Methods and systems for detecting and / or quantifying manufacturing inaccuracies

Abstract

This disclosure relates to a method for detecting and / or quantifying manufacturing inaccuracies caused by photolithography processes. The method includes: providing at least one design for fabricating a structure on a substrate using a set of photolithography processes, wherein the fabricated structure defines an array of metrology sensors, wherein each metrology sensor is adapted to generate one of a known and finite set of different possible physical events when a physical process is applied, wherein the generated physical event is: unknown prior to the application of the physical process, depends on the manufacturing inaccuracy caused by at least one of the set of photolithography processes, and is a displacement state of the fabricated structure, or has one or more physical entities associated with a structure that did not exist prior to the application of the physical process; applying the set of photolithography processes; applying the physical process; and reading out the physical events generated by all the metrology sensors. This disclosure also relates to a metrology system.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to a method for detecting and / or quantifying manufacturing inaccuracies resulting from a lithographic process. The present disclosure also relates to a metrology sensor assembly and a metrology system. BACKGROUND

[0002] Lithographic apparatuses are commonly used in the manufacture of integrated circuits and are employed to apply a desired pattern to a substrate, typically a semiconductor substrate such as Si. The transfer of the pattern is performed by irradiating a radiation-sensitive layer of a material (resist) that is uniformly deposited on the substrate.

[0003] Known lithographic apparatuses include electron beam writers, laser beam writers, steppers, and scanners. Each of these tools translates a computer-aided design (CAD) into a pattern on resist on a substrate, which is an essential part of semiconductor device manufacturing. Electron beam writers use rasterization of a single spot of an electron beam focused to a few nanometers to expose the portion of the resist that will form the features of interest, while scanners use a broad field of light modulated by a mask to achieve selective exposure of the resist. The image of the mask that is projected onto the substrate is reduced by optics such as deep ultraviolet lenses (DUV) or extreme ultraviolet mirrors (EUV). In the case of DUV, it is important that the light is transmitted through the mask, while in EUV the mask is a mirror and it is important that the light is reflected by the mask. In the semiconductor industry, electron beam lithography is primarily used to pattern features on a mask. Scanners use masks to perform lithography steps at high speed and wafer throughput. Masks are also typical substrates on top of which write-based lithography is performed to transfer the design of a pattern to a substrate that will yield a device. There are multi-beam variants of beam writers.

[0004] All lithographic apparatuses suffer from a variety of inaccuracies that introduce small differences between what is designed in the CAD file and the pattern on the substrate. These inaccuracies affect the placement, shape, size, and roughness of the pattern that is produced on the wafer, which ultimately affects the performance and yield of the semiconductor devices. Some of these types of inaccuracies are referred to as edge placement error (EPE) and include overlay (OV; a measure of pattern-to-pattern shift between layers), critical dimension (CD, including local critical dimension LCD and critical dimension uniformity CDU), registration (REG; pattern position relative to an absolute grid), and any roughness types (including line edge roughness LER and line width roughness LWR) and stochastic indices.

[0005] Any EPE present at any level (from sub-device level, device level, system level, die level, up to wafer and lot level) caused by any mask, lithography apparatus, lithography process, and other equipment or process steps involved in the creation of the structure accumulates after several masks to cause potentially large mismatches between the intended design and the actual design on the substrate, which can strongly impact the yield of the manufactured devices. Defects are defined as any physical, electrical, or parametric deviations that can impact yield. Due to the increasing number of process steps and masks and the need for more dense features, maintaining a low number of defects requires increasingly stringent EPE budgets, which require strict metrology, control, and mitigation schemes for EPE at each step and level for each mask and lithography apparatus.

[0006] E-beam lithography suffers from various sources of inaccuracies that introduce EPE, such as imperfect beam focusing and astigmatism, diffraction and spherical and chromatic aberrations, beam deflection hysteresis, stage errors (including stitching, non-orthogonality, and calibration errors), overlay misalignment, registration errors, proximity effects, charge effects, fogging effects, thermal effects, etc., as well as drifts in several beam properties (such as focusing, astigmatism, XY positioning, and beam deflection calibration, etc.) during exposure.

[0007] In addition to EPE originating from the written mask, mask-based lithography using a scanner suffers from various sources of inaccuracies that can further impact EPE, such as shot noise, overlay misalignment, optical proximity effects, phase shift effects, source mask optimization errors, illumination errors, pupil errors, flare effects, shadow effects, imperfect focusing, wafer and reticle warping and thermal expansion, wafer clamping, reticle clamping, pellicle-induced distortions, dose errors, imperfect rastering, stage (reticle and wafer stage) errors, tool matching, etc.

[0008] In addition to inaccuracies that occur during resist exposure on the substrate, a variety of other inaccuracies and defects in the resist can be found due to resist baking and development, resist erosion during dry etching, and local random defects such as microbridges and broken lines, missing contacts and breaks, random effects, material quality, etc. Dry etching also introduces inaccuracies that affect overlay, CD, and LCDU, etc., such as wafer edge effects, non-uniformity and drift in etch rate and etch direction, redeposition, resist hardening, micro-loading, pitch walking, etch bias, etc. Additionally, 3D effects also impact EPE when many layers are stacked on top of each other.

[0009] All of these sources of inaccuracies accumulate after each processing step and introduce various EPEs in the final device, which impact the manufacturing yield and density of the manufactured devices and must be detected, investigated, and minimized to increase manufacturing yield and overall device performance.

[0010] Metrology tools capable of detecting and quantifying any of these errors (EPEs) play a crucial role in process control and risk mitigation in semiconductor device manufacturing. They also play a critical role in research and development activities and yield improvement, aiming to optimize operating conditions of lithography equipment and other equipment involved in semiconductor device manufacturing to maximize yield, reduce costs, increase device density, and improve performance. As the density of nanofabrication features increases and feature sizes decrease, the development of metrology tools for detecting and quantifying these errors at the single-digit nanometer and sub-nanometer levels is essential for determining whether the tools are functioning as intended (monitoring) and, if not, for identifying the source of the error or inaccuracy (troubleshooting).

[0011] There are two main categories of technologies relevant to this context: inspection and metrology. Inspection techniques typically use optical microscopes to search for defects, detecting those at the nanometer, micrometer, or larger (such as particles), while metrology techniques measure the critical dimensions and coordinates (absolute and relative) of specific features or sub-features on a patterned substrate to assess the fidelity of pattern transfer or ensure proper overlap of subsequent patterns with pre-existing patterns. As semiconductor manufacturing technologies continue to advance to increasingly smaller critical dimensions, metrology must be performed on nanoscale features, and it must be possible to achieve extremely high resolution and accuracy at high throughput; otherwise, significant sources of pattern variability may be detected too late in production, resulting in defective chips. It is crucial for the semiconductor industry to be able to measure EPE (Extended Physical Dimension) to perform process control or to utilize any operable process corrections (such as high-order scanner correction) at any tooling or process step, which can improve the final yield and performance of manufactured devices.

[0012] Two broad types of techniques are currently used to perform critical size and overlap measurements: (1) imaging-based (optical or scanning electron microscopy) and (2) scattering-based (scattering measurement, OCD).

[0013] Optical scattering-based metrology (called scattering measurements) derives information about shape, size, and composition from the scattering pattern observed in light that has interacted with a sample. This requires a regular array of similar features, which can be an in-circuit feature (e.g., a line array) or a specially designed measurement target (also known as a scribing) typically located in a region between dies on a wafer. Scattering measurements are typically performed using a dedicated grating structure (a regularly spaced array of lines and spaces) as the measurement target, where the interference pattern of monochromatic light interacting with the grating is analyzed to derive information about critical dimensions and overlap with sub-nanometer precision. Typical metrological applications involve performing a series of scattering measurements at different wavelengths of optical excitation.

[0014] Critical-size scanning electron microscopy (CD-SEM) is a direct measurement technique that works by using computational image processing of scanning electron microscope (SEM) images of patterns (typical examples being the width and spacing of a linear array, or the diameter and spacing of an array of holes). This technique, operating in a vacuum, is often used to measure the critical dimensions of patterns as well as to quantify other types of significant lithographic defects, such as random errors (missing holes, broken lines, etc.) and LER (Last-Rate Ergonomics). Optical microscopes using wavelength-tunable light sources with specific design goals are used to perform overlap metrology in a technique called image-based overlap.

[0015] Scattering measurements and other OCD techniques require specialized optical equipment, including coherent wavelength-tunable light sources and spectrometers, to operate. They typically operate at a single point rather than over an area. Some of these techniques also require special types of grating structures at certain micrometer dimensions, which limits their practicality for device metrology. CD-SEM typically operates on capital-intensive equipment and has very low throughput due to the extremely high magnification required to resolve sub-nanometer features. Furthermore, each of these techniques suffers from inherent noise, inaccuracies or origin errors in the imaging and detection components of the instrument, instrument calibration, and calculations that lead to the final metrological data (e.g., image processing, modeling, and fitting).

