Random perceptual lithography model for mask synthesis
By employing a compact model in the lithography process, considering local random variations and combining empirical data regression, the suboptimal mask design problem caused by local random variations in the lithography process is solved, achieving more accurate mask design and more efficient model prediction.
Patent Information
- Application Number
- CN202180068699.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2021-11-09
- Filing Date
- 2021-11-10
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2041-11-10
AI Technical Summary
Existing lithography process models have failed to effectively address the suboptimal mask design problem caused by local random variations in lithography processes, especially in extreme ultraviolet lithography.
By employing a compact model and considering local random variations through parameterized model properties, combined with empirical data regression, the characteristics of printed patterns, including photon density and light signal intensity, are predicted to form a deterministic model for improving mask design.
It improves the prediction accuracy of photolithography processes, reduces model running time, and allows for flexibility to adapt to different process conditions, thereby improving the accuracy and efficiency of mask design.
Smart Images

Figure CN116324617B_ABST
Abstract
Description
[0001] Related applications
[0002] This application claims priority to U.S. Provisional Patent Application Serial No. 63 / 112,733, filed November 12, 2020, entitled “Stochastic-Aware Lithographic Models For Mask Synthesis,” and U.S. Patent Application Serial No. 17 / 522,574, filed November 9, 2021, entitled “Stochastic-Aware Lithographic Models For Mask Synthesis,” filed under 35 U.S. SC § 119(e). All the foregoing technical solutions are incorporated herein by reference in their entirety. Technical Field
[0003] This disclosure relates to lithographic modeling and mask synthesis, including for extreme ultraviolet (EUV) masks. Background Technology
[0004] One step in manufacturing semiconductor wafers involves photolithography. In a typical photolithography process, a light source generates light, which is collected and guided by collecting / illuminating optics to illuminate a photomask. Projection optics relay the pattern generated by the illuminated mask onto the wafer, exposing a photoresist on the wafer according to the illuminated pattern. The patterned photoresist is then used in processes that fabricate structures on the wafer.
[0005] Various techniques involve improving photolithography processes, including the design of photomasks. In many of these techniques, the photomask design is used as input to a process model that then predicts certain process outcomes. These outcomes can be used to modify the photomask design. In many cases, the process model can be regressed relative to data from actual manufacturing experiments. Different photomask patterns, including test patterns, are used throughout the photolithography process. The resulting structures are measured and used to calibrate the process model.
[0006] Current process models typically do not directly address variations that may arise due to stochastic variations inherent in the lithography process itself. Instead, the lithography mask pattern is used as input to the model, which models the lithography process as a fully deterministic process. However, as lithography moves to shorter wavelength ranges (e.g., extreme ultraviolet (EUV) at approximately 13.3–13.7 nm) and smaller geometries (e.g., 10 nm, 7 nm, and smaller technology nodes, with minimum property sizes of approximately 20 nm, 14 nm, and smaller), stochastic variations (local stochastic variations) within relatively small regions become more significant, and conventional methods can lead to suboptimal mask designs. Summary of the Invention
[0007] In some aspects, mask patterns are accessed. Mask patterns are used in the photolithography process to print patterns on a wafer. The mask pattern is taken as input to a deterministic model of the photolithography process to predict the characteristics of the printed pattern. The deterministic model is deterministic (e.g., not a Monte Carlo simulation), but it takes into account local stochastic variations in the properties of the printed pattern.
[0008] In some embodiments, the deterministic model is a compact model that predicts hot spots in a printed pattern based on the mask pattern and also on stochastic model properties related to dose, focus, and local stochastic variations in the predicted characteristics. Examples of stochastic model properties may include diffuse spatial image signals; acid concentration or density; quencher concentration or density; inhibitor concentration or density; gradients of spatial image, acid, quencher, or inhibitor signals; mask pattern density; secondary electron concentration; ligand concentration; and resistance to surface tension. The compact model regresses against empirical data that includes local stochastic variations in the predicted characteristics. Examples of empirical data may include line edge roughness, critical size (CD) variation, worst-case CD, and the number of empirical measurements of CD being less than or greater than a specific criterion.
