Virtual metrology system components and associated methods

Abstract

A specialized machine learning model configured to minimize and / or eliminate the need for motorized hardware to block higher order diffracted radiation from corrupting alignment measurements in semiconductor manufacturing metrology processes. For example, a radiation source illuminates a metrology target in a layer of a patterned substrate with radiation, the metrology target diffracts the radiation. A radiation sensor outputs phase data and intensity data of the diffracted radiation. The phase data includes different energies associated with different diffraction orders of the diffracted radiation. The model is configured to determine a subset of the intensity data associated with positive first order diffracted radiation and negative first order diffracted radiation incident on the radiation sensor based on the phase data, and determine an intensity difference or intensity imbalance between the positive first order diffracted radiation and the negative first order diffracted radiation based on the subset of the intensity data. An alignment is determined based on the intensity imbalance.

Description

Cross-references to related applications

[0001] This application claims priority to US application 63 / 544,657, filed October 18, 2023, which is incorporated herein by reference in its entirety. Technical Field

[0002] This description relates to virtual measurement system components and associated methods. Background Technology

[0003] Photolithography projection equipment can be used, for example, in the manufacture of integrated circuits (ICs). A patterning apparatus (e.g., a mask) can include or provide a pattern corresponding to a single layer (“design layout”) of the IC, and can transfer the pattern onto a target portion (e.g., a silicon wafer) already coated with a layer of radiation-sensitive material (“resist”) by methods such as irradiating the target portion through the pattern on the patterning apparatus. Typically, a single substrate comprises multiple adjacent target portions, and the photolithography projection equipment continuously transfers the pattern onto the target portions one at a time. In one type of photolithography projection equipment, the pattern on the entire patterning apparatus is transferred to a single target portion in a single operation. This type of equipment is often referred to as a stepper. In alternative equipment (often referred to as a step-scanning apparatus), a projection beam scans across the patterning apparatus along a given reference direction (“scanning” direction) while the substrate moves synchronously parallel to or antiparallel to that reference direction. Different portions of the pattern on the patterning apparatus are progressively transferred to a single target portion. Typically, since the photolithography projection equipment will have a reduction ratio M (e.g., 4), the substrate moving speed F will be 1 / M times that of the projection beam scanning pattern forming apparatus.

[0004] Before a pattern is transferred from a patterning apparatus to a substrate, the substrate can undergo various processes, such as primer coating, resist coating, and soft baking. After exposure, the substrate can undergo other processes (“post-exposure processes”), such as post-exposure baking (PEB), development, hard baking, and measurement / inspection of the transferred pattern. This series of processes serves as the basis for fabricating individual layers of a device, such as an IC. The substrate can then undergo various processes, such as etching, ion implantation (doping), metallization, oxidation, deposition, chemical mechanical polishing, etc., all of which are designed to complete the individual layers of the device. If several layers are required in the device, the entire process or its variations are repeated for each layer. Ultimately, the device will exist in each target portion of the substrate. These devices are then separated from each other using techniques such as dicing or sawing so that individual devices can be mounted on a carrier, connected to pins, etc.

[0005] Therefore, manufacturing devices (such as semiconductor devices) typically involves processing a substrate (e.g., a semiconductor wafer) using multiple manufacturing processes to form various features and multiple layers of the device. These layers and features are typically fabricated and processed using techniques such as deposition, photolithography, etching, chemical mechanical polishing, and ion implantation. Multiple devices can be fabricated on multiple chips on a substrate and then separated into individual devices. This device fabrication process can be viewed as a patterning process. A patterning process involves patterning steps such as optical and / or nanoimprint lithography using a patterning apparatus in a photolithography device to transfer a pattern from the patterning apparatus to the substrate, and typically, but optionally, involves one or more associated patterning processing steps, such as developing with a resist in a developing apparatus, baking the substrate using a baking tool, etching the pattern using an etching apparatus, deposition, etc.

[0006] Photolithography is a central step in the fabrication of devices such as integrated circuits (ICs), in which patterns formed on a substrate define the functional elements of the device, such as microprocessors and memory chips. Similar photolithography techniques are also used to form flat panel displays, microelectromechanical systems (MEMS), and other devices.

[0007] As semiconductor manufacturing processes continue to advance, the size of functional components continues to shrink, while the number of functional components (such as transistors) in each device has steadily increased for decades, following a trend commonly known as "Moore's Law." In the current state of technology, photolithography projection devices are used to fabricate the layers of devices. These devices project a design layout onto a substrate using irradiation from a deep ultraviolet light source, resulting in individual functional components with dimensions well below 100 nm (i.e., less than half the wavelength of radiation from the irradiation source, such as a 193 nm source).

[0008] The process of printing features with dimensions smaller than the classical resolution limit of photolithography projection equipment is often referred to as low-k1 lithography, according to the resolution formula CD = k1 × λ / NA, where λ is the wavelength of the radiation used (currently, in most cases 248 nm or 193 nm), NA is the numerical aperture of the projection optics in the photolithography projection equipment, CD is the "critical size" (typically the smallest feature size to be printed), and k1 is an empirical resolution factor. Generally, the smaller k1 is, the more difficult it becomes to reproduce patterns on a substrate that resemble the shape and size planned by the circuit designer to achieve specific electrical functions and performance. To overcome these difficulties, complex fine-tuning steps can be applied to the photolithography projection equipment, layout design, or patterning apparatus. These steps include, but are not limited to: optimization of NA and optical coherence settings, self-limiting illumination schemes, use of phase-shifting patterning apparatus, optical proximity correction (OPC, sometimes also called "optical and process correction") in the layout design, or other methods generally defined as "resolution enhancement techniques" (RET). Accurate measurement is important for these operations. Summary of the Invention

[0009] Specialized machine learning models are configured to block higher-order diffraction radiation, which can disrupt alignment measurements during semiconductor manufacturing metrology processes. In some implementations, this model can be configured to minimize and / or eliminate the need for motored hardware to block higher-order diffraction radiation. Motored hardware is a complex and expensive component of metrology systems. Furthermore, motor rotation is not fast enough to meet the requirements of all metrology system use cases, and motors reduce the lifespan of the motored hardware, among other drawbacks.

[0010] In contrast, in this system and method, a radiation source irradiates a measurement target in a layer of a patterned substrate, which diffracts the radiation. A radiation sensor outputs phase and intensity data of the diffracted radiation. The phase data includes different energies associated with different diffraction orders of the diffracted radiation. A machine learning model is configured to determine, based on the phase data, a subset of intensity data associated with the positive and negative first-order diffracted radiation incident on the radiation sensor, and, based on this subset of intensity data, to determine the intensity difference or intensity imbalance between the positive and negative first-order diffracted radiation. Alignment is determined based on this intensity imbalance. Advantageously, the use of electrical hardware is eliminated (or at least significantly reduced, as described below) in this system and method. Alternatively, a measurement system is provided with virtual components configured to minimize and / or eliminate the need for electrical hardware used to block higher-order diffracted radiation from disrupting alignment measurements during semiconductor manufacturing metrology processes.

[0011] According to one embodiment, a measurement system is provided. The measurement system includes a radiation source configured to irradiate a measurement target in a layer of a patterned substrate. The measurement target is configured to diffract radiation from the radiation source. The system includes a radiation sensor configured to output phase data and intensity data of the diffracted radiation. The phase data includes different energies associated with different diffraction orders of the diffracted radiation. The system includes one or more processors operatively connected to the radiation sensor and configured to execute a trained machine learning model via machine-readable instructions. The trained machine learning model is configured to determine, based on the phase data, a subset of the intensity data associated with positive and negative first-order diffracted radiation incident on the radiation sensor. The trained machine learning model is configured to determine, based on the subset of the intensity data, the intensity difference between the positive and negative first-order diffracted radiation.

[0012] In some implementations, the intensity difference includes intensity imbalance. The one or more processors are also configured to determine the alignment of the layer based on the intensity imbalance.

[0013] In some implementations, determining the subset of intensity data includes excluding intensity data associated with destructive higher-order diffraction radiation. Data associated with the destructive higher-order diffraction radiation is identified based on the energy associated with the phase of the higher diffraction order.

[0014] In some implementations, the destructive higher-order diffraction radiation includes leakage.

[0015] In some implementations, the phase data conveys the alignment position of the layer relative to the average position from all wavelengths for all diffraction orders entering the pupil of the measurement system; and the relative diffraction efficiency corresponding to all diffraction orders entering the pupil.

[0016] In some implementations, the intensity data includes intensity imbalance with diffraction order mixing.

[0017] In some implementations, the trained machine learning model includes a multinomial regression model, a decision tree regressor, a random forest regressor, a multilayer perceptron (MLP) regressor, a stochastic gradient descent (SGD) regressor, an extreme gradient boosting (XGB) regressor, or any combination thereof.

[0018] In some implementations, the trained machine learning model includes a multinomial regression model.

[0019] In some implementations, the trained machine learning model is trained by: providing an untrained machine learning model with phase channel alignment positions relative to the average position from all wavelengths for all diffraction orders entering the pupil of the measurement system; determining substrate quality based on the relative diffraction efficiency corresponding to all radiation diffraction orders entering the pupil, and providing the substrate quality to the untrained machine learning model; providing the untrained machine learning model with a fundamental true intensity channel imbalance having order mixing, and providing the untrained machine learning model with a fundamental true intensity channel imbalance only for first-order diffraction radiation. The untrained machine learning model is provided with a new intensity channel imbalance having order mixing, and the untrained machine learning model is used to determine a new intensity imbalance of the first diffraction order in the diffraction order entering the pupil based on the new intensity channel imbalance having order mixing; in response to the difference between the determination result and the underlying true intensity channel imbalance of the first-order diffraction radiation being less than a threshold amount, the accuracy of the new intensity imbalance determination result is determined; and one or more parameters of the untrained machine learning model are adjusted based on one or more accuracy determination results to form a trained machine learning model.

[0020] In some implementations, the threshold value is 1e-4.

[0021] In some implementations, the underlying true intensity channel imbalance for first-order diffraction radiation is determined by simulation.

[0022] In some implementations, an alignment-rotating optics motor is used to determine the fundamental truth intensity channel imbalance for first-order diffraction radiation fed to the untrained machine learning model. In some implementations, the alignment-rotating optics motor is added to the system to generate training data and is removed and / or not used once the trained machine learning model has finished training.

[0023] In some implementations, the trained machine learning model is further trained by measuring multiple orthogonal polarization states for each diffraction order and providing these polarization states, along with the polarization specific to the first-order diffraction radiation, to the untrained machine learning model as further training input.

[0024] In some implementations, the system includes electrically powered hardware configured to block a portion of higher-order diffracted radiation from reaching the radiation sensor. In some implementations, the electrically powered hardware includes an alignment rotating optics motor.

[0025] In some implementations, the trained machine learning model is configured to minimize the need for the system to use the electrical hardware for blocking portions of the higher-order diffracted radiation from reaching the radiation sensor.

[0026] In some implementations, the system is configured such that the intensity is combined with the phase of the diffracted radiation to detect and correct sub-nanometer alignment shifts caused by asymmetry of the measurement target.

[0027] In some implementations, the measurement target includes a grating.

[0028] In some implementations, the intensity imbalance between the positive first-order diffraction radiation and the negative first-order diffraction radiation is configured to be used by the one or more processors to determine the alignment of the layer, and then to adjust the semiconductor device manufacturing process based on the alignment.

[0029] According to another implementation, a corresponding measurement method is provided that includes one or more of the operations described above.