[0016] In summary, available metering techniques are low-throughput, require complex and expensive tools, and rely on sophisticated data analysis, which limits the frequency and density of metering performed for process control and monitoring. Summary of the Invention

[0017] Photolithography systems inevitably introduce errors in the placement and manner in which the edges of patterns are positioned on the substrate, known as edge placement errors (EPEs). The detection and quantification of EPEs are critical because they undesirably affect all micron- and nanon-fabricated structures and have historically been a major cause of yield problems in the semiconductor industry. This disclosure describes a novel metrology method that binarizes EPEs as discrete, identifiable physical events on the substrate being lithographically processed. This method is compatible with all photolithography systems and independent of all other existing metrology techniques, expanding the toolbox available for sensing and quantifying errors at the lithography origin, in addition to SEM, TEM, AFM, and diffraction-based techniques.

[0018] This disclosure relates to methods and systems for addressing at least some of the aforementioned limitations.

[0019] According to a first embodiment, this disclosure relates to a method for detecting and / or quantifying manufacturing inaccuracies caused by a photolithography process, comprising:

[0020] Provided is at least one design for fabricating a structure on a substrate using a set of photolithography processes, wherein the fabricated structure defines an array of metrology sensors, wherein each metrology sensor is adapted to generate one of a known and finite set of different possible physical events when a physical process is applied, wherein the generated physical events are:

[0021] It was unknown before the physical process was applied.

[0022] Depending on the manufacturing inaccuracies caused by at least one of the set of photolithography processes, and

[0023] It is the displaced state of the manufactured structure, or has one or more physical entities associated with the manufactured structure that did not exist before the application of the physical process;

[0024] The fabricated structure is obtained by applying the aforementioned set of photolithography processes;

[0025] By applying the physical process, one of the known and limited set of different possible physical events is generated for each metering sensor;

[0026] Read out all physical events generated by the metering sensors; and

[0027] Process all physical events generated by the metrology sensors to detect and / or quantify manufacturing inaccuracies caused by the photolithography process.

[0028] Those skilled in the art are familiar with lithography systems and processes and will be able to perform examples of such processes mentioned in this disclosure.

[0029] This disclosure also relates to a method for detecting and / or quantifying manufacturing inaccuracies caused by photolithography equipment, comprising the following steps:

[0030] Provided at least one pattern for fabricating a structure on a substrate, the at least one pattern defining an array of metering sensors, each metering sensor including a displaceable material configured to be in a stationary state and distributed on a local region, wherein the displaceable material is adapted to reach a displaced state toward a predetermined discrete position within the local region during displacement when a physical stress source is applied.

[0031] The fabricated structure on the substrate is obtained from the photolithography equipment, wherein the displaceable material of the metering sensor is in a stationary state;

[0032] Apply the physical stress source to each metering sensor for at least a predetermined time period; and

[0033] The metering sensor is read out to detect and / or quantify manufacturing inaccuracies caused by the lithography equipment.

[0034] By imaging the metrology sensor described in the final step of the previously disclosed method, metrology related to semiconductor manufacturing can be performed at nanometer resolution using a basic optical microscope. This is a cost- and time-efficient method compared to existing metrology techniques previously described. The disclosed method can also be configured to generate a new metrology data stream independent of all other metrology techniques previously described, which can be used to efficiently detect, study, control, and minimize any of the aforementioned sources of EPE in masks, lithography equipment, etching and deposition processes, and tooling.

[0035] The disclosed method can be used in the scanner authentication process. It can be further used in the patterning control process.

[0036] Advantageously, this method can be described as conducting nanoscale experiments, each of which is a displacement of a substance (liquid or solid) to its displaced state, or the creation of one or more physical entities, the number or placement of which is sensitive to the lithographic pattern and its errors. The geometry and size of the lithographic pattern can be considered as an analog signal containing the EPE, which will be experimentally processed to detect the EPE. The displacement of the substance or the number or placement of the physical entities constitutes the experimental result generated by each metrology sensor and depends on the lithographic pattern and its deviation from the ideal design in binary (when the set of all possible physical events equals 2) or discrete (when the set of all possible physical events is greater than 2) signal form. After performing the different steps of this method, equivalent to an analog-to-digital converter, each metrology sensor contains data in binary or discrete format about the inaccuracies introduced by the entire lithography process, including the contribution of the lithography equipment. When one of these lithography processes has a dominant contribution to the metrology sensor, the metrology sensor contains data about that particular lithography process. By fabricating at least one pattern as described and applying physical processes, this method can be said to replace computer-based data processing with multiple parallel physical processes that perform analog-to-digital conversion on the substrate. This method enables the selective use of imaging to collect data generated by metrology sensors into a computer. As long as the resolution and signal-to-noise ratio of the imaging system (determined by magnification, illumination intensity, camera, and exposure settings used, etc.) are sufficient to detect the discrete data generated by the metrology sensors on the substrate, the metrology data becomes independent of the imaging technique, its associated calibration, and computer-based data processing.

[0037] In the currently disclosed method, information related to EPE is advantageously processed and binarized on the substrate by means of experiments on the pattern sensitivity on the substrate defined by the photolithography process and equipment. Imaging can be used to collect the already discretized data on the substrate for statistical analysis in a computer. Therefore, the described method can be described as performing EPE measurement before data digitization in a computer.

[0038] In this method, EPE data processing can therefore be considered to precede the digitization of data in imaging or computing. This contrasts with conventional metrology techniques, where EPE data processing is performed algorithmically or using machine learning methods after the data has been digitized into a computer, and thus follows the imaging or inspection steps. This method generates universally readable metrological data, with no additional source of error or inaccuracy introduced into the metrological data after the binarization experiment involving the displaceable material or the creation of the physical entity has occurred on the substrate. The only requirement is that the result of each binarization experiment can be correctly registered to the corresponding pattern in the original design. This contrasts sharply with conventional metrology techniques, which are susceptible to various errors and inaccuracies introduced by sensing / imaging systems as well as computational modeling and analysis. The final measurement extracted depends on the steps involved in the creation of the metrological data, from the arrangement of the excitation and imaging systems, the digitization process to computer-based data processing, such as individual components (e.g., lenses and detectors), calibration methods, measurement conditions and parameters, and models and algorithms. For example, unlike this approach, model-based measurement techniques (such as interferometry and scattering measurements) use experimental data and simulations from fundamental physics to extract quantitative estimates of the measured variables of a sample based on a parametric model of that sample. The uncertainty of these estimates is based not only on the uncertainty of the experimental data, but also on the data's sensitivity to model parameters, the parametric correlations between those parameters, and fit, etc.

[0039] In this disclosure, EPE measurement data written on a chip in a stable physical bitmap format can be collected by imaging at a spatial resolution much lower than the length scale of the error, because they are encoded and binarized in a much larger number of physical bits. Conceptually, this is similar to the principle of barcodes: data is encoded as bit strings and presented in a simple, machine-readable visual format. The method of this disclosure is similar to printing two-dimensional barcodes (such as QR codes) on paper with ink, where the actual printed barcode depends on errors generated by the printing press and the ink is only displayed after a specific process (in our case, a stress source, or more generally, a physical process) is applied. The deviation of the displayed barcode from the expected barcode can be easily determined and used to calculate the EPE. In this analogy, the printing press is a lithographic printing tool, the paper is a semiconductor substrate with one or more layers of material that can be patterned, and the ink pattern encoding information is the physical bits generated by binarization experiments on each metrology sensor. Although the type of encoded EPE information (CDU, overlap, etc.) is determined by the design of a specific type of pattern printed on the substrate, the form of the physical bits remains the same for the same combination of substrate and physical process. In other words, the encoding and printing process of the barcode remains the same, while allowing the encoding of different types of EPE information. Although physical bits can be larger than 200 nm and therefore can be read with a basic optical microscope (such as those found in some wafer inspection equipment), the data stored in the bitmap can encode metrological data about the EPE with an accuracy better than 10 nm (i.e., even EPEs of 10 nm and smaller sizes can be detected and quantified using this method). Furthermore, as long as the readout technology can distinguish individual physical bits and register them to the pattern in the design used to generate them, the entire metrological data can be retrieved without error, regardless of the imaging technique or microscope used, provided that appropriate instructions are known or given to correctly interpret the bit string. Since the entire metrological data resides in the physical bits on the substrate, the lack of improvement in calibration, magnification, or contrast when imaging the physical bits will improve measurement accuracy and precision: the measurement data becomes free from photon and electron bias or origin noise in image acquisition. Different microscopes imaging the same bitmap will be able to read exactly the same metrological data. In this way, it can be said that the measurement of the EPE is performed before imaging. Imaging only requires reading or digitizing already processed metrological data and processing it into qualitative or quantitative measurements.

[0040] The metrological data generated by this method is independent of many limitations that typically plague optical techniques. For example, in this invention, optical microscopes, whose resolution is typically limited by the wavelength of light and the lenses used, can sense localized EPE at the sub-10 nm level on a single isolated structure. This is because if a metrological sensor is designed in a way that we expect to pinpoint, but after binarization we find different positions, then we know that a patterned error has occurred in that particular metrological sensor.