[0009] Other aspects include components, devices, systems, improvements, methods, processes, applications, computer-readable media, and other technologies related to any aspect. Attached Figure Description
[0010] This disclosure will be more fully understood from the following detailed description and the accompanying drawings illustrating embodiments of the present disclosure. The drawings are intended to provide knowledge and understanding of embodiments of the present disclosure, but are not intended to limit the scope of the disclosure to these particular embodiments. Furthermore, the drawings are not necessarily drawn to scale.
[0011] Figure 1 EUV lithography processes applicable to embodiments of this disclosure are described.
[0012] Figure 2A A flowchart for simulating photolithography processes is described.
[0013] Figure 2B A flowchart is described using a compact model that considers local random variations to simulate a photolithography process.
[0014] Figure 3 A flowchart is described to add stochastic model properties to a compact model that does not take into account local stochastic variations.
[0015] Figure 4A and 4B The simulation of the photolithography process is described.
[0016] Figure 5 Depicting from Figure 3 The process window axis diagram and simulation image of the process.
[0017] Figure 6 Flowcharts depict various processes used during the design and manufacture of integrated circuits according to some embodiments of the present invention.
[0018] Figure 7 An abstract diagram of an example computer system in which embodiments of the present disclosure may operate is depicted. Detailed Implementation
[0019] Various aspects of this invention relate to a “random-aware” lithography model for mask synthesis. Local random effects within the lithography process itself become more pronounced, particularly at shorter wavelengths and higher photon energies. Local random variations are changes resulting from processes that possess some degree of randomness at a local level. For example, higher photon energies at EUV wavelengths mean fewer photons are needed to achieve a given energy exposure compared to longer wavelengths. Photon events possess some degree of randomness (i.e., probability distribution), which becomes more significant if the total photon count is low. As a result, variations in the photon count at any particular location on the wafer can lead to random defect mechanisms that were previously negligible or nonexistent for higher numbers of photons. Other components of a lithography process that can have local random variations include photon distribution, secondary electron distribution, photoacid generator (PAG) distribution, quencher distribution, and inhibitor distribution. These random effects are local, meaning they can occur in relatively small areas, such as within a 10µm × 10µm area or smaller of the printed pattern.
[0020] In various embodiments of the invention, a deterministic model of the lithography process addresses these local stochastic variations in the lithography process. In one approach, the model is a compact model that can be used for mask synthesis. The compact model is a parametric model. These parameters can be referred to as model properties. The values of the model properties can be determined by regression on empirical data. The model properties include at least one stochastic model property associated with local stochastic variations in the lithography process. Examples of possible stochastic model properties include photon density, pattern density, optical signal intensity, optical signal gradient, secondary electron density, secondary electron gradient, photoacid concentration, resist quencher concentration, resist inhibitor concentration, resist inhibitor gradient, and metal resist ligand concentration. Although the model considers stochastic variations, the model itself is deterministic. That is, for any given input, the output predicted by the model is always the same (there are no stochastic variations in the model's output), but the output has been adjusted to address the stochastic variations in the lithography process.
[0021] The advantages of this disclosure include, but are not limited to, the following. Consideration of local random variations yields more accurate predictions because random effects are not ignored. Furthermore, using a deterministic model to do so reduces model runtime. In other approaches, random variations can be modeled using stochastic models, where the same simulation is run multiple times using statistically varying conditions (e.g., Monte Carlo simulations). However, this approach requires multiple runs to obtain results based on different predictions from different runs, whereas a deterministic model produces a prediction in a single run. Parametric models, regressing against empirical data, also produce more accurate predictions. It also allows for some flexibility in model form. The constants and parameters used in the model do not need to be calculated with exact precision, and the physical processes occurring do not need to be modeled with exact precision, as model accuracy can be achieved through calibration against empirical data.
[0022] Figure 1 An EUV lithography process suitable for embodiments of this disclosure is described. In this system, a light source 110 generates EUV light, which is collected and guided by a collection / illumination optics 120 to illuminate a mask 130. A projection optics 140 relays the pattern generated by the illuminated mask onto a wafer 150, exposing a resist on the wafer according to the illuminated pattern. The exposed resist is then developed, creating a patterned resist on the wafer. This printed pattern is used to fabricate structures on the wafer, for example, by deposition, doping, etching, or other processes.
