Method and system for systematic error compensation across a fleet of measurement systems based on a trained error estimation model

Abstract

Described herein are methods and systems for compensating for systematic errors across a fleet of measurement systems based on a trained error estimation model to improve matching of measurements across the fleet. In one aspect, the error estimation model is a machine learning-based model trained based on a set of composite measurement matching signals. The composite measurement matching signals are generated based on measurement signals generated by each target measurement system and corresponding model-based measurement signals associated with each target measurement system and a reference measurement system. The training data set also includes an indication of whether each target system is operating within specifications, an indication of the values ​​of system model parameters for each target system, or both. In some embodiments, the composite measurement matching signals that drive the training of the error estimation model are weighted differently based, for example, on measurement sensitivity, measurement noise, or both.

Description

[Technical Field]

[0001] (CROSS-REFERENCE TO RELATED APPLICATIONS) This patent application claims priority under 35 U.S.C. § 119 of U.S. Provisional Patent Application No. 63 / 396,240, filed August 9, 2022, entitled "Matching Harmonics Generation of Nanosheet logic, DRAM, and 3D-Flash for Systematic Error Optimization between Ellipsometry Optical Metrology Systems by General Machine Learning," the subject matter of which is incorporated herein by reference in its entirety.

[0002] The described embodiments relate to metrology systems and methods, and more particularly to methods and systems for improving measurements of parameters that characterize semiconductor structures. [Background technology]

[0003] Semiconductor devices, such as logic and memory devices, are typically fabricated by a series of processing steps applied to a specimen. The various features and structural levels of the semiconductor device are formed by these processing steps. For example, lithography, among others, is one semiconductor fabrication process that involves creating patterns on a semiconductor wafer. Additional examples of semiconductor fabrication processes include, but are not limited to, chemical-mechanical polishing, etching, deposition, and ion implantation. Multiple semiconductor devices may be fabricated on a single semiconductor wafer and then separated into individual semiconductor devices.

[0004] Metrology processes are used at various steps during the semiconductor manufacturing process to detect defects on wafers and facilitate improved yield. Optical metrology techniques offer the potential for high-throughput measurements without the risk of sample destruction. Several optical metrology-based techniques, including implementations of scatterometry, reflectometry, and ellipsometry, and associated analysis algorithms, are commonly used to characterize the critical dimensions, film thickness, composition, and other parameters of nanoscale structures.

[0005] In general, the semiconductor industry strives to produce smaller and smaller devices with increasing structural complexity and material variety. Examples of devices exhibiting such complexity include gate-all-around (GAA) field-effect transistors (FETs), current dynamic random access memory (DRAM) structures, and current three-dimensional flash memory structures.

[0006] In one example, GaAs FETs fabricated using nanosheet fabrication techniques offer improved device performance and lower power consumption, but are difficult to fabricate due to their nanoscale size and complex geometry. Nanosheet structures contain several layers of material. The fabrication process for nanosheet structures begins with growing a superlattice of silicon and silicon germanium. These layers comprise the basic structure of the nanosheet. Maintaining control of the fabrication process by measuring the characteristics of each layer, such as film thickness, is crucial.

[0007] In another example, flash memory architectures are transitioning from two-dimensional floating gate architectures to fully three-dimensional geometries. In some examples, film stacks and etched structures are very deep (e.g., greater than 3 micrometers deep) and contain a large number of layers (e.g., greater than 400 layers). High aspect ratio structures create challenges for film and CD metrology. The ability to measure the critical dimensions that define the shape of holes and trenches in these structures is important to achieving desired performance levels and device yields. Metrology must be able to measure the CD of continuous profiles through deep channels to determine CD position variation and profile inflection point variation.

[0008] As devices (e.g., logic and memory devices) move toward smaller nanometer-scale dimensions, characterization becomes more challenging. Devices that incorporate complex three-dimensional shapes and materials with diverse physical properties contribute to the difficulty of characterization. In addition to accurate device characterization, measurement consistency across a range of measurement applications and across metrology systems serving the same measurement objective is also important. If measurement consistency is compromised within a manufacturing environment, consistency among processed semiconductor wafers will be lost, reducing yield to unacceptable levels. Matching measurement results across applications and across multiple systems (i.e., tool-to-tool matching) ensures that measurements on the same wafer for the same application produce the same results.

[0009] A typical calibration approach for a model-based measurement system consists of measuring several film / substrate systems with known thicknesses and dielectric functions. Regression is performed on the machine parameters until the parameter combination returns expected values ​​for thickness and / or dielectric function. In one example, a set of film wafers having a silicon dioxide layer on crystalline silicon over a range of thicknesses is measured, and regression is performed on the machine parameters until the machine returns the best match of thickness and / or refractive index for the given set of films. Another example is described in U.S. Patent No. 5,929,233, entitled "Methods and Systems for Determining a Critical Dimension and a Thin Film Characteristic of a Specimen," the contents of which are incorporated herein by reference as if fully set forth. This calibration procedure may be applied across a fleet of measurement systems using the same set of wafers. These wafers are sometimes referred to as transfer standards.

[0010] Calibration of a fleet of measurement systems using transfer standards suffers from several drawbacks. To obtain highly accurate results, calibration experiments involving reference wafers must be performed in a carefully controlled environment that matches the environmental conditions at the same location when the reference wafers were originally characterized. This can be difficult to achieve in a manufacturing environment and can lead to inconsistencies between measurement systems. In addition, expensive reference wafer sets must be maintained in the manufacturing environment. The risk of wafer breakage or degradation can jeopardize the integrity of the calibration process, and this risk is increased when the metrology systems being calibrated are in different manufacturing facilities.

[0011] Machine parameters are often calibrated based on thin-film measurements because thin-film systems (e.g., silicon dioxide on crystalline silicon) can be fabricated with known optical constants, clean interfaces, and low surface roughness, which allow measurements on wafer properties with repeatability approaching the sensitivity of the measurement system being calibrated. However, the accuracy of metrology systems calibrated based on reference wafers is typically limited to wafers with properties that closely match those of the reference wafer. Therefore, the effectiveness of calibrations based on thin-film measurements may be limited in different measurement applications.