[0041] The metrological data generated by this method is free from many limitations that typically plague CD-SEM and CD-AFM. For example, in this invention, complex image processing is not required to obtain information about the size and shape of the pattern from the image, as is the case in CD-SEM and CD-AFM. In this invention, only simple image processing is needed, which can detect binary or discrete data from the metrological sensor in a known format. Furthermore, in this invention, the size of the pixels used for imaging only needs to be large enough to resolve the binary or discrete data from the metrological sensor, and can be several orders of magnitude larger than the pixels in CD-SEM and CD-AFM that need to resolve the nanoscale features of interest. Therefore, in the present case, the effective physical size of each imaging pixel can be several orders of magnitude larger than the length scale for resolving EPE information. This results in significant savings in data acquisition throughput, data storage, and data processing. Moreover, due to compatibility with high-throughput inspection optical microscopes or wafer inspection tools used in semiconductor foundries, the detection and measurement of nanoscale or sub-nanometer EPEs can be performed at the wafer level. Using this method, wafer-level features of nanoscale EPEs can therefore be captured with high wafer throughput. Furthermore, using the method described herein, different imaging systems will be able to read the same metrological data from the same EPE measurement without added noise or bias. This contrasts sharply with all other metrological techniques, where measurement data is inevitably sensitive to all downstream hardware components involved in the digitization of the pattern information containing the EPE and the software layer that computes the measurement data from the digitized data. Such hardware may include illumination sources, filters, lenses, beam grating systems, mechanical stages, CMOS sensors, etc., and the software may include physical equations, physical constants, CAD, mesh generation, edge detection algorithms, parameterization, fitting, etc. In other metrological techniques, any systematic or random error introduced by any of these components or steps affects the accuracy and precision of the final metrological data. Thus, if not properly calibrated, two different metrological tools can produce significantly different measurements with the same error. Two different image processing algorithms or physical models can produce different measurements from the same raw digitized data. Aberrations caused by tool optics are arguably the primary source of systematic error in image-based metrology, which our method is insensitive to in almost the same way as barcode readers.

[0042] The ability to generate data independent of other metrology techniques and their inherent errors strongly supports hybrid metrology and can be used to enhance, monitor, or calibrate the execution of any existing metrology technique used in conjunction with photolithography patterning and photolithography processes.

[0043] In one embodiment, the metering sensor includes a plurality of mechanical actuators connected by at least one connecting element in a strain state representing the rest state, wherein each mechanical actuator is adapted to trigger mechanical actuation at the start of the etching process to reach a final state within a predetermined time period, and wherein when one of the mechanical actuators reaches its final state, each connecting element reaches an unstrained state representing the displacement state, and wherein the step of simultaneously applying the physical stress source to each metering sensor for at least a predetermined time period includes simultaneously etching an array of the metering sensors for at least the predetermined time period.

[0044] In this embodiment, the expected result of actuating the mechanical actuator is known because the actuator is designed to behave in a specific way. The mechanical actuator can be thought of as a timer. When the first timer reaches its end state, it can trigger mechanical actuation, which will put the connecting element in a displaced state. When mechanical actuation is triggered, the connecting element can physically displace toward one of the mechanical actuators (the one that is not actuated or actuated later than the other). In this state, an optical microscope can provide an image covering the entire area of ​​multiple metrology sensors, thereby obtaining immediate detailed information about manufacturing inaccuracies caused by the lithography equipment by analyzing deviations from the expected results.

[0045] This disclosure also relates to a metering sensor assembly, which includes:

[0046] Substrate; and

[0047] An array of metering sensors, each metering sensor comprising a plurality of mechanical actuators connected by at least one connecting element in a strained state, wherein each mechanical actuator is adapted to trigger mechanical actuation at the start of an etching process to reach a finished state within a predetermined time period, and wherein each connecting element reaches an unstrained state when one of the mechanical actuators reaches its finished state.

[0048] As those skilled in the art will recognize, any embodiment of the currently disclosed metrology sensor assembly can be used in any embodiment of the currently disclosed methods to detect and / or quantify manufacturing inaccuracies caused by lithography equipment, and vice versa.

[0049] This disclosure also relates to a measurement system comprising:

[0050] A photolithography apparatus configured to pattern a radiation-sensitive resist, such as a polymer, on a substrate to fabricate a structure using an array of metering sensors, wherein each metering sensor is adapted to generate one of a known and finite set of different possible physical events when a physical process is applied, wherein the generated physical events are:

[0051] It was unknown before the physical process was applied.

[0052] Depending on the manufacturing inaccuracies caused by the lithography equipment, and

[0053] It is the displaced state of the manufactured structure, or has one or more physical entities associated with the manufactured structure that did not exist before the application of the physical process;

[0054] A system adapted to apply the physical process to each metering sensor.

[0055] This disclosure also relates to a measurement system comprising:

[0056] A photolithography apparatus configured to pattern a radiation-sensitive resist, such as a polymer, on a substrate using an array of metering sensors, each metering sensor including a displaceable material configured to be in a stationary state and distributed over a local area, wherein the displaceable material is adapted to reach a displaced state toward a predetermined discrete position within the local area during displacement when a physical stress source is applied.

[0057] Physical stress source devices, such as etching apparatuses or systems, are adapted to apply physical stress for at least a predetermined period of time; and

[0058] An imaging device for imaging the metrology sensor to detect and / or quantify manufacturing inaccuracies caused by the lithography equipment.

[0059] Advantageously, by using the metrology system described above, the metrology information obtained from the metrology sensor can be derived by fast image recognition processing (such as template matching) applied to a subset of pixels in the total imaging area, due to prior knowledge of the position of the metrology sensor on the surface of the patterned substrate being observed. Attached Figure Description

[0060] The invention will now be described with reference to the accompanying drawings. The drawings are examples of embodiments and are not limited to the currently disclosed metrology sensor assemblies, metrology systems, and methods for detecting and / or quantifying manufacturing inaccuracies produced by lithography equipment. As an example, metrology sensors are generally considered suitable for generating one of a known and limited set of different possible physical events during the application of physical processes. Examples of mechanical actuators may be shown in the drawings.

[0061] Figures 1A-1C Examples of metering sensors in different states are shown.

[0062] Figures 2A-2D An embodiment of a currently disclosed metering sensor using a mechanical actuator is shown.

[0063] Figure 3 Another embodiment of the currently disclosed metering sensor is shown.

[0064] Figures 4A-4B An example of an array of metering sensors distributed on a substrate after the application of physical processes is shown, along with the resulting bitmap transformation of the binarized physical events after the application of Boolean logic.

[0065] Figures 5A-5H An embodiment of a currently disclosed metering sensor is shown, wherein the displaceable material is composed of a liquid or gel.

[0066] Figure 6 The illustration shows how a timer used in a currently disclosed metering sensor can be represented using a mechanical actuator or end region of gel or liquid.

[0067] Figures 7A-7B Further illustrations are shown of how timers can be used to obtain physical records or more generally physical events on a substrate, which can be used to detect and / or quantify manufacturing inaccuracies caused by photolithography processes.

[0068] Figures 8A-8B An example of a pattern cut in two separate but complementary sections is shown, which can be made into a complete metering sensor for detecting misalignment between two separate exposures.

[0069] Figure 9 An example flowchart of a currently disclosed method for detecting and / or quantifying manufacturing inaccuracies caused by photolithography processes is shown. Detailed Implementation

[0070] This disclosure relates to a method for detecting and / or quantifying manufacturing inaccuracies caused by photolithography processes. Manufacturing inaccuracies can be, for example, edge placement errors, including CD, CDU, LCDU, overlap, registration, roughness, and randomness index.

[0071] Preferably, in the first step, at least one design is provided for fabricating a structure on a substrate using a set of photolithography processes. The fabricated structure may define an array of metrology sensors, each of which is adapted to generate one of a known and finite set of different physical events when a physical process is applied. The generated physical event may be unknown before the physical process is applied, depending on the manufacturing inaccuracies generated by at least one of the set of photolithography processes, and may be a displaced state of the fabricated structure, or have one or more physical entities associated with the fabricated structure that did not exist before the physical process is applied. The set of photolithography processes can then be applied to obtain the fabricated structure. Physical processes can then be applied to generate one of a known and finite set of different physical events for each metrology sensor. The generated physical events of all metrology sensors can then be read out, and the generated physical events of all metrology sensors can be processed to detect and / or quantify the manufacturing inaccuracies generated by the photolithography process. The generated physical events can be stored digitally. Imaging, for example, can be used to read out the generated physical events of all metrology sensors.

[0072] The metering sensor can be configured such that the generated physical event is suitably larger than the inaccuracy itself. In this way, the inaccuracy can be read indirectly, for example, by imaging at a spatial resolution much lower than the length scale of the error. The manufacturing inaccuracy can originate from edge placement errors, where the edge placement error is less than 5 nm, or where the edge placement error is less than 1 nm.

[0073] Photolithography generally involves transferring a design pattern onto a substrate using, for example, an exposure tool and subsequent resist development. In a single run and without prior knowledge of metrology, the manufacturing inaccuracies caused by the entire photolithography process can be determined. When used holistically in conjunction with other metrology tools, or with prior knowledge of the relative magnitudes of different error sources contributing to manufacturing inaccuracies, the exact contribution of a particular equipment or process step can be quantified. In this way, specific photolithography processes and physical procedures can be performed such that the contribution of the exposure tool to the overall manufacturing inaccuracies is dominant. Similarly, when using highly calibrated and stable exposure tools, the inaccuracies caused by dry or wet etching processes can be quantified. In one embodiment, this set of photolithography processes includes at least exposure and development of a radiation-sensitive resist.