[0023] exist Figure 1 In this process, the light is in the EUV wavelength range, approximately 13.5 nm or in the range of 13.3–13.7 nm. At these wavelengths, components are typically reflective rather than transmissive. Mask 130 is a reflective mask, and optics 120 and 140 are also reflective and off-axis; this is just one example. Other types of lithography systems can also be used, including at other wavelengths, using transmissive masks and / or optics, and using positive or negative resists.
[0024] Figure 2A The simulation of the photolithography process is described. Illumination model 222 simulation. Figure 1The light source 110 and illumination optics 120 are used to predict the light pattern 225 illuminating the mask. Model 242 illustrates the effects of the photomask 130 and projection optics 140 to predict the illumination 245 for photoresist exposure. This is sometimes referred to as a spatial image 245. The photoresist model 252 may include effects such as photoresist exposure from the spatial image, secondary electron generation, chemical reactions, and subsequent removal, whether by chemical development, etching, or other processes. The removal of the photoresist leaves a patterned layer 254 on the wafer, also known as a printed pattern. The wafer fabrication model 256 then simulates subsequent processes, such as etching, deposition, doping, implantation, etc., to create device structures 258 on the wafer.
[0025] For convenience, Figure 2A The diagram shows individual boxes corresponding to physical components or processes, but it's not necessary to implement the model in this way. For example, all these effects could be combined into a single compact model, which could be used for full-chip evaluation, such as... Figure 2B As shown. This model predicts the patterned resist or equivalent spatial image (in the mask pattern 230) from the mask pattern 230. Figure 2B This is referred to as printed pattern 234. This single model incorporates effects from the light source, optics, and resist, and in the embodiments described herein, also considers localized random variations in the photolithography process. A compact model 232 can be designed to process the mask pattern across the entire die in a rapid manner. Up to 10 18 Or more simulations to perform full-chip calibration on devices such as microprocessors.
[0026] The compact model 232 can be a parameterized deterministic model that predicts the characteristics of the printed pattern 234 and / or the printed pattern based on model parameters, also referred to as model properties. For example, the compact model can predict hot spots (defects in the printed pattern) or failure rates in the pattern based on the mask pattern 230, the dose and focus of the lithography system, and also based on certain model properties related to local random variations. These model properties will be referred to as stochastic model properties.
[0027] A compact model can be regressed against empirical data. Measurement data for different mask patterns are collected. The resulting patterns fabricated on the wafer are measured. These are typically critical-size scanning electron microscopy (CDSEM) measurements of the photoresist and / or etched structures. Other measurements, such as atomic force microscopy (AFM), may also be used.
[0028] One measure of the quality of patterned resist 234 is the critical dimension (CD). CD is the size of a significant feature in the patterned resist or the size of a significant spacing between features. Typically, the minimum CD is the minimum linewidth or spacing width printed in the resist. Therefore, it is a measure of the resolution of the resist and the photolithography process. (See again...) Figure 2A and 2B The simulation results 254 and 234 in these figures can include CD predictions. Another common output is the predicted spatial image, or a contour of constant intensity in the spatial image.
[0029] Compact models can be used for mask synthesis or mask correction. Mask correction includes optical proximity correction, subresolution auxiliary features, phase-shifting masks, reverse lithography, and other types of resolution enhancement techniques. In optical proximity correction (OPC), the geometry in the mask pattern is perturbed based on the prediction results. In subresolution auxiliary features, subwavelength features are added to or removed from the mask pattern to introduce beneficial diffraction and scattering. In phase-shifting masks, different mask shapes introduce different amounts of phase shift to introduce beneficial interference in the resulting illumination pattern.
[0030] Figure 3 A flowchart is described to describe adding stochastic model properties to a compact model that does not account for local stochastic variations. The flowchart will be explained using examples shown in Figures 4 and 5. Compact model 310 predicts the location of hot spots (e.g., defects) on the wafer, but does not account for the increased failure rate due to local stochastic variations. For example, this “stochastic unconscious” compact model 310 can predict patterned resist CD based on dose, focus, and mask pattern. It can also predict the failure rate or identify hot spots within a process window (PW) defined by the range of these parameters. Compact model 310 can be a parametric model that regresses relative to empirical data.