[0012] In another approach, system parameter calibration to achieve measurement consistency over time and across different measurement applications is improved by matching measurement spectra across a fleet of metrology systems rather than specimen parameter values. System parameter values ​​are optimized to minimize differences between measured spectra produced by a reference system and a target system for measurements of the same metrology target. The updated system parameter values ​​are used in subsequent measurement analyses performed by the target metrology system (e.g., CD metrology, thin film metrology, CD matching applications, etc.). Further description of this approach is provided in U.S. Patent Nos. 5,929,999 and 5,929,999, assigned to KLA-Tencor Corporation, the contents of each of which are incorporated herein by reference in their entirety.

[0013] In yet another approach, system parameter calibration to achieve measurement consistency over time and across different measurement applications is improved by matching spectroscopic errors across a fleet of metrology systems. System parameter values ​​of a target metrology system are calibrated based on matching spectroscopic errors with a reference metrology system. In this approach, the spectroscopic error is the difference between the measured spectrum and the modeled spectral response of the specimen being measured. One or more system parameters of the target metrology system are calibrated to minimize the difference between the spectroscopic error associated with measurements of one or more metrology targets measured by the reference metrology system and the spectroscopic error associated with measurements of the same metrology targets measured by the target metrology system. Further description of this approach is provided in U.S. Patent No. 6,279,999, assigned to KLA-Tencor Corporation, the contents of which are incorporated herein by reference in their entirety.

[0014] Matching spectroscopic errors across a fleet of metrology systems relative to a reference metrology system, e.g., a "best of breed" tool, poses several limitations. For example, the fleet of all target metrology systems must be recalibrated to maintain systematic errors within desired tolerances when the reference metrology system undergoes hardware changes or maintenance work. In another example, spectroscopic error data loses important signal information specific to the measurement metrology target because the data is based on the difference between the measured and theoretical signals, rather than the measured signals themselves. In another example, system parameter optimization to minimize spectroscopic errors is complex and computationally intensive. [Prior art documents] [Patent documents]

[0015] [Patent Document 1] US Patent Application Publication No. 2004 / 0073398 [Patent Document 2] U.S. Patent No. 9,857,291 [Patent Document 3] U.S. Patent No. 10,605,722 [Patent Document 4] U.S. Patent No. 10,006,865 Summary of the Invention [Problem to be solved by the invention]

[0016] Matching and maintaining tool measurement consistency between tools over time, over maintenance cycles, and across a wide range of measurement applications is a central challenge in developing metrology systems that meet semiconductor industry customer requirements. Process and yield management in both R&D and manufacturing environments requires tool-to-tool consistency of measurement results on the order of measurement repeatability. Therefore, methods and systems for improving tool-to-tool matching and consistent measurement performance across a wide range of measurement applications are desirable. [Means for solving the problem]

[0017] Described herein are methods and systems for compensating for systematic errors across a fleet of metrology systems based on trained error estimation models. Compensating for systematic errors improves matching of measurement results across a fleet of metrology systems across a range of metrology targets and measurement applications.

[0018] In one aspect, the error estimation model is a machine learning-based model trained based on a set of composite measurement matching signals. The composite measurement matching signals are generated based on measurement signals generated by each target system and corresponding model-based measurement signals associated with each target measurement system and a reference measurement system. The trained error estimation model enables rapid systematic error monitoring and optimization across multiple metrology tools. The trained error estimation model enables optimization of system parameters among a fleet of metrology tools without simulating measurement system model parameters. Thus, computational effort is significantly reduced.

[0019] The composite measurement matching signal associated with each metrology tool incorporates measurement information specific to each target metrology tool, reference metrology tool, and metrology target with reduced complexity compared to conventional tool matching approaches. For purposes of error monitoring across the metrology tool fleet, an indication of whether each target system is operating within specifications is included as part of the training data set. For purposes of optimizing system parameters across the metrology tool fleet, system model parameters for each target system used to provide measurement data are included as part of the training data set.

[0020] In further aspects, the optimized system parameters are then used for further measurement analysis. In some examples, critical dimension (CD) measurements are performed by the target measurement system using the optimized subset of system parameters. For example, structural parameters of the metrology target may be estimated based on regression of a target system measurement model updated with spectroscopic data associated with measurements of the metrology target.

[0021] In another further aspect, the composite measurement matching signals that drive the training of the error estimation model may be weighted differently. In one example, the relative weighting is based on the measurement sensitivity for any of the multiple measurement sites, multiple measurement samples, multiple illumination wavelengths, and multiple measurement subsystems. In this way, particular measurement sites, samples, subsystems, or illumination wavelengths that have particularly high measurement sensitivity are emphasized. In another example, the relative weighting is based on the measurement noise associated with any of the multiple measurement sites, multiple measurement samples, multiple illumination wavelengths, and multiple measurement subsystems. In this way, particular measurement sites, samples, subsystems, or illumination wavelengths that have particularly high measurement noise are not emphasized.

[0022] The foregoing is a summary and, as such, necessarily contains simplifications, generalizations, and omissions of detail. Consequently, those skilled in the art will appreciate that this summary is merely illustrative and is not intended to be in any way limiting. Other aspects, inventive features, and advantages of the devices and / or processes described herein will become apparent in the non-limiting detailed description set forth herein. [Brief explanation of the drawings]

[0023] [Figure 1] 1 is a simplified diagram illustrating a metrology system 100 operable in accordance with the method for monitoring and correcting systematic errors across a fleet of metrology systems as described herein. [Figure 2] FIG. 1 is a diagram illustrating an error estimation model-based flock matching engine in one embodiment. [Figure 3] FIG. 1 is a diagram illustrating an error estimation model training engine in one embodiment. [Figure 4A] 10 is a plot showing spectroscopic measurement signal matching between a target measurement system and a reference measurement system over a range of illumination wavelengths, before and after calibrating the target system parameters using a trained error estimation model. [Figure 4B] 10 is a plot showing spectroscopic measurement signal matching between a target measurement system and a reference measurement system over a range of illumination wavelengths, before and after calibrating the target system parameters using a trained error estimation model. [Figure 5] 1 is a flowchart illustrating a method for systematic error monitoring across a fleet of metrology systems as described herein. DETAILED DESCRIPTION OF THE INVENTION

[0024] Reference will now be made in detail to certain background and embodiments of the present invention, examples of which are illustrated in the accompanying drawings.