[0030] According to another implementation, a non-transitory computer-readable medium is provided that stores instructions which, when executed by a computer (e.g., including one or more processors), cause the computer to perform one or more of the operations described above. Attached Figure Description

[0031] The above aspects, as well as other aspects and features, will become apparent to those skilled in the art after reading the following description of specific embodiments in conjunction with the accompanying drawings.

[0032] Figure 1 A lithography apparatus according to one embodiment is schematically depicted.

[0033] Figure 2 An embodiment of a photolithography unit or cluster according to one embodiment is schematically depicted.

[0034] Figure 3 An exemplary inspection system according to one embodiment is schematically depicted.

[0035] Figure 4 An exemplary measurement technique according to one embodiment is schematically depicted.

[0036] Figure 5 The diagram illustrates the relationship between the radiation irradiation spot and the measurement target of an inspection system according to one embodiment.

[0037] Figure 6The illustration depicts a measurement system with a virtual component according to one embodiment, the virtual component being configured to minimize and / or eliminate the need for motorized hardware for blocking higher-order diffraction radiation from damaging alignment measurements during semiconductor manufacturing measurement processes.

[0038] Figure 7 The illustration shows a virtual component of a measurement system according to one embodiment.

[0039] Figure 8 The figure illustrates a measurement method according to one embodiment.

[0040] Figure 9 This is a block diagram of an exemplary computer system according to one embodiment. Detailed Implementation

[0041] In semiconductor device manufacturing, metrology operations typically involve determining the location of measurement markers (or multiple markers) and / or other targets within layers of the semiconductor device structure. This location is usually determined by irradiating the measurement markers with radiation and comparing the characteristics of different diffraction orders of the diffracted radiation reflected from the measurement markers. Such techniques are used to measure alignment and / or other parameters.

[0042] Some alignment sensors are configured to output diffraction radiation intensity information. This intensity information is configured to measure and correct sub-nanometer alignment offsets caused by mark asymmetry. This correction can be performed by calculating the intensity imbalance between the positive and negative first-order diffraction radiation from the mark. However, for certain combinations of diffraction radiation wavelengths and the periodicity of the measurement mark, the output intensity information is disrupted by higher-order diffraction radiation (e.g., the positive and negative second, third, fourth, etc. diffraction orders). Electrical hardware (e.g., alignment rotation optics motors and / or other components) can be used to prevent leakage by moving to physically block higher-order diffraction radiation from reaching the alignment sensor. However, electrical hardware, in addition to the alignment sensor, is often complex and expensive. Furthermore, the motor, often part of the electrical hardware, typically rotates too slowly for many alignment sensor use cases and generally reduces the lifespan of the sensor module in the measurement system.

[0043] Advantageously, this paper describes measurement systems (and methods, software media, etc.) with virtual components (e.g., trained electronic models and / or other components) configured to minimize and / or eliminate the need for electrical hardware. These systems include specialized machine learning models developed to reduce the use of motors and / or other electrical hardware components for aligning rotating optics. The model is configured to determine a subset of intensity data associated with the positive and negative first-order diffraction radiation incident on a radiation sensor, based on phase data of the same diffraction radiation. Determining the subset of intensity data includes excluding intensity data associated with destructive higher-order diffraction radiation. Data associated with destructive higher-order diffraction radiation can be identified based on the energy associated with the phase of the higher diffraction order. The model is also configured to determine the intensity difference between the positive and negative first-order diffraction radiation (based only on this subset of intensity data). This intensity difference includes an intensity imbalance that can be used to determine the alignment of a substrate layer (e.g., a layer with measurement markings). In this way, a trained model is used to minimize and / or eliminate the need for electric hardware.

[0044] For brevity, the following description relates to semiconductor device fabrication and patterning processes. The following paragraphs also describe several components of systems and / or methods used for semiconductor device measurement. For example, these systems and methods can be used to measure alignment during semiconductor device fabrication or for other operations.

[0045] While specific references may be made herein to measurements of alignment or other parameters, and to the fabrication of integrated circuits (ICs) for semiconductor devices, it should be understood that the description herein has many other possible applications. For example, other possible applications include the fabrication of integrated optical systems, guiding and detection patterns for magnetic domain memories, liquid crystal displays, thin-film magnetic heads, etc. Those skilled in the art will understand that, in the context of such alternative applications, any use of the terms “mask,” “wafer,” or “die” herein should be considered interchangeable with the more general terms “mask,” “substrate,” and “target portion,” respectively.

[0046] The term "projection optics" as used herein should be interpreted broadly to encompass various types of optical systems, including, for example, refractive optics, reflective optics, aperture optics, and refractive-reflective optics. The term "projection optics" may also include components for any of these design types used to guide, shape, or control a projected radiation beam, either jointly or individually. The term "projection optics" can include any optical component in a lithography projection apparatus, regardless of its location in the optical path of the apparatus. Projection optics can include optical components for shaping, adjusting, and / or projecting radiation from a source before it passes through a patterning device, and / or for shaping, adjusting, and / or projecting radiation after it has passed through the patterning device. Projection optics typically do not include a source and patterning device.

[0047] Figure 1 An embodiment of a photolithography apparatus LA is schematically depicted. The apparatus includes: an irradiation system (irradiator) IL configured to modulate a radiation beam B (e.g., UV, DUV, or EUV radiation); a support structure (e.g., a mask stage) MT configured to support a patterning apparatus (e.g., a mask) MA and connected to a first positioner PM configured to accurately position the patterning apparatus according to specific parameters; a substrate stage (e.g., a wafer stage) WT (e.g., WTa, WTb, or both) configured to hold a substrate (e.g., a resist-coated wafer) W and connected to a second positioner PW configured to accurately position the substrate according to specific parameters; and a projection system (e.g., a refractive projection lens system) PS configured to project a pattern imparted by the patterning apparatus MA to the radiation beam B onto a target portion C (e.g., comprising one or more dies, commonly referred to as a field) of the substrate W. The projection system is supported on a reference frame RF. As shown, the apparatus is transmissive (e.g., employing a transmissive mask). Alternatively, the device can also be reflective (e.g., using a programmable array of mirrors or a reflective mask).

[0048] The irradiator IL receives a radiation beam from the radiation source SO. The source and the lithography apparatus can be separate entities, such as when the radiation source is an excimer laser. In this case, the source is not considered part of the lithography apparatus, and the radiation beam is transmitted from the radiation source SO to the irradiator IL by means of a beam delivery system BD (e.g., including suitable directional mirrors and / or beam expanders). In other cases, the source can be a component of the apparatus, such as when the source is a mercury lamp. The source SO, the irradiator IL, and the beam delivery system BD (if desired) can be collectively referred to as the radiation system.

[0049] An illuminator IL can alter the intensity distribution of the beam. The illuminator can be arranged to limit the radial range of the radiation beam such that the intensity distribution is non-zero within an annular region in the pupil plane of the illuminator IL. Alternatively or additionally, the illuminator IL can be operated to limit the beam distribution within the pupil plane such that the intensity distribution is non-zero within multiple equally spaced sectors in the pupil plane. The intensity distribution of the radiation beam within the pupil plane of the illuminator IL can be referred to as the illumination mode.

[0050] The illuminator IL may include a modulator AD configured to adjust the (angular / spatial) intensity distribution of the beam. Typically, at least the range of the outer and / or inner diameter of the intensity distribution within the pupil plane of the illuminator (often referred to as σ-external and σ-internal, respectively) can be adjusted. The illuminator IL may be operable to change the angular distribution of the beam. For example, the illuminator may be operable to change the number and angular range of sectors in the pupil plane where the intensity distribution is non-zero. Different illumination modes can be achieved by adjusting the beam intensity distribution within the pupil plane of the illuminator. For example, by limiting the radial and angular range of the intensity distribution within the pupil plane of the illuminator IL, the intensity distribution can have a multipole distribution, such as a dipole, tetrapole, or hexapole distribution. This illumination mode can be obtained, for example, by inserting an optics providing the desired illumination mode into the illuminator IL or by using a spatial light modulator.

[0051] The illuminator IL can be operated to change the polarization of the beam and can be operated to adjust the polarization using a modulator AD. The polarization state of the radiation beam passing through the pupil plane of the illuminator IL is called the polarization mode. Using different polarization modes can result in higher contrast in the image formed on the substrate W. The radiation beam can be unpolarized. Alternatively, the illuminator can be arranged to linearly polarize the radiation beam. The polarization direction of the radiation beam can vary across the pupil plane of the illuminator IL. The polarization direction of the radiation can be different in different regions within the pupil plane of the illuminator IL. The polarization state of the radiation can be selected according to the illuminator mode. For multi-pole illuminator modes, the polarization of each pole of the radiation beam can generally be perpendicular to the position vector of that pole within the pupil plane of the illuminator IL. For example, for dipole illuminator modes, the radiation can be linearly polarized in a direction substantially perpendicular to the line bisecting the two opposing sectors of the dipole. The radiation beam can be polarized in one of two different orthogonal directions, which can be called the X-polarization state and the Y-polarization state. For quadrupole illuminator modes, the radiation within the sector of each pole can be linearly polarized in a direction substantially perpendicular to the line bisecting that sector. This polarization mode can be called XY polarization. Similarly, in a hexapolar illumination mode, the radiation within the sector of each pole can be linearly polarized in a direction substantially perpendicular to the line bisecting that sector. This polarization mode can be called TE polarization.

[0052] In addition, the irradiator IL typically includes various other components, such as an integrator IN and a concentrator CO. The irradiation system may include various types of optical components for guiding, shaping, or controlling radiation, such as refractive, reflective, magnetic, electromagnetic, electrostatic, or other types of optical components, or any combination thereof. Therefore, the irradiator provides a modulated radiation beam B with a desired uniformity and intensity distribution across its cross-section.

[0053] The support structure MT supports the patterning apparatus in a manner dependent on the orientation of the patterning apparatus, the design of the lithography equipment, and other conditions, such as whether the patterning apparatus is held in a vacuum environment. The support structure can use mechanical, vacuum, electrostatic, or other clamping techniques to hold the patterning apparatus. The support structure can be a frame or stage, and may be fixed or movable as needed. The support structure ensures that the patterning apparatus is in a desired position, such as relative to the projection system. Any use of the terms "mask" or "mask" herein may be considered consistent with the more general term "patterning apparatus".

[0054] The term "patterning apparatus" should be interpreted broadly as any apparatus that can be used to apply a pattern to a target portion of a substrate. In embodiments, a patterning apparatus is any apparatus that can be used to impart a pattern to a radiation beam in its cross-section to form a pattern on a target portion of the substrate. It should be noted that the pattern imparted to the radiation beam may not perfectly correspond to the desired pattern in the target portion of the substrate, for example, if the pattern includes phase-shifting features or so-called auxiliary features. Typically, the pattern imparted to the radiation beam will correspond to a specific functional layer in a device (such as an integrated circuit) that will be formed in the target portion of that device.

[0055] Pattern forming apparatuses can be transmissive or reflective. Examples of pattern forming apparatuses include masks, programmable mirror arrays, and programmable LCD panels. Masks are well known in photolithography and include mask types such as binary, alternating phase-shift, and attenuation phase-shift masks, as well as various hybrid mask types. Examples of programmable mirror arrays use a matrix arrangement of small mirrors, each of which can be individually tilted to reflect the incident radiation beam in different directions. The tilted mirrors impart a pattern to the radiation beam reflected by the mirror matrix.

[0056] The term "projection system" should be interpreted broadly to encompass any type of projection system suitable for the exposure radiation used or for other factors such as immersion in liquids or vacuum, including refractive, reflective, reflective-refractive, magnetic, electromagnetic, and electrostatic optical systems, or any combination thereof. Any use of the term "projection lens" herein may be considered synonymous with the more general term "projection system."