[0031] The compact model 310 may include observable PW mask or scanner parameters, such as dose, focus, and mask bias. These are set to correspond to physically measurable quantities set during experimental photolithography exposure. The compact model 310 may also include non-physical model parameters or parameters in the model that are not visible on the mask or scanner, such as chemical blurring, inhibitor concentration, or mechanical stress. These parameters cannot change in a directly observable manner during photolithography exposure, and their parameter coefficients and weights are suitable for averaging wafer measurements observed with different resists or etch patterns.
[0032] Figure 4A and 4B A simulation of the photolithography process using this compact model 310 is described. Figure 4A In the diagram, the straight lines represent the mask pattern, and the curves represent the printed resist shapes obtained on the wafer under different dosages and focal lengths, such as... Figure 4A The elliptical process window diagram on the right-hand side is shown. In this example, the process window is + / - 50nm defocus and + / - 3% dose. Figure 4A The curve in the graph corresponds to a point in the PW plot.
[0033] Model 310 can be relative to, for example Figure 4B The empirical data shown is a regression. Figure 4B Two different scenarios are shown. The left example applies to optimal process window conditions: no focal spot or dose deviating from the ideal value. Figure 4B The left side shows the actual printed structure on a wafer without any defects. The correct case is for conditions at the edge of the process window. In this case, the actual printed structure has defects, with breaks in the printed lines. The results predicted by model 310 can be improved by comparing them with the measured results.
[0034] Model 310 in Figure 4 illustrates variations in dose and focus, but not local random variations. Return to... Figure 3 The stochastic unknown model was modified in steps 320, 330, and 340 to account for local stochastic variations. Some model forms or parameters, collectively referred to as Stochastic Model Properties (SMP), were added to the compact model in step 320. Empirical data, reflecting the effects of local stochastic variations, were collected in step 330. The SMP was regressed against the empirical data with local stochastic variations in step 340. Some statistical local variation on the wafer was used for regression in step 340. This could be worst-case CD, a measure of CD variation, or characteristics of line edge roughness or defect location, etc.
[0035] The stochastic unknown model 310 defines what constitutes observable process parameters, such as scanner dose, focus, or feature CD on the mask. The model 310 can account for variations in observable process parameters, such as if the scanner dose is slightly greater or less than the nominal value, or if there is slight defocus, or if the feature CD is slightly greater or less than the nominal value. It is unknown whether the stochastic compact model 310 can account for these types of variations.
[0036] exist Figure 3 In this process, the stochastic unknown model 310 is modified to address locally occurring stochastic variations, such as within a single CDSEM image view. The compact model 310 is modified to include these local stochastic variations, and further modified by adjusting these additional SMPs. This adjusts the model form so that the new “stochastic-aware” model 350 form predicts not only deterministic variations but also local stochastic variations. Examples include CD changes across lines or multiple repeating patterns across a local region of the wafer.
[0037] The random sensing model 350 can add "random sensing" lithography non-physical model parameters or parameters in a model that is not observable on a mask or scanner, such as secondary electron density, chemical interaction range, inhibitor yield, additional chemical blur, additional mechanical stress, etc., which cannot be changed during lithography exposure, but whose parameter coefficients and weights are suitable for the statistical changes observed in repeated local wafer measurements for each of several different individual resists or etch patterns.
[0038] For example, the stochastic unknown model 310 can predict the location of a line edge on a wafer. However, due to local random variations, there actually exists a probability distribution function for the location of the line edge. Perhaps the distribution is Gaussian, with the expected location and some standard deviation from the expected location. The stochastic unknown model 310 can predict the expected location but ignores the probability distribution. The stochastic perceptual model 350 considers the fact that a probability distribution exists. It may not explicitly predict what the distribution is, but it takes into account the effect of the distribution. For example, it can predict the line edge location based on one standard deviation from the expected location.
[0039] Note that the “random perception” model 350 itself does not have to be random. It can be a deterministic model that always produces the same predictions for the same input, rather than a prediction with some random elements, but the predictions take into account local random effects. In this example, model 350 takes into account a Gaussian distribution by always predicting the position of the line edge based on the standard deviation of the probability distribution.
[0040] Note that these variations are localized—they occur in very localized areas, such as within the field of view of a CDSEM image, which typically has a size between 0.5 and 10 micrometers per side. In some cases, variations may be visible on a 10×10 patterned array, a 20×20 patterned array, or a 1×40 1D patterned array. As a counterexample, dose and defocus can also vary, but they are not localized because these variations affect much larger areas, such as one thousand micrometers per side, or even the entire exposure field on the wafer.