[0025] A method and system for compensating for systematic errors across a fleet of metrology systems is described herein based on a trained error estimation model. The trained error estimation model enables rapid systematic error monitoring and optimization across a large number of metrology tools. Compensation for systematic errors improves matching of measurement results across a fleet of metrology systems across a range of metrology targets and measurement applications. The trained error estimation model enables optimization of system parameters within a fleet of metrology tools without simulating measurement system model parameters. Thus, computational effort is significantly reduced.

[0026] In one aspect, the error estimation model is a machine learning-based model trained based on a set of composite measurement matching signals. The composite measurement matching signals are generated based on the measurement signals generated by each target system and corresponding model-based measurement signals associated with each target measurement system and a reference measurement system. For each target measurement system, the composite measurement matching signal associated with a particular measurement target is a mathematical function of: 1) the actual measurement signals generated based on measurements of the measurement target by the target measurement system; 2) the model-based measurement signals predicted by a model of measurements of the measurement target by the target measurement system; and 3) the model-based measurement signals predicted by a model of measurements of the measurement target by the reference measurement system.

[0027] The composite measurement matching signal associated with each metrology tool incorporates measurement information specific to each target metrology tool, reference metrology tool, and metrology target with reduced complexity compared to conventional tool matching approaches. For purposes of error monitoring across the metrology tool fleet, an indication of whether each target system is operating within specifications is included as part of the training data set. For system parameter optimization across the metrology tool fleet, system model parameters for each target system used to provide measurement data are included as part of the training data set.

[0028] Metrology targets used to train the machine learning-based error estimation model include, but are not limited to, nanosheet logic structures, DRAM structures, 3D flash memory structures, etc. A group of metrology systems matched using the trained machine learning-based error estimation model is used to measure structural and material properties (e.g., material composition, dimensional properties of structures and films, etc.) associated with different semiconductor fabrication processes.

[0029] FIG. 1 illustrates a metrology system 100 for measuring semiconductor wafer properties according to exemplary methods presented herein. As shown in FIG. 1, the system 100 may be used to perform spectroscopic ellipsometry measurements of one or more structures 114 on a semiconductor wafer 112 disposed on a wafer positioning system 110. In this embodiment, the system 100 may include a spectroscopic ellipsometer (SE) 101 equipped with an illuminator 102 and a spectrometer 104. The illuminator 102 of the system 100 is configured to generate and direct illumination in a selected wavelength range (e.g., 150-850 nm, 190-850 nm, 240-850 nm, etc.) toward the structures 114 disposed on the surface of the semiconductor wafer 112. The spectrometer 104 is then configured to receive illumination reflected from the surface of the semiconductor wafer 112. It is further noted that light emitted from the illuminator 102 is polarized using a polarization state generator 107 to generate a polarized illumination beam 106. Radiation reflected by structures 114 disposed on wafer 112 is passed through polarization state analyzer 109 to spectrometer 104. Radiation received by spectrometer 104 in collected beam 108 is analyzed with respect to polarization state, enabling the spectrometer to perform spectral analysis of the radiation passed by the analyzer. These spectral spectra 111 are passed to computing system 116 for analysis of structures 114.

[0030] In a further embodiment, metrology system 100 is a target measurement system 100 that may include one or more computing systems 116 used to perform calibration of system parameter values ​​of target measurement system 100 according to the methods described herein. One or more computing systems 116 may be communicatively coupled to spectrometer 104. In one aspect, one or more computing systems 116 are configured to receive measurement data 111 associated with measurements of structures 114 of specimen 112. In one example, measurement data 111 includes an indication of the spectroscopic response of the specimen measured by target measurement system 100 based on one or more sampling processes from spectrometer 104.

[0031] Additionally, in some embodiments, the one or more computing systems 116 are further configured to receive model-based measurement data 113 from the reference measurement source 103. In one example, the model-based measurement data 113 includes spectroscopy associated with measurements of the structure 114, simulated by the reference metrology system. In some examples, the set of parameter values ​​is stored in a carrier medium 118 and retrieved by the computing system 116.

[0032] It should be appreciated that various elements described throughout this disclosure may be performed by a single computer system 116 or, alternatively, by multiple computer systems 116. Furthermore, different subsystems of system 100, such as spectroscopic ellipsometer 101, may include computer systems suitable for performing at least some of the steps described herein. As such, the foregoing description should be construed as merely exemplary and not limiting on the present invention. Furthermore, one or more computing systems 116 may be configured to perform any other step(s) of any of the method embodiments described herein. Furthermore, some or all of one or more computing systems 116 may be located remotely from the site of wafer measurement. For example, elements of computing system 116 configured to perform any of the methods described herein may be located at a separate facility located remotely from the site where wafers are measured.

[0033] In this regard, the acquisition of the spectroscopy and the subsequent analysis of the spectroscopy data are not required to be performed simultaneously or in close spatial proximity. For example, the spectroscopy data may be stored in a memory for later analysis. In another example, the spectroscopy results may be acquired and transmitted to a remotely located computing system for analysis.

[0034] Additionally, computer system 116 may be communicatively coupled to spectrometer 104, illumination subsystem 102 of ellipsometer 101, or reference measurement source 103 (e.g., external memory, reference metrology system, etc.) in any manner known in the art. For example, one or more computing systems 116 may be coupled to the computing system of spectrometer 104 and the computing system of illumination subsystem 102 of ellipsometer 101. In another example, spectrometer 104 and illuminator 102 may be controlled by a single computer system. In this manner, computer systems 116 of system 100 may be coupled to a single ellipsometer computer system.