[0057] A projection system PS may include multiple optical (e.g., lens) elements and may also include an adjustment mechanism configured to adjust one or more optical elements to correct aberrations (phase changes across the pupil plane throughout the field). To achieve this, the adjustment mechanism may be operable to manipulate one or more optical (e.g., lens) elements within the projection system PS in one or more different ways. The projection system may have a coordinate system whose optical axis extends along the z-direction. The adjustment mechanism may be operable to perform any combination of operations: shifting one or more optical elements; tilting one or more optical elements; and / or deforming one or more optical elements. Shifting of optical elements can be performed in any direction (x, y, z, or a combination thereof). Tilting of optical elements is typically achieved outside a plane perpendicular to the optical axis by rotation about axes along the x and / or y directions, but for non-rotationally symmetric aspherical optical elements, rotation about the z-axis may also be used. Deformation of optical elements may include low-frequency shapes (e.g., astigmatism) and / or high-frequency shapes (e.g., freeform aspherical surfaces). Deformation of an optical element can be performed, for example, by applying force to one or more sides of the optical element using one or more actuators, and / or by heating one or more selected areas of the optical element using one or more heating elements. Typically, it may be impossible to adjust the projection system PS to correct vignetting (transmittance variation across the pupil plane). When designing the patterning apparatus (e.g., mask) MA of a lithography apparatus LA, the transmittance map of the projection system PS can be used. Using computational lithography techniques, the patterning apparatus MA can be designed to at least partially correct vignetting.

[0058] A lithography apparatus can have two (dual) or more stages (e.g., two or more substrate stages WTa, WTb, two or more patterning apparatus stages, substrate stage WTa, and a dedicated stage WTb located below the projection system and without a substrate, for purposes such as facilitating measurement and / or cleaning). In such a "multi-stage" machine, additional stages can be used in parallel, or one or more other stages can be used for exposure while preparatory steps are performed on one or more stages. For example, alignment measurements can be performed using an alignment sensor AS, and / or level (height, tilt, etc.) measurements can be performed using a level sensor LS.

[0059] Photolithography apparatuses can also fall into the category where at least a portion of the substrate can be covered by a liquid (e.g., water) with a relatively high refractive index to fill the space between the projection system and the substrate. Immersion liquids can also be applied to other spaces within the photolithography apparatus, such as the space between the patterning apparatus and the projection system. Immersion techniques are well-known in the art for increasing the numerical aperture of projection systems. The term "immersion" as used herein does not mean that a structure such as a substrate must be submerged in a liquid, but only that the liquid is located between the projection system and the substrate during exposure.

[0060] In the operation of a photolithography apparatus, a radiation beam B is regulated and provided by an irradiation system IL. The radiation beam B is incident on a patterning apparatus (e.g., a mask) MA and patterned by the apparatus, which is held on a support structure (e.g., a mask stage) MT. After passing through the patterning apparatus MA, the radiation beam B passes through a projection system PS, which focuses the beam onto a target portion C of a substrate W. The substrate stage WT can be accurately moved, for example, to position different target portions C within the path of the radiation beam B, by means of a second positioner PW and a position sensor IF (e.g., an interferometer device, a linear encoder, a 2-D encoder, or a capacitive sensor). Similarly, a first positioner PM and another position sensor (…) can be used, for example, after mechanical retrieval from a mask library or during scanning. Figure 1 (Not explicitly depicted) The pattern forming apparatus MA is positioned accurately relative to the path of the radiation beam B. Typically, the movement of the support structure MT can be achieved using long-stroke modules (coarse positioning) and short-stroke modules (fine positioning) that form the portion of the first positioner PM. Similarly, the movement of the substrate stage WT can be achieved using long-stroke modules and short-stroke modules that form the portion of the second positioner PW. In the case of a stepper (relative to a scanner), the support structure MT may be connected only to the short-stroke actuator, or it may be fixed. The pattern forming apparatus MA and the substrate W can be aligned using pattern forming apparatus alignment marks M1, M2 and substrate alignment marks P1, P2. Although the illustrated substrate alignment marks occupy dedicated target portions, these marks can be located in the space between the target portions (these marks are called scribing alignment marks). Similarly, in cases where more than one die is disposed on the pattern forming apparatus MA, the pattern forming apparatus alignment marks can be located between the dies.

[0061] The described apparatus can be used in at least one of the following modes. In stepping mode, while the support structure MT and substrate stage WT are kept substantially stationary, a pattern imparting a radiation beam is projected onto the target portion C in a single exposure (i.e., a single static exposure). The substrate stage WT is then shifted along the X and / or Y directions, allowing exposure of different target portions C. In stepping mode, the maximum size of the exposure field limits the size of the target portion C imaged in a single static exposure. In scanning mode, while the support structure MT and substrate stage WT are scanned synchronously, a pattern imparting a radiation beam is projected onto the target portion C (i.e., a single dynamic exposure). The velocity and direction of the substrate stage WT relative to the support structure MT can be determined by the (reduced) magnification and image inversion characteristics of the projection system PS. In scanning mode, the maximum size of the exposure field limits the width of the target portion in a single dynamic exposure (along the non-scanning direction), while the length of the scanning motion determines the height of the target portion (along the scanning direction). In another mode, the support structure MT used to maintain the programmable patterning apparatus remains essentially stationary, and the pattern imparted by the radiation beam is projected onto the target portion C while the substrate stage WT is moved or scanned. In this mode, a pulsed radiation source is typically used, and the programmable patterning apparatus is updated as needed after each movement of the substrate stage WT or between consecutive radiation pulses during scanning. This mode of operation can be readily applied to maskless lithography utilizing a programmable patterning apparatus (e.g., a programmable mirror array of the type described above).

[0062] Combinations and / or variations of the above usage patterns can also be used, or completely different usage patterns.

[0063] The substrate can be processed before or after exposure in, for example, a track cell (a tool typically used to apply a resist layer to the substrate and develop the resist after exposure), a measurement tool, or an inspection tool. Where applicable, the disclosure herein can be applied to these and other substrate processing tools. Additionally, the substrate can be processed more than once, for example, to form a multilayer IC, such that the term substrate as used herein can also refer to a substrate that already includes multiple processed layers.

[0064] The terms “radiation” and “beam” used in this article for lithography are used to cover all types of electromagnetic radiation, including ultraviolet (UV) or deep ultraviolet (DUV) radiation (e.g., wavelengths of 365, 248, 193, 157, or 126 nm) and extreme ultraviolet (EUV) radiation (e.g., wavelengths in the range of 5–20 nm), as well as particle beams (such as ion beams or electron beams).

[0065] Various patterns on or provided by a pattern forming apparatus can have different process windows, i.e., the processing variable space that produces patterns within specifications. Examples of pattern specifications associated with potential systematic defects include necking checks, line retraction, line thinning, CD, edge placement, overlap, resist top loss, resist undercut, and / or bridging. Process windows for patterns or regions on a pattern forming apparatus can be obtained by merging (e.g., overlapping) the process windows of each individual pattern. The process window boundaries of a set of patterns include the boundaries of the process windows of some individual patterns. In other words, these individual patterns limit the process windows of the set of patterns.

[0066] like Figure 2 As shown, a lithography apparatus LA can form part of a lithography unit LC (sometimes also called a lithography cell or cluster), which also includes equipment for performing pre-exposure and post-exposure processes on the substrate. Typically, this equipment includes one or more spin coaters SC for depositing one or more resist layers, one or more developers for developing the post-exposure resist, one or more chillers CH, and / or one or more bake plates BK. A substrate transport device or robot RO picks up one or more substrates from input / output ports I / O1, I / O2, moves the substrates between different process units, and delivers the substrates to the lithography apparatus's feed stage LB. These devices, often collectively referred to as tracks, are under the control of a track control unit TCU, which in turn is controlled by a management control system SCS, which in turn controls the lithography apparatus via the lithography control unit LACU. Therefore, different devices can be operated to maximize throughput and processing efficiency.

[0067] To correctly and consistently expose a substrate exposed by a photolithography apparatus, and / or to monitor a portion of a patterning process (e.g., a device fabrication process) that includes at least one pattern transfer step (e.g., an optical lithography step), it is desirable to examine the substrate or other object to measure or determine one or more properties, such as alignment, overlap (which may be, for example, between structures in a stacked layer, or between structures in the same layer that have been separately provided to said layers by, for example, a double patterning process), line thickness, critical dimension (CD), focus offset, or material properties, etc. Therefore, the fabrication facility in which the photolithography unit LC is located typically also includes a metrology system that measures the substrate W (which has been processed in the photolithography unit) Figure 1 Some or all of the objects in the lithography unit or other objects in the lithography unit. The measurement system can be part of the lithography unit LC, for example, it can be part of the lithography equipment LA (such as the alignment sensor AS). Figure 1 )).

[0068] For example, one or more measurement parameters may include alignment, overlap between successive layers formed in or on a patterned substrate, critical dimensions (CD) (e.g., critical linewidth) of features formed in or on a patterned substrate, focusing or focusing errors of the optical lithography step, dose or dose errors of the optical lithography step, optical aberrations of the optical lithography step, and so on. Such measurements are typically performed on one or more dedicated measurement targets set on the substrate. The measurement may be performed after resist development, but before etching, after etching, after deposition, and / or at other times.

[0069] Various techniques exist for measuring structures formed during the patterning process, including the use of scanning electron microscopy, image-based measurement tools, and / or various specialized instruments. A rapid and non-invasive specialized measurement tool is one that directs a radiation beam onto a target on a substrate surface and measures the properties of the scattered (diffracted / reflected) beam. By evaluating one or more properties of the radiation scattered by the substrate, one or more properties of the substrate can be determined. Traditionally, this can be referred to as diffraction-based measurement. Applications of this diffraction-based measurement include measurements of alignment, overlap, etc. For example, alignment and / or overlap can be measured by comparing portions of the diffraction spectrum (e.g., comparing different diffraction orders in the diffraction spectrum of a periodic grating).

[0070] Therefore, during device fabrication processes (e.g., patterning processes, or photolithography processes), substrates or other objects may undergo various types of measurements during or after the process. These measurements can determine the presence of defects in a particular substrate, establish adjustments to the processes and equipment used in them (e.g., aligning two layers on a substrate or aligning a patterning apparatus with a substrate), measure the performance of processes and equipment, or be used for other purposes. Examples of measurements include optical imaging (e.g., optical microscopy), non-imaging optical measurements (e.g., diffraction-based measurements, such as ASML Orion metrology tools, ASML YieldStar metrology tools, ASML SMASH metrology systems), mechanical measurements (e.g., profilometry using styluses, atomic force microscopy (AFM)), and / or non-optical imaging (e.g., scanning electron microscopy (SEM)).

[0071] Measurement results can be provided directly or indirectly to the management control system (SCS). If an error is detected, adjustments can be made to the exposure of subsequent substrates (especially if the inspection can be completed quickly and sufficiently so that one or more other substrates in the batch are still exposed) and / or to the subsequent exposure of already exposed substrates. Furthermore, already exposed substrates can be stripped and reworked to increase yield, or discarded, thus avoiding further processing of substrates known to be defective. In cases where only some target portions of the substrate are defective, further exposure can be performed only on those target portions that meet specifications. Other manufacturing process adjustments are conceivable.

[0072] A measurement system can be used to determine one or more properties of a substrate structure, and in particular, how one or more properties vary between different substrate structures, or how different layers of the same substrate structure vary between layers. The measurement system can be integrated into a lithography apparatus (LA) or a lithography unit (LC), or it can be a stand-alone device.