[0041] In other words, "random" hotspots can be defined as patterns observed on a chip that predict a much higher failure rate than traditional "random unconscious" hotspot detection methods. Stochastic model properties (SMPs) are model properties that are associated with or predict random hotspot failures.
[0042] Figure 5 Describes the use of predicted signal strength images Figure 3The process involves analyzing the signal intensity images. These images can be spatial images or chemical properties of the resist, such as photosensitive acid concentration. In this example, the left column 510 represents the random, undefined compact model 310, the center column 540 is a correction for local random variations, and the right column 550 represents the resulting random, undefined model 350. The bottom row represents the predicted signal intensity images, and the ellipses in the top row represent the parameters considered in the process window.
[0043] In 510, a compact model 310 is developed for a process window with dose and focus variations, but local stochastic variations are not considered. In 540, SMPs (random param1 in the top row of the figure) are added to the model to account for local stochastic process variations. They add correction factors to the signal image. The resulting signal image is shown in 550. In this example,
[0044] Modified image signal = Original image signal + coef * photoacid gradient signal (1) The photoacid gradient is the gradient of photoacid concentration in the resist after exposure, and coef is a negative or positive parameter determined based on comparison with empirical data. Other examples can use a nonlinear combination of the original image signal and SMP to create a modified image signal.
[0045] In this example, the photoacid gradient is an SMP. Other SMPs can also be used. Examples may include:
[0046] Photon density or diffuse spatial image signal
[0047] Nonlinear filtering of aerial image signals
[0048] Acid concentration or density
[0049] quencher concentration or density
[0050] Inhibitor concentration or density
[0051] Secondary electron concentration or density
[0052] ligand concentration or density
[0053] Spatial image signal gradient
[0054] Gradient of acid, quencher, or inhibitor signaling
[0055] Logarithmic gradient of spatial image, acid, quencher, or inhibitor signal
[0056] The square of the gradient of the spatial image, acid, quencher, or inhibitor signal.
[0057] Mask pattern density
[0058] surface tension resistance
[0059] Figure 6 A set of exemplary processes 600 are illustrated during the design, verification, and manufacturing of an article of art, such as an integrated circuit, to transform and verify design data and instructions representing an integrated circuit. Each of these processes can be configured and enabled as multiple modules or operations. The term "EDA" stands for "Electronic Design Automation." These processes begin with the creation of a product concept 610 with information provided by a designer, which is transformed to create an article of art using a set of EDA processes 612. When the design is completed, the design is tape-out 634, which is when the pattern (e.g., geometric pattern) of the integrated circuit is sent to a manufacturing facility to create a mask set, which is then used to manufacture the integrated circuit. After tape-out, semiconductor dies are fabricated 636, and packaging and assembly processes 638 are performed to produce a finished integrated circuit 640.
[0060] Specifications for circuits or electronic structures can range from low-level transistor material placement to high-level description languages. High-level abstractions can be used to design circuits and systems using hardware description languages (“HDLs”) such as VHDL, Verilog, SystemVerilog, SystemC, MyHDL, or OpenVera. HDL descriptions can be translated into logic-level register-transfer-level (“RTL”) descriptions, gate-level descriptions, placement-level descriptions, or mask-level descriptions. Each lower level of abstraction, as a description of less abstraction, adds more useful details to the design description, such as more details about the modules described. Lower-level abstractions with fewer abstractions can be computer-generated, exported from design libraries, or created by another design automation process. An example of a specification language used to specify a more detailed lower-level abstraction is SPICE, which is used for detailed descriptions of circuits with many analog components. The description at each level of abstraction can be used by the corresponding tools at that level (e.g., formal verification tools). The design process can use… Figure 6 The order shown. The EDA product (or tool) should enable the described process.
[0061] During system design phase 614, the functionality of the integrated circuit to be manufactured is specified. The design can be optimized for desired characteristics such as power consumption, performance, area (physical and / or lines of code), and cost reduction. At this stage, the design can be divided into different types of modules or components.