[0074] According to one example, a photolithography apparatus provides at least one photolithographic pattern for fabricating a structure on a substrate, defining an array of metrology sensors. In this example, each metrology sensor includes a displaceable material configured to be in a static state and distributed over a local area, wherein the displaceable material is adapted to reach a displaced state toward a predetermined discrete location within the local area during a displacement process when a physical stress source is applied. The physical stress source can be an etching process or any suitable physical process that causes the displaceable material to reach the displaced state, including, for example, changing the temperature or phase, or applying vibration or evaporation. The displaceable material of the fabricated metrology sensors is initially fabricated in a static state. In another step, a physical stress source is applied to each metrology sensor for at least a predetermined time period. In yet another step, the metrology sensors are imaged to collect data for detecting and quantifying any inaccuracies generated by the photolithography apparatus.

[0075] Photolithography equipment can be an electron beam writer, a laser beam writer, a nanoimprinter, a stepper, a scanner, or other equipment used in the field of device nanofabrication. Photolithography equipment is not necessarily an exposure tool; it can also be a pattern transfer tool, such as a tool for applying dry or wet etching. The photolithography process should be interpreted broadly to encompass any suitable use of one or more steps using a photolithography equipment. An advantage of the currently disclosed methods, systems, and sensors is that any photolithography equipment can be used, and therefore, metrological data from any photolithography equipment can be generated. The technology is not dependent on a specific photolithography equipment. It can be said that the currently disclosed methods and systems can transform any photolithography equipment into a metrology device that generates metrological data about itself, thereby eliminating the need for any metrology tools other than itself to generate the metrological data; that is, the metrological data is generated solely using the photolithography process. By moving data processing to a substrate, imaging such as that of an optical microscope can be used to evaluate any photolithography equipment or the operation of a photolithography equipment. The substrate on which the structure is placed can be a rigid semiconductor substrate, such as silicon or germanium, a metal (such as aluminum or gold), an oxide (such as SiO2 or sapphire or quartz), or any substrate used in the fields of device nanofabrication and semiconductor manufacturing.

[0076] In the context of this disclosure, the structure manufactured may be a resist, a polymer, a solid, a liquid, a gel, or a stack or a combination of these materials.

[0077] Physical events can be changes in the optical properties of the structure or an added visible entity, changes in shape, changes in size, changes in placement in the XY or Z directions, buckling / bending patterns, etc. They may originate from the removal or addition of material. The actual type of physical event is known to the operator before the physical process is applied and depends on the pattern, substrate, and physical process. Different event types can be used individually or simultaneously, and the same combination of pattern, substrate, and physical process can trigger more than one type of physical event. As described above, each metrology sensor is adapted to generate one of a known and limited set of different possible physical events when the physical process is applied. Examples of physical events include events where the structure in the metrology sensor does not generate a visible entity in the vicinity of a structure that was not present before the physical process was applied, generates one visible entity, or generates several visible entities, such as light scattering points, such as cracks, holes, or particles. Other examples of physical events include visible and discretizable changes in the structure itself or another structure derived from it, such as shrinkage-expansion of area, left or right movement, clockwise or counterclockwise twisting, upward or downward buckling, etc. In one embodiment, the physical event is a change in the shape, size, or position of a portion or all of the structure, preferably wherein the change in the shape, size, or position of a portion or all of the structure is greater than 10 nm.

[0078] Physical processes in the context of this disclosure may include mechanically applying stress, releasing internal stress, spin coating, selective deposition, selective exposure, selective etching, heating, freezing, applying stress by bending a substrate, displacing a dual piezoelectric wafer, exposure to radiation of a specific wavelength, ultrasonic or mega-frequency ultrasonic vibration, and one or more combinations of the above physical processes, such as selective etching followed by selective deposition.

[0079] The process of reading out the physical events generated by all metrological sensors can be accomplished in several ways. Readout can involve imaging-based readout, discrete sampling, including, for example, lasers used in the same manner as barcode readers, scattering measurements, or SEMs used in local “probe” mode.

[0080] In one embodiment of the method of this disclosure for detecting and / or quantifying manufacturing inaccuracies arising from photolithography processes, the resulting physical event is clearly distinguishable from other physical events in a known and limited set of possible distinct physical events. This can be achieved, for example, by having a metrology sensor comprising a displaceable material configured to be stationary and distributed over a localized region, wherein the displaceable material is adapted to reach a displaced state within the localized region during the displacement process when the physical process is applied. The displaceable material may be stationary when the fabricated structure on the substrate is obtained from the photolithography apparatus. As described above, the physical process can involve many techniques. In one embodiment, the physical process is applied to each metrology sensor for a predetermined time period.

[0081] The steps for processing generated physical events may include computationally processing the generated physical events. For example, processing may include applying Boolean logic to generated physical events from at least one design based on expected physical event outcomes, and / or constructing probability distributions of the counts of generated physical events for one or more varying design parameters of the design and comparing them with expected probability distributions based on the nominal design.

[0082] The steps for reading out generated physical events may include searching for physical events only at predetermined locations on the manufactured structure.

[0083] In one embodiment of the method of this disclosure for detecting and / or quantifying manufacturing inaccuracies caused by photolithography processes, at least one design is configured such that a particular physical event is advantageous. This can be accomplished, for example, by providing two or more designs, wherein at least one of the designs can be distinguished from another design by making at least one edge-shifting lithography apparatus an integer multiple of the minimum controllable step size. Alternatively, at least one design can be configured such that a known and finite set of different physical events are not advantageous.

[0084] Figure 1A A non-limiting conceptual example of a metering sensor 100 in a stationary state is shown. The figure illustrates two possible displacement states 104 and 105, corresponding to different possible physical events. The displaceable substance 101 is in a stationary state. From this position, physical processes can be applied to move the displaceable substance 101 in predetermined directions 102 and 103 to reach one of the two possible displacement states 104 and 105. Figure 1B and Figure 1C A metering sensor 100 is shown, wherein the displaceable substance 101 is in each of the possible end states 104 and 105.

[0085] As described above, the step of reading out all the physical events generated by the metering sensors can be accomplished, for example, by imaging. Imaging the metering sensors after the physical process has been applied to displace the displaceable substance allows the location of the displaceable substance to be identified in the final state. This imaging includes, preferably, determining the location of the connecting elements by means of an optical microscope. Other microscopy techniques, such as electron, ion, ultraviolet, or X-ray microscopy, can also be used for the imaging step.

[0086] In embodiments, the material including the mechanical actuator is isotropic or anisotropically etchable when in contact with an etchant in a wet or dry etching process. It is understood that isotropically etchable materials are materials etched at a constant rate, regardless of properties such as size, crystal orientation, crystal polarity, or roughness. For example, alumina is an isotropically etchable material in contact with liquid HF or aluminum etchant. It is understood that anisotropically etchable materials are materials etched at a rate dependent on crystal properties such as crystal orientation, crystal polarity, or roughness. For example, specific crystal faces in crystalline Si can be etched at a much slower rate with certain etchants, such as the (111) face in contact with KOH or TMAH.

[0087] Figure 2A An embodiment of a metering sensor 200 is shown, which includes a mechanical actuator in the form of vertical nanopillars 202 and 203 disposed on a substrate 204. The mechanical actuator is fixed to the rigid substrate 204 and the connecting element 201, as shown. Figure 2B As shown in the diagram, the connecting element 201 that connects the two mechanical actuators includes a rigid platform made of a material that does not react with the etchant, thereby allowing the mechanical actuators to be selectively etched away.

[0088] initial, Figure 2A and Figure 2B The connecting element 201 is in a pre-strained state and is connected to the mechanical actuators 202 and 203. Such a pre-strained state can be induced, for example, by manufacturing the connecting element at a temperature above room temperature and by making the connecting element a material having a higher coefficient of thermal expansion than that of the substrate. Figure 2C The results are shown after an etching process removes at least one of the mechanical actuators of the metering sensor 210, and the connecting element 201 is preferably made of a material that is almost insensitive to the etchant. When a physical stress source is applied to the metering sensor 210, the connecting element shifts a distance 211 from the strained state toward the remaining mechanical actuator 212 of the metering sensor to an unstrained position, as shown. Figure 2CAs shown in the diagram. The driving force for the displacement of the connecting element can be, for example, tensile stress, such as thermal strain generated in the connecting element during the manufacturing process. The termination state of the mechanical actuator corresponds to the partial or complete etching of the mechanical actuator to the point where it releases the connecting element it is connected to. This can include the physical separation of the connecting element from the mechanical actuator due to high tensile force before complete release. The mechanical actuator is preferably etched laterally and isotropically in a wet etching process, preferably by an etchant (e.g., hydrofluoric acid or aluminum etchant). Depending on the material composition of the nanopillars, other wet etchants such as KOH, TMAH, H2SO4, or piranha etchant can be used. The etching process can be stopped at a desired time by drying or immersing the substrate in an etching-stopping liquid (e.g., deionized water).