[0035] The computer system 116 of the system 100 may be configured to receive and / or acquire data or information from subsystems of the system (e.g., spectrometer 104, illuminator 102, etc.) via a transmission medium, which may include wired and / or wireless portions. In this manner, the transmission medium may serve as a data link between the computer system 116 and other subsystems of the system 100. Additionally, the computing system 116 may be configured to receive measurement data via a storage medium (i.e., memory). For example, spectroscopic results obtained using the spectrometer of the ellipsometer 101 may be stored in a permanent or semi-permanent memory device (not shown). In this regard, spectroscopic results may be imported from an external system.

[0036] Additionally, computer system 116 may send data to external systems via a transmission medium. Computer system 116 of system 100 may be configured to receive and / or acquire data or information from other systems (e.g., test results from an inspection system or measurement results from a metrology system) via a transmission medium that may include wired and / or wireless portions. In this manner, the transmission medium may serve as a data link between computer system 116 and other subsystems of system 100. Additionally, computer system 116 may send data to external systems via a transmission medium.

[0037] Computing system 116 may include, but is not limited to, a personal computer system, a cloud-based computer system, a mainframe computer system, a workstation, an image computer, a parallel processor, or any other device known in the art. In general, the term "computing system" may be broadly defined to encompass any device having one or more processors that executes instructions from a memory medium.

[0038] Program instructions 120 implementing the methods as described herein may be transmitted over or stored on carrier medium 118. The carrier medium may be a transmission medium such as a wire, cable, or a wireless transmission link. The carrier medium may also include a computer-readable medium such as a read-only memory, a random-access memory, a solid-state memory, a magnetic or optical disk, or a magnetic tape.

[0039] 1 may be further configured as described herein. In addition, system 100 may be configured to perform any other block(s) of any of the method embodiment(s) described herein.

[0040] As shown in FIG. 1 , a beam of broadband radiation from illuminator 102 is linearly polarized in polarization state generator 107, and the linearly polarized beam then impinges on sample 112. After reflection from sample 112, the beam propagates with the changed polarization state toward polarization state analyzer 109. In some examples, the reflected beam has an elliptically polarized light. The reflected beam propagates through polarization state analyzer 109 to spectrometer 104. In spectrometer 104, beam components with different wavelengths are refracted (e.g., in a prism spectrometer) or diffracted (e.g., in a grating spectrometer) in different directions to different detectors. The detectors may be linear arrays of photodiodes, each measuring radiation within a different wavelength range.

[0041] In one example, computing system 116 is programmed with software to receive measurement data (e.g., raw measurement data) from each detector and process the data it receives in an appropriate manner. The measured spectroscopic response for the specimen may be determined by analyzing the change in polarization of radiation reflected from the sample in response to incident radiation having a known polarization state in any number of ways known in the art.

[0042] Any of the polarization state generator 107 and the polarization state analyzer 109 may be configured to rotate about their optical axes during measurement operations. In some examples, the computing system 116 is programmed to generate control signals for controlling the angular orientation of the polarization state generator 107 and / or the polarization state analyzer 109, or other elements of the system 100 (e.g., the wafer positioning system 110 on which the specimen 112 rests). The computing system 116 may also receive data indicative of the angular orientation of the polarization state analyzer 109 from an analyzer position sensor associated with the polarization state analyzer 109. Similarly, the computing system 116 may also receive data indicative of the angular orientation of the polarization state generator 107 from a polarizer position sensor associated with the polarization state generator 107. The computing system 116 may be programmed with software for processing such orientation data in an appropriate manner.

[0043] In one embodiment, the polarization state generator 107 is a linear polarizer that is controlled to rotate at a constant speed, and the polarization state analyzer is a non-rotating linear polarizer ("analyzer"). The signal received at each detector of the spectrometer 104 (i.e., raw measurement data) is a time-varying intensity given by:

number

number

number

[0044] In general, the spectral response of a sample to a measurement is a function of the spectrometer data S and the system parameter values ​​P, as shown in Equations (4) and (5). sys1 is calculated by the measurement system based on a subset of

number

[0045] System parameter value P sys1 are the system parameters necessary to determine the spectral response of the specimen to measurements performed by the metrology system.

[0046] For the embodiment described with reference to FIG. 1, the system parameter P sys1 The subset of includes the machine parameters in equations (1) to (3). meas and β meas The value of is determined based on measurements made by metrology system 100 for a particular specimen and a subset of system parameter values ​​as described by equations (1)-(3).

[0047] In general, ellipsometry is an indirect method for measuring the physical properties of a specimen of an inspection target. In most cases, measurements (e.g., α meas and β meas ) cannot be used to directly determine the physical properties of the specimen. The nominal measurement process is a process where the measured values ​​(e.g., αmeas and β meas ), which characterizes the specimen's interaction with the measurement system. The measurement model includes structural (e.g., film thickness, critical dimensions, etc.) and mechanical (e.g., wavelength, angle of incidence, polarization angle, etc.) parameterizations. As shown in equations (6) and (7), the measurement model estimates the mechanical (P machine ) and specimens (P specimen ) and related parameters.

number

[0048] Machine parameters are parameters used to characterize the metrology tool (e.g., the ellipsometer 101) and may include some or all of the subset of system parameters described with reference to Equations (4) and (5). Exemplary machine parameters include the angle of incidence (AOI), analyzer angle (AO), polarizer angle (PO), illumination wavelength, numerical aperture (NA), etc. Specimen parameters are parameters used to characterize the specimen (e.g., the specimen 112 including the structure 114). For a thin film specimen, examples of specimen parameters include the refractive index, dielectric function tensor, nominal layer thicknesses of all layers, layer order, etc. For measurement purposes, the machine parameters are treated as known, and the fixed and specimen parameters are treated as unknown, floating parameters. The floating parameters are determined by an iterative process (e.g., regression) that produces a best fit between theoretical predictions and experimental data. The unknown specimen parameter P specimen is the ratio between the model output and experimental measurements (e.g., α meas and β meas ) and the model output values ​​(e.g., α model and β model ) is calculated.