[0073] To achieve measurement, one or more targets are typically fabricated on a substrate. These targets are usually specially designed and may include periodic structures. For example, a target on the substrate may include one or more 1-D periodic structures (e.g., geometric features, such as gratings) printed such that, after development, the periodic structure features consist of solid resist lines. As another example, a target may include one or more 2-D periodic structures (e.g., gratings) printed such that, after development, the one or more periodic structures consist of solid resist pillars or vias in the resist. Alternatively, the strips, pillars, or vias may be etched into the substrate (e.g., etched into one or more layers on the substrate).

[0074] Figure 3 An exemplary measurement system 10 is depicted that can be used to detect alignment, overlap, and / or perform other measurement operations. This measurement system includes a radiation or illumination source 2 that projects or otherwise irradiates radiation onto a substrate W (e.g., which may typically include measurement markers). The redirected radiation is passed to a sensor 4 (such as a spectrometer detector and / or other sensors) that measures the spectrum (intensity as a function of wavelength) of the specularly reflected and / or diffracted radiation, for example... Figure 4 The graph on the left is shown. The sensor can generate a measurement signal that transmits measurement data indicating the properties of the reflected radiation. Based on this data, the structure or profile of the detected spectrum can be generated by one or more processors (a generalized example of which is shown in...). Figure 4 (As shown in the image) or through other operations to rebuild.

[0075] like Figure 1The lithography equipment LA shown in the image can provide one or more substrate stages ( Figure 3 or Figure 4 (Not shown) to hold the substrate W during measurement operations. One or more substrate stages may be in form with Figure 1 The substrate stages WT (WTa or WTb or both) are similar or identical. In the example where the inspection system 10 is integrated with a lithography apparatus, they can even be the same substrate stages. Approximate and fine locators can be provided and configured to accurately position the substrate relative to the measurement optics. For example, various sensors and actuators are provided to obtain the position of the target portion of the structure (e.g., a measurement mark) and bring it to a position below the objective lens. Typically, many measurements will be performed on the target portion of the structure at different locations across the substrate W. The substrate support can move in the X and Y directions to obtain different targets and in the Z direction to obtain the desired position of the target portion relative to the focus of the optics. For example, in practice, it is convenient to envision and describe operations such as bringing the objective lens to different positions relative to the substrate when the optics can remain substantially stationary (typically in the X and Y directions, but possibly also in the Z direction) and the substrate moves. Assuming the relative positions of the substrate and the optical system are correct, it is in principle irrelevant which one of them is moving, or if both are moving, or if a combination of parts of the optical system is moving (e.g., in the Z and / or tilt directions) while the rest of the optical system is stationary, and the substrate is moving (e.g., in the X and Y directions, but optionally also in the Z and / or tilt directions).

[0076] For typical metrological measurements, the measurement target 30 on the substrate W can be a 1-D grating, which is printed such that, after development, the strip is formed of solid resist lines (e.g., which may be covered by a deposition layer) and / or other materials. Alternatively, the target 30 can be a 2-D grating, which is printed such that, after development, the grating is formed of solid resist pillars and / or other features in the resist.

[0077] Strips, pillars, vias, and / or other features may be etched into or on a substrate (e.g., etched into one or more layers on the substrate), deposited on the substrate, covered by a deposited layer, and / or have other properties. Target 30 (e.g., stripes, pillars, vias, etc.) is sensitive to processing variations during the patterning process (e.g., optical aberrations, focus variations, dose variations, etc., in a photolithography projection device in a projection system), causing these process variations to manifest as variations in target 30. Therefore, measurement data from target 30 can be used to determine adjustments to one or more manufacturing processes and / or serve as the basis for making actual adjustments.

[0078] For example, measurement data from target 30 can indicate the alignment and / or overlap of layers in a semiconductor device. The measurement data from target 30 can (e.g., via one or more processors PRO and / or other processors) be used to determine one or more semiconductor device fabrication process parameters based on overlap, and to determine adjustments to semiconductor device fabrication equipment based on one or more determined semiconductor device fabrication process parameters. For example, in some implementations, this can include stage position adjustment, or it can include determining adjustments to mask design, measurement target design, semiconductor device design, radiation intensity, angle of incidence of radiation, wavelength of radiation, pupil size and / or shape, resist material, and / or other process parameters.

[0079] Figure 5 The diagram shows... Figure 3 A planar diagram of the typical measurement target 30 (e.g., a measurement marker) and the range of a typical radiation irradiation point S in the system. Typically, to obtain a diffraction spectrum without interference from surrounding structures, in one embodiment, the target 30 is a periodic structure (e.g., a grating) larger than the width (e.g., diameter) of the irradiation point S. The width of the point S may be smaller than the width and length of the target. In other words, the target is irradiated "unfilled," and the diffraction signal essentially contains no signal from product features or the like outside the target itself. For example, the irradiation arrangement can be configured to provide uniform intensity irradiation across the back focal plane of the objective. Alternatively, irradiation can be provided from both on-axis and off-axis directions, for example, by including a spot size selector in the irradiation path.

[0080] Figure 6 The illustration shows a virtual component (in) Figure 7 A measurement system 600 (shown and described below) is provided, in which the virtual component is configured to block higher-order diffraction radiation to prevent disruption of alignment measurements during semiconductor manufacturing measurement processes. For example, in some implementations, system 600 may be configured to minimize and / or eliminate the need for electrical hardware for blocking higher-order diffraction radiation. It should be noted that system 600 is merely one representative example of several different possible types of systems that can utilize diffraction radiation with orbital angular momentum (which may or may not have some or all of the same components and / or may function in slightly different ways). Examples of such systems include ASML's Orion system, ASML's YieldStar system, ASML's Aurora system, and / or other systems. System 600 is related to the above description... Figure 3 The system described is the same as or similar to system 10, wherein one or more components of system 600 are similar to and / or the same as one or more components of system 10 (and Figure 6Several possible additional components of the system are shown. In some implementations, one or more components of system 600 may replace one or more components of system 10, be used with one or more components of system 10, and / or otherwise extend one or more components of system 10. System 600 includes a radiation source 612 (e.g., with...). Figure 3 Source 2 shown is similar to and / or the same as the source 2 shown), sensor 604 and / or sensor 610 (e.g., with the source 2 shown to be similar to and / or the same as the source 2 shown to the sensor 604 and / or sensor 610 shown to the source 610). Figure 3 The sensor shown is similar to and / or the same as 4), and one or more processors PRO (with Figure 3 The processor PRO shown is similar to and / or identical to that shown, as well as various lenses, beam splitters, and / or other components (see, for example, the various unlabeled boxes in system 600). One or more processor PROs are operatively connected to sensor 604, sensor 610, and / or other components of system 600.

[0081] Figure 6 The illustration shows an illumination branch 625 of system 600, which includes a radiation source 612; an overlap detection branch 660, which includes a sensor 604 and one or more processors PRO; a focusing branch 650; an alignment branch 680, which has a sensor 610 (which may be the same as or similar to sensor 604) and one or more processors PRO (which may be similar to and / or the same as the processor in overlap detection branch 660); an objective lens 690; and / or other components. In some implementations, the components of system 600 are formed as part of an alignment sensor configured for use in a semiconductor manufacturing process.

[0082] Figure 6 The diagram also illustrates a measurement target 30, which may include one or more measurement marks (such as diffraction grating targets) formed in a substrate 602 (such as a semiconductor wafer), collectively referred to as measurement target 30. Target 30 may include one or more structures in a patterned substrate capable of providing diffraction signals. For example, one or more targets 30 may be included in a layer of a substrate in a semiconductor device structure. In some implementations, features include geometric features (such as 1D or 2D features), and / or other geometric features. As several non-limiting examples, features may include gratings, lines, edges, a series of precision ground lines and / or edges, and / or other features. Further information regarding measurement target 30 is described below.

[0083] Various lenses (in) Figure 6The system 600 includes an exemplary objective lens 690, reflectors, and other optical components configured to receive, transmit, reflect, focus, and / or perform other operations on illumination generated by illumination source 612, illumination focused by focusing branch 650, illumination received by detection branch 660, illumination received by alignment branch 680, and / or illumination used by other parts of the system 600. These various lenses, reflectors, and / or other optical components may include any type of lens, reflector, and / or optical component configured to allow the system 600 to function as described herein. For example, objective lens 690 may be formed of any transparent material and has a curved surface configured to focus one or more radiation spots or otherwise focus them onto target(s)30(s). Various lenses, reflectors, optical elements, beam splitters, and other optical elements may be positioned at any location and / or at any angle relative to each other that allows the system 600 to function as described herein. This may include positioning at a specific relative distance between elements, a specific angle between elements, etc. In some implementations, various lenses, reflectors, optical elements, beam splitters, and other optical components are positioned relative to each other in system 600 via structural members, clips, jigs, screws, nuts, bolts, adhesives, and / or other mechanical means. In some implementations, various of the lenses, reflectors, optical elements, beam splitters, and other optical elements can move relative to each other. For example, movement can be configured to adjust the position of a corresponding illumination spot on one or more targets 30. In some implementations, movement includes tilting, translating, or otherwise changing the distance between the various lenses, reflectors, and other optical components. Other examples of movement are conceivable.

[0084] In some implementations, movement can be electronically controlled by a processor (such as a processor PRO). The processor PRO can be included in the computing system CS ( Figure 9 In, and can be based on computer or machine-readable instructions (e.g., as follows regarding) Figure 9 As described above, the system operates by transmitting electronic signals between individual components, transmitting data between individual components of system 600, transmitting values ​​between individual components, and / or other communications. Components of system 600 can communicate wired or wirelessly via a network (such as the Internet or an Internet combined with various other networks such as local area networks, cellular networks or personal area networks, internal organizational networks, and / or other networks).

[0085] In some implementations, one or more actuators ( Figure 6(Not shown) One or more components of system 600 may be coupled to and configured to move the one or more components. Actuators may be coupled to one or more components of system 600 via adhesives, clips, jigs, screws, collars, and / or other mechanisms. Actuators may be configured to be controlled electronically. Each actuator may be configured to convert an electrical signal into a mechanical displacement. This mechanical displacement is configured to move a component of system 600. As an example, one or more actuators may be piezoelectric. One or more processors PRO may be configured to control the actuators. One or more processors PRO may be configured to individually control each of the one or more actuators.

[0086] Figure 6 The number of lenses, reflectors, and / or other optical components shown is not intended to be limiting. The principles described herein can be extended such that in some implementations, system 600 includes additional or fewer lenses, reflectors, and / or other optical components.

[0087] Radiation source 612 is configured to generate radiation. The radiation may include illumination such as light and / or other radiation. In some implementations, the radiation from radiation source 612 includes a Gaussian radiation beam and / or other radiation. The radiation may have a target wavelength and / or wavelength range, a target intensity, and / or other characteristics. The target wavelength and / or wavelength range, target intensity, etc., may be input and / or selected by the user, determined by system 600 based on previous measurements, and / or otherwise determined. In some implementations, the radiation includes light and / or other radiation. In some implementations, the light includes visible light, infrared light, near-infrared light, and / or other light. In some implementations, the radiation may be any radiation suitable for interferometric measurements.

[0088] A measurement target 30 in one or more layers of a patterned substrate 602 is configured to be irradiated with radiation. The measurement target 30 is configured to diffract the radiation. In some implementations, the measurement target 30 (which is configured to diffract radiation from a radiation source) is different from structures in one or more layers that are adjacent to, close to, and / or surrounding the measurement target, which are unable to produce diffracted radiation.