[0089] In another embodiment, the connecting element (which may alternatively be referred to as a beam) is a pre-strained element connected to at least two mechanical actuators within the metering sensor. The connecting element may be positioned pre-strained on top of the actuators. When the time required for the first actuator to activate has elapsed, the connecting element is released from the substrate, and the tensile stress pulls it toward another actuator that has not yet elapsed and thus remains attached to the substrate during displacement. The displacement experienced by the connecting element from the actuator that has elapsed first leaves a physical record on the substrate and simultaneously releases stress in the metering sensor, preventing the other actuator from creating its own physical record once it has elapsed. The displacement of the connecting element from the actuator that has elapsed first can later be detected, allowing it to be known which actuator elapsed first. When the stress is compressive, the displacement direction can be reversed. When the connecting element has a stress gradient or is composed of dual piezoelectric wafers, the displacement direction can be out-of-plane. Therefore, this disclosure describes a photolithography method for defining “timers” on a chip and creating a system that makes them mutually exclusive, such that in two or more connected actuators, only the actuator that disappears first can generate a displacement from its predetermined position and thus leave a physical record, while the other connected actuators will remain in place even after they eventually disappear.

[0090] Physical events can be fractures caused by a fracture mechanism resulting from the movement of displaceable material from a resting state to a displaced state. Fractures can be generated by a mechanical actuator being pulled by a connecting element. This pulling may cause tearing in the attached layer, leaving a fracture gap in that layer. This fracture gap can be used for selective etching of the underlying layer, forming optically observable pits. The fracture gap can also be used for selective growth of material using the underlying layer as a seed material, resulting in optically observable particles. Therefore, the resulting fracture gaps can be used to selectively and spatially amplify data about the displaced state, which simplifies data collection via imaging. Such gaps, pits, and particles are non-limiting examples of physical entities generated in association with a manufactured structure that did not exist prior to the application of the physical process. Physical entities generated in association with the manufactured structure and the displaced state of the manufactured structure are examples of a known and limited set of different possible physical events triggered when the physical process is applied.

[0091] One-dimensional displacement is achieved when two mechanical actuators are connected to a single connecting element metering device. However, multiple actuators can be connected to multiple connecting elements, thereby allowing physical displacement in any direction within the plane of the recording substrate surface.

[0092] In another embodiment, the displaceable material of each metering sensor includes a liquid or gel, wherein the liquid or gel is displaced to a displaced state when a physical stress source is applied, and wherein the step of simultaneously applying the physical stress source to each metering sensor for at least a predetermined time period includes simultaneously changing the temperature or phase of the liquid or gel or the substrate, or applying vibration or evaporation.

[0093] Figure 5 illustrates a schematic example of a metering sensor 500, wherein the displaceable material includes a liquid or gel. The sensor is photolithographically defined as 500, as shown below. Figure 5A As shown in the top view, it is arranged in two end regions 501 and a connecting region 502 connecting the two end regions 501. "End regions" and "connecting regions" should be interpreted broadly to include any suitable shape having two ends toward which the liquid or gel can displace. This can include, for example, a single line region. Figure 5B It shows Figure 5A A side view of the sensor, which is defined on the substrate by photolithography. Figure 5C A defined sensor 500 is shown, in which a liquid or gel 502 is deposited on top of the sensor and occupies all available space. Figure 5D yes Figure 5C The sensor and the liquid or gel shown are shown in a side view. Figure 5EThe diagram illustrates the shape of a defined sensor with a liquid or gel 502 deposited on top, allowing the liquid or gel to shift toward one of the end regions during the application of a physical process 504. The physical process can be, for example, a drying process, temperature change, phase transition, or vibration. Physical stress sources can cause tension or instability in the liquid or gel. Figure 5F yes Figure 5E A side view. The application of a physical stress source forces a liquid or gel to move toward one of its end regions, typically driven by the minimization of the liquid or gel's surface energy. This effectively creates an observable physical record of the displacement. Figure 5G The diagram shows the displacement of liquid or gel 502 toward the left end region 505, while the right end region 506 is free of liquid or gel due to the displacement. The liquid or gel can be adapted to return to a static state via physical processes (e.g., increasing the water content in the air or sonication) upon release or deactivation of the physical stress source. This effectively allows for multiple reuses of the metering device by resetting it to its initial state. Figure 5H It shows Figure 5G Side view.

[0094] Preferably, in embodiments where the displaceable material is composed of a liquid, gel, or solid substance, the displaceable material can be displaced in a plane on the surface of the substrate. Furthermore, the displacement of the material can create an observable physical record in the displaced state. This can effectively translate nanodefects into changes in the presence or absence of observable structures or structures larger than 100 nm, which can be easily detected optically.

[0095] In the case of mechanical actuators, each mechanical actuator or every two end regions of the gel or liquid can represent a timer. Each timer can define the expected trigger time for the mechanical actuator or end region to reach the displaced state when a physical stress source is applied. The concept of a "timer" can be described as follows: Figure 6 As shown in the example. Figure 6 In the process, an actuator is etched. The etching process begins at a given time. The actuator is etched laterally. Since the etching rate and the width of the actuator are known, an expected time for etching the entire actuator can be provided. Figure 7 shows another illustration of how a timer can be used to obtain a physical record on the substrate, which can be used to detect and / or quantify manufacturing inaccuracies caused by the photolithography equipment. Figure 7A In this process, the two actuators 202 and 203 are designed such that 203 has a larger width, and therefore its "timer" has a larger predetermined timer value. Therefore, the expected result of the etching process is that actuator 202 reaches its final state before actuator 203, as... Figure 7B As shown in the diagram. When the connecting element 201 reaches its unstrained state, it creates an observable physical record.

[0096] The application of a physical stress source places the metrology sensor in a state where the displaceable material is displaced toward a mechanical actuator with a longer trigger time, thereby effectively imprinting the finished state of the metrology sensor after actuation. The displacement direction can be reversed depending on the physics of the experiment, for example, in the case of compressive stress on the displaceable material. The displacement direction can be in the plane of the substrate, out of the plane, or a combination of both. A single metrology sensor includes at least two actuators, with the possibility of adding additional actuators in different directions on the substrate surface, allowing the fabrication of metrology sensors in which the displaceable material can be displaced in two dimensions of the surface.

[0097] Figure 3 A diagram of a metering sensor 300 is shown, in which two actuators 302 and 303 are designed with different widths and therefore have a predetermined trigger time difference upon activation. Defining actuators with different trigger times allows for the quantification of manufacturing inaccuracies in the metering sensor. Due to the higher trigger time of the right actuator 303, the displaceable material 301 will preferentially move to a right-shifted state 305 upon actuator activation, rather than to a left-shifted state 304. If the shifted state is to the left, it can be known that lithographic errors in the metering sensor are caused by the lithography system overcoming the trigger time deviation, which can be used to quantify spatially generated errors. Defining actuators with progressively smaller differences until they are identical allows for the identification of unknown physical deviations, such as edge placement errors caused by the lithography equipment during the manufacturing process. In one embodiment, each mechanical actuator or end region is designed such that its size is proportional to the predetermined amount of time taken to reach the end state (corresponding to the expected trigger time of the actuator).

[0098] Figure 4A An example of an array 400 of binary metrology sensors 401 distributed on a substrate is shown. This array comprises metrology sensors arranged in an M×N matrix configuration, where M≥2 and N≥2. The pattern defines an array of at least 10 metrology sensors, allowing measurement of the final states of several metrology sensors to collect information on lithography equipment inaccuracies until statistically significant results are obtained. Within the same array, different biases can be designed on each sensor, allowing for the quantification and testing of the precision limits of lithography. Ideally, the precision limits of the lithography equipment are achieved when the bias difference between actuators is sub-1 nm, thereby giving the sensors a random shift upon activation. An array can refer to a collection of discrete metrology sensors that are not necessarily distributed over a continuous region and are not necessarily arranged in any particular periodic or aperiodic pattern, with or without other intermediate structures. Figure 4BA bitmap or matrix barcode representation of the array 400 of the metering sensor 401 is shown. Similar to a barcode reader that reads a matrix barcode, any optical system capable of distinguishing the final state of the metering sensor can read all the encoded data therein, provided that the appropriate instructions to interpret the bit strings are known or given.

[0099] The metrology sensors disclosed herein can be run in parallel multiple times on a substrate, such as at least 5 or at least 10 times, by patterning many metrology sensors and simultaneously activating all sensors. The number of runs on a single substrate can exceed 100, one million, or even up to one billion or more. By detecting the results of multiple experiments, statistics that can be used to measure the inaccuracies of the lithographic patterning can be obtained. This is because it is known which actuator will vanish first, as their preset times are lithographically designed based on their dimensions. If the statistics deviate from the expected results, the actuators have preset times that are slightly different from the design, indicating that the tooling has introduced errors in lithography. For example, as shown in the first row of devices in Figure 4, a sensor containing actuators with the same preset times can be designed such that the probability of each actuator vanishing first is expected to be 50%. In this example, 50% of the actuators are shown as being in the up or 50% in the down final state. If an imbalance is detected in the statistics, where, for example, one of the actuators vanishes first with an 80% probability, it can be understood that the lithography apparatus has introduced a bias, causing one of the actuators to have a larger island than the other.