[0049] In model-based measurement applications such as spectroscopic ellipsometry, a regression process (e.g., ordinary least squares regression) is used to identify specimen parameter values ​​that minimize the difference between the model output values ​​and experimental measurements for a fixed set of machine parameter values.

[0050] Measurement consistency across multiple metrology applications and across multiple tools relies on a properly calibrated set of machine parameter values ​​for each measurement system. The estimates of one or more parameters of interest estimated by each of the target metrology tools must match, within a desired tolerance, the estimates of one or more parameters of interest measured by a reference metrology tool. This is called tool-to-tool matching. System model parameters are optimized for each target metrology tool to achieve tool-to-tool matching.

[0051] In one embodiment, computing system 116 is configured as an error estimation model-based swarm matching engine 150, as shown in Figure 2. Error estimation model-based swarm matching engine 150 uses a trained error estimation model that enables monitoring and optimization of systematic errors across a fleet of metrology tools.

[0052] 2, the error estimation model-based swarm matching engine 150 includes a composite measurement matching signal module 151 and a trained error estimation module 152. The composite measurement matching signal module 151 receives 1) measurement signals associated with measurements of one or more measurement targets by a target measurement system; MEAS S153; and 2) a model-based measurement signal associated with a model-based simulation of the measurement signal associated with measurements of one or more measurement targets by the target measurement system. T-MOD S154; and 3) a model-based measurement signal associated with a model-based simulation of the measurement signal associated with measurements of one or more measurement targets by the reference measurement system. R-MOD S155. The composite measurement matching signal module 151 receives: MEASS153, T-MOD S154, and R-MOD S155. Generally, the composite measurement matching signal module 151 includes: MEAS S153, T-MOD S154 and R-MOD Implementing the mathematical function of S155 generates CMMS 156. In one embodiment, the mathematical function is as shown by equation (8): MEAS S153 and T-MOD S154 and R-MOD S155 and the difference between S155 and S156.

number

[0053] However, in general, the composite measurement and matching signal module 151 MEAS S153, T-MOD S154, and R-MOD Any suitable mathematical function of S155 may be implemented to generate CMMS156.

[0054] 2, the CMMS 156 is communicated to a trained error estimator module 152. The trained error estimator module 152 generates an indicator of the matching condition of the target measurement system, COND 157, an indicator of the target system parameter value, P-SYS 158, or both. COND 157 is a measure of the matching condition of the measurement signal, MEAS S153 is a signal that indicates whether the target measurement system utilized to generate S153 matches the reference measurement system within an acceptable tolerance. In some examples, status signal COND157 is a binary signal that indicates whether the target system matches the reference measurement system within a tolerance. In some other examples, status signal COND157 is a numeric value that indicates not only whether the target measurement system matches the reference measurement system, but also the degree to which the target measurement system matches the reference measurement system.

[0055] P-SYS 158 is a signal indicating target system parameter values ​​for bringing the match between the target measurement system and the reference measurement system within a tolerance range. In this example, the trained error estimation model suggests system parameter values ​​for the target measurement system that reduce the systematic error of the target measurement system and bring the target measurement system into match with the reference measurement system within an acceptable tolerance range.

[0056] In a further aspect, the computing system 116 is configured as an error estimation model training engine 160, as shown in FIG.

[0057] 3, the error estimation model training engine 160 includes a composite measurement matching signal module 151, a machine learning module 162, and an error estimation module 163, as described with reference to FIG. 2. The composite measurement matching signal module 151 receives 1) Design of Experiment (DOE) measurement signals associated with measurements of one or more metrology targets by multiple target measurement systems; MEAS S DOE 164, and 2) a DOE model-based target measurement signal associated with a model-based simulation of the measurement signal associated with each of the measurements of the one or more measurement targets by the target measurement system. T-MOD S DOE 165, and 3) a DOE model-based reference measurement signal associated with a model-based simulation of the measurement signal associated with measurements of one or more measurement targets by the reference measurement system. R-MOD S DOE 166. The composite measurement matching signal module 151 receives a DOE composite measurement matching signal associated with each corresponding set of DOE measurement signals, DOE model-based target measurement signals, and DOE model-based reference measurement signals. 1…M CMMS DOE In one example, the set of DOE composite measurement matching signals includes M different DOE measurements, where M is any positive integer value. 1…M CMMS DOE167 is communicated to machine learning module 162. Machine learning module 162 generates a current status signal indicating whether the target measurement system associated with each of the M different DOE measurements matches the reference measurement system within an acceptable tolerance range. 1…M COND * 168. The machine learning module 162 also generates a current system parameter signal 168 indicative of the current target system parameter value of the target measurement system associated with each of the M DOE measurements. 1…M P-SYS * The error estimation module 163 receives the current state signal and the current system parameter signal generated by the machine learning module 162. In addition, the error estimation module 163 generates a DOE state signal 169 associated with each of the M DOE measurements. 1…M COND DOE 170, and DOE system parameter signals 1…M P-SYS DOE 171. The DOE state signals and DOE system parameter signals are received from a reference signal source 161, e.g., a database of DOE measurement data. The DOE state signals indicate the actual match between the target system and the reference system associated with each of the M DOE measurements. The DOE system parameter signals indicate the actual target system parameter values ​​associated with each of the M DOE measurements. The error estimation module 163 minimizes the difference between the DOE state signals and the current state signals and the difference between the DOE system parameter signals and the current system parameter signals by generating updated values ​​of the weighting parameters 169 of the error estimation model 162 being trained. In the next iteration of model training, new current state signals and new current system parameter signals are generated by the machine learning module 162 based on the values ​​of the weighting parameters 169 generated in the previous iteration. The training process continues until the difference between the DOE state signals and the current state signals and the difference between the DOE system parameter signals and the current system parameter signals become acceptably small. At this point, the trained error estimation model 172 is stored in a memory, for example, memory 132 .