[0062] During logic design and functional verification 616, modules or components in a circuit are specified using one or more description languages, and the functional accuracy of the specifications is checked. For example, components of a circuit can be verified to generate outputs that match the specification requirements of the designed circuit or system. Functional verification can use simulators and other programs, such as testbench generators, static HDL checkers, and formal verifiers. In some embodiments, a specific system of components, referred to as a “simulator” or “prototype system,” is used to accelerate functional verification.
[0063] During the synthesis and design phases of Test 618, HDL code is transformed into a netlist. In some embodiments, the netlist can be a graphical structure, where the edges of the graphical structure represent components of the circuit, and the nodes of the graphical structure represent how the components are interconnected. Both HDL code and netlist are hierarchical artifacts that EDA products can use to verify that the integrated circuit performs as specified during manufacturing. The netlist can be optimized for a target semiconductor manufacturing technology. Furthermore, the completed integrated circuit can be tested to verify that it meets specification requirements.
[0064] During netlist verification (620), the netlist is checked to ensure it conforms to timing constraints and HDL code. During design planning (622), the overall layout of the integrated circuit is constructed and analyzed for timing and top-level routing.
[0065] During layout or physical implementation 624, physical placement (location of circuit components such as transistors or capacitors) and routing (connection of circuit components via multiple conductors) occur, and cells can be selected from a library to enable specific logic functions. As used herein, the term "cell" can specify a set of transistors, other components, and interconnections that provide Boolean logic functions (e.g., AND, OR, NOT, XOR) or storage functions (e.g., flip-flops or latches). As used herein, a circuit "block" can refer to two or more cells. Both cells and circuit blocks can be referred to as modules or components and are implemented as physical structures and simulations. Parameters, such as size, are specified for selected cells (based on "standard cells") and accessed in a database for use in EDA products.
[0066] During analysis and extraction (626), circuit functionality is verified at the layout level, allowing for improvements to the layout design. During physical inspection (628), the layout design is checked to ensure that manufacturing constraints are correct, such as DRC constraints, electrical constraints, lithographic constraints, and that the circuit device functionality matches the HDL design specifications. During resolution enhancement (630), the geometry of the layout is transformed to improve how the circuit design is manufactured.
[0067] During the tape-out process, data is created for the production of the photomask (if appropriate, after the application of photolithography enhancement). During mask data preparation 632, the "tape-out" data is used to generate the photomask, which is then used to produce the finished integrated circuit.
[0068] Computer systems (such as) Figure 7 The storage subsystem of the computer system 700 can be used to store some or all of the programs and data structures used by the EDA products described herein, as well as the physical and logical designs of the products used for developing library units and using libraries.
[0069] Figure 7 An exemplary machine of computer system 700 is shown, within which a set of instructions can be executed to cause the machine to perform any or more methods discussed herein. In alternative implementations, the machine may be connected (e.g., networked) to other machines in a LAN, intranet, extranet, and / or the Internet. The machine may operate as a server or client machine in a client-server network environment, as a peer-to-peer (or distributed) network environment, or as a server or client machine in a cloud computing infrastructure or environment.
[0070] The machine can be a personal computer (PC), tablet PC, set-top box (STB), personal digital assistant (PDA), cellular phone, web device, server, network router, switch or bridge, or any machine capable of executing a set of instructions (sequential or otherwise) specifying the action the machine is to take. Furthermore, although a single machine is shown, the term "machine" should also be understood to include any collection of machines that individually or jointly execute a set (or more) of instructions to perform any one or more methods discussed herein.
[0071] Example computer system 700 includes processing device 702, main memory 704 (e.g., read-only memory (ROM), flash memory, dynamic random access memory (DRAM), such as synchronous DRAM (SDRAM), static memory 706 (e.g., flash memory, static random access memory (SRAM) and the like), and data storage device 718, which communicate with each other via bus 730.
[0072] Processing device 702 represents one or more processors, such as microprocessors, central processing units, etc. More specifically, processing device may be a Complex Instruction Set Computing (CISC) microprocessor, a Reduced Instruction Set Computing (RISC) microprocessor, a Very Long Instruction Word (VLIW) microprocessor, or a processor implementing other instruction sets, or a processor implementing combinations of instruction sets. Processing device 702 may also be one or more special-purpose processing devices, such as application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), network processors, etc. Processing device 702 may be configured to execute instructions 726 for performing the operations and steps described herein.