[0100] The disclosed metrology sensor can also be configured to not only sense but also acquire quantitative measurements of inaccuracies introduced by photolithography equipment, such as edge placement errors. This is accomplished by patterning a sensor with a known bias over a preset time, or equivalently using... Figure 6 This is accomplished using a timer-based implementation (with known size bias). For example, patterning a sensor with a 4nm bias between end states and observing one side vanishing first with 100% probability means that the accumulated error in the tooling is less than 4nm. Data from many sensors with different known nanometer or sub-nanometer biases can be compiled to compile statistical output curves that can evaluate the noise floor of the metrology technique under given process conditions involved in the construction of individual metrology sensors.

[0101] Publicly available metrology sensors can generate independent statistics about the different EPEs caused by the lithography equipment. For example, statistics may be generated that quantize the overlap error of X but are independent of the overlap error of Y, as shown in Figure 8. Overlap may also be quantified independently of the CDU quantization, as shown in Figure 8. In some cases, it is not possible to quantize a component of the EPE independently of another component. In such cases, statistics from different metrology sensors designed to quantify a specific set of EPE components can be used to remove the contributions of those EPE components to obtain an accurate measurement of the remaining EPE components.

[0102] The step of imaging metrology sensors to detect and / or quantify manufacturing inaccuracies caused by photolithography equipment may include microscopic imaging of structures on a substrate. Such imaging characterization techniques are preferably microscopy techniques, such as optical microscopy and alternatively, electron or ion microscopy, electrical characterization, ellipsometry, or scattering measurements. Compared to, for example, electron microscopy, the use of optical microscopy allows for the measurement of multiple metrology sensors in a single frame due to the typically large field of view of the frames acquired using this technique. The advantage of the disclosed metrology sensors and their methods of use is that, by using said sensors, nanoscale differences induced during the photolithography process can be macroscopically detected using optical microscopy, which initially lacks spatial resolution for detecting original nanoscale differences. Advantageously, this technique is compatible with generating metrology data at the single-line level, which can be accomplished by CD-SEM rather than by using statistical measurements such as interferometry, scattering measurements, and optical imaging measurements that sample a large area of ​​at least several micrometers in side lengths surrounding at least several individual structures. Furthermore, the metrology data generated by the metrology sensors becomes independent of resolution limitations and sources of noise and uncertainty from the optical microscope used to read the data.

[0103] In one embodiment, the metering sensor assembly includes: a substrate; an array of metering sensors, each metering sensor including a plurality of mechanical actuators connected by at least one connecting element in a strained state. In one embodiment, each mechanical actuator is adapted to trigger mechanical actuation at the start of an etching process to reach a finished state within a predetermined time period, wherein each connecting element reaches an unstrained state when one of the mechanical actuators reaches its finished state. The substrate preferably comprises a defined out-of-plane crystal orientation semiconductor substrate having a defined out-of-plane crystal orientation, wherein each metering sensor includes two mechanical actuators or two liquid actuators.

[0104] The metrology sensors disclosed herein can be used in a “noise sensing mode”, in which many sensors are patterned with actuators that are identical or have multiple discrete known biases for quantization, the smallest of which is the smallest controllable increment that the lithography apparatus can reliably produce. This can be the smallest step size or the smallest controllable stage step size in an electron beam lithography apparatus. The smallest step size can also be broadly interpreted as the typical minimum grid resolution used to design patterns in CAD, but can also be caused by controllable and known variations in the CD or CDU of the pattern or a portion of the pattern due to a specific lithography process. This sensing mode can be used to characterize the noise floor of the lithography apparatus in terms of EPE, and can be used to compare the performance of different lithography apparatuses or processes involved in the construction of the disclosed metrology sensors. For example, if changing specific process parameters in the lithography apparatus improves statistics (i.e., results in less error in the sensor’s final state), this implies higher edge placement fidelity, and the same parameters can be used in other manufacturing processes to achieve results closer to the ideal. By extension, noise sensors can be used to optimize any process parameter of any lithography equipment or process step, resulting in less error in the sensor's final state, in order to minimize CD, LCDU, random index, and roughness such as LER and LWR. Noise sensing modes can be further used to characterize the noise substrate of metrology sensors patterned by lithography equipment, and their statistics can be used as a reference for sensors in "displacement sensing mode" (described below) to achieve higher accuracy and precision in EPE quantization.

[0105] The metrology sensor disclosed herein can be used alternatively or additionally in a "displacement sensing mode," wherein the pattern of the metrology sensor is intentionally divided into two or more separate but complementary patterns, which do not individually function as metrology sensors, but can form a complete pattern that functions as a metrology sensor when combined in two or more separate exposures in the same or independent photolithography steps, such as... Figure 8A and 8BAs shown in the diagram. For example, the first pattern may define a portion of the actuator and the connecting elements connecting the actuator, while the second exposure 802 defines only the remaining actuator. When a complete pattern is obtained in two exposures, the translational error of the alignment between the first exposure defining pattern 801 and the second exposure defining pattern 802 will statistically produce a detectable imbalance due to, for example, overlap error, because the top island will always be smaller or larger than the bottom island depending on the displacement direction. Complementary patterns can exist at different locations in the same mask, or at any location within two or more different masks. Complementary patterns can be stitched together by shifting a wafer stage or a mask stage or both. When complementary patterns exist in the same mask, the resulting metrology sensor can sense and quantify registration errors and CDUs, etc. When they exist in different masks, the metrology sensor can be used to sense overlaps, etc. A variety of options are possible for dividing a pattern into two or more complementary parts, each potentially depending on or independent of other EPEs sensing different EPEs. For example, the EPE contributions in X and Y can be sensed independently. Furthermore, complementary patterns can be exposed in many ways. For example, complementary patterns can be exposed to make the complete pattern exactly as originally designed, but they can also be exposed with controlled misalignment to generate a bias that can be used to quantify errors. For example, a complete pattern designed in CAD can be expected to produce a statistically balanced result, such as due to mirror symmetry in the pattern, but can actually instead produce a systematic imbalance indicating overlap, registration, or stage error. Exposing complementary patterns using intentionally controlled translation in the X or Y between exposures can produce a more balanced statistic indicating a final pattern closer to the CAD, which can be used to identify and quantify the errors that produce the imbalance. Furthermore, complementary patterns can be used to form many different metrology sensors by repeated complementary exposures. Several complementary patterns can also exist to form several complete metrology sensors in parallel. There can also be more than one way to stitch different complementary patterns in different exposures to obtain a complete pattern that can be used as a metrology sensor. This shift sensing mode can be used to sense EPE by dividing the pattern into one or more parts that affect metrology data when the components are misaligned relative to each other (e.g., due to overlap error) or when the parts have different dimensions (e.g., due to CD error). Displacement sensing patterns can be used in a variety of applications, including sensing of EPEs such as overlap, rotation errors, registration errors, stage or beam drift over time, stage stitching errors, CDUs, and LCDUs. Displacement sensing patterns may be of further interest for characterizing one of the components of a lithography apparatus or its sub-components (such as a stage, beam deflection system, overlap alignment system, or mask). Strategies for dividing patterns into multiple parts using both single-patterning and multi-patterning methods, as well as noise sensing, can also be used to optimize proximity effect correction strategies in optical and electron beam lithography.

[0106] Figure 8A andFigure 8B An example of a cut pattern for a metrology sensor with rectangular and square actuators in a positive resist (a radiation-sensitive polymer used in photolithographic patterning) is shown, which can sense and quantify misalignments (including overlap and registration) and dimensional differences (including CDUs, etc.) when exposed separately to fabricate a complete metrology sensor.

[0107] Figure 8A A metrology sensor design 803, consisting of cut patterns 801 and 802, is shown, capable of sensing and quantifying misalignment in the Y direction independently of misalignment in the X direction. Rotating this structure by 90° allows for sensing and quantifying misalignment in the X direction independently of misalignment in the Y direction. Another feature of this design is that the misalignment facilitates alteration of the two actuators in opposite directions. In other words, the design amplifies the misalignment by a factor of two, thereby doubling the sensitivity of the metrology sensor. Intentionally controlled stage shifts, used to induce or compensate for imbalances between actuators, can be used to quantify errors. Yet another feature of this design is its insensitivity to CDU due to its double mirror symmetry, enabling the design to sense and quantify alignment inaccuracies independently of either CD or CDU inaccuracies.

[0108] Figure 8B Examples of cut patterns 811 and 812 of a metering sensor 810 with square actuators in a positive resist are shown, which can sense and quantify CDU independently of small misalignment inaccuracies (such as overlap and registration) when exposed separately to the complete metering sensor.

[0109] The metrology sensor disclosed herein can be used alternatively or additionally in a “fault sensing mode”, where displaceable material that has not reached a displaced state can be used to sense and quantify manufacturing inaccuracies, such as EPE. For example, if a manufacturing inaccuracy causes a physical discontinuity in a connecting element, the experiment cannot produce a displacement of material that can be interpreted as a displaced state of the metrology sensor, and is therefore interpreted as a fault of the metrology sensor. A fault of the metrology sensor produces useful data because it indicates that the pattern has not been produced as expected. For example, it is well known that very narrow lines used in IC manufacturing can break randomly, which can negatively impact the yield of semiconductor devices, especially in EUV scanners. The fault sensing mode can be used to sense randomly broken lines because a break in a connecting element will not result in the expected displaced state, such as when the connecting element is a beam. The metrology sensor can be intentionally designed to produce experiments that fail or have a chance of failure because any deviation from expected faulty or non-faulty behavior produces useful metrological data about the lithography equipment or process. For example, it is known that when a specific size of the structure is less than 30 nm, the experiment cannot produce a displacement of the substance that can be interpreted as a displacement state of the metrology sensor; if a metrology sensor designed to be larger than this size consistently fails to produce metrological data, this can be interpreted as the actual CD being less than 30 nm, which is less than the expected CD. Fault sensing modes can be used in noise sensing modes or displacement sensing modes, or vice versa; noise and displacement sensing modes can be used in fault sensing modes. By combining different sensing modes, the detection of metrology sensors that fail to produce metrological data enables further sensing and quantification of EPE.