[0058] Although the error estimation model training engine 160 is described for training an error estimation model suitable for evaluating group matching and for compensating for systematic errors by determining system parameter values, in general, the error estimation model training engine 160 may be configured to evaluate only group matching.

[0059] 4A-B show the spectroscopic measurement signal matching between the target and reference measurement systems over a range of illumination wavelengths, before and after calibrating the target system parameters using the trained error estimation model.

[0060] 4A shows plot 180 illustrating spectroscopic measurement signals associated with measurements of a measurement target as measured by a target measurement system and a reference measurement system over a range of illumination wavelengths before calibration of the target system parameters. Plot line 181 shows the spectroscopic measurement signal associated with measurements of the measurement target as measured by the reference measurement system. Plot line 182 shows the spectroscopic measurement signal associated with measurements of the measurement target as measured by the target measurement system before calibration of the target system parameters.

[0061] 4B shows plot 185 illustrating spectroscopic measurement signals associated with measurements of a measurement target as measured by the target measurement system and the reference measurement system over a range of illumination wavelengths after calibrating the target system parameters as described herein. Plot line 186 shows the spectroscopic measurement signals associated with measurements of the measurement target as measured by the reference measurement system. Plot line 187 shows the spectroscopic measurement signals associated with measurements of the measurement target as measured by the target measurement system after calibrating the target system parameters. As shown in FIGS. 4A-B, the agreement between the spectroscopic signals measured by the reference measurement system and the target measurement system is much closer after calibrating the target system parameters as described herein.

[0062] As shown in Figures 4A-B, calibration of system parameter values ​​based on a trained error estimation model significantly improves tool-to-tool matching and measurement stability across a wide range of measurement applications.

[0063] The application of the aforementioned method is not limited to a particular spectroscopic signal, i.e., the method is applicable regardless of the spectroscopic signal under consideration, e.g., cos(Δ), tan(Ψ), α and β harmonic signals, Mueller matrix coefficient signals, etc. In one example, an index of the measured spectroscopic response is α, which is derived from the measurement data by methods known in the art, such as those described above with reference to equations (1)-(5). meas and β meas In other examples, other indices of the measured spectral response may be considered (e.g., tan Ψ and Δ, etc.). The aforementioned spectral response indices are provided as non-limiting examples. Other indices or combinations of indices may also be considered. It is important to note that the spectral indices are based on the spectral response of the specimen, rather than on specific metrics that may be derived from the specimen's spectral response (e.g., film thickness, refractive index, dielectric constant, etc.).

[0064] Furthermore, the application of the aforementioned method is not limited to a particular range of measured wavelengths, i.e., the method applies regardless of the range of wavelengths measured, for example, a range including any of VUV, UV, visible, near-infrared, and mid-infrared wavelengths.

[0065] It should be further noted that the application of the aforementioned method is not limited to spectroscopic ellipsometry. In general, the method and system for system parameter calibration may be applied to improve tool-to-tool matching and measurement stability of any metrology tool, both in online and offline implementations. Such systems are used to measure structural and material properties associated with different semiconductor fabrication processes (e.g., material composition, dimensional properties, etc., of structures and films).

[0066] 5 illustrates a method 200 suitable for implementation by the metrology system 100 of the present invention. It is recognized that, in one aspect, the data processing blocks of method 200 may be performed via pre-programmed algorithms executed by one or more processors of computing system 116. While the following description is presented in terms of metrology system 100, it is recognized herein that particular structural aspects of metrology system 100 should be construed as illustrative only and not limiting.

[0067] In block 201, a target measurement signal is received by a computing system, such as computing system 116. The target measurement signal is indicative of measurements for one or more structures disposed on a wafer by a target metrology system, such as metrology system 100. The target measurement signal is determined based at least in part on a quantity of raw measurement data collected by the target metrology system and one or more system parameter values ​​associated with the target metrology system.

[0068] At block 202, a model-based target measurement signal is determined that is indicative of measurements on one or more structures simulated by a target metrology system.

[0069] In block 203, model-based reference measurement signals are determined that are indicative of measurements on one or more structures simulated by a reference metrology system.

[0070] In block 204, a composite measurement matching signal is generated based on the target measurement signal, the model-based target measurement signal, and the model-based reference measurement signal.

[0071] At block 205, a measure of agreement between the target and reference measurement systems is determined based on the composite matching signal, including a trained error estimation model operating on the composite matching signal.

[0072] At block 206, an indication of the match is stored in memory, for example, memory of carrier medium 118.

[0073] The terms reference measurement system and target measurement system generally refer to a measurement system state (i.e., a target) that requires adaptation of system parameters to obtain measurement consistency with another measurement system state (i.e., a reference). In this way, the target is adjusted relative to the reference.

[0074] In some examples, the target metrology system and the reference metrology system are different tools. For example, in manufacturing situations, it may be advantageous for a family of metrology systems to each be calibrated to a single reference metrology system. In this way, each of the family of metrology systems is consistent with a single reference tool. In another example, it may be advantageous for one or more metrology systems to each be calibrated to a family average for multiple metrology systems. In this way, each of the metrology systems is consistent with the entire family of metrology tools. In another example, the reference system and the target system are the same system measured at different times (e.g., before and after a hardware maintenance operation).

[0075] In general, any suitable metrology system may be utilized as the high-integrity metrology system within the scope of this patent document. For example, any of a beam profile reflectometer, a reflectometer, and a suitable X-ray-based metrology system may be utilized as the high-integrity metrology system. Furthermore, the high-integrity metrology system need not be integrated with the target metrology tool. In some examples, the high-integrity metrology system may be a separate metrology tool.

[0076] In a further aspect, an optimized subset of system parameters is loaded into a target metrology system. These optimized parameters are then used for further measurement analysis, including a measurement model (e.g., the measurement model described with reference to equations (6) and (7)). In some examples, critical dimension (CD) measurements are performed by the target metrology system using the optimized subset of system parameters. For example, structural parameters of a calibration specimen may be estimated based on a regression of the updated target system metrology model against spectroscopic data associated with measurements of the calibration specimen. In this example, the spectroscopic data is also calculated based on the underlying raw measurement data and the optimized subset of system parameters.