[0089] Sensors 604 and / or 610 (they can be separate sensors or sensors within the same sensor suite, even if they are in...) Figure 6 (Drawn separately) is configured to receive diffracted radiation from target 30 and generate a measurement signal. The measurement signal is generated based on the diffracted radiation received from target 30 and / or other information. In some implementations, radiation sensors 604 and / or 610 are configured to detect the phase, intensity, wavelength, and / or polarization of the diffracted radiation received from target 30 and generate a measurement signal.

[0090] In some implementations, the measurement signal includes measurement information about the target 30. For example, the measurement signal may be an alignment signal that includes alignment measurement information, and / or other measurement signals. The measurement information can be determined using the principles of interferometry and / or other principles. The measurement signal includes electronic signals that represent and / or otherwise correspond to radiation reflected from the target 30. For example, the measurement signal may indicate measurements and / or other information associated with a diffraction grating. Generating the measurement signal includes sensing diffracted radiation and converting the sensed diffracted radiation into an electronic signal. In some implementations, generating the measurement signal includes sensing different portions of diffracted radiation from different regions and / or different geometries of the target 30, and / or from multiple targets, and combining the different portions of reflected radiation to form the measurement signal.

[0091] Sensors 604 and / or 610 are configured to output phase and intensity data of the diffracted radiation. For example, the phase and intensity data of the diffracted radiation can be part of the measurement information in the measurement signal. The phase data includes different energies associated with different diffraction orders of the diffracted radiation. Different diffraction orders can be “decoupled” so that the individual energy for a single diffraction order can be determined. This information can also be used to similarly decouple different diffraction order intensity data (as described below). The phase data conveys the alignment position of the layer relative to the average position from all wavelengths for all diffraction orders entering the pupil of the measurement system; and the relative diffraction efficiency corresponding to all diffraction orders entering the pupil. The intensity data includes intensity imbalance with diffraction order mixing.

[0092] Intensity information is configured to measure and correct for sub-nanometer alignment offsets caused by mark asymmetry. This correction can be performed by calculating the intensity imbalance between the positive and negative first-order diffraction radiation from the mark, as described herein. Undesirably, for certain combinations of diffraction wavelengths and the periodicity of the measurement mark (target 30), the output intensity information is corrupted by higher-order diffraction radiation (e.g., positive and negative second, third, fourth, etc. diffraction orders). Electrical hardware 699 (e.g., alignment rotation optics motor and / or other components) can be used to prevent leakage by physically blocking higher-order diffraction radiation from reaching the alignment sensor through movement. However, system 600 has a virtual component 700 configured to minimize and / or eliminate the need for electrical hardware 699 (see [link to documentation]). Figure 7 ).

[0093] Figure 7 The diagram shows... Figure 6 An exemplary embodiment of the virtual component 700 of the measurement system 600 shown herein. The virtual component 700 may be composed of one or more processors PRO (as described herein). Figure 3 , Figure 6 , Figure 9 ) and / or computing system CS ( Figure 9 One or more processors PRO and / or computing system CS can be provided with a radiation sensor (e.g., Figure 6 The sensors 604 and / or 610 shown are operatively connected and configured to execute a trained machine learning model 702 via machine-readable instructions. The virtual component 700 is and / or includes the trained electronic model 702 and / or other components. The model 702 is configured (e.g., from...) Figure 6 The sensor 604 and / or sensor 610 shown receive phase data 701, intensity data 703, and / or other information from the diffracted radiation. As described above, the phase data 701 includes different energies associated with different diffraction orders of the diffracted radiation. The phase data 701 conveys information for input into the measurement system (e.g., Figure 6 The system 600 shown has the alignment positions of all diffraction orders of the pupil, layers relative to the average positions from all wavelengths; and the relative diffraction efficiency corresponding to all diffraction orders entering the pupil. Intensity data 703 includes intensity imbalance with diffraction order mixing.

[0094] Model 702 is configured to determine a subset of intensity data 703 associated with the positive and negative first-order diffraction radiation incident on a radiation sensor, based on phase data 701 of the same diffraction radiation. Determining the subset of intensity data 703 includes excluding intensity data associated with destructive higher-order diffraction radiation (e.g., "leakage"). Data associated with destructive higher-order diffraction radiation can be identified based on the energy associated with the phase of the higher diffraction order. Model 702 is configured to determine (and output) the intensity difference between the positive and negative first-order diffraction radiation based (only) on the subset of intensity data. This intensity difference includes an intensity imbalance 704 (e.g., only for first-order diffraction radiation), which can be used to determine the substrate layer (e.g., having measurement markers (e.g., Figure 6 Alignment and / or other information of the layers of measurement markers in target 30 shown.

[0095] In some implementations, the trained machine learning model 702 includes a multivariate (or polynomial) regression model, a decision tree regressor, a random forest regressor, a multilayer perceptron (MLP) regressor, a stochastic gradient descent (SGD) regressor, an extreme gradient boosting (XGB) regressor, one or more neural networks, one or more deep learning elements, or any combination thereof. A multivariate regression model is a machine learning model configured to determine and / or otherwise describe a nonlinear relationship between variables by fitting a nonlinear regression line. A decision tree regressor is a regression model configured to use a tree structure to subdivide an input dataset into increasingly smaller parts to generate predictions. A random forest generator is an algorithm that includes multiple decision trees for generating predictions based on an input dataset. A multilayer perceptron includes neural networks configured for class classification, regression problems, and / or other uses. A stochastic gradient descent regressor includes learning routines with various loss functions and / or penalties for fitting a linear regression model. An extreme gradient boosting regressor includes a collection of individually weak predictive models that make very few assumptions about the input data; when these predictive models operate together, they can make more accurate predictions. Figure 7 In the example shown, model 702 may include a multinomial regression model.

[0096] In some implementations, a trained machine learning model 702 is trained 710 by providing a trained machine learning model 712 with phase channel alignment positions relative to the average position from all wavelengths for all diffraction orders entering the pupil of the measurement system. Training 710 also includes determining substrate quality based on the relative diffraction efficiency corresponding to all radiative diffraction orders entering the pupil, and providing the substrate quality 714 to the untrained machine learning model. A fundamental truth intensity channel imbalance with order mixing 716 is also provided to the untrained machine learning model, and a fundamental truth intensity channel imbalance for first-order diffraction radiation only 718 is provided. In some implementations, the fundamental truth intensity channel imbalance for first-order diffraction radiation only provided to the untrained machine learning model is determined by simulation and / or other methods. Simulations can be performed using a rigorous coupled-wave analysis simulator and / or other simulators. In some implementations, an alignment-rotating optics motor is used to determine the fundamental truth intensity channel imbalance for only the first-order diffraction radiation, which is provided to the untrained machine learning model (e.g., by operating the measurement system in a conventional manner but physically blocking higher-order diffraction radiation from reaching the sensor, and providing the sensor output signal data (which does not include any data associated with higher-order diffraction radiation) as training input data to train the model). For example, the aforementioned alignment-rotating optics motor can be added to system 600 (or turned on / used) to generate training data, and removed and / or not used once the trained machine learning model 702 has been trained. In some implementations, the trained machine learning model 702 is trained by measuring multiple orthogonal polarization states for each diffraction order and providing these polarization states, along with the polarization specific to the first-order diffraction radiation, to the untrained machine learning model as further training input.

[0097] Then, a new intensity channel imbalance with order mixing can be provided to an untrained machine learning model (e.g., for testing model training), which is configured to determine a new intensity imbalance of the first diffraction order in the diffraction order entering the pupil based on the new intensity channel imbalance with order mixing. Training includes determining whether the new intensity imbalance is accurate in response to the difference between the determined result and the underlying true intensity channel imbalance of the first-order diffraction radiation being less than a threshold amount; and adjusting one or more parameters of the untrained machine learning model based on one or more accuracy determination results to form a trained machine learning model 702. In some implementations, for example, the threshold amount is 1e-4 (i.e., if the difference between the predicted result and the corresponding underlying true intensity imbalance of the first-order diffraction radiation is less than 1e-4, the predicted result is marked as accurate). In some implementations, the multinomial regression model degree of the model can be changed to generate another version of the model. Furthermore, for example, a neural network can be used as a different version of the model. In some implementations, the threshold 1e-4 indicates the accuracy with which the model can predict relative to the true value, and increasing the threshold can affect the quality of the prediction.

[0098] In some implementations, System 600 ( Figure 6 ) including electric hardware 699 ( Figure 6 -It may include an alignment rotating optics motor and / or other components), the electrical hardware being configured to block a portion of higher-order diffraction radiation from reaching the radiation sensor and virtual component 700, as described above. However, the trained machine learning model 702 is configured such that system 600 requires minimal use of electrical hardware 699 to block a portion of higher-order diffraction radiation from reaching the radiation sensor (e.g., Figure 6 The sensors 604 and / or 610 shown herein are examples of this. As described above, system 600 can be configured such that the intensity, combined with the phase of the diffracted radiation, is used to detect and correct sub-nanometer alignment offsets caused by asymmetries in the measurement target. For example, the intensity imbalance between positive first-order and negative first-order diffracted radiation can be configured to be used by one or more processors PRO to determine the alignment of the layer, and further, based on the alignment, to adjust the semiconductor device fabrication process.

[0099] Figure 8 The diagram illustrates a measurement method 800. In some implementations, for example, method 800 is performed as part of an alignment sensing operation in a semiconductor device manufacturing process. In some implementations, for example, one or more operations of method 800 can be performed at any time. Figure 6 The system 600 shown Figure 3 System 10 shown, computer system (e.g., such as...) Figure 9This method is implemented in or by the computer systems shown and described below, and / or in other systems or by said other systems. In some implementations, method 800 includes irradiating (operation 802) a measurement target with radiation such that the measurement target diffracts radiation; outputting (operation 804) phase data and intensity data of the diffracted radiation; determining (operation 806) a subset of intensity data associated with the positive first-order diffracted radiation and the negative first-order diffracted radiation; determining (operation 808) the intensity difference between the positive first-order diffracted radiation and the negative first-order diffracted radiation; and determining (operation 810) alignment. The operation of method 800 is intended to be illustrative. In some implementations, method 800 may be done with one or more additional operations not described, and / or without the one or more operations discussed. For example, in some implementations, method 800 may include additional operations including determining adjustments to the semiconductor device manufacturing process. Additionally, in Figure 8 The order of operations of the method 800 shown and described herein is not intended to be limiting.

[0100] In some implementations, one or more portions of method 800 may be implemented in and / or controlled by one or more processing devices (e.g., digital processors, analog processors, digital circuits designed to process information, analog circuits designed to process information, state machines, and / or other mechanical devices for electronically processing information). The one or more processing devices may include one or more means that perform some or all of the operations of method 800 in response to instructions electronically stored on an electronic storage medium. The one or more processing devices may include one or more means configured by hardware, firmware, and / or software specifically for performing one or more operations of method 800 (e.g., see below with...). Figure 9 (Related discussion).

[0101] At operation 802, a measurement target in one or more layers of a patterned substrate is irradiated with radiation from a radiation source. Operation 802 may include generating radiation using a radiation source and one or more lenses and directing the radiation toward the measurement target. In some implementations, the radiation from the radiation source includes a Gaussian radiation beam and / or other radiation. For example, the patterned substrate may be a semiconductor wafer and / or other substrate. In some implementations, method 800 is configured for use with a semiconductor wafer and in a semiconductor manufacturing process. The measurement target is configured to diffract radiation. In some implementations, the measurement target (which is configured to diffract radiation from the radiation source) differs from structures in one or more layers that are adjacent to, close to, and / or surrounding the measurement target (which are unable to generate diffracted radiation).