[0073] The computer system 700 may also include a network interface device 708 that communicates via a network 720. The computer system 700 may also include a video display unit 710 (e.g., a liquid crystal display (LC) or a cathode ray tube (CRT)), an alphanumeric input device 712 (e.g., a keyboard), a cursor control device 714 (e.g., a mouse), a graphics processing unit 722, a signal generation device 716 (e.g., a speaker), a graphics processing unit 722, a video processing unit 728, and an audio processing unit 732.
[0074] Data storage device 718 may include machine-readable storage medium 724 (also referred to as non-transitory computer-readable medium) on which one or more sets of instructions 726 or software embody any one or more methods or functions described herein. During execution by computer system 700, the instructions 726 may also reside wholly or at least partially in main memory 704 and / or processing device 702, which also constitute machine-readable storage media.
[0075] In some implementations, instruction 726 includes instructions that implement the functions corresponding to this disclosure. Although machine-readable storage medium 724 is shown as a single medium in the example implementation, the term "machine-readable storage medium" should be understood to include a single medium or multiple media (e.g., a centralized or distributed database, and / or associated caches and servers) storing one or more sets of instructions. The term "machine-readable storage medium" should also be understood to include any medium capable of storing or encoding a set of instructions for execution by a machine and causing the machine and processing device 702 to perform any one or more methods of this disclosure. Therefore, the term "machine-readable storage medium" should be understood to include, but is not limited to, solid-state memories, optical media, and magnetic media.
[0076] Certain parts of the foregoing detailed description are presented based on algorithms and symbolic representations of operations on data bits within computer memory. These algorithmic descriptions and representations are the methods used by those skilled in the art of data processing to most effectively communicate the substance of their work to others skilled in the art. An algorithm can be a sequence of operations that leads to a desired result. These operations are those that require physical manipulation of physical quantities. These quantities can take the form of electrical or magnetic signals that can be stored, combined, compared, and otherwise manipulated. Such signals can be referred to as bits, values, elements, symbols, characters, items, numbers, etc.
[0077] However, it should be remembered that all these and similar terms will be associated with appropriate physical quantities and are merely convenient notations applied to those quantities. Unless otherwise stated, as is apparent from this disclosure, it should be understood that throughout the specification, certain terms refer to the actions and processes of a computer system or similar electronic computing device that manipulate and convert data representing physical (electronic) quantities within the registers and memories of the computer system into other data representing physical quantities similarly represented within the computer system's memory or registers or other such information storage devices.
[0078] This disclosure also relates to an apparatus for performing the operations herein. The apparatus may be specifically constructed for the intended purpose, or it may comprise a computer selectively activated or reconfigured by a computer program stored in the computer. Such a computer program may be stored in a computer-readable storage medium, such as, but not limited to, any type of disk including floppy disks, optical disks, CD-ROMs and magneto-optical disks, read-only memory (ROM), random access memory (RAM), EPROM, EEPROM, magnetic cards or optical cards, or any type of medium suitable for storing electronic instructions, each coupled to a computer system bus.
[0079] The algorithms and demonstrations presented herein do not inherently relate to any particular computer or other device. Various other systems can be used with the program based on the teachings herein, or it may prove convenient to construct more specialized devices to perform the method. Furthermore, this disclosure is described without reference to any particular programming language. It should be understood that the teachings of the invention described herein can be implemented using a variety of programming languages.
[0080] This disclosure can be provided as a computer program product or software, which may include a machine-readable medium having instructions stored thereon, which can be used to program a computer system (or other electronic device) to perform the process according to this disclosure. Machine-readable media include any mechanism for storing information in a machine-readable (e.g., computer-readable) form. For example, machine-readable (e.g., computer-readable) media include machine-readable (e.g., computer-readable) storage media, such as read-only memory (“ROM”), random access memory (“RAM”), disk storage media, optical storage media, flash memory devices, etc.
[0081] In the foregoing disclosure, implementations of this disclosure have been described with reference to specific example implementations thereof. It will be apparent that various modifications may be made thereto without departing from the broader spirit and scope of the implementations of this disclosure as set forth in the following claims. Where elements are referred to in the singular in this disclosure, more than one element may be depicted in the drawings, and the same elements are labeled with the same numerals. Therefore, this disclosure and the drawings are to be considered illustrative rather than restrictive.