[0077] In yet another aspect, the composite measurement matching signals that drive the training of the error estimation model may be weighted differently. In one example, the relative weighting is based on the measurement sensitivity for any of the multiple measurement sites, multiple measurement samples, multiple illumination wavelengths, and multiple measurement subsystems. In this manner, a particular measurement site, sample, subsystem, or illumination wavelength with particularly high measurement sensitivity may be emphasized. In another example, the relative weighting is based on the measurement noise associated with any of the multiple measurement sites, multiple measurement samples, multiple illumination wavelengths, and multiple measurement subsystems. In this manner, a particular measurement site, sample, subsystem, or illumination wavelength with particularly high measurement noise may be de-emphasized.

[0078] Metrology systems configured to measure the shape and material properties of dielectric and metallic films and structures may utilize the methods described herein. Such measurements include, by way of non-limiting example, measurements of film properties and dimensions, CD, overlay, and composition. Such metrology systems may include any number of illumination sources, including, but not limited to, lamps, lasers, laser-driven light sources, X-ray sources, and EUV sources. Such metrology systems may utilize several measurement techniques, including, but not limited to, all implementations of ellipsometers (including broadband spectroscopic or single wavelength, single or multi-angle, or angle-resolved, with fixed or rotating polarizers and compensators), all implementations of reflectometers (including spectroscopic or single wavelength, single or multi-angle, or angle-resolved), differential measurements such as scatterometers and interferometers, and all implementations of X-ray-based metrology.

[0079] As described herein, the term “metrology system” includes any system used at least in part to characterize a specimen in any manner. Exemplary terms used in the art may include a “defect inspection” system or an “inspection” system. However, such terminology does not limit the scope of the term “metrology system” as described herein. Additionally, the metrology system 100 may be configured for inspection of patterned and / or unpatterned wafers. The metrology system may be configured as an LED inspection tool, an edge inspection tool, a backside inspection tool, a macro inspection tool, or a multi-mode inspection tool (including data from one or more platforms simultaneously), as well as any other metrology or inspection tool that would benefit from calibrating system parameters based on the difference in error spectra between a reference metrology tool and a target metrology tool.

[0080] Various embodiments are described herein in terms of semiconductor processing systems (e.g., metrology systems or lithography systems) that may be used to process specimens. The term "specimen" is used herein to refer to a wafer, reticle, or any other sample that may be processed (e.g., printed or inspected for defects) by means known in the art.

[0081] As used herein, the term "wafer" generally refers to a substrate formed from a semiconductor or non-semiconductor material. Examples include, but are not limited to, monocrystalline silicon, gallium arsenide, and indium phosphide. Such substrates may be commonly found and / or processed in semiconductor fabrication facilities. In some cases, a wafer may include only a substrate (i.e., a bare wafer). Alternatively, a wafer may include one or more layers of different materials formed on a substrate.

[0082] One or more layers may be formed on the wafer. For example, such layers may include, but are not limited to, resist, dielectric material, conductive material, and semiconductor material. Many different types of such layers are known in the art, and the term wafer as used herein is intended to encompass wafers on which all types of such layers may be formed.

[0083] One or more layers formed on a wafer may be "patterned" or "unpatterned." For example, a wafer may include multiple dies with repeatable pattern features. The formation and processing of such material layers may ultimately result in a completed device. Many different types of devices may be formed on a wafer, and the term wafer as used herein is intended to encompass a wafer having any type of device fabricated thereon known in the art.

[0084] A typical semiconductor process involves wafer processing in lots. As used herein, a "lot" is a group of wafers (e.g., a group of 25 wafers) that are processed together. Each wafer in a lot is composed of multiple exposure fields from a lithography processing tool (e.g., a stepper, scanner, etc.). Multiple dies may be present within each field. A die is a functional unit that ultimately becomes a single chip. One or more layers formed on a wafer may be patterned or unpatterned. For example, a wafer may include multiple dies, each of which has repeatable patterned features. The formation and processing of such material layers may ultimately result in a completed device. Many different types of devices may be formed on a wafer, and the term wafer as used herein is intended to encompass a wafer having any type of device fabricated thereon known in the art.

[0085] A "reticle" may be a reticle at any stage in the reticle fabrication process or a completed reticle that may or may not be released for use in a semiconductor fabrication facility. A reticle, or "mask," is generally defined as a substantially transparent substrate having substantially opaque regions formed thereon and arranged in a pattern. The substrate may comprise, for example, a glass material such as quartz. The reticle may be disposed on a resist-coated wafer during an exposure step of a lithography process, thereby transferring the pattern on the reticle to the resist.

[0086] In one or more exemplary embodiments, the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functions may be stored on or transmitted via one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media, including any medium that facilitates transfer of a computer program from one place to another. Storage media may be any available medium that can be accessed by a general-purpose or special-purpose computer. By way of example, and not limitation, such computer-readable media may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code means in the form of instructions or data structures and that can be accessed by a general-purpose or special-purpose computer, or a general-purpose or special-purpose processor. Also, any connection is properly termed a computer-readable medium. For example, if software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of medium. Disk and disc, as used herein, include compact discs (CDs), laser discs, optical discs, digital versatile discs (DVDs), floppy disks, and Blu-ray discs, where disks typically reproduce data magnetically while discs reproduce data optically using lasers. Combinations of the above are also necessarily included within the scope of computer-readable media.

[0087] Although certain specific embodiments are described above for instructional purposes, the teachings of this patent document have general applicability and are not limited to the specific embodiments described above. Accordingly, various modifications, adaptations, and combinations of the various features of the described embodiments may be made without departing from the scope of the invention as set forth in the claims.