[0102] For example, the measurement target may include a grating and / or other geometries. In some implementations, the measurement target includes a first grating in a first layer of one or more layers of a patterned substrate and a second grating in a second layer of one or more layers of a patterned substrate, which together form the measurement target. In some implementations, operation 802 is performed by... Figure 3 The radiation source 2 shown is Figure 6 The same or similar radiation source as source 612, the measurement target (such as target 30 shown in these same figures), and / or the other components mentioned above are used to perform the measurement.

[0103] At operation 804, phase and intensity data of the diffracted radiation are output from the radiation sensor. For example, the output may be in the form of an electronic measurement signal and / or other forms. In some implementations, the measurement signal includes measurement information about (multiple) targets. For example, the measurement signal may be an alignment signal and / or other measurement signals that include alignment measurement information. The measurement information (e.g., alignment values ​​and / or other information) can be determined using the principles of interferometry and / or other principles.

[0104] The measurement signal includes an electronic signal that represents and / or otherwise corresponds to radiation diffracted from the targets(s). For example, the measurement signal may indicate measurements and / or other information associated with a diffraction grating target. Generating the measurement signal involves sensing the diffracted radiation and converting the sensed reflected radiation into an electronic signal. In some implementations, generating the measurement signal involves sensing different portions of reflected radiation from different regions and / or different geometries of the target, and / or from multiple targets, and combining the different portions of the reflected radiation to form the measurement signal.

[0105] In some implementations, operation 804 may include detecting the phase, intensity, wavelength, and / or polarization of diffracted radiation received from the measurement target using a radiation sensor, and generating a measurement signal based on the orbital angular momentum and the phase, intensity, wavelength, and / or polarization. In some implementations, detecting reflected radiation includes detecting one or more phase and / or amplitude (intensity) shifts in reflected radiation from one or more geometric features of the target(s). One or more phase and / or amplitude shifts correspond to one or more dimensions of the target. For example, the phase and / or amplitude of reflected radiation from one side of the target may be different from the phase and / or amplitude of reflected radiation from the other side of the target.

[0106] Detecting one or more phase and / or amplitude (intensity) shifts in reflected radiation from a target involves measuring local phase shifts (e.g., local phase increments) and / or amplitude variations corresponding to different portions of the target. For example, reflected radiation from a specific region of the target may include a sinusoidal waveform with a specific phase and / or amplitude. Reflected radiation from different regions of the target (or from targets in different layers) may also include sinusoidal waveforms, but with different phases and / or amplitudes. The detected reflected radiation also includes measuring phase and / or amplitude differences in reflected radiation at different diffraction orders. The detection of one or more local phase and / or amplitude shifts can be performed using, for example, the Hilbert transform and / or other techniques. Interferometry techniques and / or other operations can be used to measure phase and / or amplitude differences in reflected radiation at different diffraction orders.

[0107] As described above, phase and intensity data of the diffracted radiation are output from the radiation sensor. The phase data includes the different energies associated with different diffraction orders of the diffracted radiation. The phase data conveys the alignment position of the layer relative to the average position from all wavelengths for all diffraction orders entering the pupil of the measurement system; and the relative diffraction efficiency corresponding to all diffraction orders entering the pupil. The intensity data includes the intensity imbalance with diffraction order mixing. Operation 804 can be controlled by... Figure 3 , Figure 6 and Figure 9 The detector 4, sensor 604, sensor 610, and / or processor PRO (described below) and / or other components shown are similar to and / or identical to the radiation sensor to perform this function.

[0108] Operations 806, 808, and / or other operations can be performed by one or more processors operatively connected to the radiation sensor and configured to execute a trained machine learning model (e.g., included in a computing system such as, as described below). Figure 9 The computational system described is executed in CS. In some implementations, the trained machine learning model includes multinomial regression models, decision tree regressors, random forest regressors, multilayer perceptron (MLP) regressors, stochastic gradient descent (SGD) regressors, extreme gradient boosting (XGB) regressors, or any combination thereof.

[0109] Operation 806 includes using a trained machine learning model (e.g., a multinomial regression model) to determine, based on phase data, a subset of intensity data associated with the positive and negative first-order diffraction radiation incident on the radiation sensor. Determining this subset of intensity data includes excluding intensity data associated with destructive higher-order diffraction radiation (e.g., "leakage"). For example, data associated with destructive higher-order diffraction radiation may be identified based on the energy associated with the phase of the higher diffraction order. In some implementations, operation 806 includes physically blocking, or controlling electrically blocked, a portion of the higher-order diffraction radiation reaching the radiation sensor using the electrical hardware of the measurement system (e.g., alignment rotating optics motor and / or other components). In some implementations, operation 806 is performed by... Figure 3 and Figure 6 The processor PRO shown and described above (and) Figure 9 The processor (PRO) shown in the diagram and described below is the same as or similar to one or more processors. Note that, ideally, the trained machine learning model is configured such that the measurement system requires minimal or no electrical hardware to block portions of higher-order diffracted radiation from reaching the radiation sensor.

[0110] Operation 808 includes using a trained machine learning model to determine the intensity difference between positive and negative first-order diffraction radiation based on a subset of intensity data. For example, the intensity difference may include intensity imbalance. As described below, operation 810 of method 800 may include using one or more processors to determine layer alignment based on intensity imbalance and / or other information. For example, intensity may be combined with the phase of diffraction radiation to detect and correct sub-nanometer alignment shifts caused by measurement target asymmetry. In some implementations, operation 808 is performed by... Figure 3 and Figure 6 The processor PRO shown and described above (and) Figure 9 One or more processors that are the same as or similar to the processor PRO shown in the diagram and described below will execute the processor.

[0111] In some implementations, the trained machine learning model used in operations 806 and / or 808 is trained by providing the untrained machine learning model with phase channel alignment positions relative to the average positions from all wavelengths for all diffraction orders entering the pupil of the measurement system. Substrate quality is determined based on the relative diffraction efficiency corresponding to all radiative diffraction orders entering the pupil, and this substrate quality is provided to the untrained machine learning model. A fundamental truth intensity channel imbalance with order mixing is added to the untrained machine learning model, and a fundamental truth intensity channel imbalance for only first-order diffraction radiation is also provided to the untrained machine learning model. For example, the fundamental truth intensity channel imbalance for only first-order diffraction radiation provided to the untrained machine learning model can be determined by simulation, using an alignment rotating optics motor, and / or other methods. The alignment rotating optics motor can be used to generate training data, and then, once the trained machine learning model has been trained, the alignment rotating optics motor is removed from and / or not used in the measurement system.

[0112] A new intensity channel imbalance with order mixing is provided to an untrained machine learning model, which then determines a new intensity imbalance for the first diffraction order entering the pupil based on this new intensity channel imbalance with order mixing. The accuracy of the new intensity imbalance determination is determined in response to the difference between the determined result and the underlying true intensity channel imbalance of the first-order diffraction radiation being less than a threshold amount (e.g., 1e-4). One or more parameters of the untrained machine learning model are adjusted based on one or more accuracy determination results to form a trained machine learning model. In some implementations, the trained machine learning model is further trained by measuring multiple orthogonal polarization states for each diffraction order and providing these polarization states, along with the polarization specific to the first-order diffraction radiation, to the untrained machine learning model as further training input.

[0113] At operation 810, the alignment of one or more layers of the substrate is determined based on intensity imbalance, additional information in the measurement signal, and / or other information. In some implementations, operation 810 includes determining adjustments to the semiconductor device fabrication process. In some implementations, operation 810 includes determining one or more semiconductor device fabrication process parameters. One or more semiconductor device fabrication process parameters may be determined based on one or more detected phase and / or amplitude changes, alignment values, and / or other information. One or more parameters may include parameters of radiation (radiation used for measurement), alignment values, measurement inspection positions on layers of the semiconductor device structure, radiation beam trajectories across targets, and / or other parameters. In some implementations, process parameters may be broadly interpreted to include stage position, mask design, measurement target design, semiconductor device design, intensity of radiation (for exposure of resist, etc.), angle of incidence of radiation, wavelength of radiation (for exposure of resist, etc.), pupil size and / or shape, resist material, and / or other parameters.

[0114] In some implementations, operation 810 includes determining process adjustments based on one or more determined semiconductor device manufacturing process parameters, adjusting the semiconductor device manufacturing equipment based on the determined adjustments, and / or other operations. For example, if a determined measurement result is outside the process tolerance, the out-of-tolerance measurement result may be caused by one or more manufacturing processes where the process parameters have drifted and / or otherwise changed, causing the process to no longer produce acceptable devices (e.g., the measurement result may violate an acceptable threshold). One or more new or adjusted process parameters may be determined based on the measurement results. The new or adjusted process parameters may be configured to enable the manufacturing process to produce acceptable devices again.

[0115] For example, new or adjusted process parameters can bring previously unacceptable measurements back into acceptable ranges. The new or adjusted process parameters can be compared to existing parameters of a given process. For example, if a difference exists, that difference can be used to determine adjustments to the equipment used to produce the device (e.g., parameter "x" should be increased / decreased / changed to match the new or adjusted version of parameter "x" determined as part of operation 1506). In some implementations, operation 1506 may include electronically adjusting the equipment (e.g., based on the determined process parameters). For example, electronically adjusting the equipment may include sending electronic signals and / or other communications to the equipment, which results in changes to the equipment. For example, electronic adjustment may include changing settings and / or other adjustments on the equipment. In some implementations, operation 810 is performed by... Figure 3 and Figure 6 The processor PRO shown and described above (and) Figure 9One or more processors that are the same as or similar to the processor PRO shown in the diagram and described below will execute the processor.

[0116] Figure 9 A schematic diagram of an exemplary computer system CS that can be used for one or more of the operations described herein. The computer system CS includes a bus BS or other communication mechanism for communicating information, and a processor PRO coupled to the bus BS for processing information (or...). Figure 3 The processor PRO shown is similar to and / or is one of several processors. The computer system CS also includes main memory MM (such as random access memory (RAM)) or other dynamic storage devices coupled to the bus BS for storing instructions and information to be executed by the processor PRO. The main memory MM may also be used to store temporary variables or other intermediate information during instruction execution by the processor PRO. The computer system CS also includes read-only memory (ROM) or other static storage devices coupled to the bus BS for storing static information and instructions for the processor PRO. A storage device SD, such as a disk or optical disk, is provided and coupled to the bus BS for storing information and instructions.

[0117] A computer system (CS) can be coupled via a bus (BS) to a display (DS) for showing information to the computer user, such as a flat panel display, touchpad display, or cathode ray tube (CRT). Input devices (ID), including alphanumeric keys and other keys, are coupled to the bus (BS) for communicating information and command selections to the processor (PRO). Another type of user input device is a cursor controller (CC) for communicating directional information and command selections to the processor (PRO) and for controlling cursor movement on the display (DS), such as a mouse, trackball, or cursor direction keys. Such input devices typically have two degrees of freedom on two axes (i.e., a first axis (e.g., x) and a second axis (e.g., y)), allowing the device to specify its position in a plane. Touchpad (screen) displays can also be used as input devices.

[0118] In some embodiments, a computer system CS may perform some or all of the operations described herein in response to a processor PRO executing one or more sequences of one or more instructions included in main memory MM. These instructions may be read into main memory MM from another computer-readable medium, such as storage device SD. Execution of the instruction sequence included in main memory MM causes processor PRO to perform the process steps (operations) described herein. One or more processors arranged in a multiprocessor configuration may also be employed to execute the instruction sequence included in main memory MM. In some embodiments, a hard-wired circuit system may be used instead of or in combination with software instructions. Therefore, the description herein is not limited to any particular combination of hardware circuitry and software.