Claims
1. A method for processing data, comprising: Access the mask pattern used in the photolithography process for printing patterns on a wafer; as well as Applying the mask pattern to a deterministic model of the photolithography process via a processor to predict the characteristics of the printed pattern includes: The mask pattern is applied to a compact model that predicts the characteristics of the printed pattern; as well as A correction is applied to the predicted characteristics from the compact model, wherein the correction takes into account local random variations of the characteristics in the printed pattern. The compact model is regressed relative to first empirical data obtained by measuring patterns fabricated on a wafer, and the correction is regressed relative to second empirical data including local random variations of the predicted characteristic.
2. The method according to claim 1, wherein the compact model does not take into account local random variations of the property.
3. The method of claim 1, wherein applying the correction to the predicted property from the compact model comprises (a) a linear combination of the predicted property from the compact model and (b) the correction, wherein the correction is a function of a stochastic model property related to the local stochastic variation of the predicted property.
4. The method of claim 1, wherein applying the correction to the predicted property from the compact model comprises (a) a nonlinear combination of the predicted property from the compact model and (b) the correction, wherein the correction is a function of a stochastic model property related to the local stochastic variation of the predicted property.
5. The method of claim 1, wherein the compact model predicts the characteristics of the printed pattern based on the dosage, focus, and the mask pattern.
6. An electronic system comprising: Memory that stores instructions; as well as A processor, coupled to the memory, executes the instructions, which, when executed, cause the processor to perform operations including: Access the mask pattern used in the photolithography process for printing patterns on a wafer; as well as The mask pattern is applied to a deterministic model of the photolithography process to predict the characteristics of the printed pattern, wherein the deterministic model takes into account local random variations in the characteristics of the printed pattern. The deterministic model includes a compact model that predicts the characteristics of the printed pattern based on dosage, focus, the mask pattern, and random model properties, wherein the random model properties are related to local random variations in the predicted characteristics. The operation further includes regressing the compact model relative to empirical data that does not include the properties of the random model.
7. The electronic system of claim 6, wherein the predicted characteristics of the printed pattern are used for mask synthesis and / or for mask correction.
8. The electronic system of claim 7, wherein the mask pattern comprises a mask pattern for the entire die.
9. The electronic system of claim 6, wherein the local random variation comprises random variation within a 10µm × 10µm area of the printed pattern.
10. The electronic system according to claim 6, wherein: At least one of the following stochastic model properties is: diffuse spatial image signal; acid concentration or density; quencher concentration or density; inhibitor concentration or density; gradient of the spatial image, acid, quencher, or inhibitor signal; mask pattern density; secondary electron concentration; ligand concentration; and resistance to surface tension; and The compact model predicts hotspots in the printed pattern based on the mask pattern.
11. A non-transient computer-readable medium including stored instructions, which, when executed by a processor, cause the processor to perform operations including: Access to mask patterns used in photolithography processes for printing patterns on wafers; and The mask pattern is applied to a deterministic model of the photolithography process to predict the characteristics of the printed pattern, wherein the deterministic model predicts the characteristics of the printed pattern based on model attributes, the model attributes including at least one stochastic model attribute related to local stochastic variations in the predicted characteristics. The deterministic model is regressed relative to empirical data that includes local stochastic variations of the predicted property.
12. The non-transient computer-readable medium of claim 11, wherein the at least one stochastic model property is one of the following: diffuse spatial image signal; acid concentration or density; quencher concentration or density; inhibitor concentration or density; gradient of spatial image, acid, quencher, or inhibitor signal; mask pattern density; secondary electron concentration; ligand concentration; and resistance to surface tension.
13. The non-transient computer-readable medium of claim 11, wherein the empirical data includes one of line edge roughness, critical dimension (CD) variation, and worst-case CD.
14. The non-transient computer-readable medium of claim 11, wherein the deterministic model predicts hot spots in the printed pattern based on the mask pattern.
15. The non-transient computer-readable medium of claim 11, wherein the deterministic model predicts the failure rate in the printed pattern based on the mask pattern.
16. The non-transient computer-readable medium of claim 11, wherein the deterministic model comprises a compact model that predicts the characteristics of the printed pattern based on dose, focus, the mask pattern, and the random model properties.
Citation Information
Patent Citations
Figurative models calibrated to correct errors in process models
US20180322228A1