Claims

1. receiving, by a target metrology system, target measurement signals indicative of measurements for one or more structures disposed on a wafer, the target measurement signals being determined at least in part based on an amount of raw measurement data collected by the target metrology system and one or more system parameter values ​​associated with the target metrology system; determining a model-based target measurement signal indicative of measurements on the one or more structures simulated by the target metrology system; determining a model-based reference measurement signal indicative of measurements for the one or more structures simulated by a reference metrology system; generating a composite measurement matching signal based on the target measurement signal, the model-based target measurement signal, and the model-based reference measurement signal; determining an indicator of match between the target measurement system and the reference measurement system based on the composite matching signal, the determining step including a trained error estimation model operating on the composite matching signal; storing the indication of the match in a memory; A method comprising:

2. 2. The method of claim 1, further comprising training the error estimation model, the training step being based on a plurality of Design of Experiments (DOE) measurements of one or more structures by a group of metrology systems and a DOE composite measurement matching signal associated with a corresponding indicator of agreement between each of the group of metrology systems and the reference metrology system.

3. The method of claim 1 , further comprising determining a set of values ​​for one or more system parameters of the target measurement system, said determining including the trained error estimation model.

4. 4. The method of claim 3, further comprising training the error estimation model, the training step being based on a plurality of Design of Experiments (DOE) measurements of one or more structures by a group of metrology systems and a DOE composite measurement matching signal associated with corresponding indicators for values ​​of the one or more system parameters associated with each of the group of metrology systems.

5. The method of claim 1 , wherein the composite measurement matching signal is the sum of the target measurement signal and a difference term, the difference term being the difference between the model-based reference measurement signal and the model-based target measurement signal.

6. The method of claim 1 , wherein the reference measurement system is a single measurement system of a group of measurement systems.

7. The method of claim 1 , wherein the reference measurement system is an average of a plurality of measurement systems in the group of measurement systems.

8. The method of claim 1 , wherein the target measurement system and the reference measurement system are spectroscopic ellipsometers.

9. The method of claim 1 , wherein the reference measurement system is a measurement system measured a first time, and the target measurement system is the measurement system measured a second time after the first time.

10. 10. The method of claim 1, wherein the measurements for the one or more structures include spectroscopic measurement data associated with any of a plurality of measurement sites, a plurality of measurement samples, a plurality of illumination wavelengths, and a plurality of measurement modalities.

11. an illumination source configured to provide an amount of illumination light to one or more metrology targets disposed on the wafer; a detector configured to detect an amount of light from the one or more measurement targets in response to the amount of illumination light and generate a measurement signal in response to the amount of detected light; one or more computing systems; Equipped with the one or more computing systems; receiving, by a target metrology system, target measurement signals indicative of measurements for one or more structures disposed on a wafer, the target measurement signals being determined at least in part based on an amount of raw measurement data collected by the target metrology system and one or more system parameter values ​​associated with the target metrology system; determining a model-based target measurement signal indicative of measurements for the one or more structures simulated by the target metrology system; determining a model-based reference measurement signal indicative of measurements for the one or more structures simulated by a reference metrology system; generating a composite measurement matching signal based on the target measurement signal, the model-based target measurement signal, and the model-based reference measurement signal; determining an indicator of agreement between the target measurement system and the reference measurement system based on the composite matching signal, the determining step including a trained error estimation model operating on the composite matching signal; storing the indication of the match in a memory; 1. A measurement system configured to:

12. the one or more computing systems; 12. The metrology system of claim 11, further configured to perform a step of training the error estimation model, the training step being based on a plurality of Design of Experiments (DOE) measurements of one or more structures by a group of metrology systems and a DOE composite measurement matching signal associated with a corresponding indicator of agreement between each of the group of metrology systems and the reference metrology system.

13. the one or more computing systems; 12. The measurement system of claim 11, further configured to perform the step of determining a set of values ​​for one or more system parameters of the target measurement system, said determining step including the trained error estimation model.

14. 14. The metrology system of claim 13, further comprising training the error estimation model, the training step being based on a plurality of Design of Experiments (DOE) measurements of one or more structures by a group of metrology systems and a DOE composite measurement matching signal associated with corresponding indicators for values ​​of the one or more system parameters associated with each of the group of metrology systems.

15. 12. The metrology system of claim 11, wherein the composite measurement matching signal is a sum of the target measurement signal and a difference term, the difference term being a difference between the model-based reference measurement signal and the model-based target measurement signal.

16. The metrology system of claim 11 , wherein the reference metrology system is a single metrology system of a group of metrology systems.

17. The measurement system of claim 11 , wherein the reference measurement system is an average of a plurality of measurement systems in the group of measurement systems.

18. The measurement system of claim 11 , wherein the reference measurement system is a measurement system measured a first time, and the target measurement system is the measurement system measured a second time after the first time.

19. an illumination source configured to provide an amount of illumination light to one or more metrology targets disposed on the wafer; a detector configured to detect an amount of light from the one or more measurement targets in response to the amount of illumination light and generate a measurement signal in response to the amount of detected light; a non-transitory computer-readable medium storing instructions; Equipped with The instructions, when executed by one or more processors, cause the one or more processors to: receiving, by a target metrology system, target metrology signals indicative of measurements for one or more structures disposed on a wafer, the target metrology signals being determined at least in part based on an amount of raw measurement data collected by the target metrology system and one or more system parameter values ​​associated with the target metrology system; determining a model-based target measurement signal indicative of measurements for the one or more structures simulated by the target metrology system; determining a model-based reference measurement signal indicative of measurements for the one or more structures simulated by a reference metrology system; generating a composite measurement matching signal based on the target measurement signal, the model-based target measurement signal, and the model-based reference measurement signal; determining an indication of agreement between the target measurement system and the reference measurement system based on the composite matching signal, the determining step including a trained error estimation model operating on the composite matching signal; storing the indication of the match in a memory; A measurement system that performs the following.

20. The non-transitory computer-readable medium further stores instructions that, when executed by the one or more processors, cause the one or more processors to:

20. The measurement system of claim 19, further comprising the step of determining a set of values ​​for one or more system parameters of the target measurement system, said determining including the trained error estimation model.

Images