[0119] As used herein, the terms "computer-readable medium" or "machine-readable medium" refer to any medium that participates in providing instructions to a processor (PRO) for execution. Such media can take many forms, including, but not limited to, non-volatile media, volatile media, and transmission media. Non-volatile media include, for example, optical discs or magnetic disks, such as storage devices (SDs). Volatile media include dynamic memory, such as main memory (MMs). Transmission media include coaxial cables, copper wires, and optical fibers, including wires comprising a bus (BS). Transmission media can also take the form of sound waves or light waves, such as those generated during radio frequency (RF) and infrared (IR) data communications. Computer-readable media can be non-transitory, such as floppy disks, floppy disks, hard disks, magnetic tape, any other magnetic media, CD-ROMs, DVDs, any other optical media, punched cards, paper tape, any other physical media with a perforated pattern, RAM, PROMs and EPROMs, FLASH-EPROMs, any other memory chips, or cartridges. Non-transitory computer-readable media may have instructions recorded thereon. When executed by a computer, the instructions can perform any of the operations described herein. For example, a temporary computer-readable medium may include a carrier wave or other medium that propagates electromagnetic signals.

[0120] Various forms of computer-readable media can be used to carry one or more sequences of one or more instructions to the processor PRO for execution. For example, the instructions can initially be carried on the disk of a remote computer. The remote computer can load the instructions into its dynamic memory and transmit them over a telephone line using a modem. A modem local to the computer system CS can receive data over the telephone line and convert the data into an infrared signal using an infrared transmitter. An infrared detector coupled to the bus BS can receive the data carried in the infrared signal and place the data on the bus BS. The bus BS carries the data to the main memory MM, and the processor PRO retrieves and executes the instructions from the main memory MM. The instructions received by the main memory MM can optionally be stored on the storage device SD before or after execution by the processor PRO.

[0121] The computer system CS may also include a communication interface CI coupled to the bus BS. The communication interface CI provides bidirectional data communication coupling with a network link NDL connected to a local area network (LAN). For example, the communication interface CI may be an Integrated Services Digital Network (ISDN) card or a modem to provide data communication connectivity to a corresponding type of telephone line. As another example, the communication interface CI may be a LAN card to provide data communication connectivity to a compatible LAN. Wireless links may also be implemented. In any such implementation, the communication interface CI transmits and receives electrical, electromagnetic, or optical signals carrying digital data streams representing various types of information.

[0122] A network link (NDL) typically provides data communication to other data devices via one or more networks. For example, a network link NDL can provide a connection to a host computer (HC) via a local area network (LAN). This can include data communication services provided via a global packet data communication network (now commonly referred to as the "Internet" INT). A LAN (Internet) can use electrical, electromagnetic, or optical signals to carry digital data streams. Signals passing through various networks and signals on the network data link (NDL) and through the communication interface (CI) (which carries digital data to and from the computer system (CS)) are example forms of carriers for transmitting information.

[0123] A computer system CS can send messages and receive data, including process code, via a network, network data link (NDL), and communication interface (CI). In the Internet example, a host computer HC can transmit requested code for an application via the Internet (INT), network data link (NDL), local area network (LAN), and communication interface (CI). For example, such a downloaded application can provide all or part of the methods described herein. The received code can be executed by the processor PRO upon reception and / or stored in storage device (SD) or other non-volatile storage device for later execution. In this way, the computer system CS can obtain application code in carrier form.

[0124] Various embodiments of the system and method are disclosed in the subsequent list of numbered entries. Other features, characteristics, and exemplary technical solutions of this disclosure will be described below according to entries that may optionally be claimed in any combination: 1. A measurement system comprising: a radiation source configured to irradiate a measurement target in a layer of a patterned substrate with radiation, the measurement target configured to diffract radiation from the radiation source; a radiation sensor configured to output phase data and intensity data of the diffracted radiation, the phase data including different energies associated with different diffraction orders of the diffracted radiation; and one or more processors operatively connected to the radiation sensor and configured to execute a trained machine learning model via machine-readable instructions, the trained machine learning model being configured to: determine, based on the phase data, a subset of the intensity data associated with positive first-order and negative first-order diffracted radiation incident on the radiation sensor; and determine, based on the subset of the intensity data, an intensity difference between the positive first-order and negative first-order diffracted radiation. 2. The system according to item 1, wherein the intensity difference includes intensity imbalance, and wherein the one or more processors are further configured to determine the alignment of the layer based on the intensity imbalance. 3. The system according to any one of the preceding clauses, wherein determining the subset of the intensity data includes excluding intensity data associated with destructive higher-order diffraction radiation, the data associated with the destructive higher-order diffraction radiation being identified based on the energy associated with the phase of the higher diffraction order. 4. The system according to any one of the preceding clauses, wherein the destructive higher-order diffraction radiation includes leakage. 5. The system according to any one of the preceding clauses, wherein the phase data conveys the alignment position of the layer relative to the average position from all wavelengths for all diffraction orders entering the pupil of the measurement system; and the relative diffraction efficiency corresponding to all diffraction orders entering the pupil. 6. The system according to any one of the preceding clauses, wherein the intensity data includes intensity imbalance with diffraction order mixing. 7. The system according to any one of the preceding clauses, wherein the trained machine learning model comprises a multinomial regression model, a decision tree regressor, a random forest regressor, a multilayer perceptron (MLP) regressor, a stochastic gradient descent (SGD) regressor, an extreme gradient boosting (XGB) regressor, or any combination thereof. 8. The system according to any one of the preceding clauses, wherein the trained machine learning model includes a multinomial regression model. 9. The system according to any one of the preceding clauses, wherein the trained machine learning model is trained by: providing an untrained machine learning model with phase channel alignment positions relative to the average positions from all wavelengths for all diffraction orders entering the pupil of the measurement system; determining substrate quality based on the relative diffraction efficiency corresponding to all radiation diffraction orders entering the pupil, and providing the substrate quality to the untrained machine learning model; providing the untrained machine learning model with a fundamental truth intensity channel imbalance having order mixing, and providing the untrained machine learning model with a fundamental truth value only for first-order diffraction radiation. Intensity channel imbalance; providing a new intensity channel imbalance with order mixing to the untrained machine learning model, and using the untrained machine learning model, based on the new intensity channel imbalance with order mixing, determining a new intensity imbalance of the first diffraction order in the diffraction order entering the pupil; determining whether the new intensity imbalance determination result is accurate in response to the difference between the determination result and the underlying true intensity channel imbalance of the first-order diffraction radiation being less than a threshold amount; adjusting one or more parameters of the untrained machine learning model based on one or more accuracy determination results to form a trained machine learning model. 10. The system according to any one of the preceding clauses, wherein the threshold quantity is 1e-4. 11. The system according to any one of the preceding clauses, wherein the fundamental true intensity channel imbalance for first-order diffraction radiation provided to the untrained machine learning model is determined by simulation. 12. The system according to any one of the preceding clauses, wherein an alignment rotating optical device motor is used to determine the fundamental true intensity channel imbalance provided to the untrained machine learning model for first-order diffraction radiation only. 13. The system according to any one of the preceding clauses, wherein the alignment rotating optics motor is added to the system to generate training data, and is removed and / or not used once the training of the trained machine learning model is complete. 14. The system according to any one of the preceding clauses, wherein the trained machine learning model is further trained by measuring a plurality of orthogonal polarization states for each diffraction order and providing these polarization states, together with the polarization specific to the first-order diffraction radiation, to the untrained machine learning model as further training input. 15. The system according to any one of the preceding clauses further includes electrical hardware configured to block a portion of higher-order diffracted radiation from reaching the radiation sensor. 16. The system according to any one of the preceding clauses, wherein the electric hardware includes an alignment rotating optics motor. 17. The system according to any one of the preceding clauses, wherein the trained machine learning model is configured such that the system requires minimal use of the electrical hardware for blocking portions of the higher-order diffracted radiation from reaching the radiation sensor. 18. The system according to any one of the preceding clauses, wherein the system is configured such that the intensity is combined with the phase of the diffracted radiation to detect and correct sub-nanometer alignment shifts caused by asymmetry of the measurement target. 19. The system according to any one of the preceding clauses, wherein the measurement target includes a grating. 20. The system according to any one of the preceding clauses, wherein the intensity imbalance between the positive first-order diffraction radiation and the negative first-order diffraction radiation is configured to be used by the one or more processors to determine the alignment of the layer, and thereby adjust the semiconductor device manufacturing process based on the alignment. 21. A measurement method comprising: irradiating a measurement target in a layer of a patterned substrate with radiation from a radiation source, the measurement target being configured to diffract radiation from the radiation source; outputting phase data and intensity data of the diffracted radiation using a radiation sensor, the phase data including different energies associated with different diffraction orders of the diffracted radiation; and executing a trained machine learning model using one or more processors operatively connected to the radiation sensor and configured by machine-readable instructions, the trained machine learning model being configured to: determine, based on the phase data, a subset of the intensity data associated with positive first-order and negative first-order diffracted radiation incident on the radiation sensor; and determine, based on the subset of the intensity data, an intensity difference between the positive first-order diffracted radiation and the negative first-order diffracted radiation. 22. The method according to clause 21, wherein the intensity difference includes an intensity imbalance, and wherein the method further includes using the one or more processors to determine the alignment of the layer based on the intensity imbalance. 23. The method according to any one of the preceding clauses, wherein determining the subset of the intensity data includes excluding intensity data associated with destructive higher-order diffraction radiation, the data associated with the destructive higher-order diffraction radiation being identified based on the energy associated with the phase of the higher diffraction order. 24. The method according to any one of the preceding clauses, wherein the destructive higher-order diffraction radiation includes leakage. 25. The method according to any one of the preceding clauses, wherein the phase data conveys the alignment position of the layer relative to the average position from all wavelengths for all diffraction orders entering the pupil of the measurement system; and the relative diffraction efficiency corresponding to all diffraction orders entering the pupil. 26. The method according to any one of the preceding clauses, wherein the intensity data includes intensity imbalance with diffraction order mixing. 27. The method according to any one of the preceding clauses, wherein the trained machine learning model comprises a multinomial regression model, a decision tree regressor, a random forest regressor, a multilayer perceptron (MLP) regressor, a stochastic gradient descent (SGD) regressor, an extreme gradient boosting (XGB) regressor, or any combination thereof. 28. The method according to any one of the preceding clauses, wherein the trained machine learning model comprises a multinomial regression model. 29. The method according to any one of the preceding clauses, wherein the trained machine learning model is trained by: providing an untrained machine learning model with phase channel alignment positions relative to the average positions from all wavelengths for all diffraction orders entering the pupil of the measurement system; determining substrate quality based on the relative diffraction efficiency corresponding to all radiation diffraction orders entering the pupil, and providing the substrate quality to the untrained machine learning model; providing the untrained machine learning model with a fundamental truth intensity channel imbalance having order mixing, and providing the untrained machine learning model with a fundamental truth value only for first-order diffraction radiation. Intensity channel imbalance; providing a new intensity channel imbalance with order mixing to the untrained machine learning model, and using the untrained machine learning model, based on the new intensity channel imbalance with order mixing, determining a new intensity imbalance of the first diffraction order in the diffraction order entering the pupil; determining whether the new intensity imbalance determination result is accurate in response to the difference between the determination result and the underlying true intensity channel imbalance of the first-order diffraction radiation being less than a threshold amount; adjusting one or more parameters of the untrained machine learning model based on one or more accuracy determination results to form a trained machine learning model. 30. The method according to any one of the preceding clauses, wherein the threshold quantity is 1e-4. 31. The method according to any one of the preceding clauses, wherein the fundamental true intensity channel imbalance for first-order diffraction radiation provided to the untrained machine learning model is determined by simulation. 32. The method according to any one of the preceding clauses, wherein an alignment rotating optical device motor is used to determine the fundamental true intensity channel imbalance provided to the untrained machine learning model for first-order diffraction radiation only. 33. The method according to any one of the preceding clauses, wherein the alignment rotating optics motor is used to generate training data and is removed and / or not used once the training of the trained machine learning model is complete. 34. The method according to any one of the preceding clauses, wherein the trained machine learning model is further trained by measuring a plurality of orthogonal polarization states for each diffraction order and providing these polarization states, together with the polarization specific to the first-order diffraction radiation, to the untrained machine learning model as further training input. 35. The method according to any one of the preceding clauses further includes using electrically powered hardware to block a portion of the higher-order diffraction radiation from reaching the radiation sensor. 36. The method according to any one of the preceding clauses, wherein the electric hardware includes an alignment rotating optics motor. 37. The method according to any one of the preceding clauses, wherein the trained machine learning model is configured such that the system requires minimal use of the electrical hardware for blocking portions of the higher-order diffracted radiation from reaching the radiation sensor. 38. The method according to any one of the preceding clauses, wherein the intensity is combined with the phase of the diffracted radiation to detect and correct sub-nanometer alignment shifts caused by asymmetry of the measurement target. 39. The method according to any one of the preceding clauses, wherein the measurement target includes a grating. 40. The method according to any one of the preceding clauses, wherein the intensity imbalance between the positive first-order diffraction radiation and the negative first-order diffraction radiation is configured to be used by the one or more processors to determine the alignment of the layer, and further to adjust the semiconductor device manufacturing process based on the alignment. 41. A non-transitory computer-readable medium storing instructions that, when executed by a computer, cause the computer to perform operations including: determining, based on phase data, a subset of intensity data associated with positive and negative first-order diffracted radiation incident on a radiation sensor by executing a trained machine learning model, wherein a radiation source is configured to irradiate a measurement target in a layer of a patterned substrate with radiation, the measurement target being configured to diffract radiation from the radiation source; and a radiation sensor configured to output the phase data and intensity data of the diffracted radiation, the phase data including different energies associated with different diffraction orders of the diffracted radiation; and determining, based on the subset of intensity data, an intensity difference between the positive and negative first-order diffracted radiation by executing the trained machine learning model. 42. The medium according to clause 41, wherein the intensity difference includes an intensity imbalance, and wherein the operation further includes using the computer to determine the alignment of the layer based on the intensity imbalance. 43. The medium according to any one of the preceding clauses, wherein determining the subset of the intensity data includes excluding intensity data associated with destructive higher-order diffraction radiation, the data associated with the destructive higher-order diffraction radiation being identified based on the energy associated with the phase of the higher diffraction order. 44. The medium according to any one of the preceding clauses, wherein the destructive higher-order diffraction radiation includes leakage. 45. The medium according to any one of the preceding clauses, wherein the phase data conveys the alignment position of the layer relative to the average position from all wavelengths for all diffraction orders entering the pupil of the measurement system; and the relative diffraction efficiency corresponding to all diffraction orders entering the pupil. 46. ​​The medium according to any one of the preceding clauses, wherein the intensity data includes an intensity imbalance having a mixture of diffraction orders. 47. The medium according to any one of the preceding clauses, wherein the trained machine learning model comprises a multinomial regression model, a decision tree regressor, a random forest regressor, a multilayer perceptron (MLP) regressor, a stochastic gradient descent (SGD) regressor, an extreme gradient boosting (XGB) regressor, or any combination thereof. 48. The medium according to any one of the preceding clauses, wherein the trained machine learning model comprises a multinomial regression model. 49. The medium according to any one of the preceding clauses, wherein the trained machine learning model is trained by: providing an untrained machine learning model with phase channel alignment positions relative to the average positions from all wavelengths for all diffraction orders entering the pupil of the measurement system; determining substrate quality based on the relative diffraction efficiency corresponding to all radiation diffraction orders entering the pupil, and providing the substrate quality to the untrained machine learning model; providing the untrained machine learning model with a fundamental true intensity channel imbalance having order mixing, and providing the untrained machine learning model with a fundamental true intensity channel imbalance only for first-order diffraction radiation. Intensity channel imbalance; providing a new intensity channel imbalance with order mixing to the untrained machine learning model, and using the untrained machine learning model, based on the new intensity channel imbalance with order mixing, determining a new intensity imbalance of the first diffraction order in the diffraction order entering the pupil; determining whether the new intensity imbalance determination result is accurate in response to the difference between the determination result and the underlying true intensity channel imbalance of the first-order diffraction radiation being less than a threshold amount; and adjusting one or more parameters of the untrained machine learning model based on one or more accuracy determination results to form a trained machine learning model. 50. The medium according to any one of the preceding clauses, wherein the threshold amount is 1e-4. 51. The medium according to any one of the preceding clauses, wherein the fundamental true intensity channel imbalance for first-order diffraction radiation provided to the untrained machine learning model is determined by simulation. 52. The medium according to any one of the preceding clauses, wherein an alignment rotating optical device motor is used to determine the fundamental true intensity channel imbalance provided to the untrained machine learning model for first-order diffraction radiation only. 53. The medium according to any one of the preceding clauses, wherein the alignment rotating optics motor is used to generate training data and is removed and / or not used once the training of the trained machine learning model is complete. 54. The medium according to any one of the preceding clauses, wherein the trained machine learning model is further trained by measuring a plurality of orthogonal polarization states for each diffraction order and providing these polarization states, together with the polarization specific to the first-order diffraction radiation, to the untrained machine learning model as further training input. 55. The medium according to any one of the preceding clauses, the operation further includes controlling the electrical hardware to block a portion of the higher-order diffracted radiation from reaching the radiation sensor. 56. The medium according to any one of the preceding clauses, wherein the electric hardware includes an alignment rotating optical device motor. 57. The medium according to any one of the preceding clauses, wherein the trained machine learning model is configured such that the system requires minimal use of the electrical hardware for blocking portions of the higher-order diffracted radiation from reaching the radiation sensor. 58. The medium according to any one of the preceding clauses, wherein the intensity is used in combination with the phase of the diffracted radiation to detect and correct sub-nanometer alignment shifts caused by asymmetry of the measurement target. 59. The medium according to any one of the preceding clauses, wherein the measurement target includes a grating. 60. The medium according to any one of the preceding clauses, wherein the intensity imbalance between the positive first-order diffraction radiation and the negative first-order diffraction radiation is configured to be used by the computer to determine the alignment of the layer, and thereby adjust the semiconductor device manufacturing process based on the alignment.