[0110] Figure 9 A flowchart illustrating an example of a currently disclosed method 900 for detecting and / or quantifying manufacturing inaccuracies is shown. The method includes the steps of: providing at least one design for fabricating a structure on a substrate using a set of photolithography processes, at least one pattern defining an array of metrology sensors, wherein each metrology sensor is adapted to generate one of a known and finite set of different possible physical events when a physical process is applied 901; applying the set of photolithography processes to obtain the fabricated structure 902; applying a physical process, thereby generating one of a known and finite set of different possible physical events for each metrology sensor 903; reading out the physical events generated by all metrology sensors and processing the generated physical events 904.

[0111] Further details

[0112] 1. A method for detecting and / or quantifying manufacturing inaccuracies caused by photolithography equipment, comprising:

[0113] Provided at least one pattern for fabricating a structure on a substrate, the at least one pattern defining an array of metering sensors, each metering sensor including a displaceable material configured to be in a stationary state and distributed on a local region, wherein the displaceable material is adapted to reach a displaced state toward a predetermined discrete position within the local region during displacement when a physical stress source is applied.

[0114] The fabricated structure on the substrate is obtained from the photolithography equipment, wherein the displaceable material of the metering sensor is in a stationary state;

[0115] Apply the physical stress source to each metering sensor for at least a predetermined time period; and

[0116] The metering sensor is read out to detect and / or quantify manufacturing inaccuracies caused by the lithography equipment.

[0117] 2. The method according to Project 1, wherein the substrate is a semiconductor substrate.

[0118] 3. The method according to any one of the preceding items, wherein the physical stress source is applied simultaneously to each metering sensor.

[0119] 4. The method according to any one of the preceding items, wherein the metering sensor is read out by imaging.

[0120] 5. The method according to any one of the preceding items, wherein each metering sensor includes a plurality of mechanical actuators connected by at least one connecting element in a strain state representing the stationary state, wherein each mechanical actuator is adapted to trigger mechanical actuation at the start of the etching process to reach an end state within a predetermined time period, and wherein when one of the mechanical actuators reaches its end state, each connecting element reaches an unstrained state representing the displaced state, and wherein the step of simultaneously applying the physical stress source to each metering sensor for at least a predetermined time period includes simultaneously etching an array of the metering sensors for at least a predetermined time period.

[0121] 6. The method according to item 5, wherein the step of imaging the metering sensor includes imaging the connecting element.

[0122] 7. The method according to any one of the preceding items, wherein the mechanical actuator is made of an isotropic or anisotropic etchable material in contact with an etchant.

[0123] 8. The method according to any one of the preceding items, wherein the mechanical actuator is a vertical nanopillar disposed on the substrate.

[0124] 9. The method according to any one of the preceding items, wherein the termination state of the mechanical actuator corresponds to the extent to which the mechanical actuator is partially or completely etched to release the connecting element to which it is connected.

[0125] 10. The method according to any one of the preceding items, wherein the mechanical actuator is etched laterally.

[0126] 11. The method according to any one of the preceding items, wherein the etching process is a wet etching process.

[0127] 12. The method according to any one of the preceding items, wherein the etching process removes at least one of the mechanical actuators of the metering sensor.

[0128] 13. The method according to item 12, wherein the connecting element is displaced from the strain state toward the remaining mechanical actuator of the metering sensor.

[0129] 14. The method according to any one of the preceding items, wherein the connecting element is made of a chemically stable material in contact with an etchant used for mechanical actuator materials.

[0130] 15. The method according to any one of the preceding items, wherein the connecting element is a pre-strain element connected to at least two mechanical actuators within the metering sensor.

[0131] 16. The method according to any one of the preceding items, wherein the connecting element is disposed on the top of the actuator.

[0132] 17. The method according to any one of the preceding items, wherein the connecting element is a beam.

[0133] 18. The method according to any one of the preceding items, wherein the connecting element is a mechanical strain element that is released to an unstrained state when one of the mechanical actuators reaches its end state.

[0134] 19. The method according to any one of the preceding items, wherein the connecting element reaching its unstrained state creates an observable physical record.

[0135] 20. The method according to any one of the preceding items, wherein the displaceable material is photolithographically patterned.

[0136] 21. The method according to any one of items 1-2, wherein the displaceable material of each metering sensor comprises a liquid or gel, wherein the liquid or gel is displaced to a displaced state when the physical stress source is applied, and wherein the step of simultaneously applying the physical stress source to each metering sensor for at least a predetermined time period comprises simultaneously changing the temperature or phase of the liquid or gel or applying vibration or evaporation.

[0137] 22. The method according to item 21, wherein the liquid or gel is disposed in two end regions and a connecting region connecting the two end regions.

[0138] 23. The method according to item 22, wherein, when the physical stress source is applied, the liquid or gel shifts toward one of the end regions.

[0139] 24. The method according to any one of items 21-23, wherein the liquid or gel is adapted to return to a static state when the physical stress source is released or deactivated.

[0140] 25. The method according to any one of the preceding items further includes the step of interpreting the failure of the displaceable material to reach a displaced state as a further manufacturing inaccuracy.

[0141] 26. The method according to any one of the preceding items, wherein the displaceable material is displaceable in the plane of the substrate.

[0142] 27. The method according to any one of the preceding items, wherein the displaceable material creates an observable physical record in a displaced state, the method further comprising the step of physically magnifying the observable physical record.

[0143] 28. The method according to any one of the preceding items, wherein each mechanical actuator or every two end regions of the gel or liquid represents a timer.

[0144] 29. The method according to item 28, wherein the timer is defined as the expected triggering time when the mechanical actuator or end region reaches the displacement state when a physical stress source is applied.

[0145] 30. The method according to any one of items 28-29, wherein each metering sensor includes at least two timers.

[0146] 31. The method according to any one of items 30, wherein the at least two timers are designed to have the same expected trigger time.

[0147] 32. The method according to any one of items 30, wherein the at least two timers are designed to have a predetermined difference in trigger times.

[0148] 33. The method according to any one of items 30, wherein the difference is used to quantify manufacturing inaccuracies.

[0149] 34. The method according to any one of items 32-33, wherein the predetermined difference corresponds to an unknown physical deviation, such as edge placement error.

[0150] 35. The method according to any one of items 28-35, wherein the pattern comprises a plurality of timers.

[0151] 36. The method according to any one of items 28-35, wherein each mechanical actuator or end region is designed such that its size is proportional to a predetermined amount of time taken to reach the end state, corresponding to the expected trigger time of the timer.

[0152] 37. The method according to any one of the preceding items, wherein the array of the metering sensors is distributed on the substrate.

[0153] 38. The method according to item 37, wherein the metering sensors are arranged in an M×N configuration, where M>=2 and N>=2.

[0154] 39. The method according to any one of the preceding items, wherein the pattern defines an array of at least 10 metering sensors.

[0155] 40. The method according to any one of the preceding items, wherein the step of imaging the metrology sensor to detect and / or quantify manufacturing inaccuracies generated by the lithography apparatus includes microscopy techniques, such as optical microscopy, for imaging the structure on the substrate.

[0156] 41. A metering sensor assembly, comprising:

[0157] substrate;

[0158] An array of metering sensors, each metering sensor comprising a plurality of mechanical actuators connected by at least one connecting element in a strained state, wherein each mechanical actuator is adapted to trigger mechanical actuation at the start of an etching process to reach a finished state within a predetermined time period, and wherein each connecting element reaches an unstrained state when one of the mechanical actuators reaches its finished state.

[0159] 42. The metering sensor assembly according to item 41, wherein the substrate comprises a semiconductor substrate having a defined out-of-plane crystal orientation.

[0160] 43. The metering sensor according to any one of items 41-42, wherein the substrate is a semiconductor substrate having a defined out-of-plane crystal orientation, and wherein each metering sensor includes two mechanical actuators or two liquid actuators.

[0161] 44. A measurement system comprising:

[0162] A photolithography apparatus configured to pattern a radiation-sensitive resist, such as a polymer, on a substrate using an array of metering sensors, each metering sensor including a displaceable material configured to be in a stationary state and distributed over a local area, wherein the displaceable material is adapted to reach a displaced state toward a predetermined discrete position within the local area during displacement when a physical stress source is applied.

[0163] Physical stress source devices, such as etching apparatuses or systems, are adapted to apply physical stress for at least a predetermined period of time; and

[0164] An imaging device for imaging the metrology sensor to detect and / or quantify manufacturing inaccuracies caused by the lithography equipment.

[0165] 45. The metrology system according to item 44, wherein the imaging device includes a microscope, preferably an optical microscope.

