trained AI algorithm 110 and / or the generated machine learning AI model 112 to predict the planned time horizon for the polishing endpoint before the overfill of the material begins to be removed from the substrate surface, rather than at the same time as the overfill of the material is removed from the substrate surface. Based on the predicted planned time, the composition of the polishing liquid on the surface of the polishing pad can be changed to provide better local planarization performance.

[0125] In some embodiments, changing one or more of the plurality of polishing parameters based on the machine learning AI model 112 in activity 318 includes changing the composition of the polishing liquid disposed on the surface of the polishing pad based on the functional relationship. In some embodiments, changing the composition of the polishing liquid includes starting, stopping, or changing the flow rate of the individual polishing liquid components supplied to the surface of the polishing pad.

[0126] In some embodiments, the training data 111 used to train the machine learning AI algorithm 110 further includes any portion or combination of the substrate tracking data 128, the equipment system data 130, and the electrical test data 132 as previously described in FIGS. 1B and 1C.

[0127] FIG. 5 is a diagram showing a method 500 for matching polishing performance between polishing systems.

[0128] In activity 502, method 500 includes receiving training data including a plurality of training data sets in an artificial intelligence (AI) training platform 30. Here, various data sets among the plurality of training data sets correspond to substrates polished using various combinations of polishing stations of the polishing system and substrate carrier assemblies. Each of the training data sets includes process system data related to each of the substrates polished using the polishing system.

[0129] Here, each of the training data sets includes process system data 114, and the process system data 114 includes polishing recipe data 118 and control parameter data 120. The polishing recipe data 118 includes a plurality of polishing parameters and corresponding target values. 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.

[0130] In 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 identify differences between various substrate carrier assemblies and / or between various polishing stations of the polishing system.

[0131] In activity 506, method 500 includes performing one or more corrective actions based on the identified differences.

[0132] In some embodiments, method 500 is used to identify differences between various substrate carrier assemblies and / or between various polishing stations across a plurality of polishing systems and perform one or more corrective actions based thereon.

[0133] Advantageously, the machine learning AI system and the AI algorithm training method described herein are used to better understand and utilize the combined capabilities of the devices and subsystems of an advanced CMP processing system, resulting in improved polishing results, a wider desired process window, and improved uniformity of polishing system processing.

[0134] The foregoing is directed to embodiments of the present disclosure, but other and further embodiments of the present disclosure may be devised without departing from the basic scope thereof, and the scope of the present disclosure is defined by the following claims.

Claims

1. A method of polishing a substrate, comprising: polishing the substrate using a polishing system, wherein a system controller of the polishing system:[[]] (a) flowing a polishing liquid onto a surface of a polishing pad according to a polishing recipe, the polishing recipe being defined by a plurality of polishing parameters and corresponding target values; (b) pressing the substrate against the surface of the polishing pad according to the polishing recipe; (c) adjusting a first control parameter among a plurality of control parameters for controlling operations of respective components of the polishing system, the first control parameter being a control parameter that causes a change in a value of a first polishing parameter among the plurality of polishing parameters, thereby maintaining the first polishing parameter at or near its target value; (d) generating process system data including time-series data of the polishing recipe and the first control parameter; and (e) simultaneously with (a) to (d), generating time-series in-situ result data using measurement values obtained from an in-situ substrate monitoring system causing the polishing system to perform operations including the above, and polishing the substrate; the system controller of the polishing system repeatedly causing the polishing system to perform (a) to (e) for a plurality of substrates to obtain a plurality of corresponding training data sets, each of the training data sets including the process system data and the in-situ result data for the polished substrate; receiving training data including the plurality of training data sets in an artificial intelligence (AI) training platform, at least a portion of the plurality of training data sets being received sequentially in a timely manner; based on analysis of the received training data performed by a machine learning AI algorithm, the system controller of the polishing system changing one or more of the target values of the plurality of polishing parameters; A method comprising the above.

2. The method according to claim 1, wherein the target value for each of the polishing parameters includes a desired set point, 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 according to claim 1, wherein the in-situ result data includes data obtained from a signal provided by a camera arranged to observe a change in temperature of at least a part of the surface of the polishing pad and configured to detect the change in temperature.

4. The first polishing parameter includes the temperature of the surface of the polishing pad. The method according to claim 3, wherein the first control parameter includes the flow rate of the coolant supplied to the surface of the polishing pad or the flow rate of the polishing liquid supplied to the surface of the polishing pad.

5. The in-situ result data is data obtained from a signal provided by a camera arranged to detect the position where the polishing liquid is dispensed onto the surface of the polishing pad, or includes data obtained from a signal provided by a camera arranged to detect the coverage amount of the polishing liquid dispensed from the polishing liquid supply nozzle onto the surface of the polishing pad. The method according to claim 1.

6. The first control parameter is the flow rate of the polishing liquid supplied to the surface of the polishing pad, or the position of the polishing liquid supply nozzle relative to the surface of the polishing pad The method according to claim 5 includes.

7. The in-situ result data is data obtained from a signal provided by a camera arranged to detect the temperature of at least a part of the surface of the polishing pad, and data obtained from a signal provided by a sensor configured to detect the composition of the polishing liquid The method according to claim 1 includes.

8. The first polishing parameter includes the temperature of the surface of the polishing pad. The method according to claim 7, wherein the first control parameter includes the flow rate of the coolant supplied to the surface of the polishing pad or the flow rate of the polishing liquid supplied to the surface of the polishing pad.

9. The in-situ result data includes data obtained from a signal provided by a camera arranged to detect the roughness of the surface of the polishing pad or arranged to detect the optical characteristics of the surface of the polishing pad. The first polishing parameter includes the pad conditioning parameter of the surface of the polishing pad. The method according to claim 1, wherein the first control parameter includes the rotational speed of the conditioning disk, the downward force applied to the conditioning disk in contact with the polishing pad, the residence time of the conditioning disk on one or more portions of the surface of the polishing pad, or the sweep speed of the conditioning disk across the surface of the polishing pad.

10. Maintaining the first polishing parameter at or near its target value includes i. determining the difference between the actual value of the first polishing parameter and its target value; ii. changing the first control parameter of the first control system based on the determined difference; iii. continuously repeating i. and ii. to provide closed-loop control for the first polishing parameter. The method according to claim 1, comprising:

11. The method according to claim 10, wherein the first polishing parameter includes the temperature of the surface of the polishing pad.

12. The polishing liquid includes a slurry composition, The method according to claim 11, wherein the first control parameter includes the flow rate or amount of the slurry composition supplied to the surface of the polishing pad.

13. The method according to claim 12, wherein the first control parameter includes the flow rate of the coolant supplied to the surface of the polishing pad.

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

15. The method according to claim 14, further comprising starting, stopping, or changing the flow rate of the individual polishing liquid components supplied to the surface of the polishing pad in response to the change in the target value of the composition of the polishing liquid.

16. the training data used to train the machine learning AI algorithm is substrate tracking data including the processing history of one or more of the plurality of substrates and / or information regarding devices formed on the one or more substrates, equipment system data including information generated using one or more equipment supply systems, the equipment system data including analysis information of polishing liquid supplied from a remote polishing liquid distribution system to the polishing system, and electrical test data including electrical test information generated from one or more of the plurality of substrates in an electrical test measurement process after polishing The method according to claim 1, further comprising one of, or a combination of these.

Images