[0125] The concepts disclosed herein can be associated with any general imaging system used for imaging subwavelength characteristics and are particularly applicable to emerging imaging techniques capable of generating increasingly shorter wavelengths. Emerging techniques already in use include EUV (Extreme Ultraviolet) and DUV lithography, with DUV lithography capable of generating 193 nm wavelengths using ArF lasers and even 157 nm wavelengths using fluorine lasers. Furthermore, EUV lithography can generate wavelengths in the 20–5 nm range by using synchrotrons or bombarding materials (solid-state or plasma) with high-energy electrons, thereby generating photons within this range.

[0126] While the concepts disclosed herein can be used for imaging on substrates such as silicon wafers, it should be understood that the disclosed concepts can also be used in any type of lithography imaging system, such as systems for imaging on substrates other than silicon wafers. Furthermore, combinations and sub-combinations of the disclosed elements may include individual embodiments.

[0127] The above description is intended to be illustrative only and not restrictive. Therefore, those skilled in the art will understand that modifications can be made in the described manner without departing from the scope of the claims set forth below.

Claims

1. A measurement system, comprising: A radiation source configured to irradiate a measurement target in a layer of a patterned substrate with radiation, the measurement target being configured to diffract radiation from the radiation source; A radiation sensor configured to output phase and intensity data of diffracted radiation, the phase data including different energies associated with different diffraction orders of the diffracted radiation; as well as One or more processors, operatively connected to the radiation sensor and configured to execute a trained machine learning model via machine-readable instructions, the trained machine learning model being configured to: Based on the phase data, a subset of the intensity data associated with the positive first-order diffraction radiation and the negative first-order diffraction radiation incident on the radiation sensor is determined; as well as The intensity difference between the positive first-order diffraction radiation and the negative first-order diffraction radiation is determined based on the subset of the intensity data.

2. The system of claim 1, wherein the intensity difference includes intensity imbalance, and wherein the one or more processors are further configured to determine the alignment of the layer based on the intensity imbalance.

3. The system of claim 1 or 2, wherein determining the subset of the intensity data comprises excluding intensity data associated with destructive higher-order diffraction radiation, wherein the data associated with the destructive higher-order diffraction radiation is identified based on the energy associated with the phase of the higher diffraction order.

4. The system of claim 3, wherein the destructive higher-order diffraction radiation includes leakage.

5. The system according to any one of claims 1-4, wherein the phase data conveys the alignment position of the layer relative to the average position from all wavelengths for all diffraction orders entering the pupil of the measurement system; and the relative diffraction efficiency corresponding to all diffraction orders entering the pupil.

6. The system according to any one of claims 1-5, wherein the intensity data includes an intensity imbalance having a mixture of diffraction orders.

7. The system according to any one of claims 1-6, wherein the trained machine learning model comprises a multinomial regression model, a decision tree regressor, a random forest regressor, a multilayer perceptron (MLP) regressor, a stochastic gradient descent (SGD) regressor, an extreme gradient boosting (XGB) regressor, or any combination thereof.

8. The system of claim 7, wherein the trained machine learning model comprises a multinomial regression model.

9. The system according to any one of claims 1-8, wherein the trained machine learning model is trained in the following manner: Provide the untrained machine learning model with the phase channel alignment position relative to the average position from all wavelengths for all diffraction orders of the pupil entering the measurement system; The substrate quality is determined based on the relative diffraction efficiency corresponding to all radiation diffraction orders entering the pupil, and the substrate quality is provided to the untrained machine learning model. Provide the untrained machine learning model with a base true intensity channel imbalance having order mixing, and provide the untrained machine learning model with a base true intensity channel imbalance only for first-order diffraction radiation; The untrained machine learning model is provided with a new intensity channel imbalance with order mixing, and the untrained machine learning model is used to determine a new intensity imbalance of the first diffraction order in the diffraction order entering the pupil based on the new intensity channel imbalance with order mixing. In response to the fact that the difference between the determined result and the fundamental true value of the intensity channel imbalance of the first-order diffraction radiation is less than a threshold amount, it is determined whether the new intensity imbalance determination result is accurate. as well as One or more parameters of the untrained machine learning model are adjusted based on one or more accuracy determination results to form a trained machine learning model.

10. The system of claim 9, wherein the threshold quantity is 1e-4.

11. The system of claim 9 or 10, wherein the fundamental true intensity channel imbalance for first-order diffraction radiation provided to the untrained machine learning model is determined by simulation.

12. The system of claim 9 or 10, wherein an alignment rotating optical device motor is used to determine the fundamental true intensity channel imbalance provided to the untrained machine learning model for first-order diffraction radiation only.

13. The system of claim 12, wherein the alignment rotating optics motor is added to the system to generate training data, and is removed and / or not used once the training of the machine learning model is complete.

14. The system according to any one of claims 9-13, wherein the trained machine learning model is further trained by measuring a plurality of orthogonal polarization states for each diffraction order and providing these polarization states, together with the polarization specific to the first-order diffraction radiation, to the untrained machine learning model as further training input.

15. The system according to any one of claims 1-14, further comprising electrical hardware configured to block a portion of higher-order diffraction radiation from reaching the radiation sensor.

Images