Claims

1. A method for detecting and / or quantifying manufacturing inaccuracies resulting from a lithographic process, the method comprising the steps of: providing at least one design for manufacturing a structure on a substrate using a set of lithographic processes, wherein the manufactured structure defines an array of metrology sensors, wherein each metrology sensor is adapted to produce one of a known and limited set of possible different and discrete physical events upon application of a physical process, wherein the produced physical event: is unknown prior to application of the physical process, depends on a manufacturing inaccuracy resulting from at least one of the set of lithographic processes, corresponds to a shifted state of the manufactured structure, or creates one or more physical entities associated with the manufactured structure that did not exist prior to application of the physical process, and is larger than the inaccuracy; applying the set of lithographic processes to obtain the manufactured structure; applying the physical process, whereby one of the known and limited set of possible different and discrete physical events is produced for each metrology sensor, wherein the physical process comprises mechanically applying stress, releasing internal stress, spin coating, selective deposition, selective etching, freezing, bimorph wafer shifting, exposure to radiation of a specific wavelength, ultrasound or megasonic vibrations, and combinations of one or more of the above physical processes; reading out the produced physical events of all metrology sensors; and processing the produced physical events of all metrology sensors to detect and / or quantify manufacturing inaccuracies resulting from the lithographic process, wherein after the steps have been performed, each metrology sensor contains data about inaccuracies resulting from the lithographic process.

2. The method of claim 1, wherein, the produced physical event is visually distinguishable from other physical events of the set of possible different and discrete physical events.

3. The method according to any of the preceding claims, wherein, the manufacturing inaccuracy results from an edge placement error, wherein the edge placement error is less than 5 nm.

4. The method of claim 1, wherein, the step of processing the produced physical events comprises computationally processing the produced physical events.

5. The method of claim 4, comprising the step of applying Boolean logic to the produced physical events based on an expected physical outcome from the at least one design produced by each metrology sensor.

6. The method of any of claims 4-5, comprising the step of constructing a probability distribution of counts of produced physical events for one or more varied design parameters of the design.

7. The method of claim 1, wherein, the set of lithographic processes comprises at least exposure and development of a radiation-sensitive resist.

8. The method of claim 1, wherein, the step of reading out the produced physical events comprises searching for the physical events only at predetermined locations of the manufactured structure.

9. The method of claim 1, wherein, the produced physical events are digitally stored.

10. The method of claim 1, wherein, imaging is used to read out the produced physical events of all metrology sensors.

11. The method of claim 1, wherein, each metrology sensor comprises a shiftable substance configured to be in a rest state and distributed over a local area, wherein the shiftable substance is adapted to reach a shifted state within the local area during a shifting process upon application of the physical process.

12. The method of claim 11, wherein, the shiftable substance is in a rest state when the manufactured structure on the substrate is obtained from the lithographic process.

13. The method of claim 1, wherein, applying the physical process to each metrology sensor for a predetermined period of time.

14. The method of claim 1, wherein, the at least one design is configured such that one particular physical event is favorable.

15. The method of claim 14, wherein, providing two or more designs, wherein at least one of the designs can be distinguished from another design by at least one edge shift of an integer multiple of a minimum controllable step length of the lithography apparatus.

16. The method of claim 1, wherein, the at least one design is configured such that none of the known and limited set of different and discrete physical events is favorable.

17. The method of claim 1, wherein, the physical event is a change in shape, size or placement of part or all of the structure.

18. The method of claim 11 or 12, wherein, each metrology sensor comprises a plurality of mechanical actuators connected by at least one link element in a strained state representing the rest state, wherein each mechanical actuator is adapted to trigger a mechanical actuation to reach an end state in a predetermined amount of time at the beginning of the etching process, and wherein, when one of the mechanical actuators reaches its end state, each link element reaches an unstrained state representing the shifted state, and wherein the step of applying the physical process simultaneously to each metrology sensor for at least a predetermined period of time comprises simultaneously etching the array of metrology sensors for at least a predetermined period of time.

19. The method of claim 18, wherein, the end state of the mechanical actuator corresponds to a partial or complete etching of the mechanical actuator to an extent that it releases the link element it is connected to.

20. The method of claim 11, wherein, the shiftable substance of each metrology sensor comprises a liquid or a gel, wherein, upon applying the physical process, the liquid or gel shifts to a shifted state, and wherein the step of applying the physical process simultaneously to each metrology sensor for at least a predetermined period of time comprises simultaneously changing the temperature or the phase of the liquid or gel or applying a vibration or evaporation.

21. The method of claim 20, wherein, the liquid or gel is arranged in two end regions and a connecting region connecting the two end regions, wherein, upon applying the physical process, the liquid or gel shifts towards one of the end regions.

22. The method of claim 18, wherein, each mechanical actuator represents a timer.

23. The method of claim 20, wherein, each two end regions of the gel or liquid represent a timer.

24. The method of claim 22, wherein, the timer defines an expected trigger time at which the mechanical actuator triggers the shiftable substance to reach a shifted state upon applying the physical process.

25. The method of claim 23, wherein, the timer defines an expected trigger time at which the end region triggers the shiftable substance to reach a shifted state upon applying the physical process.

26. The method of claim 22 or 24, wherein, each mechanical actuator is designed such that its size is proportional to a predetermined amount of time it takes to reach an end state, and the predetermined amount of time corresponds to the expected trigger time of the timer.

27. The method of claim 23 or 25, wherein, each end region is designed such that its size is proportional to a predetermined amount of time it takes to reach an end state, and the predetermined amount of time corresponds to the expected trigger time of the timer.

28. The method of claim 17, wherein, the change in shape, size or placement of part or all of the structure is larger than 10 nm.

29. The method of claim 3, wherein, the edge placement error is smaller than 1 nm.

30. The method of claim 1, wherein, mechanically applying stress is applying stress by bending the substrate.

31. A method for detecting and / or quantifying manufacturing inaccuracies resulting from a lithography process, the method comprising the steps of: There is provided for manufacturing at least one design for fabricating structures on a substrate using a set of lithography processes, wherein the fabricated structures define an array of metrology sensors, wherein each metrology sensor is adapted to produce one of a known and limited set of possible different and discrete physical events upon application of a physical process, wherein the produced physical event: is unknown prior to application of the physical process, depends on a fabrication inaccuracy produced by at least one of the set of lithography processes, corresponds to a shifted state of the fabricated structure, or creates one or more physical entities associated with the fabricated structure that did not exist prior to application of the physical process, and is greater than the inaccuracy; applying the set of lithography processes to obtain the fabricated structures; applying the physical process, thereby producing for each metrology sensor one of the known and limited set of possible different and discrete physical events, wherein the physical process comprises selective exposure or heating; reading out the produced physical events of all metrology sensors; and processing the produced physical events of all metrology sensors to detect and / or quantify the fabrication inaccuracy produced by the lithography processes, wherein after the steps have been performed, each metrology sensor contains data about the inaccuracy produced by the lithography processes.

32. A metrology system comprising: a lithography apparatus configured to pattern a radiation-sensitive resist on a substrate using a design comprising an array of metrology sensors to fabricate structures, wherein each metrology sensor is adapted to produce one of a known and limited set of possible different and discrete physical events upon application of a physical process, wherein the produced physical event: is unknown prior to application of the physical process, depends on a fabrication inaccuracy produced by the lithography apparatus, corresponds to a shifted state of the fabricated structure, or creates one or more physical entities associated with the fabricated structure that did not exist prior to application of the physical process, and is greater than the inaccuracy, wherein after the steps have been performed, each metrology sensor contains data about the inaccuracy produced by the lithography processes; a system adapted to apply the physical process to each metrology sensor, wherein the physical process comprises mechanically applying stress, releasing internal stress, spin coating, selective deposition, selective etching, freezing, bimorph piezoelectric wafer shifting, exposure to radiation of a specific wavelength, ultrasound or megasonic vibrations, and combinations of one or more of the above physical processes.

33. The metrology system of claim 32, wherein, Each metrology sensor comprises a shiftable substance configured to be in a resting state and distributed over a local area, wherein the shiftable substance is adapted to reach a shifted state towards a predetermined discrete location within the local area in a shifting process upon application of the physical process.

34. The metrology system according to any of claims 32-33, further comprising an imaging device for imaging the metrology sensors to detect and / or quantify the fabrication inaccuracy produced by the lithography apparatus.

35. The metrology system of claim 32, wherein, The radiation-sensitive resist is a polymer.

36. The metrology system of claim 32, wherein, Mechanically applying stress is applying stress by bending the substrate.

37. A metrology system comprising: A lithographic apparatus configured to pattern a radiation-sensitive resist on a substrate using a design comprising an array of metrology sensors, wherein each metrology sensor is adapted to produce one of a known and finite set of possible different and discrete physical events upon application of a physical process, wherein the produced physical event: is unknown prior to application of the physical process, depends on a manufacturing inaccuracy produced by the lithographic apparatus, corresponds to a shifted state of the manufactured structure, or creates one or more physical entities associated with the manufactured structure that did not exist prior to application of the physical process, and is greater than the inaccuracy, wherein after the steps have been performed, each metrology sensor contains data about the inaccuracy produced by the lithographic process; a system adapted to apply the physical process to each metrology sensor, wherein the physical process comprises selective exposure or heating.

Images